SYNTHETIC AlND MECHANISTIC STUDIES OF
RUDOLPH MILTON ALBERT, JR.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
This Dissertation is Dedicated to
Elizabeth, Laura, Robert,
Pebecca and Daniel
AC K':O'..LE DGE E TS
I wish to express my deep appreciation to Dr. G.B. Butler
for his friendship and guidance as my research director. The advice
and suggestions of the other members of my supervisory committee
are also acknowledged with gratitude.
I:y colleagues in the laboratory made working more pleasant
and for this I am grateful. A special note of thanks is extended
to Dr. Richard L. Veazey for many valuable discussions and helpful
I wish also to thank the management of the Organic Chemicals
Group, Glidden-Durkee Division of the SCt; Corporation and especially
Mr. R.P.T. Young for a very generous leave of absence which made
it possible to come here at all.
The patience and skill of my typist, Mrs. Jimmie McLeod, was
of inestimable value in the preparation of this manuscript and I
would like to give special recognition to her.
Finally, I wish to give a special expression of gratitude
to my family. All have sacrificed much to make the completion of
this program possible and their abiding love has been a source of
strength and determination.
TABLE OF CONTENTS
ACKNIOWLEDGEMENTS . . . . . . . . .
LIST OF TABLES . . .
LIST OF FIGURES . .
ABSTRACT . . . . . . . . . .
I. INTRODUCTION . . . . . . .
II. STUDIES RELATED TO THE PREPARATION OF
3,3-DItMETHOXYCYCLOPROPENE . . . . . . 7
A. Reaction Ey-Products . . . . . . 10
B. Extension to Other Systems . . .. . . 14
C. Preparation of Other Ketals . . . ... 16
D. Cyclization of the Ethylene Ketal ...... 19
E. Improved Procedure for Preparation of
Dimethoxvcvclopropene . . . . . ... 23
F. Other Cyclization Processes . . . ... 25
III. CHARGE TRANSFER COMPLEXATION . . . ... 26
IV. 4+2 CYCLOADDITIOI. THE DIELS ALDER REACTION . 33
A. Reaction with Butadiene Derivatives . . .. 36
B. Reaction with Cyclic Dienes . . . ... 47
V. 2+2 CYCLOADDITION REACTIONS . . . . .
A. Reaction with Hexafluoroacetone . . .
B. Reaction with Enamines . . . .
C. Reaction with Triazolinediones ..
D. Reaction with Dimethylacetylenedicarbo::ylate
VI. REACTIONS WITH AMINES . . . . . . . .
. . . vi
VII. SU:-::-IARY AND CONCLUSIONS
VIII. EXPERIMENTAL . . .
A. Equipment and Data
B. Synthesis . . .
C. Procedures and Data
Couple:: Studies .
REFERENCES CITED . . . .
BIOGRAPHICAL SKETCH . . .
for Charge Transfer
. . . . . . . .
. . . . . . . .
. . . . . . . .
LIST OF TABLES
I. Mass Spectral Fragments from 3,3-Eth'lenedioxv-
cyclopropane . . . . . . . .... . 20
II. Variation of the Chemical Shift of the Ring Protons
of 3,3-Dimethoxycyclopropene with Solvent . . .. 27
III. IIMR Determination of the Equilibrium Constants of
Charge Transfer Complexes . . . . . .... 32
IV. Relative Abundances of Mass Spectrumr Peaks from
Diels Alder Adduct . . . . . . . . 42
V. Data for Dimethox:,'cy,'clopropene-Styrene Complex Study 132
VI. Data for Dimethox:ycyclopropene-Divinyl Ether Complex
Stud, . . . . . . . . .... . . .. 133
LIST OF FIGURES
1 NMR Spectrum of C3H BrC13 .............. 12
2 Ultraviolet Spectrum of 3,3-Dimethoxycyclopropene 28
3 iNR Study of the Complex Between Dimethoxycyclopropene
and Styrene . . . . . . . . ... .. 30
4 M.? Study of the Complex Between Dimethoxycyclopropene
and Divinyl Ether . . . . . . . ... 31
5 -iR Spectrum of the Dimethoxycyclopropene-Dimethyl-
butadiene Adduct . . . . . . . . ... .37
6 !.MR Spectrum of Dimethoxycyclopropene-Isoprene Adduct 38
7 M1IR Spectrum of the Adduct Formed Between Dimethoxy-
cyclopropene and Tetracyclone . . . . ... 51
8 U.R Spectrum of Dinethoxycyclopropene-Hexafluoro-
acetone Adduct . . . ... . . . .. . .60
9 NIR Spectrum of l,l-Dimethoxy-2-diethylaminocyclo-
propane . . . . . . . . ... .. . .75
J0 UfIR Spectrum of N,N-dipropyl-B-alanine . . ... 77
11 I-:R Spectrum of N,N-diphenyl-B-alanine . . ... 81
12 NRP. Spectrum of the Intermediate Leading to N,NI-diphenyl-
B-alanine . . ... . ... . . . .83
Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy
SYNTHETIC AND MECHANISTIC STUDIES OF
Rudolph Milton Albert, Jr.
Chairman: Dr. George B. Butler
Major Department: Chemistry
The preparation of 3,3-dimetho:-:ycyclopropene in pure form was
first reported in 1970. It was synthesized by treating l-bromo-3-
chloro-2,2-dimetho:x:ypropane with potassium amide in liquid ammonia.
The precursor was obtained by the acid-catalyzed reaction of 2,3-
dichloropropene with N-bromosuccinimide and methanol.
This research program has been directed toward expanding and
improving the reactions leading to dimetho::ycyclopropene as well
as toward defining the reactivity of this unusual intermediate.
Direct synthesis of different ketals was attempted by substituting
other alcohols for methanol. Bromine addition was a major competing
reaction. Likewise, 2,3-dibromopropene gave only 1,2,2,3-tetra-
bromopropane when reacted with 11-bromosuccinimide and methanol.
Successful preparation of the ethylene ketal of l-bromo-3-
chloro-2-propanone was realized by acid-catalyzed exchange of ethylene
glycol on the dimethyl ketal. Cyclization of this intermediate was
also attempted since the resulting cyclopropene might be more reactive
than dimethoxycyclopropene in situations where steric crowding is
critical. Two products were obtained from its reaction with potassium
amide, however. One was the expected cyclopropene which could not
be purified. The second product was obtained in pure form and was
identified as the cyclopropane derivative L ,7-dioxaspiro[2.u]heptane.
Significant improvement in both yield and operational convenience
was realized by modifications in the reaction procedures. Instead
of adding the ketal intermediate to the aide solution, potassium
amide was prepared in one flask and then transferred by means of
nitrogen pressure to a second flask containing the intermediate in
Dimethoxycyclopropene was found to be an electron-deficient
olefin that forms weak charge transfer complexes with styrene and
with divinyl ether.
In Diels-Alder reactions, dimethoxycyclopropene reacted quanti-
tatively with derivatives of butadiene to give the expected 7,7-
dimethoxy-3-norcarenes. With cyclic dier.es, such as cyclopentadiene
and furan, reaction was inhibited by steric interaction. In both
these cases only dimerization of dimethoxycyclopropene was observed.
Its reaction with tetracyclone gave the ket.al of a 2t5H)-furanone
by addition of the opened cyclopropene system to the carbonyl function
of the diene.
Cycloaddition reactions with olefins to give cyclobutane
products were unsuccessful with simple carbon-carbon double bonds.
Eoth electron-rich (enamines) and electron-poor (acetylene dicarboxvlate)
oltfins were completely unreactive. Species containing a hetero atom
did react, however. Hexafluoroacetone gave a bicyclic adduct by
formal 2t2 cycloaddition of the olefin and the carbonyl group. With
4-phenyl-1,2,4-triazoline-3,5-dione and the corresponding 4-methyl
derivative the products were copolymers which did not retain the
Reaction of dimethoxycyclopropene with diethylamrine gave only
1,l-dimethoxy-2-diethylaminocyclopropane. With dipropylamine,
however, II,11-dipropy'l-B-alanine methyl ester was obtained along with
the expected cyclopropylamine. 'When diphenylamine was used there
was no c,,clopropylamrrine observed and only the methyl ester of
tN.,l-diphenyl-B-alanine was recovered. An intermediate was detected
but could not be purified due to the ease with which it was converted
to the alanine derivative. A mechanism involving a ketene acetal
as the intermediate is consistent with the data available.
Although cyclopropene and many of its derivatives have been
available for many years, cyclopropene chemistry has been the object
of considerable interest only in the past two decades. The upsurge
of activity which took place in the mid-fifties was due to three
main factors. These factors were (1) the developments in carbene
chemistry which aided synthesis of cyclopropene systems, (2) the
applications of molecular orbital theory which made attack on theo-
retical problems a desirable goal, and (3) the discovery of the
cyclopropene system in natural products.
Many systems were designed and studied, but cyclopropene deriv-
atives unsubstituted in the double bond positions are relatively
uncommon. The parent hydrocarbon was first synthesized in 19222
but it was not until the facile preparation by Closs3 in 1966
that it was used for extensive studies. Geminal dimethyl cyclopropene
was prepared by Closs4 in 1961 by base catalyzed thermal decomposition
of 3-methyl-2-butenal tosylhydrazone. This procedure also yielded
the very unstable 3-methylcyclopropene5 although in poor yield.
Studies of the chemistry of these derivatives is limited and deal
mainly with isomerization reactions. A kinetic study of the thermal
isomerization of 3-methyl-, 3,3-dimethyl-, 1,3,3-trimethyl-, and
1,2,3,3-tetrainethylcyclopropene showed that increased substitution
caused an increase in thermal stability.6
Synthesis of 3,3-diphcnylcyclroropene was observed upon irrad-
iation of 3,3-diphenyl-3H-pyrazole7 but rearrangement to 3-phenyl-
indene occurred on gas chromatography.
A mixture of the halogenated derivatives, 3-chloro-, 3,3-dichloro-
and 1,3-dichlorocyclopropene is produced by reduction of tetrachloro-
cyclopropene with tri-n-butyltin hydride. Breslow used 3-chloro-
cyclopropene in the preparation of unsubstituted cyclopropenyl cation
and 3,3-dichlorocyclopropene was hydrolyzed to cyclopropenone.
The mixture of reduction products was used as a probe into the
mechanism of Diels-Alder adduction of perhalocyclopropenes0 but
no adducts with 3,3-dichlorocyclopropene were observed.
Bertrand and Monti devised the following scheme for the
preparation of cyclopropenyl ketones:
IH3 HCBr, C 3 Zn 3
R-CH=C-C-CH > R-CH-L-CH R-CH-C-CECH
tBuOK H, HAc \C
Br Br Br H
Ht+ It 1. tEu. L 3 CH
HE > R-CH---C- CH > -
glycol 3 2. H 0
Er H H
They studied the thermal rearrangement of this ketone to 3-acetyl-
1-methylcyclopropene2 and also the Diels-Alder adducts which it3
and its ethylene ketall form with cyclopentadiene. These adducts
are to be described in Chapter IV.
In the course of studies directed toward the synthesis of
cyclopropenone, Pratt prepared 3,3-dimethoxycyclopropene and
presumably cyclopropenone itself for the first time. His plan
was to e::trude the cyclopropene derivative from the Diels-Alder
adduct formed between tropone dimethylketal and dimethyl
acetylenedicarboxylatc. Purification of the adduct was a major
problem in this procedure and it was, in fact, during an attempted
distillation that the formation of dimethoxycyclopropene was first
suggested. Subsequent experiments indicated that this "batch
pyrolysis" was the preferred process for the extrusion reaction.
OC 2C 3 302
+: ). .. 3,
OCH I OCH3
CO CH 3 CH
Although he was never able to isolate dimethoxycyclopropene in
pure form, he did report MIR and I. spectra which support its presence
among his reaction products. Attempts were made to determine the
molecular weight by mass spectrometry and by a gas density method
but neither gave satisfactory results. Reaction with cyclopentadiene
was attempted and the lack of reactivity was attributed to inadvertent
hydrolysis to cyclopropenone. The evidence presented fcr cyclo-
propenone formation was by inference only and a later evaluation of
this worki6 did not claim its preparation.
In the course of attempting to synthesize compounds potentially
capable of charge-transfer complexation prior to polymerization,
Eaucom discovered an easy and relatively simple procedure for the
preparation of l-bromo-3-chloro-2,2-dimetho::ypropane.17 Subsequent
investigation of this reagent led to the successful isolation of
pure 3,3-dimethoxycyclopropene. (Scheme I)
The reactivity of this ketal was investigated in a preliminary
series of experiments.8 Baucom found that it gave a Diels-Alder
adduct with 1,3-diphenylisobenzofuran but not with tetracyclone.
In the latter case, a product was isolated which was the result of
solvent carbon tetrachloride addition to the dimertho.:ycyclopropene.
CH i OCH OCH3
3 -s- 3 CH O C4 OCH3
3 CC1 3
--. CHC1=CH-C" + *CC1
CH2= -CH2C1 I-Br t Cil OH H2 4 BrCH2--C-CH2Cl
C11 3 H SO BC
CH 30 OCH
K 1 1 H 2
H C CH Br
CH 0 OCH3
CH 30\ OCH 3
CH3O 0 CH3
Scheme 1. Synthesis of 3,3-Dimethoxycycloopropene7
When a pure sample of dimethoxycyclopropene was added to
deuterium oxide, it was hydrolyzed immediately to cyclopropenone.
No attempt was made, however, to isolate the pure ketone.
In an attempt to prepare a Diels-Alder adduct with a-pyrone,
the major product isolated was a dimer of dimethoxycyclopropone.
Preliminary evidence was also obtained for the formation of the
Diels-Alder adduct but this was not confirmed.
CH OC.3 CH 0 OCHI 3
Treatment of l-bromo-3-chloro-2,2-dimerhoxypropane with potassium
tert-butoxide gave only 1,l-dimethoxy-2-t-butoxycyclopropane, even
in experiments designed to trap the cyclopropene as it was formed.
This suggested that dimethoxycyclopropene experiences facile attack
by nucleophilic reagents. The addition of other nucleophiles was,
therefore, of interest and Baucom reacted the ketal with excess
dimethvlamine and did, indeed, obtain the expected dimethylamino
The results of these experiments indicate that 3,3-dimethoxy-
cyclopropene is a highly reactive reagent possessing a somewhat
electron-deficient double bond. Its ready availability in relatively
large quantity suggests its use as a source of cyclopropenone and
also as a means of incorporating a cyclopropanone moiety into
previously inaccessible systems by preliminary adduct formation and
subsequent hydrolysis to the free carbonyl function.
The present research program was undertaken to investigate this
system on several fronts. First, the reactions leading to 3,3-di-
methoxycyclopropene are unique and their generality was to be esta-
blished. Second, only MINR and I. properties of dimethoxycyclopropene
were reported by Baucom and other characteristics of this reagent
were desired. The third and most general objective was to establish
the chemical reactivity of this compound toward a variety of systems
and consequently to suggest the most efficient use of 3,3-dimethoxy-
cyclopropene as a synthetic reagent and as a possible monomer in
Studies Related to the Preparation of 3,3-Dimethoxvcvclopropene
The series of reactions leading to 3,3-dimethoxycyclopropene
which Baucom discovered involves rather unique procedures and one
objective of this research program was to evaluate the scope of
these reactions. Also, while the chemical processes involved are
the same as those of the original work, significant improvements
in yield and operational convenience have resulted from modifications
When Baucom added Jl-bromosuccinimide to 2,3-dichloropropene
in refluxing methanol containing a catalytic amount of sulfuric
acid, l-bromo-3-chloro-2,2-dimethoxypropane was obtained in up to
40 percent yield as a pure crystalline material.17 A Second major
reaction product was proposed to be l-bromo-2,2,3-trichloropropane.
The proposed sequence of reactions is shown in scheme 2. Because
of the complexity of side reactions possible, the 33-40 percent
yield of the desired ketal was thought to be quite good and incapable
of significant improvement without resorting to very expensive or
One possibility which was investigated, however, was blowing
air through the reacting mixture. This would serve to remove some
of the by-product HCI and to possibly air-oxidize Br and promote
further reaction. A definite increase in yield was noted when this
-Er i H+
Er + CH=C-CH C1 >
ErCH -C-CH Cl
2 + 2
ErCH -C-CH2C 1
MleH E BrC.- -T-CH2Cl
+ CH H ErCH -2CHC1 + C1
2 Ii1 H 2"
-> ErCH2CCl 2CH2C1
Scheme 2. Reaction of 2,3-Dichloropropene with iE.S.'/Methanoll7
was done but a subsequent series of reactions revealed that the
effect was due to a decrease in reaction temperature and not to
any action of the air itself. Thus, modification of Eaucom's
procedure resulted in routine yields of around 65 percent by slow
addition of il5S at 25-30' without the use of air-blowing. Slight
increases could be realized to 5-10 but the time required for each
portion of IBS to react increased and made the overall process too
lengthy. One experiment using perchloric acid as catalyst gave
a slightly higher yield but the added hazard of the reagent out-
weighed this advantage.
One explanation for the effect of lower temperature is that
solvolysis of l-bromo-2-methoxy-2,3-dichloropropane (reaction 4
of scheme 2) is slower. This would decrease the rate of formation
of by-product HC1 and favor production of the intermediate (reaction
3) over the formation of l-bromo-2,2,3-trichloropropane (reaction 5).
If this were the case, then very rapid addition of NBS with subsequent
rapid formation of the intermediate would further favor the desired
reaction. It had been felt that rapid addition of the reagent would
have an adverse effect on the yield by favoring Br2 addition and
at the same time make control of the reaction e::otherm difficult.
However, experience with the process began to indicate that if
the initial temperature were low, an adequate ice bath should control
the temperature and allow an evaluation of the yield.
Thus, N-bromosuccininide was added as rapidly as possible to
2,3-dichloropropene in cold methanol containing a catalytic amount
of sulfuric acid. The temperature, originally at -100, rose to -30
but quickly fell to -100 again. This temperature was held for
three hours and then gradually warmed to room temperature overnight
as the ice/acetone bath melted and equilibrated. The yield of
l-bromo-3-chloro-2,2-dimethoxypropane was 71 percent of theory.
Wiile the increase in yield may be moderate, the added operational
convenience is significant. Typically, the apparatus can be assembled
and the reagents charged in the last half-hour of the day. After
stirring overnight, the product is isolated and purified by standard
procedures. The older procedure required addition times of two to
five hours and then storage overnight before work-up.
A. Reaction By-Products
An evaluation of the by-products was also attempted in an
effort to better understand the reactions taking place. Analysis
by Ill-IM of the by-product mixture revealed that very little direct
bromination takes place. A sample of the oil was fractionally
distilled under reduced pressure and 85 percent of it was collected
as a single fraction (b.p. 103-1040/58 mm. Hg). Chromatography on
silica gel removed the last traces of impurity and gave what was
apparently pure compound.
The mass spectrum gave a molecular ion peal: at m/e 224 with
the proper isotopic distribution ratio for C3H BrCl3. The base
peak at m/e 145 corresponds to the expected loss of a bromine radical
and further fragmentation by loss of HC1 gives an abundant peak
at m/e 109 (55 percent of base peak). These observations are consistent
with the structure proposed in scheme 2 but do not allow any con-
clusions about the structure. The major inconsistency in correlating
the spectrum with the proposed structure is an abundant (53 percent)
peak at m/e 83 with isotopic distribution ratio for CHC1 Rearrange-
ment pathways leading to this ion are difficult to rationalize.
The only fragments retaining bromine are relatively small (12 percent)
and appear at m/e 189 and 175 corresponding to loss of -Cl and -CH2Ci
respectively. The loss of these radicals is most easily visualized
as a-cleavage of a molecular ion bearing bromine on the central
Ir^2 ej- 2
CH Cl Br+
2 2. -Cl. II
2. -Cl CICH2-C-CH2Cl
The NtIR spectrum of this sample is also inconsistent with the
suggested l-bromo-2,2,3-trichloropropane as seen in Figure 1.
The three singlets appear to have a ratio of 2:1:1 and 1-bromo-
1,2,3-trichloropropane was considered a possibility. This structure,
however, should have a more complex: spectrum and greater differences
in the chemical shifts of the methylene and the methine protons.
The observed spectrum could arise if the rotation about one of the
C-C bonds were sufficiently hindered to place the two protons affected
in different environments. There is no indication of such hindrance
in either the product ketal or the dibromodichloropropane which was
prepared by addition of bromine in carbon tetrachloride to 2,3-dichloro-
CH 2=CC1-CH Cl t Br2/CC1 ---> BrCH -CClBr-CH Cl
It appears likely, therefore, that two by-products are produced
that differ only in the positions of the bromine atom and one of
the chlorines. This condition could arise by attack of Cl with
equal facility on either end of a bromonium ion intermediate.
_i__--------_____ 5 -cps
, 7I0 *60-- I 5.0
Figure 1. NMR Spectrum of C3H PrC13
Cl Br C1
It is interesting to note that there is very little, if any, of the
corresponding attack at C-1 by CH30H to give l-methoxy-2-bromo-
B. Extension to Other Systems
Extension of these reactions to other systems could serve three
purposes. First, their scope could be evaluated to determine whether
they are general or specific processes. Second, a comparison of
the reactivity of various derivatives would be a probe of the
mechanism of the cyclization reaction (viz., 1,3-dibromo- and
1,3-dichloro- vs. l-bromo-3-chloro-2,2-dimethoxypropane). Third,
the reactivity of other ketals (especially the ethylene ketal)
might be favored over the dimethyl derivative for steric reasons.
Application of the IJBS/methanol reaction to 2,3-dibromopropene
would yield 1,3-dibromo-2,2-dimethoxypropane if the same sequence
as scheme 2 were followed. Such was not to be, however, as the
dichloro and dibromo olefins were found to react differently in
our system. When 2,3-dibromoropene was reacted in methanol with
NES, the major product had only a sharp singlet at 4.266 in the in.R
spectrum. This is very near to the chemical shift of the -CHBEr
protons of the starting clefin which indicates that the major reaction
involved is probably direct bromination:
CH,Br-C=CH + 1IES + CH3OH H BrCH -C-CH,Br
S2 ---- 2
Similarly, substitution of ethylene glycol for methanol failed
to yield the ethylene ketal directly. Analysis of the crude reaction
mixture indicated that some ketal formation did occur but that
direct bromination was again the major reaction pathway. Similar
results h2ve been observed using anhydrous ethanol as solvent.
These results point to a high degree of specificity in this
reaction which requires chlorine on the starting olefin (although
probably only on C-2) and only methanol for efficient solvolysis
of the intermediates. These factors as well as elucidation of
the reaction mechanism require a separate program of study.
C. Preparation of Other Ketals
Successful preparation of l-bromo-3-chloro-2,2-ethylenedioxy-
propane was achieved in essentially quantitative yield by acid-
catalyzed exchange of ethylene glycol on the dimethoxy derivative.
CH3 CH 6 0
\V 3 H
BrCH -C-CH C1 + HOCH2CH,.OH 0 BrCH2-C-CHCl + 2 CH OH
The new ketal is a colorless liquid (m.p. 100) boiling at 70 under
1.2 mm. Hg pressure. In the InIR spectrum there are two 2-proton
singlets at 3.626 and 3.476 and one 4-proton singlet at 4.056.
The infrared spectrum and elemental analysis are consistent with
the proposed structure. The mass spectrum shows no parent peak.
The base peak at m/e 121 shows an unusual isotopic abundance ratio
of 1.65 to 1 since it can arise from two sources, one with bromine
and one with chlorine (scheme 3). The ions are at m/e 49 and 93 and
correspond to CH2Cl and CH2Br+ respectively, but the expected
metastable peaks are not observed to confirm that they. arise from
the oxonium ion species as indicated in the scheme.
Furthe-r extension of this process to the sulfur analog was
next attempted. When ethane dithiol was substituted for ethylene
glycol and the mixture was heated on the steam bath, a very vigorous
exothermic reaction initiated which erupted from the flask. Due
to the extremely disagreeable conditions which resulted, this reaction
was not repeated, but there seems to be little doubt that l-bromo-
3-chloro-2,2-ethylenedithiopropane could be prepared by this route
Scheme 3. Mass Spectral Fragmentation Pattern for 1-Bromo-
if its production were required and the proper precautions were
taken to control the vigor of the reaction.
Three other ketals were prepared by more conventional routes
from the appropriate ketones. These were
\ClCH / 3
CICH -C-CO Cl
2- 2 1
BrCH2 -C -C12 Br
A fourth, 1,3-dibromo-2,2-ethvlenedioxypropane, would complete the
series but was not prepared.
The dimethyl ketals were prepared by refluxing the ketone with
methanol containing a trace of acid catalyst. Water of reaction was
removed by passing the condensed solvent through a bed of molecular
sieve before returning it to the reaction flask. In the case of
the glycol reaction, a benzene azectrope was adequate for removing
the water. Because these ketals were difficult to obtain and are
available in very limited quantity, the comparison of their reactivity
in producing cyclopropene derivatives was not attempted. This study
should be revived when an acceptable material balance on the cyclization
process is realized or when a dependable method of derivitizing all
cyclopropene products becomes available.
D. Cyclization of the Ethylene Ketal
Since l-brono-3-chloro-2,2-ethylenedioxypropane was readily
available in large quantity, its cyclization was attempted using
the procedures available. The reaction was conducted as described
by Baucom for the dimethoxy derivative and a yield of distilled
product of 13 percent was obtained. Analysis of the material by
NMR showed three singlets at 7.72, '.01, and 0.906 with integration
intensities of 15, 140, and 110 respectively. The absorbance at
7.726 was assigned to the cyclopropene ring protons of the expected
product and the signal at 4.01 is very near that of the methylene
protons of the ketal moiety in the starting material (4.056). The
intensity of this absorbance, however, should only be twice that
of the 7.726 absorbance or thirty units. Since the excess was exactly
equal to the intensity of the absorbance at 0.906, the presence of
a cyclopropane derivative20 was suggested although none had ever
been detected in the dimethoxy reactions and its preparation would
be difficult to rationalize. Relative area ratios indicated that
the product mixture was 20 percent cyclopropene and 80 percent
Treatment of the mixture with water removed all of the cyclo-
propene but left the cyclopropane unaffected. A pure sample of the
colorless liquid was obtained by preparative gas chromatography.
In the NMR there were two singlets of equal intensity at 4.016 and
at 0.916. The mass spectrum showed an abundant molecular ion at
m/e 100 which had 74 percent the intensity of the base peak at m/e 99.
Each of these ions, then, apparently experiences a similar frag-
mentation pattern as indicated by the following table.
Table I. Mass Spectral Fragments from 3,3-Ethylenedioxycyclopropane
m/e relative intensity molecular formula fragment lost
100 74 C H 0 +molecular ion
56 62 C3H40 C2H
44 48 C2H O C3H0
99 100 C5H702 H
55 92 C3H30 C2H4O
43 90 C2H30O+ H C
40 42 C3H C2HO02
The appearance of ions at m/e 44 and 43 by loss of C H 0 from
the molecular ion and the base peak ion respectively suggests
rupture of the dioxolane ring and extrusion of the elements of
cyclopropanone. Such involvement of the dioxolane ring is unusual
and reflects the structural peculiarity of this spiro ketal.21
Support for this process is found in the metastable peaks produced.
The transition m/e 99 to 55 by loss of C2H 0 is a common reaction of
ethylene ketals and gives an observable metastable peak at 30.6.
A metastable peak at 18.7 is only slightly less intense and cor-
responds to the transition m/e 99 to 43 by loss of C H 0.
A characteristic feature of the infrared spectra of dioxolane
derivatives is a group of four or five peaks between 1000 and
1200 cm. This pattern is prominent in the spectrum of 1-bromo-
3-chloro-2,2-ethylonedio:ypropane with peaks at 1030, 1095, 1130,
and 1140 cm. The spectrum of the cyclization product, however,
does not have this characteristic pattern. In the region of interest
there are only two strong peaks at 1030 and 1185 cm.- Apparently
the molecular vibrations responsible for absorptions at these
frequencies are constrained by the small ring attached as a spiro
derivative to the dioxolane ring.
SKINH2I 0 0
BrCH -C-CH 2Cl IIH > 2
H H H. H
Formation of this product, 4,7-dio:aspiro[2,4]heptane, is
difficult to rationalize. The preparation of the potassium amide
was carefully conducted to insure that no unreacted metal remained.
Thus, when the bromochloroketal was added, the reaction medium
should not have been a reducing system and it is unlikely that
the cyclopropene derivative is being hydrogenated. Likewise, a
dehalogenation reaction involving the free metal is not possible.
Baucom obtained a similarly unexpected product retaining all four
methylene protons from the reaction of l-bromo-3-chloro-2,2-dimetho:.:y-
propane with lithium hydride in the presence of tetracyclone.
The reaction scheme postulated to account for his product could
apply in this case although the amjide displacement on bromine does
not have compelling precedence.
BrCH -C-CH C1
H., -C -H -C1
E. Improved Procedure for Preparation of Dinethoxvcyclopropene
Cyclization of l-bromo-3-chloro-2,2-dimethoxypropane with
potassium amide in liquid arnmonia yields 3,3-dimetho:-:ycyclopropene
as shown in scheme 1. The procedure developed by Baucom7 is
essentially the same as that described earlier for the ethylene
ketal. Yields were reported to be 30 to 50 percent but in the
current program this process yielded 0 to 50 percent with 10 to
15 percent being the most frequent result.
Significant improvement was realized by reversing the order
of addition of reactants in the cyclization process. The yield
was 58 percent for the first run and has been consistently above
40 percent in later runs. In addition to higher yield and fairly
consistent results, this procedure is more convenient and is a
safer process which has allowed scaling up to three times the
usual run. In the case of the first reaction using this procedure,
the starting material was completely consumed and there was little
residue after distillation of the product. Nevertheless, a total
material balance was still not achieved. Only about 75 percent of
the starting material was accounted for. This has been a persistent
problem that makes comparison of reagents and mechanism probes
unreliable and is one reason none have been attempted.
The apparatus used was similar to that of Schlatter.22
Essentially, it consisted of two flasks arranged side by side with
appropriate connecting tubes and provision for pressurizing one of
them. The potassium amide was prepared in the usual fashion in
one of them and then transferred by means of nitrogen pressure to
the other flask which contained the intermediate in excess ammonia.
A relatively short reaction time at the temperature of refluxing
ammonia was required to complete the cyclization and excess amide
was destroyed with ammonium chloride. Isolation of the product
then followed the established process.
F. Other Cyclization Processes
Several other systems were evaluated in an effort to find an
alternate method for the preparation of dimethoxycyclopropene.
Hexamethylphosphoramide (HMPA) has been used to effect dchydro-
halogenation23 and the following reaction can be visualized.
CH 0 OCH3 CH30 OCH
+ 2 HMPA )
BrCH2-C-CH2CI HMPA 2 H3)2
It should be possible to distill the product from the mixture as
it is formed. When a sample of the bromochloroketal was added to
HMPA and slowly heated to 2200 an exothermic reaction seemed to
take place and the mixture became very dark. A few milliliters
of a clear liquid distilled from the reactor but it rapidly boiled
away when warmed to room temperature. Apparently the ketal was
reactive toward HMPA and produced decomposition of the solvent to
give dimethylamine which was the liquid collected. The conditions
employed here were too severe for practical use, however.
When the bromochloroketal was heated for two hours at 750 with
sodium hydride in HMPA it was recovered unchanged in quantitative
Charge Transfer Complexation
The use of 3,3-dimethoxycyclopropene in copol.ymerization
processes affords the interesting possibility of incorporating
cyclopropanone units in polymer chains by subsequent hydrolysis
of the ketal moiety. Modification of polymer properties by further
reaction at these sites could lead to materials of considerable
A good deal of evidence has accumulated for the participation
of charge transfer complexes in many copolymerization reactions.2a
The ease with which 3,3-dimethoxycyclopropene experiences nucleo-
philic addition (eg. tert-butoxide and dimethylamine) suggests
that it is a rather electron-deficient olefin. Therefore, the
formation of charge transfer complexes between this reagent and
two electron-rich monomers was investigated. These monomers were
s-yrene and divinyl ether.
Of the two methods commonly used to study charge transfer
complexation, namely ultraviolet spectroscopy and nuclear magnetic
resonance, the latter is clearly preferred in this case for two
reasons. First, it is extremely difficult to obtain 3,3-dimethoxy-
cyclopropene of spectroscopically pure form in sufficient quantity
for use in UV studies and secondly, the IMRP, singlet of the ring
protons appears to be sensitive to changes in the electronic nature
of various solvents (Table II). Also, the UV spectrum of dimethoxy-
cyclopropene (rig. 2) shows no absorption above 210 mu.
Table II. Variation in the chemical shift of the ring protons
of 3,3-dimethoxycyclopropene with solvent.
Solvent 6 Value for Cyclopropene Protons
carbon tetrachloride 7.75
A study of charge transfer complexation between maleic anhydride
and both styrene and divinyl ether by the INMR method was reported
by Butler and Campus in 1970. Thus, a semi-quantitative measure
of the electron affinity of 3,3-dimetho:xycyclopropene relative to
maleic anhydride should be available from a comparison of the
equilibrium constants of the two systems.
The NMR studies were done on a Varian Associates analytical l.IMR
spectrometer. The shift of acceptor protons (singlet of 3,3-dimethoxy-
cyclopropene) was observed. The concentration of acceptor was kept
constant while the donor concentration (always in large excess) was
increased. Solutions were prepared at room temperature and the
spectra were obtained at the normal operating temperature of the
spectrometer. Peagents and solvents were all purified by distillation
just prior to use.
I I I I I I
180 190 200 210 220 230
Wa lengthgt, mw
Figure 2. Ultraviolet Spectrum of 3,3-Dimethox:ycyclopropene
The I;:R technique utilizes the linear relationships derived by
Hanna and Ashbaugh:25
1 1 1 1
A = A CD + A
SQA D AAD
obs AD AD
A A A
A = 6 -6 is the difference between the shift of the
obs obs 0'
acceptor protons in completing media and the shift
of the acceptor in uncomplexed form.
D A A
A = 6 -6 is the difference in the shift of the acceptor
AD AD 0
protons in pure complex.
CD is the concentration of the donor (which must always be
much greater than the acceptor concentration in order
that the quotient YAD/yA yD remains constant over the
range of solutions studied and thus 0 = K, the equi-
librium constant of comple:.ation).
In these experiments the acceptor concentration was kept constant
at 0.05 mole liter while the donor concentration was increased
from 0.4 to 8.8 moles liter- A
from 0.4 to 8.8 mols liter By plotting 1/A as a function of
1/CD a straight line was obtained in both cases. (Figures 3 and 4).
The slope of the line and its intersection with the ordinate permit
a first approximation of the equilibrium constant of co.plexation and
of th- shift of acceptor protons in the pure complex. For a more
exact determination of K and A the method of least squares was
applied to equation 1, and the results obtained are shown in Table Ill.
The corresponding values for maleic anhydride complexes are shown for
I I I I I
0.2 0.4 0.6 0.8 1.0
)n 1 -1-1
Figure 3. J11R Study of the Complex Between Dimerhoxycyclopropene
0.2 0.4 0.6 0.8 1.0
Figure 4. :MIIR Study of the Complex Between Dimetho:-:ycyclopropene
and Divinyl Ether
lWhile the values of K are quite low, they appear to be real
and of the same approximate order of magnitude as other weak complexes.
(Butler and Campus report K = 0.008 and AD = 127.5 for the complex
between divinyl ether and fumaronitrile.) As expected, the
electron affinity of 3,3-dimethoxycyclopropene is considerably
less than that of maleic anhydride.
Table III. IMRE Determination of the Equilibrium Constants of Charge
Complex Solvent Temp., OC AD' cps K, 1 M-
DVE:DMCP hexane 38 125.0 0.005
DVE:MA CDC13 24 33.5 0.036
St:DMCP CC14 38 37.0 0.093
St:.A CC1 38 125.0 0.216
1 + 2 Cycloadditions. The Diels-Alder Reaction
The use of cyclopropene derivatives ad dienophiles in Diels-
Alder reactions is a process of synthetic utility as well as
mechanistic interest. The parent hydrocarbon reacts rapidly and
quantitatively with cyclopentadiene at 00 to form the tricyclic
adduct (I). The reaction is stereospecific and only the endo isomer
H H o HH
is formed.26 Substitution at the double bond positions of cyclo-
propene affects the stability and reactivity of the olefin but does
little to alter the course of the Diels-Alder reaction. A single
substituent in the 3-position directs the geometry of the product
but as long as C-3 carries only one substituent the reaction proceeds
smoothly and the adducts have the endo configuration. Thus,
3-methylcyclopropene adds to cyclopentadiene to give only the endo-
anti isomer -nd none of the other three possibilities, while 3,3-
dimethylcyclopropene does not add even at 100. Battiste investigated
the reactivity of a number of derivatives and found that geminally
substituted cyclopropens ae unreactive ard any diene.27
substituted cyclopropenes are unreactive toward any diene.
In 1968 Tobey demonstrated that all tetrahalocyclopropenes
undergo facile 1,4 addition to cyclopentadiene, furan, and 1,3-
butadiene.28 His experiments indicate that electronic rather than
steric effects would account for the observed results. Again the
products were all of endo configuration. The adducts from the
cyclic dienes, however, were rearranged bicyclic derivatives (II).
C1 + 1 - 1 >
C+ Cl C
\ Z Lc i C l -Cl C
Sprouse reacted cyclopropene and some of its derivatives with
1,3-diphenylisobenzofuran (III) and obtained adducts with exo-
stereochemistry.29 Sargent also obtained an exo adduct when he
reacted 1,2-bis(trifluoromethyl)-3,3-difluorocyclopropene with
cyclopentadiene. Sprouse found that reaction of tetrachloro-
cyclopropene with CIII) gave a stable tricyclic adduct (IV) which
could be rearranged to a structure like (II) upon heating. Adduct
IV also had the exo stereochemistry like the other adducts from
c Ci Ci
The only example of a geminally disubstituted cyclopropene
derivative giving rise to the exclusive formation of an exo adduct
with any diene other than III is the report of Monti and Bertrand
in 1970.1 They reacted cyclopentadiene with the ethylene ketal of
3-acetyl-3-methylcyclopropene at 1400 in a sealed tube and obtained
exo-3-methyl-3-acetyltricyclo[126.96.36.199 2' -6-octene ethylene ketal.
The observation by Baucom and Butler8 that the reaction of
3,3-dimethoxycyclopropene with III yielded the exo adduct was
consistent with the work of Sprouse. Baucom also obtained evidence
for adduct formation between dimethoxycyclopropenc and a-pyronel
but this was not confirmed.
If3CO OCH 3
A. Reaction with Butadiene Derivatives
When dimetho:.xycyclopropene was mixed with an excess of butadiene,
isoprene, or 2,3-dimethyl butadiene and the solution allowed to
react at room temperature for several days, the expected adducts
were formed in high yield. Because of the symmetry of the system,
only one isomer was possible in each case. After evaporating excess
diene, the products were analyzed by gas chromatography and the
adduct peak was greater than 90 percent of the crude reaction mixture.
Pure samples were obtained by preparative GC while none of the by-
products could be isolated in sufficient quantity for identification.
CH 30 OCH3 R R R ICH3
CO OCX P ^________ ^l DCH3
1 FF:2 CH 3
VII R1=H; R2 CH3
Figure 5 shows the IMR spectrum of the dimethylbutadiene adduct.
The two methoxy groups are different at 3.3 and 3.46. The ring
protons are centered at 2.16. The singlet at 1.66 is due to the
methyl groups in positions 3 and 4, while the bridgehead protons
give a multiple at 1.36. If the methyl groups are replaced by
hydrogen as in the butadiene adduct, the vinyl protons give a broad
singlet at 5.56 and the peak due to the methylene protons is shifted
just slightly downfield. (Note shaded areas in Fig. 5.)
In the spectrum of the adduct with isoprene (Fig. 6) both of
these groups are present on the double bond and the resulting increase
in coupling produces much broader peaks. Upon expanding the scale,
the two different sets of protons at positions 2 and 5 can be
The mass spectra of these adducts indicated that analogous
fragmentation pathways were followed as would be expected. The
usual procedure for interpreting mass spectra involves localizing
the charge on the hetero atom and then considering logical reorgan-
ization of the resulting ion-radical. The spectra of these adducts
showed an abundant molecular ion which would be predicted by this
procedure of analysis. Clearly the weakest bond in every case is
the 1,7 bond and initial 0-cleavage of this bond in A leads to the
very stable ion-radical B.
I OCH3 1
R2 OCH3 R %CH3
From here, however, it is difficult to rationalize the remainder of
the spectrum. The base peak in every case corresponded to the loss
of a dimethoxymethyl radical. There was also a fairly abundant
peak corresponding to the loss of a methyl radical and fragments
attributable to consecutive reactions of the resulting ester ion
These processes were more readily accommodated by considering
an alternate representation of the ion-radical, C, which is actually
a resonance form of B.
CH DCH <
P2 'CH 2 \CH3
Using these two structures, then, the major features of the spectra
follow very logical rearrangement pathways. The base peak is
adequately rationalized from C as involving a 1,3 hydrogen migration
to give an allylic radical, D, and then a homolytic cleavage of the
6,7 bond to lose the requisite radical and give an ion with extended
conjugation, E. The metastable peak for the transition C (or D) ->
+ 1 .+ H
OCHL -> *3- 3
JCH 2 3
H OCH3 H JCH3
C D E
was just detectable, but it was present. The loss of a methyl
radical from C is also a very logical process and many features
of the spectrum arise from further reaction of the ester ion (scheme
4)--including alternate pathways to the base-peak ion. Further,
the metastable peaks (indicated by in scheme 4) observed for the
proposed transitions are prominent and offer further evidence for
the proposed sequence of reactions. The relative abundances of the
ions resulting from the proposed transitions are given for each adduct
in Table IV.
-CH "1+ R1
OCH 3 3- -CH OH
P CH3 3
c 3 F
m/e 59 E
Scheme 4. Mass Spectral Fragmentation Pattern for Diels-Alder
Adducts Between Dimethoxycyclopropene and Butadienes
Table IV Relative Abundances of Mass Spectrum Peaks from
Adducts Between Dimethoxycyclopropene and Butadiene
2,3-Dimethylbutadiene (VI), and Isoprene (VII).
Transition Ion V VI
molecular ion B or C 66 63
loss of methyl radical F 15 32
loss of CH3 then CH30H G 27 30
loss of C3H 02 E 100 100
loss of H2 from base peak H 60 20
loss of CH4 from base peak I 7 43
loss of cyclohexadiene J 67 33
The reactivity of these adducts was partially evaluated in
a preliminary series of experiments. One interesting reaction
would be hydrolysis of the acetal to give 7-norcarenone derivatives
which would be a unique series of cyclopropanones. This was attempted
by refluxing the 2,3-dimethylbutadiene adduct with aqueous dioxane
containing a drop of sulfuric acid. After eight hours under these
conditions, the adduct was recovered unchanged.
Hydrogenation of the butadiene adduct over palladium on carbon
led to cleavage of one of the exocyclic bonds of the cyclopropane
ring as well as saturation of the olefin. The major product identified
from this reaction was the methyl ester of cyclohexane carboxylic acid.
Likewise attempted bromination of the double bond with bromine
in carbon tetrachloride ruptured the cyclopropane ring as evidenced
by the loss of the signal at 1.36 in the UIMP and the evolution of
hydrogen bromide gas frcm the reaction mixture. The isoprene adduct
was used for this reaction and none of the products were identified.
When the reaction of dimethoxycyclopropene with l-methoxy-
butadiene (50 percent excess) was attempted in THF, there was no
indication of adduct formation. However, mixing the two liquid
reactants (threefold excess of diene) without solvent resulted
in high conversion to the adduct. Analysis of the product fraction
after distillation (by gas chromatography) showed two components
in the approximate ratio of 30:70. Isolation of the major component
by preparative GC gave a product whose I.1R, IR, and elemental analysis
were completely consistent with the expected adduct. In the NMR,
CH3 O CH3 >
the methoxy group at position 2 gave a peak which coincided with the
low-field methoxy of the ketal function. The corresponding methine
proton in position 2 appeared as a poorly resolved multiple at
3.9 to 4.06. All of the other absorptions were shifted downfield
by one to two parts per million from their counterparts in the
adducts obtained from the butadiene hydrocarbons.
The mass spectrum, while not inconsistent with the proposed
structure, was difficult to rationalize. No molecular ion was
evident, but this was not unexpected. The species corresponding
to the molecular ion of the hydrocarbon derivatives should readily
lose a methoxy radical. Further loss of molecular hydrogen to
give the ion M could account for the highest m/e value exceeding
20 percent of the base peak. The base peak appeared at m/e 105
and the second most abundant peak (89 percent of base) was at
m/e 77. Since m/e 77 corresponds to C6H the base peak is likely
OCH3 ICH 3 CH3
K L 3 1.1
to be the oxonium ion 1, although its formation from M is difficult
to rationalize. There was an abundant metastable peak at 56.5
which arises from the transition -- -CJH + CO. The remaining
features of the spectrum indicated that a variety of competing
reactions were occurring and that no single fragmentation pathway
m =5 6.6 I +
+ I + CO
m/e 105 m/e 77
Another possible e::planation for the difficulty in interpreting
the mass spectrum is that the adduct is not thermally stable.
Operating techniques for the mass spectrometer involve storage of
the sample in a reservoir at about 2000 before injection into the
ionization chamber. Thermal stability of samples must therefore
be considered and additional samples of the adduct were required
in order to investigate this possibility.
When another analysis of the distilled product was made on the
gas chromatograph, a lower rate of flow was used for the carrier
gas. Four components were observed this time which were sufficiently
well separated to allow isolation. Peak four was the major component
(66 percent) and was determined to be 2,7,7-trimethoxy-3-norcarene.
Peaks one and two were suspected decomposition products while
peak three might possibly be an isomer of the adduct.
All of the mixture available was subjected to preparative gas
chromatography and these four components were collected. There was
not enough of peak one to permit its identification. Peak two was
identified as methyl benzoate by its NMR and IR spectra. Peak three
was not obtained in pure form and could not be identified. Its
infrared spectrum gave a carbonyl absorbance but comparison with
that of peak two indicated that this was due to methyl benzoate
as an impurity. Instead of the characteristic three or four bands
for a ketal between 1000 and 1200 cm. there was only a single
broad absorption centered at 1100 cm. In the NMR there were the
expected absorbances due to methyl benzoate. In addition there were
6-value Protons Description
5.93 1 multiple
3.33 2 singlet
2.68-1.30 3 broad multiple
1.18-0.96 2 singlet
These results proved that peak three is not an isomer of the
adduct but assignment of a structure cannot be made without further
The thermal stability of 2,7,7-trimethoxy-3-norcarene was checked
by reinjection of a sample of peak four into the gas chromatograph.
In addition to unchanged adduct there were "lights" and peaks
corresponding to one and two. There was none of the compound giving
peak three, however.
Thus, thermal decomposition of 2,7,7-trimethoxy-3-norcarene
produced methyl benzoate as a major product. The main features
of the mass spectrum were then easily explained by considering
this ester to be the species actually being analyzed.
B. Reaction With Cyclic Dienes
Diels-Alder reactions with cyclic dienes could conceivably lead
to derivatives of cyclopropyl norbornene, possibly with exo- stereo-
chemistry if the analogy to the diphenylisobenzofuran study were
followed. Such adduction seemed feasible since the reactivity of
Scheme I'. Diels-Alder Reactions of Dimethoxycyclopropene with
cyclopentadiene (toward maleic anhydride) is much greater than any
of the other dienes used in our work and in the light of recent
reports of successful adduct formation using other geminally substi-
tuted cyclopropenes. However, when dimethoxycyclopropene and cyclo-
pentadiene were allowed to stand together at room temperature no
reaction occurred at all. At 700 the two appeared to react indepen-
dently as the IIR showed a change in the diene pattern, presumably
due to formation of dicyclopentadiene, and the characteristic peaks
of the dimethox:ycyclopropene dimer appeared. There was no indication
of adduct formation. Apparently steric factors inhibit this reaction
as proposed for other systems.
Because of the very facile reaction observed by Eaucom with
diphenylisobenzofuran, it seemed reasonable to expect dimethoxycyclo-
propene to react with furan as well. Eut when a solution of the
cyclopropene in furan was stored in the dark for several days at
room temperature, there was again no evidence for adduct formation.
Instead, there was almost quantitative dimerization of the dimethoxy-
Diels-Alder reaction of dimethoxycyclopropene with tetracyclone
would be expected to give first the bridged carbonyl adduct VIII.
This structure should be readily decarbonylated to give the norcara-
diene IX or possibly its valence isomer X. Compounds with structures
VIII or IX should show two different U-IF. signals for the methoxy
groups and a singlet for the ring protons. The tropone ketal should
show two equivalent protons in the vinyl region but possibly a
Scheme 5. Diels-Alder Reaction of Dimethoxycyclopropene with Tetra-
singlet for the methoxy groups. The actual spectrum (Fig. 7) of
the product isolated showed a six-proton singlet at 36 for the
nethoxy groups and an AB quartet between 6.0 and 6.36 as well as
a 20-proton multiple in the aromatic region. This was clearly
inconsistent with any of the expected structures. Further, elemental
analysis corresponded to a 1:1 adduct without the loss of carbon
monoxide while the infrared spectrum was free of carbonyl absorption.
Treating a sample of this adduct with aqueous acetone and a trace
of acid resulted in the loss of methoxy signals in the 11l1F, and the
appearance of an intense carbonyl peak at 1780 cm. in the infrared
spectrum. The IMIR spectrum also showed a widely separated AB
quartet with a coupling constant, JAB, of 5.5 cps. One doublet
of this system was centered at 7.456 and the other at 6.106.
These observations are consistent with a structure in which
an opened cyclopropene has added across the carbonyl of the tetra-
cyclone. Two modes of addition appear possible, giving either XI
Acid catalyzed hydrolysis of these compounds would give derivatives
of a 3(2H)-furanone (XIa) and a 2(5H)-furanone (XIIa) respectively.
I o C
H h Phe4
Spectral data for the unsubstituted furanones, available from
the literature, are tabulated below and compared with the data obtained
for the hydrolysis product.
Chemical Shift, 6
Compound IR absorbance A-protons B-protons AB' cps
3(2H)-furanone, XIa33 1706 5.70 8.23 2.5
2(5H)-furanone, XIIa 1775, 17453 7.6335 6.15 5.2
H/drolyzed adduct 1780 7.45 6.10 5.5
On the basis of these comparisons, structure XII can be assigned to
the adduct. A similar structure was proposed by Breslow36 in the
reaction of cycloheptenocyclopropene with tetracyclone, but no mechanism
was suggested for its formation. The only spectral data relevant
to the present situation were the infrared absorbances which included
a strong band at 1765 cm.-
Unsubstituted cyclopropenone reacted with tetracyclone in
methylene chloride in the normal Diels-Alder fashion to give tropone
after decarbonylation of the initial adduct. Thus, it appears
that the steric effect imposed by geminal substitution in dimethoxy-
cyclopropene is sufficient to inhibit its reaction with tetracyclone
while the carbonyl group of this dine is sufficiently reactive to
allow an alternate reaction pathway to become operative.
The dimerization of cyclopropenone as well as its reaction
with diphenylcyclopropenone gave 2(5H)-furanone derivatives. The
mechanism suggested for this process was nucleophilic attack on the
carbonyl carbon of cyclopropenone, followed by 1,2 cleavage and
S^ -0 Ph- 0
pY \ / /10 --"
.Ph Ph xC=-C-H Ph H
cyclization. Operation of a similar mechanism in the tetracyclone
reaction with dimethoxycyclopropene would involve nucleophilic
attack on the ketal carbon accompanied by 1,3 cleavage and then
CH30 0 OCH3 OCH 3
-C +-OCH > OCH
/I' 1 HC=CH
Ph Ph/ H
Alternatively, since tetracyclone has an electron-deficient
carbonyl group, the mechanism could be viewed as nucleophilic attack
by the olefin on that carbonyl carbon, followed by 1,3 bond cleavage
3 C 3
t D OCH 3 C
\ / >- >
CH 0 OCH3
Scheme 7. Diels-Alder Reaction of Dimethox',c,clopropene with
,C H 3
A third possibility is that the 1,3 bond of dinethoxycyclopropene
is cleaved first to give either a dipolar of a diradical species
which is then trapped by the tetracyclone carbonyl system. It is
not possible to distinguish the actual mechanism from the data
available from these experiments.
Another interesting possibility for using dimethoxycyclopropene
as a synthetic intermediate is its reaction with tetrazines. The
initial adduct (XIII) should readily lose nitrogen to give the
diazanorcaradiene XIV or possibly the diazatropone ketal XV.
Hydrolysis of this ketal would then give a route to diazatropone.
Sauer and Hinrichs found that cyclopropene and 1,2,3-tri-
phenylcyclopropene both added readily to tetrazines to give the
diazanorcaradiene products. Rearrangement to the diazacyclohepta-
triene system was observed for the triphenyl cyclopropene adduct.
Reaction of dimethoxycyclopropene was attempted with two
tetrazines. Preliminary 1MR evidence indicated that for the case
of 3,6-dicarbomethoxy-s-tetrazine, the diazanorcaradiene was present
in the reaction mixture but the complexity of the mixture and the
paucity of material available did not permit its isolation. There
is no indication of adduct formation at all when 3,6-diphenyltetrazine
and dimetho:,xycyclopropene were reacted for 18 hours in refluxing
Tetrachlorocyclopentadiene affords the interesting possibility
of reacting with dimethoxycyclopropene in two ways. Diels-Alder
reaction could lead to the adduct XVI.
In addition, if the methylene protons of the diene were sufficiently
active, a very unusual fulvene derivative could result from reaction
CH+ 3i -- > + 14CI + H 2 CH OH
at the ketal function. The formrjation of fulvenes has been observed
for reactions of tetrachlorocyclopentadiene with aromatic aldehydes.
Cl -- --=Ar
However, when the two reagents were allowed to stand for several
days at room temperature in'just enough benzene-d6 to effect solution,
the tetrachlorocyclopentadiene was recovered unchanged while the
2+2 Cycloaddition Reactions
A few reactions are known where cyclopropenes undergo cyclo-
additions with formation of a four-membered ring. The initial
adducts often rearrange, however, and only secondary products are
actually isolated. 1 When 1,3,3-trimethylcyclopropene was irradiated
with ultraviolet light in acetone solution containing benzophenone
as sensitizer, the two isomeric tricyclic dimers were isolated in
15 percent yield.40
H C HC3 H3 C CH
3 3 HG CH
HHhv 3 t 3
HC (C 6H5) 2CO H 3C-
H365-2 3 H3
H3C CH3 H3C CH3
Baucom found, however, that contrary to the reaction of 1,3,3-
trimethylcyclopropene, the dimerization of 3,3-dimethoxycyclopropene
was not a photolvtic process but was a thermal reaction probably
proceeding by a two-step mechanism.7
Since the dimerization appears to be a very facile process
undr mild conditions, and in the light of recent observations in
the condensed phase, other cycloaddition reactions of dimethoxycyclo-
propene were studied. A number of interesting and novel systems
could be approached in this manner by the proper choice of olefin.
The charge transfer studies discussed earlier indicate that
dimethoxycyclopropene is an electron-deficient olefin. Its facile
reaction with electron-rich dienes in the Diels-Alder reaction
supports this conclusion. However, the formation of a dimer via
2+2 cycloaddition requires that it react as an electron-rich species
also. In addition, the unusual reaction of dimethoxycyclopropene
with tetracyclone suggests that the presence of a hetero atom in
the reacting system is important. Therefore, 2+2 cycloaddition
reactions were investigated using an electron-deficient carbonyl
species (hexafluoroacetone), electron-rich olefins Cenamines) and
electron-poor olefins (triazolinediones and methyl acetylenedicarbox-
ylate. The results indicate that dimethoxycyclopropene does not
undergo 2+2 cycloaddition reactions with carbon-carbon double bonds
regardless of their electronic character, but that appropriately
activated hetero-olefins react readily.
A. Reaction with Hexafluoroacetone
This reagent is electrophilic in nature and two modes of addition
to dimethox:ycyclopropene could be predicted. The first is analogous
to the tetracyclone reaction, involving opening of the cyclopropene
CH 30 OCH 0 CH 0 OCH
3 3 Lp 3H
A + CF -C-C 3--> E > 0
N P ----
+ C-CF3 H F
^ --- ^" r^ .on
With the tetracyclone adduct in mind, the I;MR spectrum of a compound
with such a structure would be predicted to show an AB quartet for
the vinyl protons and a singlet for the methoxy groups. The tri-
fluoromethyl groups are also equivalent and should appear as a
singlet in the F19 1-,R spectrum. Hydrolysis of such a compound
should proceed readily to give 5-bis(trifluoromethyl)-2(5H)-furanone.
Alternatively, the reaction could retain the three -mebered
ring and terminate by closure at C-2 to give a bicyclic system.
CHCH 3 -
CH30 3 CH30 C-F3 H2 CF3
C 3 CF3
A compound of this structure should have an IM.R spectrum indicative
of non-equivalent methoxy groups and also non-equivalent trifluoro-
methyl groups. The ring protons are also different and coupling
would be expected. Hydrolysis of the ketal would produce a cyclo-
propanone and therefore should be difficult to achieve.
%hen dimetho:ycyclopropene and excess he::afluoroacetone were
mixed and allowed to react for three days at room temperature
there was smooth and complete conversion to a single compound.
The small amount of impurity present was easily removed by treatment
with silica gel.
The NMP. spectrum (Fig. 8) gave the following data:
-- --- ` .-- I-. ~ -- -- .
6 Protons Description
5.66 1 multiple
'1.86 1 broad singlet
3.81 3 singlet
3.50 3 singlet
These absorbances are consistent with the bicyclic structure and the
multiplet at 5.666 was assigned to the ring proton at C-4 because
it is possible to realize long-range coupling with the trifluoro-
methyl groups. The broad singlet at 4.866 was assigned to the other
ring proton and the other singlets were attributed to non-equivalent
The F 19 IR spectrum indicated two non-equivalent trifluoromethyl
groups which experience coupling with each other and with something
else, probably one or more protons. No other features of the structure
could be determined.
In the infrared, there were absorbances between 1300 and 1500 cm.-
characteristic of trifluoromethyl groups and between 1000 and 1200 cm.
for the ketal chromophore. There was also an intense absorbance at
1670 cm.1 which could be in the carbonyl range. An aldehyde could
result from rearrangement of the bicyclic product:
IOCH3 c Cr H3OCH3
CrF 43 OCH C> CH H
JT 3 CF 0
However, no aldehyde proton was observed in the IHP. and if this were
the structure, the methox:y groups would be equivalent.
The mass spectrum was very clean, showing only two fragments
with an intensity greater than 20 percent of the base peak. A
very small molecular ion peak was detected at m/e 266. The base peak
occurred at m/e 197 corresponding to the loss of a trifluoromethyl
radical. Other fragments were due to loss of a metho:-:y radical
(m/e 235) and to the trifluoromethyl ion (m/e 69).
Hydrolysis of a sample of the adduct was also attempted in
order to provide additional structural information. After 17 hours
in refluxing aqueous dio:-:ane containing a trace of acid, the adduct
was recovered unchanged.
Thus, the second mode of addition appears to be operative and
the adduct produced was 3,3-bis(trifluoromethyl)-5,5-dimethoxy-2-
H q 3/CF3
B. Reaction with Enamines
Enamines are electron-rich olefins and should be ideal for reaction
with an electron-deficient system such as dimetho::ycyclopropene. A
number of cyclopropenone derivatives have been reacted with a variety
of enemies to give both "C-C" and "C-N" insertion.1-4 The
former, observed only in low yield, results from 2+2 cycloaddition
and then reopening of the four-membered ring. The latter involves
cleavage of the 1,2 bond of the cyclopropenone and insertion into
the C-1i bond of the enamine. Eicher and Bohm obtained stable
ketaines from the reaction of phenyl-methyl-cyclopropenone and
diphenylcyclopropenone with acyclic enamines. '5,6 These last
reactions require the participation of the carbonyl oxygen of the
cyclopropenone and would be impossible for dimethoxycyclopropene.
However, reactions analogous to the studies of Dreiding would provide
interesting results regardless of the mode of insertion and the effect
of the ketal instead of a ketone could be valuable from a theoretical
point of view.
Two enamines were selected for reaction with dimethoxycyclroropene.
They were N,IN-diethylstyrylamine and l-(2-methyl-l-propenyl)-pyrrolidine.
Ph\ CHI H
H C =C CH "1 /
C2H15 C113 I
The reactions were conducted neat, at 530, under nitrogen. In the case
of the styrylamine, decomposition of the dimethoxycyclopropene occurred
but the enamine was unaffected. Both the pyrrolidine enamine and
the dimethoxycyclopropene remained after 16 hours of reaction time
in the second case.
These results indicate that dimethoxycyclopropene does not
add to electron-rich olefins.
C. Reaction with Triazolinediones
4-Phenyl-A -1,2, '-triazoline-3,5-dione (XVII) is a cis-locked
azodicarbonyl compound and a very electrophilic cyclophile.7
Cycloaddition of XVII with simple nonoolefins is limited to 1,2-
addition leading to diazetidines. For example,7 the diazetidine
XVIII was obtained from PhTD and indene.
+ o -
0 J -Ph
The proposed dipolar intermediate was trapped with water.
Reaction of XVII with alkenylidenecyclopropanes (e.g., XIX)
gave two 1:1 adducts in stereochemically pure form. A concerted
cycloaddition across carbon atoms 2 and 4 was implicated.
3 C 6 5 ,C3
l> C=C 4-+
2 3 "H3
133C Ci3 0
H2C I- Ph
The same products could arise from a dipolar intermediate such as
XX by two different modes of isomerizing the cyclopropyl cation
portion, but their data argue against its intervention.
R =-- 0
Dimetho>ycyclopropene, reacting as a simple olefin in a 2+2
cycloaddition,would give the very unusual diazetidine XXI. The steric
+ :'. I---
requirements for this adduct should be no more severe than those of
the dimer. Participation of a dipolar intermediate such as XXII
would very likely lead to rearranged products resulting from opening
of the cyclopropyl ring (XXIII).
H 30 OCH3
-- > 1-- _h
XX I II
The alternate ring closure at the oxygen atom would produce a cyclo-
heptadiene with one trans double bond--a prohibitively strained system.
When XVII was reacted with dimethoxycyclopropene, however,
neither of these adducts was formed. Instead, a spontaneous copoly-
merization reaction occurred giving a product with a number average
molecular weight of 12L'4 as determined by vapor phase osmometry.
Elemental analysis was inconsistent with either a 1:1 or a 2:1
copolymer composition. Wagener has obtained evidence for incor-
poration of additional triazolinedione into the copolymer, presumably
by nucleophilic attack of the dipolar intermediate on the monomer.
Postulating a product consisting of four units of 1:1 copolymer and
one extra phenyltriazolinedione molecule results in a molecular
weight of 1275 and elemental analysis is consistent with this composition.
Assignment of the structure of the copolymer rests mainly on
the infrared spectrum. Carbonyl peaks at 1720 and 1780 cm.- indicated
that reaction had occurred through both the nitrogen and the oxygen
of the triazolinedione moiety. In unreacted XVII the carbonyl
double appears at 1780 and 1760 cm. Reaction through the oxygen
atom leads to the formation of a C-N double bond in the product and
an absorbance at 1620 cm.- can be assigned to this chromophore.51
Also observed are bands at 1655 cm.-1 (C-C double bond, cis disub-
stituted) and the characteristic bands for a ketal between 1000 and
The !IMR spectrum provides much less information. There is a
broad absorption around 7.56 and a very broad doublet centered at 3.56.
The relative areas of these two regions is 8:6.
Thus, the most likely structure for the major portion of the
product is C OCH
3 \/ 3
0< i--- -
with some units going through both nitrogen atoms and with about one
extra XVII unit for each four units of the copolymer.
The dipolar intermediate in the copolymerization of XVII and
ethyl vinyl ether has been trapped with acetone and with cyclo-.
hexanone. No evidence for a similar termolecular adduct could
be detected when XVII and dimethoxycyclopropene were reacted in acetone.
H-CH 3 / C3
O JN--0 C 330
I II 3 -3 CHIV
Trapping of the 1,5 dipole (XXIII) leading to a seven-membered ring
in the product was also considered possible but only polymer was
S OCOCH3 N=J
X CH3 XXIV
It was felt that the less reactive N-nethyltriazolinedione (XXIV)
night be a more effective substrate for interception of a dipolar
intermediate and this possibility was investigated. To provide a
basis for comparison with the PhTD e::periments, this reaction was
also run in methylene chloride. The results in both solvents were
the same. There was no evidence for monomeric species in either
case, although the yield of purified polymer product was lower than
in the XVII study. Molecular weights were determined by VPO and
found to be about 990 when methylene chloride was solvent and 760
when the reaction was run in acetone. Elemental analyses suggested
the monomer ratio in the produce was 1.3:1 (XXIV:dimethoxycyclopropene)
in both instances and spectroscopic analyses were consistent with
the ring-opened structure proposed earlier.
D. Reaction with Dimethylacetylenedicarboxylate
The cycloaddition of dimethoxycyclopropene in dimethylacetylene-
dicarboxylate was attempted as both a thermal and a photolytic
process. Because of the interesting possibility for a 2:1 adduct,
excess acetylenedicarboxylate was used and the reactions were run
without solvent. Two different routes can be visualized, depending
upon whether or not the initial 1:1 adduct forms a four-merrbered ring.
S. CH CH CH
R-C-C-R + 3 3 R --
OCH3 R R
CH O OCH
C3 O3 R-C-C-R
R = -CO CH3 R
CH 3 OCH
CH30 OCH3 HCCH-d 3
S P.-C-C-iR ,n
-;;:~^ / \np
The thermal reaction was patterned after the room temperature dimer-
izations in which the reagents are allowed to stand together for
several days under ordinary laboratory conditions. The photolytic
process was attempted by exposing the mixture to the sunlight in
a quartz UV cell. In both cases the course of the reaction was
followed by periodic sampling for !;I-R analysis and in both experiments
decomposition of the dimethoxycyclopropene was observed with no
indication at all of adduct formation. It is likely that traces
of free acid catalyzed the decomposition.
Reactions With Amines
The observation by Baucom that nucleophiles apparently add
very readily to dimetho::ycyclopropene led him to the preparation
of 1 ,-dimethoxy-2-dimethylaminocyclopropane.
+ (CH3)2JH >
3 \ N(CH3
If a similar reaction could be effected using diallylamine
as the nucleophile, the very unusual 1,6 diene produced would afford
the possibility of producing a polymer with pendant cyclopropane groups.
+ (CH2CHCH-CH2 2H
Modification of polymer properties might then be expected to arise
by cross-linking through or chemical reactions of the very reactive
However, exposure of these reagents to conditions that led
to quantitative production of the dinethylaminocyclopropane failed
to give any indication at all of the addition compound. Diallyl-
aminocyclopropane would be expected to show a multiple near 1.36
as well as a methoxy doublett". Neither of these features was
observed in the spectrum of the crude reaction mixture while typical
decomposition patterns dominated. Distillation of this material
failed to produce any fractions indicative of the anticipated
course of reaction.
This result raised the question of why the two amines should
react so differently and suggested that reaction with a variety
of amines might reveal something of the electronic character of
When dimetho::ycyclopropene was added to excess diethylamine
and the mixture stored at room temperature for several days, the
major product formed (60 percent) and the only one isolated was 1,1-
dimethoxy-2-diethylarriinocyclopropane. The pure product was isolated
by preparative gas chromatography but there was not enough sample
available to allow collection of the two observed minor components.
CH30^ ___CH3 CHOOCH
CH30 CH3 + (C2115)2rH CH3 CH
The NMR spectrum (Figure 9) gave the following data:
6-value Protons Description
3.42 3 singlet
3.32 3 singlet
2.89-2.51 4 quadruplet
2.14-1.90 1 multiple
1.18-0.94 6 triplet
1.05-0.75 2 multiple
The two singlets at 3.42 and 3.326 were assigned to the nonequiv-
alent methoxy protons. The quadruplet at 2.706 and the triplet
at 1.066 were assigned to the diethylamino protons. The absorbances
at 2.12 and 0.906 were assigned to the cyclopropyl ring protons.
The mass spectral analysis was completely consistent with the
proposed structure and produced a fragmentation pattern analogous
to that observed by Baucom for the dimethylamino derivative.
Continuation of the series to di-n-propylamine was next attempted
and the course of the reaction was found to be considerably more
complex. As in the case of diallylamine, the crude reaction mixture
did not appear to contain any of the addition product. However, this
time one fraction from the product distillation showed INMR absorption
which could be assigned to the nonequivalent methoxy groups.
Analysis of this fraction by gas chromatography indicated two major
components comprised about 85 percent of the mixture and that these
components were present in a ratio of 40:60. Pure samples of each
one were then obtained by preparative gas chromatography.
The first component gave absorptions in the NMR at 3.45 and 3.356
for the non-equivalent methoxy groups and a one-proton multiple at
2.08 characteristic of the lone proton at C-2 of the cyclopropane
ring. The absorbance of the geminal protons on the ring overlapped
with the characteristic multiplets for ths propyl groups and could
be detected only by analysis of the integration areas. The infrared
spectrum of this component was almost superimposable on that of the
diethl,,laminocyclopropane which confirmed that it was 1,1-dimethoxy:
CH30 3 CH 3 OCHa new
+ (CH3-CH2-CH22 H + compound
The NtTR spectrum (Figure 10) of the second component gave the
6-value Protons Description
3.65 3 singlet
2.97-2.62 2 multiple
2.56-2.17 6 multiple
1.78-1.17 4 multiple
1.07-0.66 6 multiple
In the infrared spectrum there was an intense carbonyl absorption
at 5.72u which was interpreted with the corroboration of the C-0
stretching at 8.30p as being indicative of an ester. Elemental
analysis gave an empirical formula of C 1H2 102 and in the mass
spectral analysis there was a molecular ion at m/e 187 (13 percent
of the base peak). The base peak was at m/e 158 and another fragment
appeared at in/e 114 which was 65 percent of the base. These data
indicated that the second component of the mixture was n1,iI-di-n-propyl-
amino-f-alanine methyl ester (X:XV).
(e) (d) (c)
CH 3-CH -CH
3X2 -CH -CH -OCH
CH3-CH2-CH2 (c) (b) (a)
(e) (d) (c)
Assignments of the UMR absorbances of Figure 10 to the protons of
this structure can be made as follows:
6-value Protons Assignment
3.62 3 a
2.97-2.62 2 b
2.59-2.17 6 c
1.67-1.17 4 d
1.07-0.66 6 e
The mass spectral fragmentation pattern arises from the expected
two modes of cleaving a C-C bond next to the nitrogen which is
directing the fragmentation to the virtual exclusion of all other
possibilities because of its greater electronegativity.
The change in reactivity through this series of amines cannot
be attributed to differences in basicity since all are of nearly
equal value. Aromatic amines are considerably less basic than
CH- CII -CH "'C-'C CH C-OCH
3 2 2 2 2 3
-C HO i
-C3H502 1 2 -CH
C3} I7 +
i=CH2 CH -=I-CH -CH -C-OCH
/32 C12 2 3
C3 7 3
m/e 114 m/e 158
aliphatic amines so a study of their reactivity could provide additional
information about the chemistry of dimethoxycyclopropene.
Diphenylamine was selected for first evaluation since examination
of the products should be less complicated due to the presence of
only aromatic protons in the IJ.R. Because this amine is a solid,
a solvent (CH Cl ) was required. The amine was used in only 10 percent
molar excess due to its low bulk density and the resulting high
volume of material required.
Analysis of the crude reaction product by NMR did not reveal
the presence of cyclopropyl protons, but some adduct formation was
indicated by the characteristic absorbances of two non-equivalent
methoxy groups. A large new singlet at 3.186 was also observed
along with several smaller peaks.
Separation of this ni:-:ture by silica gel chromatography was
attempted but when the sample was placed on the column it immediately
experienced a vigorous exothermic reaction. Development of the
column yielded a weight recovery of approximately 65 percent which
appeared to be mostly a single component plus some free diphenylamine.
A second chromatography yielded pure samples of this component.
The NMR spectrum (Figure 11) gave the following data:
6-value Protons Description
7.32-6.65 10 multiple
4.15-3.80 2 multiple
3.51 3 singlet
2.75-2.38 2 multiple
This appeared to also be an amino acid ester--the methyl ester of
N,N-diphenyl-B-alanine. The low-field absorbances were assigned
to the aromatic protons and the singlet at 3.56 to the methyl group
of the ester. The two 2-proton multiplets form an A 2B pattern
which would be expected for the four methylene protons of structure
The infrared spectrum showed carbonyl absorption at 5.721 and strong
asymmetrical C-C-0 stretching at 8.50U indicating an ester. Elemental
analysis gave an empirical formula of C16 H17NO and the mass spectral
analysis was consistent with the proposed structure. The molecular
ion at m/e 255 was 40 percent of the base peak which appeared
at m/e 182. The next largest fragment was 16 percent of the base
at m/e 77 (C6.H ) and all others were less than 10 percent.
C 6H5 +. 0 C3H50 C6H5, t
R "0 -CH 6 =CH2
-CH2- CH2-C-OCH > /UCH
/ 2 3 / 2
m/e 255 m/e 182
Identification of the intermediate which was isomerized on
silica gel was next attempted using the second half of the same
mixture. The first procedure tried was to cool a pentane solution
of the material to -780. The oil which separated confirmed that
the non-equivalent methoxy groups were present without cyclopropyl
protons but no purification of components was realized. After a
few days, these samples were blended back together and the NIHR
spectrum was taken. Very little of the absorbances assigned to
the methoxy groups remained and the singlet at 3.186 dominated
This material was dissolved in methanol and then divided into
two portions. The first portion was allowed to stand at room temper-
ature in an open flask to slowly evaporate solvent in an effort to
effect crystallization. No crystals ever formed and when all of
the solvent had evaporated the resulting oil was again analyzed
by NMR. The result was a pure sample of !I,N-diphenyl-6-alanine
methyl ester with no sign of free diphenylamine.
The second portion of the methanol solution was treated with
a few drops of water and an oil separated. It was dried under
vacuum and analyzed by IM'iR (Figure 12). Absorbances due to the
alanine ester were obvious. An absorbance at 2.30-1.936 appeared
to be one portion of a second A2 2 pattern. The second portion
Figure 12. NMR Spectrum of the Intermediate Leading to
would be expected in the range of 4.10 to 3.406 but absorbances due
to the alanine ester also appear in this range. However, if that
portion of the integration attributable to the known compound was
subtracted from the total integration in this area, the difference
was exactly equal to the integration of the 2.30-1.936 multiple.
Thus, another A2B2 system seemed likely. If the singlet at 3.185
was due only to a compound containing this A2B2 system, then this
singlet was equivalent to nine protons. In the aromatic region the
absorbance was due to the S-alanine derivative and some free diphenyl-
amine as well as the unknown intermediate. Subtracting the known
compounds from the total left thirteen aromatic protons for the
intermediate which implied that it contained only one diphenylamiino
group. When this IMIR sample was poured onto silica gel and then
extracted with ether the next day, N,II-diphenyl-3-alanine methyl
ester was again obtained with some contamination by the free amine.
In the infrared spectrum of the sample rich in the intermediate there
were few absorbances not present in the alanine derivative although
some intensities were different. Weak absorbances at 8.90, 10.55,
and 10.95p and a strong absorbance at 13.30p with a shoulder at 13.55p
appeared to be unique with the unknown intermediate. The carbonyl
absorption was considerably less intense as would be expected if
the intermediate lacked this chromophore. The absorbances due to
the aromatic system were only slightly less intense.
No further purification of this sample could be effected and
continued manipulation saw it all be slowly converted to the -
alanine ester. Thus, no positive identification of the intermediate
was possible. One structure which could satisfy the observed phenomena
is the ortho ester XXVI although its formation would be difficult
C6 5 OCH
N-CH -CH -OCH
2 22 3
Baucom observed the formation of methyl orthoacrylatel7 under
certain conditions during the preparation of dimethoxycyclopropene
and found that the orthoester moiety gave a nine-proton singlet
at 3.226. In the present case the predominant single occurs at
Although structure XXVI is consistent with the main features
of the spectra, its formation would require a very unlikely series
of reactions. First, a portion of the dimethoxycyclopropene would
have to react with diphenylamine in an exchange reaction of some
sort at the ketal center to produce methanol. Then all of the
methanol would have to react with additional dimethoxycyclopropene
to give methylorthoacrylate. This reaction has been observed
but a number of other products are also obtained. Addition of di-
phenylamine to the methyl orthoacrylate52 would then lead to compound
XXVI. The probability of such a sequence occurring is very low and
there was also no evidence for any of the by-products to be expected
A more likely mechanism for the formation of the alanine ester
involves a ketene acetal as the intermediate. The simple addition
H3CO OCH3 H3CO OCH3 H CO CH
R "H + 7 >
2H 2 H
reaction is nucleophilic attack at C-i by the electron pair on
nitrogen with production of negative charge at C-2 where protonation
completes the addition. Alternatively, 1,3 cleavage could occur
followed by reorganization of the electron system and protonation
at C-I to give the ketene acetal.
CH 0 OCH CH 0
3 3 3 H Uj CH 3C- C H-
-r I' ?, 2r-CH a2-a Hsr
2 F. 2 1 --U 2Cr H
S2>\H H OCH3
Hydrolysis of the ketene acetal would then give the alanine esters
observed. One possible reason for the difference in reactivity in
the series of aliphatic amines is the increasing bulk of the
substituents which could hinder the approach to C-2 to complete
the simple addition process.
Corey53 reported that the lJMR spectrum of l,l-dimetho::yethylene
consisted of two singlets at 3.08 and 3.466 in the ratio of 3:1.
These absorbances are in the general region of the spectrum observed
for the intermediate but more detailed comparison is not possible
without a purer sample of the compound.
The possibility of further reaction of the cyclopropyl amines
was also considered. At the same time it was possible to investigate
the case of adding amines to methyl orthoacrylate. A sample of
dimethoxycyclopropene containing eighteen mole percent methyl
orthoacrylate was treated with excess dimethylamine for four days
at room temperature. Following evaporation of the amine, the crude
reaction product was analyzed by NMR. The only absorbances observed
were those for l,l-dimethoxy-2-dimethylaminocyclopropane and eighteen
mole percent methyl orthoacrylate.
Diphenylamine was then added to the U11R tube to attempt a
ring opening reaction. No change at all was observed after 3 hours
at 40-450 and after eighteen hours at 60-650 there was only partial
reaction of the orthoacrylate. Thus, the cyclopropyl amines appear
to be stable once formed and it seems unlikely that methyl ortho-
acrylate would add amines under the conditions where formation of
the alanine esters was observed.
Summary and Conclusions
The objectives of this research program were (1) to improve
and expand the reactions leading to 3,3-dimethoxycyclopropene,
(2) to further evaluate its physical and spectral properties, and
(3) to establish its chemical reactivity toward a variety of systems.
Significant improvement was realized in both yield and operational
convenience by modifications in the reported procedures. Attempts
to expand the scope of the reactions involved indicated that they
are very specific for the reagents used and that other materials
behave quite differently in the same system.
The ultraviolet spectrum of dimethoxycyclopropene shows no
absorbance above 210 mp. The observed spectrum was a triplet
centered at 187 with = 1670. The NMR absorbance for the
cyclopropene protons was found to be sensitive to the electronic
nature of the solvent and charge transfer studies were conducted
using the technique of Hanna and Ashbaugh. The results of these
studies indicated that dimethoxycyclopropene is an electron-deficient
olefin in that weak complexes were formed with electron-rich species.
The donors in this case were sty,'rene and divir.,rl ether. The use
of more powerful donors such as enamines would be a logical extension
of this work. It would be especially interesting to see if stronger
charge transfer complexation followed the use of this particular
type of donor since enamines failed to give chemical reactions with
As a dienophile, dimethoxycyclopropene is only moderately
reactive. Adduct formation was realized in excellent yield with
acyclic diencs, but an extended reaction time of a week or more
was required. No reaction at all could be effected using cyclic
dienes, presumably due to the steric interference of the geminal
substitution. The unusual adduct observed to form between tetra-
cyclone and dimethoxycyclopropene indicates that the olefin is also
an electron source since the most likely mechanism for its formation
involves nucleophilic attack on the carbonyl carbon of the diene.
Formation of a normal Diels-Alder adduct was indicated but not
proved in the reaction of dimethoxycyclopropene with 3,6-dicarbo-
methoxy-s-tetrazine. The tricyclic adduct in this case is [188.8.131.52]
instead of [184.108.40.206] which should ease the steric requirements,
but it is also interesting to note that the diene system includes
hetero atoms. The importance of this feature would be an interesting
program of study.
Dimerization of dimetho::ycyclopropene, observed early in the
program, is still the only successful 2+2 cycloaddition reaction
with a carbon-carbon double bond. The product of the reaction with
hexafluoroacetone is a four-membered ring from what is formally a
2+2 cycloaddition reaction, but the most likely mechanism is not
an electrocyclic process. As in the case of tetracyclone, the
mechanism probably involves nucleophilic attack on the carbonyl
group. In the hexafluoroacetone reaction, however, the cyclopropane
ring is preserved since ring closure at C-2 occurs instead of
Cycloaddition with electron-rich olefins should be a facile
process for dimethoxycyclopropene if the conclusions about its
electron-deficient character are correct. The lack of reactivity
toward enamines was therefore unexpected. The reason for this
result is probably steric rather than electronic, however, since
no adduction was observed with electron-poor species either.
Dimethylacetylenedicarboxylate gave no indication of reaction at
all. With phenyl triazolinedione, a very electrophilic cyclophile,
none of the diazetidine which would result from cycloaddition was
observed. Instead a copolymer was formed, presumably by nucleophilic
attack of the olefin on the triazolinedione and coupling of the
resulting dipolar intermediate. It is interesting to note also
that again the reactive center contains a hetero atom. Other
reactive species involving oxygen, nitrogen, or sulfur should
provide an enlightening area of investigation. Possibly effective
reagents include isocyanates, thioketones, diazoalkanes, isothio-
cyanates, and diimides.
Nucleophilic addition of secondary amines to dimethoxycyclc-
propene supports the conclusion that it is an electron--deficient
olefin. Simple addition was observed for lower aliphatic amines
to give 1,l-dimethoxy-2-dialkylaminocyclopropanes. As the size of
the substituents on the amine increased, however, a second process
became increasingly important. This process involved destruction
of the three-membered ring and formation of 1N,1-dialkylalanine