Full text of "Organic seminar abstracts"

See other formats

"L I B HAHY

OF THE
UN IVERSITY
Of ILLINOIS

Q, 54-7

libs

966 -67

pt.2

The person charging this material is re-
sponsible for its return on or before the
Latest Date stamped below.

Theft, mutilation, and underlining of books
are reasons for disciplinary action and may
result in dismissal from the University.

University of Illinois Library

jm-nrm

L161 — O-1096

Digitized by the Internet Archive

in 2012 with funding from

University of Illinois Urbana-Champaign

http://archive.org/details/organicsemi1966672univ

ORGANIC SEMINAR ABSTRACTS
1966-67

Semester II

Department of Chemistry and Chemical Engineering
University of Illinois

0 7

SEMIHAR TOPICS

II Semester 1966-67

Mechanisms in the Biological Chemistry of Pyrophosphate Esters

W. Tc Shier 221

The Configurational Stability of ALkenyl Radicals

Peter M» Harvey 230

Isomerization of Organic Thioeyanates to Isothiocyanates

Joseph Co Stickler 238

The Physiologically Active Constituents of Marihuana

Donald Co Schlegel 2V7

Benzene Photolysis

Warren J« Peascoe 255

The Photodimerization of Thymine

Sheldon A. Schaff er 263

Thermal Rearrangements of Cycloheptatrienes

W. Do Shermer 272

Cyclization Reactions of N-haloamines , -Amides , and -Imines

Daniel RB Bloch 28l

The Thermal Endo-Exo Isomerization of Some Diels-AIder Adducts

Tommy L. Chaffin 290

Reactions of NH Radicals

Terry Go Burlingame 296

Geometric Isomerism in Diazoketones

Daniel B» Pendergrass 305

Eo S. R. Studies of Organic Ground-State Triplet Molecules

Robert Jo Basalay 313

Recent Studies Concerning the Mechanism of the Favorskii Rearrangement

Peter Ao Gebauer 322

Prostaglandin Syntheses

Edward Bertram 331

Possible Yinyl Cation Intermediates

David Ao Simpson 338

Sigmatropic Reactions

Rs H. Watson 3^7

Photochemistry of Cyclobutanones and Cyclobutanediones

Edward F» Johnson 356

- 2 «•

The Mechanism of Papain Catalysis

Paul Elliot Bender 365

The Photosensitized Cis-Trans Isomerization of Olefins

Robert Kalish 373

The Abnormal Claisen Rearrangement

James E0 Shaw 382

-221-

MECHANISMS IN THE BIOLOGICAL CHEMISTRY OF PYROPHOSPHATE ESTERS
Reported by W. T. Shier February 23, 19&1

Introduction;

For decades physical organic chemists have studied the mechanisms of the
reactions they conducted in their flasks , often determining to a high degree
of certainty the detailed path of the atoms and electrons involved in these
conversions. In contrast the mechanisms of the reactions of many biologically
important functional groups have been largely ignored even to the present day.
One such group is the pyrophosphate group.

The pyrophosphate esters found in nature may be represented by the for-
mula

0 0
II II

r„0-P-0-P-0-R' ,

1 I
OH OH

where, depending on the stratagems of nature, R or R' may be as simple as a
proton or as complex as a nucleoside. The pivotal role of phosphorus compounds
in the chemistry of life was first recognised by Fritz Lipmann, and represents,
perhaps, one of his greatest contributions to the understanding of the chem-
istry of biological systems. As will be shown, pyrophosphate esters and the
pyrophosphate ion itself are endowed with unique biochemical properties.

These compounds have heretofore invariably been considered in various other
contexts, usually based on some non-functional portion of the molecule. When
considered as pyrophosphate esters, they represent a large and disjointed group
of natural products. A method of organizing and correlating this heterogeneous
array is offered by a rapidly emerging branch of Organic Chemistry, Bioorganic
Mechanisms -■ that hybrid of Physical Organic Chemistry and Enzymology1. On
the basis of mechanistic aspects of the biological function of these compounds
they have been correlated into three groups and their biological reactions or-
ganized by three general equations.

The Role of the Pyrophosphate Ion as a Leaving Group:

Th.e pyrophosphate ion functions as a leaving group in many biosynthetic
coupling reactions. Typical is the biosynthesis of nucleoside coenzymes2,
which can be represented schematically as:

nucleoside-O-P-P-P + R-O-P ^ nucleoside-0-P-P-OR + PP(*

The role of the pyrophosphate ion as a leaving group in a similar substitution
reaction has been studied by Lipmann.3 In the biosynthesis of adenosine-5 ' -
phosphosulfate (APS) in a yeast extract preparation, the thermodynamics of the

Abbreviations and structural symbols used in this work:

-OPP, pyrophosphate monoester; P<*, inorganic phosphate; OPP, inorganic pyro-
phosphate; Enz-SH, sulfhydryl-containing enzyme; WAD, nicotinamide adenine
dinucleotide; MDP, ( TPN) nicotinamide adenine dinucleotide phosphate; PRPP,
5-phospho-ribosyl-l-pyrophosphate; -OP, phosphate monoester; CDP, cytidine di-
phosphate; CMP, cytidine monophosphate; UDP, uridine diphosphate; CoA, coen-
zyme A; TPP, thiamine pyrophosphate; FAD, flavin adenine dinucleotide; ATP,
adenosine triphosphate; ADP, adenosine diphosphate; AMP, adenosine monophos-
phate; DCC, dicyclohexylcarbodiimide.

-cLdei-

equilibrium

ATP + S04= ATP-sulfurylase Aps + pp,

strongly favor the reverse reaction, When pyrophosphatase was added to a pur-
ified ATP-sulfurylase system, the equilibrium concentration of APS increased
from O.OluM./ml. to 0«23uM,/ml,3 Thus, the pyrophosphatases, which catalyse
the very exothermic hydrolysis of the pyrophosphate ion, provide an energy
coupled system with a sufficient over-all free energy drop to favor APS synthesis,

Pyrophosphate Esters as Coupling Intermediates:

Polyisoprenoids can be considered to be biosynthesized by means of the poly-
merization in defined modes of A^isopenteny] pyrophosphate monomer.4 The inter-
mediacy of these pyrophosphate esters in the biosynthesis of squalene has been
demonstrated by Lynen and his collaborators. 5>6 When 2i4C»mevalonate-5-phosphate
was added to a crude enzyme system from yeast, labeled farnesyl pyrophosphate
and geranyl pyrophosphate were isolated by paper chromatography and identified
by comparison with synthetic standards,5 When the system was inhibited with
iodoacetamide A-isopentenyl pyrophosphate accumulated.6 A sulfhydryl-containing
enzyme was isolated in crude form which converted A-isopentenyl pyrophosphate
into dimethylallyl pyrophosphate. The proven and postulated products of A3~
isopentenyl pyrophosphate polymerization are summarized in the following table:7

Degree of Polymerization:

n=l

-OPP

A-isopentenyl pyrophosphate

-OPP

n=2

~ DPP

Geranyl pyrophosphate

-^ Monoterpenoids

n=3

Farnesyl pyrophosphate

^ Sesquiterpenoids
Squalene

n=k

n=10

n=n

Geranyl-geranyl pyrophosphate

CH3
H(-CH2-C=CH-CH2)-ioOPP

CH3
H(-CH2-C=CH-CH2>

-a, Diterrenoids

C>> Complex Lipids 15
^iphytoene

-^. Ubiquinones

-^ Gutta percha

The driving force for the condensation of monomer units has been attributed
to the unique effectiveness of the pyrophosphate ion as a leaving group.8
It is contended here that this ability may result not so much from any chemical
property of the pyrophosphate ion, as from its rapid removal from the system
by highly active, widely distributed pyrophosphatases. The pyrophosphate con-
centration can reasonably have a profound effect on the pyrophosphate ion-elim-
inating mechanism in the enzyme -substrate complex, thus inhibiting the overall
reaction, Again, the pyrophosphatases may energy couple the hydrolysis of the

jj^jL-«jphutijJua.bc ICii uu caxuuj.i-uctx'LujLi bond f Oj.iuaLiui.u

These condensation reactions can be summarized by general equation I:

R-OPP + H-S: qrr-rt R-S:

+ HOPP

Pyrophosphatase

pp(-

In this equation it is tempting to consider H-S: a general nucleophile.
Since Yuan and Block observed a 63^ conversion of l-3H-A3-isopentenyl pyrophos-
phate to squalene in a yeast autolysate,9 the rate of the biological reaction
is greater than the rate of non-enzymatically assisted dissociation, which would
result in hydrolysis to isopentenol. Hence, a strict SnI mechanism is ruled
out, although an anchimerically assisted dissociation of the pyrophosphate
ester as the enzyme-substrate complex is note The absence of any isomeric
reaction product4 arising from allylic rearrangement in squalene biosynthesis
also argues against any SmI mechanism, although Ingraham has suggested that
the cationic species may be bound to the surface of the enzyme in a manner that
permits further attack only at the terminal carbon.10 It has been also sug-
gested8 that a cationic species is formed, but that its formation is concerted
with the formation of a carbon-carbon bond; that this is anything other than
an Sn2 mechanism is not made clear.

General equation I can be seen to describe the polymerization of A^iso-
pentenyl pyrophosphate by considering head-to-tail condensation as resulting
from nucleophilic attack by the v. electrons of the double bond of the monomer
either directly on the ester if ied carbon of the growing chain12, or on an enzyme-
growing-chain complex:

OPP

HOPP

The head-to-head reductive condensation of farnesyl pyrophosphate to pro-

duce squalene according to Cornforth's hypothetical scheme13 can be seen to
follow general equation I:

CH3 jOPP H
RCH2C=CH-CH2 ^ S-Enz

CH3
"> RCH2C=CHCH2- S-Enz

CH2-CH— CCH2R

£opp

CH-

CH*

H
I

-> RCH2C=CH^CH- S-Enz

Steven1 s
Rearrangement

CH-
I

-H©

CHpCH— CCHpR

PPO-CH2CH=CCH2R

CH3 C CH3

RCHpC=CHCH-S-Enz

CH2CH=CCH2R

l~\<-\ ),

CH2CH=CCH2R
I

CH3 w CH3

RCH2C=CHCH-?S~Enz

(t CH2CH=C-CH2R
H^H I

H2N0CO>S CH^

CH<

CH2CH=CCH2R

RCH^CH-CH^ + S-Enz

CH2CH=CCH2R
»
CH3

(R=Geranyl; B=proton acceptor, possibly part of the active site)

This mechanism satisfies a considerable body of compelling but not con-
clusive evidence obtained largely by Popjak and collaborators,14 The obser-
vations and conclusions are:

(1) The biosynthesis of squalene from farnesyl pyrophosphate by washed rat
liver microsomes was powerfully inhibited by p-chloromercuri-benzoate, N-ethyl-
maleimide and Cu ions but not by iodoacetamide. Hence, the active site of the
enzyme contains a functional S-atom, but not a free~SH, which would permit in-
hibition by iodoacetamide.14

(2) When squalene was biosynthesized in the same system from 5-2H2-mevalonate,
11 atoms of 2H, not the theoretically possible 12, were retained. Mass spec-
troscopic analysis of succinate derivatives obtained from the centre of the
chain by ozonolysis of the deutero squalene showed mostly trideutero iaolecules.
Hence, the labeling at the centre is -CHD-CD2-.14

(3) In the biosynthesis of squalene from farnesyl pyrophosphate no tritium
from H3HO was incorporated into squalene, while incorporation of up to 0.8
ug-atom of the tritium per ug»mole of squalene from labeled TPN3H was observed.14
Hence, the condensation is a reductive process involving TPNH.

In Cornforth's mechanism, in the steady state all the experimental obser-
vations are satisfied. It may further be noted that in each -0PP eliminating
step general equation I is followed. That is

Enz
I

S: + ROPP

•OPP

Enz
S-R

-H

S-Enz

Where R=farnesyl and R' = H or farnesyl.

Reactions of pyrophosphate esters other than the synthesis of carbon-carbon
bonds can be shown to follow general equation I. In the biosynthesis of a com-
plex phospholipid from Halobacterium cutirubrum an ether linkage is formed by
phytanyl pyrophosphate and an -OH of glycerol-1-phosphate.15

i.e. R-OPP + H-0-C- > R-0-C- + HOPP

Since the biosynthetic studies were carried out in a crude cell-free extract
with the above reaction monitored only by analysis of 32P incorporation into
the final product, no detailed mechanism for the formation of the ether linkage
has been put forward. It was observed, however, that the use of phytanyl pyro-
phosphate removed the requirement for ATP in the biosynthesis. Hence, the es-
terified pyrophosphate group supplies the energy for condensation. It can read-
ily be seen, also, that the reaction conforms to general equation I.

In the biosynthesis of nucleotides, PRPP condenses with glutamine, orotic
acid, purines, or pyrimidines as follows:

P-0-H2C 0

\ l /

POH

opp

OPP

OH OH

The problem of strict application of the terms of physical organic chemistry
arises- Despite the fact that the leaving group is a to a heteroatom16 a strict
SnI reaction is highly unlikely. According to the Michaelis-Menten hypothesis,17
in the biological reaction only PRPP is bound to the enzyme; the N-containing
species also binds to the enzyme surface where reaction occurs. An SH2 displace-
ment of OPP by the lone pair of the N-containing species, while it and PRPP
are both bound to the enzyme is the simplest mechanism, although not the only
one possible.

It is observed that inversion of configuration occurs establishing the
stereochemistry observed in nucleotides.18 This inversion of configuration
is consistent with an Sm2 mechanism, but not conclusive evidence for it.11

Group Transfer Intermediates:

The nucleoside pyrophosphate diesters represent a large and rapidly increas-
ing group of biosynthetic intermediates. While the basic forms - the nucleoside
diphosphate -sugars, -alcohols and -diglycerides - were known a decade ago, and
extensively reviewed then,19 studies demonstrating their universal involvement
in the biosynthesis of complex lipids, oligosaccharides, homo- and hetero-poly-
saccharides in all levels of life, from the cell walls of Neurospora crassa20
to the complex lipids of the human brain,21 have occupied the time of hundreds
of research workers in the intervening years. To introduce order into this
thriving jungle of biosynthetic pathways, the following general equations are
presented:

0 0
H II

R-0-P-0-P~OCH2

1 ' s 0- 5
OH OH

HO Y

+ R'OH

0 0
II B

R-O-ROR' + HO-P-OCH2

1 I
OH OH

III

0 0
II 'I
R-O-P-0-P-0CH2 B

" 0

OH OH

OH Y

+ R'OH

0 0
W II
-^R-O-R1 + HO-P-0-P-OCH

OH OH

Oh y

Y= -H or -OH; B= purine or pyrimidine

-d.cLO-

Neither of these type reactions has received direct mechanistic study.
Results obtained for analogous systems will be considered, and the results ob-
tained tentatively applied to the two general equations pending the appearance
of systematic studies of these systems.

The alcoholysis of anhydride phosphate linkages has not been studied in
biological systems,22 but Cohn showed that 180 labeled phosphate underwent
isotope exchange with the carboxylate oxygen atoms of 3-phospho-glycerate in
the following oxidative phosphorylation of ADP23

0 0*

'i * li *e

: *■ R-C-0 -P-0 + H + + DPMI

R-

II
-CH

* 11 ■

+ HO -P-0

0 H

X© 1

+ DPN

-CH( OH) CH2

0P03H2

R-

II
-C-i

* ll *Q

D -P-0 +
U
0 H

(ADP)-0H

I *
0 H

* *
0 0

s N R-C> 0 + (ADP>0-P-0

y0 OH

The reverse of the second reaction was interpreted as nucleophilic attack
by the carboxylate oxygens on the terminal P atom of ATP22 That is,

0
0 0 0 HO 0 0 0

/: w \\ \/, il II N e

-CI6 + H-0-P-0(-ADP) » -C-0-P-0(-ADP) N -C-O-P-OH + (ADPfO

\\ 1 < I ^ I

0 OH OH OH

Extending this finding to the alcoholysis of anhydride phosphate linkages,
nucleophilic attack by the oxygen of the alcohol on the appropriate phosphor-
us atom can be put forward as a working hypothesis for further research. One
experiment which suggests itself consists of feeding R180H to an active cell free
extract containing the appropriate nucleoside diphosphate ester; identification
of 180 in excess in the phosphate diester and not in the nucleoside monophosphate
would support the hypothesis.

In a transfer reaction analogous to general equation HI Douderoff, Barker
and Hassid explained the isotope exchange between KH232P04 and glucose -1-phosphate
in a sucrose phosphorylase preparation from Pseudomonas saccharophila in the
absence of fructose, the phosphate acceptor, in terms of a glucosyl-enzyme
complex.24

i.e. Glucose -I-OPO3H + Enzyme ^ Glucose -Enzyme + H2P04~
(Glucose-1-032P03H) (H232P04")

The analogous substrate-enzyme complex has been suggested for systems described
by general equation^H.25 The irreversibility of the reaction prevents the use
of 32P-labeled nucleoside pyrophosphate in an experiment analogous to that of
Douderoff et al.24

The interesting question of whether P-0 or C-0 cleavage occurs could be
answered by preparing the appropriate nucleoside triphosphate labeled with
180 in the phosphate( s) , adding it to a cell-free extract, and analysing R-0-R'
for excess 180; excess l80 would indicate P-0 cleavage.

The nucleoside pyrophosphates function as transfer intermediates in some reactions,

For example, the following kinase catalysed reaction2 fits general equation II:

2 nucleoside -OPP - — *» nucleoside-OP + nucleoside-OPPP

This reaction is general to both nucleoside and deoxynucleoside pyrophosphates.'

Phosphatides are biosynthesized from CDP-diglycerides26 according to the
scheme of general equation II:

CH2OCOR

CHOCOR,

CH20-P-0-P-QCH2 0>

0 0

R'-OH

CH20C0R
CHOC OR _
CHoO-P-OR' + CMP

where R = CH3(-CH2)-i4, for example, and R'OH = L-a glycerol, L-serine, myo-inositol
or phosphatidyl glycerol.

Other cytidine diphosphate alcohols also undergo enzymatic conversions26
according to the scheme of general equation II:

0
II
CDP-OR + R'OH rrr^ R-O-P-OR' + CMP

where ROH is ethanolamine or choline in an alternate synthesis of lecithins26
or a-glycerol or 5-ribitol in the biosynthesis of bacterial cell wall polymers.27
The nucleoside diphosphate-pentoses and -hexoses represent the largest group
of transfer intermediates1 - they are found for most of the known nucleosides
and deoxynucleo sides.19 The sugar moieties include hexoses, pentoses, glycur-
onic acids, hexosamines, mucopeptides and oligosaccharides2 attached to the
pyrophosphate moiety at the anomeric carbon. They undergo reaction according
to the scheme of general equation III. For example, the synthesis of glycogen
from UDP-glucose in rat liver:29

CH2OH

glucosyl
transferase

UDP)V

n UDP

"Non-Functional" Pyrophosphate Esters:

This group of pyrophosphate esters consists largely of the coenzymic f orm( s)
of the B vitamins.30 The term "non-functional" is somewhat of a misnomer, for
the pyrophosphate moiety serves as a linkage in some cases (e.g. CoA, NAD, FAD,
etc.) , it may serve as a binding site to attach the coenzyme to the enzyme
(e.g. TPP) , or it may become functional after conversions have been made on other
parts of the molecule (e.g. mevalonate-5-PP) .12 On any account, however, the
pyrophosphate group does not undergo a permanent change in the normal biological
function of the molecule.

In the conversion of the B vitamins to their coenzymic forms the vitamin
is usually incorporated intact:30

ooH

Vitamin Bx -» Thiamine Pyrophosphate

0 0

II 11
CH2CH2-OrP-0-P-OH

1 I
OH OH

Pantothenic Acid -» Coenzyme A
NH2

„ 0 OH CH3 i 0 0
■ II I i ill )\

HS(-CH2> NfC(-CH2)- N-C-CH-C-CH2Of-P-OPOH2 C

Riboflavin -> Flavin Adenine Dinucleotide

r---"o !

^_J

OH OH

_oj

i\i

Niacin -> Nicotinamide Adenine Dinucleo-

NH2 tlde

CH2 i

I I

HCOH ,

I i

HCOH ,

HCOH ' 0 Q

CH20l-P-0-P-0CH2 n
OH OH

ONH;

0 0^^

II 11

CH20-P-0-P-0CH2

1 I
OH OH

■H OR

R = H : NAD

R = -P03H2: NADP

BIBLIOGRAPHY

1. For a more complete definition cf. the Preface of T. C. Bruice and S. J.
Benkovic, "Bioorganic Mechanisms/' Vol. 1, W. A. Benjamin, Inc., New York
(1966).

2. A. Kornberg, Advances in Enzymol. , 18, 191 (1957).

3. P. W. Robbins and F. Lipmann, J. Am. Chem. Soc, z£, 6409 (1956).

k. J. R. Richards and J. B. Hendrickson, "The Biosynthesis of Steroids, Terpenes,
and Acetogenins," W. A. Benjamin, Inc., New York, 1964, p. 198.

5. F. Lynen, B. W. Agranoff , H. Eggerer, U. Henning and E. M. Moslein, Angew.
Chem., II, 657 (1959).

6. F. Lynen, H. Eggerer, U. Henning, and I. Kessel, Angew. Chem,, 70, 13$
(1958).

7. Modified from ref . k.

8. Ref. k, p. 201.

9. C. Yuan and K. Block, J. Biol. Chem., 2j&, 2605 (1959).

10. L. L. Ingrahan, "Biochemical Mechanisms," John Wiley and Sons, Inc., New
York, 1962, p. 96.

11. D. E. Applequist, Personal communication, Feb. k, 1967-

-229-

12. For a general review cf., G. Popjak and J. W. Cornforth, Adv. Enzymol. , 22,
281 (i960).

13. cited in G. Popjak, "Proceedings of the Fifth International Congress of
Biochemistry/' G. Popjak, ed. , Pergamon Press, New York, 1963> P* 207.

14. G. Popjak, D. Goodman, J. W. Cornforth, R. H. Cornforth, and R. Ryhage,
J. Biol. Chem. , 2J>6, 1934 (I96l).

15. M. Kates, Personal communication, Jan. 5 , 1967 •

16. E. S. Gould, "Mechanism and Structure in Organic Chemistry," Holt, Rinehart,
and Winston, New York, 1959, P« 284-5 and footnote 6l.

17. For a general discussion of the Michaelis-Menten hypothesis cf .,a. White, P.
Handler, E„ L. Smith, "Principles of Biochemistry," 3^d. ed. , McGraw-Hill
Book Co., New York, 1964, p. 221-8.

18. L. Warren in "Metabolic Pathways," Vol. 2, D. M. Greenberg, ed. , Academic
Press, New York, 1961, p. 4-59.

19. J. Baddiley and J. G. Buchanan, Quart. Revs., (London), 12, 152 (1958).

20. Ref. 17, p. 412.

21. Ref. 17, p. 468.

22. D. E. Koshland in "The Mechanism of Enzyme Action, "W. D. McElroy and B.
Glass, eds. , Johns Hopkins Press, Baltimore, 1954. p. 608.

23. M. Cohn, J. Biol. Chem., 201, 735 (1953).

24. M. Doudoroff, H. A. Barker, and W. Z. Hassid, J. Biol. Chem., 168, 725 (1947)

25. W. Z. Hassid in "Chemical Pathways of Metabolism," D. M. Greenberg, ed. ,
Vol. 1, Academic Press Inc., New York, 1954, p. 235.

26. E. P. Kennedy in ref. 13, p. 113.

27. J. Baddiley, J. G. Buchanan, and B. Carss, Biochim. Biophys. Acta, 27, 220
(1958).

28. a) E. Cabib, Ann. Rev. Biochem. , 32, 321 (1963) •
b) L. F. Leloir, Biochem. J., £1, 1 (1964).

29. L. F. Leloir and C. E. Cardini, J. Am. Chem. Soc, 79, 6340 (1957).

30. R. J. Williams, R. E. Eakin, E. Beerstecher, and W. Shive, "The Biochemistry
of the B Vitamins," Reinhold Publishing Corpn. , 1950, pp. 123-215.

•230-

THE CONFIGURATIONAL STABILITY OF ALKENYL RADICALS
Reported by Peter M. Harvey February 27 , 19&7

INTRODUCTION

There is evidence in the literature that aliphatic radicals generated at
optically active carbon atoms are unable to maintain their initial configurations,
and that they yield racemic products.1 In view of the known configurational stabil-
ities of alkenyllithiums2 relative to alkyllithiums,3 it is of interest to examine
whether alkenyl radicals are configurationally more stable than alkyl radicals.
This seminar will review the evidence for the nonlinearity of vinyl and substituted
vinyl radicals and will discuss chemical studies in which isomeric alkenyl radicals
are generated and captured.

If it can be established that the vinyl radical is nonlinear and does not main-
tain its configuration, the question of whether loss of configuration occurs by
rotation or by inversion is not a trivial one. Although no experiments have been
reported which distinguish between the two pathways, one might expect the isomeriza-
tion to proceed by inversion, with a simple migration of the a-substituent to an
alternate site of high electron density, rather than by rotation, in which the jt-
bond of the vinyl group must be broken. The question of inversion vs. rotation will
not be dealt with further in this review, and for convenience the term "inversion"
will be used to refer to the Interconversion of isomeric alkenyl radicals, regardless
of the mechanism.

ELECTRON SPIN RESONANCE STUDIES

The ESR spectrum of the vinyl radical should show eight lines if the three
hydrogen atoms are nonequivalent or six lines if the two g_-hydrogen atoms are
equivalent. Fessenden and Schuler irradiated liquid ethylene with 2.0 Ilev electrons
in the cryostatted cavity of an ESR spectrometer.4 Near 10U°K they observed super-
imposed on the twelve -line spectrum of the ethyl radical a doublet of doublets which
they attribute to the vinyl radical; the apparent hyperfine splittings of 102. kk and
13=39 gauss were not significantly temperature dependent. The formation of butane,
1-butene, and 1-hexene during the electron irradiation of ethylene at 105-l63°K has
been cited as evidence that ethyl and vinyl radicals are indeed formed,5 Further-
more^, vinyl radicals have been trapped with labelled methyl radicals generated from
iodomethane-C14 during the electron irradiation of liquid ethylene,6

The center of the nine -line pattern attributed to the trideuteriovinyl radical
from the irradiation of ethylene ~d4 was displaced 0.40 gauss from that of the
unlabelled species. Fessenden and Schuler attribute this shift upon isotopic sub-
stitution to a second -order effect which can be rationalized if the 102. W- gauss
splitting represents the sum of two approximately equal coupling constants. If the
a-proton of the vinyl radical is inverting with a frequency comparable to the
difference of two nonequivalent p-proton splittings, then the inversion effectively
interchanges the two [3-protons and causes a change in the spin state when the two
j3»protons are antiparallel. As a result, the inner four lines corresponding to
these spin states are broadened,, leaving a spectrum whose major splitting corresponds
to the sum of the two g-proton splittings. If this interpretation is correct, the
13»39 gauss splitting can be assigned to the a-proton.

The vinyl radical spectrum was also observed when liquid ethane containing 0.5$
acetylene was irradiated. No signal attributable to the 1,2-dideuteriovinyl radical
was observed in a similar experiment employing acetylene-d^J this result is consist-
ent with an inverting radical in which all the lines are broadened beyond detection.
Hence Fessenden and Schuler conclude that the vinyl radical is nonlinear and that
the a-proton is rapidly inverting. They feel that the hybridization on the a-
carbon atom is intermediate between sp2 and sp and probably lies close to the former.
By estimating the combined line widths of the broadened lines to be on the order of
fifty gauss, they place limits of 3xl0~8 and 3xl0~10 sec on the lifetimes of the
individual configurations. Assuming a classical model and a normal prcexponential
factor of about 1013 sec"1 (i.e., AS'^0) , this range of lifetimes corresponds to a
barrier to inversion of about 2 kcal/mole; If the lifetime of a single configuration

-231-
is limited by quantum-mechanical tunnelling,7 this figure represents a lower limit
on the estimated barrier height.

The electron-beam irradiation of a 2.5-mole-$ solution of allene in liquid
ethane at 101°K gave rise to an ESR signal which was interpreted as a superposition
of the spectra of the ethyl, 2-propenyl ( a-methylvinyl) , 3-propynyl, and allyl
radicals.4 The sixteen lines attributed to the 2-propenyl radical can be grouped
into four equally intense, overlapping Ii2i2i± quartets j this pattern is consistent
with a slowly ( on the ESR time scale) or noninverting nonlinear radical. The lower-
ing of the inversion rate when a methyl group is substituted for the a-proton of the
vinyl radical is consistent with a tunnelling mechanism for inversion. The hyperfine
splittings for the 2-propenyl radical are 19.48, 32.92, and 57*89 gauss^ the latter
two values are assigned to the two g-protons since their sum is similar to the sum
of the two g-proton hyperfine splittings in the vinyl radical,

Cochran, Adrian, and Bowers generated the vinyl radical at 4.2°K by ultraviolet
irradiation (2537 A0) of a solid argon matrix containing 1$ hydrogen iodide and 9$
acetylene.8 They observed a complex unsymmetrical spectrum but were able to assign
the eight-line gross pattern to three nonequivalent protons. Other workers observed
a similar spectrum in the solid phase at 77°K after atomic hydrogen generated in a
silent electrical discharge had been allowed to react with acetylene at 20°C in the
gas phase or at -196°C in the solid phase.9 The complexity of the spectrum pre-
sumably arises from the failure of anisotropic dipolar interactions to average to
zero through rapid thermal tumbling. The hyperfine splittings due to the three
protons are 15.7* 3^.2, and 68.5 gauss °910 these values are in good agreement with
the values given above for the a-proton and the sum of the g-proton hyperfine
splittings for the vinyl radical in the liquid phase. In the solid phase at 4.2°K
the inversion is apparently slowed or frozen so that two distinct g-proton couplings
are observed.

When a mixture of hydrogen iodide and acetylene -d2 was photolyzed under similar
conditions, the ESR spectrum consisted of two groups of lines separated by approx-
imately 64 gauss. The spectrum became weaker as the temperature was raised to 32°K,
but no new lines appeared. Addition of a hydrogen atom to the acetylene molecule
can give rise to either or both isomeric vinyl radicals I and II. The absence of any

H^ R^ ^D"

C=C ^=C.

I II

lines with splittings of 3^ or 16 gauss was rationalized8 to require that only one
species, the isomer (I or II) whose g-proton gives rise to the ok gauss splitting,
is present in detectable amounts. If inversion of the vinyl radical is a tunnelling
phenomenon,7 the rate of inversion should be relatively temperature independent,
since there is no classical activation energy requirement. An inversion process
that is rapid near 100°K in the liquid phase but slow enough to maintain configuration
at 4.2°K in the solid phase is not consistent with a tunnelling mechanism unless the
transition from liquid to solid at 4.2°K is accompanied by an additional stabilization
of the predominant radical relative to the transition state for inversion of at least
150 cal/mole.

A single predominant isomer might be the kinetically-determined product of
specific cis or trans addition of the hydrogen atom, or it might be the thermodynam-
ically more stable radical. Cochran, Adrian, and Bowers feel that neither of these
explanations is entirely satisfactory. They believe that the addition of a hydrogen
atom to a molecule of acetylene is exothermic and should give rise to a vibrationally
excited vinyl radical, which should invert rapidly before being cooled to 4.2°Kj
they do not consider the possibility that the rigid matrix may hinder the atomic
motion necessary for inversion, or that it may provide a lattice -relaxat ion mechanism
for rapidly deactivating an initially-formed excited cis or trans radical. If the
isomer ratio i/ll is thermodynamically controlled, the predominance of a single
isomer even at 32°K would require at least a 15$ difference in the total zero-point
vibrational energies of I and II in the absence of a significant difference in the
steric interactions of the matrix with the cis and trans radicals.8 The fact that

170

objections can be raised to both kinetic and thermodynamic control of the i/ll ratio
suggests that the original interpretation of the ESR spectrum as indicative of only-
one isomer may need reexamination.

Several attempts have been made to determine the geometry at the a-carbon atom
of the vinyl radical by comparing theoretically calculated spin densities for
different configurations with the observed hyperfine splitting constants. Using a
calculation based on the hyperconjugation theory of Coulson and Crawford,11 Dixon
calculated that the carbon-carbon-a-hydrogen bond angle 0 (see structure III) lies
between 120° and I5O0.12 Adrian and Karplus narrowed this range to I3O0 - 15O0 by

using a valence bond calculation and by assuming
that the 68 gauss splitting corresponds to the
trans -f3 -proton.13 However, the agreement of
calculated and experimental hyperfine splittings
for any value of 0 may be fortuitous in this
case) Strauss and Fraenkel have shown that the
same type of valence bond method gives a poor
fit between calculated and observed C13 hyper-
fine splittings.14 An extended Hilickel molecular
orbital treatment by Petersen gives the best fit between theory and experiment for
0 . i^0.15

RADICAL ADDITIONS TO SUBSTITUTED ACETYLENES

Alkenyl radicals have generally been assumed to be intermediates in free radical
additions to alkynes. Radical additions to mono- and disubstituted acetylenes have
been briefly discussed in several reviews concerned chiefly with radical additions
to olefins.16"18 The pertinence of individual stereochemical studies to the question
of the configurational stability of alkenyl radicals must be carefully considered,
since conclusive evidence for a hemolytic mechanism is often lacking. Even in
reactions for which a radical mechanism is established, free alkenyl radicals may
not be involved. In several cases, interpretation of the results is complicated by
incomplete product studies, ambiguous stereochemical assignments, and the absence of
proof that product ratios are kinetic ally controlled.

A large number of radical additions to substituted acetylenes, most of which
are initiated by peroxides or UV- irradiation, yield products corresponding to
preferential trans addition to the triple bond. Included in this group are the
liquid -phase addition of hydrogen bromide to propyne,19*20 the bromination of 2-
butynoic acid,21 the addition of trichlorosilane to mono- and dialkylacetylenes,22'23
the addition of alkyl- and arylthiols to ethoxyacetylene,24 propiolic acid,25'26 and
arylacetylenes,25*27*28 the reaction of thiolacetic acid with phenylacetylene27 and
1-hexyne,29 the addition of triphenyltin hydride to phenylacetylene,30 and the
addition of ditin and diarsine compounds to hexafluoro--2-butyne.31 On the other
hand, the brominations of several terminal and internal alkynes,32'35 the reaction
of hydrogen bromide with 1-bromoalkynes,36 and the addition of perfluoroalkyl
iodides to acetylenic compounds37*38 appear to proceed predominantly by cis addition
to the triple bond.

Skell and Allen have rationalized the exclusive trans UV-catalyzed hydro-
bromination of propyne in the liquid phase at -780 to -60u by 1) a configurationally
stable cis alkenyl radical, 2) a bridged bromine radical, or 5) bromine atom
addition to an initially -formed alkyne -hydrogen bromide complex, closely followed by
hydrogen atom transfer.19'20 Oswald, Griesbaum, Hudson, and Bregman have suggested
that rapid equilibration of cis and trans alkenyl radicals and stereoselective
hydrogen atom transfer to the less -hindered cis radical can account for the apparent
trans addition of alkyl- and arylthiols to phenylacetylene.27 Bergel'son has con-
cluded that in the bromination of mono- and disubstituted acetylenes, the reaction
stereochemistry is determined mostly by the relative thermodynamic stabilities of
the cis and trans radicals, which in turn depend on the steric repulsions between
the substituents on the radicals.32'35 In contrast to this result, Truce, Klein,
and Kruse have found that the steric requirements of the mesityl groups in the
transition state of the product -forming step in the addition of 2-thiomesitylene to
mesitylacetylene are not sufficiently large to reverse the normal trans stereochemistry
of thiol addition.28

-233-

RADICAL REACTIONS OF ALKENYLMETALLIC COMPOUNDS

Beletskaya, Karpov, and Reutov have claimed a radical mechanism for the reaction
of S-chlorovinylmercuric chloride with iodine in benzene and in carbon tetrachloride
to form l-iodo-2-chloroethylene„39 They have also studied the stereochemistry of
the j3-bromostyrenes (IV) produced by the reaction of bromine with cis and trans -2-
phenylethenylmercuric bromides (V) ,40 In methanol in the presence of added bromide

j#CH=CH-HgBr + Br2 - — > jfcH=CHBr + HgBr2

V IV

ion, the reaction proceeds with high (88.5%-cisj 93 .5% -trans ) retention of stereo-
chemistry at the double bond. In carbon tetrachloride the reaction with either
cis or trans V gives nearly equal amounts of cis and trans-(3-bromostyrenes , as
evidenced by the index of refraction of the distilled isomer mixture j the
equilibrium c is/trans ratio is 3>1» ^.f the reaction does indeed proceed by a
radical mechanism and if the product -forming step is the bromination of a B-styrenyl
radical, the similarity of the product ratios from cis and trans organomercurials
indicates that the p-styrenyl radical either is linear or, if nonlinear, is rapidly
inverting. In the latter case, the loss of original stereochemistry in the presence
of as good a radical trap as molecular bromine requires that the isomerization be
exceedingly rapid.

Gloc.kling has shown that the thermal decomposition of l-( 2-methylpropenyl)
silver(I) in ethanol at -20° probably proceeds by a radical chain mechanism. 41
Whitesides and Casey have studied the thermal decompositions of cis and trans-1-
propenyl-and 2~but~2-enylcopper( I) and silver( I) and the corresponding tri-n-
butylphosphine complexes in ether at ambient temperature j„42 The sole organic pro-
ducts are 2,^-hexadienes and 3>^-dimethyl-2,4-hexadienes, formed in high yields with
greater than 99% retention of configuration at the double bond. For example, eis-1-
propenyl(tri-n-butylphosphine) silver (I) (VI), prepared from 1-propenyllithium of 97%
cis stereochemistry, gave cis ,cls-2,Wnexadiene ( VII) in 95% yield and cis , trans-
2,4~hexadiene (VIII) in 4."5%~yield, corresponding to a total yield of cis propenyl
groups of 97% and a stereospecificity of 100%, based on the isomeric purity of the
starting organolithium.

K CH

Li^ /CH3 1' {fU3?P£U nBu3PA* CH,
\„_n^ ether, -78° .- J °\„__/ *

CT CH3
II I
V.C C. _ VII

;c=c' -^; f ' — -* jc=c.

IT ^H 2° dl0xane H^ "H CH3^ ^H

VI /C— C\ /CH<

H C=C^

VIII

If free alkenyl radicals are formed in these reactions, product formation
probably occurs by coupling of an alkenyl radical with a molecule of unde composed
organometallic , present in ca 0.1 M or lower concentration under the reaction con-
ditions. In order to account for the high stereospecificity of diene formation,
coupling must proceed at least 102 times faster than inversion of the radical:

k( coupling) x [alkenyl radical]x[alkenylmetallic] ^>

102 x k( inver si on) x[ alkenyl radical]

If the configurational lifetime of the alkenyl radical is approximately that of the
vinyl radical, on the order of 10~8 or 10~10 r.ef ;4 then k( coupling) must be greater
than 1011 l-mo.le"1sec~1. Whitesides and Casey argue that this minimum value for

.OX)i

k( coupling) is unreasonably large and conclude that the reaction does not proceed
with the formation of free alkenyl radicals. They also rule out geminate combination
of radicals within a solvent cage because they feel that the 100% efficiency of cage
combination necessary to explain the product yields is unlikely. They do not exclude
the possibility that alkenyl radicals are formed which are configurationally
stabilized by jt-complexation to the metal atoms.

GENERATION OF ALKENYL RADICALS BY PERESTER DECOMPOSITION

Bartlett has proposed a nonconcerted radical mechanism, in which the initial
step is cleavage of an oxygen-oxygen bond, for the thermal decomposition of tert-
butyl percinnamate.43 Kampmeier and Fantazier decomposed tert -butyl cis and trans -
a,P«dimethylpercinnamates (IX and X) in cumene at 110°. 44 The observed products are
consistent with the following reaction scheme:

■OH
?=C^ XI

CH3 CH3

XIII

c=c^

CH3 CH^

XIV

0 C-0-O-tBu

CH3 CH3

-£■» tBuO» + ^=C^

CH3 CH3

CH3COCH3

+
( CH3) 3COH

C02 + ,£=Q

CHq

3 CH3

— 1>C-CH3
CHo

C=C^ -A> tBuO + ^=C^
df3 g-0-O-tBu CH3 C-0*

K /CH3

co2 + c=c

^C=C^ XII
CH3 C-OH

A J^3

^C=C^ XV
CH3 H

Dicumyl is also produced in significant ( cis -43.4%; trans -65.5$) yields, cis and
trans -a ,g -Dimethylcinnamic acids (XI and XII) are obtained in low (1-2%) yields and
have the same stereochemistry as the parent perester. 3 A-Dimethylcoumarin (XIII)
is formed exclusively from the cis perester in 12.8% yield. If the acids and the
coumarin arise through the acyloxy radicals rather than by a minor heterolytic
pathway, these observations indicate that the cis and trans acyloxy radicals do not
inter convert.

In separate experiments, Kampmeier and Fantazier showed that trans perester
(X) recovered after partial decomposition contains no cis perester ( IX) , that cis
and trans -a ,8 -dimethylcinnamic acids added to decomposing trans perester are
recovered nearly quantitatively and unisomerized, and that cis and trans -2 -phenyl -
2-butenes (XIV and XV) are not isomerized under the reaction conditions. The sum of
the yields of cis and trans -2-phenyl-2-butenes accounts for most of the alkenyl
radicals formed, as measured by the evolution of carbon dioxide, so that hydrogen
abstraction is the major reaction of xhe alkenyl radicals. Tne cis and trans

peresters give mixtures of olefins XIV and XV with the same c is/ trans ratios, 1.1-
1.2. At 100° the equilibrium cis/trans ratio is approximately four."45 The common
kinetically-controlled isomer ratio again supports the intermediacy of a linear or
rapidly-inverting nonlinear radical.

If nonlinear alkenyl radicals are stereospecifically generated from the cis
and trans peresters, it might be possible to trap the radicals with an efficient
scavenger before they can isomerize. The first-order rate constant for decomposition
of the trans perester is not affected by the addition of an equimolar quantity of 2-
thiomesitylene, even though the yield of trans -a ,ft -d imethylcinnamic acid increases
from 1.9$ to 36. 3$. 46 This result requires separate product- and rate -determining
steps and suggests that Bartlett's one-bond cleavage mechanism43 is also operating in
this case. As the concentration of added 2=thiomesitylene is increased, the cis/
trans ratio of 2=phenyl-2-butenes from both cis and trans peresters increases toward
the equilibrium value. The failure of the thiol to trap a mixture of radicals
richer in the trans isomer from the trans perester does not rule out the stereo-
specific formation of nonlinear radicals, since isomerization of the olefinic pro-
ducts by the added thiol would obscure this observation. A precedent for such
olefin isomerization is found in the work of Oswald, Griesbaum, Hudson, and Bregman,27
who found that ethanethiol and thiophenol catalyze the isomerization of cis-ft-thio-
phenylstyrene and cis-ft-thioethylstyrene to the more stable trans isomers.

Singer and Kong have investigated the thermal decomposition of the tert -butyl
peresters of cis and trans -a-methylc innamic acids (XVI) and of cis and trans -q-
phenylcinnamic acids (XVII) in several solvents at 110°. 47>48 in a single solvent,

jZ5CH=C(CH3)C03tBu -£» j0CH=CHCH3

XVI XVIII

jZ$CH=C(j#)C03tBu -^> J0CH=CH0

XVII XIX

the cis and trans peresters give the same kinetically-controlled (nonequilibrium)
cis/trans ratios of 1-propenylbenzenes (XVIII) or stilbenes (XIX) °9 the cis/trans
olefin ratios increase as the solvent is changed from toluene to cyclohexene to
cumene. The authors interpret these results in terms of stereoselective capture of
rapidly-equilibrating cis and trans alkenyl radicals by the solvent. Cumene has
the largest steric requirement in the transition state for hydrogen atom transfer,
so that this solvent shows the greatest preference for hydrogen transfer trans to
the a-phenyl group and yields an olefin mixture containing the greatest proportion
of the cis isomer. Similar reasoning has been used to explain the variation of the
cis/trans decalin ratios observed when cis or trans-9-carbo-terjt-butylperoxydecalin
is decomposed in different solvents.49

GENERATION OF ALKENYL RADICALS BY THE HUNSDIECKER REACTION

The brominative decarboxylation of the silver salts of carboxylic acids is
regarded as a reliable method for generating alkyl radicals.50 Berman and Price
have reported that the bromination of silver cis and trans -cinnamates in refluxing
carbon tetrachloride gives trans -ft -br omostyrene (10$ and 17.5$, respectively), 1,1,2-
tribromo-2-phenylethane ( 25% and 35$), and the c innamic acid of unchanged stereo-
chemistry (12$ and 8$).51 Since at equilibrium g-bromostyrene contains an appreciable
amount of the cis isomer,40 the exclusive formation of the trans isomer from both
cis and trans silver salts indicates that the free energy of the transition state
for bromination of the cis radical is significantly (>3.2 kcal/mole if 1$ of cis -ft -
bromostyrene could have been detected) higher than that of the trans radical. If
the thermodynamic stabilities of the cis and trans radicals are comparable, this
explanation corresponds to a high stereoselectivity for bromine transfer to the
intermediate vinyl radical. Conclusions drawn from this study must be regarded as
tentative, since large percentages of starting silver salts were not accounted for.

The Hunsdiecker reactions of silver cis and trans -a-phenylcinnamates give
different cis/trans ratios of a-bromostilbenes.5i=; Studies of product stabilities
under the reaction conditions are not reported.

GENERATION OF ALKENYL RADICALS BY THERMAL DECOMPOSITION OF DIACYL PEROXIDES

Simamura, Tokumaru, and Yui decomposed cis and trans-die innamyl peroxides
carbon tetrachloride and in bromotrichloromethane,53 The stereochemistry of the g-
halostyrene products is given below;

isomer of c is/trans ratio of product

peroxide solvent jfcH=CHC.l jfciI=€HBr

cis CC14 18/82

BrCCl3 27/73

trans CCI4 19/81

BrCCl3 14/86

Separate experiments showed that the g_-halostyrenes are not isomerized under the
reaction conditions „

A common linear intermediate radical, is ruled out as the sole source of pri
ducts by the different cis /trans ratios from cis and trans peroxides in bromot]
chloromethane o The partial retention of configuration in bromotrichloromethane bui
not in carbon tetrachloride is consistent with a competition between inversion of
stereospecifically -generated alkenyl radicals and trapping of the radicals by
solvent °9 the smaller activation energy requirement for breaking a bromine -cart-
bond in the chain transfer step is reflected in the greater ease with which bromo-
trichloromethane adds to olefins o54 In carbon tetrachloride the equilibration of
the isomeric alkenyl radicals is complete j if the equilibrium constant for the
radical isomerization is near unity , the 1:4 c is/trans product isomer ratio suggests
that the phenyl group may sterically hinder halogen transfer to the cis radical,

BIBLIOGRAPHY

1. J„ Hine, "Physical Organic Chemistry," McGraw-Hill Book Co. , Inc., New York,
1962, pi 473 .

2. Do Yo Curtin and J. W. Crump, J. Am. Chem. Soc., 80, 1922 (1958).

3. R. L„ Letsinger, J. Am. Chem. Soc., 72, 4842 (1950}".

4. R. W. Fessenden and R. H. Schuler, J. Chem. Fnys., 39, 2147 ( I963) .

5. R. A. Holroyd and R. W„ Fessenden, J0 Phys. Chem., gf, 274 5 (1963).

6. R. A. Holroyd and G„ W. Klein, Intern. J. Appl. Radiation Isotopes, 15, 633
(1964).

7. Both Fessenden and Schuler4 and Cochran, Adrian, and Bowers8 suggest the
possibility of tunnelling. Experimental methods for detecting tunnelling of
protons have been reviewed by M. T„ Link, MIT Organic Seminar, I Semester 1965-
1966, p. I96.

8. E. L. Cochran, F. J. Adrian, and V. A. Bowers, J. Chem. Phys., 40, 213 (1964).
9o N. Ya„ Buben, R. V. Kolesnikova, N, L, Kuznetsova, and V. I. Trofimov, Bull.

Acad. Sci. USSR, Div. Chem. Sci., I99O (1964).

10. These values are revised values attributed to Adrian and coworkers and quoted
by Fessenden.4

11. C. A. Coulson and V. A. Crawford, J. Chem. Soc, 2052 (1953).

12. W. T. Dixon, Mol. Phys., 9, 201 (I965).

13. F. J. Adrian and M. Karplus, J. Chem. Phys., 4l, 56 (1964).

14. H. L. Strauss and G. K. Fraenkel, J. Chem. Fnys., 35., I738 (I96I).
15c H. Petersen, Ph.D. Thesis, University of Illinois, I967, pp. 82-92.
I60 B. A. Bohm and P. I. At ell, Chem. Rev., 62, 599 (1962).

17. C. Walling and E. S. Huyser, Org. Reactions, 13_, 91 ( 1963) .

18. F„ W. Stacey and J. F. Harris, Org. Reactions, 15, 150 (1963).

19. P. S. Skell and R. G. Allen, J. Am. Chem. Soc. ,80, 5997 (1958) .

20. P. S. Skell and R. G. Allen, J. Am. Chem. Soc, W, 1559 (1964).

21. A. Pinner, Ber. , 28, 1884 (1895).

077

22. R. A. Benkeser and R. A, Hickner, J. Am. Chem. Soc., 80, 5298 (1958).
2.3„ R„ A« Benkeser, M. L. Burrous, L. E. Nelson, and J. V. Swisher, J. Am. Chem.
Soc, 85, 4385 (1961).

24. H. J. Alkema and J. F. Arens, Ree. Trav. Chim. , 79, 1257 (i960).

25. Y. Liu and H. Wang, Hua Hsueh Hsueh Pao, 31, ^51("l965) > Chem. Abstr. , 64,
17391 (1966}.

26. L. N. Owen and M. J. S. Sultanbawa, J. Chem. Soc., 3109 (1949).

27. A. A. Oswald, K. Griesbaum, B. E. Hudson, and J. M. Bregman, J. Am. Chem.
Soc, 86, 2877 (1964).

28. W. E. Truce, H. G. Klein, and R. B. Kruse, J. Am. Chem. Soc, 83, 4636 (1965) .

29. J. A. Kampmeier and G. Chen, J. Am. Chem. Soc., 87, 2608 (1965)7

30. R. F. Fulton, Ph.D. Thesis, Purdue University, 196*0, cited by H. G. Kuivila,

in F. G. A. Stone and R. West, "Advances in Organometallic Chemistry," Academic
Press, New York, 1964, Vol. 1, p. 65.
31o W. R. Cullen, D. S. Dawson, and G. E. Styan, J. Organometal. Chem., 3, 4o6
(1965)o

32. L. D„ Bergel'son, Bull. Acad. Sci. USSR, Div. Chem. Sci., 995 (i960).

33. I. N. Nazarov and L. D. Bergel'son, J. Gen. Chem. USSR, 27, l6l4 (1957).

34. I. N. Nazarov and L. D. Bergel'son, Bull. Acad. Sci. USSR, Div. Chem. Sci.,
826 (i960).

35<> 1" N. Nazarov and L. D. Bergel'son, Bull. Acad. Sci. USSR, Div. Chem. Sci.,

834 (i960).
36. L. D„ Bergel'son, Bull. Acad. Sci. USSR, Div. Chem. Sci., 1145 (1960).
37o K. Leedham and R. N. Haszeldine, J. Chem. Soc, 1634 (1954).
38. N. 0. Brace, J. Org. Chem., 27, 449 (1962).
39- lo P. Beletskaya, V. I. Karpov, and 0. A. Reutov, "Proceedings of the Conference

on the Problems of the Application of Correlation Equations in Organic Chemistry,"

Tartu ^ 1962, p. 355 9 cited in reference 40.

40. I. P. Beletskaya, V. I. Karpov, and 0. A. Reutov, Bull. Acad. Sci. USSR, Div.
Chem. Sci., 1615 (1964).

41. F. Glockling, J. Chem. Soc, 716 (1955) i 3^40 (1956).

42. G. M. Whitesides and C. P. Casey, J. Am. Chem. Soc, 88, 454l (1966).

43. P. D. Bartlett and R. R. Hiatt, J. Am. Chem. Soc, 80, 1398 (I958).

44. J. A. Kampmeier and R. M. Fantazier, J. Am. Chem. Soc, 88, 1959 (1966).

45. D. J. Cram and M. R. V. Sahyun, J. Am. Chem. Soc, 8j?, 1257 (I963).

46. R. M. Fantazier and J. A. Kampmeier, J. Am. Chem. Soc., 88, 5219 (I966).

47. L. A. Singer and N. P. Kong, Tetrahedron Letters, 2089 (1966).

48. L. A. Singer and N. P. Kong, J. Am. Chem. Soc., 88, 5213 (1966).

49. P. D. Bartlett, R. E. Pincock, J. H. Rolston, W. G. Schindel, and L. A. Singer,
J. Am. Chem. Soc, 87, 2590 (1965).

50. D. E. Applequist and A. H. Peterson, J. Am. Chem. Soc, 82, 2372 (i960).
51o J. D. Berman and C. C. Price, J. Org. Chem., 23, 102 (I95H).

52. E0 L. Sukman, Dissertation Abstr., 1£, 1211 (1958).

53. 0. Simamura, K. Tokumaru, and H. Yui, Tetrahedron Letters, 5l4l (1966).

54. C. Walling, "Free Radicals in Solution," John Wiley and Sons, Inc., New York,
1957, Po 250.

-238-

ISOMERIZATION OF ORGANIC THIOCYMATES TO ISOTHIOCYANATES

Reported by Joseph C. Stickler March 2, I.967

INTRODUCTION

The observation that organic thiocyanates isomerize to isothiocyanates
(equation l) was made about a century ago.1'2 However, not until this decade

R-I-C5NI ^ R-N=C=^ (1)

have serious investigations been initiated to obtain a mechanistic description
of these isomerizations. Experimental evidence in these investigations has in-
dicated that they may proceed by at least four pathways depending upon the struc-
ture of the organic substrate and reaction conditions. If the organic moiety
contains an allylic double bond, the reaction may proceed by a shift in that
bond and a change in the site of anion attachment.3 6 Also, conditions can
be controlled so that either a direct displacement7 or ionization mechanism8
is responsible for the isomerization of saturated systems. The isomerization
of thiocyanates upon irradiation is believed to proceed by a radical pathway.9
An interesting outgrowth of these isomerization studies was the information they
yielded on the ambident reactivity of the thiocyanate anion.7

ALLYLIC SYSTEMS

The isomerization of allylic thiocyanates to isothiocyanates has been known
since I875.1"'2 The fact that the site of attachment to the organic substrate
changed in allylic isomerization was substantiated when it was demonstrated that
crotyl thiocyanate upon heating to l40° yielded quantitatively a-methylallyl
isothiocyanate.4 The proposed mechanism for this rearrangement contains a six-
membered-ring transition state, which may be represented by contributing struc-
tures I through IV or alternatively by V. This is a characterization similar
to that attributed to the rearrangement of a,a-dimethylallyl chloride by Young,
Winstein and Goering.10

S— c =N S= C-=JJ ( SCN.)

I II III

c c cr -c

(SCN)6 L .A

Sr^1

This mechanism is supported by several experimental considerations. The
entropy of activation for the allyl and 2-methylallyl thiocyanate rearrangement
are -3*k + 1 and -8.7 + 1 cal./deg.-mole6 respectively which is indicative of
a high degree of order in the transition state relative to the ground state.
These values are the same sign and order of magnitude as several reactions be-
lieved to proceed through a six-membered-ring transition state such as the
Claisen rearrangement of allyl vinyl ether.11 These rearrangements obey first-order

-239-
kineticso The rate of conversion in the allylic case is faster than in the sat-
urated case. For instance, allyl thiocyanate6 in toluene at 86.4° isomerizes
at a rate of 3.81 X 10~4 sec"1 as compared to I.87 X 10~6 sec"1 for benzhydryl
thiocyanate at 90° 0° in benzene,8 The smallness of charge separation in the
transition state is indicated by the fact that neither changes in solvents towards
greater polarity nor added salts increased the rate of reaction significantly.
The changes in rate appear especially small when compared to the data compiled
by Streitwieser12 for the solvolysis of other allylic systems believed to occur
by way of an ionization route. For instance, for the solvolysis of Y/Y-dimethylallyl

chloride, the ratio k.rrar ,, . ..

5 Op ethanol ethanol

the corresponding thiocyanate , 5 k

= 3300, but for the isomerization of

3« Thus the con-

acetonitrile cyclohexane
tributions of structures III and IV are believed to be small compared to I and
II, with V perhaps being the best representation.

Although the isomerization of a pure isothiocyanate to thiocyanate has not
been clearly established, it is known that an equilibrium state is obtained in
the isomerization of allylic thiocyanates.5'*13'14 From the results of these
equilibria in Table I, several points of interest have been made.

TABLE I5
Equilibrium thiocyanate/isothiocyanate at
100° in different media.

% of thiocyanate

! at

equilibrium

Compound

Pure

cyclo-
hexane

aceto-

nitrile

Cone,
m/l

Benzhydryl

2-3

0.1-0.75

allyl

9-11

10~3-10_1

Y-methylallyl

10" 3

y /y-dimethylallyl

10"3

Analysis by U. V. and I.R. spectrophotometry.

As the polarity of the medium is increased, the percent of thiocyanate increas-
es, which is consistent with the fact that the thiocyanate has a higher dipole
moment and therefore would be more stable in a polar solvent. If the rearrange-
ment causes the organic substrate to lose stability by loss of hyperconjugative
stabilization of a y -methyl group, then there is a greater amount of the thio-
cyanate in equilibrium. Contrasting with this result is the fact that the more
methyl substitution at the gamma position, the faster the rate of conversion
to the isothiocyanate as seen from Table II. Apparently the greater the ability
of the organic substrate to stabilize a positive charge the more contribution
the ionic structures III and IV make to the transition state. In what can appar-
ently be considered an extreme case of the equilibrium favoring the thiocyanate,
Y-phenylallylthiocyanate does not rearrange at all with an allylic shift to
a-phenylallyl isothiocyanate which would no longer have the allylic double bond
conjugated to the phenyl group, but instead isomerizes to "y-phenylallyl iso-
thiocyanate.15 The mechanism believed to be responsible is the same as that
involved in the isomerization of saturated thiocyanates and will be discussed in
the next section.

-2 Ro-
table ii5

First-order rate coefficients at 60° for the
isomeric rearrangement of allylic thiocyanates;
RR'C^HCHaSCN ) RR'C(NCS)C=CH2

Compound

105 k/w)/ "i\
60u(sec x)

W6O0

R R>

Ab Cb

H H

CH3C H
CH3 CH3

1.8 3.3
2? 31
270d 96d

0.5
0.9

Determined by U. V. spectrophotometry.

A and C stand for acetonitrile and cyclohexane solvents
respectively.

trans

Extrapolated from lower temperature data.

SATURATED SYSTEMS

Many saturated thiocyanates are known to isomerize to the corresponding
isothiocyanates. The mechanism believed to apply to most of these systems is
a unimolecular ionization to an ion pair, which can return to starting material
or go to the isothiocyanate. Before considering the details of the above mech-
anistic description, the experimental evidence which excludes other pathways
of reasonable a priori probability will be presented.

The likelihood of a concerted pathway in which there would be some bond
making of the nitrogen to the a-carbon before complete bond breaking of the sul-
fur to a-carbon bond seems small. This route would involve a four -member ed-
ring transition state as shown in equation (2) and would be analogous to the
Chapman rearrangement of imino esters.16 Such a transition state would probably

R-S-C=N

-> R-N=C=S

(2)

have a higher energy of activation than the allylic isomerization; however, the
energies of activation for saturated and allylic systems are quite similar.
Also, one might expect retention of optical purity, but extensive racemization
is found when an optically active substrate is employed.17 The argument that
the geometry of the C-S-C=N group is prohibitive to the concerted mechanism has
been made.8

A possible second route is represented by equation (3). This route is
discredited by the fact that the kinetics are first order.

-2kl~
R

2RSCN > M — )- RNCS + RSCN ( 3)

RSC SCN9

Again, one would expect retention of optical purity in the product, which is
not found.

A third pathway might he an a-elimination to a carhene and thiocyanic acid,
and then recombination. However, attempts to identify the acid by spectroscopy
and to trap a carbene failed.6

A fourth possible mechanism would be homolytic cleavage followed by simple
recombination (equation k) which would yield the necessary first-order kinetics,
but the more likely chain process would result in a more complicated expression.

R-SCN y R. + *SCN ^ R-NCS (k)

Besides the solvent and salt effect which are more indicative of an ionization
mechanism, no side products were isolated which would be expected to accompany
a radical pathway. However, it appears that conditions can be controlled so as
to lead to an isomerization of benzyl and benzhydryl thiocyanates to the iso-
thiocyanates by a radical pathway. Mazzucato9 recently reported that when ben-
zyl thiocyanate is irradiated with a low pressure mercury arc, using the 2537 &
line, at a concentration of about 10 % in n-hexane, at room temperature and
with the exclusion of oxygen from the system, a photoequilibrium with the iso-
thiocyanate was established. The photostationary equilibrium mixture was about
70% thiocyanate and 3C$> isothiocyanate. On the basis of detection of the fluor-
escence emission of the benzyl radical in the ^60-530niM- region starting from
both the thio- and isothiocyanate in low temperature photolyses, a radical mech-
anism was suggested.

Having disposed of the above pathways, evidence supporting the ionization
mechanism will be presented. Perhaps the most persuasive evidence is the com-
parison of this system to the many similar systems studied by Winstein,18 Smith,19
Goering,20 Darwish21 and others. The proposed mechanism is an ionization to form
an ion pair intermediate through which isomerization occurs. The evidence
overwhelmingly points toward this pathway and the purpose here will be to see
how well a detailed picture of this mechanism is fulfilled. As a point of ref-
erence, equations (5) and (6) representing the detailed process of ionization-
dissociation described by Winstein and coworkers22 will be employed. Several

ionization dissociation
I ) II ) III VlV (5)

external
intimate or

or solvent-

internal separated dissociated

ion pair ion pair ions

li + - ^ +

RX -^ R X ^ R

k , k

x" , R + X (6)

-1 "-2 "-3

types of organic substrates have been studied, but because of its suitability for
extensive studies attention was mainly focused on the benzhydryl substrate.

olio

~<—-TC— -

As previously stated, the reaction obeys first-order kinetics. The energy
and entropy of activation for the isomer ization of 4,4' -dimethylbenzhydryl thio-
cyanate in acetonitriie are 20.67 kcal./mole and -10.0 cal./mole-deg. , respectively.8

In a study of solvent effects on the rate of isomer: ization , Fava and co-
workers demonstrated that an increase in polarity accelerated the rate, which
is compatible with the ionization mechanism. For example, at 90° for the benz-
hydryl case, the rates in methyl ethyl ketone, acetonitriie and dimethylform-
amide relative to benzene were 10, 150 and 280. This solvent sensitivity is
similar in order of magnitude to that of other reactions believed to occur by
a unimolecular ionization. For instance, the ratio of rates in acetonitriie and
benzene are 120 for the rearrangement of camphene hydrochloride and 2b for the
rearrangement of 9-decalyl perbenzoate.8

The effect of structure on reactivity revealed that with increasing stab-
ility of the carbonium ion formed by ionization, the rate of isomerization also
increased. A qualitative study of the rates of the series n-butyl, sec -butyl,
and t-butyl thiocyanates indicated that rate increased in going toward the more
highly branched substrate.6 In a study of para- substituted benzhydryl thiocy-
anates, Fava and coworkers demonstrated that electron releasing substituents
facilitated the rate, while electron withdrawing groups hindered the rate.
Using Brown's u values23 a Hammett plot of log k/ko vs.cr resulted in a linear
relationship with p = -3.40 in acetonitriie at 70°. This result compares well
with p = -^.05 for the solvolysis of benzhydryl chloride in ethanoL.8 Hence
structurax modifications of the organic moiety which aid the stabilization of
a positive charge increases the rate of isomerization.

Salts added to the reaction medium also increased the rate of isomerization,
thus supporting the ionization mechanism. Fava and coworkers8 investigated
in particular the effects of two salts, sodium perchlorate and sodium thiocyanate.
The investigation was carried out in two solvents, methyl ethyl ketone and acet-
onitriie } the effect being greater in the less polar methyl ethyl ketone. The
results fit well the Winstein equation24( 7) for "normal" salt

k = ko (1 + b [salt])

(7)

effects. The b values are summarized in Table III. The possibility that

TABLE III

Salt effects on isomerization of benzhydryl thiocyanate at 90°

in different media.

b values from Winstein equation (7)

Salt

Methyl ethyl
ketone

Acetonitriie

bMeEtC0

bMeCN

NaSCN
NaC104

16.7
11.7

4.27
2.93

3.92
3.99

a concurrent direct displacement by the nitrogen end of the ionic thiocyanate
might account for its greater effect as compared with the sodium perchlorate
was not completely ruled out. However, the ratio of the b values in the two
solvents was independent of the salt, which is expected if the effect is a
specific salt effect. Also, the degree of specificity of the different salts
is not uncommon.8

Although the solvent, structural, and the salt effects described above lead
strongly to the conclusion that an ionization pathway is operative, they do not
give a detailed picture of the mechanism. Fava and coworkers further particularized

the mechanism "by performing exchange experiments with sulfur labeled sodium
thiocyanate.8'25 The main portion of these exchange experiments was carried out
on the 4,4'-dimethylbenzhydryl thiocyanate where it was shown that direct dis-
placement by either the sulfur or nitrogen end of the ionic thiocyanate on the
organic substrate is negligible (equations 8 and 9) > since the kinetics indicated

NCS* + RSCN

*SCN + RSCN

,NCS*R + SCN

*SCER + SCN

(8)
(9)

essentially unimolecular exchange with ionic thiocyanate. Therefore all radio-
activity must enter the organic substrate by way of the ionization route. The
first conclusion drawn from these experiments was that since the amount of label,
on the ionic thiocyanate always remained far greater than that on the organic
thiocyanate or isothiocyanate, the organic thiocyanate could not have become
completely dissociated upon ionization or the labeled ionic thiocyanate would
have become equilibrated with the organic thiocyanate. Thus it was proposed
that the isomerization occurred through an ion pair.

The salt effect on both the rate of isomerization and exchange of kfhl
-dimethylbenzhydryl thiocyanate in acetonitrile at 25° were examined. The
results are shown in Table IV. As can be seen from Table IV the rate of exchange

TABLE IV

Salt effects on rates of exchange and isomerization of

h, h '-dimethylbenzhydryl thiocyanate in acetonitrile

b values from Winstein equation (7)

Isomerization

Exchange

NaC104
NaSCN

.38

a 25.0 + 0.1°
b 0o2 + 0.1°

is more greatly effected by added salts than the isomerization rate. The author
interpreted this to mean that a more advanced degree of ionization (i.e. greater
charge separation) was involved in exchange than in isomerization. Further
indication that exchange occurs at a more advanced stage of ionization is the
p value of -4.5 for exchange at 70° in acetonitrile.26 This is more negative
than the p = -J.kQ measured under the same conditions for isomerization indicating
a more highly polar transition state. Whether exchange might occur at the ex-
ternal ion pair or upon dissociation or both cannot be determined from this
data. However, the authors have speculated that for the highly polar acetonitrile
and the benzhydryl moiety, the dissociated carbonium ion seems the more probable
intermediate for the first -order exchange process. Assuming that isomerization
occurs at the internal ion pair stage, and that if in some instances the ionization
proceeds further and exchange takes place with labeled ionic thiocyanate, then
these exchanged species will proceed back to the internal ion pair and then
partition between thiocyanate and isothiocyanate, more information may be ob-
tained. The ratio of incorporated label in thiocyanate and isothiocyanate was
made and it was determined that the ion pair returned to covalency with the sulfur
end five times more often than the nitrogen end. With this information a lower

-2kk-

limit may be set on the rate of ionization by the expression (10) where kj.
is the rate of ionization and k. the isomerization rate. The determination

1 s

ki = k_, + k

6k,

(10)

of the rate of ionization in this manner assumes that the ion-pair stage at which
isomerization occurs is not preceded by another ion pair which returns exclusive-
ly to the thiocyanate. The stereochemical evidence indicated that racemization
of p-chlorobenzhydryl thiocyanate and isomerization takes place at the same
intermediate, thus supporting the assumption that no ion pair precedes the ion
pair stage responsible for isomerization. 17>2G This method is similar to that
employed by Goering and coworkers in their study of the jD-nitrobenzoate-carbonyl
-018 system*20 the main exception being that partition between the isotopic
oxygens is not dependent upon reaction conditions as is the ratio k /k...

From the ratio of the rate of exchange to the rate of ionization (equation
11) , it is possible to calculate an upper limit to the percent of dissociation
that occurs. Using 4,4' -dimethylbenzhydryl thiocyanate in acetonitrile at 0°
with 0.01M NaSCN,

(11)

the upper limit is calculated to be 5°3/°»The authors make the following sum-
mary to the result: Out of 100 internal ion pairs formed, about 5 undergo furth-
er ionization and 95 return to the covalent state. From the partition data mentioned
above it was concluded that of the 95 returning to covalency, about 79 return
to thiocyanate and 16 to the isothiocyanate. Also concluded was that under
conditions favoring ionization to a smaller extent (less polar solvent or less
stable carbonium ion) , return to covalency from the internal ion pair predom-
inates even more completely.

An outgrowth of these exchange experiments employed to particularize the
isomerization mechanism was an extensive mechanistic study on the isotopic
exchange between substituted benzhydryl thiocyanates and ionic thiocyanates.
Fava and coworkers26 >srT concluded that with strongly electron attracting sub-
stituents a bimolecular pathway prevailed and with strongly electron donating
substituents a unimolecular process was obtained. However , with substituents
having intermediate electron-donating' abilities , concurrent bimolecular and
unimolecular mechanisms were operative.

The isomerization of cyclopropylcarbinyl thiocyanate (VI) is of special
interest, since this is the only example known in which the saturated organic
moiety undergoes a rearrangement of the carbon skeleton during isomerization.
Spurlock and Newallis28 have reported that the isomerization of cyclopropyl-
carbinyl thiocyanate at 155°C in acetonitrile yielded compounds VII through XI,
the respective percent yield indicated below each structure. The kinetics were

SCN

'SCN

VII (12)

sew

VIII (6)

ix (75)

KCS

x (5)

'NCS

XI (2)

^NCS

-2li-5-
first order and the rate of isomerization at 15^.7° was 2,55 X 10 5 sec -1 in
dimethylf ormamide . The authors have reported that control experiments indicated
each of the products VII -XI was stable to the reaction conditions. In the pres-
ence of added potassium thiocyanate in dimethylformamide solutions, the isomer-
ization rate was increased, "but the products and their distribution were un-
affected. Since direct displacement by the ionic thiocyanate would favor the
cyclopropylcarbinyl isothiocyanate, the authors concluded that the acceleration
of the rate could be attributed mainly to "normal" salt effects.

The isomerization of methyl thiocyanate is somewhat unexpected since exper-
iments have indicated that isomerization of n-butyl thiocyanate is negligible.
However, at 131° , the boiling point of methyl thiocyanate, it isomerized and was
aided by dissolved salts.29 Although the isomerization proceeds in the neat methyl
thiocyanate, it is prevented in non-polar solvents.6 The kinetics, which were
measured in the pure state, were found to fit a first-order rate law; and the rate
of isomerization at 120° was 1.02 X 10 7 sec 1. From the above rate data, an
energy and entropy of activation were calculated to be kl kcal./mole and 31
cal./mole-deg. , respectively.30 The concerted four-membered ring,30 bimolec-
ular6 and ionization10 mechanisms have all been suggested on the basis of this
sparse experimental data. A more systematic and comprehensive study of this
system seems to be necessary for classification of the mechanism involved.

AMBIDENT NATURE OF THE THIOCYANATE ANION

One of the interesting aspects of the isomerization studies is the infor-
mation they yield about the ambident reactivity of the thiocyanate anion. So
far we have seen that there are three general mechanisms by which these isomer-
izations proceed; by allylic rearrangement, ionization and radical pathways.
In order to provide further scope to the study of this ambident reactivity, Fava
and coworkers7 were able to control the conditions of the isomerization so as
to provide the isomerization with another quite general pathway. By using the
benzyl substrate which 'undergoes direct substitution easily and maintaining a
substantial amount of ionic thiocyanate in the reaction media, the isomerization
was found to proceed mainly by an S .2 mechanism. Radioactive thiocyanate was
employed and the rates of exchange, which proceeded either by a direct displace-
ment of the S or N end of ionic thiocyanate as well as the rate of isomerization,
direct displacement by the N end, were measured. Since the rate of isomerization
was on the order of 10 6 and the isotopic exchange rate on the order of 10"4, the
rate of exchange is approximately equal to the reactivity of the sulfur end.
The rate of isomerization is equal to the reactivity of the nitrogen end of the
thiocyanate anion. Therefore the ratio of k /k„ equals the reactivity ratio
kq/k . In acetonitrile and methyl ethyl ketone at temperatures ranging from
70-100°, the values of k /k were on the order of 102 to 103 for the benzyl sub-
strate. For instance, at 70° the ratios of k /k are 1000 and 725 for the methyl
ethyl ketone and acetonitrile respectively; while values of 65O and 460 are ob-
tained at 100° for the same two solvents. On the other hand, in reactions stud-
ied by Cannell and Taft31 in which carbonium ions were generated in aqueous sol-
utions independently from the thiocyanate ions, k /k values ranging from 2 to 9
were determined from product ratios. In Fava's work concerned with the iso-
merization by the ionization mechanism, a value for k /k of 5 was obtained for
the isomerization of 4,4' -dimethylbenzhydryl thiocyanate in acetonitrile, which
is on the same order of magnitude as values obtained from the Cannell and Taft work.

As indicated from the above results, this system provides a good method for
determining how various factors affect the relative reactivities of the two ends
of the thiocyanate anion. Further investigation into the solvent effects on this
ambident anion might prove useful in ascertaining the role solvation plays in
determining relative reactivities. Also a systematic study of the thiocyanate
radical, in cases where the radical pathway is operative, could reveal more about
the nature of the relative reactivity of the two ends of the thiocyanate radical.
An extensive review related to nucleophilic ambifunctional reactivity by Gompper32
has been published.

BIBLIOGRAPHY

1. G. Gerlick, Ann., 17_8, 80 (1875)-

2. 0. Billeter, Ber. , 8, 462 (1875)-

3. 0. Billeter, Helv. Chira. Acta, 8, 337 (1925).

4. 0. Mumm and H. Richter, Ber., 7£B, 843 (1940).

5. A. Iliceto, A. Fava, and U. Mazzucato, Tetrahedron Letters, No. 11 , 27

(I960),

6. P. A. S. Smith and D. W. Emerson, J. Am. Chem. Soc, 82, 3076 (i960).

7. A. Fava, A. Iliceto, and S. Bresadola, ibid. , 87, 4791 (1965).

8. A. Iliceto, A. Fava, U. Mazzucato, and 0. Rossetto, ibid. , 83, 2729 (1961).

9. U. Mazzucato, G. Beggiato, and G. Favaro, Tetrahedron Letters, 5455 (1966).

10. W. G. Young, S. Winstein, and H. L. Goering, J. Am. Chem. Soc, 73, 1958

(1951).

11. A. A. Frost and R. G. Pearson, "Kinetics and Mechanism," John Wiley and

Sons, Inc., New York, N.Y. , 1953, PP- 104-105.

12. A. Streitvieser, Jr., Chem. Rev., j>6, 65O (1956).

13. A. Iliceto and G. Gaggia, Gazz. chim. ital. , 90, 262 (i960).

14. A. Iliceto, A. Fava, U. Mazzucato, and P. Radici, ibid. , 90, 919 (i960).

15. E. Bergmann, J. Chem. Soc, 1361 (1935).

16. A. W. Chapman, ibid., 127, 1992 (1925).

17. A. Fava, U. Tonellato, and L. Congiu, Tetrahedron Letters, 1657 (1965).

18. (a) S. Winstein and A. H. Fainberg, J. Am. Chem. Soc, 80, 459 (1958);

(b) S. Winstein and J. S. Gall, Tetrahedron Letters, No. 2, 31 (i960) ;

(c) S. Winstein, J. S. Gall, M. Hojo, and S. Smith, J. Am. Chem. Soc,
82, 1010 (i960); (d) S. Winstein, M. Hojo, and S. Smith, Tetrahedron Let-
ters,Nc22, 12 (i960); (e) S. Winstein, A. Ledwith, and M. Hojo, ibid. ,
341 (I96I).

19. S. G. Smith, ibid. , 979 (1962).

20. (a) H. L. Goering and J. T. Doi, J. Am. Chem. Soc, 82, 5850 (i960) ; (b)

H. L. Goering and J. F. Levy, Tetrahedron Letters, 6^+4 (1961)5 (c) H. L.
Goering and J. F. Levy, J. Am. Chem. Soc, 84, 3853 (1962); (d) H. L. Goering,
R. G. Briody, and J. F. Levy, ibid. , 8£, 3059 (1965).

21. (a) D. Darwish and R. McLaren, Tetrahedron Letters, 1231 (1962); (b) D. Dar-

wish and E. A. Preston, ibid. , 113 (1964); (c) D. Darwish and J. Noreyko,
Can. J. Chem. , 4£, 1366 (1965) •

22. S. Winstein, E. Clippinger, A. H. Fainberg, R. Heck, and G. C. Robinson,

J. Am. Chem. Soc, 78, 328 (1956).

23. Y. Okamoto and H. C. Brown, J. Org. Chem., 22, 485 (1957).

24. S. Winstein, E. Clippinger, A. H. Fainberg and G. C. Robinson, J. Am. Chem.

Soc, 36, 2597 (195*0-

25. A. Fava, A. Iliceto, A. Ceccon, and P. Koch, ibid. , 87, 1045 (1965).

26. A. Ceccon, I. Papa, and A. Fava, ibid. , 88, 4643 ( i960) .

27. A. Fava, A. Iliceto, and A. Ceccon, Tetrahedron Letters, 685 (1963).

28. L. A. Spurlock and P. E. Newallis, ibid. , 303 (1966).

29. J. Gillis, Rec trav. chim., jg, 330 (1920).

30. C. N. R. Rao and S. N. Balasubrahmanyam, Chemistry and Industry, 625 (i960) .

31. L. G. Cannell and R. W. Taft, Abstracts of the 129th A. C. S. Meeting,

46N ( 1956) .

32. R. Gompper, Angew. Chem. Int. Ed., J5, 560 (1964).

nli <~r

THE PHYSIOLOGICALLY ACTIVE CONSTITUENTS OF MARIHUANA

Reported by Donald C. Schlegel March 6, 1967

INTRODUCTION

Cannabis sativa Linn., more commonly known as hemp, is a plant which grows
wild or is cultivated throughout much of the world. The greenish resin known as
marihuana (The Americas), hashish (Middle East), or charas (Far East) can be
extracted from the flowers and seeds of the female plant and the leaves of both
sexes.1 The oil is composed of at least 25 recognized compounds, the major con-
stituents being caryophyllene ^5.7$, (3-humulene 16$, cc-selinene 8.6$, f3-farnesene
5.1$, a-bergamotene 5$, ^-phellandrene 2.7$. 2 In smaller content, 1.2$ or less
depending on the geographical origin, are two tetrahydrocannabinols, the active
constituents which make marihuana a psychotomimetically active drug.3 The search
for these compounds, their structural elucidation and most recently the resynthesis
of their optically inactive forms are the topics of this seminar.

PHYSIOLOGICAL EFFECTS

Marihuana is a psychotomimetically active drug. It is known to have a profound
effect on the central nervous system. It can be taken into the body by smoking or
direct ingestion. The effects have been "described as a feeling of well being
alternating with depression, distortions of time and space, and double consciousness."1
Illusion and fanciful hallucinations are common and disorientation and delirium may
follow. An increased sensitivity of the eyes to light is also observed.4

STRUCTURE ELUCIDATION

In the 1870's, chemists initially began to investigate marihuana.5 They met
with little success, however, and it wasn't until the 1890' s and the work of Wood,
Spivey and Easterfield5 that much was known about the supposedly active constituent
or constituents of marihuana. At that time they isolated a material C2;LH2602, I, from
the higher boiling fraction of the marihuana resin. Treatment of this compound with
cold fuming nitric acid yielded a yello\/ trinitro derivative C21II23N308, II. Further
oxidation of this product with hot fuming nitric acid gave among other products a
mononitro derivative CnHuNC^, III. The nitro group was reduced to the amine,
diazotized, treated with potassium iodide to form the iodide and then reduced with
sodium amalgum to form Ci:iH1202, IV, called cannabinolactone. 'This material was
reoxidized with basic permanganate to give a CnH1004 compound, V, cannabinolactonic
acid. Reduction of this gave a dibasic acid C1:LH1204, VI. Thirty years later Cahn6
further oxidized cannabinolactonic acid with hot dilute nitric acid and obtained
trimellitic acid, VII, thus showing the aromatic ring substitution. He also
demonstrated that the acid formed a lactone with a tertiary alcohol by collecting
acetone and 3-hydroxy-Wnethylbenzoic acid, VIII, on potassium hydroxide cleavage
of hydroxy-cannabinolactone.

Cahn applied the name cannabincl, earlier used to denote the high boiling
fraction of marihuana, to the compound C2iH2602, I. The material formed both mono-
methyl ether and monoacetyl derivatives. Measurement of the critical oxidation
potential gave a value, Ec = 0.995 ± 0.10, characteristic of substituted phenols but
not carbinols. Recovery of •►hexanoic acid on potassium permanganate oxidation
indicated the presence of a straight chain moiety. Empirical formula considerations
suggested that the chain was a n-pentyl side chain on the phenolic ring with the
sixth carbon atom coming from oxidation of the ring. Assembling the information
gave IX as the basic structure for cannabinol. It was not until the general
structure of cannabidiol was determined by groups led by Todd and Adams that the
complete structure of cannabinol was known.

5^11

Fuming
HNO^

cold CH3

^ ^g

cf/hi r^

KMnQ4
KOH

CH3-

dilute HNO3

I hot

Cannabidiol, another material found by Adams and Todd in the high boiling
fraction of marihuana was seen immediately to possess a molecular formula C2iH3002
quite similar to cannabinol.7 It had two acidic hydroxy 1 groups which were readily
acetylated, Ultraviolet spectral comparison showed the phenolic hydroxyls to be in
a resorcinol arrangement. Under mild hydrogenation conditions it rapidly consumed
two moles of hydrogen without great change in the UV chromophore, indicating that
the phenolic ring remained unchanged.8 Treatment of cannabidiol with pyridine
hydrochloride yielded conclusive proof of its general structure when p-cymene and
olivetol were formed.9

olio.

Py.'HClj

5HH

C5H11
CH2 CH3

The point of attachment of the two rings was determined by oxidizing tetrahydro-
cannabidiol. The resulting menthane carboxylic acid was isolated as its anilide/

JHa CH3

KMnQ4

■>

acetone
r.t.

/\0H-
CH3 CH3

CkH

5nll

CH3 CH3

Assuming cannabinol to be similar to cannabidiol, the last data suggested the
correct structure for cannabinol. Synthesis of this structure confirmed the identity
of cannabinol.10'11

CH3

5H11

Recently, using modern analytical techniques Mechoulam and Shvo12 have completed
the structural elucidation of cannabidiol. Earlier work by Adams et al.13 on place-
ment of the two non-aromatic double bonds had demonstrated that a terminal double
bond was present by recovering formaldehyde on ozonization of cannabidiol. Ultra-
violet studies of model compounds further showed that the second double bond was
neither conjugated with the aromatic ring nor with the terminal isopropylene double
bond.13 Consequently, the remaining double bond could be located only in positions
A5, A6, or A1. On NMR analysis, Mechoulam and Shvo observed 3 olefinic protons and

5Hn

5H11

2 olefinic methyl groups in the spectrum of cannabidiol. These data eliminated the
A5 isomer as a possibility. The NMR spectrum further showed the C3 proton at t6.15.
Normally a benzyl proton appears around T7tl3.14 This datum indicated the proximity
of another deshielding force, the double bond. The C3 proton appeared as a doublet
(J = 11) which is not appreciably coupled with the C2 proton. If cannabidiol has
the double bond in the A1 position, then 02 would be at an angle of approximately
850 and consequently not appreciably coupled to C3. On hydrogenation the product,
tetrahydrocannabidiol, shows the C3 proton at t7.^0. If the A6 isomer were the
correct structure, little change in the C3 proton position would be expected. A
final experiment was a selective epoxidation on the ring double bond of cannabidiol

-250-
bisdinitrobenzoate with perbenzoic acid. The NMR spectrum of the product showed no
change in the position of the methyl group or protons surrounding the side chain
double bond but the Ci, C2 and C3 groups did change. In particular the C3 proton
was now observed as a doublet at t6.70, again indicating the proximity of the
epoxide linkage. From the data it was concluded that cannabidiol is the A1
isomer.12

Evidence leading to the stereochemistry at carbon 3 and k , the final structural
unknown, initially came when Adams oxidized tetrahydrocannabidiol with potassium
permanganate in acetone and obtained menthane carboxylic acid as one of the pro-
ducts.9 He also found that L-raenthyl chloride could be converted via Grignard
formation and carbonation to the same menthane carboxylic acid. Later work by
Roberts, Shoppee and Stephenson15 has shown that both 3a and 3p-bromocholestane are
converted to cholestane-3P-carboxylic acid. Assuming a similar reaction with
menthyl chloride, an equatorial carboxyl group would be expected thus giving the
thermodynamically most stable eee isomer. In other evidence, the coupling constant
J = 11 for the C3 proton coupling with C4 proton is indicative of a diaxial con-
formation and thus a trans ring junction.12 These data then indicate that canna-
bidiol has a 3 >^ -"trans ring junction. The total structure is seen below.

7CH3

Ci8

/\<k

CH3 CH2

Early investigations by Adams et al. had shown that naturally occurring
cannabidiol could be converted with acids to two optically active, psychotomimetically
active products.16 Sulfur dehydrogenation converted these materials to cannabinol
thus proving their structures as tetrahydrocannabinols.17 Furthermore, both of the
isomers when hydrogenated yielded the same two epimers, proven similar by IR, KMR
and chromoplate comparison.18 Recent investigations have shown that treatment of
cannabidiol with a catalytic amount of p -toluene sulfonic acid produces the A6-^>,k-
trans -tetrahydrocannabinol, XIV, while treatment of cannabidiol with dilute hydro-
chloric acid in ethanol produces XV, the A1-3,4-trans-tetrahydrocannabinol.18"21

JH3

CH3 CH

5H11

0^^— CgHn

Lii3
OH XV

5nll

XIV

Structure proof of the double bond position in the isomers came from an NMR
analysis. In non-rigid cannabidiol the C3 proton is at t6.15 while the C2 proton
is at tk.kl. The fixed ring system of tetrahydrocannabinol, however, causes
deshielding by the aromatic ring of the C2 proton. Therefore, the A1 double bond
was assigned to the isomer with a C2 proton signal at T3.58 and the A6 structure to
the isomer with olefinic C6 proton at t^.55, a normal, unaffected olefinic proton
value . 21

-251-

Both the A1 and A6 isomers are psychotomimetically active. Taylor et al.22
have found the A1 isomer to isomerize partially to the A6 isomer when chromatographed
at 280° (column 10$ GE-SE-30 on Diatoport S) . From this observation, they suggest
that on smoking marihuana, the psychotomimetic effect may be due to the A6 isomer
instead of the A1. Gaoni and Mechoulam20 disagree after finding no isomerization of
the A1 isomer to the A6 isomer in gas chromatographic experiments up to 300° using
a different column (SE-30 on Chromosorb W) . Recent experiments by Lerner and
Zeffert3 have shown by using a smoking machine and analyzing the smoke by GC that
the A6 isomer increased from 3$ to only 9$ of the total A1, A6-tetrahydrocannabinol
content. This would suggest that although some isomerization does occur on
smoking, the major psychotomimetically active component is still the A1 isomer.

By treatment with p -toluene sulfonic acid in toluene, the A1 isomer can be
converted 9Cf?o to the A6 isomer.19 These same workers have also noted a slow
isomerization of the A1 to A6 isomer on long standing.19 Gaoni and Mechoulam20
suggest the facile isomerization occurs because there is much less steric crowding
in the A6 isomer between the phenolic hydroxyl and the C2 protons.

TOTAL SYNTHESIS OF TETRAHYDROCANNABINOLS

Mechoulam and Gaoni recently reported the first total synthesis of a
psychotomimetically active constituent of marihuana, dl-A1-3,^-trans -tetrahydro-
cannabinol.23 Initially citral a XVI was condensed with the lithium derivative of
olivetol dimethyl ether XVII. This condensation led to a complicated mixture which
was dissolved in pyridine and treated with p-toluenesulfonyl chloride to form the
sulfonate ester followed by ring closure and sulfonate elimination. Chromatography
of the mixture over 10$ silver nitrate and alumina yielded a fraction with polarity
similar to natural cannabidiol dimethyl ether. Re chromatography of this material
yielded dl-cannabidiol dimethyl ether XVIII in low yield. Demethylation with methyl-
magnesium iodide gave an QCffo yield of dl-cannabidiol which upon treatment with 0.05$
hydrochloric acid gave dl-A1 -3,^4- -trans -tetrahydrocannabinol.

CHO

/ ^5-^11

XVI

OR OCH3

R - H

R = SO2C6H4CH3

0.05$ HC1
.in EtOH

C5H11

<CH3*

155-65u
15 min

5^11

A second synthesis of both dl-A1 and A6-3,k -trans -tetrahydrocannabinol was
carried out by Fahrenholtz, Lurie and Kierstead.24 This synthesis involved a
Pechmann condensation of olivetol XIX with diethyl -cc-acetoglutarate XX to form the
expected coumarin XXI. Cyclization of XXI using sodium hydride in an aldol type
condensation gave XXII. Ethylene glycol converted XXII into the corresponding
ketal XXIII which exists in two polymorphic forms. Reaction of the ketal with
methyl magnesium iodide followed by acidic hydrolysis gave XXIV. Birch reduction
converted XXIV to the trans ketone XXV. Carbinol XXVI was obtained by conversion of
XXV to its tetrahydropyranyl ether followed by treatment with methyl magnesium
iodide and subsequent cleavage of the protecting group. Dehydration of XXVI using

.o^o^

a catalytic amount of p -toluene sulfonic acid in "benzene gave dl-A6-3,4-trans-
tetrahydrocannabinol as a single isomer shown by glpc. Treatment of XXVI with Lucas
reagent gave the corresponding chloride XXVII. Dehydrohalogenation of XXVII with
sodium hydride in refluxing tetrahydrofuran gave in quantitative yield a mixture of
the A1, lh1of and the A6, 26$, tetrahydrocannabinol isomers as shown by glpc.

QJJEt

<L PH3 OH

C5H11" v

XIX

1. CH3MgI
x2. H"1"

C=H

5nll

XXI

DMSO
NaH
,15-20°

XXIV

c (H0CH2)2
XXIII

5H11

XXII

XXIV

Li/Ma.
-78°

1 Conversion to ether

2 CHa^gl

5! Removal of hydroxyl

C5H11 protecting group

XXV

CrH

5-n-ll «.

ZnCl2

TIC1

C=H

5nll

^05^11

XXVI

71$

-tol-S03H

O^vx^CgHn

Recently Taylor, Lenard and Shvo carried out a one step synthesis utilizing
citral and olivetol in the presence of a 10$ benzene solution of boron trifluoride
etherate.22 Chromatography of the mixture over Florisil followed by preparative
VPC separation gave a 20$ yield of dl-A6-3,4-trans -tetrahydrocannabinol. , Some of
the A6-cis isomer was also obtained and identified by comparison with a sample of
the cis isomer prepared by an independent, unequivocal route.

-PS^-

53-

CHO

BF3 etherate
benzene

CsH

5^11

CH3^.0/ ^CsHu
CH3

BIOGENESIS

No labelling studies have been performed as yet, but by examination of the
other constituents of marihuana, the following biogenetic pathway has been postulated.
Olivetol probably is formed via the usual head to tail linking of acetate units to
form phenolic compounds.

CH3CCH2CCH2

^6=0

CH2 CH2-CC>2H

0=C

I
C=0

CH3CCH2CCH2

X/CO2H

oit^'Sdh

c=h

5^11

red,

-CO;

CH;

It is uncertain whether the pentyl side chain is formed from acetate units or
whether hexanoic acid is incorporated directly.25 Also in question is when the
decarboxylation occurs.

Once formed, olivetol XXVIII or the acid precursor XXIX probably condenses
with geranyl pyrophosphate to form either cannabigerol XXX or cannabigerolic acid

.OH

'5H11

p2or3

XXVIII R = H
XXIX R = CO2H

XXX R = H
XXXI R = COsH

PPi

Both cannabigerol and cannabigerolic acid have been found in extracts of marihuana,21
These two materials then condense further to the corresponding cannabidiol XXXII
and tetrahydrocannabinol XXXIII or cannabidiolic acid XXXIV and tetrahydrocannabinolic
acid XXXV.26'27 Whether the carboxylate derivatives are converted in part in any
of the preceeding steps to the decarboxylated derivatives is still uncertain. The
carboxylated derivatives have a high antibiotic activity against gram positive
bacteria which marihuana lacks and they may only compose another metabolic pathway
of the plant.26

■cy'-r-

C=H

5-^11

R = H XXXII R = H XXXIII R = H

R = COaH XXXIV R = C02H XXXV R = CO2H

SUMMARY

Recent studies have fully elucidated the structures of the natural psycho-
tomimetic ally active constituents of marihuana, A1 and A6-3^^ -"trans -tetrahydro-
cannabinol and a likely precursor, cannabidiol. Total synthesis of the dl forms has
also been accomplished. Speculations on the biogenetic pathway have been advanced.

BIBLIOGRAPHY

1. D. Downing, Quart. Revs., 16, 152 (1962).

2. M. C. Nigam, K. I. Handa, I. C. Nigam, and L. Levi, Can. J. Chem. , 4j5, 3372-6

(1965).

3. Chem. and Eng. News, 44, Dec. 26 (1966).

4. "Drug Addiction," Collier's Encyclopedia, 1966, VIII, 395.

5. T. Wood, W. Spivey, and T. Easterfield, J. Chem. Soc., 20-36 (1899).

6. R. S. Cahn, J. Chem. Soc, 1342-53 (1932).

7. R. Adams, M. Hunt and J. H. Clark, J. Am. Chem. Soc, 62, 196-200 (1940).

8. R. Adams, C. K. Cain and H. Wolff, J. Am. Chem. Soc, S|, 732-3 (1940).

9. R. Adams, C. K. Cain and J. H. Clark, J. Am. Chem. Soc, 62, 735-7 (1940).

10. R. Adams and B. Baker, J. Am. Chem. Soc, 62, 2405-8 (194*67.

11. R. Ghosh, A. R. Todd and S. Wilkinson, J. Chem. Soc, 1393-6 (1940).

12. R. Mechoulam and Y. Shvo, Tetrahedron, 12, 2073-8 (I963).

13. R. Adams, H. Wolff, C. Cain and J. Clark, J. Am. Chem. Soc, 62, 2215-9 (1940).

14. L. M. Jackman, Applications of NMR Spectroscopy in Organic Chemistry, Mac-
millian Co., N.Y. , N.Y. , I96TJ 5c~

15. G. Roberts, C. W. Shoppee and R. J. Stephenson, J. Chem. Soc, 27O5-I5 (1954).

16. R. Adams, D. Pease, C. K. Cain and J, H. Clark, J. Am. Chem. Soc, 62, 2402-
4 (1940).

17. R. Adams and B. Baker, J. Am. Chem. Soc, 62, 2401-2 (1940).

18. Y. Gaoni and R. Mechoulam, Tetrahedron, 22, 1481-8 (1966).

19. R. Hiveiy, W. Mosher and F. Hoffmann, J. Am. Chem. Soc, 88, 1832-3 (1966).

20. Y. Gaoni and R. Mechoulam, J. Am. Chem. Soc, 88, 5673-5 ^1966).

21. Y. Gaoni and R. Mechoulam, J. Am. Chem. Soc, B£, 1646-7 (1964).

22. E. Taylor, K. Lenard and Y. Shvo, J. Am. Chem. Soc., 88, 367-9 (I966).

23. Y. Gaoni and R. Mechoulam, J. Am. Chem. Soc., 87, 3273-5 (1965).

24. K. Fahrenholtz, M. Lurie and R. Kierstead, J. Am. Chem. Soc, 88, 2079-80
(1966). "~"

25. T. A. Geissman, Biogenesis of Natural Compounds, Ed. by P. Bernfeld, Macmillian
Co., N.Y. , N.Y., 1963, 580-9.

26. R. Mechoulam and Y. Gaoni, Tetrahedron, 21, 1223-9 (I965).

27. F. Korte, M. Haag and V. Classen, Angewande Chemie Int. Ed., 4, 872 (1965).

-255-

EENZENE PHOTOLYSIS

Reported by Warren J. Peascoe March 9 > 19&7

Many examples of reactions involving electronically excited "benzene have
recently been reported. This seminar will review these reactions. A brief sum-
mary of the general processes which occur upon absorption of a quantum of ener-
gy is presented.

Absorption of a quantum of energy by a molecule in the ground state So
produces an excited singlet which may return to the ground state by various pho-
tophysical processes.1 Singlets in the higher excited states S2, S3, . . .
undergo internal conversion to the first excited singlet state Si or undergo
inter system crossing to produce triplets. The Si singlet may also undergo in-
ter system crossing to produce a triplet, or it may return to the ground state
by internal conversion or fluorescence. The higher excited triplets rapidly
undergo internal conversion to the Ti state and may return to the ground state
by intersystem crossing or by phosphorescence.

The irradiation of benzene i,n^or with light in the 2300-2700 A region has
recently been reviewed by Noyeir/ 58who concluded that the absorbed energy is
efficiently dissipated by photophysical processes,, Irradiation at a shorter
wavelength, 1.0^9 A., has been reported by several groups to lead to the formation
01 polymer deposited on the walls of the reactor, two transient intermediates,
and minor amounts of hydrogen, methane, acetylene, ethane , toluene, and a C2
or C3 substituted benzene,3 8}55 The quantum yield (number of molecules which
undergo a specific reaction/number of molecules electronically excited) of the
formation of hydrogen was reported to be about 0.02 by Lipsky and coworkers,3
and Foote and coworkers reported the formation of acetylene with a quantum yield
of O.OI5.4 The major transient intermediate has been identified as fulvene 1
independently by Ward and coworkers5 and by Kaplan and Wilzbach55.

It was purified by preparative gas chromatography and had the same
uv, ir, nmr, and mass spectra as authentic fulvene. Fulvene was
shown to be the same intermediate detected but not identified by
the Foote4 and Lipsky6 groups by a comparison of the vapor phase
ultraviolet spectra and retention, time on gas chromatography.
The minor intermediate ir approximately l/6 the concentration of
1 fulvene was detected by Kaplan and Wilzbach by gas chromatography
and by its uv spectrum.55 Since it absorbed in the same region
as fulvene, its uv spectrum was determined from the difference between the uv
spectrum of irradiated benzene, vs. benzene blank, and the spectrum of fulvene
vapor in the same concentration as in irradiated benzene. On the basis of its

uv absorption (k ' 2500 A) and extinction coefficient (log e^ 4.5) ,

max

l,3-hexadiene-5~yne "was proposed as a possible structure.

The Foote and Lipsky groups both observed that fulvene formed rapidly and
reached a low steady-state concentration. The initial rate of formation of ful-
vene was found to be inversely proportional to the pressure of added nitrogen
by Foote and coworkers for nitrogen pressures of less tnan one atm;4 Lipsky and
coworkers found that the rate of formation decreased with increasing nitrogen
pressure from 0.1 to 50 atm. 6 The quantum yield for the disappearance of benzene
was also inversely related to the pressure; it was 1 at benzene pressures of less
than 0.1mm4 ; fell to 0.25 at 1mm, and decreased further if nitrogen was added.6
This dependence on pressure indicated that the reaction was collision quenched.
The polymer seemed to be formed by a further reaction of fulvene since the ratio
of the quantum yield of benzene disappearance to the quantum yield of the formation
of fulvene was a constant as the nitrogen pressure was varied.

Lipsky and coworkers calculated from the absorption oscillator strengths
that the second and third excited benzene singlets had fluorescence lifetimes
of less than 10 12 sec.9 Since no fluorescence was observed from these states,
the state lifetime must be even shorter; and a pressure of ca. 100 atm is required

-256-

in order to have the mean time between collisions equal the calculated fluores-
cence lifetime. Fluorescence and triplet sensitized emission of biacetyl were
observed upon excitation of benzene to the Si state with light of wavelength
greater than 2200 A. However, neither fluorescence from any singlet state
nor sensitized emission of biacetyl was observed upon excitation with light of
wavelength in the 2200-1600 A region corresponding to excitation to the S2
or S3 state. Thus either a direct photochemical reaction or very efficient in-
ternal conversion to the ground state is required from the S2 and S3 states.
The observed collision quenching at low pressures ruled out a direct photoreaction
upon irradiation at 1849 A. Lip sky and coworkers were led to propose that the
initially excited S3 or S2 benzene very rapidly underwent internal conversion
to vibrationally excited ground state which either lost its vibrational energy
to produce benzene or isomerized to fulvene.6

The irreversible formation of fulvene and polymer upon irradiation of liq-
uid benzene at wavelengths greater than 2000 A under nitrogen atmosphere was
reported by Bryce-Smith and coworkers,10-'11 with an approximate quantum yield
of 0.01. The uv spectrum of the product resembled that of authentic fulvene,
and the one-to-one adduct formed between it and maleic anhydride was found to
have the same infrared spectrum and mixed melting point as the adduct prepared
from authentic fulvene.

The photolytic isomerization of orthodi substituted benzenes has been re-
viewed through 1.964 in a previous seminar.12 The isolation of substituted
dewar benzenes and isomerizations through interconvertible dewar benzenes were
discussed. Additional examples of isomerizations of substituted benzenes have
been reported. The formation of perfluorodewar benzene from the irradiation
at 2537 A of perfluorobenzene in the vapor phase has been independently reported
by Haller13 and by Camaggi and coworkers.14 Lemal and Lokensgard15 have con-
verted hexamethyldewar benzene into a mixture of 20$ hexamethylprismane and hex-
amethylbenzene upon irradiation at 2537 A in butane solution. The hexamethyl-
prismane could be phot ochemic ally converted to the dewar benzene and hexamethylbenzene,

Wilzbach and Kaplan have reported that 1,3^5-tri-t-butylbenzene 2 photo-
isomerized at 2537 A in isohexane to l,2,4-tri-t~butylbenzene J5 through a benz-
vaiene intermediate 4. 16 Irradiation of 2 or ^ led to a photostationary mixture
with 7° 3°/° 2, 20.67c j5, less than 0.7$ h, 7.1$ £, and 64.8$ 6. Two fulvenes were
also produced in low yield from 2. The product mixtures were analyzed by nmr

(0.04)

I 6

utilizing the characteristic t-butyl peaks, and the quantum yields shown in

the diagram were determined for separate isomers 2, 3, 4> and 5* The prismane 6
could not be completely separated from the dewar benzene 5, a^d the quantum yields
for its reactions are approximate. The products were identified on the basis
of their spectral properties and rearomatization reactions on heating. The
benzvalene 4 showed a nmr methyne proton ABX pattern, t 8,27, 8.35; and 4.95
for Ay B. and X respectively, with J.=6.6cj< J -2. 45 , and J==1.25 cps. The

A.D AX. £3A

t-butyl protons were observed at 1 8.94 (18 H) and 8.99 (9 fi) . The benzvalene
rearomatized upon heating exclusively to j5 leading to the assignment of structure 4.

The simplest method for photoconversion of ^ into 4 involves bending the

-2.57-
planar ring of J5 along with the formation of transannular bonds; a similar ring
bending of 2 along with the formation of transannular bonds leads to 7 rather

than k. Wilzbach and Kaplan did not report the isolation and identification
of the intermediate initially formed from the irradiation of 2, and thus 7 rather
than h may be the intermediate in the photoisomerization of 2 into 3« It is
possible that J may not have been detected since the nmr chemical shifts of the
t-butyl protons of 2 would be expected to be very similar to the chemical shifts
of the t-butyl protons of h. A second possibility is that 4 was the intermediate
and that it was formed from the isomerization of 7 in the ground or excited state.
As a member of the class of valene compounds , the isomerization of benzvalene
7 to benzvalene h might be expected.17 Evidence that irradiation of 2 leads
to the initial formation of 7 in either an excited state or in the ground state
has been reported by Kaplan and coworkers o18 They found that 2 irradiated in
methanol produced 8 with a quantum yield of 0.15 and that no isomeric tri-t-

butylbenzenes were formed.

nv
eOH

)

L 2

OMe

The product was identified
by its spectral properties,
and the proton in position
6 was assigned the endo-
configuration on the basis of
its nmr chemical shifty T 9«27>
and coupling constants

Ji,6=4.6 and J5.,6=2.7 ops. The t-butyl groups were shown not to be on adjacent
caroons by thermal conversion of b into l^^-tri-t-butylfulvene.

The formation of benzvalene upon irradiation of benzene in. the liquid phase
at 2537 A has been reported by Kaplan and coworkers56. Concentrations of product
(ifo based on. benzene) high enough for nmr spectroscopy and preparative gas chrom-
atography were obtained by irradation of dilute solutions of benzene in hexa-
decane at 650 - Treatment with methanolic HCT of either irradiated benzene or
the product purified by preparative gas chromatography and absorbed in isohexane
produced two products identified by methods not described as 6-endo-methoxy
(3.1.0)bicyclo-2-ene> 10 (R^CHa), and one of the 4-methoxy isomers, 9 (R=0CH3).
The uv spectrum of the irradiated benzene product showed no maximum above 2100 A
but had a broad shoulder between 2200 A and 2J00 A ( S-^OO) . The nmr spectrum
of the product showed three resonances of equal area* an unsymmetrical triplet (1.5-and
1.7-cps couplings) at T k.Ok? a symmetrical triplet (l.5-eps couplings) at T 6.47, and

a quintet (L J=6.2 cpsj at T 8.16. The product rearomatized only slowly at
room temperature and was assigned the benzvalene structure.

Photoaddition reactions have been reported which may be rationalized as
formal additions to benzvalene and to dewar benzene. Farenhorst and Bickel19
reported that benzene .irrarHqted in acetic acid with a low pressure mercury lamp
produced polymer, £ (*R--0C0CH3) , and two other unidentified products in lesser amounts.
The acetoxy group of Q (R=-X0CH3) is reportedly in the exo-position from nmr
analysis, but only chemical shifts without any splitting patterns were reported.
Without nmr data for the endo- and exo-compounds the stereochemistry must remain
tentative. Compound £ (R=-0C0CH3) could not be hydrolyzed in either mild acid
or base but in both cases produced polymer and a yellow discoloration due to
fulvene as indicated by the uv spectrum. Irradiation of aqueous phosphoric
acic saturated with benzene produced two products tentatively identified as alde-

"*"° ° /p- OH) with the hydroxy group in the

-258-
exo -posit ion. The stereochemistry of the hydroxy group was based only on report-
ed nmr chemical shifts.

5 k

+ R

Kaplan and coworkers18 reported that benzene irradiated in trifluoroethanol
yielded 2. and ±2 (R=0CH2CF3) in the ratio of 2:1 with a total quantum yield of
0.05° The identification of the products was based on spectral data. Nmr double
resonance analysis of 10 (R=0CH2CF3) showed H6 to be coupled with Hi and H5
( J=7cps) indicating an exo -posit ion for H6 to the authors. The coupling constant
is closer to the reported vicinal cyclopropyl coupling constants 5*2 to 8.0 cps
for trans hydrogens than to 8.0 to LI. 2 cps for cis hydrogens,,57 and the opposite
stereochemistry , H6-endo, is indicated at C6. The stereochemistry of 9 (R=0CH2CF3)
was not determined.

The photolytic formal addition of olefins to benzvalene has also been re-
ported by two groups. Wilzbach and Kaplan20 reported that products correspond-
ing to 11 could be isolated by gas chromatography from the product mixtures from
the irradiation at 2537 A of benzene and ethylene, cis-2-butene, cyclopentene ,
or 2,3-dimethylbut-2-ene. The products were characterized by their spectral
properties and extensive nmr decoupling experiments. Bryce-Smith and coworkers21
reported that equimolar mixtures of cyclooctene and benzene at room temperature

R2
\

R-

R-3 R.4

^i*^C j/

R2
R-

rTT

2—f

or at -60° irradiated in the 2350-2580 A region produced two 1:1 addition products
in the ratio of 8:1. Benzophenone and acetone were ineffective as sensitizers but
|3-propiolactune increased the rate of addition two-fold. The major product 11
was characterized by spectral studies } products from catalytic hydrogenation, and
formation of 1:1 adducts with tetracyanoethylene or phenyl azide. But-1-ene, oct-1-ene,
cyclohexene, cycloocta-l,5-diene , and ethyl vinyl ether also gave 1:1 photoadducts
analogous to 11. The structure of the minor product was suggested to be 12 on the
basis of a sharp olefinic singlet at T k.kj. Srinivasan and Hill22 reported the
-formation of 12 in 50'° yield from the photoaddition of benzene and cyclobutene at
2537 A. The product was identified on the basis of its spectra (a sharp nmr reso-
nance for two olefinic protons at T 4.25) and the products produced upon catalytic
hydrogenation. Upon heating to 200° , the product decomposed to produce benzene and
butadiene.

Bryce-Smith and coworkers23 have noted that both 11 and 12 provide an olefinic
functional group and should be able to add a second molecule of benzene in a step
towards photopolymerization. They irradiated benzene and small amounts of
cyclooctene (0.35 mole-fa) or the cyclooctene-benzene 1:1 adduct (0.06 mole-fo)
for 100 hr with a medium pressure mercury lamp and found a polymer to be produced.
The polymer was fractionated by its solubility in hexane, acetone, benzene, and
chloroform. The major fraction in benzene had a molecular weight of about 1500.
There were no aromatic protons in the nmr spectrum, and the ir spectrum showed six
of the seven bands observed in monomeric 1: , The trace amount of olefin was
sufficient to suppress effectively the formauxon of fulvene. Polymers with different
properties were reported to be formed with light of wavelength of ca. 2000 A.

Koltzenburg and Kraft24-*25 reported the photoaddition of 1,3-ciienes to benzene
upon irradiation at 2537 A. Gas chromatography showed at least ten products from
the reaction of isoprene with benzene with a hBp ;yield of 1/$ and a 2j°r yield of 14.

-259-
The products 1J5 and ih were identified on the basis of their chemical and spectral

h3L

15 14 H

properties. Toluene, o- xylene, perfluorobenzene, and 1,2,4,5-tetrafluorobenzene
have also been found to form dimeric adducts with isoprene. The photoadduct
from butadiene and benzene dimerized to for:: 15. Attempts to sensitize the reaction
with benzophenone led only to the formation of dimeric dienes.

A 2:1 photoadduct formed from maleic anhydride or N- substituted maleimide
and benzene has been reported.26 29 These reactions which proceed by photo-
activation of the substrate or a charge-transfer complex between substrate and
benzene, rather than by direct photoactivation of benzene, will not be discussed.

Two more photoaddition reactions for which no evidence is available for
determination of the erMted rpqctine; species must be considered. The photo-
addition of methyl acetylenecarboxylate or dimethyl acetylenedicarboxyiate to
benzene has been reported by two groups to produce substituted cyclooctatetraenes
17. 30 32 The reaction is thought to proceed by 1,2-addition to benzene to form
lb which isomerizes to 17. The cyclooctatetraene acid formed by saponification

-R

■R1

16 17

of the ester was identical with the acid prepared by an alternate route. The
diacid was assigned structure 17 on the basis of its spectra and products produced
by catalytic hydrogenation. The formation of 18 by photoaddition of 2-methylbut-2-ene

18 12

to benzonitrile has been reported by Atkinson and coworkers.33 They found that
benzophenone effectively quenched the reaction and that aliphatic acetylenes
added to give cyclooctatetraenes 19.

To account for the photoisomerizations and photoaddition reactions, Bryce-
Smith and Longuet-Higgins have proposed a mechanism which allows the rationalization
of the observed products in terms of the lowest benzene singlet and triplet.34 The
first benzene singlet which has B 5^/mmetry35 is antisymmetric about a plane

through any opposite pair of caroun atoms ; the lowest electronic configuration

of the singlet biradical 21 is also antisymmetric about its plane of symmetry.

An orbital correlation diagram36 indicates that 20 may pass adiabatically into 21.

) Fulvene

in addition

-) R

-} Benzvalene

4 fa

"^^ — > rrrrg,

> Pr i smane

d\ 1>1 "^ "> 1,4-adduct

-260-
Similarly one component of the lowest benzene triplet, which has B symmetry37

and is antisymmetric about a plane bisecting any opposite pair of carbon-carbon
bonds, may adiabatically pass into the triplet state of either of the diradicals
23 or 24o The lowest electronic configuration of both 23 and 2J+ in the trip-
let state is antisymmetric about a plane bisecting the terminal bonds of 23
and the double bonds of 24. The diradicals are proposed to react as indicated
to form the observed products from electronically excited benzene. There is
no experimental evidence which requires the diradicals 21, 2J5, and 2k to be on
the reaction path leading from electronically excited benzene to products, and
it is possible that benzene in the lowest singlet or triplet state reacts dir-
ectly to form the observed products. The experimental determination of the mult-
iplicity of the electronically excited state leading to the formation of dewar
benzene, dewar benzene addition products, prismane, and 1,4- addition products
would be useful in evaluating this mechanism since triplet states are predicted.
The formation of cyclooctatetraenes by the 1,2-addition of acetylenes to benzene
and the 1, 2~add.it ion to benzonitrile were not explicitly considered by the authors.
It has not been established that these reactions must proceed by attack of excited
benzene on the substrate. An. alternate mechanism, attack of excited substrate
on benzene, would not require modification of the proposed reaction scheme.

An alternate mechanism involving a single highly reactive intermediate has
been proposed by Farenhorst,38 van Tamelen39 has pointed out that the Woodward-
Hoffmann rules40 predict preferential conrotatory ring opening of cyclobutene
systems, and that conrotatory ring opening of dewar benzene leads to trans -benzene 25.

conrotatory

ring opening

The trans-benzene pi-orbitals are topologically equivalent to a conjugated six-
membered Moebius ring. The HMO energy calculated by Heilbronner41 for a ground
state six-membered Moebius system was shown to be equal to the HMO energy of
benzene in its first excited state. Farenhorst thus proposed that benzene in
an excited state isomerized to trans -benzene in the ground electronic state with
a Moebius pi-electronic structure and that the trans -benzene reacted to form the
observed products. He also pointed out that a possible transition state 26
leading to the formation of dewar benzene and a transition state 27 leading to
benzvalene would have nearly the same energy as trans-benzene. These transition

/ i

/ 1 2
26

states consist of a localized double bond (C5-C6) and a conjugated four-membered
Moebius ring (Ci,C2,C3,C4) \ the transition state 27 also contains transannular
bonds between Ci and C3 and between C2 and C4.

Irradiation at 2.537 A of benzene in rigid organic glasses at liquid nitro-
gen temperature has been studied by many groups.42 54 A product proposed by
Anderton and coworkers to be a substituted hexatriene was detected by its uv
spectrum.43 The position of the three observed peaks varied with the solvent
used to form the rigid glass.43 46 Migirdicyan and coworkers reported that the
product from the reaction in ethanol 28 (R* CHOHCH3) dehydrated on preparative

Cg) mSd-> H '• iR //(f^C^j/i

alass

-261-

gas chromatography to produce 1,3,5,7-octatetraene identified by its uv spec-
trum. 45'j46 After gas chromatography the product from the reaction in methanol
28 (R= -CH2OH) still showed the same three uv peaks. The formation of the sub-
stituted hexatriene has been reported to be first order in the intensity of the
exciting light by Migirdicyan and coworkers.47"49 There is no evidence which
allows the distinction between a four-centered reaction and one involving a hex-
atriene diradical.

Several groups have observed the esr spectra of solvent radicals upon ir-
radiation of benzene in rigid glasses.50 53 The esr spectra of the radicals
from the 2537 A irradiation of benzene in the glass and esr spectra of the rad-
icals produced by the y-radiolysis of the pure solvent glass were the same.
The formation of molecular hydrogen with a yield greater than twice the yield
of radical formation was reported by Shelimov and coworkers.51 The source of
most of the hydrogen was the solvent since the use of benzene-d6 led to a ratio
of H2/HD of 9»5 a"t benzene-d6 concentration of 1.8X10 3 M and 8.1 at benzene-d6
concentration of 6.0X10 2 M as determined by mass spectroscopy. The formation
of both hydrogen and solvent radicals was reported to be second order in the
intensity of irradiation. 50>52j'53 The reaction is interpreted in terms of trip-
let benzene absorbing a second quantum of energy to produce a doubly excited
triplet. The doubly excited triplet transfers its energy to a solvent molecule
(RH = 3-methylpentane , methylcyclohexane, 2-methyltetrahydrofuran, methanol,
isopropyl alcohol, or cyclohexane) producing ground state benzene and a solvent
triplet which dissociates into two radicals, H* and R° . The hydrogen radical
abstracts a hydrogen from another solvent molecule and produces a second trapped
solvent radical R°.52 Support for the biphotonic process involving a doubly
excited triplet comes from the recent observation by Godfrey and Porter54 of
the absorption spectrum of triplet benzene which was observed after flash photol-
ysis of benzene in a rigid matrix at 77° K.

SUMMARY

Benzene irradiated in the vapor phase at wavelengths shorter than 2200 A
produces polymer , f ulvene , an unidentified intermediate „ and decomposition prod-
ucts. Fulvene and benzvalene have been isolated from irradiated benzene solutions
and the presence of benzvalene and dewar benzene in solution has been detected
by formal addition reactions to both benzvalene and dewar benzene. Two mechan-
isms have been proposed to account for the benzene photoreactions. The Bryce-
Smith and Longuet-Higgins mechanism is based on the correlation of the lowest
benzene singlet and triplet states with singlet or triplet states of diradicals
which can lead to the observed products. The Farenhorst mechanism is based on
a trans-benzene intermediate which has pi-orbitals topologically equivalent to
a Moebius ring. In rigid media at 77° K benzene forms substituted hexatrienes
by a monophotonic process and leads to the formation of solvent radicals and
hydrogen by a biphotonic process.

BIBLIOGRAPHY

1. N. J. Turro, "Molecular Photochemistry," W. A. Ben(iamin, Inc. , New York,
N. Y. , I965.

2. W. A. Noyes Jr. and I. linger, Adv. in Photochemistry, hs h$ (1966).

3. F. Mellows and S. lipsky, J. ' Phys. Chem. , JQ, i+076 ( 1966).

k. J. K. Foote, M. H. Mallon, and J. N. Pitts Jr., J. Am. Chem. Soc. , 88,
3698 (1966).

5. H. R. Ward, J. S. Wishnok, and P. D. Sherman Jr., J. Am. Chem. Soc, 89,
162 ( 1967) .

6. K. Shindo and S. Lipsky, J. Chem. Phys., 4£, 2292 (I966).

7. J. N. Pitts Jr., J. K. Foote, and J. K. S. Wan, Photochemistry and Photo-
biology, k, 323 (1965)0

8. J. E. Wilson and W. A. Noyes Jr., J. Am. Chem. Soc, 6^3, 3025 (1941).

-262-

9. C. L. Braun, S. Kato, and S. Lipsky, J. Chem. Phys. , ^, 1645 (1963).

10. H. J. F. Angus, J. Mo Blair, and Do Bryce- Smith, J. Chem. Soc, 200.3 (i960).

11. Jo Mo Blair, and D. Bryce-Smith, Proc. Chem. Soc, 287 (1957).

12. M. McDaniel, U. of I. Organic Seminar, 59 (Summer 1965) .

13. I. Haller, J. Am. Chem. Soc, 88, 2070 (1966).

14. G. Camaggi, F. Gozzo, and G. Cevidalli, Chem. Comm. , 313 (1966).

15. D. M. Lemai and J. P. Lokensgard, J. Am. Chem. Soc, 88 , 593^ (1966).
I60 K. E. Wilzbach and L, Kaplan, J. ' Am. Chem. SocM 87, 4~004 (I965).

17. H. Go Viehe, Angew. Chem. Internat. Ed. , 4, 746 (1965).

18. L, Kaplan, J. Ritscher, and K. E. Wilzbach, J. Am. Chem. Soc, 88, 2881
(1966).

19. E. Farenhorst and A. F. Bickel> Tetrahedron Letters., 59H (1966).

20. K. E. Wilzbach and L. Kaplan, J. Am. Chem. Soc, 88, 2066 (1966).

21. D. Bryce-Smith, A, Gilbert, and B. H. Orger, Chem, Comm. , 512 (1966).

22. R. Srinivasan and K. A. Hill, J. Am. Chem, Soc, 87, 4653 (1965).

23. D. Bryce- Smith and A. Gilbert, Chem Comm., 643 (i960).

24. G. Koltzenburg and K. Kraft, Tetrahedron Letters, 389 (1966) .

25. G. Koltzenburg and K. Kraft, Angew. Chem. Internat. Ed., hj 981 (1965).

26. D. Bryce-Smith and J. £. Lodge, J. Chem. Soc, 2675 (I962)

27. Do Bryce-Smith and A. Gilbert,' J. Chem. Soc, 918 (1965).

28. D. Bryce-Smith and M. A. Hems, Tetrahedron Letters, 1895 (1966).

29. J. S. Bradshaw, Tetrahedron Letters, 2039 (1966).

30. E. Grovenstein Jr. and D. V. Rao, Tetrahedron Letters, 148 (1961).

31. D. Bryce-Smith and J. E. Lodge, J. Chem. Soc, 695 (1963).

32. D. Bryce-Smith and J. E, Lodge, Proc. Chem. Soc, 333 (196l).

33. J. G. Atkinson, D. Eo Ayer, G. Biichi, and E, W. Robb, Jo Am. Chem. Soc,
85, 2257 (1963).

34. D. Bryce-Smith and H. C. Longuet-Higgins, Chem. Comm., 593 (1966).
35 » S. F. Mason, Quarterly Reviews, 15, 287 (1961).

36. H. C. Longuet-Higgins and E. W. Abrahamson, J. Am. Chem. Soc, 87, 2045

(1965).

37. G. W. King and E. H. Pinnington, J. Mol. Spec try. , 15, 39^+ (1965)*
380 E. Farenhorst, Tetrahedron Letters, 6465 (1966).

39. E. E. van Tamelen, Angew. Chem. Internat. Ed., 4, 738 (1965).

40. Ro B. Woodward and R. Hoffmann, J. Am. Chem. Soc, 87, 395 (1965).

41. E. Hielbronner, Tetrahedron Letters, 1923 (1964).

42. G. E. Gibson, N. Blake, and M. Kalm, J. Chem. Phys., 21, 1000 (1953)-

43. E. J. Anderton, H. T. J. Chilton, and G. Porter, Proc Chem. Soc, 352
(I960).

44. I. Norman and G. Porter, Proc Roy. Soc, A230, 399 (1955).

45. E. Migirdicyan, J. Chim. Phys,, 6j>v 520 ( 1966) .

46. E. Migirdicyan and S. Leach, J. Chim. Phys., _5j3, 762 (1961).

47. S. Leach, E. Migirdicyan, and L. Grajcar, J. Chim. Phys., 56, 748 (1959).
48o E. Migirdicyan and S. Leach, J. Chim. Phys. , 58, 4l6 (I96TJ".

h-9. E. Misrirdicyan , J. Chim. Phys. . 6j5, 535 (1966XT

50. V. G. Vinogradova, B. N. Shelimo^rLV.Fok, and V, V, Voevodskii, Dokl. Akad.
Nauk. S. S. S. R. , 154, i88 (1964).

51. B. N. Shelimov. N. V. Fok, V. V. Voevodskii, Dokl. Akad. Nauk. S. S.
S. R. , 144, 59o (1962).

52. B. Brocklehurst , W. Ao Gibbons, F. T. Lang, G. Porter, and M. I. Savadatti,
Trans, Faraday Soc, 62, 1793 (1966).

53. E. Migirdicyan, J. Chim. Phys., 6^, 5^3 (1966).

54. T. S. Godfrey and G. Porter, Trans. Faraday Soc, 62, 7 (1966).

55. L. Kaplan and K. E. Wilzbach, J. Am. Chem. Soc, '8£, 1030 (1967).

560 K, E. Wilzbach, J. S. Ritscher, and L. Kaplan, J. Am. Chem. Soc, 8g, 1031

( 1967) •
57. D. J. Patel, M. E. H. Howden, and J. D. Roberts, J. Am. Chem. Soc, 85,

3218 ( 1963) 0

560 w. A. JMoyes Jr, and D. A. Barter, J. Chem. Phys., 46, 674 (I967).

-263-
THE FHOTODIMERIZATION OF THYMINE

March 1.6, 1967

Reported by Sheldon A. Schaffer

INTRODUCTION

It has long been hypothesized that ultraviolet (UV) radiation damage to micro-
organisms has involved alterations in the deoxyribonucleic acids (DNA) of the
organism,1 but until fairly recently a molecular interpretation of the radiation
damage was unavailable. Since the UV spectrum of DNA is due almost entirely to the
absorptions of its purine and pyrrolidine base components, most of the investigations
of UV photodamage have involved studies of the photochemistry of the most common
purines, adenine (I) and guanine (II), and the most common pyrimidines, thymine (III),
cytosine (IV) and uracil (V) (Note that thymine is found only in DNA and uracil is
found only in ribonucleic acids (RNA)).

NHp

H=K

NHo

IT
H

O^^N-
H

Early studied2"4 indicated that the pyrimidine bases were much more sensitive
to UV radiation than' the purines, and therefore most investigations have been
directed toward the photochemistry of the pyrimidines.

ISOLATION OF THE THYMINE PHOIODIMER

The first photoproduct isolated from UV irradiation of solutions of pyrimidines
was the hydration product of uracil, 5,6-dihydro=.6=hydroxyuracil.2j,5-?6 Berends and
co-workers7 pointed out that the slow reversible hydration of uracil could not
account for the rapid lethal effect of UV radiation on bacteria. Their investigations
showed that besides the reversible hydration of uracil an irreversible reaction takes
place when the irradiation is carried out on an aqueous solution at neutral pH.
Later work on the irradiation of pyrimidine bases in frozen aqueous solutions showed
that thymine, which exhibits only weak susceptibility to UV radiation in liquid
aqueous solutions, was highly sensitive to such radiation in frozen solutions.
Thus, irradiation of a frozen aqueous solution of thymine with" a 4w germicidal lamp
(with maximum intensity at 25^ mu) for 15 minutes resulted in a kOfo loss of thymine
absorptivity at 264 mu. Uracil under the same conditions lost only 6$ (corrected
for the formation of reversible photoproduct) of its absorptivity at 260 m^i, while
cytosine was totally unaffected. The reason for the not too obvious experiment of
irradiation of frozen solutions of pyrimidines was that DNA in aqueous solution is
known9 to be surrounded by an ordered lattice of water, and it was thought that
frozen aqueous solutions of the pyrimidine bases would approximate the native state
of DNA better than liquid solutions and thus provide a better model for the inves-
tigations of the photochemistry of DNA.

Irradiations of DNAX and apurinic acid (DNA which has undergone a mild acid
hydrolysis to remove all the purine bases) gave products chromatographicaliy
identical to the so-called "first irreversible product"7 obtained from the irradia-
tion of thymine in frozen solutions.

Beukers and Berends12^13 reported the first physical data on the thymine photo-
product. Their compound gave an elemental analysis for CsHeO^T^, the molecular
formula of thymine, thus showing that it was either a dimer or polymer of thymine.
A molecular weight determination by the method of isothermic distillation14 gave
values between 24-0 and 270 indicating that the molecule was a dimer of thymine
(calc. mol. wt. ■ 252). The compound lacked UV absorption at 264 mu, indicating
the lack of the C5-Cs double bond of thymine, and its infrared spectrum showed weak
peaks at 960 cm -1 which were taken to be characteristic of a cyclobutane ring system

■cut-

(in the light of recent evidence16 this assignment is somewhat in doubt). The nmr
spectrum17 (DMSO»d6) showed signals (relative to external water) at -326 (2H,
singlet,, -W) , -179 (2H* singlet, -NH) , 4±3 (2H, singlet, -CH) , and +185 cps (6H,
singlet, C-CH3) . On the basis of these data Beukers and Berends13 suggested
structure VI for the thymine photodimer. Wulff and Fraenkel17 pointed out that
there are four' structures, VI-IX, which fit the data for the thymine photodimer.

CH3 H

cAn

H CH3
0

CH3 CH3

/ff

H H
H H

VIII

0 H

. H CH;
IX

•nh

They also showed that molecular models indicate that only the cis-syn Isomer (VII)
could reasonably be expected to be formed from the dimerization of adjacent thymines
on the same DNA. strand, and therefore the cis-syn structure should be assigned to the
dimer isolated from DBA. An enzymatic digestion of a sample of UV irradiated DNA
gave various trinucleotides of the type pXpTpT17 (where X is any one of the four
bases occurring in DMA and TpT is the thymine dimer joined by the normal phospho-
diester bond between the ribose moieties of the thymine nucleosides) . This implies
that at least some of the thymine dimers are formed between adjacent thymines on
the same DNA strand.. Furthermore, they suggested that one of the trans isomers
(VIII or IX) might be formed in cross-linked dimers between two DNA strands.18*19

TEE STRUCTURE DETERMINATION OF THE THYMINE PHOTODIMER

In 1965 Blackburn and Davies20 offered the first chemical proof of structure
for the thymine dimer. They found20*21 that the treatment of the thymine dimer
(formed by the irradiation of a frozen solution of thymine) with 10$ NaOH gave a
disodium salt believed to have structure X because it exhibited a maximum UV
absorption at 237 eu« The monoanion of thymine has its maximum absorption at 23C rnu.
Increasing the NaOH concentration to k-Cffo resulted in the formation of a. white
precipitate which showed only end absorption in the UV and whose IR spectrum showed
bands at 1310, I56O (COi) , 1655 ( -NHCQNH2) , 315 0, 3280 and 3370 cm"1 (amide NH's).
This material was assinged the structure of the disodium salt of the bisureido acid
(XI). Such ureido acids are known to be formed from 5^6-dihydrouracils by treatment

CH3 CH3
Na 02&„ V 17- . € 02Na

NaOH

°^< H

NaOH

H2N

^Sr^S

with base.22*23 When XI was dissolved in water it reverted to X, and X could be
converted back to the thymine dimer by treatment with acid. The facile recyclization
of XI in water is in contrast to the behavior or the p-ureido acid salts or the

-2op-

5,6-dihydrouracils which recyclize only in the presence of acid22-"23 and is
probably due to the all cis arrangement of groups on the cyclobutane ring.

Treatment of XI with bromine (an attempted Hoffmann rearrangement) resulted in
the formation of a rearranged product, isomeric with the thymine dimer. This com-
pound exhibited only UV end absorption in both alkaline and neutral solutions 5, its
IR spectrum showed bands at iGjk (WHCGNH2) , 1723, 1770 (COOTCO) , 32^0, 337Q and
3^20 cm'1 (amide NH's) j the nmr spectrum (in trifluoroacetic acid) gave signals at
tO.02 (IB, singlet, W) , 2.55 (IH, singlet, MH) , h.96 (IE, doublet, J = 9 cps,
cyclobutane hydrogen), 5.55 (XH, quartet, J± = 9 cps, J2 ■ 2 cps, cyclobutane
hydrogen), 8.38 and 8. if 5 ( 3H each, singlets, C-CH3) . The weak coupling of the
signal at T5.55 could be eliminated by running the spectrum in deuterated TFA. In
DMSO solution an additional broad band appeared at t2„9 (2H, 33H) . On the basis of
these data, Blackburn and Davie s offered structure XII as the rearranged compound. „

9 CH3 H H

The chemical proof offered to support structure XII is as follows : treatment
of the rearranged product with nitrous acid results in the loss of the carbamoyl
moiety with a simultaneous collapse in the nmr AB pattern to an A2 system. This
requires the location of the carbamoyl group on one of the nitrogens of the
imidazole ring to provide the molecular dissymmetry which makes the cyclobutane
hydrogens magnetically non-equivalent.

Pyrolysis of XII, or its decarbamoylated derivative, gave 2,3-dimethylmaleimide,
XIII, and 2-imidazolone, XIV. The production of XIII shows that both the rearranged

xrv

product and the thymine dimer had methyl groups on vicinal carbons, and therefore
the dimer must have one of the syn structures VII or VIII. A syn structure is also
indicated by the Jc13H satellite spectrum in the nmr of the dimer.24

The formation of 2-imidazolone (XIV) from the pyrolysis, taken with the nmr
data of XII, confirm that the cyclobutane hydrogens of the dimer were vicinal. The
coupling constant of 9 cps for these hydrogens is consistent25 with their being cis
oriented, but their orientation cannot be rigorously proved in this way as the
Karplus relation26 is not strictly applicable to systems containing strong electro-
negative groups.27 Further proof for the cis ring junction was based on the failure
of similar systems to give trans fused rings28 and the fact that only cis fused
dimethyl bicyclo[4.2.0]octane-7,8-dicarboxylate gives cyclohexene on pyrolysis.29

Chemical proof of structure for the DMA derived thymine photodimer was obtained
by Blackburn and Davie s30 by growing E^ coli on thymidine -6T, irradiating the
bacteria with UV light, and isolating the photoproducts chromatographically. The
photoproducts were mixed with carrier thymine "ice "■dimer" and repeatedly recrystallized
without loss of any radioactivity. Treatment of an NaOH solution of the dimer with
bromine gave the rearrangement product XII with ^Ojo of the initial activity. Since
Blackburn and Davies feel that the oxidative rearrangement is only allowed for the
cis-syn isomer, they conclude that the DHA. derived dimer must then have structure VII.

:66~

STUDIES ON SUBSTITUTED THYMINES

Studies on the UV Induced photodimerizations of substituted thymines 9X7 s 31
thymidine,32 and thymidyl (31 "*" 5') thymidine33 have shown that more than one
cyclobutane type thymine dimer may be formed. The identification of the photodimers
from substituted thymines has been presented by Blackburn and Davies31 and rests on
the following evidence ; irradiation of a frozen solution of 1,3-dimethylthymine
leads to two cyclobutane type photodimers17 the higher melting of which was shown
to be identical with the tetramethyl derivative of the thymine "ice -dimer",17 and
thus has structure XXb. Similarly, 1-methylthymine gives two cyclobutane photo-
dimers when irradiated in a frozen aqueous solution.31 Both of these photodimers
are alkylated smoothly with dimethylsulfate to give the tetramethyl compounds. One
of these compounds was shown to be identical to the tetramethyl derivative of the
thymine "ice-dimer" and thus the original compound had structure XXc. The other
dimethylated -derivative of the 1-methylthymine dimer was different from the
unas signed, lower melting dimer from 1,3-dimethylthymine and thus represents a third
isomer.

0 0 0 0

(a:R=R'=H)

(b:R=R'=CH3)

(c;R=CH3,R»=H)

(asR=R»=H)
(bsR=R»=CH3)
(c:R=CH3,R»=H)
(dsR=H>R'=CH3)

XXI

(a:R=R'=H)
(b;R=R»=CH3)

XXII

(a;R^R5=H)
(b;R=R«=CH3)

Blackburn and Davies pointed out that 1-methylthymine exists in two crystal
modifications34 and that topochemical arguments (vide infra) presented by Stewart35
show that irradiation in the solid state of the more stable of the crystal forms
should give the dimer XIXc. On this basis the second dimer of 1-methylthymine was
assigned the trans -syn structure XIXc. This implies that the unassigned dimer from
1,3-dimethylthymine must have one of the anti structures, either XXIb or XXIIb.
An assignment cannot be made for this compound at this time.

The structure of the photodimer of 3~niethylthymine was deduced from the partial
methylation of the thymine "ice-dimer" to give a symmetrical dimethyl derivative
(determined by nmr) which undergoes photoreversion (vide infra) to 3-methylthymine
and is identical in its spectroscopic and chromatographic properties with the 3 =
methylthymine dimer. This establishes structure XXd for the 3-ine'thylthymine dimer.

Weinblum and Johns32 have carried out a similar treatment to prove the
structures of four different thymine dimers obtained from thymine, thymidine, DNA and

thymidyl (31 •»,5") thymidine.
MECHANISM OF THE FHOTODIMERIZATION

Phot odimerizat ions of the type undergone by thymine have been observed for
other pyrimidine bases.36"38 Waeker38 has attempted to correlate the tendency of
pyrimidines to undergo a phot odimerizat ion with the polarity of the C5-C6 double
bond. Mantione and Pullman39 have pointed out that this argument considers only
the ground state properties of the molecules and does not take into account the
role of the phot oactivat ion. An alternate explanation based on the unpaired
electron density in the C5-C6 double bond in the excited triplet state of the
pyrimidines was offered by Mantione and Pullman to explain the relative order of
dimerization of the pyrimidine bases.

The calculations of Mantione and Pullman, presented in Table 1, were made
using simple Mckel molecular orbital theory. They acknowledge the fact that this
simple theory does not distinguish between excited state singlets or triplets, but
they assume that the calculations nearly represent the thymine triplet.

Table I39

Cone, of odd electrons in
Ability to the C5~C6 bond in the excited

Compound dimerize triplet state.

Orotic acid Good 1.120

Thymine " 1.207

Uracil " 1.252

6~Methyluracil " 1.208

Isocytosine Fair 1.159

5-Aminouracil " 1.053

5-Methylcytosine Weak O.879

Cytosine " O.857

2-Thiothymine Ml 0. 95 3

5-Nitrouracil " O.639

6 -Azathymine 1 . 142

The data in Table 1 show good qualitative agreement with experimental fact
(except for those compounds which contain an additional heteroatom) and while the
results indicate that high electron spin density in the C5-C6 bond promotes
dimerization ^ they do not prove that the triplet excited state is a precursor to the
photodimer.

The reason that the triplet state is favored by Mantione and Pullman seems to
be linked to the observation by Beukers and Berends7^37 that oxygen and paramagnetic
salts decrease the production of photoproducts in the irradiation of liquid solu-
tions of uracil and increase the amount of phot ore version of previously formed
dimers. It is known that paramagnetic substances increase the number of singlet ■>
triplet transitions allowed by increasing the spin-orbit coupling in the molecules
undergoing the transitions.40

Recently Lamola and MLttal41 have studied the dimerizations of thymine and
uracil in acetonitrile. They found that the dimerization of thymine could be
completely quenched by the addition of isoprene (a triplet quencher for thymine).
The uracil dimerization was only partially quenched by isoprene. This work implies
that the dimerization of thymine in acetonitrile proceeds entirely through a
triplet state while the dimerization of uracil proceeds partially through a triplet.
Chromatographically the dimer formed from thymine in acetonitrile is different from
the thymine "ice -dimer" and was tentatively assigned a trans -anti structure.

Lamola and Mittal also investigated the UV irradiation of uracil in liquid
aqueous solution. The major products formed from such irradiations are the uracil
hydration product and uracil dimers. Upon adding 2,4- hexadienol (HDE), an H20
soluble triplet quencher, to the solution, they observed an increase in the ratio of
hydrate to dimers. They could identify two dimers among the photoproducts, one of
which was identical with the uracil "ice -dimer" while the other was a. new dimer.

-268-

In the presence of HDE the ratio of "ice-dimer" to new diraer increased. Thus it
appears that the new uracil dimer is formed through a triplet species while the
uracil "ice -dimer" and the hydration product are not.

An analogous situation occurs in the dimerization of coumarin42 where singlet
state excited coumarin is thought to lead to a cis~"head -to-head" dimer while
triplet excited coumarin gives the trans -"head -to-head" and "head -to-tail" dimers.

D6nges and Fahr43 have recently studied the structure of the uracil "ice-dimer"
and have assigned to it a cis - "head -to -head " structure. Thus it appears that the
dimerization of uracil in liquid aqueous media is controlled in the same way as the
dimerization of coumarin, and the dimerization of thymine in liquid solutions may
be under this same control.

Excited state triplets have been observed both by optical44 and ESR45546
measurements on UV irradiated thymine and DMA samples in H20 methylene glycol glasses
at 77°K. The triplet species found in all the cases studied could be assigned to
the conjugate base of thymine.

Free radicals have also been in UV irradiated samples of thymine and DNA47=49
at 77°K, and here again the radical could be assigned to a thymine species, XXIII.

'CH3

XXIII

These data show that the thymine triplet and free radical are important species
present in UV irradiated samples of thymine and DNA, but they do not implicate
either species directly in the photodimerization reaction.

CONTROL OF THE STEREOCHEMISTRY OF THE THYMINE "ICE-DIMER"

It was mentioned above that one of the isomeric dimers obtained from the
irradiation of a frozen solution of l~methylthymine was assigned its structure on
the basis of the crystal structure of 1-methylthymine.35 Schmidt and co-workers50
have shown that the photodimerizations of olefins in the solid state depends on the
orientation of the molecules in the crystal lattice and occur with a minimum amount
of atomic and molecular motion. Among the examples of this topochemical control is
the dimerization of trans -cinnamie acid. This compound exists in two crystal forms ,
a and p. Irradiation of the more stable, a, form gives only a-truxillic acid, XXIV,
while irradiation of the p form leads only to p-truxinic acid, XXV.51'52

COaH

a„form_
C02H

Jb£orm
hv

CO^H

Wang53"55 has suggested that thymine in frozen aqueous solutions exists in a
high state of aggregation, being excluded from the ice crystals and forced into
clusters of solid thymine. The UV spectrum of thymine in a frozen solution56 most
nearly resembles the spectrum of solid thymine and helps to confirm Wang's theory.

-2o9-

If thymine does exist in a high state of aggregation in a frozen solution,
then the arguments of Schmidt and his co-workers55 would require that the stereo-
chemistry of the dimer formed from the UV irradiation of the frozen solution should
be determined by the crystal structure of solid thymine.

An X-ray analysis of thymine monohydrate57 shows it to have a structure which
can be represented as in Figure 1. If we restrict the photodimerization of thymine
to that course which requires the least amount of molecular movement, it is clear
that the cis-syn isomer will be the only dimer formed. Evidence that thymine exists

Figure 1

5-f

.- -H-^

-CH3

-ch3 y

H H

H3C-U

Crystal Structure of Thymine Monohydrate58

as the monohydrate in frozen aqueous solutions comes from experiments by Wang58
where thin films of thymine were irradiated under conditions of varying humidity.
Table 2 illustrates the importance of humidity on the thymine dimerization, and
Wang regards these data as proof that thymine monohydrate is the species undergoing
the photodimerization.

Table 2

Humidity (#)

Time (hr) 98 71 30 PgOs

1 15.0 17.3 11.4 7.5

2 31.0 30.9 21.0 16.3

3 55.0 55.O 27.O 21.7

io Dimerization of Thymine Under Conditions of
Varying Humid ity( 58) .

FHOTOREVERSION

One of the major reasons for the interest shown in the photodimerization of
thymine is its connection with the biological inactivation of bacteria and phages
by UV light. 19,59~63 It has been well established that some of the effects of UV
radiation may be reversed by some photochemical process,64 and one of the first
properties of the thymine photodimer to be determined was its ability to revert
back to thymine when it was reirradiated with UV light in a liquid aqueous solution.
In fact, it was shown that irradiation of thymine led to the establishment of a
photostationary state between the photodimer and the monomer.65 At short wavelengths
the equilibrium favors the monomer while long wavelengths favor the dimer.

It has also been found that a crude enzyme extract from yeast catalyzes the
splitting of the thymine dimer63'66 or the splitting out of the dimer from the DNA
chain.67

The thymine dimer now has a firmly established place in the area of photo-
damage and photorepair in biological systems.

BIBLIOGRAPHY

1. M. R. Zelle and A. Hollaender in, "Radiation Biology," ed. by A. Hollaender,
McGraw, New York, 1955^ vol. 2.

2. R. L. Sinsheimer and R. Hastings, Science, 110, 525 (1949).

3. C. E. Carter, J. Am. Chem. Soc, 72, I835 (1950).

4. E. Christensen and A. C„ Giese, Arch. Biochem. Biophys., £1, 208 (195*0 .

5. R. L. Sinsheimer, Radiation Research, 1, 5O5 (1954).

6. A. M. Moore and C. H. Thomson, Science 122, 594 (1955).

7. A. Rorsch, R. Beukers, J. Ijlstra, and W. Berends, Rec. trav. chim. , 77, 423
(1958).

8. R. Beukers, J. Ijlstra, and W. Berends, Rec. trav. chim., 77, 729 (I958).

9. B. Jacobson, J. Am. Chem. Soc, 77, 2919 (1955).

10. R. Beukers, J. Ijlstra, and W. Berends, Rec. trav. chim., 79, 101 (i960) .

11. R. Beukers, J. Ijlstra, and W. Berends, Rec. trav. chim., 7H, 247 (1959).

12. R. Beukers and W. Berends, Biochim. Biophys. Acta, 4l, 550 (i960).

13. R. Beukers and W. Berends, Biochem. Biophys. Acta, ¥9, 101 (1961).

14. R. Signer, Ann., 4j8, 2h6 (19.30).

15. L. W. Barrison, J. Chem. Soc, l6lh (I95I).

16. H. E. Ulery and J. R. McClenon, Tetrahedron, 1£, 749 (I963).

17. D. L. Wulff and G. Fraenkel, Biochim. Biophys. Acta., £1, 332 (I961).
17b. H. Dellweg and A, Wacker, Photochem. Photobiol. , 5, 119 (I966) .

18. D. Shugar and G. Baranowska, Nature, 185 , 33 (i960).

19. J. Marmur and L. Grossman, Proc. Nat. Acad. Sci. , 47, 778 (196l).

20. G. M. Blackburn and R. J. H. Davies, Chem. Commun. , 215 (1965).

21. G. M. Blackburn and R. J. H. Davies, J. Chem. Soc. (C), 2239 (1966).

22. R. D. Blatt, J. K. Martin, J. Mo T« Ploesser, and J. Murray, J. Am. Chem. Soc.,
76, 3663 (195^).

23. D. Keglevic and A. Kornhauser, Croat. Chem. Acta, 31, 47 (1959).
2k. R. Anet, Tet. Letters, 3713 (1965),

25. R. Steinmetz, W. Hartmann, and G. 0. Schenk, Ber. , £8, 3854 (1965).

26. Mo Karplus, J. Chem. Phys., 30, 11 (1959).

27. M. Karplus, J. Am. Chem. Soc., 8£, 2870 (I963).

28. E. R. Buchman, A. 0. Reims, T. Skir, and M. J. Schatter, J. Am. Chem. Soc, 64,
2696 ( 19^2) .

29. P. deMayo, S. T. Reid, and R. W. Yip, Can. J. Chem., 42, 2828 (1964).

30. G. M. Blackburn and R. J. H. Davies, Biochem. Biophys. Res. Commun., 22, 704
(1966). —

31. G. M. Blackburn and R. J. H. Davies, J. Chem. Soc. (C), 1342 (1966).

32. D. Weinblum and H. E„ Johns, Biochim. Biophys. Acta, 114, 450 (1966).

33» H. E. Johns, M. L. Pearson, J. C. LeBIanc, and L. W. Helleiner, J. Mol. Biol.,
9, 503 (1964).

34. K. Hoogsteen, Acta Cryst., 16, 28 (1963).

35. R. F. Stewart, Biochim. Biophys. Acta, 7^, 129 (I963).

36. D. Shugar and E. Chargaff in "The Nucleic Acids," ed. by J. N. Davidson,
Academic Press, New York, i960, vol. 3, p. 40.

37. R. Beukers, J. Ijlstra, and W. Berends, Rec. trav. chim., 78, 883 (1959).

38. A. Wacker in "Progress in Nucleic Acids Research," ed. by J. N. Davidson and
W. E. Cohn, Academic Press, New York, 1963, vol. 1, p. 369.

39. J. M. Mantione and B. Pullman, Biochim. Biophys. Acta, 91, 387 (1964) „

40. N. J. Turro, "Molecular Photochemistry," W. A. Benjamin Inc., New York, I965,
chapter 4.

41. A. A. Lamola and J. Mittal, Science, 1^4, 1560 (1966).

42. G. S. Hammond, C. A. Stout, and A. A. Lamola, J. Am. Chem. Soc, 86, JJ03 (1964).

~271-

43. K. H. Dtfnges and E. Fahr, Z. Naturforsch. , 21b, 87 (1966).

44. J. W. Longworth, R. 0. Rahn, and R. G. ShuLman, J. Chem. Rhys., 4£, 2930 (1966).

45. R. 0. Rahn, R. G. Shulman, and J. W. Longworth, J. Chem. Phys., 4j, 2955 (1966).

46. R. G. Shulman and R. 0. Rahn, J. Chem. Phys., 4£, 2940 (1966).

47. J. Eisinger and R. G. Shulman, Proc. Nat. Acad. Sci., 50, 69k (1963) .

48. P. S. Pershan, R. G. Shulman, B. J. Wyluda, and J. Eisinger, Physics, 1, 163
(1964).

49. R. G. Shulman, R. 0. Rahn, P. S. Pershan, J. W. Longworth, and J. Eisinger in
"Electron Spin Resonance and the Effects of Radiation on Biological Systems,"
National Academy of Sciences- National Research Council, Washington, D. C. ,
1966, p. 46.

50. M. D. Cohn and G. M. J. Schmidt, J. Chem. Soc., 1996 (1964).

51. M. D. Cohn, G. M. J. Schmidt, and F. I. Sonntag, J. Chem. Soc., 2000 (1964).

52. G. M. J. Schmidt, J. Chem. Soc, 2014 (1964).

53. S. Y. Wang, Nature, 190, 690 (I96I).

54. S. Y. Wang, Photochem. Photobiol. , 3, 395 (1964).

55. S. Y. Wang, Fed. Proc., 24, Part III, Supp. 15, S-71 (1965).

56. S. L. Davis and I. Tinoco, Jr., Nature, 210, 1286 (I966).

57. R. Gerdil, Acta Cryst., 14, 333 (I961).

58. S. Y. Wang, Nature, 200,~%79 (19^3).

59„ A. Wacker, H. Dellweg, and E. Lodeman, Angew. Chem., 73, 64 (I961).

60. A. Wacker, H. Dellweg, and D. Weinblum, Naturwiss., 4j_, 477 (1961).

61. A. Wacker, J. chim. phys., 58, ll4l (1961).

62. R. B. Setlow and J. K. Setlow, Proc. Nat. Acad Sci., 48, 1250 (1962).

63. D. L. Wulff and C. S. Rupert, Biochem. Biophys. Res. Commun. , 7, 237 (I962).

64. C. S. Rupert in "Photophysiology," ed. by A. C. Geise, Academic Press,
New York, 1964, p. 283.

65. R. Beukers, J. Ijlstra, and W. Berends, Rec. trav. chim., 78, 883 (1959).
GG0 A. A. Lamola, J. Am. Chem. Soc., 88, 813 (1966).

67. R. B. Setlow and W. L. Carrier, Proc. Nat. Acad. Sci., £1, 226 (1964).

March 20, 1967

_07O ..

THERMAL REARRANGEMENTS OF CYCLOHEPTATRIENES1

Reported By W. D. Shermer

Introduction:

Cycloheptatrienes, which also have the common name tropilidenes » have been found
to undergo several thermal and photolytic rearrangements. These isomerizations can be
divided into two genral classes - hydride shifts and skeletal rearrangements. The lat-
ter rearrangements are exemplified by the reactions shown in Figure 1. One example of
the photochemical, rearrangements will be discussed briefly but it is not intended to
be a complete review.

Figure 1

(Eq. 1)

^•73

r-^PK

H ^3^ \_J/ kl3
II

a.) R=H

b) R=D

c) R=4>

d) R=0CH3

Chemical studies involving thermal reactions of 7-substituted cycloheptatrienes
have often been complicated by the production of unexpected side-products resulting
from the isomerization of the substituted tropilidene . 2 f 3 This isomerization could
conceivably occur by either an intermolecular or an intramolecular process or both.
Since equation 2 Is a known reaction,4 a possible intermolecular reaction would be the
initial formation of a small amount of tropylium ion, followed by hydride-ion transfer

from another cyclohepta

(Eq. 2)

triene molecules. To

distinguish between the
inter- and intramolecular
processes, ter Borg and

his coworkers undertook a study of 7-deuterocycIoheptatrIene (l-b) by mass spectroscopy.

Table 1 gives the results of their work. The invar iance of the isotopic distribution

over the range of experimental conditions rules out the possibility of an intermolecular

rearrangement .

Table 1
Distribution of Deuterium After Heating 7-Deuterocycioheptatriene

(percentages)
Starting Material Heated Products

98 98 121 140 140

1700 3297 W) 270 kl£

6.0 6.1 5.9 6.0 6,2 6,1

93.9 93-9 9^.1 93.9 93.7 %•■ .3

Temperature
Time hours
C7H8
C7H7D
C7KSD2
C7H5D3

Calculated Binomial
Distribution Due To
Intermolecular React.

36.8

39.2

-. Q

JLO<

4.9 + D4, D5,etc,

Four intramolecular shifts are possible: ( l) from C-7 to C-l or C-6 (a position)
(2) from C-7 to C-2 or C-5 (3 position), (3) from C-7 to C=3 or C-k (y position),
and (4) a random shift from the 7 position to the a, &, and tf positions. To
elucidate which of the four possible mechanisms is in operation, Nozoe and Tak-
ahasi2 followed the isomerization of 7-methoxycycloheptatriene. The peaks in
the NMR spectrum due to the methyl and C-7 protons (6.65T and 6.83T respectively)
decrease with time at 103°C and eventually disappear completely. Simultaneously,
a singlet at 6.^3T(3H) and a triplet at 7»77f(2H) appear and increase in intensity.
On continued heating, a singlet at 6.50r(3H) and a doublet at 7«55t(2H) appear
and increase in intensity and, on prolonged heating, increase at the expense
of the signals at G.kyx and 7«77^« As equilibrium nears, two new signals appear,
a singlet at 6.58t(3H) and a triplet at 7«83t(2H),> At equilibrium the areas
under the methyl peaks is as follows: 6. ^3t:6. 50r:6. 58t=13. 6:80.5:6.0. Analysis
of the four proposed mechanisms shows that only the 7 to tf hydride shift will
give the sequence of peaks observed for the C-7 hydrogens of multiplet - triplet -
doublet - triplet.

A. P. ter Borg and coworkers3 made a kinetic study of the isomerization
of I-b. It was noted that a shift of a deuterium atom from the 7 position yields
the same compound and, consequently, no first order isotope effects are observed.
Neglecting second order effects, which are generally small, only one rate constant
is involved in the reactions. This constant is defined as the rate constant
for the reaction in which a hydrogen of the CH2 group shifts.; i.e. k73=k37=k31
etc. (for I-b a statistical factor of 1/2 is necessary) . Differential rate
equations were developed to describe the rearrangements taking place for the 7
to a' shift as well as for each of the other three possible mechanisms. Theoretical
calculations, based upon the experimentally determined number of hydrogens in
the 7 position and the length of time of heating, were carried out to determine
the number of protons in the a, (3, and tf positions. These calculated values
were then compared to the experimental values to determine which mechanism best
fit the experimental data. Table 2 gives the results of one such set of calculations.
For broad peaks, the error in integration of NMR spectra can be as high as 10$ >

Table 2
Proton Distribution for 7-Deuterocycloheptatriene
(After heating at i*fO°C for kl:2 hours)
.1 Calculated For Mechanism

Position

Experime

1.66

1.82

1.86

1.72

7">#

7->6

7-X2

7-*Random

(1.66)

1.81

1.86

1.73

1.80

1.86

1.73

1.81

1.80

1.73

1.81

1.86

1.80

decreasing as the peaks get sharper and more intense . 7 No mention is made in
the paper of what was done to minimize this error, Consequently, the significance
of these numbers is in doubt and the tables can not be accepted at full face
value. In their NMR studies, ter Borg and KLoosterziel found the same sequence
of signals, mentioned above in Nozoe1 s work, develop for the proton in the 7
position in I-b and ter Borg cites this as evidence for the 7 to 5" shift. The
data given in Table 2 appear to substantiate this conclusion, but a more positive
statement cannot be made. In accordance with the f- 3" hydrogen shift, ter Borg
proposed the mechanism shown in Figure 2. The multicentered, intramolecular
nature of the mechanism is further substantiated by an absence of solvent effects
on the rates of rearrangement.

Fieure 2

The rate constant for the isomerization of I-b, calculated from the observed
number of protons in the 7> Q> P> and £f positions , was k=6.0 x 10 7 sec x at
121°C. It has been observed that the rearrangement of I to II is always accelerated
by substitution. For I-d at 121°C , Nozoe and Takahasi2 found k73=2.48 x ICf5
sec"1 (extrapolated), approximately 100 times as fast as the unsubstituted tropilidene.
At 120. 2°C, Nozoe reports k73=2»31 x 1.0 5 sec x (extrapolated) for I-d and ter Borg
reports )^73-J).G0 x 10 b sec -1 for I-c,6 7~phenylcycloheptatriene, The rate constants
reported for I-d were based on MR measurements and, since no standard deviation
is reported, these numbers should be qualified in the same manner as the MR work
by ter Borg. The rate constant for I-c was based on UV measurements and should
be quite accurate but, again, no standard deviation is reported.

Further Information on these three compounds may be gained by comparing their
activation parameters (Table 3) • These values are based on MR calculations with
no standard deviation given and, therefore, can only be taken as approximate
values. The large negative AS- indicates that this reaction may proceed through
a transition state such as that pictured by ter Borg. The decrease in activation
energy in going from I-b to I-c possibly reflects the amount of conjugation
energy due to the trienic and aromatic systems in the transition state. If this

Table 3
Activation Parameters

Compound AEa(Keal. /mole) AH-(Kcal./mole) AS-(e.u.)

I-b3 31 30.2 =8.2

I-c 27.6 26.9 -11.7

I-d9 26.4 25.7 =15.0

conjugation exists, it would require some co-planarity between the phenyl group
and the trienic system in the transition state. This should result In a more
negative AS-^ which is observed experimentally. Nozoe ascribes the more negative
AS^ for I-d as compared to I-c to some contribution of the lone pair of electrons
to the ragidity of the transition state. However, Nozoe is trying to explain
an extremely small difference between two numbers , of which only the value for
I-c can be considered significant since it is based on UV measurements and not
on MR data as is the AS- for I-d. Consequently, this explanation is highly
speculative and, as Nozoe admits,2 more work is necessary. This same problem
exists in trying to explain the AAS- for I-b and I~c. While the difference can
only be regarded as the approximate value, when considered in the light of the
mechanism proposed by ter Borg and the values of the Arrhenius activation energies,
it is at least indicative of the difference which does exist between the two
entropy values.

The equilibrium concentrations of a series of R groups was studied by ter Borg
and coworkers. In compounds III, II, IV, and 1 the R group is in conjugation
with 3; 2, 1, and 0 double bonds of the trienic system respectively and their
respective concentrations reflect the ability of the isomer to be stabilized through
conjugation as the electron donating tendency of R increases, isomer III predominating.
If R is other than an electron-donating group, no overwhelming preference of one
isomer over another is shown.

Equilibrium Concentrations8

III

IV I

T °C

N(CH3)2

100

„_

100

OCH3

3 »

120

SCH

8 —

11.4

CH3

17 2

140

C6H5

1.8 -

136

2k —

142

The mechanism proposed by ter Borg for the rearrangement involves a 1,5
hydrogen shift with migration of two double bonds, the remaining double bond not
participating. If this hypothesis is true, 1,3-cycloheptadienes should also
undergo rearrangement. Kloosterziel and ter Borg10 prepared 2^7-aihydrotopone (v)and

!Eq. 3)

•275-

found that it rearranged to
2 , 3-dihydrotropone ( VI ) . The
rate of isomerization was followed
via UV and the equilibrium constant
calculated. At 60°C ;. K=k1/k„1=l,86-
kj^l.texlO""6 sec"1 and k,.1=7»6 x 10 7
sec 1. These rate constants include
no significant solvent effect s
at 60°c ki + k~i=2.l8 x 10 6sec 1 in n-heptane and 4. 5 x 10 6sec -1 in ethanol.
The slight solvent effects which do appear may be due to the solvation of the
carbonyl group. At 101°C, ki=l.l6 x 10 4sec 1 or approximately 104 times k73
for tropilidene itself. There is no spectral evidence for the presence of an
enol and, therefore, a maximum of only a few percent may be present. As a result 3
formation of the enol and subsequent hydride transfer is ruled out by ter Borg,
who proposes that the small amount of enol which might be formed would be insufficient
to account for the large rate enhancement. Without knowing any rate data for the
enol, this is not a well-grounded assumption. If the rate of reaction of the
enol was very rapid, in conjunction with a steady state approximation,, a few percent
enol could easily account for the rate enhancement. It would be necessary to
show that enolization catalysts , such as acids, do not affect the rate of reaction
to prove that enolization does not cause the rate enhancement.

The predictions of Woodward and Hoffmann11 also support the 1,5 hydride shift.
Due to the constraints imposed by the ring, suprafacial hydride shifts are the
most likely transfer mechanism and, thermally, only the 1,5 shift would be allowed
to proceed in a suprafacial manner; the 1,3 and 1,7 shifts would be antarafacial.
On the other hand, photochemically, the 1,3 and 1,7 shifts are allowed to be
suprafacial and the 1,5 shift would be allowed in an antarafacial manner,
Razenberg and ter Borg12^13 found that the 7-substituted tropilidenes do indeed
'undergo a 1,7 shift in photolytically induced rearrangements. Thus all the work
carried out substantiates the proposed 1,5 hydride shift in thermal isomerlzations.
Skeletal Reorganizations
A) Norbornadiene-Cycloheptatriene Isomerization

W. G. Wood14 studied the isomerization of norbornadiene to form cycloheptatriene
and found that toluene and the products of the reverse Diels-Alder reaction,
acetylene and cyclopentadiene , were also formed. The mechanism shown in Figure 3
was considered because small amounts of benzene and ethylene were found as side=
products and carbene is known to add to benzene to give tropilidene and toluene.

+ HoC:

Figure 3

VII

The norbornadiene was pyrolysed in the presence of a large excess of n-butane
in an attempt to trap the carbene. However, no trace of pentanes or benzene
were found while the isomerization and the reverse Diels-Alder proceeded to
give the expected yields. Two controls were run to demonstrate that carbene does
react, under the experimental conditions, with n-butane to give pentanes and
with benzene to give toluene. Wood, therefore, concluded that this was not the
correct mechanism and proposed the mechanism shown in Figure h.

<r—>

il ^

Figure h

Lustgarten and Richey15 studied the rearrangement of 7-phenyl and 7-alkoxy-
norbornadienes to cycloheptatrienes and found only the 1, 2, and ^-substituted
tropilidenes, no aromatic isomers where found. No significant solvent effects
were observed which tends to rule out charge separation in the transition state
and indicates that an intramolecular process is possible,,

(Eq. k)

VII

(in n-decane;

-$■

R=H

R=D

R=<1>

R=OMe

R=OC(CH3)3

R=0C2H.4QC2H5

AHi=34,5 Kcal./mole, A&^=1 e,u. for Vll-e)

When R=H, the isomerization occurred at ^75°C and toluene, cyclopentadiene ,
and acetylene were produced as well as tropilidene. When R was phenyl or alkoxy,
the isomerization was carried out at only 175°C and only the substituted tropilidenes
were observed. In contrast, when the methylene carbon had a ketal or thioketal
substituent attached to it, the isomerization, which can be carried out in the
same temperature range, produced only benzene and compounds derived from the ketal
or thioketal function and no tropilidenes were found. Using a capillary furnace
mounted at the inlet of a mass spectrometer, Lemal and coworkers16-5'1'7 detected
the presence of *CH3, C02, and CH3C02CH3 when 7>7-&Imethoxynorbornadiene was
pyrolysed. They proposed that benzene and dimethoxy carbene were initially
produced and the carbenes then decomposing by a radical chain mechanism to give
the products observed. To account for the formation of benzene and the dimethoxy
carbene in a manner consistent with a general mechanism for the various substituted
norbornadienes discussed above, Lustgarten and Richey proposed the intermediate
shown in equation 5« Using this general mechanism, the equilibrium shown in
equation 6 should exist for the mono substituted norbornadiene but the fragmentation

H3CO

(Eq. 5)

(Eq. 6)

+ (CH30)2C;

-2(7-

of the norcaradiene intermediate would result in a less stable carbene than in
the dimethoxy case and, therefore might be slower than the competing 1.5-hydrogen
transfer. However, direct loss of dimethoxy carbene without intermediates cannot
be ruled out, and since tropilidenes are not known to decompose to give carbenes,
Lustgarten and Richey have no evidence for the intermediates in equation 5.

Herndon and Lowry18 studied the kinetics of the isomerization of norbornadiene
to cycloheptatriene to determine if the toluene produced was generated directly
from the norbornadiene or from the cycloheptatriene or both. (See Figure 5.)
They used a gas phase stirred flow reactor designed so that the contents of the
reactor are completely mixed by diffusion, thereby leading to uniform concentrations
which become time -invariant within the reactor.

( BCH)

( CPD)

The experimental results
show that toluene is produced
from both compounds. This
result was indicated in some
previous work by Klump and
Che sick19 but not proven
conclusively. Table k gives
the data obtained by Her,don
and Lowry.

Figure 5

Table k
Rate Constants and Arrhenius Activation Energies

( 1) BCH -> CPD + A

(2) BCH -> CHT

(3) BCH -» T
( h) CHT -> T

k (sec x)

(400.6°C)
2.31 x 10~2
1.73 x I0~2

1.04 x l(f3

1.05 x 10" 3

££a (calo/mole)

50,190 + 76O
50,610 + 78O

53,1^0 ± 730
52,100 + 820

Herndon and lowry claim all of the reactions are first order and

unimolecular and present the reaction scheme shown in Figure 6 as the most likely

mechanism. The symbol [I] represents a common intermediate to tropilidene and

toluene. Birely and Chesick20 have also examined these reactions and have obtained

^ CHT

BCH > [Ij

CPD + A " T

Figure 6

almost identical values for the various rate constants and energies of activation.
Since the values of the activation energies for all of the equations in Table k
are almost identical, Birely and Chesick claim that this is evidence for a common
intermediate to all three products: cyclopentadienes (+ acetylene) , cycloheptatriene,
and toluene. Herndon and Lowry18 dispute this conclusion and argue that this
is evidence against a common Intermediate and is evidence only for similar initial
steps. If a common intermediate for the reverse Diels-Alder reaction and the
isomerization did exist, the ratio of the rate constants can be shown, to be the
ratio for the reactions which take place after the common intermediate is formed.
An Arrhenius plot of this ratio gives the difference in activation energies. In
this case the difference is extremely small, 400 cal. , and this would Indicate that
the two processes occurring after the intermediate is formed are essentially similar.
But the isomerization and reverse Diels-Alder reactions are not similar. Thus, no
common intermediate exists and it is simply fortuitous that the activation

-270-
energies are almost identical. Using this same approach, it might also be argued
that the similarity of the activation energies after the common intermediate is
fortuitous, thereby allowing no distinction to be made. Tropilidene has not,
as yet, beerj reported as a side product in the Diels-Alder synthesis of norbornadiene
as might be expected if a common intermediate did exist and this has been offered
in support of the theory of Herndon and Lowry. However,, based upon the difference
in the heats of formation ^35^36 formation of norbornadiene is favored by approximately
8 Kcal./moie and, consequently, it may be formed with complete exclusion of topilidene.

B) Norcaradiene - Cycloheptatriene 1 somer iz ation °

Some chemical reactions of tropilidene often give products which appear to
come from norcaradiene^9 °21 while others give products involving the cycloheptatriene
structure. This ambiguous nature would make it appear that tropilidene is in
equilibrium with its valence tautomer - norcaradiene.

Anet22 studied the temperature dependent MR spectra of tropilidene to determine
if it is planar or nonplanar. At -150°C, the methylene protons give rise to two
chemically shifted bands with a separation of 76 cps and increasing. The mean.
chemical shift is 7»8t, essentially unchanged from that observed at higher
temperatures. Since the methylene protons are nonequivalent at low temperatures,
cycloheptatriene is nonplanaro In similar work, Jensen and Smith23 were able to
get down to -170°C and the separation, between peaks was 86 cps. They were not
able to find any evidence for the presence of norcaradiene. An electron diffraction
study by Traettenberg24 showed that tropilidene was indeed nonplanar with the
plane comprised of the 1,2,5^ and 6 carbons making an angle of 36. 5° with the
plane of the 1,6, and 7 carbons and an angle of 40.5° with the plane of the
2,^,k, and 5 carbons.

An equilibrium between tropilidene and norcaradiene should be detected by
variations in the coupling constants with changes in temperature,, Roberts and
coworkers25 carried out a complete analysis of cycloheptatriene at -70°C and
+115° C but no change in coupling constants could be observed. Consequently,
there is less than 5$ norcaradiene present even at -7C°C. When 7^7-bistrifluor-
omethylcycloheptatriene was studied,25 the trifluoromethyl groups remained equivalent
down to -l8.5°C and there is no evidence of any of the norcaradiene structure.

Both the trifluoromethyl group and the cyano group are strongly electron
withdrawing, having negative resonance effects and negative inductive effects.26
The Hammett para-sigma constants are 0..54 and 0.66 respectively.2^ Yet, when one
of the CF3 groups is replaced by a CN group, some norcaradiene is observed.28 At
-85°C, the relative concentration of the substituted tropilidene to substituted
norcaradiene is 80s 20 as calculated from WMR spectra peaks. When both trifluoromethyl
groups are replaced by cyano groups, only the 7>7~dicyanonorcaradiene (VIII)
is observed.29 At 100°C, VTII rearranges to phenylmalonitrile and 3,7-dicyano-
cycloheptatriene (IX) which in turn undergoes a series of 1,5 hydrogen shifts
to give 1,4- and l,5~dicyanocycloheptatriene upon further heating.

C) 7 >7-Di substituted Cycloheptatriene s ;

Berson and Willcott'-' examined the norcaradiene-diradical intermediate
equilibrium proposed by Wood (Fig. k) in an attempt to determine if the rate of
recyclization of the diradical intermediate to form norcaradiene occurred at
a rate competitive with hydrogen transfer to form toluene from the diradical
intermediate. Blocking the 1,5-hydrogen transfer by di substitution at the 7
position, 3,7^7-trimethyltropilidene (x) was pyrolysed at 300°C to produce the
profusion of products shown in equation. 7; the percentages given refer to the com-
position of the reaction mixture after pyrolysis for 40 minutes, but do not include
1 - 2$ of unidentified products.

-27S-

H3<^3?5<CH3 H3VCH2 H3° CH2 H3Cf,CH3 Ife= ^CH3

it. f\ + fi ;aQ^ ^

1=/ HsC^^^R

H*C

X(35^) Xl(25#) Hl(2l) XIII(2<#) XIV(l6$) XV(2$) (XVI) (2^)

a) R=H b) R=D

The mechanism of the isomerization has been shown by deuterium labeling to
be a true skeletal rearrangement of the ring carbons rather than a superficial
series of hydrogen shifts. Examination of equation 7 shows that the 1-6 carbon
chain maintains the same sequence in each compound , while the C-7 carbon and
its geminal methyls are allowed to wander and reattach between any pair. Both
nonaromatizing rearrangements to form XI and XIV are reversible. When either XI
or XIV is isolated and resubjected to the reaction conditions, a typical mixture
of pyrolysis products results. Mass spectroscopy studies have shown that the
rearrangements do not take place by an intermolecular process. The mechanism
shown In Figure 7 , proposed by Berson and Willcott,32 incorporates the major structural
changes, the intramolecularity, and the reversibility. This mechanism indicates
that, compounds XVII, XVIII, and XIX should be formed. So far these products
have not been found but may be present in the 1-2$ of unidentified material
or may be formed after a longer period of time than Berson and Willcott ran their

XVIII

Figure 7

XIX

experiments. Production of the dienes XIII and XIV is not without precedent.
Employing the norcaradiene intermediate, this isomerization can be classified
as a Cope type rearrangement with one double bond being replaced by a cyclopropyl
ring and a hydrogen from one of the geminal methyls being transfered to the six
member ring.37-*38 There is no way to distinguish a diradical mechanism from
a concerted 1,5-carbon shift without intermediates. This latter process is
permitted, but not required,11 and requires the C-7 migration from C-l to C-5
to be suprafacial. An. examination of properly substituted optically active
tropilidenes should distinguish this from an antarafacial process and from a
mechanism involving a diradical intermediate. This work is currently in progress.

-280-

Fhotochemical Isomerization:

Chapman and Borden^'J found that Irradiation of neat 7-alkoxycycloheptatriene
produced mainly the substituted bicyclo[3.2.0]heptatriene (XX-b1) and only trace
amounts of toluene. If this irradiation is carried out in the vapor phase , both
XX-b1 and 1-alkoxycycloheptatriene (ill-b') are produced. Pyrolysis of XX-b'
produces Ill-b' and irradiation of Ill-b1 converts it to XX-b1. Srinivasan34
found that irradiation of tropilidene in the vapor phase produces mainly toluene
and a maximum of 5$ bicyclo[3.2.0]heptadiene (XX-a). The formation of neither
toluene nor XX-a was quenched by addition of oxygen or nitric oxide , which indicates
that the isomerization is intramolecular and that it does not arise from a triplet
state of tropilidene. Srinivasan has postulated that toluene is formed from a
vibrationally excited ground state and not from an upper electronic state. The
production of the isopropyl-toluenes upon pyrolysis of 3,7,7-trimethyltropilidenes
shows that this reaction also occurs thermally and this particular photochemical
example was included because of this correspondence.

Bibliography

1) W. G. DeWitt, Univ. of 111. Organic Seminar Abstracts, 1963, P. 19.

2) T. Nozoe and K. Takahashi, Bull. Chem. Soc Japan, ^8, 665, (1965).

3) A. P, ter Borg, H. KLoosterziel, and N. van Meurs , Rec. Trav, Chim., 82, 717, (1963),
k) Z. N. Parnes, M. E, Volpin, and D. N. Kursanov, Tet. Let., No. 2.1, 20, (i960).

5) A. P. ter Borg, H. KLoosterziel, and N. van Meurs, Pro. Chem. Soc, 359, (1962).

6) A. P. ter Borg and H, KLoosterziel, Rec. Trav. Chim., 82, 74l, (1963).

7) J. A. Pople, W. G. Schneider, and H. J. Bernstein, "High Resolution Nuclear
Magnetic Resonance," McGraw-Hill, New York, N.Y. , 1959, P. 78«

8) A. P. ter Borg, E. Razenberg, and H. KLoosterziel, Rec. Trav, Chim., 82, 1230,(1963)

9) E. Weth and A. S. Dreiding, Pro. Chem. Soc, 59, (1964).

10) A. P. ter Borg and H. KLoosterziel, Rec Trav. Chim., 82, II89, (1963).

11) R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc, 87, 2511, (1965).

12) A. P. ter Borg and E. Razenberg, Rec. Trav. Chim., B4, 24l, (1965).

13) A. P. ter Borg and E. Razenberg, Rec Trav. Chim., BE, 245, (1965).

14) W. G. Wood, J. Org. Chem., 23, ilO, (1958).

15) R. Ko Lustgarten and H. G. Richey, Jr., Tet. Let., 4655, (1966).

16) D. Mo Lemal, R. A. Lovald, and R. ' W. Harrington, Tet. Let., 2779, (1965).

17) D. M. Lemal, E. P. Gosselink, and S. D. McGregor, J. Am. Chem. Soc, 88,582,(1966).

18) W. C. Herndon and L. L. Lowry, J. Am. Chem. Soc, 86, 1922, (1964).

19) K. N. KLump and J. P. Chesick, J. Am. Chem. Soc, 8£, 130, (1963).

20) J. H. Birely and J. P. Chesick, J. Phy. Chem., 66^ 568, (1963).

21) W. von E. Doering, G. Laber, R. Vonderwahl, N. F. Chamberlain, and R. B.
Williams, J. Am. Chem. Soc, 78, 5448, (1956).

22) F. A. L. Anet, J. Am. Chem. " Soc. , 86, 458, (1964).

23) F. R. Jensen and L. A. Smith, J. Am. Chem. Soc, 86, 956, (1964).

24) M. Traettenberg, J. Am. Chem. Soc, 86, 4265, (19^5).

25) J. B. Lambert, C. J. Durham, P. Lepoutere, and J. D. Roberts, J. Am. Chem.
Soc, 87, 3896, (1965).

26) E. S. Gould, "Mechanism and Structure In Organic Chemistry," Holt, Rinehart,
and Winston, New York, N. Y, , 1959, P. 218.

27) J. Hine, "Physical Organic Chemistry", McGraw-Hill, New York, N. Y. , I962, P. 87.

28) E. Ciganek, J. Am. Chem. Soc, 87, 1149, (1965).

29) E. Ciganek, Private communication,

30) J. A. Berson and M. R. Willcott, III, J. Am. Chem. Soc, 87, 2751, (1965).

31) J. A. Berson and M. R. Willcott, III, J. Am. Chem. Soc, B?, 2752, (I965).

32) J. A. Berson and M. R. Willcott, III, J. Am. Chem. Soc, B8, 2494, (1966).

33) 0. J. Chapman and G. W, Borden, Proc Chem. Socl, 221, (1963).

34) R. Srinivasan, J. Am. Chem. Soc, 84, 3432, (1962).

35) A. F. Bedford, A. E. Beezer, C. T. Mortimer, and H. D. Springall, J. Chem. Soc,

3823, (1963).

36) H. L. Finke, D. W. Scott, M. E, Gross, J. E. Messerly, and G. Waddington,
J. Am. Chem. Soc, 78, 5469, (1958).

37) P. J. Ellis and II. M. Frey, Proc Chem. Sou. 221, (1964).

38) G. Ohloff, Tet. Let., 3795, (1965).

CYCLIZATION REACTIONS OF N-HALOAMINES, -AMIDES, AND --IMINES

Reported by Daniel R, Bloch March 2j5, 1967

INTRODUCTION

There have been many reports in the recent literature concerning the use of
nitrogen free radicals in organic synthesis.1 Although cyclization reactions of
N-halogenated nitrogen compounds have been known for a long time,1 only recently
has the mechanism^2"6 the influence of solvent and structure on the efficiency of
the process, 1,2>7 9 and addition of nitrogen radicals to unsaturated compounds been
studied.7"13 Preference for formation of pyrrolidines and 7 -lactones from the
photolysis of N-halo compounds has been well documented. 1"4>8 Readily obtainable
starting materials, good yields and relatively simple reaction conditions are
characteristic of these reactions.

N-HALOAMINES

The first cyclization reaction involving N-haloamines to appear in the lit-
erature was reported by Hofmann in I883.13 The reaction of N-bromoconiine (I) in
hot sulfuric acid and subsequent treatment with base afforded B-coneceine (II).

1) H5SQ4.UK)0

2) OH" '

I II

No further work appeared until the early 1900' s when Loeffler and co-workers reported
further examples of cyclization reactions of N-haloamines including an elegant
synthesis of nicotine (III).14 Thus, reactions of this nature have been called

1) H2S04,100C

2) OH'

III

Hofmann-Loeffler (H-L) reactions, although other names have been used.1 In a review
by Wolff,1 a table of reported H-L reactions has been compiled which is complete
through the 1950' s and describes reactants, products and reaction conditions.

Wawzonek3*5 was the first to study the mechanism of the H-L reaction and later
work by Lukes6 and Corey4 have shown the reaction to be a free radical chain process.
The following mechanism has been proposed.

1+ RpNHCI (5) ?© +'
— » -RI-H ^ > Cl-R-N-H + RsNH

or R2NCI

RsNHCl

+ •

RsNH

+ CI'

OH
Corey4 has shown that a 0. J>hU solution of N-ehlorodi-n- (I)

butylamine (NCBA) in 85/0 sulfuric acid at 25° is stable

in the dark for 285 minutes, although decomposition R-N-R

could be induced by the addition of ferrous ion. {5)

Photolysis of the NCBA solution showed an induction

period which could be essentially removed by purging the solution with nitrogen.

The reaction could be interrupted by removing the light source and started

immediately by irradiating again. These features are characteristic of radical

chain react! ^1=

-282 -

Corey4 prepared the N-chloro compounds by passing chlorine gas over a ligroin
(60°-90°) solution of the amine. The resulting solution was washed with dilute
acid and dilute base before extracting the chloroamine into 85/0 sulfuric acid. For
reactions in anhydrous solvent, the ligroin solution was washed, dried, and con-
centrated in vacuo. An aliquot of the residue was taken up in absolute acetic acid.
The amount of N-chloroamine assumed present was based on the total amount of active
chlorine as determined by titration.

H-L reactions involving acyclic reactants generally lead to pyrrolidines1'4'8
which are formed (eq 1) by a cyclic mechanism involving the nitrogen and a hydrogen
on a 8 -carbon. The predominance of hydrogen removal at the 6 -carbon favors an
intramolecular hydrogen abstraction which occurs preferentially thru a quasi -
six-membered transition state. If the process were intermolecular, one would expect
more random hydrogen abstraction giving a greater variety of products.

Radicals which abstract hydrogen atoms from carbon generally show a preference
for hydrogen in the order tertiary /> secondary )> primary. This same order of
reactivity is followed for H-L reactions. In the free radical decomposition (85$
sulfuric acid, 95°) of N-chlorobutylamylamine (IV), two products are possible from
hydrogen abstraction at a 6-carbon. l-n-Butyl-2-methylpyrrolidine (V) would result
from secondary hydrogen abstraction and 1-n-amylpyrrolidine (VI) would be formed
from primary hydrogen abstraction. The fact that (V) was the only tertiary amine

CH3CH2(CH2)35J(CH2)3CH3 I H [~ ~J

CI w CH3 N

n-C4H9 n-C5Hi:L

IV V VI

isolated shows the preference of secondary over primary hydrogen abstraction. For
a comparison of the reactivity of tertiary and secondary hydrogen and of tertiary
and primary hydrogen, the N-chloro

derivatives of n-butylisohexylamine and n-amylisohexylamine were prepared and
subjected to H-L conditions. No tertiary amine could be isolated although the
disappearance of N-chloroamine was very rapid and accompanied by the evolution of
hydrogen chloride. It was suggested that the tertiary chloro compound was formed
and rapidly solvolyzed in strong sulfuric acid. It was also found that t-butyl
chloride liberated hydrogen chloride when shaken with 85$ sulfuric acid.

The stereochemistry of the H-L reaction was studied by thermally decomposing
the N-chloro derivative of ( -) -methylamylamine-4-d (VII) in sulfuric acid at 95°.
The products, 1,2-dimethylpyrrolidine (Villa) and l,2-dimethylpyrrolidine-2-d
(VHIb) , isolated in kyf> yield were optically inactive. An isotope effect (%/kjj)

Jt-h(d)
:h,

CH3

"T^fe

VII Villa ,b

of 3.5^+ was observed for the reaction. This result strongly suggests the decom-
position involves an intermediate in which the 8 -carbon is trigonal.

As evidence that a 6-chloro compound is an intermediate, Corey4 treated the
solution resulting from photolysis of NCBA with silver ion. Practically no silver
chloride precipitated. When the resulting solution was made basic, silver chloride
was obtained in ca. 99$ yield. Reaction under thermolytic conditions gave a 65/0
yield of silver chloride. This suggests that the unreactive chlorine was bound to
a carbon atom prior to hydrolysis. Since basification resulted in cyclization to the
5-carbon and freeing of the chloride, it is reasonable that the chloride was bound
to the 5-carbon. Wawzonek5 was successful in isolating the 4-chloro derivative in
377o yield from the decomposition of NCBA in sulfuric acid.

-28j-

Recent work by Neale8 has shown that side reactions may also be important in the
photolytic decomposition of NCBA. He studied the reaction with respect to a) acidity,
b) degree of purity of chloroamine, c) applied irradiation and d) the rate at which
nitrogen swept the reaction mixture. Most decompositions were run at 20° in acetic
acid 1..5M in water while the molarity of sulfuric acid was varied (Table I).

Table I Photolytic Rearrangement of N-Chlorodi-n-butylamine

■ry '

Molarity

Source

H2SO4

BusjNCI

Bu^NCl

0.223

0.5

0.260

1.0

0.255

2.5

0.243

1.5

0.44

U-l

1.5

o.46

U-2

1.5

0.46

U-l

1.9

0.46

U-l

3.9

0.46

U-l

1.5

0.46

U-2

1.1

1.9

0.46

U-2

3.9

0.46

U-2

1.9

0.46

3.9

0.46

7.7

0.46

1.0

0.46

0.97

0.48

U-l

1.0

0.25

U-l

2.5

0.24

U-l

Entries

3 1=4,

ref 4^ entries

5-19, :

Decomposition

2910
62
52

^7
0.81
1.62

O.96
0.37
1.08
O.98
O.54

0.43
0.24
0.16
0.05
0.10
0.22
0.14

Yield of N-butyl-
pyrrolidine

59
80

42
41
56
88

8~7

95
49
17
60

ref 8. Irradiations entries 1-4, quartz
lamp, range 200-400 mu^ entries 5-19* Hanovia mercury arc lamp with filter,
transmition 300-400 rnu. ^Nitrogen flow? entries 1-4, reaction run under nitrogen ^
entries 6-19> nitrogen bubbled slowly through solution j entry 5* nitrogen rapidly
bubbled through solution. Solvents entries 1-4, 18 and 19 run in anhydrous acetic
acid 5 entries 5-17 run in acetic acid 1.5M in H20. Entries 1-4, prepared as in ref
15. ^Determined by loss of active chlorines entries 1-4 are half -life values *j
entries 5-19 are in mmoles of chloroamine consumed per min. , which is constant for
O-80/0 of the reaction, except entry 17:0-30$ (ref 27).

The chloroamine was prepared from NCS and amine in ether by stirring for one hour,
washing with water and dilute sulfuric acid, drying and evaporating the solvent.
Reagent grade NCS and amine gave product ( U-I) , recrystallized NCS and amine gave
product (U-II) and distilled U-I gave product (D) . Increased chloroamine purity
increased yields of N-n-butylpyrrolidine (NBP) and decreased the rate of reaction
(Table I).

Wnen the intensity of the irradiation was decreased or lower wave lengths
filtered out, the rate of reaction was suppressed and the yields of NBP were decreased.
The rate of flow of nitrogen bubbled through the reaction mixture also determined the
rate of reaction. Increased rate of flow increased the yield of KBP but decreased
the rate of reaction. Trapping spent nitrogen showed volatile substances were being
removed from solution by the nitrogen.

In the dark, acid solutions of distilled and undistilled chloroamine were
unstable and UV spectra showed formation of a new compound with X^^ at 306 m^.. At
a given sulfuric acid concentration in acetic acid 1.5M in water, the rate of formation
of this new compound in the dark was the same for distilled and undistilled chloroamine.
The rate was also found to be dependent on sulfuric acid concentration. In a solution
1M in sulfuric acid decomposition is rapid whereas in a solution 4M in sulfuric
decomposition is very slow.8 The absorbing species was proven to be volatile by

*in min.

sweeping it in a nitrogen stream from a weakly acidic solution of undistilled chloro-
amine into an acetic acid trap. Extraction with pentane gave a solution with
absorption at 307 mn. The species responsible for this absorption was shown to be
N,N-dichloro-n-butylamine by comparison with a sample formed from NCS and butylamine
( ether g ^
v max ^ ' max "^

Photolytic decomposition of NCBA was followed by periodically taking the UV
spectra of aliquots. When chloroamine of purity U-I and U-II was photolyzed in
aqueous acetic acid, new variable absorption appeared in the region 312-320 mu which
grew to a miximum near 60-70% reaction and disappeared when the active titer fell to
zero. A less pronounced maximum appeared in the region 320-340 mu in anhydrous
acetic acid. When distilled chloroamine was used the new absorption which appeared
(X1Tiax 3O6 mu) remained constant. In solutions of excess strong acid (7.7M acid:0.46M
chloroamine) no new absorption appeared during reaction. It is apparent that
dichloroamine is formed during photolysis of chloroamine solutions although it is
unstable under the reaction conditions and does not accumulate in solution. At
lower sulfuric acid concentrations, dichlorobutylamine formation may compete favorably
with the H-L reaction which could account for decreasing yields of pyrrolidines with
decreasing acid concentration (Table I).

Corey found that the rate of decomposition of chloroamine increased with
increasing sulfuric acid concentration (Table I, entries 1-4) while Neale stated his
results (Table I, entries 8-1J?) were nin direct contrast" to Corey's and that the
rate decreased with increasing acid concentration. Neale also reported that in
solutions c_a. 0. 5M in chloroamine, reactions are quite slow when the molar ratio of
sulfuric acid to chloroamine was 1:1 or 2:1 (Table I, entries 16 and 17). At low
acid -amine ratios, it appears that both researchers' data indicate an increase in
rate with an increase in sulfuric acid concentration. Neale' s values support this
conclusion up to an acid -chloroamine ratio of 3:1.

Although the mechanism of the H-L reaction is generally agreed upon, the
initiating species is subject to controversy. Wawzonek proposed that the protonated
N-chloroamine is the initiating species since the chloroamine exists mostly in the
protonated form in sulfuric acid solutions. Corey4 suggested that unprotonated N-
chloroamine is the initiator. Protonated NCBA shows no appreciable absorption above
225 mu, whereas free amine absorbs at higher wave length (A^£c267, e . 320) 17 (Fig I).

Fig I. N-chlorodi-n-butylamine:
(1) 2.65 x 10~3M in HOAc, cell
1.0 cm j (2) I.98 x 10~3M in CCI4,
cell 1.0 cm j (3) 0.46M in H2SO4-
1.5M H20-H0Ac; cell 0.105 cm; (4)
0.46M in I.5M H2SO4-I.5M H20-
HOAc, cell 0.105 cm.
N,N-Dichloro-n-butylamine : ( 5)
1.86 x 10~3M in ether, cell 1.0 cm,
from NCS and butylamine; (6)
I.56 x 10~3M in pentane, cell
1.0 cm, isolated from reaction
mixture; (7) absorbance of 2 mm.
thickness Pyrex glass vrs. air.

Absorbance

325 mu

-285-
Corey's data show that the rate of reaction increases with increasing sulfuric acid
concentration. If free chloroamine is the initiating species, its concentration
should decrease with increasing acidity and, hence, decrease the initiation rate.
Thus acid catalysis must involve acceleration of the propagation process and/or
retardation of chain termination. It seemed likely to Corey that strong acid should
inhibit chain termination. The interaction of two protonated, positively charged
radicals by coupling or atom transfer would be slower than for neutral species,
especially if the ions were solvated. The photolytic decomposition of NCBA in carbon
tetrachloride18 and anhydrous acetic acid,2'4 where amine is unprotonated , is very
slow relative to the acid catalyzed reaction (Table I, entry 1). There are at least
three possible explanations why the reaction is slow: a) Neale's suggestion that the
unprotonated chloroamine is a poor initiator, b) Corey's proposal that chain termina-
tion is more favorable for unprotonated nitrogen radicals, and c) an acid catalyzed
propagation sequence which has not yet been defined.

Neale8 suggested that the N,N-dichloro-n-butylamine is the initiating species
in the H-L reaction, since it is formed and decomposed during photolysis (^g^ 3^6 >
e 320). He further argued that if free N-chloroamine were the initiating species,
its low concentration and extinction coefficient would require it to absorb light
and dissociate very efficiently (Fig I).

N-HALOAMIDES AND N-HALOIMIDES

The mechanism proposed for the photolytic decomposition of N-haloamides and N-
haloimides is analogous to that proposed for the H-L reaction (eq 1).11,19,2° These
reactions are initiated by light and peroxides and are inhibited by bubbling oxygen
through the reaction solution. One notable exception is that, in general, acid is
not needed to catalyze the reaction. Although reaction conditions vary, hydrogen
abstraction at the /-carbon is predominant, with subsequent formation of 7 -lactones.
Barton19 has done the most extensive mechanistic study to date. In a search for a
general method of forming saturated lactones from saturated acids, photolysis of N-
iodoamides afforded 7-iminolactones which could be hydrolyzed to 7-lactones.
lodination of 33-acetoxy-ll-oxo-5a:-pregnane-20-carboxamide (IX) with lead tetra-
acetate and iodine in benzene gave c_a. 55/^ yield of lactone upon alkaline hydrolysis.
In a similar manner, orthotoluamide gave phthalide and stearamide gave 7-stearo-
lactone. The following mechanism was proposed: /)

CH3CO

Fb(0Ac)4-I.

base

HNI

(2)

X0^ XN^

Of course, one cannot rule out the possibility in any of the known N-haloamide or N-
haloimide rearrangements that intramolecular hydrogen abstraction is performed by
amide oxygen rather than nitrogen, followed by tautomeric regeneration of the normal
amide group (eq 3) • This would involve the same size cyclic transition state as
preferred by alkoxy radicals in the light induced radical chain decomposition of

-286-

RtW < > RC=NR' > .RC=NR' > .RCNHR1 (3)

t-butylhypochlorite s with side chains of three carbons or longer. These reactions
occur largely via an intramolecular path (1,5 hydrogen shift) to give S-chloro-
alcohols.24 An intermediate N-iodoamide was isolated by reacting lead tetraacetate
and iodine with benzamide. N-Iodoamides could also be prepared by reacting amides
with t-butylhypochlorite and iodine in various solvents where the iodinating species
was shown to be t-butylhypoiodite. 19 Crystalline N-iodo compounds were obtained from
benzamide, succinimide, n -butyr amide , n-hexanamide and n-octadecanamide using hypo-
halite as the iodinating agent. N-Iodo-n-octadecanamide was isolated in two forms
with melting points of Hk and 120°. Whether these were two crystal forms or two
geometrical isomers has not been determined.

Solutions resulting from photolysis of N-iodoamides showed a strong infrared
band at 1680 cm"1. This suggested the presence of an amide or iminolactone. Washing
these solutions with sodium hydrogen sulfite caused the appearance of a /-lactone
band in the infrared spectrum. The following results were obtained in order to
determine which of the two species was present after photolysis and before hydrolysis.
7-Iodobutyramide was found to be unstable at room temperature in a "humid" atmosphere
and cyclized spontaneously to give 7 -lactone. Photolysis of the higher melting N-
iodo-n-octadecanamide showed that all the iodine originally present as N-iodo was
found as molecular iodine at the end of the reaction. Yields were never greater than
50/0 unless excess iodinating agent was used. These results suggested the following
sequence :

DEI

_> (

-y

>KH-

HI K-^

tt 1 f

■\^NH

During the initial period of darkness there was no appearance of iodine. Irradiation
caused a rapid development of iodine. During a further period of darkness only a
small amount of iodine appeared which was about 10$ of the amount of iodine liberated
during the light period. Further periods of light and darkness produced the same
effects.

The solution resulting from photolysis of N-iodo-n-octadecanamide was divided
into two equal portions. One portion was hydrolyzed in the normal manner while the
other was treated with zinc dust and acetic acid before hydrolysis. Both solutions
gave the same yield of lactone upon hydrolysis. It has been shown19 that 7-iodo-
butyramide is smoothly converted to butyramide under identical treatment with zinc
and acetic acid. These results suggest that iminolactone is responsible for the
infrared band at 1680 cm"1. Isolation of a derivative of the postulated iminolactone
was finally achieved during the photolysis of 7-phenylbutyramide in the presence of
an excess of t-butyl hypochlorite and iodine. The crystalline compound which formed
during the photolysis was regarded as the iodine chloride complex (X) of N-iodo-7-
phenylbutyroiminolactone. Excess iodinating agent reacting with 7 -iminolactone
could be responsible for the formation of X. Infrared, NMR, and microanalytical
data support the structure assigned to (X) .

ft J Uni.ici

Like the H-L reaction, optically active compounds with an asymmetric 7-carbon
rearrange to give racemic product. Photolysis of optically active (+) -^-methylhexan-
amide (XI) in the presence of t-butylhypochlorite and iodine gave racemic k-methyl-k-
hexanolactone (XII) . An insertion mechanism was thus excluded.

* v> 0

iuj.2

XI XII

N-Bromoamides and N-chloroamides can also be phot olytic ally rearranged to give
7-lactones. Rearrangement is most efficient with N-t -butyl derivatives of the amides,
Treatment of a N-t-butylamide with 10$ excess t-butylhypobromite in carbon tetra-
chloride at room temperature gave the N-bromo compound. On subsequent irradiation,
active bromine was lost within ten minutes. The infrared spectra of the resulting
solution showed a strong absorption characteristic of secondary amides. Brief
heating afforded the iminolactone (eq 5) which could be precipitated by dilution with
anhydrous ether. The N-chloroamides could not be rearranged as readily as their N-

RCH2CH2CH

Jj-t-

Br
XIII

C4H9

* W

HBr
N-t-C4H9

(5)

bromo counterparts. Heating the 7-chloro compound in sulfuric acid was required for
ring closure. Table II lists reaction conditions for photolytic rearrangement of
N-iodo-, N-bromo-, and N-chloroamides.

Table II. Photolytic Rearrangement of N-Haloamides and N-Haloimides

Compound

AcO

CONHR

Temp

Light

Time,

Substituent

Solvent

°C

Source

min

io 7 -Lactone

-H

benzene

300

k6.b

-CH3

HCCI3

reflux

None

-C6H5

HCCI3

reflux

2^0

17.2

CH3(CH2)3'

CH3CH2CH

CH2CH2CH2'
C&R5

•H

benzene

120

■t-C^Hg

benzene

CC14

•C(C6H5) 3

benzene

None

•CH3

benzene

150

None

CCI4

150

None

■C6H5

benzene

120

None

■CH3

benzene

■t-C4Hg

benzene

•H

CFC13

18 hrs

■CCH3

CFCI3

8-9 hrs

-H

-CCH<

CFC13

8 hrs

11 hrs

125 -watt high -pressure mercury arc, lamp, tungsten lamp, Hanovia 100-watt
medium pressure mercury arc lamp, and Rayonet 3500 A0 lamp augmented by a Victor
500-watt mercury vapor lamp.

The lower yields for N-methyl compounds (Table II) could be attributed to dehydro-
halogenation resulting from loss of hydrogen a to nitrogen (N-C-H). Irradiated N-
chloro-N-methylacetamide (XIV) reacted completely within 30 minutes while N-t-butyl-
N-chloroacetamide (XV) was unreactive for 180 minutes under identical conditions

■288-

XIV

CH3CN-CH3

3oij*-t-C4H9

pentanoamide , which in common with the N-t-butylamides lack a hydrogens, were
unsuccessful. While N-bromo compounds could be rearranged in benzene or carbon
tetrachloride, N-chloro rearrangements were successful only in benzene and pyridine,
This solvent effect is presently being investigated,25'26

The rearrangement products of N-chloro-N-t-butylpentanoamide and N-chloro-N-
t-butylhexanoamide were reduced to the 4-chloroamines with diborane, Refluxing for
three hours gave the corresponding pyrrolidine (eq 6) . This is another way of

RCHCH2CH2CNH-t -C4H9

BpH

2n6

[RCHCH2CH2CH2NH-t-.C4.H9]

base

t-C4,Hg

(6)

preparing H-L rearrangement products without using strong acids.

The hydrogen abstracting, chain-carrying species in aliphatic,22 allylic and
benzylic23 halogenations by N-halosuccinimides is normally the halogen atom, rather
than the succinimidyl radical. Since intramolecular rearrangements are generally
more rapid than the corresponding intermolecular reactions, it is possible that
acyclic imidyl radicals might rearrange at rates fast enough to permit selective
introduction of functional groups at the 7-position of imides (eq 7).20

R— i

N-

V I

T^X

(7)

XVI

Petterson has shown that 7-chloro-N-acetylamides (XVI, X=C1, R"=CCH3) are formed
from compounds having primary, secondary, or benzylic 7-hydrogen. N-Chloroimides
were readily made Q> 9Cffo) by reacting the parent imide with t-butylhypochlorite in
methanol. 4 Photolysis under helium produced ^--chloro derivatives which were con-
verted to lactones by acid hydrolysis. Table II shows that under similar conditions
N-acetylamides give better yields of 7-lactone than do the corresponding N-hydro-
amides. The reason why the second carbonyl group on nitrogen promotes rearrange-
ment is not yet known. One might expect from Neale's work that benzene would be a
better solvent for imide rearrangements. Preliminary studies show that reaction
times in benzene are decreased but yields are also diminished.28

CONCLUSION

Respectable yields from N-halo cyclization reactions and the short series of
operations involved in the reaction further spotlights the increasing usefulness of
nitrogen radicals in organic synthesis. These reactions serve as an example of a
free radical synthesis which may be difficult or even impossible by a nonradical
approach. The mechanism of these reactions is generally agreed upon, although the
species responsible for initiating the reactions must be subject to further
investigation. Rearrangement conditions and yields are dependent upon the solvent,
groups substituted on nitrogen in amides and the substituents at the 5-carbon
(amines) or 7 -carbon (amides and imides).

-289-
BIBLIOGRAPHY

1. M. E. Wolff, Chem. Rev., 6£, 55 (1963).

2. G. R. Wright, J. Am. Chem. Soc, JO, I958 (1948).

3. S. Wawzonek and T. P. Culbertson, J. Am. Chem. Soc, 8l, 5367 (1959).
k. E. J. Corey and W. R. Hertler, J. Am. Chem. Soc., 82, 1657 (i960).

5. S. Wawzonek and P. J. Thelen, J. Am. Chem. Soc, 72, 2118 (I95O) .

6. R. Lukes and M. Ferles, Coll. Czech Chem. Comm. , 20, 1227 (1955).

7. R. S. Neale, Tetrahedron Letters, 483 (1966).

8. R. S. Neale and M. R. Walsh, J. Am. Chem. Soc, 87, 1255 (1965).

9. F. Minisce and R. Galli, Tetrahedron Letters, 167 (1964).

10. F. Minisci and R. Galli, Chim. Ind. (Milan), 46, 546 (1964).

11. R. S. Neale, N. R. Marcus, and R. G. Schepers, J. Am. Chem. Soc, 88, 305I
(1966). ~

12. R. S. Neale, M. R. Walsh, and N. L. Marcus, J. Org. Chem., £0, 3683 (1965).

13. A. W. Hofmann, Ber. , 16, 558, 586 (I883).

14. K. Loeffler and S. Kober, Ber., 42, 3427 (1909) ', K. Loeffler and C. Freytag,
ibid., 42, 3^27 (1909).

15. G. H. Coleman, G. Nichols, and T. F. Martens in "Organic Synthesis, Coll.
Vol. Ill," John Wiley and Sons, Inc., New York, N. Y. , 1955, p. 159.

16. C. Walling, "Free Radicals in Solution," John Wiley and Sons, Inc., New York,
N. Y., 1957, Chapt 8.

17. W. S. Metcalf, J. Chem. Soc, 148 (1942).

18. S. Wawzonek and J. D. Nordstrom, J. Org. Chem., 27, 3726 (1962).

19. D. H. R. Barton, A. L. J. Beckman, and A. Gossman, J. Chem. Soc, 181 (I965).

20. R. C. Petterson and A. Wambsgans, J. Am. Chem. Soc, 86, 1948 (1964).

21. R. S. Neale and R. L. Hinman, J. Am. Chem. Soc, 8£, 2o*66 (1963)) R. S. Neale,
J. Am. Chem. Soc, 86, 5340 (1964).

22. P. S. Skell, D. L. Tuleen, and P. D. Read, J. Am. Chem. Soc, 8£, 285O (I963).

23. R. E. Pearson and J. C. Martin, ibid., 8£, 354 (1963).
2k. C. Walling and A. Padwa, ibid. , 8£, 1597 (I963).

25. R. C. Petterson, A. Wambsgans, and R. S. George, 151st National Meeting of the
American Chemical Society, Pittsburgh, Pa., March 1966, Paper No. 53.

26. G. A. Russell, J. Am. Chem. Soc, 80, 4987 (1958).

27. R. S. Neale, private communication.

28. R. C. Petterson, private communication.

-290-

THE THERMAL END0-EX0 ISOMERIZATION OF SOME DIELS-ALDER ADDUCTS

Reported by Tommy L. Chaff in April 6, 1967

According to the Alder Rule1 of Diels-Alder additions, of the two stereo-
isomeric adducts of a cyclic diene with a dienophile, that one which is formed
with the maximum accumulation of double bonds will preponderate. Although there
are many exceptions to this rule, there has been controversy as to whether they
result from the isomerization of the initially formed endo adduct directly to to
the exo;

Addends^ Endo^Exo

or whether the endo product is reversibly formed from the addends which can
slowly form the thermodynamic ally more stable exo isomer;

Endc^ Addends"^ Exo

Any type of direct isomerization which does not involve dissociation into kin-
etically free addends will be termed an "internal" mechanism in contrast to an
"external" mechanism such as dissociation and recombination. This seminar will
present the evidence for these two possibilities.
ORIGIN OF THE PROBLEM

The first thermal isomerization of this type was reported in 1933 by Alder
and Stein2. They observed the isomerization of endo-dicyclopentadiene (la)
to exo-dicyclopentadiene (lb) at 170°. They felt that this was due to the existence

of an equilibrium with a small concentration of the monomer under the reaction
conditions .

It has also been observed1 >3 that addition of maleic anhydride to 6,6-pent-
amethylenefulvene at room temperature produces the endo isomer while at higher
temperatures mixtures of the isomeric adducts are found. This, too, was explained
on the basis of the reversible formation of small concentrations of the addends.
Woodward noted3 that the dihydro derivatives did not isomerize and that a solution
of the adducts in benzene or ethyl acetate turned yellow when warmed, presum-
ably from the colored fulvene. The exo isomer was considerably more stable and
therefore the reaction was thought to be kinetically controlled at low temperatures.
This dissociation-recombination mechanism has been proposed4 to be a general one
for Diels-Alder adducts of cyclic dienes.

Craig, in 1951 > reported5 that heating the endo adduct of maleic anhydride
and cyclopentadiene (Ila) to I9O0 produced the exo isomer (lib) and not the ad-
dends as had been previously reported.6 He concluded that the isomerization
proceeded by means of a non-isolable intermediate and noting that the dibromo
compound did not rearrange, he proposed the following mechanism involving the double
bond:

H 0

-291-

'.lib

C02H
C02H

III

This mechanism can explain why the addends -./ere observed for the rearrange-
ment of the fulvene -maleic anhydride adducts, since this type of mechanism would
be impossible for fulvene.7

Schroder has proposed8'9 that the endo-exo isomerization of dicyclopentadiene
is an intramolecular rearrangement.
EXPERIMENTAL EVIDENCE

Berson has studied10 the rate of formation of the endo and exo adducts
(III) from furan and maleic acid. Contrary to a previous report1 x, the reaction

does not produce only the endo isomer . Since the adducts are
quite unstable12 they were isolated by saturating the double
bond with bromine to prevent retrogression. If both isomers
are formed directly from the addends , then the rates of formation
of both should be at their maxima when the concentration of
the addends is highest, at the beginning. This is assuming
second-order kinetics for both isomers. On the other hand,
if the exo isomer is formed directly from the endo without forming the kinet-
ically free addends, then the endo isomer would have its maximum rate of formation
at the beginning and the exo isomer should have its maximum rate at some later
time corresponding to the maximum concentration of the endo isomer. The data
are shown graphically in Figure 1, and although it seems to support the direct

isomerization, the analytical
method could not be tested on
mixtures of known composition
due to the instability of the
endo adduct, and the authors
chose not to distinguish between
the proposed mechanistic paths.

It has been reported5
that the adduct of cyclopenta-
diene and maleic anhydride (II)
undergoes diene interchange at
200° with 2,3-dimethylbutadiene
and dienophile interchange with
fumaric acid. This would seem
to indicate that at least part of the rearranged product results from retrogression,,

Berson13-'14 studied the isomerization of the endo adduct of maleic anhydride
and cyclopentadiene (II) with C14 labeled carbonyl groups in the presence of
an equimolar amount of unlabeled maleic anhydride in boiling decalin (I88.50) .
Since the endo adduct exchanged rather rapidly under the reaction conditions,
it was necessary to determine the activity of the formed exo adduct at short
reaction times. This, in turn, meant that small amounts of the exo adduct were
formed. Therefore the amount and activity of formed exo adduct were determined
by isotopic dilution. If the isomerization proceeds by purely an internal or
direct isomerization, the activity of the exo adduct formed at any time will
be the same as the endo activity at that time. If the isomerization proceeds
by an external path, that is, by retrogression, then the activity of the exo
adduct formed at any time will be the same as the activity of maleic anhydride

100 150 200
TIME, hours
Figure 1

250 300

-292-
at that time. By dividing the reaction into arbitrarily small time increments
the theoretical activity to be expected of the exo adduct by each path was cal-
culated by graphical integration. The activity of the exo adduct isolated was
considerably higher than that expected by a purely external path and therefore
i t vas concluded that a significant part of the isomerization occurred by a direct
path not involving kinetically free fragments. They felt that the most likely
mechanism involved cage recombination or some intermediate complex of the ad-
dends. It has since been demonstrated15 that the rate of addition of cyclopent-
adiene and maleic anhydride is too slow to compete with diffusion and therefore
cage recombination is not a reasonable possibility,

Baldwin and Roberts16 conducted the isomerization of the endo isomer in
fche presence of tetracyanoethylene (TCNE) which had been reported to be a very
good dienophile15. The assumption was that TCNE would react with cyclopentadiene
much faster than would maleic anhydride. Only partial inhibitiion of the for-
mation of the exo isomer was observed which was taken to indicate that an inter-
nal and an external mechanism were in competition.

A number of possible internal mechanisms have been proposed in addition
to the 7^5-hydrogen shift suggested by Craig5. They fall basically into two
categories: mechanisms in which both diene -dienophile bonds are broken with the
fragments contained in a solvent cage or as a complex; and mechanisms in which
any other bonds are broken. In the second category are included: (l) an acid
catalyzed Wagner °Meerwe in rearrangement17; (2) a base catalyzed inversion
(enoiization)18; (5) formation of a nortricyclyl derivative (IV) or cyclopent-
adienyl succinic anhydride ( V) 5"19j2°- (4) and cleavage of the 2^,3-carbon-carbon
bond to form a common intermediate- possibly a double cyclopropane structure ( vT*1!

0^° ^0

IV V VI

has been observed that the rearrangement is not catalyzed by acids or bases5
which would seem to eliminate the first two proposals. Roberts and co-workers21
planned to study the internal mechanism further by rearranging the endo adduct
which was stereospecifically labeled in only one carbonyi. They planned to use
TCNE as a scavenger for cyclopentadiene so that all of the exo isomer formed
would result from the internal mechanism. It was hoped that the position of the
label in the exo isomer would eliminate some of the possible internal mechanisms.
However, it was found that maleic anhydride reacts with cyclopentadiene at a
rate comparable to that of TCNE under the reaction conditions. An alternative
would be to run the reaction in an excess of maleic anhydride so that essentially
all of the exo product formed by the internal process would be labeled and none of
that by the external process. First^ however, the isomerization was carried
out again on the adduct with uniformly labeled carbonyls22 to re-examine the case
for the internal mechanism. This was prompted by the fact that TCNE was found
not to be an effective diene scavenger for this case and by the observation that
in Berson's exchange experiment14 in boiling decalin not all of the maleic an-
hydride was in solution. Furthermore^ a considerable amount of the maleic an-
hydride sublimes out of the reaction within a few minutes. Both of these exper-
iments , then^, prejudice the results toward an internal process. When essentially
the same experiment was run with an equimolar amount of the labeled endo isomer
and unlabeled maleic anhydride in t-pentylbenzene , which gave a homogeneous
solution, the results indicated that no internal mechanism was involved in the

„29>
isomerization.

In other work, Miranov has shown23 that isomerization of the 7-methyl
(VII) , 1 -methyl (VTIl), and 6-methyl (IX) -5-norhornene-2,3-endo-cis-dicarboxylic
anhydrides gives a 2:1 ration of 1- to 6-methyl adducts and only a trace of the
7 -methyl adduct. Even though the endo-exo ratio was not determined this is
interpreted solely in terms of dissociation to the addends since substituted
cyclopentadienes have been shown to undergo double bond migration under these
conditions24. Also the endo adduct of maleic anhydride and 1,4-diphenyIcyclo-

-L:;-^\-

1-4

Q I >j

\> V^

VTI

VIII

pentadiene (Xj has been reported25 to give a 1:3 ratio of 1 -.'••■ exo isomer and 1,5-
diphenyl-5-nor.oornene-endO'°cis°2,3' dicarboxylic anhydride (XJ. Although the
author postulated that the products resulted from two competing mechanisms > it
appears that they can be explained in the same way as those of Miranov,

Baldwin26 has studied the isomerization of specifically deuterated dicyc-
lopentadienes (i). An external mechanism should produce scrambling of the deu-
terium. Statistically, one would expect 25$ unlabeled, 25$ doubly deuterated
and 50/-' monodeuterated product from monodeuterated starting material. An inter-
nal mechanism would preduct all monodeuterated product. This will obviously
be complicated by reversible formation of the monomer from one or both of the
isomers since the reaction conditions are essentially those used to generate
the monomer from the dimer27. Therefore if any labeling specificity is found
in the product it can be taken as evidence for an internal mechanism, but scram-
bling of the label could be interpreted as evidence against an internal mechanism
only if it can be demonstrated that the specifically deuterated product does
not equilibrate under the reaction conditions and that the starting material
has not equilibrated before reaction. A specifically deuterated exo dimer was
recovered after 90 minutes at 196° during which time no scrambling had occurred.
The endo dimer appears to have undergone about 60$ equilibration at the end of
7 minutes under the reaction conditions. The results of the isomerization of
the endo dimer are shown in Table I,

Table I

Deuterium Distributions in Dicjyc lopentad iene s upon endo to exo

Re ar r a njeme ntatl96cr~"

Time , Rearrangement ,

min. °/o

^endo Dimer -\ /-
'd0 dx d2 d3 ' 1

exo Dimer — \
dQ di d2 d3 I

0
7

60
90

o 6 8o k 10

1,1 26 57 16 l

11 34 42 18 5

16 67 20 9 3

30 hi 19 4
28 k-9 22 2

31 47 2.1 1

The theoretical distribution of deuterium in the product was calculated
on the assumption that both the doubly and the triply deuterated endo starting
material had the deuterium in one cyclonentadiene unit. This leads to a cal-
culated distribution of 29.4^ d0.» 44.4$ d2, , \$?o d2, and 1% d3. This is in good
agreement with the experimental results, and for short reaction times } where

equilibration is incomplete in the starting material, would seem to suggest
little if any contribution from an internal mechanism,, Herndon and co-workers30
have arrived at the same conclusions from a kinetic study of the thermal decom-
position of the dimers in the gas phase,

Berson studied the isomerization of optically active adducts of cyclo-
pentadiene with methyl acrylate and methyl methacrylate19-'28. Heating the op-
tically active exo adduct of cyclopentadiene and methyl methacrylate(XIIb) at
170° in decalin for 3° 5 hours gave 5°6/» conversion to the racemic endo adduct
(XIIa)o The recovered starting material, was only T/° racemized and it was shown
that the optically active endo isomer does not racemize under the reaction
conditions. F>jrthermore , the extent of conversion corresponds to that antic-
ipated on the assumption that the addends are common intermediates for the race-
mization of the exo isomer and its conversion to endo. Under these circumstances
the percent conversion should have proceeded to an extent equal to the percent

racemization of exo times the kinetic ratio(~^— ) for formation of adducts from
addends. exo

C02CH3

Xllaj R=CH3

Xllb;

R=CH3

XIIIa$ R=H

Xlllb;

R=H

XI Vb

Similar results were also found for the isomerization of the optically
active endo adduct of cyclopentadiene and methyl acrylate (XIIl) . These results
strongly indicate the lack of an appreciable contribution from an internal mech-
anism in this case.

In the systems considered thus far the isomers have been chemically dif-
ferent and it has been necessary to follow the behavior of both species, how-
ever for the optically active adduct of 9-phenylanthracene and maleic anhydride
(XIV) this is not the case since the isomers are also enantiomers15. If an
internal mechanism exists which allows interconversion of the enantiomers , loss
of optical activity should exceed dissociation into kinetically free fragments.
If, however, the rate of loss of optical activity is exactly the same as the rate
of dissociation, then there can be no other significant path for racemization.
The loss of optical activity of the adduct was followed as a function of time and
tee first order rate constants determined at three different temperatures.
The rates of dissociation were determined spectrophotometrically by means of the
diene ultraviolet absorption and found to be identical with those for loss
of optical activity.

One final example of this Isomerization has been studied29. When either
the endo or exo adduct of cyclopentadiene and l,4-benzoquinone=2,3-epoxide is
heated to 220°~for 10 minutes the resulting mixture consists of an approximately
equal mixture of the two isomers as evidenced by a comparison of the infrared
spectra with those of known mixtures. When either isomer is heated under the
same conditions in the presence of an equimolar amount of TONE only 1,4-benzo-
quinone-2,3=epoxide, 292,3,3~tetracyano-5-norbornene, and starting material are
produced.
CONCLUSION

In the light of these studies there seems to be no firm evidence for any
mechanism for the thermal endo°exo isomerization of Diels-Alder aducts other
than dissociation and recombination.

-295-

BIBLIOGRAPHY

1. K. Alder and Go Stein, Angew. Chem., 50, 514 (1937)'

2. K. Alder and G. Stein, Ann., 504, 216(1933) •

3. R. B. Woodward and H. Baer, J. Am. Chem. Soc. , 66, 645 (1944).

4. K. Alder and W. Trimborn, Ann. . 566 y 58 (I.950).

5. Do Craig, J. Am. Chem. Soc, 73,~TO9 (1951).

60 M. Kloetzei, Organic Reactions , Vol. IV, R. Adams, Ed., Wiley, New York,
1948, p. 9.

7. D. Craig, J. J. Shipman, J. Kiehl, F. Widmer, R. Fowler and A. Hawthorne,
J. Am. Chem. Soc. , 76, 4573 (1954).

8. W. Schroder, Angew. Chem., 72, 865 (i960).

9. Wo Schroder, Angew. Chem., jQ, 24l (1961).

10. J. A. Berson and R. Swidler, J. Am. Chem. Soc, 7£, 1721 (1953).

11. 0. Diels and K. Alder, Ann., 490, 243 (193L).

12. R. B. Woodward and H. Baer, J. Am. Chem. Soc, 70, ll6l (1948).

13. J. A. Berson and R. D. Reynolds, J. Am. Chem. Soc, 77 4434 (1955).

1.4. J. A. Berson, R. D. Reynolds and W. M. Jones, J. Am. Chem. Soc, j^, 6049

( I956) c

15. J. A. Berson and W. A. Mueller, J. Am. Chem. Soc, 83, 4940 (1961).

16. J. E. Baldwin and J. Do Roberts, J. Am. Chem. Soc. 85, 11 5 (1963).

17. P. Bartlett and A. Schneider, J. Am. Chem. Soc, 68, 6 (1946).

18. H. Kwart and I. Burchik, J. Am. Chem. Soc, 74, 3094 (1952).

19. J. A. Berson, A. Remanick and W. A. Mueller, J. Am. Chem. Soc, 82 , 5501
(I960).

20. R. B. Woodward and T. J. Katz , Tetrahedron, 5, 70 (I.959) .

21. U. Schiedegger , J. E. Baldwin and J. D. Roberts, J. Am. Chem. Soc, 89,
894 (1967).

22. C. Ganter, U. Schiedegger and J. D. Roberts, J. Am. Chem. Soc , 87. 2771

(1965).

23. Vo A. Miranov, To M. Fadeeva , A. U. Stepanyants and A. A. Akhrem, Bull.
Acad. Sci., U. S. S. R. , Div. Chem. Sci., (Eng. Trans.), 293 (1966) .

24. V. A. Miranov, E. V. Sobolov and A. N. Elizarova, Tetrahedron, ig, 1939 (1963)

25. K. Leppanen, Ann. Acad. Sci. Fennicae, Ser. A II, 131 (1965)$ Chem. Abstr. ,
64, 14112 ( 1966) .

26. Jo'e. Baldwin, J. Org. Chem., Jl, 244l (1966).

27. R. R. Moffett, Organic Synthesis, Vol. 32, Wiley, New York, 1952, p. 4l.

28. J. A. Berson and A. Remanick, J. Am, Chem. Soc, 8_3, 4947 (1961).

29. M. J. Youngquist, D. F. O'Brien and J. W. Gates, Jr., J, Am. Chem, Soc,
88, 4960 (1966).

30. W. C. Herndon, C. R. Grayson, and J. M. Manion, J. Org. Chem., J52, 526
(1967)o

-2Q6-

REACTIQNS OF NH RADICALS

Terry G. Burlingame April 10, 1967

INTRODUCTION

Imidogen, or NH, may be considered the simplest of the nltrenesj i.e., that
class of reactive species containing monovalent nitrogen, which therefore contains
six electrons in its outer shell. Several reviews on nitrene intermediates, R-N,
in general, have appeared in the literature. 1,a Imidogen is isoelectronic with the
much-studied reactive species carbene, or CHS„3 Both of these may exist either
as the triplet or as the singlet specie?, (in which case carbon and nitrogen are
quite electrophilic due to their electron deficient structure) „ The most commonly
used method to generate NH is through the phct ©decomposition of hydrazoic acid, HN3,
in which the well-established primary process is as follows :

hV
H-H-Ns-N -=-» NH + N2

Free NH was first detected in l892„4 Since thatjiime, the species has been well
characterized spectroscopically in the gas phase5 (by absorption and emission) and
in solid matrices at low temperatures6 (by UV and IR absorption) . Less definitive
evidence has been obtained for existence of NH in the liquid phase.

The electronic states of NH have been determined spectroscopically.7

Ground state ; triplet , 3IT
Lowest excited state; singlet, XA
Second excited state; singlet, 127'
The exact energy separation of these states is not known, although the 3£" and ^
levels have been estimated to differ in energy by 2] kcal.7

SCOPE

The purpose of this seminar will be to discuss in some detail the gas, solid,
and liquid phase reactions of NH radicals with other molecules 3 to examine the
mechanisms proposed for those reactions in which NH is postulated as an important
intermediate, and to discuss the validity of such mechanisms in the light of the

experimental evidence presented.

REACTIONS OF NH IN THE VAPOR PHASE

The most systematic analysis to date of the reactions of NH radicals with
various organic molecules in the gas phase has been carried out by Lwowski and co-
workers.8 Reactions of methane, ethane, ethylene, butene-1, heptene-3, and 2,3
dimethyIbutene-2 with NH radicals generated by the phot ode composition of hydrazoic
acid, Hl3, were studied. Two sets of experiments were performed; those in which NH
was generated by high energy flash photolysis of < _ hydrocarbon mixtures, and those
in which steady, slow irradiation of the mixtures was carried out. Flash kinetic
spectroscopy performed at various time intervals after irradiation in the flash
experiments gave absorption bands (A3rt <-X32T) due to the triplet ground state of the
.NH radical and bands from highly vibrationally excited °C«N radicals. The C2 and CH
transients were detected in the ethane and ethylene reactions) CH was also detected
in the methane reaction. There is a close correlation between the decay time of the
NH absorption and the appearance time of the °CSN spectrum. No bands due to singlet
NH were observed. The principal nitrogen-containing products in all the flash-
initiated reactions were N2 and HCN. No alkyl cyanides were detected. The slew
photolyses yielded in general HCN, alkyl cyanides, saturated hydrocarbons and hydrogen
in the gas phase. No analyses for NH3 or KHL4.N3, aim* ,r certainly products of
secondary NH reactions, were made. For example, in the case of ethylene-HN3 mixtures,
products obtained were HCN, CH3CN, CH*, H2 and an amorphous solid. The HCN/CH3CN
ratio of .9 was independent of ethylene pressure in the range 8O-56O mm. Photolysis
of a 0.5 ;1 mole ratio mixture of DN3 and HN3 gave CH^DCN and CH3CN in a ratio of
O.25. 'The following scheme was proposed in view of the above results to occur in
"hot.h t.Vip flash and alow reactions*

32~) + CH^CHo

H2C-CH<

H2G~CH3

Further fragmentation of the nitre ne intermediate then occurred according to its
energy content %

^D°CH3

flash

reaction

-»■ C3U + H. +'C^K

A"Ui3 reaction

•IS

,^={H2 + CH3CM

^HCN + CH4

If the above scheme is reasonable , one should expect to find CH3CN but no HCN in the
steady photolysis of HN3 + 2,3»dimethylbutene-2;

}E3 CHj

CHa-C— (

CH3 CH3

CHq-C — C — CH

HH CH3

CH3CU + (CH3)^CHCH3
CaH6 + (CH3)^C-CNN

This was found to be the case. Furthermore, in the reaction between NH and butene-1
addition at the 2-position produced C2H5CH and CH3CN in almost equal amounts, suggest-
ing that the intermediate formed after 1,3 hydrogen transfer had a lifetime long
enough to permit vibrational relaxation. In addition, flash and slow photolysis of
ethyl azide vapor, C^HsNs, gave similar results to the ethylene -NH reactions,
suggesting that the same intermediate nitre ne was being formed in both cases. Flash
photolysis gave HCN, CH3CN (ratio ^.9), H2, CH4 and a white solid. Steady irradia-
tion gave CH3CN, CJU, H2 and a polymeric gum.. , HON seems to react in forming the gum.
No ethyleneimine , or ethylamine, which result from solution photolysis of C^E5N3,9
were found,

Lwowski and co-workers have considered in detail several alternative explanations
which may also be used to rationalize their results s

The possibility exists for the formation of nitrogen atoms from HN3. Indeed,
the energy requirements of the process HN3 •> H° + °N + N2 are well within the limits
of that supplied by the 2537 A radiation. Winkler et.al.10 have proposed a "unified
mechanism" for the reactions of nitrogen atoms with simple organic molecules,
illustrated here for C3 molecules s

CH3c=CH~Cii2 \
CH3CH2CH2C1 )
CH3CH2CH2CH3 >
CH2— G.H2 J

+ N°

XCH^

N« + "CaHs
N- + CH3»

CH3='CH «'Cfl2

HCN + CH3'
HCN + 2H«

-» HCN + C^H5<

+ H<

In addition, Dubrin et.al.11 have studied the reactions of methane and ethylene
with 13N atoms produced by nuclear techniques and have concluded that the reaction
with ethylene follows the paths

CsH^ + N(TD)

>> HCN + °CH3

Cyanide radicals were ruled out on the basis of results different from the expected
H2C-CH-C=N/HCN ratio of 4~7ol obtained on addition of 1XCN to ethylene.12 Lwowski
and co-workers rule intermediacy of nitrogen atoms in their system out since (1) no
emission or absorption bands for triplet N2 resulting from If atom recombination are
observed. (2) There is close correlation of the decay and appearance times of NH
and CN respectively, making further dissociation of IfH unlikely.

One could, argue also that the flash and slow photoiyses proceed by entirely

-2Q8-
different mechanisms. The only rationale for a common intermediate given by Lwowski
is the similarity of products obtained in the flash and slow photolyses and the
corresponding photolysis of ethyl azide, differences being primarily due to dif-
ferences in the energy content of said intermediate depending on its mode of formation.
However, there seems to be no evidence at present, to rule out the formation of
vibrationally excited ethylene imine folio-wed by decomposition;

NH 4 H2C*CH2 _» HsQ^— CH2 * > CH3CN + HCN
( singlet J

triplet) (singlet or triplet)

However, in analogy to the reaction of methylene with ethylene in the gas phase to
produce excited cyclopropane which then isomerizes to propylene , we would also expect
excited ethylene inline to isomerize.

The independent work of two other research groups bears an important relation-
ship to that of Lwowski, Miller and Rice13 studied the system HN3-ethylene,
analyzing for all products in efforts to obtain a complete mass balance. Products
identified were KH4N3, ethane, and HCN in comparable amounts, and smaller amounts of
H2 and CH3CN. Formation of NH4N3 in reactions of NH with both alkanes and alkenes
suggests the competing sequence;

NH 4 CaHi (or other H atom, donor) ■> NH2 + 0^3-
NH2 + C^ (or other donor) ■> NH3 + CaH3-
NH3 + HN3 ■> NH4N3

In a series of three papers Back14"16 and co-workers have analyzed the flash
and slow photolysis of isocyanic acid vapor, H~N=C=0, in the presence and absence of
various hydrocarbons. A priori the photolysis of HNCO and HN3 should be related in
the same way that reactions of the two methylene precursors ketene, CH2=C=0, and
diazomethane CH2N2 are related. At low pressures the photolysis of isocyanic acid
gave CO, N2 and small amounts of H2 as non condensable products.14 Although C02,
HCN, C2N2, NO, N20 and N02 could have been detected, they were not found. Small
amounts of NH3 and N2H4 were detected also. The mechanism proposed accounted for
most of the observations;

HNCO -£¥— » m + co

NH + HNCO — > NH2 4 NCO

NH2 4- HNCO — * NH3 4 NCO
2NC0 _> N2 4- 2C0

The observation that added ethylene reduced the yield of CO, N2 and H2 was accounted
for by the scavenging of the initially formed NH by the ethylene.

Lwowski has given three alternative explanations in the light of his studies to
account for Back's failure to observe HCN;

(1) HCN produced from the added olefins reacts with HNCO;

-°\

C-H

oxadiazoles
(2) NH reacts rapidly with HNCO to give »NCO which then attacks ethylene;
•NCO 4 H2C=CH2 — » 0=C=N-CH2CH2-

(5) HNCO forms a relatively long lived excited state which reacts with C^.4
to give ethyl isocyanate faster than it dissociates.

Brash and Back16 have carried out a more detailed study of the steady irradiation
of HNCO vapor in the presence of olefins and paraffins. Increased amounts of olefins
reduced the N2 and H2 quantum yields to zero and the CO yield to a constant value
indicating complete scavenging of NH radicals. No imines, amines, or other nitrogen

-299
containing products could be found, HON was again not detected. HNCO irradiated
with up to 500 mm of butene-2 gave no HCN or imines. 'The results were explained on
the basis of rearrangement and polymerization of highly vibrationally excited inter-
mediates although no specific analysis of polymeric products was made. Photolysis
of HNCO in the presence of ethane, propane, and neopentane showed the same general
behavior except that the hydrogen yield increases, and products of radical coupling
are found. Small amounts of added ethylene reduce the H2 yield drastically,
presumably by efficient scavenging of H atoms. Photolysis of DNCO-C3H8 mixtures and
HNCO-D3H8 mixtures gave primarily HD, showing that each molecule of H2 contained one
hydrogen atom from HNCO and one from propane. Insertion of NH into a C-H bond to
give a vibrationally excited amine was proposed as the most likely process leading
to product formation °9 however, production of hydrogen atoms by this process seems
unlikely and other alternative mechanisms cannot be ruled out by the data.

REACTIONS OP NH RADICALS IN THE SOLED PHASE

The stabilization of reactive species and their subsequent reactions in solid
matrices at low temperatures have been reviewed in a recent seminar.17 Briefly
reviewing some of the main differences between gas phase and matrix reactions, we
find the following:

(1) Severe translational limitations exist for species formed in a matrix 9
diffusion out of the matrix "cage" is severely limited for all but the smallest of
molecules or radicals.

(2) Molecules or radicals formed in excited electronic and/or vibrational states
by photolysis or combination with other reactive species may be rapidly converted to
lower energy electronic and/or vibrational states by frequent collisions with the
inert matrix cage molecules. Alternatively, the matrix cage may be a reactive
molecule which can efficiently add to and trap a reactive intermediate.

An early study by Milligan and Jacox18 attempted to correlate reactions of CH2
and NH in inert argon matrices. Infrared spectroscopy was used to directly analyze
the products from the photolysis of mixtures of HN3-ethylene -argon and HN3-acetylene»
argon at 4°K. The spectral analysis of the HN3 -ethylene -argon system shows that
ethyleneimine is the sole product in the matrix. "This product could conceivably
arise from singlet or triplet NH reacting in the matrix.

matrix

H2C=CH2 + NHl^A) — ■* \J deactivation

H
.. . * (triplet)

HoC-ch^ + m(3z~) nigh enew » <r-7 sPin

h2^-Lh2 + m( 2, ) - \y inversion

H2C<!H2 + NH(32T)

favorable

H2C-CH2 matrix ^ H2C-CH2
ta tNH

The last pathway is probably more likely since NH should readily be deactivated to
the triplet ground state by matrix collisions. However, there is no experimental
evidence other than Lwowski's spectroscopic results in the gas phase which allows
one to distinguish unambiguously between the three pathways. Matrix photolysis of
HN3»eis-2-bute:ae and/or HN3 -trans -2-butene might help establish the identity of the
reacting species. Experiments with methylene from photolysis of diazomethane19
have shown that it is initially formed as an excited singlet which is subsequently
deactivated to the lowest singlet and after an order of magnitude more collisions,
to the ground triplet. Addition of triplet CH2 to ethylene followed by spin inversion

-z.r\r\

accounts for formation of cyclopropane in matrix reactions to be contrasted with
exclusive propylene formation in the gas phase3 where vibrationally excited singlet
cyclopropane is initially produced, then rearranges.

A somewhat more interesting result is obtained in the HN3~acetylene -argon
photolysis. Singlet NH would be predicted to react as follows:

nh + hcech

H~C

= C-H

H
H^N-ChC-H

These species could undergo further rearrangement to acetonitrile or methyl iso-
cyanide. Triplet NH could add in the following manner:

HC-C

t im

-C«H

These products could further rearrange:

CH.2=C=NH

-» HaC=C=MH

azacyclopropene

The infrared spectral analysis rules out aminoacetylene , acetyleneimine , and azacyclo-
propene 0 The same data strongly indicate the presence of acetonitrile and methyl ~
isocyanide as well as the previously unobserved species keteneimine, suggesting that
both singlet and triplet NH could be presnet. The analogous matrix reaction of CH2
with acetylene produces allene almost exclusively.

The photolysis of HNCO and DNCO in argon and nitrogen matrices at k° and 20°K
has been carried out.20 Infrared analysis reveals results quite different from
those obtained in the gas phase in that little NH or CO is spectroscopically
detectable and assignment of new bands seems consistent with the species H-O-C^N.
Two mechanisms, each assuming a different primary process, are proposed:

( 1) HNCO

[H« + °NCO]
matrix cage

diffusion

H« + [°NCO]

recombination

» HNCO + HOCN

In this scheme, enough H atoms should escape the cage to make °NCO observable by IR
spectroscopy^ however, no °NCO is observed by infrared spectral analysis.

(2) HNCO — » [NH + CO]

HNCO
NH(32) + :c=o » :c — 6

-301-

Evidence for the second pathway ven by the fact that in a separate experiment
photolysis of an argon- C0-HN3 mixture gave good yields of both HNCO and HOCN.
Analogous to this reaction in carbene chemistry is the photolysis of N2 + CO + CH2N2 in
a matrix at 20°K with the production of high yields of ketene.3

When HN3 is photolyzed in a matrix composed of solid C02 and the product
analysis is carried out by direct infrared analysis on the matrix, two distinct
groups of bands are observed.21 One group increases in intensity during the
irradiation, the second group rapidly reaches a maximum, then decreases. Control
experiments showed that NH radicals did not diffuse through the C02 cage to give
spectroscopically observable amounts of NHS. Photolysis of RN3 in an N20 matrix
gave group I bands but none of the second group, indicating that the latter arose
from an intermediate NH-C02 adduct. Further analysis identified group I bands as
H-N=0. The characteristic absorptions of CO were also identified in the C02 matrix.
Isotopic substitution using C1302 and C028 showed definitely that a carbon containing
species was responsible for group II bands 0 The observation of pairs of bands with
different growth rates in similar regions of the transient spectrum strongly suggests
rapid rearrangement of the initial intermediate or formation of an intermediate
capable of c is -trans isomerization. Possible intermediates are the following;

-C-0

Jj-N

(a)

(b)

(c)

0-0

H %

(a)

Analogies for (a) exist in the reaction of CH2 with C02 in a matrix,3 (b) could under-
go a cis -trans isomerization about the N-0 bond analogous to alkyl nitrites. Structure
(cj is ruled out by the infrared analysis and the fact that there was no comparable
product (glycxal) produced in the CH2-C02 reaction. Rearrangement to (d) would
involve unfavorable movement of heavy atoms rather than simple hydrogen transfer.

Interesting insight into the reacting NH species in matrix reactions may be given
by the photolysis of mixtures of HN3 and 02 in solid nitrogen at 20°Ko 22 Infrared
analysis of the products formed indicates that both cis and trans nitrous acid,
KO-N-0 are initially formed which undergo further trans ^cis isomerization by UV
radiation and cis^trans isomerization by the infra-red team of the spectrophotometer.
Evidently NH readily reacts with ' be produce HONO. However, the alternate sequence

HN3* + 02 — ¥ HONO + N2
could not be ruled out at the time0 Reactions of CH2 in the gas phase are not
affected by added 02 except when high pressures of inert gas are used, indicating
that singlet •» triplet deactivation occurs by collision with the gas, the reaction
of triplet CH2 with oxygen then occurring readily.

LIQUID AT© SOLUTION PHASE REACTIONS INVOLVING NH RADICALS

It appears that the evidence involving NH formation in solution reactions of
organic molecules is less conclusive than in the gas and solid phase reactions
discussed thus far. There are three reagents, the decompositions which in solution
phase are proposed in some cases to produce NH radicals. These are (1) hydrazoic
acid or azide ion under appropriate photolytic conditions (2) hydroxylamine-O-
sulfonic acid H2N-0~S0,3H under alkaline conditions, and (3) chloramine, H^NCl, under
alkaline conditions. We shall take each of these reagents in turn and compare
examples where NH has been considered an important intermediate with those where
intermediacy of NH radicals has more or less conclusively been ruled out.

The thorough work of Burak and Treinin23 has shown that NH radicals can indeed
be produced in solution. The photolysis of degassed aqueous solutions of NaN3 with
2537 A0 light produces N2, NH20H, H2, NH3 and N^Ii*. Added NH3 causes a large
increase in the N2H4 yield, paralleled by a corresponding decrease in the NH20H
yield. The following mechanism was proposed by the authors to account for the results:

hv .

jfe

»3

+ H20
Mo* —

*
*

-» HN3" 4=
NH + No

-302 -

The ultimate fate of the NH radicals is represented in the following scheme

W + H20 — — -> NH20H
H20 + Mi + N3 — > N2 + N2H2 + OH"

2E2H,

N2H4 + N2

HN3 + WH3

The authors propose that singlet NH is the reactive species, acting as a strong
Lewis acid in its reactions with H20, NH3, and Ng. The quantum yield of N2 was not
affected by such radical scavengers as N20, acetone or methanol -phosphate buffer,
which results rule out mechanisms involving chain reactions or solvated electrons.
The dependence of the quantum yields of N2 and NH20H on azide ion concentration
suggests that there is competition between Ng and H20 for the NH radical. Excess
added ammonia also exerts a marked scavenging effect as shown by its effect on the
quantum yields. Calculated ratios of rate constants for NH scavenging are H20;NH3;
NJ - l:l8;285« The observations that the quantum yields are independent of the
light intensity and that smaller concentrations of impurities including oxygen
have no effect on ID NH20H show that the NH radicals produced are quickly scavenged
by the large excess of H20 present.

An early study24 showed that reactions suggestive of NH formation could be
carried out in organic solutions. Hydrazoic acid was irradiated in the presence of
toluene solvent to yield presumably mixed toluidines. The isomer ratio was not
determined 3 derivative formation alone was used to confirm the presence of toluidines.
The only control reaction run was that of a dark reaction which gave no product
formation over a 12 hour period. In a later more systematic study25 on the solution
photolysis of azides, hydrazoic acid was irradiated in the presence of benzene to
give low yields of aniline. Photolysis of n-butyl and n-octyl azides in benzene
produced comparable amounts of N-n«butyl and N-n~octyl anilines respectively. No
mechanisms for these reactions were proposed although presumably if NH and nitrenes
were involved, reaction would proceed by insertion of N~H or N-R into a C-H bond.
It is interesting that photolysis of diazomethane in benzene solution yields 32/0
cycloheptatriene and 9$ toluene;3

+ :CH2 * > L \\ +

However, none of the analogous ring-expanded product 1-H-azepine was searched for
in the corresponding NH experiment.

Hydroxylamine-O-sulfonic acid^, H2N~0-S03H, has been found to produce some very
interesting reactions in recent years, particularly a series of so-called "imination"
reactions as illustrated by the following schemes26

. ,,, ^H

( V=0 + NH20S03H °H ->

Formation of NH followed by addition to the 00 bond could be envisioned. However,
the evidence available suggests that reaction takes place through undissociated
H2NOSO3H.

0 + HgN0S03g -*£-* ( V^J^L_ 0 — > ( Kj + S°4@

Evidence against dissociation of H^OSO.aH to NH radicals is the observation that its
rate of reaction in NaOH with various added nucleophiles depends on the nucleophile
used which would not be true of NH formation were the rate determining step. Studies
of the reaction of HI with H^OSGaH28 have shown that nucleophilic attack on nitrogen

zrcz

occurs to give INH2 initially. Alkyl groups on the nitrogen atom slow the reaction

considerably . Thus, a claim that HH from H2HOS03B' was trapped by reaction with

cyanide ion and precipitated as silver cyanamide27 may be better explained by direct

nucleophilic attack of CN~ on the acid. Furthermore, H2NOSO3H does not incorporate

radioactive sulfur from radioactive sulfate solutions thus ruling out an initial

equilibrium. q

0 OH @
HsNOSOs ^-^ SO4 + HH

In an interesting reaction reported by Appel and Btiehner29 however, we find it
hard to rule out direct participation of HH radicals. The reaction of H2NOS03H in
sodium meth oxide -methanol with butadiene gave a low yield of ^-py^sline.. This
result is suggestive of a 1,4 addition of HH:

r + MH — — -> U n-H

Other added olefins had produced no detectable amounts of aziridines. It is more
likely that 1,2 addition of HH actually occurs first followed by rearrangement of the
vibrationally excited aziridine to 3 pyrroline. This sequence is not without analogy
in carbene chemistry^ addition of CH2 to butadiene in the gas phase produces vinyl-
cyclopropane as the major product and smaller amounts of cyclopentene.3 The cyclo-
pentene arises presumably from isomerization of excited vinylcyclopropane^ at least
it has been shown that vinylcyclopropane readily undergoes this isomerization
thermally.

Chloramine, HH2C1, has also been utilized recently in a number of interesting
"imination" reactions. °

HH has been shown to' be a product of the photolysis of solid chloramine at low-
temperatures, and of the thermal decomposition of gaseous chloramine. However, in
the solution reactions of the compound, the rate of product formation is dependent
on the nature of the substrate being attacked, which would not be the case if HH
formation were rate -determining „ Thus, these reactions seem to proceed by direct
nucleophilic additions:31

<^J^H-CeH11 + HH2C1 — > (_/C| — » W^H

cf \ H

BIBLIOGRAPHY

1. R„ A. Abramovitch and B. A. Davis, Chem. Rev., &*_, lk$ (1964).

2. L„ Horner and A. Christmann, Angew. Chem. Intern. Ed. Engl., 2, 599 (I963).

3. For an extensive review see ¥. Kirmse, Carbene Chemistry, Academic Press,
Hew York, 1964, ch. 2. For a critical analysis of the gas phase reactions

of methylene see J. Bell in Prog, in Phys. Org. Chem., S. Cohen, A. Streitwieser,
Jr., R. W. Taft, Ed., Interscience, Hew York, 1964, Vol. 2, p. 1.

4. J. M. Eder, Monatsh. , 12, 86 (1892).

5. B. A. Thrush, Proc. Roy. Soc. (London), A£J£, 143 (1956).

6. D. E. Milligan and M. E. Jacox, J. Chem. Phys., kl, 2838 (1964).

7. G. Herzberg, Spectra of Diatomic Molecules, 2nd ed., Van Hostrand , New York,
1950.

8. D. W. Cornell, R. S. Eerry, and W. Lwowski, J. Am. Chem. Soc., 88, 544 (1966).

9. w. H. Saunders and E, A. Caress, J. Am. Chem. Soc. 9 86, 86l (19oTT).

■*,04-

10. H. Go S. Evans, G. R. Freeman, and C. A„ Winkler, Can. J, Chem, , 3>4, 1271

(1956)o
H. J. Dubrin, R. Wolfgang, and C. McKay, J. Chem. Phys., 40, 2208 (1966).

12. Jo Dubrin, C. McKay, M. L, Pandow, and R. Wolfgang, J. Inorg. Nucl, Chem.,
26, 2113 (1964).

13. E0 D. Miller, Ph.D. Dissertation, Catholic University of America, Catholic
University of America Press, Washington, D.Co, I96I.

14. J. Y. P. Mui and R. A. Back, Can. J. Chem., kl, 826 (I963).

15. R. A. Back, J. Chem. Phys., 40, 3493 (1964).

I60 J. Lo Brash and R. A. Back, Can. J. Chem., k%, 1778 (I965).

17. J. Billet, Univ. of 111, Organic Seminar Abstracts, 1966, p. 128,

18. M. E, Jacox and D. E. Milligan, J. Am. Chem, Soc, 85, 278 (I963).

19. G. Herzberg, Proc. Roy. Soc. (London), A262, 291 (19ol).

20. M. E. Jacox and D. E. Milligan, J, ChemTTEys., kO, 24^7 (1964).

21. D. E. Milligan, M. E. Jacox, S. W. Charles, and G. C. Pimental, J. Chem, Phys,,

2L, 2302 (1962).

22. Jo Do Baldeschwieler and G„ C. Pimentel, J, Chem. Phys., 33, 1008 (i960).

23. I. Burak and A. Ireinin, J. Am. Chem. Soc., 87, 4031 (1965)°.

24. R, K. Keller and P. A. S. Smith, J. Am. Chem. Soc, 66, 1122 (1944).

25. D. H. R. Barton and L. R. Morgan, Jr., J. Chem. Soc, 622 (1962).
2c E. Schmitz, R. Ohme, and D. Murawski, Angew. Chem., 73, 708 (I96I) .

27» G. Bargigia, Atti Accad. Nazi. Lincei, Rend. Classe Sci., Fis., Mat,, Nat.,
£, 587 (1964).

28. P. A, So Smith, H. R. Alul, and R. L0 Baumgarten, J. Am.. Chem. Soc, 86, U39
(1964) o ™

29. R. Appel and 0. Bftchner, Angew. Chem. Intern. Ed. Engl., 1, 332 (1962)0

30. E. Schmitz and R. Ohme, Ber, , %£, 2X66 (1961) .

31. E. Schmitz, Angew. Chem. Intern. Ed8 Engl., £, 333 (1964).

-305-

GEOMETRIC ISOMERISM IN DIAZOKETONES

Reported by Daniel B. Pendergrass April 13, 1967

Although diazoketones (i) vere known as early as 2.894(1), they were not
widely used "by the synthetic organic chemist until L. Wolff (2) reported their
rearrangement to the corresponding ketenes. This rearrangement is the charac-
teristic step In the Arndt-Eistert synthesis by which a carboxylic acid may be
converted to its next highest homolog or one of its derivatives^).

R-C02H ■> R-C0-C1 (a)

R-C0-C1 + 2 CH2N2 ■> R-C0=CKN2 + CH3C1 + N2 (b)

R~C0~CHN2 + H-R5^R=CH2-CO-R5 + N2 (c)

r« = -OH , -OR" , -NHR" , or -NH2

The thermolysis (4,5) and photolysis (5.) of diazoketones, as well as their
decomposition in the presence of ae.ids(6,7>8) , bases(5,9).? and various metals
(3>5>10) have been discussed. The literature and reactions of diazobxides have
also been reviewed( 5.»ll) »

The most widely known synthesis of diazoketones is the reaction of an
acyl halide with two equivalents of diazomethane or one of its substituted de-
rivatives. There exists in the literature a variety of representations of the
charge distribution and bond orders in these compounds as illustrated below.

R-CO-X + R'-CHN2 -> R-CG-CN2

L 1

$ 9 0 -© $

R-C-C=N=N i-f R-C-C-N5N <— ► R-C=C-N=N

II », I* J, f I,

0 R' 0 R! ©0 R!

These compounds have also been prepared by the oxidation of a monohydra-
sone of a dik.etone( 12,13,14,15 ,16,17) obtained in a variety of ways. They are
frequently products of nitrous acid oxidation of amines alpha to a -CO-R group
(18,19), solvolytic attack on the monotosylhydrazone of a diketone( 20,21,22,23,24)
in the presence of base, or the action of chloramine or hydroxylamine-O- sulfonic
acid on an oximinoketone(25) . G. R. Harvey(26) has reported that a series of
compounds with structure II react with pj-toluenesulfonazide in methylene chloride
to give the corresponding diazoketones.

^P^C-CD-R' + N3S02_/q\- CH3 -> R:-C0~C-N2 + <t>3P=NTs

Until recently, there were few references in the literature to the absorption
spectra of diazocarbonyl compound s( 27,28) . In the ultraviolet region, there
appeared to be a characteristic band at 245-250 m\i for diazoesters and diazo-
ketones (29,30) . In the infrared spectra, an abnormally low carbonyl frequency
(1630-1660 cm"1) has been attributed to the contributions from the species with
the charge distributions shown above. A strong band at 1535-1410 cm 1,
not observed in the diazohydrocarbons, must be the symetric stretching mode of =CM
which has been shifted to higher frequency by a high degree of conjugation(3l) ->

Earlier observations of the ultraviolet spectra of diazoketones were con-
firmed by Miller and White(32) when they investigated the spectra of a series
of diazocarbonyl compounds. Solvents were found to have only a slight effect in most
of the casBs considered.. However 1 8-bisdiazo~2 ,7— octanedione (III' gave two bands „
one at 247 mu and another at 273 &M-° The 247 mu band is observed in nonhydroxylic

-306-
solvents. In hydroxylic solvents, the "band at 2^7 mu is weakened and a second band
at 273 W- appears. See Figure 1(32). The energy difference is about 11 kcal/raole

between the two bands. A difference of 25
shift normally observed( 33) •

mu is much larger than the 5-10 mu solvent

Figure I

8
6

l.Dioxane \

2.Cyclohexane

3»CH3CN^CH2C12

4.CHCI3

5.C2H5OH

I M \

10-

6.n-C3H70H

7.(CH20H)2
8.n-C4H90H
9.H20
IO.CH3COOH

O
H

230 240 250 260 270 280 290 230 240 250 260 270 280 290
Wavelength in Mu. Wavelength in Mu.

Ultraviolet spectra of III
N2CH-C0-(CH2) 4-C0-CHN2 N2C( CH3) -C0-C(CH3) 3

III IV

In a series of mixtures of acetonitrile and water, the relative intensities of the
two peaks shift smoothly with only a very slight change in the frequency of either band.
The same curves are obtained if water is added to the acetonitrile solution or vice
versa, Fig. II. This family of curves passes through an isosbestic point which consti-
tutes proof of an equilibrium between the two molecular species responsible for the
bands( 3^->35) • The fact that the curves do not form a perfect isosbestic point has been
attributed to the superposition of changes in the refractive index of the solvent on the
shift (32).

Figure II

■p

■H
H

O
H

100

230 2^0 250 260 270 280 290
Wavelength in Mu.
Spectrum of III in mixtures of water and acetonitrile. Numbers are the volume percent

of water.

-307-

Similar plots were also obtained by Fahr(36) for several diazoketones and
diazoesters in a series of dioxane-water mixtures.

The possibility of a keto-enol tautoraerism was considered first. Miller
and White(32) prepared a series of diazoketones with substituents at the positions
a and ocl to the carbonyl. The compound (IV) chosen to represent the case with
no hydrogens available for enol formation was, unfortunately,, unstable. In each
of the four solvents chosen, it gave a band at 2.90- ^Ok mu which increased in
intensity with time and was attributed to decomposition products. The major
absorption at 247-2^9 m(i was always present } but there was no band at 273-275
mu. Those compounds of the series having enolizable hydrogens gave both bands
in the expected manner. Bromine could not be used to test for a keto-enol taut-
omerism because diazoketones decompose in its presence.

Examination of the infrared spectra of these compounds showed no hydroxyl
band at 3200-3^00 cm x in nonhydroxylic solvents. In hydroxylic solvents } where
the enol should be most prevalent, the solvent masked the region of interest.
The carbonyl absorption of III is found at 16^-0 cm x in methylene chloride. It
shifts to 1625 cm a in n-butanol, but it retains the same intensity. Since this shift
may be due to the difference in solvent effects } it does not prove that the enol
does not exist ; but it argues against this possibility.

Turning to the diazo band at 2090 cm x , they noted that the integrated in-
tensity was 21$ less in n-propanol than in methylene chloride. This corresponds
to a 20$ reduction, observed in the ultraviolet spectra. On this basis , they
postulated the existence of a diazo-isodiazo tautoraerism^ citing the observation
of similar forms in diazohydrocarbons(37) • The structure Va could then be assigned
to the absorption at 2V7 mu, while the peak at 275 niu could be attributed to
Vb or Vc either of which should show a =N-H band at 3300-3^00 cm"1. As in the
case of the keto-enol tautoraerism, the nature of the solvents hides the region
of the spectra which would provide proof of the presence of the postulated equilibrium.

9 $ #9 £ 9

N=N=CH~C0-CH2-R H-N=N=C-C0-CH2-R H-N=N=CH-C0-CH-R

Va Vb Vc

In 1959? Fahr (36) reported the spectra of a number of diazocarbonyl com-
pounds j, both diazo ketones and diazoesters , which had no hydrogens available
for either of the tautomerisms described above, but which still exhibited an
isosbestic point in dioxane-water mixtures. He suggested that a hydrogen bonded
complex was formed between the solvent hydroxyl and the carbonyl of the diazo-
carbonyl compound* This is not entirely satisfactory, because it requires that
the intensity of the carbonyl remain constant in spite of the hydrogen bond for-
mation and also that the intensity of the diazo band be lowered by reduction
of the double bond character of the carbon-nitrogen bond.

Foffani and co-workers (38} entered the discussion in 1964 with a study of
the infrared spectra of diazoacetophenone and its derivatives m a variety of
solvents. They also ruled out the possibility of a keto-enol tautoraerism on
the basis of the invariance of the carbonyl intensity. For diazoacetophenone _,
the diazo nitrogen-nitrogen stretching frequency does not change (2108-2112
cm -M over the solvent range investigated. See Table I. The integrated inten-
sity of this band is also nearly independent of solvent. Changing from apolar
to polar solvents or from apolar to commonly hydrogen bonded solvents does cause
the half width to increase.

From this data and the fact that some of their substituted diazoacetophenones
had no hydrogens available but still gave an isosbestic pointy they also ruled
out a diazo-isodiazo tautoraerism.

To test the hydrogen bonded complex suggested by Fahr^, they examined the
behavior of the phenolic hydroxyl group absorption in solutions containing diaz-
oacetophenone,, With a diazoketone: phenol ratio of 20s 1^ they found only normal
hydrogen bonding to the carbonyl. Similar results were observed with ratios of

•=308-

Table I

Solvent

-J m

&)i/.

max

X 10

Am x 10"4

hexane

CCL4

C2CI4

CH2C12

CHCI3

n-butanol

CH3.NO2

2108
2108
2108
2109
2111
2112
2108

18
18
18
22

1.27
1.12
1.15
0.93
O087

0.75
0.91

5*4
5.8
5°2

A™ is the integrated intensity in liter /mole-

10il and 1:1. In these cases , as well as those in which the phenol concentration
would he too low to form appreciable adduct, a second type of carbonyl was detected.

Having disproved each of the three possibilities presented thus far, they
suggested a rotational isomerism. The rotational isomers' stability could be
affected by the intermolecular interactions with the carbonyl. The spectral data
presented are consistent with this view. In addition;, the low carbonyl frequency
suggests a form such as VI, as does the broadening of the nitrogen-nitrogen
stretching band. They note that diazobxides, in which rotation is not possible.,
show only one band in the ultraviolet, that at 245-250 mu.

R-C

'C-R1

In a preliminary communication and later in a paper(39^4o) , Kaplan and Meloy
reported the temperature dependence of the n.m.r. spectra of diazoketones.
Previous investigations (41,42,43) had been made of the cis and trans isomers
arising from restricted free rotation about the central CVN, 0-N, or N-N bond
of amides, nitrites, and nitrosoamines. A similar situation is possible for
diazoketones. Rotation about the central C-C bond may be restricted by inter-
action of the lone pair of the a carbon with the n system of the carbonyl.
This should give rise to two isomers of the diazoketone which may be designated
cis or trans from the geometry of the jt system.

CIS

trans

The temperature dependence of the n.m.r. spectra confirmed the existence
of an equilibrium. At 30°, the spectrum of diazoacetaldehyde consists of two
broad singlets. Raising the temperature to 71° causes both the methine and
aldehyde protons to exhibit time average doublets with a coupling constant of

-309-

2.2 ops. At or below 8°, each region of the spectrum contains a singlet and
a doublet (J = 7*5 CPS) in a 7«3 ratio. In the methine region, the singlet is
at lower field than the doublet. For the aldehydic proton, this is reversed.
Other alky.1 diazoketones, for example , diazoacetone , exhibit similar behavior.
At 30° , they nave a single methine signal that broadens as the temperature is
lowered until, at some temperature, the methine peak splits into a low field
singlet and a higher field doublet. The intensity of the singlet is usually
about nine times that of the doublet. In the ten compounds reported, the low
temperature cis: trans ratio varied from 9° 1 to 1:1. A shift to lower field at
lower temperatures was observed and attributed to hydrogen bonding. The relatively
unhindered cis compounds showed a greater shift than that found for the trans.

Several aryl diazoketones exhibited only a slight broadening at the lowest
temperature at which they were examined. In addition, l-diazo-3j3-<liHiethyl-2-
butanone showed no broadening over a 70° range. Diazoesters have the same type
of temperature dependence as diazoketones, but they do not show two methine
peaks until lower temperatures are reached(-30° to -50°).

The methine region was chosen for study because it was free from other
absorptions, showed the greatest chemical shift difference between the cis and
trans form at low temperatures, and because it was a part of the constant struc-
tural feature of these compounds. Both the keto-enol and the diazo-isodiazo
tautomerisms may be eliminated from consideration by the fact that the 13C=-1H
coupling constant (J = 199 cps) for the methine proton in the time-averaged species
and that of the major species at low temperatures are the same for alkyl diazo-
ketones. In the case of either tautomerism, the same coupling constant would
not be observed for the time-averaged species because the exchange of protons
would average out. This feature of the spectra argues that the C-H bond is es-
sentially unchanged during the equilibrium. Thus the case for rotational iso-
merization is confirmed.

Because the spin-spin coupling constant of trans protons is expected to be
much greater than that for cis protons in such a system, (33*^A5) the low field
methine singlet (J<( 0.3 cpsT~may be assigned to the cis form while the doublet
is assigned to the trans species. Calculations based on the resonance lines
of the methine region were used to obtain the relative populations of the two
isomers, their mean lifetimes(^-6) , and the activation energies for conversion
of the isomers.

Table 11(1*0)

The Equilibrium Constants, Standard Free Energy Difference Between the
cis and trans Forms, the Energy of Activation, and Temperature of Coal-
escence, T , for Diazoketones: RC0CKN2.

R K ( cis-Hrans) A F K T , °C

eq. — — a c

kcal/molc kcal/mole kcal/mole

CH3 0.082(-40°) 1.16 15.5 ±0.9 13.9

C2H 0.063(-40°) 1.28 16.2 +0.6 6.5

4>CH2 0.040(-40°) 1.49 18.2 + 0.6 1.0

CH3O o.859(-50°) 0.07 12.5 + 0.9 -25.0

c2ho o.8ifO(-50°) 0.08 9.0 + 0.8 -32.5

The existence of this cis-trans isomerism must be taken into account when
describing the mechanisms of the reactions of diazoketones since it may direct
the course of the reaction. If the rate of interconversion of the two isomers
is faster than the rate of reaction, this will not be a consideration.

-310-

The demonstrated preference of diazoketones for a cis configuration leads
one to consider the possibility that the Wolff rearrangement to ketenes may pro-
ceed through a smooth concerted process in which the migrating group is trans
to the leaving group. See Figure III. If the rearrangement takes place by in-
itial carbene formation, there -would be no preference for geometry. Evidence
may be presented for both mechanisms. The appearance of hydroxy ketones in.
the decomposition of some diazoketones in water would be more likely to occur
by hydrolysis of the carbene than directly from the diaz ©ketone (^7) .

Figure III

On the other hand, decomposition of VII under thermal, photolytic, or copper-
catalysed conditions(l5) leads to VIII in Q0-927> yield with very little of the
rearrangement product IX. Models of the cis form of the diazoketone (VII) are
very hard to make because of the large steric interaction of the t-butyl groups .
Since the compound exists almost entirely in the trans form, the rearrangement
process postulated above could not readily take place.

I I

(CH3)3C-C~d-C(CH3)3

VTI
((CH3)3C)2C=C=0

R-C-CH2-R1
XII

Q CH2-R'
R-C-C-N2
X

0 CH3
il I

(CKj) C-C-C=C(CH3)2

VIII

R-CO-CH-CH-

-R!

For X, XI,

and XII

<t>-

H-

£-N024>

CH3-

£-CH304>-

£-N02<t>~
£-Cl*-

CH3-
CH3-

CH3-

In a similar manner, X undergoes photolytic and silver oxide catalyzed de-
composition to XI at room temperature, but gives appreciable amounts of XII at
elevated temperatures (48) . The cis form may be more prevalent if the additional
thermal energy is sufficient to overcome the steric interactions.

The possibilities of both keto-enol and diazo-isodiazo taut omer isms In diazo-
ketones have been investigated. They were found to be unlikely. To explain
the persistence of spectral data for an equilibrium in solutions with hydroxylic
solvents, a hydrogen bonded complex was postulated. This was also shown not
to be the case. The existence of a cis-trans isomerism has been demonstrated,
and found to be consistent with spectral data.

-311-

BIBLIOGRAPHY

1. A, Angeli, Gazz. Chem. Ital. , 24, 370 (1895).

2. L. Wolff/ Ann. , 34^, 25 (19.12).

3. W. E. Bachmann and W. S. Struve , "Organic Reactions," Vol. 1, John Wiley and
Sons, Inc., New York, N. Y. , 1942, p 38.

4. R. Huisgen, H. Konig, G. Binch, and H. J. Sturm, Angew. Chem., 7]5, 368 (1961).

5. J. Go Atkinson, M. I. T. Seminars, Spring 1959, p 4ll.

6. R. Huisgen, Angew. Chem., 67, 439 (1955).

7. Jo Lane and R. 1. Feller', J. ' Am. Chem. Soc, 73, 4230 (I95I).

80 B. Eistert, "Newer Methods of Preparative Organic Chemistry," Interscience

Publishers, Inc., New York, N. Y. , 1948, p 513.
9. P. Yates and B. L. Shapiro, J. Am. Chem. Soc, 8l, 212 (1959).

10. F. Haupter and A. Pucek, Chem. Ber. , 93, 249 (1956).

11. E. Wo Robb, Ph.D. Thesis, Harvard, 1956.

12. CD. Nenitzescu and E. Solomonica, "Organic Synthesis," Collective Vol.
2, 496 (1943).

13. 0. Diels and K. Pflaumer, Chem. Ber., 48, 223 (1915).

l4o Mo 0. Forster and A. Zimmerli, J. Chem. Soc, £7, 2156 (1910).

15. M. S. Newman and A. Arkell, J. Org. Chem., 24, 385 (1959).

16. S. Hauptmann, M. Kluge, K. D. Seidig, and H. Wilde, Angew. Chem., JJ_, 678
(1965).

17. P. Yates, H. Morrison, and S. Danishefsky, J. Org. Chem., 26, 2617 (1961).
1.8. Ro Bo Wagner and H. D. Zook, "Synthetic Organic Chemistry," John Wiley and

Sons, Inc., New York, N. Y. , 1953, P 770.

19. Ao Singh, E. R. Thompson, and F. W. Westheimer, J. Biol. Chem., 2J>7, 3OO6P
(1962).

20. W. R. Bamford and T. S. Stevens, J. Chem. Soc, 4735 (1952).

21. J. W. Powell and M. C. Whiting , Tetrahedron, 7, 305 (1959).

22. M. P. Cava and R. L. Litle, Chem* and Ind. , (London) , 367 (1957).

23. M. Regitz, Chem. Ber., 38, 1210.(1965).

24. M. Regitz, Tetrahedron Letters, 22, 1403 (1964).

25. Mo 0o Forster, J. Chem. Soc, 107/260 (I915).

26. G. R. Harvey, J. Org. Chem. ,' 31, 1587 (1966).

27. G. Kortum, Z. Chem. , jjO, 361 OSkl) .

28. A. Jo Ultee Jr. and J. B. J. Soons, Rec Trav. Chim. , 71, 565 (1952).

29. S. A. Fusari, et. al. , J. Am. Chem. Soc , j[6, 2878 (1954)".

30. So A. Fusari, T. H. Haskell, R. P. Fronardt, and Q. R. Bartz , J. Am. Chem.
Soc, 7iL 2Q®1 (1954).

51. A. Foffani, C. Pecile, and S. Ghersetti, Tetrahedron, 11, 285 (i960).

32. F. A. Miller and W. Bo White, J. Am. Chem. Soc, 79, 5974 (1957).

33« J« R« Dyer, "Applications of Absorption Spectroscopy of Organic Compounds,"

Prentice -Hall, Inc , Englewood Cliffs, N. J., 1965, p l4, (b) p 115.
34. Ho Jaffe and M. Orchin, "Theory and Applications of Ultraviolet Spectroscopy,"

John Wiley and Sons, Inc., New York, N. Y. , 1962, p 562.
35« W. West, "Techniques of Organic Chemistry," A. Weissberger, Ed., Interscience

Publishers, Inc, New York, N. Y. , 1949, Vol. 1, Part II, p 1299 .

36. E. Fahr, Chem. Ber., %2, 398 (.1959).

37. E. Muller and D. Ludsteck, Chem. Ber., 87, 1887 (1954).

38. Ao Foffani, C. Pecile, and S. Ghersetti, Tetrahedron, 20, 823 (1964).

39. F. Kaplan and Go K. Meloy, Tetrahedron Letters, 2427, (1964).

40. F. Kaplan and G. K. Meloy, J. Am. Chem. Soc, 88, 95O (1966).

41. M. T. Rogers and J. C. Woodbry, J. Phys. Chem., 66, 540 (1962).

42. P. Graves and 1. W. Reeves, J. Chem. Phys., ^2, 1B78 (i960).

43. G. J. Karabatsos and R. A. Taller, J. Am. Chem. Soc, 86, 4373 (1964).

44. A. J. R. Bourn, D. G. Gilles, and E. W. Randall, Tetrahedron, 20, l8ll (1964).

45. L. H. Piette, J. D. Ray, and R. A. Ogg, J. Mol. Specry. , 2, 66T1958) .

46. F. A. Bovey, E. W. Anderson, F. P. Hood, and R. L. Kornegay, J. Chem. Phys. , 40,
3099 (1964).

-312-

tyj. E. S. Gould,, "Mechanism and Structure in Organic Chemistry/' Henry Holt and

Company, New York, N. Y. , 1959, P 628.
48. V. Franzen, Ann., 602, 199 (1957).

^1 3-

E.S.R. STUDIES OF ORGANIC GROUND-STATE TRIPLET1 MOLECULES

Reported by Robert J. Basalay
INTRODUCTION

April 17, 1967

The following discussion of triplet-state molecules will be confined to molecules
which exist as triplets in their ground-states or are in equilibrium with ground -state
singlet species o The use of esr spectroscopy to observe triplet-states, and the
results of its use to observe the triplet -states of diradicals, methylene derivatives,
and jt -electron triplets will be discussed . A diradical is properly defined as a
molecule with two unpaired electrons whose centers of gravity do not coincide, but
here the term is used to indicate molecules with two unpaired electrons localized
on different atoms which may or may not interact to give a triplet -state 0

THE TRIPLET STATE1'2

numbers j m

The two spin states of an electron can be represented by the spin quantum
_l/2„ If more than one electron is involved, the spin states are
represented by the total spin quantum number, S, given by S- |2(ms) ^1 . If two

electrons have their spins parallel in a given state, S-j (ms) electron 1 + (ms)
electron 2Ul„ The degeneracy of a state due to electron spin is given by 2S + 1.
The S-l state is triply degenerate and a triplet state „ If the electrons are anti-
parallel, S-0 and it is a singlet state.

Two factors determine whether the ground state of a molecule is a singlet or a
triplet , the orbital energies and the strength of the exchange interactions between
electrons of parallel spln0 If we have two degenerate orbitals, the triplet state
is always lower in energy because of the exchange interactions^ but if the orbitals
are of different energy the two electrons may be found paired in the orbital of
lowest energy, to form a singlet state . A competition between the energy separation
of the orbitals and the energy gained by having parallel spins determines whether a
molecule exists as a ground -state triplet or singlet „ If the two factors are
comparable, a thermal equilibrium will exist with comparable amounts of singlet and
triplet .

The degenerate triplet state levels can be split by an external magnetic field
(the Zeeman effect). The triplet levels have the spin quantum numbers m =1,0,-1
associated with them. Mien the levels are split, there are two possible esr
transitions, Ama+1 and Zma<+2o However, the selection rule for esr transitions is
Ams+i0 Hence, the Zsns+2 transition is forbidden and only the Am=+1 transition would
be observed o The resulting spectra would be similar to that of a radical with
S<L/2. However, only exchange and electrostatic interactions between electrons
have been considered.

If the magnetic dipole-dipole
forces between the two unpaired
electrons are considered, the
degeneracy of the triplet is
removed without the application of
an external magnetic field (zero-
field) o Mien the dipole interaction
is considered and zero field splitting
of the triplet levels occurs, the
"forbidden" z^n=+2 transitions becomes
allowed o The appearance of this
transition is not a violation of the usual selection rule, £ms4-l, because it does not
apply at the low magnetic field strength (~1500 gauss, hV^OO Mc./s) where the £m=+2
transition occurs. Hence, there are two types of transitions in the esr spectra of
triplet -states, !,Am=+l" and "&e \Z

A spin Hamiltonian for the two major magnetic interactions, the electronic
Zeeman^and the dipolar, of two unpaired electrons within a molecule is X s g£H«S +
g2^ f Si'S^/r3 - 3(S1»r)(S2'r)/r5J where r is the vector joining the two electrons,
P is the electronic Bohr magneton. The isotropic electron g factor should be an

-1

** — ~

anisotropic g tensor, but for most organic triplets the g tensor anisotropy is
small and will be neglected here.

The second term of X is the dipolar interaction term )i p. Expanding Xj) by means
of the various vector products and expressing everything in terms of the total spin S,
we obtain XD ■ l/2g^2 (s|(r2-3x2)/r5 + S§( r2-3y2) /r5 + S2( r2-3y2) /r5 - (SxSy + SySx) '
3xy/r5 - (SySz + SzSy)3yz/r5 - (SXSZ + SZ5X) 3xz/r5J „ This result can be expressed
in the matrix form:;

'S\ /(r2-5x2)/rs -3xy/r5 3xz/r5 \ ( SSS)

l/2g .

3xy/r5
=3xz/r"

(r2-3y^r5 -3yz/r5
3yz/r5 (fr2-3z^r5,

S-D-S

D is a symmetric tensor called the zero-field splitting tensor which may be diag=
onalized by proper choice of coordinate systems. In terms of the new coordinate
system (x%y

z»)

X=(r2-x^22/r5, Y=(r2-y,2)/r5, and Z-(r2-z'2)/r5, and the spin

Hamiltonian becomes X = gpH-S-XSj-YS^-ZS2. Since the tensor is traceless (X+Y+Z-O),
the zero-field splitting can be expressed in terms of just two independent constants,
D=l/2(X+Y)-Z and E=-l/2(X-Y) , giving X s D(S2-l/3S2) + E(sf-S2) + gBH'St

In order to find the energy level splitting in the triplet state due to zero-
field splitting j the eigenvalue problem, W<§ =<8>fc , must be solved. The spin
functions will be chosen to diagonalize the zero-field Hamiltonian matrix where

(V»o

ff^Z*

field, H,

X-Ei

igBHz
-igpH,

-igpHz

Y~E2

igpHx

The solution of the eigenvalue problem is given by the matrix, T o

If there is no external field
( Hx=Hy=Hz=0) , the matrix is diagonalized
n and E1=X> E£=Y, and E3=Z, the energies
of the levels of the triplet at zero
magnetic field.

If we impress an external magnetic
on our system parallel to the z direction, H-Hz and Hx=Hy=0, and the
matrix has the following solutions; Ej=l/2(X+Y) + l/2(X-Y) tan 0+ g0H, Eg=l/2(X+Y)-
l/2(X-Y) tan 0 - gpH, and Ea=Z, where tan 2Q = (X-Y)/2gPH. A set of graphs
simulating the zero-field splitting when H=HZ, Hy, and Hx is given in Figure l.

Figure 1 Zero-Field Splitting of Triplet Energy Levels

Magnetic Field Strength (H)-

-315-
The arrows in Figure 1 represent electron spin resonance transitions at constant hv
while varying H.

The esr spectra of a triplet whose molecules were all oriented the same way
would have three transitions whose position would vary as the orientation of the
triplet with respect to the external magnetic field were changed . An example of
this case will be seen when the esr spectrum of diphenylmethylene is discussed.
Mien the triplet molecule is dissolved in a liquid, the rapid molecular tumbling
averages out the effect of dipole-dipole interaction on the esr spectra of randomly
oriented molecules. Mien the triplet molecule is frozen in a glass, the effect of
the dipolar interaction is not averaged out for the randomly oriented molecules. The
intensity of the esr signal is greatest when the external magnetic field is parallel
to a principal magnetic axis of the triplet. The derivative curve of an esr
absorption spectrum gives special prominence to the magnetic field strengths which
cause esr transitions to occur in the molecules which have principle magnetic axes
parallel to the external magnetic field „ This type of triplet esr spectrum is
simulated in Figure 2. Approximate zero-field splitting parameters can be obtained
from studies of the Am=±2 transitions whose anisotropy is relatively small, allowing
them to be easily observed for randomly oriented triplet molecules in frozen matrices.4

Figure 2. Simulation of ESR Spectra of Randomly Oriented Triplets3

Absorbtion
Spectrum

E=0

E^O

ST-J H

A A

Derivative
Spectrum

DIRADICALS

Certain organic compounds exhibit a chemical behavior characteristic of free
radicals, but contain even numbers of electrons. One of these compounds is
Chichibabin's hydrocarbon, p,p» -biphenylene-bis-(diphenylmethyl) (I). This compound

Triplet
(paramagnetic)

- AS-O-O-fo

Singlet
(diamagnetic)

\ //

was thought to exist in two forms, a triplet state and a singlet state. The fact
that an esr signal is observed for Chichibabin's radical in solution indicates a
paramagnetic species is present.5

Another type of diradical (II), where X is a bridging group (i.e. -0-, (-CH2-)n,)>

can not have a quinoid singlet state. If there is
X yy Ciflo spin interaction between the two halves of the diradical,
vZy there will be a ground -state triplet and an excited -

state singlet. If there is no spin interaction, the
halves of the molecule will be independent of one another and behave like two mono-
radicals. Compounds of this type (II) were examined by Jarrett, Sloan, and Vaughan.6'7
The sharpness (esr linewidth <-~ 12 gauss) of the esr spectra indicated a comparatively
small interaction between the unpaired electrons. The larger the spin interaction,
the broader the esr spectrum becomes because the spin-spin coupling produces strong
relaxation effects which decrease the relaxation time increasing the esr line-width.

-316-

Ill IV V

The spin exchange ( spin interaction between the two radical centers) in several
biradicals (I, III, IV, and V) which were labeled with C13 at the triphenylmethyl
carbon atoms was studied by means of the resultant nuclear hyperfine splitting
observed in the esr spectra. If the spin interaction is large, the esr spectra will
show nuclear hyperfine splitting due to two equivalent C13 (1=1/2) atoms. This is a
triplet -state. If the spin interaction is small, the esr spectra will show nuclear
hyperfine splitting due to one C13 atom. This is equivalent to a pair of doublets
or a triplet with very weak spin interaction. The latter case was observed.

However, Weissman8 has shown that the electron interaction between the aromatic
rings of the anions of paracyclopheaes (VI), provided n or m is 1 or 2, is large (i.e.
CqH4— (CH2)m "tne nuclear hyperfine splitting due the hydrogen atoms of
X . X both aromatic rings are observed in the esr spectra) . If n,
(CHaJn-CsH* m > 3, then the nuclear hyperfine splitting due to the

VI hydrogen atoms in one aromatic is observed in the esr spectra

indicating small electron interaction between the aromatic
rings. MeConnell9 has performed theoretical calculations for the ground -excited
state splittings in large biradicals of this type (II) and according to the calcu-
lations a larger spin interaction is indicated. The spin interaction observed for
a -CHsCH^- bridge (i.e. X in II) is equivalent to a calculated spin interaction
resulting from a bridge of about five -CH2- groups. This discrepancy between the
spin interaction observed and that one would expect is the "biradical paradox."

A possible explanation for the smaller than expected spin exchange was offered
by Bersohn. 10 One or two -CH2- groups do not present a serious barrier to the
passage of a single electron, but it is a formidable barrier to the spin interaction
of two electrons of parallel spin.

1 CI

VII VIII

An examination11 of the esr spectra of compounds I, VII, VIII, and DC indicated
that the paramagnetic species in biradical solutions is a dimer or higher polymer.
As a result the spin interaction between electrons would be much smaller than would
be expected for the monomer. It was found that the room temperature intensity of
the esr signal can be maintained at low temperature if the specimen is cooled rapidly
enough. If the temperature is lowered slowly the intensity of the esr signal decreases,
Also, when the specimen is heated, the esr signal is enhanced. If the heated solu-
tion is cooled to room temperature, the esr signal diminishes to a final intensity of
slightly less than the starting intensity. If the specimen is cooled to a temperature
at which equilibrium is slow, the intensity of the esr signal can be enhanced by
illumination. The enhancement persists until the solution is heated to a point where
the equilibrium is rapid. Thereupon the intensity diminishes again. These effects
indicate that biradical monomers associate in solution to form dimers or polymers.

In two communications, Chandross and Kreilick12 reported attempts to produce
a triplet molecule by linking two phenoxyl radicals by a -CMe2- bridge (XII). No
triplet spectra were observed, but chemical and spectral data indicated the reaction
scheme on the following page. The infrared spectrum of the oxidation product (XIII)
shoved no band at 6,h u, characteristic absorption of phenoxyl radicals, but there
was a doublet, characteristic of this type (XIII) of cyclohexadienone , at 6.08 and
6,16 u. The presence of the biradical (XII) in equilibrium with XIII is indicated
by a rapid reaction of the oxidation product with 02, The dienone (XHI) would not
be expected to react rapidly with oxygen (reaction complete in minutes at room temp.),
but the biradical (XII) would.

Fb02£
Ag20,
MnOo

To make the internal formation more difficult, they then replaced the -CMs2-
group with a 2,2' -biphenylene group (XIV). When XIV was oxidized with aqueous
alkaline f erricyanide , the infrared spectrum of the oxidation product in CCI4 had
a pair of strong absorptions at 6.0 and 6.2 u assigned to XVI and a strong band at
6.4 [i which was assigned to XV. The esr spectrum revealed the presence of triplet
species. There were two sets of /^m=il lines as well as a Am=±2 transition. The
larger splitting was attributed to XV and the smaller was thought to be due to a
dimer possibly linked by a peroxide bond (XVII) . If this smaller splitting is due
to the dimer (XVII) , it would indicate that the unpaired electrons can interact
relatively strongly across long organic molecules, possibly to the degree which could
be predicted by theory without creating the "biradical paradox."

XIV

XVI

0-0

XVII

A pyridinyl diradical has been prepared and examined by Kosower and Ikegami
Their esr studies of 1,1" ethylene -bis -(4-carbomethoxypyridinyl) ( XVIII) indicates
radical-radical association producing dimers or polymers. The ends of the n-mers
behave like monoradicals as in the case of Chichibabin's hydrocarbon. The esr
spectrum of XVIII at, 77°K at low concentration is" characteristic of a triplet with

CH302C^^n^-CH2-CH2-1<^>-C02CH3 ^ffi On30^-^\.Q^^R2

-N^)-C02CH3
XVIII

D=0.0178 cm""1 and E=0.0017 cm"1. With increasing concentration, the triplet spectrum
disappears as a strong signal due to the monoradical-like ends of the polymers appears,
Again, it would be interesting to see whether the spin interaction here could be
predicted without creating a "biradical paradox. "

Young and Castro14 prepared a stable biradical (XXI) which Kreilick15 studied by
esr. The esr spectrum at room temperature of the oxidation product of XIX initially
gave an esr signal with hyperfine splitting due to four equivalent protons, from the
formation of XX. On further reaction, an esr signal appeared with splitting due to
six equivalent hydrogen atoms, from the formation of XXVII. Apparently, the unpaired
electrons can delocalize into the three aromatic rings to interact with all six meta
protons. The biradical (XXVII) esr spectrum consists of two pairs of lines disposed
about g=2. The separation of the inner more intense lines is 38. k gauss.

-318-

Since D«: l/r3 where r is the average distance between the two unpaired electrons
and the proportionality constant is known from theory,1 a separation of 9 A between
the two unpaired electrons is calculated. The degree of spin interaction here is
less than observed for XV .

OH 0' 0*

PbO:

MTHF

XXX
METHYLENE DERIVATIVES

XXI

The first methylene compound to be studied by esr was diphenylmethylene „ Murray,
Trozzolo, Wasserman, and Yager16*17 irradiated a solution of diphenyldiazomethane in
"Flororolube " (polyehlorotrifluoroethylene) at 77°K with an Hg arc. The esr spectrum
consisted of derivative signals at 1227, 1619, ^588, 5272, and 7522 gauss. The
intensity of the spectrum did not decrease as long as the temperature was kept below
123°K, indicating the paramagnetic species was the ground -state or was at least in
thermal equilibrium with the singlet-state. The five line spectrum corresponded to
a triplet esr spectrum of randomly oriented molecules having a spin Hamiltonian with
D = 0.401 cm"1 and E = 0.018 cm"1.

Brandon, et al.18 subsequently presented the results of a detailed study of the
esr spectra of ground -state triplet diphenylene molecules oriented in benzophenone
single crystals. A preliminary examination of the esr spectra revealed that there
were four differently oriented sets of principle axes indicating four equivalent but
differently oriented positions in the benzophenone unit cell could be occupied by
the triplet species. The major magnetic axes of the triplets in one of these
equivalent sites were determined. If the magnetic field were made parallel to one
of these axes and the crystal was rotated about this axis, the esr line positions at
several fixed angles of rotation were measured. The variation of the magnetic field

Figure 3. Variation of Triplet Line
Position with Rotation about a Magnetic
Axis. (The points are determined from
separate esr spectra and curves are
drawn to fit them.)

H( magnetic fie!

strengths necessary for resonance with respect to the angle of rotation is simulated
in Fig. 3" The angles of rotation where maximum separations occur indicate the
positions of the other two axes. The principle magnetic axes were determined relative
to the crystallographic axes of the benzophenone single crystal. The measurements
of the stationary magnetic field strengths necessary for resonance versus the angle
of the external magnetic field in the plane of the principle magnetic axes were fitted
by the spin Hamiltonian X - I?«g»(0)S + D(Sf -J./3S2) + E(S| -_S|) „ Those values of
D and E which gave the best fit are 0.4050 cm"1 and 0.0192 cm"1, respectively.

The zero-field splitting parameter (D) is very large compared to diradicals
(D^O.Ol cm"1) and jt -electron triplets (D"O.20 cm"1). Since Dgel/r3, the large value
of D indicates the electrons in the diphenylmethylene are on the same carbon atom a
large fraction of the time. Higuchi19 has calculated the value of D for CH2#
methylene, where the two unpaired electrons are each in one of two orthogonal p orbitals,

-319-

to be 0,90,55 cm"1. The D value of diphenylene is smaller than Dqjj indicating the
unpaired electrons can delocalize into the phenyl rings, Skell2° has proposed a
linear structure for diphenylmethylene with D^ symmetry, with each orthogonal p
orbital conjugated with a phenyl ring. If the triplet were linear the zero-field
splitting parameter E would be zero 3, but this is not observed. Hence, the diphenyl-
methylene molecule is bent. The structure of diphenylmethylene probably has one
electron generally localized at the divalent carbon and the other conjugated with the
phenyl rings. This bent structure for methylene compounds has been confirmed by the
observation of geometric isomers of 1- and 2-naphthylmethylenes by esr.21 The esr
spectra indicated the presence of two sets of triplet species whose zero-field
splitting parameters are similar indicating similar species and the difference
between the zero-field splitting parameters of each isomer does not vary as the host
(the frozen matrix) is changed,

XXIV

XXV

XXVI

The zero field splitting parameters for many methylene compounds have been
determined .17i'18''21"26 The D (O.5I8 cm"1) value of phenylmethylene is larger than
the D of diphenylmethylene as expected since two phenyl groups would allow greater
derealization of electrons, lowering D. The compounds XXII -XXVII have D values
(0.37-0.42 cm"1) similar to that of diphenylmethylene indicating a similar electronic
structure (i0e. one electron localized at the divalent carbon atom and one delocalized)
Propargylene derivatives26 have relatively large values of D (0. 55-0. 63 cm"1, but one
would expect that the unpaired electrons could be delocalized into the triple bond.
These compounds also have E — 0 indicating a linear geometry at the methylene group.
If the unpaired electrons delocalize^, they would go into two perpendicular it systems
(see Fig. V» . Then, each carbon atom of the conjugated chain would have some fraction
of triplet methylene character according to the spin density of the jt-system at the
carbon atom. Thus, the effect of derealization of the unpaired electrons to lower
D is nulified. Theoretical calculation indicate that negative spin densities must
be used to predict the experimental value of D.

H —

1/ TPfy

c— c — c—

1<tnlll , / 1

-R

delpcaliz at i on

Liwii|yi.i»|y
H — C — C — C—

•R

Figure ka De localization of Unpaired Electrons in Propargylene (H-C-CeCH) Derivatives
jt-ELECTRON TRIPLETS

The possibility that certain derivatives of cyclic k-n jt-electron systems could
have triplet ground states is suggested by simple molecular orbital theory. This is
based on the degeneracy of the highest occupied molecular orbitals. These orbitals
are degenerate only if the system is symmetrical. If the symmetrical system is
distorted ( Jahn-Teller effect) , the degeneracy of the orbitals is lost. If two
electron are placed in the degenerate orbitals, a ground -state triplet results, but
if the orbitals are not degenerate both electrons may occupy the lowest orbital
giving a singlet.

Cyclic aromatic compounds with three fold or greater axes of rotation, have two
degenerate unfilled lowest anti-bonding it-electronic levels. If dinegative anions
were made from the addition of two electrons to anti-bonding orbitals, they could be
triplet states. Jahn-Teller effects could remove the degeneracy to allow the two
electrons to occupy one orbital and form a singlet ground-state. The simplest case,
benzene dinegative ion, has not been reported. However, the esr spectra of the
dinegative ions of triphenylbenzene27 (XXIX) , aecacylene^l'XXVTII) , and coronene28
(XXX) have been reported. The coronene dinegative ion did not give a very intense
esr spectrum in solution, but if the di-potassium coronene salt is crystallized from
tetrahydrofuran, the solid obtained gives a triplet esr spectrum whose signal

■320-

XXVIII

XXIX

XXX

intensity varies intensively with temperature. At -l60°C no triplet signal was
observed. Increase in temperature resulted in a continuous increase of esr triplet
signal intensity up to +60°C where the sample decomposed. This temperature dependence
indicates a thermally excited triplet -state. The zero-field splitting parameters (D)
for (XXVIII), (XXIX), and (XXX) were respectively 0.021 cm"1, 0.0^2 cm"1, and 0.^2.
0rO6O cm"1. Since D*cl/r , the magnitude of D should vary inversely with the size
of the jt-electronic level the unpaired electrons occupy, and it does.

Cyclopentadienyl cations, which may have triplet ground-states according to
simple molecular orbital theory, have been studied by esr.29 Penta-P-naphthyl

Energy^--

Figure 5. Energy Level
Diagrams for Symmetrical and
Distorted Cyclopentadienyl
Cation

symmetrical

distorted

cyclopentadienyl cation did not show a triplet esr spectrum at any temperature. The
pentaphenyl and pentachloro derivatives gave very distinct triplet esr spectra. If
the molecule has a triplet ground -state and the excited singlet state is not appre-
ciably populated, or the singlet and triplet states are equal in energy, the
intensity of the esr spectrum should follow Curies law (i.e. intensity is inversely
proportional to temperature). If the triplet is not the ground -state, the intensity
of the esr signal due to the triplet still obeys Curie's law, but the concentration
of the triplet varies with temperature being a thermally excited state, and thus
Curie's law is not obeyed. The esr spectrum of the pentaphenylcyclopentadienyl
cation did not obey Curie's law and the triplet is a thermally excited state 9 but
the esr spectrum of the pentachlorocyclopentadienyl cation did obey Curie's law and
the triplet is the ground -state or very close to it. The symmetry of the cyclo-
pentadienyl ring would be expected to be distorted more the larger its substituents
become. The more distorted the ring is the more likely the ground -state will be a
singlet, and it is observed as the substituents (P-naphthyl> phenyl> chloro) on the
cyclopentadienyl ring become more bulky the triplet is less stable. The zero-field
splitting parameters for the pentachloro and pentaphenyl derivatives were D = 0.1^95
cm"1 and D - O.IO5O cm"1, respectively, as would be expected since the pentaphenyl
derivative would allow the two unpaired electrons to have a greater average distance
between them than in the pentachloro derivative.

CONCLUSION

Many examples of triplet-state species have been presented which can be
observed by esr spectroscopy. The zero-field splitting parameters can be used to
infer information about the electronic structure of the triplet species. By use of
esr to study reaction mixtures, it should be possible in some case to detect and
characterize triplet reaction intermediates.

BIBLIOGRAPHY

1. A. Carrington and A. D. McLachlan, "Introduction to Magnetic Resonance,"
Harper and Row, New York, N.Y. , 1967.

-321-

2. K. G. Harbison, M.I.T. Seminar in Organic Chemistry, Fall, 1963.

3. E0 Wasserman, L. C» Snyder, and W. A. Yager, J. Chem. Phys., 4l, 1763 (1964) .

4. M. S. deGroot and J. H. van der Waals, Mol. Phys. , 2, 333 (1959) » Mol. Phys.,
3, 190 (1960)o

5„ Co A. Hutchison, Jr., A. Kowalsky, R. C0 Pastor, and G„ W. Wheland, J. Chem.
Phys., 20, 148.5 (1952).

6. H„ S. Jarrett, G. J. Sloan, and W. R. Vaughan, Jc Chem. Phys., 25_, 697 (1956) .

7. G. Jo Sloan and W. R„ Vaughan, J0 Chem. Phys., 22, 750 (1957).

8. S. I, Weissman, J. Am„ Chem. Soc, 80, 6462 (195B).

9. H. M. McConnell, J. Chem. Phys., 27, 115, 968 (1957).

10. R. Bersohn, Ann. Rev. Phys. Chem., 11, 382 (i960) .

11. R. K. Waring, Jr. and G. J. Sloan, J. Chem. Phys., 40, 772 (1964).

12. E. A. Chandross and R. Kreilick, J. Am. Chem. Soc.,~B£, 2530 (1963) j J. Am.
Chem. Soc, 86, 117 (1964).

13. E. M. Kosower and Y. Ikegami, J. Am. Chem. Soc, 89, 46l (1967).

14. N. C. Yang and A. J. Castro, J. Am. Chem. Soc, 82, 6208 (I960).

15. R. Kreilick, J. Chem. Phys., 4^, 308 (1965).

16. R. W. Murray, A. M. Trozzolo, E. Wasserman, and W„ A. Yager, J. Am. Chem. Soc.
84, 3213 (I962).

17. A. M. Trozzolo, E. Wasserman, and R„ W. Murray, J. Am. Chem. Soc, 84, 4990
(1962).

18. R. W. Brandon, G. L. Closs, C. E„ Davoust, C. A. Hutchison, Jr., B. E. Kohler,
and R. Silbey, J. Chem. Phys., 4^, 2006 (1965) .

19. J. Higuchi, J. Chem. Phys., ^8, 1237 (I963).

20. R. M. Etter, H„ S. Skovronek, and P. S. Skell, J. Am. Chem. Soc, &L, 1008

(1959).

21. A. M. Trozzolo, E. Wasserman, and W. A. Yager, J. Am. Chem. Soc, 87, 129

(1965).

22. E. Wasserman, L. Barash, A. M. Trozzolo, R. W. Murray, and W. A. Yager, J. Am.
Chem. Soc, 86, 2304 (1964).

23. I. Moritani, S. I. Murahashi, M. Nishino, Y. Yamamoto, K. Itah, and N. Mataga,
J. Am. Chem. Soc, 89^ 1259 (1967).

24. E. Wasserman, J. Chem. Phys., 42, 3739 (1965).

25. E. Wasserman, L. Barash, and W„ A. Yager, J. Am. Chem. Soc, 87, 4974 (1965).

26. R„ A. Bernheim, R. J. Kempf, J. V. Gramas, and P. S. Skell, J. Chem. Phys.,
i£, 196 (1965).

27. R„ E. Jesse, P. Biioen^, R. Prins, J. D. W. van Voorst, and G. J. Hoijtink,
Mol. Phys., 6, 633 (I963).

28. M. Glasbeek, J. D. W. van Voorst, and G. J. Hoijtink, J. Chem. Phys., 45, I852
(1966).

29. R. Breslow, H. W. Chang, R. Hill, and E. Wasserman, J. Am. Chem. Soc, 89,
1112 (I967).

-322=

RECENT STUDIES CONCERNING THE MECHANISM
OF THE FAVORSKII REARRANGEMENT

Reported by Peter A. Gebauer April 20, 1967

I. Introduction

The Favorskii rearrangement of a-haloketones to derivatives of carboxylic
acids by the action of bases has been reviewed. 1>2*3'4 A general review of reactions
of a-haloketones with bases has also appeared.5 The purpose of this seminar
will be to present the recent work related to the mechanism of the Favorskii
rearrangement, especially that concerning the question of whether a cyclopro-
panone intermediate, If formed } is generated via an intermediate zwitterion or
delocalized species,, or whether it is formed in a concerted manner.

II o Historical

The labeling experiments by Loftfield which indicated that the a and a'
carbon atoms of 2-chlorocyclohexanone become equivalent during the course of the
Favorskii rearrangement limit the possibilities for the mechanism of the Favorskii
rearrangement. Although one case will be discussed in which the a and a' carbons
do not become equivalent } it appears that generally the reaction must proceed
through either a cyclopropanone (i), or through a planar species (II)7-*8 which is
probably best described as a diradicalj, but in keeping with the usage of most
of the workers in this field, this mechanism will be referred to as the zwitter-
ionic mechanism. Usually the intermediate (II) is considered to form a cyclo-
propanone in the course of the rearrangement.8

0 0

II II

_c/CxcV~~*-c/C

"1 I ' I

I II

Burr and Dewar8 argued against the concerted formation of the cyclopropa-
none on the grounds that the geometry of the p orbital on the carbanion is not
correct for nucleophilic displacement of the halogen.

III. Reactions of Cyclopropanone s

Breslow and co-workers have reported the interception of a cyclopropanone
under conditions similar to those of the Favorskii rearrangement. Cycloprope nones
such as IV were obtained by treatment of the dibromides III, Vs and VI with tri-
ethylamine in methylene chloride or chloroform.

RCHBr-CO-CHBrR -Et3N j
III R=<!>
V R=n-C4H9

Turro and Hammond10 found that cyclopropanones will react to form products
analogous to those of the Favorskii rearrangement. Treatment of 2,4-dimet-hyl-
2-bromo=3-pentanone ( VTl) with sodium methoxide in ether produced only 12$ of
the rearrangement product VIII while tetramethylcyclopropanone, IX, the expected
intermediate in the Favorskii rearrangement of VTl and the hemiketal of tet-
ramethylcyclopropanone (X) produced 97$ and )>98$ of VIII respectively. The hemi-
ketal (X) in refluxing methanol also produced 24$ of VIII. In all cases, XI
was the only other product reported.

-323-

CH3

0 OMe
(CH3)2CH-C-C(CH3)2

(CH3)2CH-C-C02Me
1

CH3 VIII

^a0Me > 973$
Me OH or DME yu

Me<X /OH

>9&$ <1#

MeOH(refl.)
>

2k%

(CH3)2CH-CO-CBr(CH3)2 ^gg ) 12$

VII

Similar results were obtained with 2,2-dimethyIcyclopropanone.11
Turro and Hammond10 concluded that VII does not form either the cyclopro-
panone or the hemiketal as the major product when treated with "base. They
also conclude that the cyclopropanone is in equilibrium with an acyclic species
such as II. This is not necessarily required since it is possible that methanol
might add directly to the cyclopropanone as suggested by House.12

It would seem from a comparison of the results of the reactions of the
bromide in base and the hemiketal in refluxing methanol that cyclopropanones
might form ketonic solvolysis products such as XI if the concentration of base
were low enough. However, other means of formation of these products must, in
general, be operative also.

IV. Tb.e Question of a Delocalized Intermediate

Fort13 suggested that the intermediate (II) proposed by Aston and Newkirk7
and Burr and Dewar8 might not close to a cyclopropanone but react directly with
solvent to form ketonic solvolysis products or with base to form rearrangement
products. He suggested that XII might be the most efficient way of describing
the intermediate.

^ £K -*-

L I ~ I I I I

XII

As indicated in XII, Fort meant to imply that there was some overlap between
the radial, carbons and oxygen but he conceded14 that this would be slight due
to the large distances between the radial atoms.

Fort13 found that a-chlorodibenzyl ketone (XIII) reacted with lutidene in
methanol to produce only a-methoxydibenzyl ketone (XEV). Desyl chloride (XV) and
a-chloroacetone (XVI) did not react under these conditions.

-324-

-CH-

WCH3 ?Me

45CH2-CO-CHCl<t>

OH *

$CH2-C0-CH4>

XIII

XIV

CH3-CO-CH2Cl

XVI

It was assumed that the delocalized intermediate vas formed in the first case
hut not in the last hecause the base is not strong enough to form it in the
absence of stabilization by phenyl groups.

Fort14 also treated compound XVII with varying concentrations of base.

NaOMe

MeOH

C02CH3

XVII

XVIII

XIX

He found that as the concentration of base decreased, the yields of rearrangement
products also decreased., and concluded that a delocalized intermediate } which
can react with base to form rearranged products (XIX) or with solvent to form
ketonic solvolysis products ( XVIII.) was being formed.

Fort's13'14 conclusions seem to rest on the tacit assumption that a cyc-
lopropanone would not give ketonic solvolysis products in the absence of strong
base. However, as noted earlier, the work of Turro and Hammond11'12 seems to
indicate that all the reactions of the delocalized intermediate could as easily
be explained by a eye lopropanone.

Fort,15 Richey and co-workers,16 and Cookson and co-workers17 have trapped
adducts which they all assumed were formed by the addition of an acyclic species ,
such as XEI, to furan. Hammond and Turro11 have also obtained a furan adduct of
2 ,2-dimethylcyc lopropanone. Although these workers18 suggest that the cyclopro-
panone may be in equilibrium with some acyclic species, it might be possible for
the cyclopropanone itself to form this adduct.

V. Solvent Effects and Stereospecificity

The conflict between Loftfield's6 concerted mechanism and the stepwise
zwitterion mechanism7'8 cannot be resolved by kinetic determinations, but the
stereochemistry of the products is different for the two mechanisms.1 Where
two stereoisomeric products are possible , the concerted mechanism predicts in-
version of configuration at the a carbon bearing the halogen while the zwitterion
mechanism should proceed with racemization.

Stork and Borowitz19 found that the cis and trans isomers of 1-chloro-l-
acetyl-2-methylcyclohexane (XX) underwent Favorskil rearrangement with nearly
complete inversion and concluded that the cyclopropanone formation was concerted.

House and Gilmore20 found, however, that the stereospecificity of the reaction
seemed to depend on the polarity of the solvent. The cis isomer of XX produced
essentially racemized product when the rearrangement was carried out in methanol,

-325-

and mainly inverted product in dimethoxyethane (DME). It was suggested that this
solvent effect was due to polar solvents facilitating the loss of chloride to
form the planar intermediate (II) while the concerted mechanism predominates
in nonpolar solvents.

Similar results were obtained21 using piperitone oxide XXI and isophorone
oxide XXII as models for a-haloketones.

XXI XXII

Tchoubar and co-workers22'23 found a similar solvent effect with compound XX,
shown in Table I.

T,ule I

E-Vect of Solvent on Stereospecificity of Favorskii Rearrangement

Base Solvent fo Inversion jo Retention

NaOMe DMSO 75-5 2^.5

(")* (DME) (95)* (5)

t~BuOH 83.5 1.6.5

KOH H20-Pyridine 37 63

^Values in parenthesis are from House20 for purpose of comparison.

Tchoubar22 argued that solvation of the anion was the factor which deter-
mined the solvent dependence. They assumed that in a nonpolar solvent the negative
charge of the anion (XXIIl) resides on the a! carbon which, they say, would be
tetrahedral. In a polar, hydrogen bonding solvent, the solvent would be pre-
sumed to hydrogen bond with the oxygen and localize the negative charge on the
oxygen atom. In this case they expected the a1 carbon to be hybridized sp2 and
planar.

0
II

VC\ ci

-fa'

ay\

XXIII

They therefore concluded that the critical difference between the suggestions

of Loftfield6 and Burr and Dewar8 is whether the a' carbon is tetrahedral or planar.

It is unlikely that a tetrahedral carbanion would have a very long life-
time when the possibility of derealization of the charge exists. However, the
carbanion might react rapidly before rehybridization. This would require that
the anion displace the chloride either very soon after, or simultaneously with, loss
of the proton. This possibility can be checked by labeling studies. If the
reprotonation is faster than or as fast as the loss of chloride, the anion prob-
ably has a lifetime long enough to form the more stable planar species.

House and Thompson24 found that essentially no deuterium was incorporated
into unreacted 9-chloro-trans -I-decalone when treated with sodium methoxi.de in
deuterated methanol.

Olson25 found that no deuterium was lost from the C-k position in unreacted
2-a-bromo- anS 2-a-iodocholestan-3-one-2,4-d3 (XXIV) , but that the corresponding
chloro compound lost 20/6 of the label after one half life of the Favorskii rearrangement

-326-
when treated with 0.2 M sodium ethoxide in absolute ethanol.

XXIV

X=C1, Br, I

Olson concluded that in the bromo- and iodo- compounds loss of a proton
is rate determining while in the chloro compound, loss of chloride occurs at
approximately the same rate as reprotonation. This conclusion was verified by
the observation that the trideuterated compounds above lost halide at one-fifth
the rate of their hydrogen analogues.

Mayer26 and Tchoubar and co-workers27 studied the Favorskii rearrangement
of 1-acetyl-l-chlorocyclohexane (XXV), using sodium phenoxide as the base in
the presence of deuterated phenol.

:och3

XXV

They found that about half a mole of deuterium was incorporated per mole of
starting material which was recovered and concluded that the first step, at
least in this case, of the Favorskii rearrangement is a pre-equilibrium between
the anion and starting material. This conclusion is not necessarily valid since
it is possible that the reaction which is responsible for the deuterium in-
corporation may not be related in any way to the Favorskii rearrangement.

House and Gilmore21 also found deuterium incorporation in unreacted
isophorone oxide and piperitone oxide.

It appears that the rate of reprotonation of the anion is quite close to
that of loss of halide. The anion discussed earlier therefore probably has a
fairly long lifetime and is planar with the negative charge delocalized on to
both the carbon and oxygen atoms.

House and Thompson24 suggested that an equatorial halogen would favor the
concerted formation of a cyclopropanone while an axial halogen should favor
formation of the zwitterion or delocalized intermediate. However, work by
House and Frank12 ■,24 using both the cis and trans isomers of 9-chloro-l-decaIone
(XXVl) led to the conclusion that the ketonic solvolysis products did not arise
from a planar intermediate in either case since the cis and trans isomers pro-
duced different product ratios. Since the same planar intermediate should arise
from both isomers, the products also should have been the same. Ketonic sol-
volysis products were obtained from both isomers in methanol and rearrangement
products were obtained in DME. The possibility of direct nucleophilic displace-
ment was ruled out since the 9-methoxy compounds were formed mainly with retention
rather than inversion.

The trans isomer produces mainly the products (XXVII) which would be formed
by the formation of the most stable carbanion, while the cis isomer produces
mainly the unexpected products (XXVIIl).

-327-

C02Me

H C02Me

XXVII

H
b

H
a

XXVIII

COpMe

The stereochemistry at the 9-Posi'tion in "the 9-carbomethoxy derivatives
(XXVTl) was consistent with a concerted formation ox the cyclopropanone. That
is, the cis isomer formed XXVIIa and the trans isomer formed XXVIIb. The ke-
tonic solvolysis products were suggested to be formed via either an alkylidine
epoxide (XXIX) or the hemiketal of the cyclopropanone (XXX).

3--

XXIX

XXX

House concluded that the conformation of the halogen did not affect the
mechanism of the Fa-vorskii rearrangement.

Smissman and co-workers28 also have studied this effect. They subjected
>a-hromo-trans-2-decaIone ( XXXI) ^ >e-bromo-tranS"2-decalone (XXXIIa) and
2-e-bromo-9-methyl-trans-3-decalone (XXXIlb) to Pavorskii rearrangement con-
ditions in polar (EtpH) and nonpolar (DME) solvents. The axial coumpouni ( XXXI)

XXXLI

a R=H

b P.=CH3

in both solvents produced only products which were presumed to arise from the
epoxide (.XXXIII). No rearrangement products were obtained.

■-CEt

XXXIII

The non-methylated equatorial isomer ( XXXIIa) formed both the products found
in the axial case and some rearrangement product in both solvents. The methyl-
ated equatorial isomer ( XXXIlb) produced 38 to k-hfo rearrangement product (XXXIV)
a diacetic acid (XXXV) and 9-methyl-trans~decaiin-2 , 3-dione (XXXVl) . This
last compound was presumed to be formed by oxidation and to be the precursor of
the diacetic acid.

-328-

COpMe

C02H
C02H

CH3
XXXV

Rowland29 suggests that products such as these arise from air oxidation of in-
itially formed products, hut it is unlikely that this is occurring in this case.

Smissman concluded that, in the case of the compounds which were studied,
the cyclopropanone mechanism is operative, and the conformation of the halogen
rather than the polarity of the solvent determined whether rearrangement will
occur or not. He explains the solvent effects observed hy House and Gilmore21
by saying that the a,P-epoxyketones cannot reach a true equatorial position and
so this probably effects the route of the reaction. He harks back to Wendler's30
suggestion that halogen migration might occur before rearrangement in order
to rationalize the solvent dependence in the case of l-acetyl-l-chloro-2-methyl-
cyclohexane.20 Kende1 claims that halomethyl compounds such as XXXVTII react
slower than those having the halogen on a carbon in the ring as in XXXVII.

CHpX

XXXVII

xxxvrii

If this is true, then the chlorine cannot be migrating before reaction. On the
other hand House20 says that XXXVIII qualitatively reacts faster thai XXXVII,
in which case no conclusion may be drawn.
VI. Rearrangement of 2-Bromocyclobutanone

The Favorskii rearrangement of 2-bror.iocyclobutanone (XXXIX) is known
to occur under conditions, such as carbonate as the base or simply in water
solution,31 which are much milder than the usual Favorskii conditions. This
fact, coupled with the fact that cyclopentanones generally do not undergo Favorskii
rearrangement, seems to suggest that the mechanism which is operative in this
case is not the general one for the Favorskii rearrangement.

TBr

CO-

H20

(>-co£

Two possible mechanisms were considered for this rearrangement:

A. Semibenzilic

B. Cyclopropanone

The predicted labeling results for these two mechanisms when run in D20 are
shown in Fig. 1.

C02D

COoD

The labeling studies32 indicate that the semibenzilic mechanism is operative
since no deuterium is incorporated into the ring. It has previously been shown1'6
that this mechanism is not the general one for the Favorskii rearrangement
since this mechanism is unsymmetrical. However, the rearrangement of XXXIX
does indicate that, although the cyclopropanone mechanism is generally the pre-
dominating one, other mechanisms are operative in cases where the formation
of a cyclopropanone is unfavorable or where the other mechanisms are particu-
larly favored o

Conclusion

The Favorskii rearrangement appears usually to go through a cyclopropanone
intermediate, although other intermediates cannot be strictly ruled out. Other
points concerning this mechanism such as solvent effects and the relationship
of the conformation of the halogen to the course of the reaction, appear to
depend to a great extent on the system under consideration.

Bibliography

1. A. S. Kende, Organic Reactions, 11, 261(1960).

2. J. G. Shell, M. I. T. Seminars' in Organic Chemistry, Dec, 14, I960.

3. R. Jacquier, Bull. soc. chim, France, 17, D35 (1950).

4. D. W. Lamson, Univ. of Illinois Seminars in Organic Chemistry, May 21,
196U.

5» B. Tchoubar, Bull. soc. chim. France, 1363 (1955).

6. R. B. Loftfield, J. Am. Chem. Soc, 73, 4707 (1951).

7. J. Aston and J. Newkirk, J. Am. Chem. Soc, 73, 39OO (1951).

8. J. G. Burr and M. J. S. Dewar, J. Chem. Soc, 1201 (1954).

9. Ro Breslow, J. Posner, and A. Krebs, J. Am. Chem. Soc, 85, 234 (1963).

10. N. J. Turro and W. B. Hammond, J. Am. Chem. Soc, 87, 3258 (1965).

11. W. B. Hammond and N. J. Turro, J. Am. Chem. Soc, 88, 2880 (1966).

12. H. 0. House and G. A. Frank, J. Org. Chem., 30, 29E8 ( 1965) .

13. A. W. Fort, J, Am. Chem, Soc, 84, 2620 (1962)".

14. A. W. Fort, J. Am. Chem. Soc, "cW, 2625 (I962).

15. A, V. Fort, J, Am. Chem. Soc, BE, 1+979 (1962).

16. H. G. Richey, Jr., J. M. Richey, and D. C. Clagett, J. Am. Chem. Soc,

86, 3906 ( 1964) .

17. Ro C. Cookson, M. J. Nye, and G. Subrahmanyan, Proc Chem. Soc, 144 (1964)

18. N. J. Turro, W, B. Hammond, and P. A. Leermakers, J. Am. Chem. Soc,

87, 277^ (1965).

19. G. Stork and I. J. Borowitz, J. Am. Chem. Soc 82, 4307 (1962).
20.

H, 0. House and W. F. Gilmore, J. Am. Chem. Soc,

., 3980 (1961)

-330-

21. H. 0. House and W. F. Gilmore, J. Am. Chem. Soc, 83, 3972 (1961).

22. B. Tchoubar, Bull. soc. chim. France, 1533 (1963).

23. A. Gaudemer, J. Parcello, A. Skrobek, and B. Tchoubar, Bull. soc. chim.
France, 2^05 (1963) .

2k. H. 0. House and H. W. Thompson, J. Org. Chem. , 2_8, l6k (1963).
23. B. A. P. Olson, Dissertation Abstracts, 2jj, MkL3 (1965).

26. M. Mayer, Compt. Rend., 233, ^88 (1961).

27. M. Charpentier-Morize, M. Mayer, and B. Tchoubar, Bull. soc. chim. France,
529 (1963).

28. E. E. Smissman, J, L. Lemke, and 0. Kristiansen, J. Am. Chem. Soc, 88,
335 (1966),

29. A. J. Rowland, J. Org. Chem., 27, 1135 (1962).

30. N. L. Wendler, R. P. Graber, and G. G. Hazen, Tetrahedron, £, ikk (1958).

31. Jo-M Conia and J. -M. Ripoll, Bull. soc. chim. France, 755 (19&3) •

32. J. -Mo Conia and J. Salaun, Tetr. Letters, 1175 (1963).

PROSTAGLANDIN SYNTHESES

Reported by Edward Bertram

April 2k, I967

Prostaglandins are C20 carboxylic acids with hormone -like qualities. They
show varying pharmacological, effects1 such as smooth muscle activation, lowering or
raising of blood pressure, mobilization of lipids and affecting the reproductive
tract o They are found in almost all parts of the body, but mainly in the vesicular
glands of man and animals . This seminar will deal with their recent laboratory
syntheses and the proposed mechanisms for biosynthesis .

The initial structure determination was carried out by Samuelsson and Berg-
strom.2'3 By the utilization of mass spectra, infrared, ultraviolet, optical
rotation and chemical degradation, they proposed the mirror image of I as the
absolute stereochemistry of prostaglandin Ex (PGEx) „ This structure was basically
confirmed by Abrahamsson4 by an X ray analysis of the tris-p-bromobenzoate methyl
ester of prostaglandin F^a (PGFxp) Ho

0 H, rm 0

( -) -ll-a-15(s) -dihydroxy-9-°xo-13

trans prostenoic acid

Since then Samuelsson, Van Dorp and co-workers5 have re -evaluated the evidence
and the determination of the optical rotations of some by-products and degradation
products, and compared them to other alcohols of known absolute stereochemistry and
decided that the absolute stereochemistry for PGEx is as shown in I0

Other prostaglandins are PGE2 which contains an additional cjLs double bond at

C5 and PGE3 which contains two additional cis double bonds at C5 and Cx?. The PGFq,

and the FGFg ser5.es are the prostaglandins obtained by reduction of the C9 oxo group,

3 has the hydroxy! trans to the other ring hydroxy! and a contains the two hydroxyl

groups ciso Other series are the PGE-217 III and PGE-278 IV. These are named after

their ultraviolet spectra and are obtained as natural products or by dehydration of

the Cn hydroxy! group. The C19 hydroxyl. compounds of some of these a ,6 -unsaturated

derivatives have also been isolated.6
OH

The prostaglandins are derived from essential fatty acids,' PGEx from all cis

8, 11, lk eicosatrienoic acid V

11, lk eicosatetraenoic

, PGE2 from all c_is
8, 11, lkf 17 eiccsapentaenoic acid VII
determined by ^ or 14C labeling studies.8

acid VI and PGE3 from all cis

This was

VII

770
-JJ1-'

It was also shown that the hydrogens at 0Qf Cu and C12 were not lost in the
formation of PGEX using vesicular sheep enzyme. Kleriberg and Samuelsson9 proved
this by specifically labeling the three positions with ^ using a 14C labeled 8,
11, Ik eicosatrienoic acido

Samuelsson10'5'11 and later VanDorp12 showed conclusively that the oxygen on the
ring and the C15 hydroxy! group came from molecular oxygen. They also proved that
both the ring oxygens came from the same molecule of oxygen, Samuelsson incubated
all cis 8, 11, lk eicosatrienoic acid in an oxygen atmosphere of 180~180 5670,
160-T50 1% and 160-160 hjS and then reduced the VGEX to the PGFxp compound with
NaBH^ in EtOH, thus preventing the exchange of the keto oxygen on workup. He then
methylated the hydroxyl groups and cleaved the double bond using KMh04 and periodate
to obtain the diacid which he then ethylated. The mass spectrum, of this compound
was run and the P + 2 and the P + k peaks were compared for ions which still con-
tained both oxygens. It was determined that if both oxygens in the ring resulted
from the same oxygen molecule then the ratio P + 2/P + k would be 0.02, but if two
oxygen molecules were involved then the ratio would be I.5« He obtained a ratio of
0.06 and so concluded that both ring oxygens must be from the same molecule. Van-
Dorp12 used an oxygen mixture of 180-18Q kj$>, 160-180 6i and 160-160 h 7$ and a much
cleaner enzyme preparation (hardly any endogenous ¥GEX present) and treated the
sample in three different ways, (A) kept all original oxygens (B) exchanged the C9
oxo group for all 160 and (C) removed the oxygen at C1S by oxidation. He obtained
the following table of values.

jo of molecules with18©

18/

(A)
(B)

(C)

k-6%

one 180
25. 5#

two

three 180

25.5%

hQi

These experiments show that all of the oxygen functions are obtained from
molecular oxygen and that both ring oxygens are from the same molecule. Van Dorp
also measured the up -take of oxygen and estimated that two moles of oxygen were
consumed per mole of PGEx formed. He also determined that Glutathione was a co-
factor (almost exclusively) and that there was a need for an anti -oxidant (propyl
gallate or £-hydroquinone) to obtain maximum yields of BGrElo

The Van Dorp group also identified several by-products obtained by varying the
reagents of incubation. In a normal reaction using Glutathione and an anti-oxidant,
they obtained PGEx I and compound VIII ( ll-hydroxy-8°trans -10°trans -l^-cis -
eicosatrienoic acid). If the antioxidant was left out, I, VIII, IX (il-a-hydroxy-
9,15-clioxo~13prosteneic acid) and X (15=hydroperoxy-ll^»hydroxy-9-oxo-13-prostenoic
acid) resulted. If Glutathione was not added, there was a significant drop in PGEX
formation and compounds XI ( 12~hyaroxy~8-trans -IQ-trans-heptadecadienoic acid), XII
(9-=Q!"15"dihydroxy-ll-oxo-I3"prostenoic acid) and XIII PGF^ were isolated.

From the labeling experiments Samuelsson proposed two possible mechanisms, one
(Scheme l) was the direct molecular oxygen addition across the Cu, Ce position and
a conrotatory Cs, C1S bond formation. This would be done by a dioxygenase, a sub-
sequent monooxygenase for the addition of a second oxygen at C15 followed by
isomerization of the double bond to the C13 trans configuration and, then opening of
the peroxide.

dicxygena.se

COgH

monooxygenase

^^GO^H

-7 "2-7

A second mechanism (Scheme 2) he proposed was the formation of a hydroperoxide
at Cu> isomerization of the double bond from A11 to A12, then attack on the Cg by
the hydroperoxide to form a cyclic peroxide followed by a concerted ring closure to

form the C&$ 0^2 bond with isomerization of the A bond to trans A , then the
addition of a hydroxyl group at C15 and opening of the peroxide to give PGE1(, With
only the use of the labeling experiment 9 Samuelsson could not determine if initial
oxygenation was at Cll9 Cg, or C15o

^^^CO^L

-i I

Scheme 2

Van Dorp's proposed mechanism (Scheme 3) is similar to Scheme 2 except that he
uses a peroxide radical and supports the C^x attack with the fact that compound VIII
is formed, but no compounds have been isolated with only the C15 hydroxyl group or
Cg hydroxyl group,, He supports the radical mechanism because of an esr band which
forms when the enzyme and substrate are added and slowly disappears on incubation 0
He also shows that compound XI is formed from an intermediate in the PGEX synthesis
since he has isolated a mole to mole ratio of XI and malonic aldehyde, the latter
being trapped by a thiobarbituric reaction. He suggests that since no intermediates
are obtained when the normal reaction is run, the process is of a concerted nature
and all of it takes place while the substrate is attached to the enzyme. Also, none
of the products isolated could be transformed into PGEx on further incubation.

OH
VIII

R - (CHs)6eoc/°j

R'*e CsHxx

Scheme 3

The esr data favor the peroxide radical, but they do not rule out the hydro-
peroxide since the esr band may be due to other sources. The initial attack at Cn
seems to be favored by the isolation of VIII, in slight contrast uo uue fact uhau tne

-z.-z.lt

w>6 double bond compounds (double bond 6 carbons from the CH3 end of the chain) are

the most specific for the enzyme.1-3 A free carboxyl group is also required and a
C20 fatty acid is favored over C18> C19j) C21 or C22 fatty acids.15*16 A systematic
substrate specificity search has not yet been done. There has been extensive work
done on inhibitors and co-f actors by Van Dorp and co-workers „12

A simple synthesis of the prostaglandin structure was used in the structural
determination by Samuelsson and Stallberg.17 They succeeded in synthesizing XIV by
two methods using Grignard and condensation type reactions. No report on the
biological activity was given.

XIV

Since then there have been several syntheses reported. Bagli18 and co-workers
have succeeded in making XV, a derivative of PGEl0 They started with the potassium
salt of ethyl 2»cyclopentanone earboxylate and alkylated it using 7-bromoethyl-
heptanoate. It was then monobrominated and after refluxing the monobromo derivative
in 20$ HsS04 XVI resulted. On reaction of XVI with acetone cyanhydrin in NaC03 and
H2Q/MeOH, a nitrile was obtained which yielded the trans carhoxylic acid XVII on
hydrolysis, A monoesterification to XVIII was then accomplished by refluxing with
a 0.5 mole excess of p_-toluenesulfonie acid in MeOH for 55 minutes. This ester was
transformed to the acid chloride XIX and the acid chloride was reacted with acetylene

in the presence of A1C13 in CCI4 to form a vinyl chloride XX which yielded the methyl
ketal XXI on basic hydrolysis in MeOH.

H *QH

~0oRi

XVI

XVIII
XIX

jLa,

XXI

,c-c:

„c-c>

I !

H H

*C1

,0Me

v0Me

The reaction of XXI with NaBRg. yielded a mixture from which XXII was obtained
after refluxing in 2N H2S04(> This compound was then reacted with n-pentyl magnesium
bromide to yield compound XV.

Since then the Bagli group have modified their synthesis.19 They reacted 1-
heptyne with XIX $ the acid chloride, to form, the vinyl chloride XXIII which was
transformed by NaOH in MeOH to a vinyl methoxy compound.. A NaBEg. reduction followed
by hydrolysis yielded a ketone XXIV which on further NaBHi reduction resulted in
the formation of XV.

The stereochemistry of the product was shown by the reduction of XVIII with
NaBEj, to yield a mixture, 85$ of which could be transformed into a lactone XXV. This
same lactone could be formed from the product of oxidative cleavage of XV thus
placing the hydroxyl and carboxyl group cis to one another.

The trans stereochemistry of the two chains was suggested by analogy to the reaction
in which only the trans compound is obtained when cis 2,3 dialkyl cyclopentanone is
re fluxed in base , as was XVII when it was formed from the nitrile „

AJvJLJL

XXIII

Hs i£H

CO^Me

& A0.

COpMe

XXIV

XXV

The compound that was synthesized (XV) has not yet been identified as a natural
product, but it still possesses a vaso depressor effect s which is amazing in view of
the lack of oxygen at the C1± poistion. A method for adding this Cn oxygen function
has been reported by Bagli,20 but no details axe available.

The first total synthesis of a naturally occurring prostaglandin has been
accomplished by Beal and co-workers22 who succeeded in synthesizing the dihydro
TGE± XXXII which is a metabolite present after the incubation of TGE1 in pig lung

enzyme preparation

Beginning with a formylation of 3=ethoxy-2-cyclopentenone ,

they obtained almost a quantitative yield of the sodium salt XXVI. This compound
then underwent an in-situ Wittig reaction to give compound XXVII which on catalytic
reduction and a second formylation gave the sodium salt XXVIII.

EtO

COsEt

CO^Et

XXVI

XXVII

XXVIII

This sodium salt was then put through a second modified Wittig reaction using
n-hexanoylmethylenetriphenylphosphonium chloride and gave XXIX which on catalytic
reduction resulted in 85$ XXX. Acid catalyzed solvolysis of XXX in benzyl alcohol
resulted in the corresponding benzyl enol ether. The ketone was reduced to the 15
hydroxy compound using tri-t-butyl aluminum hydride and this compound was then con-
verted to the diketone XXXI by hydrogenolysis. XXXI was rigorously reduced using
30fo rhodium on carbon to give a mixture, 11$ of which corresponded by tic to XXXII<

OEt 0

OsEt

JOsEt

XXIX

AVvA

.^6-

^-^—

HCL ,H

OsEt

CO^Et

AJLa,J=

xxxii

Compound XXXII was compared with the natural dihydro PGEX using ultraviolet,
infrared, mass spectra, nmr and tic. It was also shown by a radio isotope dilution
experiment using the NaB3H4 reduced product and crystalline optically active natural
(biosynthetic) dihydro PGFxp methyl ester to contain at least 22$ of a compound with
all asymmetric centers identical to the natural compound. If the reaction had been
random, it would have contained only 6$ of the desired product.

A third method of synthesis by Just and Simonovitch23*24 has been reported in
Chemistry in Canada. They oxygenated the hydroboration product from cyclopentadiene
and protected the resulting cyciopenta-3-en°l as the tetrahydropyranol ether XXXIII.
A reaction of XXXIII with diazoacetic ester over copper powder resulted in a mixture
°£ exo-endo, syn-anti compounds which on refluxing in methanolic sodium methoxide
resulted in an exo, syn-anti mixture. The exo compound was reduced with MAIH4 and
the alcohol obtained was oxidized to the aldehyde XXXIV. The aldehyde underwent a
Wittig reaction with hexyltriphenyl phosphonium bromide to yield four isomers which
could be separated by tic. Hydrolysis in refluxing 0.5$ oxalic acid in methanol
resulted in the alcohol which was oxidized using Jones reagent to the ketone XXXV
as a mixture of cis -trans isomers. These isomers were separated by tic, the trans
having a 957 cm"x band in the infrared.

XXXIII

XXXIV

XXXV

The alkylation of XXXV was accomplished using seven equivalents of methyl -7-
iodoheptanoate in dimethoxye thane and potassium t-but oxide as the base. The enamine
alkylation was also attempted but gave a lower yield. The alkylation was made
difficult by the problems of separation of XXXV from the product XXXVI, and the
instability of compound XXXV. An effective separation was accomplished by reduction
of the ketone with NaBH* to yield XXXVII.

CO-Me

C02Me

XXXVI

XXXVII

The acid obtained by NaOH hydrolysis in aqueous methanol of XXXVII was
oxidatively solvolyzed using one equivalent of H^Og in a NagCOs buffered solution
of formic acid, followed by shaking in 10$ aqueous Na^Qa and resulted in dl PGF^
XXXVIII. PGEX was also obtained by a similar oxidative solvolysis of XXXVI. The
PGFjq, was identical bv mass spectra and tic to natural PGF^.

Several biological tests were run and the dl PGFxq, was found to have one -half the
smooth muscle activity of natural PGF^ while the activity of the synthesized PGEX
compound was only of the order of 10-25$ of natural PGElo

7.
8.

9.
10,

11.

13.
14.

15.
16.

17.
18.

19.
20.

21.
22.
23.

24.

BIBLIOGRAPHY

W. Horton, Experientia, 21, 113 (19^5) <•

Samuelsson, Angew. Chem. Internat. Edit., 4, klO (I965).

Bergstrom and B. Samuels son, An. Rev. Biochem. , 34, 101 (1965)
S. Abrahams son, Acta Crystallogr . , l6, k0$) (1963).
D. H. Nugteren, D. A. Van Dorp and S. Bergstrom, M

E.
B.
S.

Nature, 212, 38 (1966).

Hamberg and B. Samuels son,
257 (1966).

M. Hamberg and B. Samuels son, J. Biol. Chem. , 24l,

H. Vonkeman, Chem. Weekblad., 62, 361 (I966).

(a) S. Bergstrom, H. Danielsson and B. Samuelsson, Biochim. Biophysic. Act.,

90, 207 (1964) 1 (b) S. Bergstrom, H. Danielsson, D. Klenberg and B. Samuelsson,

J. Biol. Chem'., 239, PC k006 (1964); (c) D. A. Van Dorp, R. K. Beerthius,

D. H. Nugteren and H. Vonkeman, Nature, 20j5, 839 (1964); (d) D. A. Van Dorp,

R. K. Beerthius, D„ H. Nugteren and H. Vonkeman, Biochim. Biophysic. Acta.,

90, 204 (1964),

D. Klenberg and B. Samuelsson, Acta. Chem. Scand., 19 (I965).

R. Ryhage and B. Samuelsson, Biochim. Biophysic. Res. Coram., 192, 79 (19^5).

B. Samuelsson, J. Am. Chem. Soc, 87, 3 Oil (1965).

D. H. Nugteren, R. K. Beerthius and D. A. Van Dorp

405 (I966).

Rec. Trav. Chim. , 82,

Struizk, R. K. Beerthius, H. J. J. Pabon and D„ A. Van Dorp, Rec. Trav.

Dorp, Rec. Trav. Chim., 85,
Samuelsson, J. Biol. Chem.,

Vonkeman, Nature, 203,

C. B

Chim., 85, 1233 (1966).
H. J. J. Pabon, L. Van der Wolf and D. A
1251 (1966).

S„ Bergstrom, H. Danielsson, D. Klenberg and B
259, PC 4006 (1964).

D. A. Van Dorp, R„ K. Beerthius, D. H. Nugteren and H
839 (1964).
B. Samuelsson and G„ Stallberg, Acta. Chem. Scand., 17, 810 (1963).

J. F. Bagli, T. Bogri and R„ Deghenghi and K. Wiesner, Tetrahedron Letters,

465 ( 1966) .

J. F. Bagli and T. Bogri, Tetrahedron Letters, 5 (1967).

J. F. Bagli, 10th Annual Medicinal Chemistry Symposium,

June I966. (C. and E. News, 44, No. 27, 32 (196 "

E. Anggard and B„ Samuelsson, J. Biol Chem. , 239
P. F. Beal, J. C. Babcock and F. H. Lincoln, J
G. Just and C. Simonovitch. Chem. in Can. , 19.

4097

Bloomington, Indiana,
1964) .

G. Just, private communication, March 20, 1967<

Am. Chem. Soc.,
Jo. 1, 41 (I967).

88, 3131 (1966)

-338-

POSSIBLE VINYL CATION INTERMEDIATES

Reported by David A. Simpson

.May 1, I.967

Within the past few years vinyl cations have been postulated as inter-
mediates in a variety of reactions. The textbook rule regarding the reputed
instability of these ions is based primarily on the unreactivity of vinyl halides
toward alcoholic, silver nitrate. 1>2 Thus vinyl chloride on long heating with
solutions of silver nitrate in ethanol gives no silver chloride. However,
recent studies have questioned the general instability of vinyl carbonium ions.
It is the purpose of this seminar to review those reactions which have been
claimed to proceed through such intermediates.

The first postulation of a vinyl cation intermediate was made in 1951
by Newman and associates as a result of studying the alkaline decomposition
of 3--nitroso-2-oxazolidones (l)° The following mechanism was proposed;3

aldehyde s_9 ketones, acetylenes"
(vinyl ethers) *

Ri P-2

C#
I

(h)

R3— R4— H

HoO

)COPH

'NH-NO

R2"
R3"

(2)

OC02H

N=N-OH

(3)

R.3,R4/H

aldehydes, allylic alcohols,1
glycols, cyclic carbonates
(acyclic alky! carbonates.,
acyclic dialkyl carbonate
ethers) *

R2
R3"

OC02H

(5)

* reaction carried out in anhydrous alcohol, using
the corresponding sodium alkoxide

The first step was thought to involve ring opening of (l) to a nitrosamine (2)
followed by tautomerism to an hydroxyazo intermediate ( 3) . Two paths of reaction
are now open to (3) : path. A involves the initial formation of an unsaturated
diazonium hydroxide intermediate (by the base catalyzed elimination of carbonic
acid) followed by loss of nitrogen to yield the vinyl carbonium ion (k) , and
then formation of products; or (3) could lose nitrogen first yielding the sat-
urated carbonium ion (s) which then yields products. All products obtained from
the 4,4-disubstituted oxazolidones could be explained by path B, but it was
necessary to invoke path A to account for the products resulting from the

-339-

5,5-disubstituted derivatives.5 Although path A seems to explain all the exper-
imental observations adequately , a mechanism involving an intermediate methylene (6)
cannot be rigorously excluded*6

RK /* OH" R< •• -lb \

C=<3 — - — > ^C=C=N=<N. — iH-4 ^C=Cj ^ PRODUCTS

R2 NW=U~0H R/ R2

(6)

At any rate , there does not seem to he enough evidence to support any of the pos-
sible mechanisms very strongly relative to the others.

Acetylene derivatives have served as the source for a number of proposed
vinyl cation intermediates. The earliest example was reported by Jacobs and
Searles who claimed that the hydration of acetylenic ethers appeared to be sim-
ilar to that of vinyl ethers. That is, the hydration involved a rate-determining
formation of a carbonium ion.7 They measured the rates of hydration of several
acetylenic ethers in alcohol-water solutions by a dilatometric method. The
rates were found to be first order with respect to acetylenic ether and to
hydronium ion. A mechanism consistent with the observed kinetics was formulated
as follows.8

I. R-Q~CrCH + H30+ + [R-0~OCH2]+ + R20

II. [R-0-C=CH2]+ + H20 —^ [R-Q-G<SH2j+

6h2

iii. [r-c~c=€h2]+ + h20 > r-0-c=ch2 + h30+

6h2 6h

iv. r-0- — > r-o-c-ch3

OH 8

Tne authors believed that the first step was rate-determining and that the reaction
was specific acid-catalyzed.

Sixteen years later Drenth and co-workers undertook a study of the hydra-
tion of acetylenic ethers and thioethers. Their rate studies largely confirmed
the findings of Jacobs and Searles, however s the reactions were found to be general
acid-catalyzed. 9j>lc The observation of general acid catalysis indicated that
the first step of Jacobs and Searles9 mechanism was rate-determining. Most of
the work of Drenth was aimed at confirming the rate-determining first step and
educing the timing of the addition of the water molecule. Only the results of
their investigations of the acetylenic thioethers will be summarized in this sem-
inar. However, many of the same experiments and results were also shown to ap-
ply in studies of the acetylenic ethers.

.Evidence presented for a rate-determining proton transfer in the general
acid-catalyzed hydration is condensed as follows ; (l) the reaction is faster in
H20 than in D2C>for ethylthioethyne the ratio of k-. .JK, 0 amounts to O.475

(2) substitution of a. tertiary butyl group for the ethyl group in ethylthioethyne
results in a rate enhancement.10 If step II is rate-determining,, it would be
expected that the t-butyl compound would hydrate slower- than the ethyl derivative,
since the addition of a water molecule in 11 would be more sterically hindered
in the case of the t-butyl derivative. The fact that the latter compound is hyd-
rated even faster than the ethyl derivative indicates that the proton transfer
must be the kineticaily important step (application of the Taft equation to a
series of thioethers further shows that steric effects are not essential).11
And ( 3) , infrared analysis of recovered ethylthioethyne from a reaction in heavy
water shows that a pre -equilibrium is not important in this reaction.12

Since the addition of the water molecule could occur simultaneously with the
proton transfer or in a subsequent step, Drenth !s group set out to determine the

-3^0-

timing of addition. Abbreviating their findings: the activity postulate of
Grunwald for reactions in alcohol-water mixtures; the relation of Zucker and
Hammetty and entropy considerations all seem to indicate that a water molecule
is not covalently involved in the rate -determining step of the reaction. There-
fore, the authors concluded that addition of water takes place in a step subse-
quent to protonation. 12 The evidence presented render both a it -complex mech-
anism and a cyclic transition state involving the ether, hydronium ion, and
water improbable. Thus, the rate -determining formation of a vinyl cation inter-
mediate is strongly indicated.

Peterson has reported evidence for a rate-determining formation of a vinyl
cation in the acid-catalyzed hydration of phenylpropiolic acids (7) and phenyl-
acetylenes. Pseudc-first-order rate constants for the acids (7a-d) were measured

(7)

:5C-C02H

a, X=0CH3 c, X=E
b, X=CH3 d, X=C1

spectrophotometrically and the rate data were found to correlate with cr, p = -4.79+
0.02* The authors claim that the large negative value of p indicates a high
degree of positive charge on the benzylic carbon in the transition state and there-
fore that the C-OH2 bond formation occurs after proton transfer. In addition
a solvent kinetic isotope effect implies a rate -determining proton transfer.13
Further studies of ortho, meta, and para substituted phenylacetylenes by Bott,
Eaborn, and Walton again suggest a rate-determining proton transfer. This is
followed by a rapid reaction of the formed vinyl carbonium ion with solvent to
form the ketone.14

Recently, vinyl cations have been implicated in the reactions of trifluoro-
acetic acid with alkynes. In a preliminary report Peterson and Duddey regarded
the vinylic trifluoroacetates formed as arising from intermediate substituted
vinyl cations.15 This communication was followed by a more detailed study of the
reactions of unsubstituted hexynes and 5-substituted-l-pentynes. Evidence for
the cationic nature of the transition states was obtained by comparing the reactions
of the substituted alkynes to the previously studied reactions of identically
substituted alkenes<>x6 A Hammett-Taft plot (Taft's a values for trifluoroacetic

acid derived from fluorine nmr frequencies) 17 showed the rate constants for the
alkynes were only slightly smaller than those of the alkenes, and demonstrated
nearly the same pattern of substitueni effects for both, indicating similar
cationic like transition states. The study of the reaction of 3-hexyne is es-
pecially interesting.

OCOCF

CF3C02H

OCOCF3

DIMERS
POLYMERS

The vinyl trifluoracetates (9) and (10) were formed in nearly identical amounts

and were shown to be stable under the reaction conditions. Thus, they were reported

to have arisen from nonstereospecific addition., The yields of (9), (10) and
(12) were found to vary with the hexyne concentration as shown in Table I,18

Tablg_I»,

Molarity of (9) and (10) (12)

3-hexyne ^ mole <p mole $

0.107 98 2

lo03 2k 21

2o00 18 30

Besides a sweeping generalization that vinyl cations are readily accessible by
addition of protons to alkynes, the authors had little to say concerning the
above findings. However , it is tempting to interpret the observation of non-
stereospecific addition as rendering mechanisms involving a jt -complex or con-
cerned, type transition state improbable for this system. The product depen-
dence on the concentration of 5"hexyne could be rationalized by attack of the
vinyl cation derived from 3-hexyne on a second molecule of substrate.

Peterson and Kamat have reported the formation of a transition state re-
sembling a vinyl cation, in the related trifluoroacetolysis of 6~heptyn-2-yl
p-toluene sulfonate , -1 9

Fahey and Lee have recently studied the reaction 1-phenylpropyne with hyd-
rogen chloride in acetic acid. The observed kinetics (reaction found to follow
a third-order rate law, first-order in acetylene and secord -order in hydrogen
chloride) and the stereochemistry of the products lead the authors to propose
a mechanism involving intimate and solvent separated vinyl carbonium-hydrogen
dichloride ion pairs as intermediates.20

Another reaction involving protonation of an unsaturated system, which
may lead to vinyl cations, is the addition of compounds of the form HX to allenes.
The chemistry of allene has recently been reviewed.21 The addition of HX to al-
lene has been observed to follow Markovnikov ' s rule. For a long time it was
thought that 2-substituted propenes and/or 2 ,2-disubstituted propanes were the
only products of such additions.22 However,, Griesbaum and associates have recently

HX
HgC^C^CHg — — — ^ H3C-CX—CH2 + H3C-CX2-CH3

(13) (lh)

found such reactions to be more complex. It was reported that in the electro-
philic addition of hydrogen bromide to allene considerable amounts of 1,3-dibromo-
l,3~dimethylcyclobutane (two isomers.) were formed, as well as the conventional
products (13.) arid (l4)»23 Structural proof for the cyclic compounds consisted
of mass spectral, nuclear magnetic resonance „ and infrared data. Additional
evidence was obtained by reducing the dibromo compound with tri-butyltin hydride
to yield a mixture of the isomeric 1,3-dimethylcyclobutanes.

Formation of the cyclic products by a thermal reaction was considered im-
probable since thermal dimerization of allene leads to a 1,2-1,2 adduct, while
hydrogen bromide addition causes a formal 1,2-2,1 dimerization. Reaction of 2-
bromopropene (13, X=Br) with hydrogen bromide under the conditions of the allene
addition led exclusively to 2,2-dibromopropane. Ultraviolet irradiation of
2-bromopropene gave essentially starting material and showed no traces of the
cyclic products. Thus, the authors concluded that a simple dimerization of
2-bromopropene was not Involved. Furthermore, since the free radical addition
of hydrogen bromide to allene yields no cyclic products,24 the cyclodimerization
was thought to be an ionic reaction. Since the reaction of methylacetylene with
hydrogen bromide also led to cyclo-dimerization,25 and because the rearrangement
of allene to methylacetylene in the presence of hydrogen bromide could not be
ruled out, the following path was proposed.26

-3^2-

CH2=C=CH2

CH2 -=0 =CH2

CH3-C— CH2
(15)

CH3-CX=CH2
X

1. X

CH-

2. HX

CH-

CH3-C=CH

CH3C^CH

1. X

2. HX

While the intermediate steps in the above scheme remain unclear, the authors
felt that the vinylic cation (15) was involved in the cyclizations. Some data
for the addition of hydrogen bromide to allene and methylacetylene is reported
in Table II.

Table II.26

Relative amounts of components in adduct mixture, wt. 70
Substrate BrCH=CHCH3 CH3CBr=CH2 CH3CBr2CH3

Run#1 ch£Sh2

Rnn Jio* CH3C=CH
Run #2* CH2=C=CH2

23
33

.56
60

25
27

9
10

3
.3

* run 7r2 contained hydroquinone as inhibitor

Evidence that the l=bromopropene is formed by a concurrent free radical addition
is shown by the effect of the inhibitor on its yield. If one ignores the free
radical addition product, the methylacetylene and allene additions show the same
selectivity for the remaining products. It should be mentioned that the deter-
mination of product yields In Table II is not completely clear. The authors
claim that the components comprised 75-98$ of the total adduct mixtures (depend-
ing on the purity of the starting materials) . However, it was not indicated how
this estimate was made (no mention of internal standard use). Accordingly these
values may not be too reliable «, Furthermore, a mechanism Involving concerted
addition cannot be rigidly excluded.

A possible route to a vinyl cation could be deamination of vinyl amines,,
Curtin's group has examined several such deaminations . The reaction of 2,2-
diphenylvinylamine (16) with nitrosyi chloride is especially relevant to this

UlbL'USSl'J-l.

The formation of (17); (18) and (19) was explained in terms of a

-3^3-

<t>

.CI

$ m2
c=c

*(l6)*

4>C=C<t> + v

(17) H (18) CI
13#* #

h'(i9)no

CHgCxg
2.5°

(*)2C

30^

■H + NO

4>-C-<t>

15^

* amounts of products obtained are reported as mole per cent based on amine
employed

carbonium ion rearrangement from the ion (20) , formed by loss of nitrogen from
the diazonium ion, to the ion (21). The ion (2l) could lose a proton to form the
acetylene (17) or suffer attack by chloride ion to give the cis- (18) and trans-
(19) chlorides.

C=C

(20)

4>C=C

0x ^ra c:

/c=cv

V H

(21)

(22)

The nearly random formation of these three products is of particular interest.
The authors interpret the results as being consistent with three differently
oriented vinyl carbonium-chloride ion pair intermediates. The possibility of
(20) fragmenting into a phenyl cation and phenyl acetylene followed by readdition
was deemed improbable since gas phase chromatography of the product mixture showed
no traces of phenylacetylene or chlorobenzene. The product of replacement without
rearrangement , 2 , 2-diphenylvinylchloride^was shown to be present to a maximum

However, an intermediate bridged ion could not be rigorously excluded

28"

Of 2$

by the authors.

All attempts to intercept the diazonium intermediate (22) with the sodium
salt of P-naphthol failed. It was concluded that the diphenylvinyl system did
not provide the necessary stabilization of (22) and de stabilization of (20)
and (21) which is required for trapping. The instability of the diazonium ion
was attributed to possible participation of the aromatic ring in the transition
state for loss of nitrogen and/or freedom of the vinylcarbonium ion to assume
sp hybridization.29 A study of 3~atf1ino-2-phenylindenone (23) was initiated in
hopes that it would be free of these difficulties. Again attempts to intercept
the diazonium intermediate failed. However, during the reaction of (23) with
nitrosyl chloride at -10° an infrared absorption appeared at 2090 cm"1. Observation
of this peak over a period of time showed that it disappeared according to a first-
order rate law. The authors stated that no definite assignment could be made to
this absorption, but it might be ascribed to a diazonium or diazoaikane intermediate30'31

Studies by Grob and Cseb on the solvolysis of a-bromostyrenes {2k) have
questioned the "well-known" general unreactivity of vinyl halides. The first-order

(24)

R= -H
-NH2
-OCH3

rate constants for 0.01 M solutions of these compounds in
triethylamine were determined and appear in Table III.32

-NHCOCH3
-W02

& ethanol with 0.0.1 M

~NH2

0.00

-OCH3

100.10

-WHCOCH3

115.20

-H

170. Q0

-344-

Table III.
(24) temperature k^ec"*1) k ,

-Ms 0.00 9.57(9.119)* X 10"5 5.5 X 108

3.60(3.60) x 10"5 8.5 x 103

3.80(3.92) X 10~5 2.2 X 103
6.00(6.80) X 10"6 1

*values in parenthesis are in the presence
of 0.05 M triethylamine

The following additional observations were made concerning these reactions: (l)
preparative solvolyses in the presence of triethylamine or calcium carbonate gave
only the corresponding acetophenones from the substituted a-bromostyrene s (24s
r= -NH2, -OCH3 and -NHCOCH3) , a-bromostyrene (24; R= -H) yielded in addition
22/® phenylacetylene; (2) the reaction of the p-nitro derivative was exclusively
second-order and gave only p-nitrophenylacetylene; (3) each of the substituted
phenyiacetylenes was shown to be stable under the reaction conditions; (4) after
a short warming with silver-nitrate in 80$ ethanol the methoxy derivative (24:
R= -OCH3) gave a precipitate of silver bromide , the nitro compound (24: R= -N02)
after several hours at 100° gave no precipitate; (5) and a-bromostyrene (24:
R= -H) in 50$ ethanol reacts ten times faster than in the less ionizing 80$
ethanol solution.32

Based on these observations , the authors propose that the a-bromostyrene
derivatives (24; R= -H,-MI2, -0CH3,-NHC0CH3) react via an S^l - El type mechanism

involving the intermediate formation of a vinyl cation, and that the nitro compound
(24: R= -N02) reacts by a bimolecular elimination (E2) process. However, the
following facts must be considered. First, products were never isolated from the
kinetic runs themselves, and the preparative solvolyses were not run under identical
conditions with the former. Secondly, no data or plots of the kinetic studies
were given, only the resulting rate constants were reported. And thirdly, the
authors claim that the phenylacetylene formed from a-bromostyrene (24: R= -H)
could not arise from a competitive E2 process due to the small deviations observed
in the first-order rate constants on addition of five molarequivalents of tri-
ethylamine. However, the question remai.ns whether a competing E2 process could
have been detected by the workers. Crude calculations indicate that a competitive
E2 process here appears not unlikely. Further indication that the reactions may
be more complex than the authors claim appears in the magnitudes of the relative
rate constants (Table III). The reported accelerations do not parallel the known
electrical properties of the substltuents. In spite of these criticisms the relatively
facile reaction of alcoholic silver-nitrate with the methoxy derivative (24:
R~ -OCH3) indicates that further studies of "stacked" vinyl halide systems may prove
interesting.

In a related study Grob and co-workers have examined the solvolytic decarboxy-
lation of the potassium salts of cis and trans a,£-unsaturated-J3-halo acids.
Salts of the cis series when heated in aqueous solution yielded the corresponding
acetylenes, while those of the trams series gave ketones in addition to acetylene
derivatives. The reaction of the cis salts was explained by a concerted mechanism,
on the other hand a rate-determining ionization to an intramolecularly solvated
vinyl carbonium ion was proposed for the solvolysis of the trans salts.33

Stable cations having a contributing vinyl cation resonance form have reportedly
been observed by nmr. Richey's group has examined several systems, but only the
propynyl and ethynyl-di-p-methoxyphenylcarbinol (25) case will be discussed.
Extraction of the propynyl compound (25: R- -CH3) from carbon tetrachloride into

CH3O-4 >- C-CEC-R R- H or CH3

(85) 0

OCHc,

-3^5-

concentrated sulfuric acid gave solutions whose nrar was assumed to be that of the
corresponding alkynyl cation. Strong evidence for this was obtained by neutralizing
the sulfuric acid solutions which gave approximately yof/o of the starting alcohol.
The absorption of the propynyi methyl group (t 7° ^-0) appeared considerably downfield
from the absorption of the same group in a carbon tetrachloride solution of the
substrate (x 8.13).34 The absorption of the corresponding ethynyl derivative
(25; R= -H) snowed a much larger downfield shift, (from T 7° 35 for the alcohol to
t 4.30 for the ion} The authors suggest that this indicates that the carbon to
which the ethynyl hydrogen is attached is significantly involved in charge dereal-
ization as shown by the resonance structure (26).35

$/ $ /

-c=c-c <e- — -» -c=c=c

(26) ^

Footnotes 36-38 are additional references to reactions in which vinyl car-
bonium ions have been postulated. However, due to their similarity to systems
already discussed and the limited amount of information available they will not
be examined in this abstract.

In conclusion, evidence presented seems to indicate that in certain systems
the intermediacy of vinyl cations best explains the experimental results. Such
findings cast doubt on the inferred general instability of such intermediates.
In the past these intermediates have been largely rejected because of the unreactivity
of vinyl halides toward alcoholic silver-nitrate. However, as several workers have
pointed out,39-'40 the inertness of vinyl halides is based on ethylene derivatives
and may be less applicable to more highly alkylated systems. In any case a definitive
study of the reactivity of various vinyl halides toward alcoholic silver-nitrate
appears to be missing from the literature. As a result the chemistry of vinyl
carbonium ions is somewhat ill defined.

BIBLIOGRAPHY

1. R. L. Shriner, R, C, Fuson, and D. Y. Curt-in, "The Systematic Identification
of Organic Compounds" , John Wiley and Sons, Inc., New York, 1956, P« 1^1 •

2. J. D. Roberts and M. C. Caserio, "Basic Principles of Organic Chemistry",
W. A. Benjamin, Inc., New York, 1964, p. 321.

3. M, S, Newman and A. F. Weinberg, J. Am. Chem. Soc, JJ3, 4654 (I956).

4. M. S. Newman and A. Kutner, ibid. , 73, ^199 (1951) •'

5. M, S. Newman W. M. Edwards, ibid., 36, 1840 (1954).

6. J. Hine, "Divalent Carbon", 'The Ronald Press Company, New York, 1964, p. 89.

7. A. J. Kresge and Y. Chiang , J. Chem. Soc, 53 (1967), and references cited therein,

8. T. L, Jacobs and S. Sear.Ies, Jr., J. Am. Chem, Soc, 66 , 686 (1944).

9. E. J, Stamhuis and W. Drenth, Rec Trav. Chim, , 80, 797 (I96.I).

10. W, Drenth and H. Hogeveen, ibid. , 7£, 1002 (1960)0

11. H. Hogeveen and W. Drenth, ibid. , 82, 375 (1963).

12. H, Hogeveen and W. Drenth, 'ibid. , B|, 410 (1963).

13. D. S. Noyce 9 M. A. Matesich, M. D. Schiavel.li, and P. E. Peterson, J, Am,
Chem. Soc, 87, 2295 (19^5) •

14. R. W. Bott, C. Eaborn, and D. R. M. Walton, J. Chem, Soc, 384 (1965).

15. P. E. Peterson and J. E. Duddey, J, Am. Chem. Soc, 85, 2865 (1963).

16. P. E. Peterson and. G. Allen, J. Org. Chem, , 27., 2290 (1961) .

17. R. W. Taft, J. Am. Chem. Soc, 8£, 709 (I963T7

18. P. E. Peterson and J. E. .Duddey, ibid. , 88, 4490 (1966).

19. P. E. Peterson and R. J. Kamat, ibid. , 88, 3152 (I966).

20. R. C. Fahey and Do Jae Lee, ibid. , 88, 5555 (1966).

21. K. Griesbaum, Angew. Chem. Intern. Ed. Engl., 5, 933 (1966).

22. T. L. Jacobs' and R. N. Johnson, J. Am. Chem. Soc. , 82, 6397 (i960).

23. K. Griesbaum, ibid. . 86, 2501 (1964).

24. K. Griesbaum, A. A. Oswald, and D. N. Hall, J. Org. Chem., 29, 2404 (1964).

25. K, Griesbaum, Angew. Chem. Intern. Ed. Engl., 3, 697 (1964).

-346-

26o Ko Griesbaum, W. Naegele, and G. G. Wanless, J. Am. Chem. Soc. , 87, 3151

( 1965) •
2T» Do Yo Curtin, J» A. Kampmeier, and R. 0! Connor , ibM0 > ^L> ^63 (1965) •
280 Do Yo Curtin, private discussion, April 4, 1967*

29« J« Ao Kampmeier, Ph. Do Thesis, University of Illinois, i960, p. 24.
30. D. Yo Curtin, J. A. Kampmeier, and M. L. Farmer, J. Am. Chem. Soc, 87,

87^ (l965)o
3L Mo Lo Farmer, Ph.D. Thesis, University of Illinois, 1964, pp0 29=30o

32. Co A. Grob and Go Cseh, Heiv. Chim. Acta, 4j, 194 (1964).

33. Co A. Grob, J. Csapilla, and G. Cseh, ibid., 47, 1590 (1964),

34. H. G. Rlchey, Jr., J. C. Philips, and L. E. Rennick, J. Am. Chem, Soc, 87,
I38I (I965).

35* H. G. Richey, Jr. , L. E. Rennick, A. S. Kuchner, J. M. Richey, and J. C.

Philips, ibid. , 87, 4017 (1965).
360 Ho Vieregge , H. M. Schmidt, J. Renema, H. J. T. Bos, and J. F. Arens, Rec

Trav. Chim., 8£, 929 (1966).

37. Ho Wo Whitlock, Jr. and P. E. Sandvick, J. Am. Chem. Soc, 88, 4525 (1966).

38. R. Wo Bott, C. Eaborn, and D. Ro M. Walton, Organometal Chem., 1, 420 (1964).

39. P. E. Peterson and J. E. Duddey, J. Am. Chem. Soc, 85, 2865 (1963).

40o No Co Deno, "Progress in Physical Organic Chemistry", Volo II., Interscience

Publishers, New York, 1964, p. 181.
4l. Wo M. Jones and F. W. Miller, J» Am. Chema Soc, 8£, i960 (1967).

-2*7-

SIGMATROPIC REACTIONS

Reported by R. N. Watson
INTRODUCTION

May 11, 1967

The concept of the sigmatropic reaction was derived by Woodward and Hoffmann,1
on the basis of their molecular orbital calculations, to correlate formally a large
number of separate reactions „ The definition and general characteristics of sigma -
tropic reactions, as well as reactions which illustrate the various types of sigma -
tropic changes, are discussed. Emphasis has been placed on the geometry of the
transition state involved in these reactions, and the correspondence with molecular
orbital predictions. Although the majority of the mechanistic work has been done on
thermal sigmatropic reactions, the available evidence for photochemical reactions is
also considered,,

GENERAL CHARACTERISTICS

Woodward and Hoffmann1 define a sigmatropic reaction of order (i,j) as an
uncatalyzed, intramolecular reaction in which a sigma bond, flanked by one or more
pi -electron systems, migrates to a new position whose termini are i-1 and j-1 atoms
removed from the original bonded loci. This terminology conforms with that pre-
sently in the literature (thus a (1^5) hydrogen shift is also a (1,5) sigmatropic
shift).2 Although not specifically mentioned by Woodward and Hoffmann as criteria
for sigmatropic reactions, it is found that many of the reactions are thought to be
concerted, and many have been found to be reversible also. Some generalized illus-
trations of sigmatropic changes are given below.

^K^

A-

(1.5) ;

( 5 »3) y

2 1

(5*5) >

It is believed that orbital symmetry relationships are the main factors which
determine the course of these reactions.3 For example, the (l,j) sigmatropic
migration of hydrogen within the all-cis polyolefinic framework (I •> II) may take
place by two paths, suprafacial or antarafacial (the transition state for this
change is thought of as consisting of a hydrogen atom and a radical containing
2k + 3 pi -electrons) . In the suprafacial process, the hydrogen appears at all times

rH r

on the same face of the pi-system, with the transition state having a plane of
symmetry, G. In the antarafacial process the migrating hydrogen passes from the
top face of one carbon to the bottom face of another, with the transition state
having a two-fold axis of symmetry, C2. 1 It is found that, in order to maintain

-348-
positive overlap between the highest occupied orbital of the olefin system and the
hydrogen orbital, the isomerization I ■» II must occur thermally (ground state
orbital symmetry) by the suprafacial path when k is odd, and antarafacially when k
is 0 or even. These results are reversed for first-excited -state transformations,
and are supported by extended Huckel M. 0. calculations.3 However, if the migrating
group possesses an available low-lying pi-orbital, alternative transition state
processes may occur. The symmetry-allowed (l,j) sigmatropic transformations
(assuming a-orbital interaction of the migrating group with the n-system) for j ^7
are given below (Table I).

Table I1

Symmetry Allowed Transformations

(1.J)

(1,3)

(1,5)
(1,7)

Thermal

Antarafacial

Suprafacial

Antarafacial

Excited State

Suprafacial
Antarafacial

For sigmatropic reactions in
which i and j are greater than 1,
proceeding through transition
states with a plane of symmetry,
thermal changes are symmetry
allowed when i + j - 4n + 2,
whereas excited -state trans -

Suprafacial

formations are symmetry allowed when i + j = kn.1 Apparent sigmatropic reactions
which violate these rules may be taking place through multi-step processes, perhaps
involving diradical intermediates, but these are expected to require vigorous
conditions.

(1,3) SIGMATROPIC REACTIONS

No established examples of thermal, uncatalyzed (1,3) hydrogen shifts could be
found (antarafacial). Woodward and Hoffmann believe this is because the carbon
framework must not become so distorted during reaction as to cause impairment of
coupling within the pi-system.1 Thus, the antarafacial process would be difficult
or impossible for j a 3.

Photochemical (1,3) reactions can be postulated for at least two cases, based
on product analysis. When 4,10-dimethyl-A3*5-hexalin (III) was irradiated in pentane,
a GOf/o yield of the isomerized non-conjugated diene (IV) was isolated, along with other

hydrocarbons.4 That this reaction is
intramolecular and uncatalyzed was not
demonstrated, but when a similar system,
7,5,82 9, H-e;rgostadiene, is irradiated, it
is found that the reaction is independent

TTT ■** TV

AA± xv of protic and aprotic solvents and gives

an isomeric non -conjugated diene with the original ergostane skeleton. No other
mechanistic work was done, but a. (1,3) hydrogen shift can rationalize these products.
A (1,3) shift involving a carbon-carbon sigma bond was postulated5 for the photo-
chemical conversion of verbenone (V) to chrysanthenone (VI) .

,H.

VII

vn:

(1,5) SIGMATROPIC REACTIONS

There have been a large number of reactions6"20*42"45 which involve the (1,5)
sigmatropic migration of a hydrogen atom. Frey and Ellis6 showed that cis-2-methyl-
l,3~pentadiene (VII) undergoes a reversible, first order isomerization in the range
!97-237°C to 4~methyl"l,3-pentadiene (VIII) . The entropy of activation of this
reaction was approximately -8 e.u., the negative value being attributed to the loss
of two internal free rotations in going from the reactant to a 6-membered cyclic
transition state (IX). Wollnsky and co-workers7 observed this (1,5) migration in a
number of other 1,3-dienes with a vinyl and alkyl group in a cis arrangement, while
trans -1,3-dienes are stable in the same conditions. Thus the reversible rearrange-
ment of X gives 15$ XI at 260°C, and 100$ XI at 360°.

Since a higher temperature is required for the equilibration of non -planar 1,3-
dienes (XII does not reach equilibrium with XIII at 450°C) , and since the trans -1,3-

dienes are stable ( trans -2-methyl-l, 3 -pentad iene gave less than 1.0$ of rearrange-
ment products at k^0u) , they suggest that the transition state has 5 carbon atoms
in a plane (or nearly planar) with the migrating hydrogen above or below the plane.
This is consistent with Woodward and Hoffmann's prediction of a. suprafacial route.

XII

XIII

The rearrangement of cyclopentadienes also proceeds by (1,5) hydrogen shifts. 8'9
Mclean and Haynes8 determined the entropy of activation for the conversion of 5-
methylcyclopentadiene to 1-methylcyclopentadiene to be -10 e.u. On the basis of an
n.m.r. study of deuterated cyclopentadiene in the range k^^6^°C9 Roth9 concluded
that the first order rearrangement which resulted in a statistical distribution of
deuterium was an intramolecular succession of (1,5) shifts. Roth also investigated
the rearrangement of 1-deutero-indene (XIV) which was thought might possibly undergo
a (1,3) shift (from indene to indene) in preference to a (1,5) shift (from indene to
the non-aromatic isoindene (XV), then back to indene). By observing the relative
intensities of the three non-aromatic hydrogen positions of 1-deutero-indene (3. 18,
3.51> an^ 6.72f)in the n.m.r. until the deuterium is statistically distributed, he
concludes that a (1,3) hydrogen shift, does not occur at all, and estimates that the
energy difference between a (1,3) said a (1,5) shift in cyclopentadiene is at least
11.5 kcal./mole.

(1.5).

XIV

(l*5'K

The thermal rearrangements of cycloheptatrienes (recently reviewed by Shermer10)
involving the migration of hydrogen around the ring have been shown to involve the
suprafacial, intramolecular (l,5)sigmatropic shift of hydrogen to the apparent-
exclusion of all other mechanisms.10

Similarly, (1^5) migrations have been observed in 8-membered ring compounds.11"18
Careful integration of the n.m.r. spectrum at various intervals of the equilibration
of neat l,3=cyclo8ctadiene (XVTa°d) at 150° for 2k hours gave kinetic results which,
according to Glass, Boikess, and Winstein,11 are consistent only with a series of
successive intramolecular (1,5) hydrogen shifts (AS' s -10 e.u.). It can be noticed
that (1,5) hydrogen shifts have comparable rates and activation energies in six
seven,10 eight,13 and nine13 member ed rings and in open-chain systems.

6*14

/y \^D

XVI

v_/a

Rearrangements within the cycloSctatriene system are also thought to proceed by
a facile (1,5) suprafacial sigmatropic shift, mechanism. 13*17-*18 Roth15 demonstrated
that 1,3,6-eyeioo'ctatriene and 1,3 ,5 -cycloSctatriene are in equilibrium at 225° by
means of a (1,5) hydrogen shift. The rate of this reversible reaction is first order
and is not influenced by solvent polarity. Although bicyclo[l|-.2o0]octa-2,4~diene is
present in the equilibrium mixture, Roth believes that this is formed from the 1,3,5-
triene only and is not an intermediate in the isomerization between the two. Roth
rules out any large contribution due to a (1,3) shift in this isomerization by
observing the rearrangement of 7j>8-ciideutero«.l,3,5=cyclodctatriene. If a (1,3) shift
is present the deuterium atoms would be statistically distributed through all positions
on the ring, whereas for a (1,5) shift the deuteriums are limited to the J>9k and 7*8
positionsa The n.m.r. of rearranged material did not show a statistical distribution
through all positions. Roth also computed the least value for the energy difference

-350-

between the (1,5) shift and the hypothetical (1,3) shift to be 7.5 kcal./mole.

The rearrangement of 5 j 8 -bis -(a-cyanoisopropyl) -l,3,6~cyclo£Jctatriene to 3,Q-
bis -(a-cyanoisopropyl) -bicyclo [^„2.0]octa- 2,4-diene is believed to proceed through
the 1,3,5 isomer.16 Kice and Cantrell16 think the 1,3*6 isomer would exist
preferentially with the bulky a-cyanoisopropyl groups in a quasi -equatorial con-
figuration (XVII, R - a-cyanoisopropyl) , in this configuration the migrating hydrogen

would be located directly over one of the double bonds and

well-situated for a suprafacial process.

Photochemical (1,5) sigmatropic reactions ( antaraf acial)

have been postulated to occur in several open chain systems,18'19

but only one (8-membered) cyclic system.20 Thus, the irradia-

tCf

XVII
the allene (XIX).

XVIII

tion of allo-dcimene (XVIII) was reported by Crowley18 to give
That antarafacial (1,5) processes have not been observed for small

ring compounds would seem to substantiate
Woodward and Hoffmann's selection rules.

R-CH3

;i,T) SIGMATROPIC REACTIONS

XJDC

•24

The (1,7) sigmatropic shift should
occur by the antarafacial route in ground
state reactions and by the suprafacial
route in excited state reactions.1 A
thermal (1,7) hydrogen migration has been proposed for the interconversion of^
precalciferol and calciferol and for conversions in analogous triene systems. s
Schlatmann21 converted cis°l-oyclohexylidene-2-( 5 8 -hydroxy-2 ' -methyleneeyclo-
hexylidene) ethane ( XX) into cis~l~( cyclohex-1 ' -ene) -2-( 5 ,,-hydroxy-2,t-methylcyclohex-
l"-ene)ethene (XXI) by heating at 70-90° for a few hours. Schlatmann determined
that the reaction is found only with conjugated triene systems with a cis -configuration

at the central double bondj
that the reaction rate
(see Table II) is independ-
ent of solvent, acids,

the entropy of activation is negative ( ~

bases, arid

inhibitors

free radical
, that there is

no exchange with CH3QD
during reaction 5 and that

le II. Rate Constants of the Isomerization (XXf=^XXI)
Under Different Conditions at 60.8°C.22

Medium
decalin

0 ethanol
ethanol +
Q.8xlO"3MHCl

k in see •
If.lxlO"5

4.6xl0~5
if.5*10~5

Medium

% ethanol +
2.3xlO~3M (CaHs)3W

70 ethanol +
1.5xlO~4M hydroquinone

k in sec"1
3.7xl0°"5

4.7xl0"5

Therefore he concludes that the reaction is intramolecular and occurs via a rigid
cyclic transition state (XXII) . The configuration of these systems will easily

permit an antarafacial process.

The photochemical (1,7) hydrogen shifts of cycloheptatriene
systems10^25"27 in contrast with the (1,5) thermal behavior,
gives significant confirmation of Woodward and Hoffmann's
selection rules. Both ter Borg25 and Roth26 observed a series

.O J.i.

of (1,7) shifts in 7-deutero~i,3,5-cycloheptatriene (XXIIa) by

integrating the n.m.r. spectrum of irradiated samples. Both
report the initial formation of the 1-deutero-isomer (XXIIb) , and believe the
equilibrium is completely consistent with a set of consecutive (1,7) shifts ( XXIIa -d).
In a similar n.m.r. study, Murray and Kaplan27 found that, although l,4-bis(7-
cycloheptatrienyl) benzene undergoes a series of thermal (1,5) hydrogen shifts, the
compound undergoes (1,7) shifts on irradiation. Their analysis of the spin-spin

■351-

XXII

coupling of this compound leads them to believe that it exists in the preferred
conformation XXIII0 Thus, positive overlap could easily be maintained between the
migrating hydrogen orbital and a framework orbital at the terminus of the migration,
thereby providing easy access to a suprafacial shift in each case.

Although the (1,3) suprafacial shift is
allowed photochemically, it has not been con-
clusively demonstrated as occurring in these
systems concurrently with the (1,7) shift. However
Murray and Kaplan57" believe that the formation of
2~phenyleycloheptatriene instead of 1 -phenyl -
cyclopehtatriene from 7-phenylcycloheptatriene is
possibly indicative of a preferential (1,3) shift
over the (1,7) shift.25 To explain the observed

XXIII
Rs4-( 7-cycloheptatrlenyl)

phenyl
dominance of the (1,7) shift, Woodward and Hoffmann1 believe that the higher values
of j are preferred in order to achieve a maximum degree of linear conjugation in
the transition state.

(3,3) SIGMATROPIC REACTIONS

The most well-known examples of (3,3) sigmatropie reactions are the Cope28"32
and Claisen33"38 rearrangements. Both reaction types are thermal (thermal changes
symmetry -allowed when i + jskn + 2-61) , intramolecular, relatively insensitive to
catalysis, and show negative entropies of activation, thus satisfying the general
characteristics of sigmatropic reactions.28"30 The primary concern here will be the
influence of orbital symmetry in determining the orientation of the transition state
in these reactions.

If the transition state of the Cope rearrangement is depicted as a complex of
two allyl radicals situated in roughly parallel planes , then the question of the
geometry of the transition state is whether the two allyl radicals are bound between
all three pairs of atoms ( six-center boat form, XXIV) or only bound through the

four terminal atoms ( four -center chair form, XXV)

28 "30

1 2

// \\

XXIV

XXV

the meso-isomer would give a mixture of cis, cis and trans

If the rearrangement proceeds
by the four-center path,
then rac-3 A -dimethyl -1,5-
hexadiene will produce
only cis , cis and trans ,
trans -octadiene , however
the me s o - c propound will
produce only cis, trans -
octadiene by this path.28""31
By the six-center path,
trans -2 96 -octadiene ,

and the racemic isomer must rearrange to cis, trans -2,6-octadiene. Doering and

trans=2,6-oetadiene from meso=

Roth31 obtained almost exclusively (99.77°) cijL

3, ^-dimethyl =1,5 -hexadiene at 225°C, and found that the rac -isomer gave IQffi cis
cis- and 90fo trans , trans -octadiene . Therefore the rearrangement proceeds by the
four-center transition state (XXV) s the free energy difference between the ^-center
and 6-center arrangements was calculated as at least 5°7 kcal./mole.

Similarly, Marvell and co-workers36 determined by conformational analysis that
the transition state of the Claisen rearrangement of cis- and trails -a,7 -dimethyl
allyl phenyl ethers is best represented by a cyclic four-center chair form. From
the rates of rearrangement of cis and trans -y -substituted allyl aryl ethers and
those of p -alkylallylaryl ethers, White and Norcross33*34 also conclude that the
chair configuration is preferred.

A qualitative explanation for this preference was offered by Doering and Roth.30
In an allyl radical the energetically lowest molecular orbital contains two electrons

distributed fairly uniformly throughout the system,, whereas the next higher orbital
has only one electron with a very small electron density at the central carbon.
Therefore two aliyl radicals with this combination of orbitals will, repel each
other at all points and can bond only at the terminal carbons, which is best
represented as the chair form.

Fukui and Fujimoto32 applied a simple Huckel perturbation calculation to the
double allyl system and determined the energy increase due to weak conjugation

between atomic it orbitals at carbons 2
op j J

2s (of XXIV) to be AE a 2P,

22: . 7 9

where

^22' - 2 2 C^ Cs' is a measure of the "overlap stabilization" (C2 and C2' denote the

J
coefficients of 2pjt atomic orbitals of the j th M. 0. , the summation being carried out
over all occupied orbitals), and 7 is the resonance energy of 2,2s conjugation.
Their calculations give AE~5 °*° 6 kcal./mole which is consistent with Doering's31
experimental value of 5.7 kcal./mole.

Woodward and Hoffmann ,39 in a somewhat different manner, also suggest that
orbital symmetry relationships play a predominant role in determining the preference
for the chair -like transition state. A correlation diagram was drawn for the
hypothetical process of two allyl radicals approaching each other from infinity, in
parallel planes. These motions involve two symmetry elements ; 0lt the plane
passing through carbons 2 and 5%, and 02, a plane parallel to the radical planes in
the boat form, or a C2 axis perpendicular to O^ in the chair form. The end products

AS 4V-
SS -4+

ss 4f-

0X02

or
tfiC2

SA
AA

-Hr-SA

41- AS

4f-ss

in this hypothetical motion are a bicyclohexane in the boat case and cyclohexyl
biradical in, the chair case.39 The essential difference in the two approaches is in
the behavior of the occupied SA level in which the boat approach correlates to an
antibonding a orbital, while in the chair form it goes over to a non-bonding radical
orbital. Thus at any point, the chair-like transition state is at a lower energy
as a result of the difference in the correlation properties of the SA orbital. By
comparing this diagram with the actual correlation diagram for the (3,3) reaction,
Woodward and Hoffmann predict that the chair form of the transition state is of
lower energy.

It appears reasonably certain that the course of the (3,3) sigmatropic reactions,
such as the Claisen and Cope rearrangements is determined primarily by the orbital
symmetry requirements. This supports Woodward and Hoffmann's basic assumption that
all sigmatropic reactions are directed by the orbital, symmetry.

-353-

(3,5) SIGMATROPIC REACTIONS

The existence of the (3,5)' process has not been effectively demonstrated, but
it could be a partial explanation for the observation that 7~14C-allyl-(2,6-
dimethyl phenyl) -ether (XXVI) rearranged under irradiation for 2 l/2 days at 25° to
give l.kio of the ^ -allyl phenol ( XXVII) with label distributed fairly evenly in the
a and y positions (at 20-30° the rate of the thermal rearrangement is practically
zero, and would give 100$ retention of label in the y position) „40 Since 2,6-
dimethyl phenol was formed in 9$ yield in these same conditions, this could be a

0~CH2~CHs€H2
' & . P 7

XXVI

AJvV XX

Position Jo Label

2.35

CH2-CH-CH2
Of p 7

97.

Position Jo Label

hi
53

case of dissociation to free radicals and recombination. A more definitive examination

of this reaction was not done,

A thermal (3,5) type process has been considered by Fahrni and Schmid29*41 to

explain the isotopic distribution obtained by heating allyl-7-14C-mesityl ether

(XXVIII-is not capable of phenol formation) for 96 hours at 170° in diethylaniline .

The heated material had radioactivity almost evenly distributed between the a and

7 atoms, whereas the starting material had label exclusively in the 7 atom of the

allyl group. This result is not possible by the accepted Claisen mechanism of a

sequence of (3,3) steps. ^ A possible pathway which could account for this result

and still preserve the observed intramolecular character of the rearrangement is a

(3,5) sigmatropic ortho-ortho1 rearrangement. This type of rearrangement may also

explain small discrepancies in the isotopic distribution found in the equilibrium

of other allyl ethers.29 Although reasonable, this mechanism has yet to be fully

established .

.s^S* 0 * 0 0^

3^L

xr/111

(3,5X

I5i3)

(5,5) SIGMATROPIC REACTIONS

Woodward and Hoffmann1 as well as Fukui and Fujimoto32 predict from molecular
orbital considerations that a chair-like transition state would be preferred in the
(5,5) sigmatropic shift of cis 9 cis-decatetraene (XXIX).

XXIX

SIGMATROPIC REACTIONS INVOLVING CYCLOPROPANE RINGS AND HEIEROATOMS

It has been noted that a cyclopropane ring may replace a it -bond in the frame-
work system for sigmatropic changes.1 This observation has been very veil documented
in the literature. n*i3>i4*3o>42-45 It has been f0Und that cis-l-methyl-2-vinyl-
cyclopropane rearranges by a (1,5) shift- at temperatures above l60° to cis°hexa-1.4-
diene,42 (+) -trans ^3-hydroxymethyl°A4 -carene (XXX) rearranges stereospecifically
( suprafacial) at 200° to (' =) «6~hydroxymethyl~A2'8-p-menthadiene (XXXI),14 1,4-cyclo-
Sctadiene and bicyclo[5.1o0. ] oct~2-ene are in thermal equilibrium above l80°((l,5)
shift),43 and (1,5) sigmatropic shifts are observed in the rearrangements of

bicyclononadienes to cyclononatrienes.11,13,45
iL^CH.3 Cyclopropyl ring intermediates are proposed for
v^\^CH2CH the Abnormal Claisen rearrangement46*47 and
similar systems.48"50

The incorporation of an oxygen in the
^ ^/Sr framework system has already been noted in

^ XXX -z?\~ XXXI the Claisen rearrangement. There is evidence

that nitrogen may also participate in sigma-
tropic reactions. Staab and co-workers51"54 believe a Cope type (3.°3) sigmatropic
rearrangement (XXXII •*• XXXIII) occurs in the thermal isomerization of double Schlff
bases of 1,2-diaminocyclopropane (XXXII) to 2,3~disubstituted-lH,l,4-diazepines
(XXXIV). An extreme example of the variations possible in the (1,5) sigmatropic
reaction is given by the rearrangement of l-(p-nitrobenzoyl) ~2,2~dimethyl aziridine
(XXXV) to N-O-methallyl) -l»p-nitrobenzamide (XXXVI) in which a 3-membered ring,
nitrogen, and oxygen atoms participate in the framework system.55"56

2000

■7

N=CH-R

(3,3)

N N

XXXIII

CH3-C-CHa

■z

5.
6.

o=c=/^Wo2

11^

CH«^
I
CH^^C "CHa

:-^Ws

xxxv

H0~C=<' ">-AM^£

BIBLIOGRAPHY

XXXIV

CH3 H 0 /—x
CH2=6 -CHg-N-C =T_N >N02

XXXVI

R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc, 87, 2511 (1965).

H. M. R. Hoffmann, Annual Reports, 62, 246 (1965).

R„ E. Cunningham, Jr., Univ. of 111. Org. Seminars, I Semester, I965-I966,

p. 103.

W. G. Dauben and W. T. Wipke, Pure Appl. Chem., 9, 539 (1964).

J. J. Hurst and G. H. Whitman, J. Chem. Sec., 2864 (I960).

H. M. Prey and R. J. Ellis, J. Chem. Soc, 4770 (1965'.

7. J. Wolinsky, B. Chollar, and M. D. Baird, J. Am. Chem. Soc., 84, 2775 (1962)

8. S. McLean and P. Haynes, Tet. Let., 2385 (1964).

9. W. R. Roth, ibid., 1009 (1964).

10. W. D. Shermer, Univ. of 111. Org. Seminars, II Semester, 1966=1967, p. 272.

11. D. S. Glass, R. S. Boikess, and S. Winstein, Tet. Let., 999 (1966).

12. E. N. Marvell, G. Caple, and B. Schatz, ibid., 389 (1965).

13. Do S. Glass, J. Zirner, and S. Winstein, Proc. Chem. Soc., 276 (1963).

14. G. Ohioff, Tet. 'Let., 3795 (1965).

15. W. R. Roth, Ann., 671, 25 (1964).

16. J. L. Kice and T. S. Cantrell^ J. Am. Chem. So:.., 8^, 2298 (1963).

-355-

17. J. M. Conia, F. Leyendecker, and C. Dubois -Faget, Tet. Let., 129 (1-966).

18. K. J. Crowley, Proc. Chem. Soc. , 17 (1964).

19. R. Srinivasan, J. Am. Chem. Soc, 84, 3982 (I962).

20. J. ZIrner and S. Winstein, Proc. Chem. Soc., 235 (1964).

21. W. R. Messer, Univ. of 111. Org. Seminars, II Semester, 1965=1966, p. 10.

22. J. L. Sehlatmann, J„ Pot, and E. Havinga, Ree. Trav. Chim. , &3, 1173 (1964).

23. R. L. Autry, D. H. Barton, A. K. Ganguly, and W. H. Reusch, J. Chem. Soc, 3313
(1961).

24. R. L. Autry, D. H. Barton, and W. H. Reusch, Proc Chem. Soc, 55 (1959).

25. A. P. Ter Borg and H. Kloosterziel, Rec Trav. Chim., 84, 24 1 (1965).

26. W. R. Roth, Angew. Chem., 25_, 921 (I963).

27. R. W. Murray and M. L. Kaplan, J. Am. Chem. Soc, 88, 3527 (1966).

28. J. C. Gaal, Univ. of 111. Org. Seminars, Summer, 19&3 s P« 1°«

29. S. J. Roads in P. de Mayo, ed., "Molecular Rearrangements," Interscience
Publishers, New York, 196.3^ Part I, p. 655.

30. W. von Doering and W. R. Roth, Angew. Chem. Int. Ed., 2, 115 (I963).

31. W. von Doering and W. R. Roth, Tetrahedron, 18, 67 (19o°2).

32. K. Fukui and H. Fujimcto, Tet. Let., 251 (ISJSG) .

33. W. N. White and B. E. Norcross, J'. Am. Chem. Soc, 8j5, 1968 (I96I).

34. Ibid., 3265 (I96I) .

35. L. D. Huestis and L. J. Andrews, ibid., 1963 (1961).

36. E. N. Marvell, J. L. Stephenson, and J. Ong, ibid., 87, I267 (1965).

37. E. N. Marvell, B. J. Burreson, and T. Crandall, J. Org. Chem., 30, IO3O (I965).

38. E. N. Marvell, B. Richardson, R. Anderson, J. L. Stephenson, and T. Crandall,
ibid., 1032 (I965).

39. R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc, 87, 4390 (1965).

40. R. Schmid and H. Schmid, Helv. Chim. Acta., 36, 6^7 (1953) •

41. P. Fahrni and H. Schmid, ibid., 42, 1102 (1959).

42. R. J. Ellis and H. M. Frey, Proc. Chem. Soc, 221 (1964).

43. W. Grimme, Chem. Ber. , £8, 756 (1965).

44. W. R. Roth and J. Kflnig, Ann., 688, 28 (1965).

45. W. R. Roth, ibid., 631,' 10 (19&TJ7

46. W. M. Lauer and T. A. Johnson, J. Org. Chem., 28, 2913 (1963).

47. R. M. Roberts, R. G. Landolt, R. N. Greene, and E. W. Heyer, J. Am. Chem. Soc.,
8g, l4o4 (1967).

48. R. M. Roberts and R. G. Landolt, ibid., 8£, 2281 (1965).

49. R. M. Roberts, R. N. Greene, R. G. Landolt, and E. W. Heyer, ibid., 87,
2282 (1965).

50. D. E. McGreer, N. W. K. Chiu, and R. S. McDaniel, Proc. Chem. Soc., 415 (1964).

51. H. A. Staab and F. VSgtle, Tet. Let., 51 (I965).'

52. H. A. Staab and F. Vflgtle, Chem. Ber., £80 2691 (1965).

53. Ibid., £§, 2701 (I965).

54. H. A. Staab and C. Wtfnsche, ibid., 3479 (1965).

55. D. V. Kashelikar and P. E. Fanta, J. Am. Chem. Soc, 82^ 4930 (i960).
^6. P. E. Fanta and M. K. Kathan, J. Heter. Chem., 1, 293 (1964).

-356-
PHOTOCHEMISTRY OF CYCLOBUTANONE S AND CYCLOBUTANEDIONES

Reported by Edward F. Johnson

May 15, 196?

CYCLOBUTANONES

The photochemistry of cyclobutanone in the gas phase has been reviewed by
Srinivasan1 and compared to the known photochemistry of other cyclic ketones.
In general, the primary photochemical process for cyclic ketones is thought to
be a ring opening to a biradical which then undergoes rapid dissociation to
stable molecules. Carbon monoxide is split out, leaving cyclic hydrocarbons and
various olefins; in addition, unsaturated aldehydes are. formed by abstraction
of a 6 hydrogen. Cyclobutanone undergoes vapor phase reactions similar to those
of its higher homologs, with the exception that only a very minor amount of the
unsaturated aldehyde is formed.2

Photolysis of cyclic ketones in the liquid phase is often dependent on the
choice of solvent.3 The first step in the photochemical process can be formally
represented by the fission of a CO-^ bond to give, most commonly, the most sub-
stituted aryl/aikyl biradical. The biradical will then stabilize itself by one
of three paths; (a) hydrogen transfer from the carbon atom_a_to the carbonyl
group to form a ketene, (b) hydrogen transfer to the carbonyl group to form an
unsaturated aldehyde, or (c) loss of carbon monoxide. The loss of CO, well
known in the gas phase s is seen only for certain ketones in the liquid phase.4

The photolysis of cyclobutanone in inert solvents has been found to give
the same products as in the gas phase; namely ethylene, ketene, propylene,
and cyclopropane o5-*6 This decomposition, both in vapor and condensed phases,
can be explained in terms of one primary photochemical process (scheme 1) in
analogy to -fhebond cleavage reported for the vapor phase reactions of other cyc-
lic ketones. First, there is an a cleavage from the n -» n state of 1 to yield
the hypothetical biradical 2 which decomposes by either cycloelimination or de-
carbcnylation, pathway A or B, respectively.

Scheme 1

h V

n -> it

Path A

Path B

-> CH2=CH2 + CH2=C=0

■1

A * A

In the presence of a reactive solvent, a third decomposition pathway is
seen to be operative. Hcstellter7 irradiated the bicyclic ketone J in methanol
and isolated two products in nearly equal amounts. He proposed the carbene h
and the ketene j? as intermediates to explain the products. No direct evidence
was presented for the presence of the carbene.

-357-

Quinkert, Cimbollek, and Buhr5 have examined the photochemistry of C-lj
epimeric D-nor~l6-keto steroids 6 and 7>

2 I

In benzene, 6 and 2 Sive the decarbonylation product 8 and the cycloelimination
product £ in product ratios of 6l:39 for 6 and 8^92 for 7, On the other hand,
acetals are the major products if the steroid is irradiated in ethanol solution
as shown in scheme 2. Steroid J also gives products showing complete retention

Scheme 2

OEt

+ 8 + Q

of configuration., The configuration was determined "by spectral properties and
by chromic acid oxidation to the lactone. The resulting lactones were compared
to authentic samples prepared from the original steroids by Baeyer-Villigar
oxidation. Therefore, the retention of configuration seems to rule out a "free"
alkyi/acyl biradical.

Turro and Southam6 have also studied this ring expansion reaction. Irrad-
iation of cyclobutanone in deutero-methanol leads to nearly exclusive formation
of 10. This product offers additional support for the carbene intermediate,
and rules out a potential mechanism involving a hydrogen shift and addition of
methanol across a double bond. The cleavage has been shown to favor the most

CH30D

OCH3

CH3OD

OCH3

substituted bond a_to the carbonyl group. If 2,2-dimethylcyclobutanone is
irradiated, none of the acetal resulting from the cleavage of the C4— -0 bond
is found. Cyclopentanone, when irradiated in ethanol, does not undergo ring
expansion, but forms 4-pentenal instead. Yates and Kilmurry8 report a case
in which a tricyclic cyclopentanone undergoes ring expansion by way of an 0x0=
carbene. If d-cyclocamphanone, 11, is irradiated in cydc?ne.>sne,the carbene is
trapped, and the resulting product was isolated and shown to be 12.

■358-

h-y"

HEKZOCYCLOBUIENEDIOHES

Oxocarbene s have also been proposed in the photolysis of benzocyclobuten-
edione, 1^, in solution. Brown and Solly9 have isolated two dimers in the
photolysis of 1£ in degassed cyclohexane using natural sunlight filtered through
pyrex. The dimers, 14 and 1J? (5.5 and 38$ yield respectively) , are proposed to
result from the reaction of the oxocarbene 16 with 1^ or an excited state of 13„
More definite evidence for the oxocarbene results from the work of Staab and
Ipaktschi.10 They irradiated 1^ in pentane; methylene chloride (2:1) solutions
at 20° using a mercury high pressure lamp and isolated three dimers^ Ik (25%) ,
the cis isomer of 14 (5$), and 17 (h$) . No evidence is presented to show whether
1J is formed as a result of a photochemical process on 1^ or as a result of
rearrangement of another product . The oxocarbene 16 is trapped as the ethyl
acetal by irradiating 13 in refluxing alcohols while a control experiment carried
on without irradiation yielded no acetal after refluxing in alcohol for 12 hours.

±2

0 ±k

Xl R2R3 R4
18

Irradiation of 1£ in excess alkenes (propylene, isobutylene^, cyclohexene , butadiene
and ethyl vinyl ether) at 20° for 2k hours gave the spiro-lactone structure 18.

CYCLOBUTANE-1 , 3-DIONES

An examination of the ultraviolet spectrum of tetramethylcyclobutane-1,3-
dione^ 20, showed an n -» it transition in the 30C mu region.12 Kosowerls observed
a second n -> it transition in the ^>k0 mM- region. The low intensities of the
transitions as well as both the direction and the magnitude of the solvent ef°
fects are consistent with n -» «* transitions. The bands are observed at j>48 mM-
(€ 18) and 308 mi-i (6 39) in iso-octane; 3kk nm (6 18) and 30^ mH (6 30) in
ethanol. The two different transitions are explained by an interaction between
the carbonyl groups in the excited state and are described as follows;

*+ *2* (3^0 mu band)

A
B

n -> it-

n -> m'

Jtg

(300 mu band)

-359-
The excited state A may be pictured as shown below.

+
h v \

20 A

The photolysis of tetramethylcyclobutane-l,3-dione in the vapor phase had
been carried out by Turro, Leermakers, Wilson, Necker, Byers, and Vesley.14
They found that complete photolysis at low pressure yields 2.0 moles of carbon
monoxide , O.O78 moles of propylene, 0.0024 moles of methane, and a trace of
propane for every mole of the dione. There is a slight induction period in the
evolution of carbon monoxide and a rather obvious induction period for propylene.
This difference, and also the low yield of olefins compared to the production
of carbon monoxide, is explained by the early formation of high molecular weight
polymers which were observed on the sides of the reaction vessel. The photolysis
was carried out using pyrex vessels and mercury arc lamps. These conditions were
used for all the experiments reported in the remainder of the seminar, unless
otherwise noted.

Several groups of workers have looked at the photolysis of the dione in
inert solvents. Cookson, %e, and Subrahmanyam15 have irradiated 20 in benzene
solution. They obtained tetramethylethylene (80$ yield) and carbon monoxide.
Turro and co-workers16 observed that a ketene is also a product of the reaction.
Evidence for the presence of dimethylketene is suggested by the yellow color of
the reaction mixture, the strong infrared band at 2124 cm 1 „ and the disappear^
ance of both the yellow color and the 2124 cm"1 band when iso=propyl alcohol
was added. The quantum yield for the disappearance of 20 was 0. 38 + 0.01.

The formation of an olefin and carbon monoxide seems to be a general reaction
for tetra- substituted cyclobutane~l,3~diones. The exhaustive photolysis of di~
spiro[5.1.5<>3j-"fcetradecane~7>l4-dione, 21, in degassed benzene and methylene
chloride yields bicyclohexlidene (70$ yitJ-d) and CO.14 If a benzene solution

h v v < >==/ \ + 2C0

of tetramethyl- and tetraethylcyclobutane-l,>dione is photolyzed, high yields
of the expected tetramethyl- and tetraethyl- olefins are obtained., but no di-
methyldiethylethylene could be detected (limit of detection about 0.1$). The
yellow color of a ketene was also observed in the latter reaction.

If the photolyzed solution of 20, after low conversion, is analyzed by
vapor phase chromatography (vpc), the presence of a new component is detected
in significant amounts. This new component corresponds to isopropenyl isopropyl
ketone, 22, and after isolation by vpc, was shown to be identical to an authentic
sample. However, an examination of the reaction mixture by spectral methods showed
the ketone 22 to be absent. Therefore, it was thought that the ketone 22 was
formed as a result of decomposition of an unknown precursor upon analysis by vpc.
As the photolysis of 20 proceeded, 22 reached a steady state but dropped to l/4
if the reaction mixture was allowed to stand in the dark for 15 hours.

If dione 20 is photolyzed in benzene under 540 mm of oxygen and the reaction
is followed by mass spectroscopic and vpc studies, the products shown in scheme
3 are observed (with appropriate mole ratios).17 No isopropenyl isopropyl ketone22
was detected by vpc. It was shown that both the ketone 22 and tetramethylethylene
were stable to both light and oxygen under the reaction conditions. Therefore,

hV,

<t>H. 05

=360»
Scheme 3

CH3CCH3 + CO + C02 +
1.5 0.8 0.4

0.07

the precursor of ketone 22 forms tetramethylethylene under degassed conditions
and acetone and tetramethylethylene oxide in the presence of oxygen,,

Much evidence has been presented that the precursor of isopropenyl isopropyl
ketone is tetramethylcyclopropanone, 2J5, or its excited state , for which many
resonance forms can be drawn as shown below. Several workers have been successful
in trapping the cyclopropane ne intermediate and finally in isolating it from the

0 0.

Ac <-* jK <—> etc

< — >

< — > J/

reaction mixture . Richey, Richey, and Clagett18 irradiated tetramethylcyclo-
butane-l,3=dione in 5$ solutions cf ethanoi and were able to isolate tetramethyl-
cyclopropanone ethyl hemiketal, 24 , in 35$ yield by recrystallization and 55$
yield by vpc. Also detected were the esters 2£ (20-25$), 26 , and 2J and ketone 28.
The last three were found in 10-15$ yield and were believed to result from

OLOCH2CH3

A-

0
II
C-0CH2CH3

■C-OCH2CH3
21

0CH2CH3

decomposition of the hemiketal. Isolation and the resulting decomposition of
the pure hemiketal have shown this assumption to be correct.19

Turro and co-workers14'16 obtained corresponding products from the photol-
ysis of the dione in methanol. They also detected a 5$ yield of tetramethyl-
ethylene. The quantum yield for the disappearance of 20 in methanol was found
to be 0.49 + 0.03. Cookson and co-workers15 carried out the photolysis in the

presence of furan and isolated adduct 2£ in 15$ yield after distil-
lation. Turro and co-workers20 were able to isolate tetramethyl-
cyclopropanone, 2J>, and examine its properties. Saturated pentane
solutions of the dione 20 were photolyzed in a pyrex Hanovia 450-w
immersion apparatus at ~^E° for 1 to 2 hours. Longer reaction times
cut down the yields of the cyclopropanone. The resulting solution
was stripped of the tetramethylethylene and pentane until an approx-
imately 10$ solution of 2J> was obtained (estimated by infrared spectroscopy).
This remaining pentane solution of 2^ was purified by bulb to bulb distillation
atJL mm and 20°. A band for the C=0 stretching frequency was found at 1840
cm , which would be expected for such a small ring carbonyl compound. They
found that the infrared band disappears when either furan, oxygen, or methanol
is added to the pentane solution. When injected into the vpc, 2J rearranges to
isopropenyl isopropyl ketone. The authors concluded that tetramethylcycloprop-
anone "is a fairly stable, distillable compound which can be handled in pentane
solution." Furthermore, all the various trapping experiments to prove the

-361-

intermediacy of 2^ in the photolysis of the dione 20, are valid. In fact, all
the reactions reported for tetramethylcyclopropanone under the photolysis con-
ditions will also proceed in the absence of light starting with pentane solutions
of the cyclopropanone. The oxidation of 2J> in the presence of oxygen is thought
to proceed through the peroxides ^0 and 31 which decompose to the observed products,
acetone and tetramethylethylene oxide. 14>XT

=0 + CO /"\

By using pyrex glass the excitation was primarily located in the lowest
energy transition of the dione ( the 3^0 mu region) . The use of filters to
excite only the lowest energy transition did not appear to alter the course
of the reaction.14 No fluorescence or phosphorescence from 20 was detected at
77°K in an ethanol: ether: isopentane matrix. The lack of emission from 20 implies
an extremely rapid path of deactivation from the lowest excited state. There is
some indirect evidence which indicates that the chemically active state is the
n -> n* singlet. Since the quantum yield for decomposition of 20 is high, the
most likely paths for deactivation of the lowest excited state are (a) photochem-
ical reaction, or (b) intersystem crossing to the lowest triplet^ followed by
decomposition. The photolysis of the dione is neither sensitized by benzophenone
(E = 69 kcal per mole) nor quenched by 1,3-pentadiene (0.3M). These facts imply

that the lowest triplet is not involved. However 5 this result is not conclusive
because the triplet of benzophenone may not be of sufficient energy to excite
the dione, and the decomposition of dione in the triplet state may be faster than
energy transfer to the quencher.

In order to decide what the primary photochemical processes in the photol-
ysis of tetramethylcyclobutane-l,3-dione were, Haller and Srinivasan2x took the
infrared spectrum of the reaction mixture as soon as possible after photolysis
by using a cell which allowed direct infrared analysis at variable temperatures.
The reaction was studied at room temperature and in a nitrogen matrix at k°K, and
conversion of 20 to products was held to 10$ or less to insure that no secondary
processes would take place. The workers were able to follow the production of
tetramethylcyclopropanone by a band at 1840 cm"1 and the dime thy lketene by a
band at 2124 cm 1. Both bands were observed at both temperatures indicating two
primary photochemical processes were taking place, as shown in processes C and D.

From the extinction coefficient of dimethylketene (at 2124 cm"1), it is
estimated that this product accounts for at least 20$ of the disappearance of
the dione. Concurring results also were obtained by an ultraviolet study.
The yield of esters formed when alcohol reacts with the dimethylketene has already

-362-

been shown to be between 20-30$, 14?1S which agrees with the physical, data. How-
ever, process D severely complicates the photochemistry of 20 for the following
reasons; l) If dimethylketene is irradiated at 2537 ft in cyclohexane using a
vessel with KBr windows, tetramethylethylene , carbon monoxide, and tetramethyl-
cyclopropanone are observed. Since conversion was kept negligibly small, the
cyclopropanone must come only from dimethylketene. The reaction can be explained
in the following manner. The photolysis of dimethylketene in the vapor phase

>=« ^-^ >;

>; + >— c=c

+ CO

> A + co

yields dimethyl carbene at either 2537 or 366O $., although the quantum yield is
much lower when the higher wavelength radiation is used.22 An analogy for the
latter reaction is the known procedure for making cyclopropanones from diazometh-
ane and ketenes.23 Therefore, photolytic decomposition of dimethylketene may also
be a source of tetramethylethylene and tetramethylcyclopropanone when the dione
is photolyzed in an inert solvent. 2) If oxygen is bubbled through a solution
of dimethylketene in cyclohexane, the band at 2124 cm"1 disappears and new bands
are seen at 2324 cm""1 (C02) and 1720 cm""1 (acetone). 3) Dimethylketene is known
to undergo dimerization - xclusively to tetramethylcyclobutane-l,3-dione at
room temperature.20-'24

If oxygen is added to a partially photolyzed solution of the dione, the ke-
tene disappears quickly, but the cyclopropanone is found to diminish rather slowly.21
This reaction casts doubt on the efficiency of oxygen as a trapping agent for
the cyclopropanone. However, Turro and co-workers20 claim that when oxygen is
added to a pentane solution of isolated tetramethylcyclopropanone, it disappears
very quickly. No apparent reason is offered for this conflict.

By assuming that the extinction coefficient for cyclopropanone is the same
as for cyclobutanone , Haller and Srinivasan21 estimate that the primary process
C_ is of no more importance than process D. Therefore since there are 2 moles
of dimethylketene for every mole of tetramethylcyclopropanone, primary process
D accounts for no more than 40$ of the disappearance of the dione.

These same workers found still a third primary process which is important.25
This process involves the loss of 2 moles of CO to form t etramethylethylene dir-
ectly. Evidence for this process was a band at II76 cm 1 in the infrared spectrum

+ 2C0 E

of the photolyzed dione, at low conversion, in a nitrogen matrix. This band
was assigned to the ethylene derivative. Better evidence for process E was pro-
vided by an ultraviolet study of a cyclohexane solution of the dione 20. The
spectrum showed a marked increase in absorption in the region of 240 to 200 mu
on photolysis, at low conversion, in a quartz cell. The contribution to this
band by dimethylketene was determined to be minor, while the contribution due to
the cyclopropanone derivative was even less. In control experiments it was shown
that this absorption was not due to products from the reaction of ketene with water
or oxygen.

A summary of the reactions taking place when tetramethylcyclobutane-1,3-
dione is photolyzed is shown in the following schemes

.565-

>=<

+ 2C0

h-r

=c=o

+ CO

It is difficult to estimate the extent of processes C and E. for several
reasons. First, the secondary processes give products identical to those formed
in the primary processes. Second, the nature of the process which converts
tetramethylcyclopropanone to tetramethylethylene is unknown. However, direct
photolysis of the cyclopropanone is unlikely at low conversion of 20, so that the
presence of tetramethylethylene in the early stages of the reaction, requires
the third primary process. As reported earlier in the seminar, trapping experi-
ments showed 5% substituted ethylene present when the cyclopropanone was trapped
by an alcohol and in the presence of a high pressure of oxygen. However, as it
has been pointed out, oxygen may or may not be a good trapping agent.

It has been shown earlier that there is an interaction between the carbonyl
carbon atoms in the excited state of the dione.13 This effect could decrease
the distance between the other two carbon atoms which would be required for the
simultaneous loss of two carbon monoxide molecules.

Process E^ according to Srinivasan,25 accounts for 40$> of the disappear-
ance of the dione. However, Turro14 maintains that the trapping of cyclopropanone
with oxygen, furan or an alcohol is complete, and this trapping process accounts
for a majority of the disappearance of the dione. Process E accounts for only
5$. Srinivasan21 points out the trapping experiments may not be valid for the
following reasons:

a) Oxygen, furan, or an alcohol may react with an excited state precursor of
tetramethylcyclopropanone, which otherwise may lead to other products such as the
substituted ethylene, b) A solvent effect may change the course of the reaction
in a more polar solvent.

Prolonged photolysis of tetramethylcyclobutane-l,3-dione leads to still another
product, the lactone jj2, in low yields.20 Product J52 might be thought of as a

dimerization product of dimethylketene. However, the thermal
dimerizati on of dimethylketene leads exclusively to the tet-
ramethylcyclobutane-l,3-dione.24 A photochemical reaction of
dimethylketene under the same conditions of the photolysis
reaction of the dione, did not lead to significant amounts of
the lactone.20 A possible mechanism for the lactone reaction
is shown below. Cookson and co-workers26 have also reported a low yield of the

hv"

V- ?

lactone, along with other products, in the photolysis of 20.

They have also shown that this lactone formation takes place in other cyclic

=364-
diones. For example, when 2,2,4,4-tetramethyl~cyclohexane-l,3-dione is irradia-
ted^ lactone j>4 is found in 35-56$ yield.

21 &

CONCLUSION

CyclolxitarDne has "been found to undergo most reactions expected for a cyclic ketone j
however, in alcohol solvents , a ring expanded acetal product has been observed. This
product is believed to result from a carbene intermediate. An oxocarbene intermediate
is also proposed in the photolysis of benzocyclobutenediones to yield ring expanded
acetals. The photolysis of cyclobutane~l,3-diones has been found to go by three dif-
ferent paths. The secondary processes combine to make the understanding of the various
primary processes difficult.

BIBLIOGRAPHY

1. R. Srinivasan, Advan. Photochem. , 1, 83 (1963).

2. R. K. KLemm, D. N. Morrison, P. Gilderson, and A. T. Blades, Can. J. Chem.,
i£, 193^ (1965).

3. R. 0. Kan, "Organic Photochemistry," McGraw-Hill, New York, N. Y. I9660

4. J. E. Starr and R. H. Eastman, J. Org. Chem. , ~%ko !593 (19o6).

5. G. Quinkert, G. Cimbollek, and G. Buhr, Tet. Letters, 4573 (1966).

6. N. J. Turro and R. M. Southam, Tet. Letters, 545 (1967).

7. H. U. Hostellter, Tet. Letters, 687 (1965).

8. P. Yates and L. KLlmurry, Tet. Letters, 1739 (1964); J. Am. Chem. Soc, 88,
I563 (1966).

9. R. F. G. Brown and R. K. Solly, Tet. Letters, 169 (1966).
10. Ho A. Staab and J. Ipaktschi, Tet. Letters, 583 (1966).

12. E. A. LaLancette and R. E. Benson, J. Am. Chem. Soc, 8^, 4867 (1961).

1> E. M. Kosower, J. Chem. Phy. , ^8, 2813 (I963).

l4. N. J. Turro, P. A. Leermakers, H. R. Wilson, D. C. Necker, G. W. Byers, and G. F.

Vesley, J. Am. Chem. Soc, 87", 2613 (1965).
15- R. C. Cookson, N. J. Nye, and G. Subrahmanyam, Proc Chem. Soc, 144 (1964).
l6. N. J. Turro, G„ W. Byers, and P. A. Leermakers, J. Am. Chem. Soc, 86, 955 (1964).
17- P« A. Leermakers, G. V. Vesley, N. J. Turro, and D. C. Neckers, J. Am. Chem.

Soc, 86, 4213 (i$6h).

18. H. G. Richey, J. M. Richey, and D. C. Clagett, J. Am. Chem. Soc, 86, 3906 (1964).

19. N. Jo Turro, W. B. Hammond, P. A. Leermakers, and H. T. Thomas, Chem. and Ind.

990 (1965).
20» N. J. Turro, W. B. Hammond, and P. A. Leermakers, J. Am. Chem. Soc, 87,
2774 (1965).

21. I. Haller and R. Srinivasan, J. Am. Chem. Soc, 87, 1144 (1965).

22. R. A. Holroyd and F. E. Blacet, J. Am. Chem. Soc, 7£, 4830 (1957).

23. W. B. DeMore, H. 0. Pritchard, and N. Davidson, J. Am. Chem. Soc, 8l, 5874
(1959); W. B. Hammond and N. J. Turro, J. Am. Chem. Soc, 88, 2880 JV)GG) y

N. J. Turro and W. B. Hammond, J. Am. Chem. Soc, 88, 3672 (I966).
2K W. E. Hanford and J. C. Sauer/Org. Reaction, ^, 127 (1946) 5 R. H. Hasek,

Research (London) l4, 74 (l§6l).
25- I» Haller and R. Srinivasan, Can. J. Chem., 4^, 31^5 (1965).
26. R. C. Cookson, A. G. Edwards, J. Hudec, and M. KLngsland, Chem. Comm. , 98 (1965).

■z£c

-pop-

THE MECHANISM OF PAPAIN CATALYSIS
Reported by Paul Elliot Bender
INTRODUCTION*

May 18, 196\

Papain is a crystalline proteolytic enzyme found in highest concentrations in
the secreted fluid of latex vessels under the skin of the tree Caprica papaya.1
Named in 1879 by Wurtz and Bouchut,2 who performed the first veil controlled
experiments on the crude extract, it has played an important role in the establish-
ment of many basic facets in our present conceptions of enzyme action.1 The
literature concerning papain has been extensively reviewed to 1962.1-?3i'4"s A review
of the work on primary structure appeared in 19647 and some findings of more recent
studies, (I965), have been outlined.8 This seminar will emphasize the material of
the last four years concerning the mechanism of the papain catalyzed hydrolysis of
synthetic substrates.

ACYL ENZYME INTERMEDIATE

The first direct evidence of an acyl papain derivative was the observation by
Lowe and Williams,9,910 in a difference spectrum,, obtained thirty seconds after
adding methyl thionohippurate shown in Figure I to the activated buffered enzyme,

NH2

C-NH
I

(CH2)3

NH— CH

BAEE

-OCaH5

NH2

C=NH

J s

(CH2)3

NH- CH

BAA

■NH;

NH2

I
C^O

I
NH

I
(CH2)3

:«cnh— CH

OCH.

BCME

NH2

( QH2) 4

C6H5CH20-.CNH— CH

■OCgHspNOa CeasCKHCHa-C-OCHs

Z-L-lysine-p-nitrophenyl ester

methyl
thionohippurate

Figure I. Structural Formulas

CsHsCNHCHaC-OCaHs

ethyl hippurate

of a single U.V. absorption band at X 313 rap, (log es4.3 + C.3) whose O.D. dropped

iiislx ■"

max

to 0 in 12.5 minutes. From a comparison to the model compounds shown in Table I10

below, the authors suggested that the group most comparable in terms of X and

Cbromophore

ethyl dithioacetate
methyl thionohippurate

Table

Acyl Enzyme

Models

1 C1)

-"max
(rap)

log e
0 max

x ('

^max
(mu)

305

4.08

460

230

»K 26

291

4.00

2p0

log e

max

1.25

4.00

*In this paper a-N~benzoyl~L-arginine ethyl ester, a~N-benzcyl~L~argininamide, a-N-
benzoyl-L-citrulline methyl ester, and a-N-benzyloxycarbonyl will be abbreviated as
BAEE, BAA, BCME arid Z, respectively.

-366-

Table
X_ W
(mu)

I (Conto)

log e
& max

358

1.25

278

k.71

(2)
Ghromophore Xmax"' log e^ X^ log e^

(mu)

=UHS 358 1.25 268 ito05

papain at pH 7

log e, whose presence in the intermediate could account for the observed difference
max

spectrum, is the dithioaeylester moietyD It is quite obvious that both the substrate

and the enzyme absorb well below 313 mu<> Upon adjustment of the pH from the initial

60 0 to 2.5 (where denaturation is known to occur)11 the 313 mu band shifted to

309 mu and the absorption was maintained.

The spectral evidence for the second acyl enzyme , trans -cinnamoyl papain #12>13

also had to be obtained via a difference spectrum, since for this acyl intermediate ,

the rate of deacylation is comparable to that of acylation. To prepare a partially

acylated enzyme, excess trans -cinnamoyl imidazole was added to a pH 3«^3 buffered

activated papain solution. After five minutes the mixture was chromatographed or a

Sephadex G~25 column to isolate the acyl enzyme. The difference spectrum of a

fraction, so obtained s containing the acylated enzyme was scanned spectrophoto-

metrically from 39O to 2^0 mu, revealing an absorption at X . 326 mu (log e*k.k2k) .

Comparison of the X and log e of activated native trans -cinnamoyl papain to

that of trans -cinnamoyl cysteine (the thiol ester model) and N«acetylserinamide

(the hydroxyl ester model) as shown in Table II13 below indicates X of the thiol
x J J ■ max

Table II

Acyl Enzyme Models

trans -cinnamoyl derivative X (mu) log e (mu)

papain 326 kek2k

N-acetylserinamide 281.5 ^0385

Cysteine 306 h0^k

ester model to be closer to that observed for the acyl enzyme. Upon denaturation of
trans -einnamoyl papain in 4.8M guanidinium chloride X shifts to 3OI-309 mu which,

TUSLX.

as in hippuryl papain, is in good agreement with the acylated thiol model in a more

uniform nonperturbed environment. Indication of the intermediacy of the observed

species is provided by the following evidences (1) A complete system showed an

absorption at 35^ mu which first increased to a maximum and then decreased to zero.

(2) The isolated species has the characteristic extinction of the trans -cinnamoyl

moiety and a X which differs from the reactant trans -cinnamoyl .imidazole (X

max J v max

307 mu) and the product trans -cinnamate ion (X 269 mu) „ (3) The trans -cinnamoyl

max

moiety is not separated from the native or denatured enzyme on Sephadex filtration,

implying eovalent bonding. (V> Addition of the specific substrate, BAEE (shown in

Figure I) , to the trans -cinnamoyl enzyme showed that the rate of deacylation of

the trans -cinnamoyl group was coincident with the rate at which papain catalyzed

hydrolysis of the specific substrate reappears.14 In the light of the kinetic

evidence for BAEE hydrolysis involving a thiol group s (see section on kinetics) and

the implication of a thiol group at the active site, Brubacker and Bender concluded

that observation (If) implied a bonding of the trans -cinnamoyl moiety at the same

active site, (probably the thiol group) responsible for BAEE hydrolysis.

Also indicative of a eovalent acyl enzyme intermediate was the work of Kirsch

and Katchalski.15 The investigators compared the enzyme catalyzed to hydroxide

catalyzed ratio of the hydrolysis rate to the 180 exchange rate of acyl-thiol

carbonyl labeled 180»ethyl hippurate (shown in Figure I). Assaying the remaining
ester vs. time, they found a hydroxide ion catalyzed ratio of lks but the enzyme
catalyzed ratio of 80 was within experimental error of no detectable exchange.
The authors concluded that the enzymatic pathway for hydrolysis either passes
through an acyl enzyme intermediate or through an enzyme substrate complex, which
hinders 180 exchange sterically.

The further finding by Brubacher and Bender13 that in the deacylation of
trans -cinnamoyl papain in the presence of added nucleophiles, the rate of nucleo-
philic catalysis by amines was much greater than for the oxygen analogues, was
taken by the authors to implicate a thiol ester as the enzyme intermediate.

ET^Bffi KINETICS

Papain has a broad specificity toward amides , and wiH hydrolyze a polypeptide
more completely than either pepsin or trypsin.4 Kinetic investigations of papain
catalyzed hydrolysis have relied mainly upon synthetic substrates such as N-acyl
a -amino derivatives of esters and amides of a-amino acids. Kinetic studies on these
substrates have yielded much significant information as to the mechanism of
catalysis.

Since Mf.chaelis-Menton kinetic parameters have been employed throughout this
paper, they are derived below assuming both the kinetic path shown and steady state
conditions.

E + S ^=* ES — k* E + P

0 * ka(S)(E) - (k=a+ kb)(ES)

where (E)»

of free enzyme

Ejs (E) + (ES) where (E)« concentration

(Ec)s enzyme added
initially

dt (S) + k„a + kb

V -4* k

definitions^- k . « k, 9 Ks -^=^-r- — k., k - ES dissociation constant -

a s

Smith and coworkers g,1 studying the pH and temperature dependencies of BAEE

hydrolysis of the Michaelis Menton kinetic parameters, found k^ ,/k to be a concave

cat m

downward bell shaped curve as a function of pH. The limbs of the BAEE curve appear

to represent the titration of two prototropic groups in the enzyme with pKx«4.3 and

pK2®8. 02 at 37°o The shift in the descending limb with temperature indicated a

heat of ionization of 5.1 kcal/mole at 0°, which could correspond to either an a-

amino or a thiol group in the opinion of the authors. They suggest the latter

alternative due to the proven necessity of a free thiol group to enzyme activity.

A lack of significant shift in pKx with temperature, implicated , to the authors, the

titration of a carboxylate ion other than an a-carboxyl group, as shown by the

negligible heat of ionization.

Although it was shown that positively charged substrates are hydrolyzed more

readily than neutral ones, and that negatively charged substrates show inhibition,

the shape of the k ./k vs. pH curves for basic BAEE, BAA. (shown in Figure I), and

cat m

neutral hippur amide were nearly identicals only the value of k ,/K (lim) was

greatly different. This indicated to the authors that acylation proceeded by the
same mechanism in all of these eases (i.e., all dependent upon two prototropic
groups). A plot of k vs. pH for BAEE showed this rate constant to be pH

independent down to pH 5.0„ Below this value, a k , decrease was seen, which was

Cav

-368-

th ought to be indicative of a single titratable group of pKa 3° 5 at 25° 9 active in

the decomposition of an acyl intermediate to products. However, correcting Smith's

pH-Stat data for the state of ionization of the product at low pH, Sluyterman16

found no pH dependence . Williams17 has also found no pH dependence in the hydrolysis

of methyl hippur ate .

In a study of papain catalyzed hydrolysis of a series of hippurate esters,18

Lowe and Williams found that nine of the eleven esters studied have essentially the

same value of k (2-4 sec"1)., Kirsch and Igelstronr19 found similar k . independence
cat cat

of leaving group in a series of carbobenzoxyglycine esters „ To account for this

observation, in light of an acyl enzyme intermediate , Lowe and Williams proposed the

following kinetic scheme (where k2, k3, Vx and P2 are the acylation rate constant,

deacylation rate constant, an alcohol, and hippuric acid, respectively).

E + S

assuming steady state conditions

ES!

•f

■> E + Ps

(Eo)(S)

L$LL

ko,

(k^-k3)

(k..i+ka)

definitions s k

cat

k ,/K

cat/ m

k.k-

(k»!+ka)

in which k2 )>^> k3 and therefore k

k3„ According to these kinetics, if k ,/

cat'

m
with p

l0 A plot of log (k. ,/K ) vs. a gave a good linear fit

c ax m

k ,/K to be ^

cat' m

»/K where K

* s s

'cat ■' "3
= ka/K , then k2 « k , „ A -plot of log (k.^/K ..]

!, indicating, according to the authors,

a constant and therefore k2 <^ k^:i.

The functional dependence of the deacylation rate constant upon pH was determined
by Brubacher and Bender13 using the isolated trans -cinnamoylated papain aided to the
desired buffer solution, and the absorption of the difference spectrum at 330 mu.
followed with time0 The authors observed the deacylation rate constant (k3) to
increase in a sigmoid fashion with pH to a k3 (lira) value cf 3«68 x 10"30
Deacylation was apparently dependent upon the ionization of a single protctropic
group of pKa ^.69, which the authors proposed to be a carboxyl group,,

In a comparison of papain catalyzed BAA to BAEE hydrolysis, Whitaker and
Bender20 spectrophotometrieally reinvestigated the kinetics of hydrolysis, extending
the pH range employed by Smith and coworkers1 and analyzing the kinetic data in
terms of the rate constants and prototropic equilibria illustrated below by making

EH2

11 Ki

EH +

kiClimt
^ k_,

EH2S
EHS

J[k2

EH2S:

Mlim);

EHS5

k3(lim.)

^ EH + P;

+ Pi

certain assumptions s (1) Kg is pH independent above pH 5.0, (2) BAA and BAEE share
a common mechanistic pathway in their papain catalyzed hydrolysis, (3) k3 is
independent of pH at high pH, and (k) steady state conditions. The authors observed

/Km vs., pH dependencies of BAA and BAEE are identical in form,

are very different. They conclude

that although the k
the ]

vs,

cat/ m
pH dependencies of the two substrat-

cat

from this, that granting assumption (2), two rate steps are involved in these
hydrolyses, and of the two one is slower in amide hydrolysis whereas the other is
slower In ester hydrolysis. A complete analysis in terms of the proposed rate

-369-

constants demonstrates the following; (1) a sigmoid dependence of k3 to pH^ (2)

the pKi8 for BAEE deacylation is the same as that for BAA deacylation ^ which in
conjunction with (1) supports assumption (3) 9 and shows the dependence of deacylation
upon a group of pKa 3.91^ (3) k3 (lim) for both BAA. and BAEE are equal within
experimental error 9 which is to be expected for the deacylation of a common aeyl
moiety °9 (4) k2 (lim) for BAEE is 3D2 fold greater than k3 (lim)§ (5) k3 (lim) for
BAA is three fold greater than ks (lim) 3 (6) the values of pKx and pK2 for the
aeylation step of BAEE, 4.29 and 8.49 respectively,, are essentially indentical to
those for BAA9 and similar to the values obtained earlier by Smith and coworkers.1
Employing these values^ internal consistency of the data to the proposed scheme of
equation (2) was shown by using the derived values of the limiting rate constants s
prototropic dissociation constants and the value of the substrate dissociation con-
stant to calculate k ,/K s k , and K vs. pH profiles which displayed good

C8iU 233. C £tXi lu

correspondence to the empirical data^ except in the region below pH 4.5 io the K
and k , vs. pH plots for BAA. Bender and Brubacher21 offer an explanation for

these discrepancies as perhaps representing either an increase in K with pH (in
this range) due to the increased repulsion of the positively charged substrate by
the progressively more positively charged enzyme 9 or a perturbation of the enzyme of
such a nature as to shift the pKx of the enzyme substrate complex below the pK2 of
the free enzyme.

In conflict with the bell shaped pH dependencies of k . obtained by Whitaker

and Bender^20 Sluyterman26 and Williams17 both reported that k , was pH independent

for ethyl hippurate hydrolysis down to pH 4.2 and 3=8 respectively. Lowe and
Williams31 suggest that the pH dependence of deacylation in benzoylargininyl papain
is due simply to binding of the positively charged guanido group by a carboxylate
ion ^ assisting deacylation by orienting the thiol ester bond and perhaps modifying
the conformation of the acyl enzyme. They note that although the k3 of benzoyl-
argininyl papain is 7 times greater than k , of ethyl hippurate (taken as

k, , +.son) at pH 6.0^, the former is approximately equal to the latter at pH 3*0.?

where the oarboxyl group is largely prctonated. This hypothesis would predict that
the masking of the basic group (e„g.5 with a N-formyl group) should reduce k^ .

(lim) to the area of 2.7 sec"1 (the k . of ethyl hippurate} . The recent work by

Bender and Brubacher21 has shown that for Z-L-lysine p-nitrophenyl ester (shewn in
Figure I) s the value of k + (lim) drops from 45 sec"1 to 32 sec"1. The authors

consider these two values to be the same. Furthermore^ substitution of BCME (shown
in Figure I) for BAEE,, (which provides a substrate with an isosteric acyl function) s
has been recently found by Cohen and Petra23 to exhibit no effect (20.2 vs. 20.15
sec"1) on k3 (lim) calculated similarly. The latter results imply that the presence
of a positive charge on the substrate does not increase the deacylation rate.

Kinetic studies by Bender and Brubacher21 with the series p-nitrophenyl^, benzyl
and methyl Z-L-lysine esters as a function of pH^, analyzed in terms of the scheme
of equation (2)$ similar to the treatment of Whitaker and Bender^20 have followed
similar k ,s k^ ,/K 9 and K vs. pH dependencies to those shown by BAEE. For this

09X CQ.X HI HI

series of esters 3 k3 was in all cases at least 3.5 fold smaller than k2. However^

it was found that K decreased from a value of 10.7mM for the methyl ester to

authors state that a change in K., of this magnitude strongly implies that the

substrate leaving group is bound at some enzymatic site. This dependence of Ks upon

the structure of F^ is in conflict with the conclusion of Lowe and Williams18 based

upon the linear dependence of log (k ,/K ) vs. 6 for four aryl hippurate esters.

cat m

MECHANISTIC PROPOSALS

Several mechanisms .have been proposed in the course of papain investigations in
order to explain some of the experimental observations. Most of those proposed agree
that an acyl enzyme is formed s which is later deacylated by various means. Lowe and

-370-

Williams18 based their suggestion that the k2 step is subject to nucleophilic
catalysis combined with acid catalysis on the observations that (1) for substituted
aryl hippurate esters log (k^ ,/K ) gives a better fit to Q than to o*~ (p^l.2)

indicating relative insensitivity in k2 and (2) for BAA hydrolysis the data of

Smith and coworkers1 show k , ,TT ^/k ,,_ ~, to be 0.8,, which Lowe and Williams

cat(H20,r cat(D20j '

consider to be an indication of acid catalyzed ester hydrolysis . It is their view

that the alkaline limb of the k ./k vs. pH curve represents not thiols but

C3.X xn

imidazolium titration (for histidine pKa - 5.6-7.0 at 25°^ AH ~ 6. 9-7. 5 kcal/mole) 0
Their proposal^ then^ is that the thiol group is aeylated by proton abstraction from
a sulfhydryl group by a carboxylate ion with sulfur attack at the acyl moiety and
proton donation to the leaving group by imidazole . It should be noted that there
are only two histidine residues in papain^ at 106 and 175 • Further,, the direct
evidence for imidazole implication (it is bound by earboxymethylation with irrevers-
ible inhibition)2* has not been observed by other workers55*26 who have found that
only cysteine is bound by earboxymethylation.

Whitaker and Bender20 employed k , as a measure of k2 in the hydrolysis of

BAA. and reported k„ J\. 0 = 1°35<> They clearly expected a larger value in

accordance with general base catalysis 9 but this low value lends no support to such
catalysis. The authors suggest that the carboxylate ion acts as a base and the
thiol group as an acid which is aeylated in the k2 step.

A mechanism for the deacylation reaction offered by Smith1 involved carboxylate
attack on the thiol ester intermediate. Lowe and Williams22 tested this proposal
for nucleophic carboxylate catalysis on the ester by comparing intramolecularly
catalyzed hydrolysis rates and activation parameters of some appropriate models
shown in Figure II with the k . and activation parameters calculated from the

temperature dependency of k . (from four points) for methyl hippurate hydrolysis.

The rate constants for the hydrolysis of II and III are measured from the pH

S-fl-CH« ^S-C-CHsHHCCqHs /w S' 5

OH ™2

I II III

Figure II „ Models of the acyl enzyme

independent region of k , vs. pH. The best model of hippuryl papain (II) is

ODS

3 x 10°5 fold slower than the k . for the enzyme catalyzed hydrolysis. The other

models are also 10~5 fold slower. Activation parameter comparison for III shows
that the entropy of activation for the enzymic k , is 20 e.u./mole lower than the

model^ but the enthalpy of activation is nearly 7 kcal/mole lower for the enzymic
reaction then the model. The authors conclude that the large difference in rate
constants suggests that nucleophilic catalysis of deacylation by a carboxylate ion
is unlikely.

Lowe and Williams31 proposed that deacylation is catalyzed by nucleophilic
imidazole catalysis at the thiol ester followed by a rapid hydrolysis step. Kinetic
support for imidazole participation in deacylation has come only from the results of
Cohen and Petra23 on a~N-benzoyl<=L-eitrulline methyl ester s where a sigmoid dependence
of k3 upon pH with a pKa of 7 was calculated. The form of this dependence implies
the catalytic group is active in the protonated form.

The observation of large deuterium isotope effects by Whitaker and Bender g20

sat,

at.

-371-
2.V? for BAEE hydrolysis , and Brubaeher and Bender ,12*2S

"H20 ^T^O
k„ -Vk- 0 ss 3„35 for deacylation of trans -cinnamoyl papain, all support general

base catalysis on the thiol ester by a. carboxylate ion as the mechanism of the
deacylation step by implying proton transfer in the transition state of this step.
The complicity of a carboxylate ion to deacylation has been discussed above „ In a
study of the deacylation of trans -cinnamoyl papain by added amine nucleophiles i

Brubaeher and Bender13 observed that a plot of log k

amine eat0 deacylation

\ i. ,, ■." P

log k*) vs. pKa of the amine produced no correlation to any simple relationship. A
plot of the observed deacylation rate constant k = k3 + k^ (added nucleophile) vs.

the added nucleophile concentration gave a straight line with no evidence of binding
of the nucleophile. The authors , however, felt that certain comparisons of k& for
appropriate nucleophiles of simlliar pKa or structure necessitated that a specific
binding interaction exists, and that basicity has little , if any, influence upon
k^. The authors considered this supportive evidence for general base catalysis in
the deacylation of trans -cinnamoyl papain. They reasoned that the less basic the
nucleophile , the more readily it will, release its proton, but the more basic the
nucleophile, the better , so that these two effects would tend to cancel. In regard
to the general applicability of these observations, it must be noted that, both the
added nucleophile effects and the largest observed deacylation isotope effects
were observed in trans =cinnamoyl papain hydrolysis, where k3 (lim) - 3»68 x 10~3
sec"1, or 700 fold smaller than the k + for methyl hlppurate hydrolysis, and about

10,000 fold smaller than N~benzoylargininyl papain deacylation. Lake and Lowe27
have interpreted this as indicative of either the involvement of a different rate
determining step, or the employment of another mechanism In the hydrolysis of this
nonspecific trans -cinnamoyl papain. They have also found k^ rJ^-s^. 0 for p-

nitrophenyl hlppurate to be 1„75 which they have interpreted as a secondary isotope
effect.

A fourth deacylation pathway has been recently proposed by Lake and Lowe27 as
involving a slow conformational change from the acyl thiol enzyme (ES3) to a con-
former (ES!J), in comparison to the relatively rapid subsequent breakdown step
giving product. To test the pathway involved in general base catalysis (Scheme I)
against this fourth proposal ( Scheme II) , the authors studied the effect of added
methanol upon k . for the production of Px from p-nitrophenyl hlppurate, and also

for the production of hippuric acid (P^) from methyl hlppurate. The k . for

CQ.X

product formation, according to each scheme, is the following % for Px in Scheme I

Scheme I E + S

E + P2

E + P3

Scheme II E + S

ES!
+Pi

Jia

» ES3

E + P2

E + P.

(assuming k2 » k3 + k^ (Me OH)) k . s k3» + k4 (Me OH), in Scheme II (assuming k2 »

k3 and k3 « 14 + k5 (MeCH)) k

cat

k3£ for P2 in Scheme I (assuming k2 )> k:3 + k^

(Me OH)) l/k . ■ l/k3, and in Scheme II (assuming k3 « k| + k5 (MeCH)) l/k.

cat

(k2 + kaJ/kaks + (ks + k3)(k5 (MeOHj/ksksk^. It was observed for p~nitrophenal (Px)
that k , was independent of the methanol concentration up to 2M methanol, and

-372-
decreased slightly above this point. Following hippuric acid. production, .

was seen to be linearly related to the methanol concentration. These results are
in agreement with any pathway of the kinetic form of Scheme II with k3 <(<( k!4 + k5
(Me OH) , but do not provide a description of the k3 stepc Lake and Lowe state that
these observations provide strong support for the absence of general base catalysis
in the deacylation of hippuryl papain,,

SUMMARY

A large collection of experimental evidence has accumulated to specify certain
aspects of the pathway of papain catalyzed hydrolysis of synthetic substrates.
Although the evidence supporting an acyl thio-enzyme is significant, the detailed
mechanism of its formation and deacylation is subject to the uncertainty of con-
flicting observations and interpretations. Particularly evident, has been the
dependence of the observations and hence the proposed mechanism, upon the substrate
employed. It would appear that to make any generalized statement as to the nature
of papain catalyzed hydrolysis would at this point be premature due to a lack of
knowledge concerning possible variations or discontinuities in the mechanism with
a spectrum of substrates,

BIBLIOGRAPHY

1, E, Smith and J. Kimmel in "The Enzymes," Vol, 4, P, Boyer, H, Lardy, and K,
Myrback, Ed,, Academic Press Inc., New York, N.Y. , I960,

2, A, Wurtz and E, Bouchut, Compt, Rend, Acad, Sei„, 8g, 425 (l8?9)o

3, K, Hwang and A, Ivy, Ann, New York Acad, Sci,, j4, l,6l (1951).
k, J. Kimmel and E, Smith, Advances in Enzymol,, 1£, 267 (1957).

5„ E. Smith, R, Hill and J. Kimmel in "Symposium on Protein Structure," A,
Neuberger, Ed,, Methuen, London, 1958.

6. E„ Smith, A, Light, and J, Kimmel, Symp. Biochem, Soc, 21, 88 (1962).

7. A. Light, R, Prater, J. Kimmel, and E„ Smith, Proc. Natl, Acad. Sci. U.S.,
£2, 1276 (1964) .

8. M. Bender and F„ Kezdy, Ann, Rev. Biochem,, Vol. 34 (1965).
9» G„ Lowe and A. Williams, Proc. Chem. Soc., 140 (1964).

10. G. Lowe and A. Williams, Biochem. J., 96, 189 (I965),

11. A. Glazer and E. Smith, J. Biol. Chem., 2^5, I9V5 (I96I).

12. M. Bender and L, Brubacher, J. Am. Ghem. Soc,, 86, 5333 (1964),

13. L, Brubacher and M, Bender, J. Am. Ghem. Soc, B5, 5871 (1966).

14. M. Bender, et. al. , J. Am. Ghem. Soc., 88, 5890T1966) .

15. Jo Kirsch and E. Katchalski, Biochem,, ¥J 884 (1965)0

16. L„ Sluyterman, Biochim, Biophys. Acta, B^, 305 (1964).

17. A. Williams, Doctoral Thesis, Oxford University, 1964.

18. G, Lowe and A, Williams, Biochem, J,, 96, 199 (1965),
19o Jo Kirsch and M. Igelstrom, Biochem., 5, 7^3 (1966).

20. J, Whitaker and M. Bender, J. Am. Chem? Soc, 87, 2728 (1965).

21. M. Bender and L. Brubacher, J. Am. Chem. Soe.,~|8, 5880 (1966).

22. G, Lowe and. A, Williams, Biochem. J., 96, 194 (1965).

23. W. Cohen and P. Petra, Biochem., 6, 104f (1967).

24. S„ Yu~Kum and T. Chen~Lu, Sci. Sinica, 12, 1845 (1963).

25. B„ Finkle and E. Smith, J. Biol. Chem., 2J0, 669 (1958).

26. A, Light, Biochem. Biophys. Res. Common., if, 781 (1964).

27. A. Lake and G„ Lowe, Biochem, J., 101, 402 (I966).

-373=

THE PHOTOSENSITIZED CIS-TRANS ISOMERIZATIOli OF OLEFINS

Reported by Robert Kalish May 22 , 196'

The light induced c is ° trans isomerization of olefins has long been known, and
has been widely used as a synthetic tool;, especially for the preparation of the
less stable member of an isomeric pair of olefins., It has only been within recent
years, however, that the mechanisms of these isomerizations, particularly those
occurring under triplet photosensitized (hereafter referred to as photosensitized)
conditions , rave been intensively studied. This seminar will be specifically
concerned with mechanistic aspects of the photosensitized cis-trans isomerization
of olefins occurring in solution, the area of vapor phase photosensitized iso-
merizations having recently been reviewed.1 The utility of such studies in en-
hancing understanding of photochemical processes in general, and triplet energy
transfer2 in particular, will also be emphasized .

NON-CONJUGATED OLEFINS

Theoretical calculations1*3 indicate the most stable configuration of the
lowest triplet state of ethylene and other non-conjugated olefins to be the
twisted, non-spectroscopic form (the so-called phantom triplet,9 p) in which the
planes defined by the two GHp groups are orthogonal thus minimizing interaction
between the two 2Cp orbitals and electrons.

Such a twisted triplet state provides a ready pathway for the cis -trans
isomerization of simple olefins, as formation of this triplet from either a cis
or trans olefin is expected to result in decay (via twisting and inter system
crossing) to both the cis and trans isomers. The existence of a twisted triplet
as a common intermediate in these isomerizations is based both on theoretical
predictions, ln3i"4 and on the finding that the sum of the quantum yields for the
cis to trans (<t>+ ) and trans to cjLs (<J> ) photosensitized isomerization of many

olefins, in which the olefin triplet is formed with unit quantum efficiency, is
one.5 Calculations** indicate that 4> + $ equals one only if isomerization

occurs solely from a common twisted triplet. If. however, isomerization occurs
from two non-interconvertible spectroscopic triplets, 0\$ , + <1> <^2. The fact

that 4> + <t> has been found to equal but never to exceed one is strong pre-
sumptive evidence for a common-intermediate process.

Two types of triplet energy transfer processes., vertical and non-vertical,
can, in principle, occur via a coupled energy transfer process2 of the form shown

ml C<)

in eq 1, S being the triplet energy donor ( sensitizer) , and A being the
(olefin) acceptor. Vertical (classical) energy transfer occurs in accordance with

ml qO qO ml

(1) S~ + Ab— ► SS + AT

the Franck~Condon principle,7 producing vibrationally excited, spectroscopic

olefin triplets which then relax vibrationally (via twisting) to the twisted

triplet with subsequent decay to the isomeric ground state olefins. Non-vertical

(non-classical) energy transfer4,5'3'99 directly to the twisted triplet, although

forbidden by the Franck-Condon principle for radiative processes, can occur in

a coupled triplet energy transfer process. Vertical and non-vertical energy

transfer are competitive processes, the less probable and hence less efficient

non-vertical process being unimportant relative to the vertical process when the

S
triplet energy of the sensitizer, E (measured by the 0»Q band) exceeds the

triplet energy of the acceptor, E2, by ^3 or mere kcal/mole. In such a case

vertical energy transfer is diffusion-controlled,2 occurring on every collision.

When E is less than E2 the classical energy transfer process is endothermic.

Non-vertical energy transfer can now effectively camp &th the vertical process
which, although intrinsically more efficient, has a higher activation energy
than the non-vertical process,, When energy transfer is endothermic and hence
not diffusion- controlled, the time required for this process to occur has been
calculated to be at least 10 8 sec It is thus not surprising that, the lower
energy non~ vertical transitions can occur, as the time required for radiative
transitions j, for which the Franck-Condon principle was formulated, is about 10~15
sec Little is known about the detailed nature of non-vertical processes , however,
which will be discussed further in connection with the c is -trans isomerization
of the stilbeneso

The acetone sensitized cis^trans isomerization of 2-pentene has been studied
in detail in a variety of solvents by Borkman and Kearnso10 These workers found
that for concentrations of 2-pentene greater than 1„0 M, <t> . + $ . equalled loO

+ Ool in accord with isomerization via a common 9 twisted triplet intermediate.
The quantum yield for energy transfer from acetone to 2-pentene was determined
to be loO + Ool indicating that every acetone triplet eventually transfers its
excitation energy to a 2-pentene molecule | this finding implies nothing about the
efficiency of energy transfer during a single collision, however,,

It was found that although 2-pentene completely quenches the phosphorescence
of acetone, it does not affect the fluorescence of acetone under the same conditions
in which the acetone sensitized 2-pentene isomerization occurs with 100$ efficiency,,
Indicating that energy transfer from acetone to 2-pentene is not vibrational in
nature and proceeds from the triplet state of acetone (ivlCf 6 sec) rather than
from the shorter-lived singlet state (t^2«5 X 10"d sec) 9 a fact which has long been
assumed in ketone photosensitized olefin isomerizations, but never before rigor-
ously proven. Although attempts to detect 2-pentene triplets spectroscopically
(e,g0 „ by esr in rigid glasses at 77° K) under conditions of the isomerization
were unsuccessful, perhaps owing to a short triplet lifetime, it is quite likely
that,, in accord with the Wigner spin conservation rule, it is the 2-pentene trip-
let which Is formed upon energy transfer „

From a study of the initial rate of the isomerization as a function of the e©n-
centration of 2-pentene,, Borkman and Kearns were able to calculate the quenching
constant^, K , for the triplet energy transfer proce.s tere K. =^r_k. „ T being

the acetone triplet lifetime in solution and k being the overall bimolecular

rate constant for energy transfer from the acetone triplet to 2-pentene,, From

the experimentally determined value of K , k was calculated to be ^lO7 m"1 sec"1,

a value about 103 times less than the theoretically predicted diffusion-controlled one,

This finding led Borkman and Kearns to conclude that the energy transfer step,

known to involve close contact between sensitizer triplet and acceptor,2 is best

T1 $r
written as two distinct steps (eq 2), [S * « olefin ] being a collision complex

£l cD & ml gO k O ml

(2) S + olefin0 f=±=±[SL ..«, olefin ] > S* + olefin

in which 2-pentene is adjacent to the acetone triplet but in which energy transfer
has not yet occurred $ this collision complex can revert back to 2-pentene without
energy transfer (k, ) or undergo dissociation with concurrent energy transfer (k

Kinetic analysis of eq 2 Indicates that k,=k k /(k + k, )» If k , the nearest
tf * t a c c d c

neighbor rate of energy transfer, is large compared with k, then k. -k , i.e.,

d t a

the process is diffusion-controlled,, If k <C<(k, then k, -k k /k, and k. is less

than diffusion-controlled as found for 2-pentene « Furthermore if it Is assumed
that the collision complex has negligible stability, l<,e«, 3 £H° of formation *^-0_9

-375=

then k /k should be temperature independent indicating that k will he a function
a d '

of temperature only if k is0 Variable -temperature kinetic studi.es in the 25°

to =78° range did indicate a temperature dependence in k. (k ) from which an

™ t c

activation energy of ^ ka 3 kcal/mole was calculated for the energy transfer stepD

The E of acetone has been estimated from phosphorescence studies to be
^v^80 kcal/mole10 (a value disputed by Cundall11 who claims a value of ^75 kcal/
mole) , whereas the 0-0 triplet energy of 2-pentene has been taken by Kearns to
be equal to that of ethylene , 4 /v 82 kcal/mole (a value which may be slightly
too high1^ o Triplet energy transfer from acetone to 2-pentene is thus predicted
to be endothermic by ™ 2 kcal/mole v in reasonably good agreement with the exper-
imentally determined activation energy, thus accounting for the observed rate of
energy transfer „

Morrison, et„ al„ , have used the photosensitized cis-trang isomerization
of olefins as a tool to investigate intramolecular triplet energy transfer,12
i0eo y light absorbed in one part of a molecule results in a chemical reaction
at another, non-conjugated part of the same molecule via energy transfer from the
initially excited chromophore to the reacting center »

The most conclusive evidence for intramolecular triplet energy transfer comes
from a study of the irradiation of trans- and cj^-l-phenyl~2~huter.ie,12 " light
of 230-280,^1 being used to insure that only the phenyl chromophore is excited
to the jc^rt triplet state; the only observable reaction was cis-trang isomer-
ization with <&n4.=$, =0o21 + O0OI60 The mechanism suggested by Morrison for this

photoisomerization is given in eqs 3=60 12 s

It is assumed that the phenyl donor

S° hv

„0 &3 -K-4 rO

(6) tfa « p — VcD

and double bond acceptor chromophore s are non-interacting in the ground and sing-
let excited states, and also that the rates of intersystem crossing (k_._) and rad-

iationless decay (kj of the phenyl donor are the same for either isomer; kx

and k2 are the rates of energy transfer from the phenyl triplet to the cis and trans

olefinic linkage, respectively,, giving a common, twisted triplet, p0 Since E_6 5"

( ^83 kcal/mole) is at least 2«3 kcal/mole greater than the 0-0 triplet excitation
energy of either isomer, ki=k2 and energy transfer probably occurs at close to
the diffusion-controlled rate to give vibrationally excited, spectroscopic
olefinic triplets which decay to p which undergoes subsequent decay with concom-
itant cis-trans isomerization (k3.k4)0

Kinetic analysis of eqs 3^6 leads to the prediction that the photo stationary
state composition, ( cis) _/trans} , is equal to (k2/kj)^fk4/k3) „ Since kx=k2,

(SjLOL) AiSH£§J "^4/^3,? ii£ej "k^e photo stationary state composition is determined

by the decay ratio of p, as. intrinsic property of the excited state in a given
solvent at a given temperature.,16 Moreover, this expression predicts that ex-
citation of the double bond by intermolecuiar energy transfer from benzene should
lead to the same (cis) ./(trans) as the intramolecular process; in accordance

with this prediction, (cis) /(trans) was found to equal 1„0 for both types of

sensitization processes.

Evidence that the observed intramolecular energy transfer is electronic and
not vibrational., i.e, ., from a vibrationally excited ground state of the phenyl
group leading to thermal cis -trans isomerlzatloc.,, comes from the observation that
the quantum yield for the trans to cis isomerization in cyclopentane is reduced
from 0o21 to 0.15 in the presence of an equimolar amount of the triplet quencher
trans °2°hexene , indirectly indicating that intramolecular transfer of electronic
energy does occur o

Cis-trans isomerizations have also been ob served, to occur upon irradiation
of trans-4°hexen"2"One12 and tranS"5~hepten=2"Oneo12 Formation of the cis
isomer upon excitation of the carbonyl group of trans =S-hepten-2~one was taken
by Morrison as fairly good evidence for intramolecular' energy transfer from the
carbonyl triplet to the double bond, although the possibility of vibrational
energy transfer was not ruled out. The situation with regard to the isomerization
of trans~4"hexen=2"One is unclear,, products arising from acetyl and 2-butenyl
radicals being observed.

Recent studies of the benzene,, toluene and xylene sensitized irradiation
of a series of l~alkylcycloalkenes have indicated the possible formation of the
highly strained trans isomers,.13 Irradiation of (+)-3~carene (l) was found to
give (+)"3(10)=-carene (2) | in the presence of methanol, ethers jj and 4 were also
formed (eq ?) • Deuterium labeling studies indicated the rearrangement to the

h*v v
"7"

CH30H, xylene

-OCH3
(12#)

exocyciic olefin to be intermolecular with respect to the proton shift «

These observations of proton incorporation,, ether formation^ rearrangement
and Markovnikov addition are suggestive of ionic processes s and led Kropp to
suggest that decay of the twisted cycloalkene triplet^ formed via triplet energy
transfer o leads to both cis and trans olefin, the highly strained trans isomer
undergoing protonation with relief of strain to give a carbonium ion which leads
to the observed products,,13 The presence of olefin triplets in these reactions
was suggested by the finding, of Carroll and Marshall that the reaction rate is
decreased by added oxygen.. 'These workers propose a mechanism involving direct
protonation of the olefin triplet to give a carbonium ion and the observed products »
Kropp believes this mechanism to be less likely in view of the finding that
exocyciic and acyclic olefins 3 the trans isomers of which are not highly strained,,
undergo neither photoinduced double bond migration nor ether formation,,

Other examples of photosensitized cis^trans isomerizations of simple olefins
include the isomerization of cis,, trans » trar;.s-l55i>9=-yclcdodecatriene to the cis,
cis, trans and trans v trans, trans isomer s/^and the isomerization of methyl
oleate to methyl elaidateT1*'

CONJUGATED OLEFINS

The photosensitized cis°trans isomerization of the stilbenes and 1^,2-diphenyl-
propenes in benzene solution has been studied in considerable detail sle"2°a:m. will be
discussed with particular reference to the stilbenes 9 the results obtained for
the lP2=diphenylpropenes being similar except when otherwise notedo

The mechanism represented by eqs 8=12 will serve as a starting point for
discussion of the isomerizationo In this mechanism only vertical energy transfer

(8) S^_i^ssli^s^

-377-
(9) S1" + £> _J^ 4J* + s^ (10) sl1,/^/^^

(11) cTl » tTl (22) (a) t^-fJES- tTl -^>c^ (b)

T1.

to cis- and trans- stilbene yielding the planar, spectroscopic transoid (t ) and

ml jl rpl

cisoid (c ) triplets is proposed?, conversion of £ to t (eq H) has been

T1
shown to occur as will be seen shortly » Eq 12b, decay of t to cis-stilbene,

is a non-vertical, radiationless decay process9 completely analogous to the

non-vertical excitation processes previously discussed^ its importance will be

assayed later. The light used was filtered to avoid direct excitation of either

of the stilbene isomers, and no indication of ojay singlet state reaction of the

stilbenes (e0g0 , phenanthrene formation) was foundo16 Furthermore, for high-energy

sensitizers, i0ec , e3 vertical excitation energy of either stilbene isomer by

1 iTll

at least 3-5 kcal/mole, radiationless decay of S was found to be negligible ,

deactivation of S occurring only via energy transfer to the stilbenes0

Kinetic analysis of eqs 8-12 leads to the expression for the photostationary
state composition given by eg 1J. For high-energy sensitizers ka, and k2 will be

( 13) ( cis) «/( trans) „ = ( kx/ks) ° ( k4/k3)

equal and diffusion-controlled, and the photostationary state composition will
be constant, being determined solely by the decay ratio, k4/k3o Under such con-
ditions the quantum yields are related to (cis) J{ trans) by eq lk. The 0-0 triplet

s s

{lk) Kj*n+ = k4As s (£is) J{ trans)

2ia,b

tc' ct * ° v-— ' s' *zses=/ s
excitation energies of the stilbenes are^v5Y kcal/mole for the cis isomer

and/^50 kcal/mole for the trans isomer,.21 As E_ approaches and falls below

57 kcal/mole, kg is predicted to decrease below the diffusion-controlled limit,

ki remaining constant ., thus leading to an increasingly cis-rich photostationary

state0 As E approaches and falls below £Q kcal/mole k± should also decrease,,

Below 50 kcal/mole the excitation ratio ki/k^ for the endothermic energy transfer
process is predicted to become constant, being determined solely by the difference

c t

in the 0-0 triplet energies of the two isomers, i.a e, 9 kx/ka = expt [ ( E--E^)

^v6 X 106 at 25°. !Ehis predicts a limiting photostationary state of essentially
pure cis- stilbene.

The experimentally determined variation of the photostationary state com-
position,8'16 extrapolated to infinite dilution with respect to the sensitizer

g
concentration, with E for the stilbenes, fig 1, does not completely follow these

predictions, however. The predicted high-energy region of constant photostationary

state and subsequent increase in $ cis- stilbene as E decreases below ™62

kcal/mole are observed^ the other
predictions are incorrect . With

(cis) q £ f ^ low-energy sensitizers (Err<(^62 kcal/

(trans) Q | 6y '^-v- , mole) the value of k2 (cis- stilbene

hfu^1 * '4o ' 'r1-' "fcfT" fr ' ' '4^' ' '^q" exci'fca'fcion) falls off much more
+} 5 J 55 ou op (L, o slovlj than predicted, a finding

E (kcal/mole) which led Hammond to propose that

fig 1 triplet energy transfer to cis- stilbene

occurs via non-vertical excitation

-378=

with synchronous distortion to give the twisted phantom triplet (p) and/or t

processes requiring less energy than transition to c ; since trans- stilbene is
more stable than eis-st liberie by»6 kcal/mole,22 non- vertical excitation of the

cis isomer to p or t should only become endothermic when E <(^W- kcal/mole in

agreement with variable -temperature kinetic studies26 which indicated no activation
energy for this process in the 53-60 kcal/mole range. Inclusion of eq 15 in the
kinetic scheme leads to the photo stationary state composition given by eq 16.
For high-energy sensitizers kx ~ k/s^fes and eq 16 reduces to eq 13. Below ^62

(15) ST + c — ^->tT ,p + SS (16) (cis) /(trans) - kxk4/(k2 + k5)k3
_ s "™ s

kcal/mole k2 decreases, kj. and k3 remaining constant in accord with the observed

increase in (cis) /(trans) . Below/v 51-52 kcal/mole, kx also decreases sharply,

,:_L " S ' £»

k3 remaining essentially constant?, ki/(k2 + k5) thus decreases explaining the general
shape of fig 1 in the ^5-50 kcal/mole region.

It is observed from fig 1 that the low-energy sensitizers eosin (j>) and 9j
10-dibromoanthracene (6) establish a photo stationary state of essentially pure
trans-stilbene „ recent studies18 have shown that photolysis of these sensitizers
produces bromine atoms which cause thermal equil.ibiatn :-r. of the stilbenes.1

Sensitizers of E <( 53 kcal/mole were found to give pronounced concentration

as the concentration of the sensitizer increased?, this finding was attributed to
reversible energy transfer to trans -stilbene (eq 17; »16 Replacement of eq 9 by
eq 17 in the kinetic scheme leads to eq 18 which is fit by the experimental data.

(17) t , + ST rjp-*- tT-t + S (18) (cis) J^fcrans) ^k1k4/(k2+k5)(k3+k„1[S I)

T1
In further accord with this idea of reversible energy transfer to t , it was

found that inclusion of the triplet quencher azulene, az (£^29=^2 kcal/mole)

in the reaction mixture also gave photostatlonary states richer in trans -stilbene
indicating eq 19 to be operative . The effect of added azuiene was the same for

all sensitizers usedo Further evidence for quenching of t comes from examination
of the <t> /<t> , ratio obtained for high-energy sensitizers . While eq Ik was found

"CC CX>

to hold for the 1,2-diphenylpropene isomerization, it was not satisfied by the

stilbenesc This deviation from eq Ik was found to be due, at least in part, to the

T1
existence of self-quenching of t (eq 20). 16

>pl c<Q cO ml rpl c<0 qO

(19) t + az — — ^ t + az (20) t + t ->» 2t

These quenching studies indicate that energy transfer to cis- or trans-
stilbene leads to the ultimate production of the same triplet species which is then
quenched to give trans-stilbene ,, i. e. , energy transfer to cjLs-stilbene produces
a triplet whichican be deactivated to tr ans " stilbene by quenching. However,
quenching of £ could not be detected. To explain these findings, Hammond

ml rpl

postulated the rapid, irreversible decay of £ to t (eq 11); this interconversion

rpl

probably occurs via p., believed to be in equilibrium with t ( eq 21) . These

rpl rpl

(21) £ _ — ^ p ^=± t

findings are also evidence for the production of an electronically excited triplet( s)
upon sensitization of cis-stilbene by low-energy sensitizers indicating that
non-vertical energy transfer does not involve transfer of vibrational energy.
Quenching effects were not observed for the 1,2-diphenylpropene system, indicating
that conversion of both the transoid and cisoid triplets to p is rapid and irreversible,
isomerization occurring only from p.ls

Becent studies by Hammond and Herkstroeter have provided direct evidence as to the

79-

nature of the energy transfer processes. 1:y These workers used flash spectroscopy
to study the rate of decay of various sensitizer triplets using cis- and trans-
stilbene and 1,2-diphenyipropene as quenchers. Attempted direct study of the
stilbene and 1,2-diphenylpropene triplets was unsuccessful due to their short
lifetimes. From these quenching studies the rates of energy transfer, L, ^ from

the sensitizer triplets to the olefin acceptors were determined.. The results
obtained confirm what has already been said about the nature of the energy transfer

processes. Classically,, if E is less than E2, energy transfer will require an

activation energy equal to E^-E , the decrease in transfer efficiency as a function

of E being given by eq 22.

(22) d(iog k^/dCE^ - f 1/2.303 FT

The experimentally observed quenching curve for trans- stilbene had the slope

indicated by eq 22 for sensitizers of E ^_48 kcal/mole, indicating that non- vertical

excitation of trans-stilbene to p does not occur. The behavior of cis- stilbene towards

sensitizers of E_ ~ 42-5$ kcal/mole did not fit eq 22 s excitation of cis-stilbene

being quite efficient even for sensitizers of Em 10 kcal/mole too low to effect

vertical excitation, in agreement with the postulated non-vertical process. As
also inferred from photostationary state studies, neither cis- nor trans-1,2"
diphenylju rpene was found to exhibit classical behavior as a triplet quencher
indicating that both of these isomers undergo efficient non-vertical excitation.

From these studies it was concluded that trans- stilbene can find no excitation
pathway of substantially lower energy requirement than vertical excitation to

ml ml ml

t , indicating p to be close to isoenergetic with t but below £ in energy.
Since neither of the 1,2-diphenylpropenes behaves classically, both the cis and
trans isomers are apparently able to undergo transitions to one or more twisted
states of lower energy than either the planar cisoid or transoid triplets, This
is probably a consequence of the relief (via twisting) of the steric strain that
exists in both the cisoid and transoid triplets owing to mnbonded interaction
between a phenyl group and, the phenyl or methyl group cis to it.

The results of the quenching studies as well as the observed non-vertical
energy transfer to cis -stilbene may be explained either by the assumption that the

rpl

two triplets, t and p, are in equilibrium (eq 21) 9 or that only a single triplet ,

ml ml

t , exists., non- vertical excitation of cis- stilbene giving only t . The
latter hypothesis, although, not rigorously disprove^ appears to be much less
likely. In order to accommodate this hypothesis it must be assumed that whereas

t is selectively deactivated to trans-stilbene in quenching reactions, spontaneous

decay of t yields both cds- and trans-stilbene » Furthermore, inclusion of p

in the isomerization mechanism allows a rationalization of the very short lifetime

of the stilbene triplets, estimated from the azulene quenching studies to be

<£__ To? X 10 8 sec. If, as seems reasonable, p is assumed to have a twisted

configuration with an angle of twists it/2, it may be very close, both in energy

and configuration, to a point on the potential surface of the ground singlet state

(fig 2)| intersystem crossing from the triplet to the ground state is expected to

be very rapid under such conditions. Saltiel believes that there may be an actual

crossing of the singlet and triplet states in this region as shown in fig 2^2° '

the energy of the twisted ground state has been estimated from thermal isomerization

studies to be~^9 kcal/mole above that of trans-stilbene . 23

Recent work by Saltiel has provided additional evidence as to the nature of p
and the decay processes involved in the r^llbene isomerization.20 The photostationary
state composition obtained from isomerization of trans- stilbene-d.12 was found to equal
that obtained from undeuterated trans-stilbene for sensitizers of E^j = 48-69 kcal/mole 5

Energy
(kcal/mole)

T1 triplet excited Tf

.580=

angle of twist

Stilbene Energy Profile

fig 2

the quantum yields of the two
isomeriz?.tions were also identical,,
Deuteration is known to decrease the
rate of triplet to singlet radiationless
decay,, the effect diminishing as the energy
separation between the two states decreas-
es.24 If p is of lover energy than t f
deuteration is predicted to affect decay
of tTl to trans - stilbene more than decay of
p to a twisted ground state 9 assuming decay
from both tT1 and p to be operative. The
absence of such a deuterium effect led

Saltiel to conclude that decay from t is negligible for both trans -stilbene and trans-
stilbene -di2« He believes p to be of lower energy than tTl as shown in fig 2 with vir-
tually all decay to the ground state occurring from p^ a point of view not fully shared
by Hammond.18 Although decay to the ground state from txl(eq 12) has not been rigorous-
ly ruled out;, decay of the very short-lived cisoid triplet to ground state stilbene is
not believed to occur ^ internal conversion to p and/or t^ being faster. If formation
of cis-stilbene is assumed to occur mainly from p^ then the earlier finding that the
photo stationary state becomes richer in cis-stilbene at higher temperatures can be at-
tributed to the existence of an activation energy in the interconversion between p and
tT1 as shown in fig 2.

Photosensitized isomerization of the stilbenes by the low-energy sensitizer phenan-
thraquinone (PAQ, E^ = 48.8 kcal/mole) has recently been studied in benzene solution by
Bohning and Weiss 02* Formation of adduct J occurred competitively with isomerization.

(n %n The kinetic results obtained are in substantial agreement with the mech-
-\Y^ "^f^CsHs anism already discussed for the isomerization, the only major difference
•>lks ^T'CgHs being the inclusion of a short-lived complex^ X_, formed via triplet ener-
\ gy transfer to either cis- or trans -stilbene , decay of X being partit-

X ioned between collapse to 2 and decay to p and PAQ. The existence of X
as a common intermediate is strongly suggested by the observation that both cis- and
trans-stilbene give the same adduct. The geometrical changes leading to p are postula-
ted to occur in X^ in which there is believed to be freedom of torsional motion in the
stilbene accounting for the formation of 7 from both the cis and trans isomers. Clas-
sical energy transfer to trjns~stiibene(but not to eis-stilbsne) was also invoked j for
in its absence the quantum yields of adduct formation become independent of (cis) s/
(transj s ., whereas such a dependence was observed.

I^ie cis-trans isomerization of stilbene has been used to examine the steric re-
quirements of triplet energy transfer26 Theory predicts that this process should be
subject to steric hindrance by bulky substituents on the donor or acceptor.2 In agree-
ment with this idea., it -was found that whereas the high-energy sensitizers 2^3i-5;6-tet-
ramethyl-M -methoxybenzophenone (E§ - 70.2 kcal/mole) and 2^„6-trimethyl»4t-methoxy-
benzophenone (E^ = 68. k kcal/mole) produced the same photo stationary state as benzopl:
one (E§ = 68.5 kcal/mole) ^ indicating diffusion-controlled energy transfer, 2^6-
triisopropyl-^'-methoxybenzophenone (E§ = 69.9 kcal/mole) and 2^6-triisopropylhenzo-
phe"none (E§ = 68.7 kcal/mole) , compounds which are more hindered about the carbonyl
group where the triplet energy is believed to be localized^2 gave cis rich photosta-
tionary states indicating that energy transfer to cis-stilbene is less efficient than
to trans-stilbene. To insure that the observed results were not due to selective energy
transfer from the low-energy triplets of the photoenols of structure 8 formed from the

sensitizer triplets, the rates of quenching of the photoenolization
of the triplets of 2^4s6-trimethyl-4J-methoxybenzophenone and 2 .'■>- 6«
triisopropyl-M-methoxybenzophenone by added cis- and trans-stil-
bene were studied! the rate of decrease of photoenolization (and
hence rate of triplet energy transfer) of 2^6-trimethyl-ii-'1-meth-
oxybenzophenone was about 15 times as great as that of 2,4,6=
triiscpropyl-M-methoxybenzophenone^ trans - stilbene being the
better quencher in each case, confirming the existence of a steric effect to energy
transfer o

The photosensitized cis-trans isomerization of the piperylenes is mechanistically

>phen=

-38l-

similar to that of the stilbenes and has been reviewed by Turro,27 and recently studied
on a silica gel-benzene matrix28 and under ferrocene photosensitized conditions.29 In
addition, the piperylene isomerization has found many uses in photochemistry including
measurement of the inter system crossing quantum yields of sensitizers whose triplet
states can effect the isomerization,5 and measurement of the triplet energies of sensi-
tizers for which spectral data is not available.16-927 Owing to their relatively low
triplet energies j, both cis- and trans -piperylene (E^ = 56.9 and 58.8 kcal/mole, respec-
tively) have been extensively used as triplet quenchers in mechanistic studies.

The photosensitized conversion of cis „cis-l , 3-cyclooctadiene to bicyclo[4.2.0]~
oct-7-ene has been found by Liu30 to proceed via the isolable cis,trans-l,3"-cycloocta-
diene which is then thermally converted into bicyclo[4.2.QJ-oct-7-ene in accord with
the Woodward-Hoffmann rules o

BIBLIOGRAPHY

lo Ro Bo Cundall, Progr. Reaction Kinetics , 2, 165 (196*0.

2. For a review, see N. J. Turro, "Molecular Photochemistry „" W. A. Benjamin,
Inc, New York, N. Y. , I965.? chapter 5°

3c R. S. Mulliken and C. C. J. Roothaan, Chem. Rev. , 4l, 219 (19^7) •

4. D. F. Evans, J. Chem. Soc, 1735 (1960).

5. (a) A. A. Lamola and G. S. Hammond., J. Chem. Fhys., 4j5, 2129 (1965)5 (b) M. A.
Golub, et al. , J. Chem. Phys., 4£, 1503 (1966).

6. Z. R. Grabowski and A. Bylina^ Trans. Faraday Soc, 60, 1131 (1964).

7. Ref 2, pp 30-42.

8. G. S. Hammond, Kiagaku to Kogyo ( Tokyo) , 18, 1464 (1965).

9. (a) G. S. Hammond and J. Saltiel, J. Am. Chem. Soc, 8j>, 25l6 (1963)5 (b) ref 2,
pp 182-3.

10. R. F. Borkman and D. R. Kearns, J. Am. Chem. Soc. , 88, 3467 (1966).

11. R. B. Cundall and A. S. Davies, Proc. Roy. Soc. (London), A290, 563 (1966).

12. (a) H. Morrison, Tetrahedron Letters,, 3653 (1964)5 (b) H. Morrison, J. Am.
Chem. Soc, 8j, 932 (1965)5 (c) H. Morrison,, et al. , Abstracts, 153rd
National Meeting of the American Chemical Society, Miami Beach, Fla. , I967.9
p 01375 (d) H. Morrison, private communication.

13. (a) P. J. Kropp, J. Am. Chem. Soc, 88, 4091 (1966); (b) J. A. Marshall and
R. D. Carroll, ibid. , 88, 4092 ( 196617

14. H. Nozaki, et al., Tetrahedron Letters, 2l6l (1965).

15. A. C. Testa, J. Org. Chem. , 2£, 246l (1964).

16. G. S. Hammond, et al. . J. Am. Chem. Soc, 86, 3197 (1964).
17- S. Malkin and E. Fischer, J. Phys. Chem., fsB, 1153 (1964).

18. G. S. Hammond, private communication.

19. W. G. Herkstroeter and G. S. Hammond. J. Am. Chem. Soc, 88, 4769 (I966).

20. (a) J. Saltiel, ibid8 8£, IO36 (1967)5 (b) private communication.

21. (a) D. F. Evans, J. Chem. Soc, 1351 (1957) J (b) R. H. Dyck and D. S. McClure,
J. Chem. Phys., j>6, 2326 (1962).

22. R. B. Williams, J. Am. Chem. Soc, 64, 1395 (.1942).

23. G. B. Kastiakowsky and W. R. Smith, ibid„ ^6, 638 (1934).

24. Ref 2, pp 69-70.

25. J. J. Bohning and K. Weiss, J. Am. Chem. Soc, 88, 2893 (1966).

26. G. S. Hammond, et al., J. Am. Chem, Soc, 88, 4777 (1966).

27. Ref 2, pp 178-81.

28. P. A. Leermakers, et al. , J. Am. Chem. Soc, 88, 3176 (1966).

29. J. J. Dannenberg and J. H. Richards, ibid, 87, 1626 (1965).

30. R. S. H. Liu, ibid, 89, 112 (1967).

xfto.

Reported by James E.
INTRODUCTION

THE ABNORMAL CLAISEN REARRANGEMENT1
Shaw

May 25, 196?

In I.936 Lauer and Filbert reported that rearrangement of y-ethylallyi phenyl
ether (I) in N,N-diethyls.niline at 201-225°C did not result in the normal Claisen
rearrangement product, £-(a-ethy!aIlyl) phenol (III), but instead gave £- (a ,7 -dimethyl -
allyl) phenol (II)

.1-3

0~CH2CH~CHCH2CH3
i- a 3 y b e

CH3

^CHCH-CHCH^
a 3 y

CHCHsCH2

III

The abnormal product II appears to arise from attachment of the p-or 5 -carbon of the
allyl chain to the ortho carbon of the benzene ring. The normal Claisen rearrange-
ment involves 7 -•attachment, Lauer and Filbert found that ozonolysis of their
product gave small amounts of formaldehyde in addition to the expected acetaldehyde j,
however, they failed to attribute the formaldehyde as possibly arising from the
normal product III. It was later shown by Hurd and Pollack4 that under the same
conditions both the normal and abnormal products are formed in a 1.3 : 1 ratio.
These workers also shewed that the aliphatic analogue, 7-ethylallyl vinyl ether
(IV), rearranged to give some abnormal product V in addition to the major normal
Claisen rearrangement product VI.

CH2=CH-0-CH2CH*=CHCH2CH3
IV

??o0r P3

-> CH3CH=CHGHCHaCHO

sealed,
be

+ CH*

CH2CH3
=CHCHCH2CHO

VI : 9W

Several other examples of the abnormal Claisen rearrangement are given in the
literature,5"'10 In all of these cases the allyl phenyl ether which undergoes
rearrangement contains a 7 -substituted methyl, ethyl, or n-propyl group. No
abnormal Claisen rearrangements have been reported for a- or 3-alkyl substituted
7-aryl substituted allyl phenyl ethers.11'12 It appears that 7 -secondary or 7-
tertiary alkyl substituted allyl phenyl ethers have not been investigated , The

:- of this seminar will be to examine the mechanism of the abnormal Claisen
rearrangement in both aromatic and aliphatic systems,

MECHANISM

For the rearrangement of 7-ethylallyl phenyl ether Hurd and Pollack proposed
the following cyclic mechanism which involves attachment of the 5-carbon of the cis
configuration of the allyl chain to the ortho carbon of the benzene ring.4

CH3\_

JIT

?H3

However, the abnormal rearrangements of ethyl £-(7-n-propylallyloxy)benzoate (VII)
and 7,7-dimethylallyl estrone ether (VIII) have shown that this mechanism is

, and that the abnormal attachment of the allyl

group.5'10 For compound VII the abnorma oduct DC could result from attachment

he p- or e-carbon. If the abnormal rearrangements of both the 7 -propyl and
7~eT Lyl phenyl ethers a oceeding by a single mechanism, 3 -attachment must

OCH2CH=CHCH2CH2CH3

213-24l°C

HCH=CHCH2CH3

-0 mm.

COCX^sHs
VII

be involved. The abnormal rearrangement of ether VIII would result in product XI
if ^-attachment occurred, and product XII is there was 6 -attachment.

CH2CH2CH3
' HCH=CHp

VIII

XII

Since the observed product was XI , it appears that the abnormal rearrangement
involves attachment of the p -carbon of the allyl chain to the benzene ring.

Lauer and co-workers investigated the possibility that the rearrangement of
the crotyl phenyl ether XIII, the simplest of the /-substituted allyl phenyl ethers,
was proceeding by the abnormal path involving (3 -attachment of the allyl group rather
than by the normal route involving /-attachment. 13 In both cases the product would
be the same. However, if the /-substituted methyl group were tagged with 14C, the
normal rearrangement would give XIV while the abnormal rearrangement should give XV.

OCH2CH=CHCH3

220-235°C
sealed
tube

CH3

CHCH«CH2

OOCaHs
XIII

C00C2H5
XIV

CH3 %.

£hch=ch;

Ozonolysis of the product, followed by a study of the radioactivity of the
formaldehyde produced, showed that 15-29$ of the product had been produced by the
abnormal rearrangement process. To account for the abnormal rearrangement, Lauer
proposed the following mechanism which involves (3 -attachment of the cis configuration
of the allyl chain.

CH2 CHp
lCH=^CH

CH3 *
CHCH=CH;

COOC^

COOC^s

However, in Lauer' s work it was not shown whether ether XIII was cis or trans . Schmid
and co-workers14'15 studied the rearrangement of the cis isomer of /-14C-methylallyl
P_-tolyl ether (XVI). Heating the ether for three hours at 230°C in N,N-diethyl-
aniline gave a rearrangement product, 2-(a-methylallyl) -4-methylphenol, which showed
by examination of the formaldehyde produced upon ozonolysis, that kCffo of the
reaction was apparently proceeding by the abnormal route in comparison to Lauer1 s
15-29$.

-38*i—

CH2^

CH3

XVI

Marvell and co-workers16 were the first to show that the abnormal Claisen
product was not formed directly from the ally! phenyl ether, but instead was due to
further rearrangement of the initially formed normal Claisen product. The rearrange-
ment of 7 -ethylallyl phenyl ether (I) in N,N-diethylaniline at 195°C was followed
by infrared measurements and gave the data shown in the graph below.

....... Ether

Normal product III

Abnormal product II

Mole/L.

10 20 30 kO

Time (hours)

Furthermore, if the normal Claisen product III was heated in N,N-diethylaniline or
neat at 200-225°C, it slowly rearranged to the abnormal product II„ Roberts and
Landolt17 have found that heating either XVII or XVIII, the normal and abnormal
products of 7 -ethylallyl g-tolyl ether, results in the same equilibrium mixture of
the two with the abnormal product being favored by a ratio of 2k 1 1.

CH2CH3
CHCH=CH2

XVII

200°C,

"pit

H3
HCH^CHCH3

XVIII

Jefferson and Scheinmann10 have reported that rearrangement of 7,7-dimethylallyl
estrone ether (VIII) in N,N-diethylaniline produces only the abnormal product XI.
However, if the rearrangement is carried out in diethylaniline containing butyric

anhydride, the normal Claisen product can be
trapped as its butyric ester XIX. The normal
Claisen product rapidly isomerized to the abnormal
product XI when it was heated in diethylaniline.
On the basis of this evidence, Marvell, Roberts, and
Jefferson conclude that the abnormal Claisen
rearrangement is really the result of two consecutive
processes? normal Claisen rearrangement of the 7-
alkylallyl aryl ether to the o-(a-alkylallyl) phenol,
followed by rearrangement of this phenol to produce the isomeric phenol. Although
the evidence strongly supports this conclusion, it cannot rule out another possibility
in which the normal product revcrsibly forms the allyl aryl ether which then reacts

CHrjCHgCH^C

XIX

-385-

by some mechanism to give the abnormal product directly. However, in equilibration
experiments such as that involving XVII and XVIII , no ether has ever been reported
found 014'17

Marvell and co-workers16 also reported that the methyl ether of o~(a-ethyl-
allyl) phenol (III) was recovered unaltered upon heating it under the conditions which
converted o-(a-ethylallyl) phenol to the abnormal product II0 This indicated that
the rearrangement of the normal product to the abnormal product depends on the
phenolic hydroxy! group. Also,, 2 ,6-dimethyl-4-(a-ethyla!lyl) phenol was stable under
the same conditions showing that the allyl side chain must be ortho to the hydroxyl
group in order for rearrangement to occur. Marvel! proposed the following mechanism
for the rearrangement between the normal and abnormal products.

► CH2
Of CH

X^chc:

HCH2CH3

H3
HCH-CHCH3

1.4,1

XXI

Infrared studies of o-allylphenols have revealed that the phenolic proton is hydrogen
bonded to the ally! it-bond, thus indicating that the molecule is in the proper
conformation for reaction.18 When the spirodienone intermediate XX is formed from
III, the ethyl group could also be trans to the methyl group, but this configuration
would not allow further rearrangement to II by the intramolecular process shown.
q The intermediate XX is similar to spiro[205]octa-I,^--dien-3-one
(XXI) isolated by Winstein and Baird.19'20 Since this compound
decomposed upon standing at room temperature, it is quite unlikely
that a spirodienone intermediate such as XX could be Isolated
under the conditions of the abnormal Claisen rearrangement.

Marvell 3s mechanism is strongly supported by experiments
dealing with 14C labeling, deuterium incorporation, and cis -trans
isomerization of substituted o-allylphenols . Lauer and Johnson^'1
have shown that heating y-14C-methyla!lyl g-carbethoxyphenyl ether (XIII) for 280
hours at 220°C results in a 50 s SO distribution of 14C between the two positions
shown in XIV and XV. This is the result predicted by Marvel! !s mechanism since as
shown in intermediate XX (replace ethyl by methyl) 9 the 14C labeled methyl group can
become one of fcwo symmetrically equivalent methyl groups. This should allow equal
distribution of the radioactive label between the methyl and methylene positions in
XIV and XV. Similar results have been obtained by Schmid.14

Schmid and co-workers25*'23 have studied the incorporation of deuterium into 2-
(a-methylailyl) -if- methylphenol (XXII). As shown below the deuterated phenol XXI.Ia
can form either the cis spirodienone intermediate XXIII or the trans intermediate
XXIV by intramolecular transfer of the deuterium atom. The trans intermediate would
probably be formed more often because it is very likely that the cis intermediate is
of higher energy, due to steric factors. In the cis intermediate, proton transfers
from the methyl groups can produce phenols XXIIb or XXIIc. However, in the trans
intermediate only phenol XXIIb can be formed by an intramolecular proton transfer.
Therefore, if 2-(a-methylally!) -4-methylphenol (XXII) is heated in D20, deuterium
should be initially incorporated more rapidly at the methylene position ( sCH2) than
at the a -methyl. However, at equilibrium the amount of deuterium incorporated at
the methylene carbon should be statistically equal to that at the methyl. In
other words, if n equals the amount of deuterium on the methylene carbon and m the
amount on the methyl, then at equilibrium 3n should equal 2m or 3n/2m should equal
one assuming no deuterium isotope effects.

-386-
OH
^^^CHCH-CHD

CH3

XXIIb

CHaD

HCH<

XXIIa

CHCH=CH2
CH2D

XXIIc

By heating 2-(a-methylallyl) -4-methylphenol (XXII) in D20 at 200°C for various periods
of time and determining the deuterium content and distribution by combustion and nmr,
Schmid obtained the results shown in Table Ie

Table I

Time heated
(hours)

fo D in side chain
100/o = 5 D

k
2k
48

k.e
k9.k

68.6

3n/2m

1.78
1.18

No deuterium was incorporated at the a or p positions of the allyl group. The fact
that 3n/2m equals approximately five after four hours heating indicates that
deuterium is initially incorporated more rapidly at the methylene position as
predicted » When the deuterated phenol which possessed a 3n/2m value of 1.18 was
heated in water for ten hours, the 3n/2m value changed to O.67, showing that
deuterium is more rapidly removed from the methylene carbon than the methyl carbon
as would be predicted.

Schmid also studied deuterium incorporation into 3>5-dimethyl-2-(a-methylallyl)
phenol (XXV).22'23

H 0 TT CH3

■113

HCH2sCH2

XXV

XXVI

In this case the trans intermediate XXVI should be disfavored due to repulsion
between the cyclopropyl methyl group and the methyl group at the 3-position of the
benzene ring. It would be expected therefore that an increased amount of cis
intermediate possessing symmetrically equivalent methyl groups should help to
equalize the rates at which deuterium is incorporated at the methylene and methyl
positions. In other words, the value of 3n/2m should approach a value of one much

more quickly for XXV than it did for XXII. It was found that when XXV was heated in
D20 at 200°C for only twelve hours, 3n/2m equaled 1.04. If this result is compared
with those for phenol XXII in Table I, it is apparent that the formation of the
trans intermediate XXVI was significantly suppressed.

By the Marvell mechanism, any _o-allylphenol should be able to form a spiro-
dienone intermediate similar to XX. For this reason compounds XXVII and XXVIII could
be expected to incorporate deuterium at the 7-carbon of the allyl group. 22*23

a p 7

H2CH*sCH2

CH2C — CH2
CH3

XXVII

XXVIII

Heating l-allyl-2-naphthol (XXVII) in D20 at 200°C for 48 hours, resulted in almost
complete incorporation of deuterium at this position only. When XXVIII was heated
in D20 at 200°C for 48 hours, I.85 deuterium atoms were found on the 7-carbon and
less than 0.1 on the 6 -methyl group.

The thermal interconversion of cis and trans o-( 7 -alkylallyl) phenols can also
be explained by the Marvell mechanism. Schmid and Frater24 found that heating either
cis- or trans -XXIX in diethylaniline at 200°C for several days resulted in the same
equilibrium mixture of the two isomers.

200°C

CH3 \/ CH3

cis -XXIX

K, / . -3.6
trans/ cis

:h3

trans -XXIX

The methyl ethers of cis- or trans -XXIX did not interconvert under these conditions,
showing that the hydroxyl group is necessary for the isomerization. Marvell 's
mechanism can account for this cis -trans interconversion, since the 7-carbon of the
allyl group becomes a saturated center in the spirodienone intermediate. Marvell25''26
has shown that heating either cis or trans o-(a-7-dimethylallyl) phenol (II) at 210°C
produces the same equilibrium mixture of 22^> cis-II and 78$ trans -II. The isomeriza-
tion of cis-II in water and D20 at 205°C followed first order kinetics with k^/k^
equal to 2.8. The rate of formation of trans -II from cis-II in D20 was found to be
equal to the rate of olefinie deuterium incorporation as measured by nmr. The nmr
of trans -II which was isolated after 0.7,? 2.0, 3«6, and 4.8 half lives showed the
presence of one olefinie proton. The 7 -methyl appeared as a clean singlet. This
showed that there was one deuterium at the 7-carbon of each molecule of trans -II.
These data show that every molecule of cis-II which is converted to trans -II
incorporates one deuterium into the allyl side chain at the 7-carbon as shown below.

CH3

rVicH H

CHa

XXX

Therefore, the proton transfer in intermediate XXX to give the trans isomer must be
completely stereoselective within the limits of the experiment.

-388-

ALIFHATIC ANALOGUES

Marvell' s mechanism can account for the formation of the abnormal product V in
the rearrangement of 7-ethylallyl vinyl ether (IV) as shown below.

normal
Claisen )

of*N

XXXII

XXXIII

The enols XXXI and XXXII are directly analogous to the o-allylphenols obtained in
the rearrangement of 7-ethylallyl phenyl ether (I).

That the above mechanism is indeed involved in aliphatic systems is supported
by deuterium isotope experiments by Roberts and co-workers. 27f2B These workers
heated ^-pentenophenone-2-d2(XXXIV)and 3-methyl-4-pentenophenone-2-d2(XXXV) in sealed
tubes at 202°C and followed the changes in the deuterium distribution by nmr.

C6H5-C-

(a) (b) (c)(d)
■CD2-CH2-CH«CH2

XXXIV

( (a) (b)(c)(d)
CgH5-C-CD2-CH-CH=CH2
" CH3(e)

XXXV

By applying the Marvell mechanism, it can be predicted that at equilibrium compound
XXXIV should have one deuterium at each of positions (a) and (d) and no deuterium
at positions (b) and ( c) . Compound XXXV at equilibrium should have 0,57 deuteriums
at each of positions (a) and (d) , 0.86 deuteriums at position (e), and no deuterium
at positions (b) and (c). The experimental results are given in Table II.

Table II

Compound

Time heated
(hours)

Deuterium (g-atom)
CH2 (a) CH2 (d) CH3 (e)

XXXIV

XXXV
11

1^5

121

1.73
1.23

0.97

1.27
O.83
O.58

0.20
0.7^

0.90

0.60
0.63
0.63

0.12
0.52
0.66

Both compounds had no deuterium in positions (b) and (c). These results are in
fairly good agreement with those predicted by the Marvell mechanism. However, the
results are complicated by the fact that mass spectroscopy of the product obtained
after heating XXXIV for 72 hours showed the presence of d0, d3, and d4 molecules
indicating that intermolecular deuterium exchange had also occurred.

Other experimental work by Roberts and co-workers supports the intermediacy
of the cyclopropyl compound XXXIII in the Marvell mechanism.28'29 l-Acetyl-2,2-
dimethylcyclopropane (XXXVI) smoothly rearranged to 5-methyl-5-hexen-2-one (XXXVII)
at temperatures above 150°C. The reaction follows first order kinetics and has an
activation energy of 30 kcal/mole and an entropy of activation of -10 eu. The

o ox uv^ wux«~ oi t-^mp^una juui«i Xt> oUt_n

Hie methyl group must be cis to the carbonyl

CH3C=0

163°C
sealed
tube

fi £-3

CH3C -CHgCHgC -CH2

XXXVII

group and thus in a favorable position for the intramolecular 1,5 -hydrogen shift as
shown in intermediate XXXIII. That a cis relationship is required between the
carbonyl and methyl groups in order for the rearrangement to occur is shown by the
thermal isomerization of the eis and trans isomers of l-acetyl-2-methylcyclo-
propane.28'29 At l60°C, the cis isomer rearranged almost completely in twelve hours
to the expected 5-hexen»2-one, but at l80°C, the trans isomer decomposed only
slightly over a period of 2k hours and the product was not 5-hexen-2-one. Other
examples of this same type of rearrangement are reported in the literature.30""32
The rearrangement of l-aeetyl-2,2-dimethylcyclopropane (XXXVI) is analogous to the
rearrangement of cis -1 -methyl- 2-vinylcyclopropane (XXXVIII) involving the homo-
dienyl-l,5»hydrogen shift. 33-»34

sjx

£H<

160°C

XXXVIII

This rearrangement follows first order kinetics and has an energy of activation of
30 kcal/mole and an entropy of activation of -12 eu. 9 which are very similar to
those values previously mentioned for XXXVI. trans °l-Msthyl-2-vinylcyclopropane is
stable at 250°C. Ohloff31 has studied the thermal rearrangement of cyclopropane
XXXIX which contains both vinyl and carbonyl groups cis to methyl groups.

300°C

50 ram

X*S

CH.3 CH3
XXXIX

XLI

The major product XL was due to reaction involving the carbonyl group. The other
product XLI, which was due to reaction of the vinyl group, was formed in approximately
ICffo yield.

SUMMARY

The abnormal Claisen rearrangement has been identified as the result of two
consecutive processes; normal Claisen rearrangement of the 7-alkylallyl aryl ether
to an £-(a-alkylallyl) phenol, followed by rearrangement of the side chain of this
phenol to produce an isomeric phenol. The mechanism of the secondary rearrangement
has been formulated as involving a substituted spiro[2.5]octa-4,6-dien-3-one inter-
mediate. Considerable experimental work has provided strong support for this
mechanism. Abnormal Claisen rearrangements involving aliphatic systems appear to
be closely analogous to the aromatic case.

BIBLIOGRAPHY

1. W. M. Lauer and W. F. Filbert, J. Am. Chem. Soc, £8, 1388 (1936).

2. W. M. Lauer and H. E. Ungnade, ibid. , 58, 1392 (1935).

_xon_

3. W. M. Lauer and L. I. Hansen, ibid. , 6l, 3039 (1939).

4. C. D. Hurd and M. A. Pollack, J. Org. Chem. , 3, 550 (1939).

5. W. M. Lauer and R. M. Leekley, J. Am. Chem. Soc, 6l, 3043 (1939).

6. W. M. Lauer and H. E. Ungnade, ibid. , 6l, 3047 (1939).

7. C. D. Hurd and M. D. Puterbaugh, J. Org. Chem., 2, 381 (1938).

8. W. L. Lauer and 0. Moe, J. Am. Chem. Soc, 65, 2B9 (I9U3).

9. W. L. Lauer and P. A. Sanders, ibid. , 6^, 198 (19^-3).

10. A. Jefferson and F. Scheinmann, Chem. Commun. , 239 (1966).

11. D. S. Tarbell, Org. Reactions, 2, 1 (1944) .

12. D. S. Tarbell, Chem. Rev., 27, ^95 (19^5).

13. W. M. Lauer, G. A. Doldouras, R. E. Hileman and R. Liepins, J. Org. Chem., 26,
W5 (1961).

14. A. Habich, R. Barner, W. von Philipsborn, and H. Schmid, Helv. Chim. Acta.,
4^, 19^3 (1962).

15. H. Schmid, Gazz. Chim. Ital. , 92, 968 (1962).

16. E. N, Marvell, D. R. Anderson, and J. Ong, J. Org. Chem., 27, IIO9 (I962).

17. R. M. Roberts and R. G. Landolt, ibid., ^1, 2699 (I966).

18. A. W. Baker and A. T. Shulgin, J. Am. Chem. Soc, 80, 5358 (I958).

19. S. Winstein and R. Baird, ibid., 7£, 4238 (1957).

20. S. Winstein and R. Baird, ibid., B|, 567 (1963).

21. W. M. Lauer and T. A. Johnson, J. Org. Chem., 28, 2913 (1963).

22. A. Habich, R. Barner, W. von Philipsborn, and H. Schmid, Helv. Chim. Acta., 48,

1297 (1965).

23. H. Schmid, Oesterr. Chemiker-Ztg. , 6^, 109 (1964).

24. G. Frater and H. Schmid, Helv. Chim. Acta., 4£, 1957 (1966).

25. E. N. Marvell and B. Schatz, Tetrahedron Letters, 67 (1967).

26. E. N. Marvell, J. L. Stephenson, and J. Ong, J. Am. Chem. Soc, 87, 1267 (1965),

27. R. M. Roberts, R. N. Greene, R. G. Landolt, and E. W. Heyer, ibid., 87, 2282
(1965).

28. R. M. Roberts, R. G. Landolt, R. N. Greene, and E. W. Heyer, ibid. , 8£, l404

(1967).

29. R. M. Roberts and R. G. Landolt, ibid., 87, 2281 (I965).

30. D. E. McGreer, N. W. K. Chiu, and R. S. McDaniel, Proc Chem. Soc, 415 (1964).

31. G. Ohloff, Tetrahedron Letters, 3795 (I965).

32. K. von Auwers and 0. Ungemach, Ann., 511, 152 (1934).

33. R. J. Ellis and H. M. Frey, J. Chem. Soc, 5578 (1964).

34. D. S. Glass, R. S. Boikess, and S. Winstein, Tetrahedron Letters, 999 (I966).

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Organic seminar abstracts"

See other formats