PREPARATION AND PROPERTIES OF SOME
RAYMOND SCOTT PYRON
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
A C1UvOWL EDGII''EN TS
For professional and personal assistance above and
beyond the call of duty as a research director, I thank
Dr. W. H. Jones. The patience of my committee was also
appreciated. Further acknowledgments must also be tendered
Mrs. Thyra Johnston, my typist, for her conscientious work
on this dissertation under difficult circumstances, to
Mrs. Penney Ansell for her usual excellent drawings, and
Darrel Muck, a truly fine laboratory partner and friend.
Financial assistance from the Sloan Foundation and
financial and moral support from the Danforth Foundation
are gratefully acknowledged.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . . . ii
LIST OF TABLES. . . . . . . . .. iv
I. CYCLOPROPErYLIDENES AND HETH-YLETNEYCLO-
PROPENES. . . ........... . .. 1
II. REACTION OF A NETHYLENECYCLOPROPENE WITH
DIAZOETHANE . . . . . . 35
III. EPERIENTAL . . . . . . . 50
LIST OF REFERENCES. . . . . . . . . 95
BIOGRAPHICAL SETCH . . . . .. . 100
LIST OF TABLES
1. Electronic Spectra of 19c and Model Compounds 18
2. Electronic Spectra of 19b . . . . . 23
3. Nuclear Magnetic Resonance Spectra of Some
Methylenecyclopropenes and Their Dihydro
Derivatives . . . . . . . .. 28
4. Ultraviolet Absorption Data for Some Styrenes
and Pyrazoles . . . . . . . . 45
CYCLOPROPENYLIDENES AND METHYLENECYCLPOPROPNES
Although interest in-the area of non-benzoid aromatic
compounds* in general has remained high, there has been a
special emphasis recently on cross-conjugated, non-alternant
i-systems. For example, although fulvenes were prepared as
early as 1900 (7) they are still objects of fairly intense
theoretical and experimental research. The nature of the
fulvene system'may be qualitatively viewed as resulting from
resonance contributions from homopolar structure la and di-
polar structure lb. Appreciable contribution from lb is not
too surprising when one recalls the strong tendency of un-
saturated five-membered rings to support a negative charge.
For reviews see references 1, 2, and 3, and
especially the more recent 4.
See references 4, 5, and 6, and the references
From this viewpoint, we would expect that substituents
capable of stabilizing positive charge at the 6-position
would increase the stability of the system and give it less
olefinic character. In this case, the predictions based on
the simple picture have very closely been paralleled by a
large number of theoretical calculations, including both
simple Hickel molecular orbital calculations (2) and more
sophisticated studies based on, for example, self-consistent-
field (SCF) approaches (6,8,9). The availability of several
classical synthetic approaches (7) and continued work on
better methods/ has fostered preparation of numerous
derivatives for comparison of their properties with theo-
retical predictions. The correlation has usually been
Heptafulvene (2) is another cross-conjugated system
of interest. Contributing structure 2b again possesses the
stabilizing six-ic-electron system, but in this case the
result is a positive charge on the ring. In contrast to the
history of fulvenes, the first theoretical studies (11) of
1For example, see reference 10.
heptafulvenes appeared almost concurrently with the report
of a dibenzoheptafulvene (12) in 1949. According to later
more refined calculations the parent system .(2, R1 = R2 = H)
was predicted to have only a small stabilization energy (7)
and a small dipole moment directed away from the ring (6).*
The prediction of low stabilization energy was consistent
with the high reactivity of heptafulvene itself which was
synthesized and reported in 1960 by von Doering and Wiley
(13). In contrast to the parent system, however, derivatives
possessing groups capable of stabilizing negative charge on
the exocyclic carbon might be expected to have greatly
lowered energy and increased dipole moment over that ex-
pected for classical Kekule structure 2a. This expectation
has also been confirmed by the syntheses of several quite
stable heptafulvenes with strong electron withdrawing
groups, such as cyano groups. (_4,1). Vinylheptafulvene, a
derivative with properties lying between the parent system
and these latter stable systems, has also been recently
In addition to the single ring non-alternant hydro-
carbons, those with two rings joined by a double bond are
also significant. Two examples of this type are fulvalene
Julg's self-consistent-field (SCF) calculations
have, however, predicted that the small dipole is directed
toward the ring (7).
(3) and heptafulvalene (4). Because of the symmetry of
Q-0 G0 - asH
3a 3b 3c
4a 4b 4c
these molecules they can have no dipole moment, and although
dipolar structure 3b (and 3c) does contain the stabilized
cyclopentadienide-type ring, the presence of the high-energy
positively charged five-membered ring must be noted. From
our very qualitative view we might suspect very little
contribution from 5b and 3c and consequent high polyene
character. Similar considerations apply to heptafulvalene
(4). Theoretical calculations have also appeared for these
systems.* As might be anticipated, the synthetic problems
are great. Although von Doering has reported the synthesis
of parent systems 3 and 4 (16), the only derivatives which
have been fully reported are polybenzo derivatives (7,17,18).
However, atwo-ring system with quite different pre-
dicted properties is the mixed fulvalene, sesquifulvalene
(5). Because dipolar structure 5b contains both the stable
For some recent ones, see references 6 and 9.
tropenylium ion and cyclopentadienide ion entities, larger
dipole moments and stabilization energies than in other
fulvalenes might be anticipated. In 1951 Tinker (12)
postulated 5 as the azulene analog of biphenyl. HUckel
calculations predict a dipole moment of about 12 D. But
even self-consistent field methods, which tend to give
among the lowest calculated moments, predicted a significant
2-3 D for the dipole moment of 5 (6,9). The Hackel delocali-
zation energy was found to be 0.33 3 per i-electron (9).
Report of a tetrabenzo derivative of 5 appeared in 1953 but
interpretation of the data is somewhat confused due to the
probable non-planarity of the system (_). Several iso-n-
electronic heterocyclyclic analogs of sesquifulvenes have
been reported since the first in 1956* and rather large
contributions from structures analogous to 5b were suggested
by observed dipole moments (25). Work on the carbocyclic
system 5, however, has not proceeded as rapidly. Prinzbach
first reported the synthesis of sesquifulvalene (in solu-
tion) and simple derivatives in 1961 (26,22) and reported
See reference 20, 21, 22, 23, and 24 for recent
examples and leading.references.
chemical, nuclear magnetic resonance, and ultraviolet and
visible spectral data for several derivatives in later
communications (28,29). He suggested from his data that
homopolar structure 5a alone was quite adequate to explain
his results (28,29), although the basis for this conclusion
has been questioned (2). Hopefully, more data will appear
soon on this highly significant system.
However, despite the relative wealth of information
that had become available relating to these many systems,
at the inception of the research reported in this disserta-
tion, there was little information related to three of the
simplest possible cross-conjugated, non-alternant systems:
methylenecyclopropene (or triafulvene) (6), cyclopropenyli-
denecyclopropene (triafulvalene (7), and cyclopropenylidene-
cyclopentadiene (8). In fact, at that time, although several
R -1 R1
il 2 R2
related theoretical studies had appeared,* (30,L1,2) only
four examples of compounds containing a methylenecyclo-
propene (6) entity had been reported (3_5,35,56) and no
reports of successful** preparations of derivatives of 7 or
8 had appeared. The preparation of two quinocyclopropenes
(9) were reported by Kende (L4) in 1963 and two-4carbalkoxy-
methylenecyclopropenes (10a and lOb, respectively) were
reported by Battiste (35) and by Jones and Denham (36) in
1963. Subsequent reports of successful preparations of a
-0 9a, x = H
>- 9b, x = Br
9 a10, R1 = C02C2H5, 2 = H
S 10b, R1 = C02CH, R2 =
>) 2CH2C02CH .
2 10c, R1 = R2 = CN
dibenzoquinocyclopropene (52), and two 4,4-dicyanomethylene-
cyclopropenes (lOc (38) and 11 (39)) have also appeared. In
a conference report (29), Prinzbach recently alluded to the
synthesis of "prototypes" of the cyclopropenylidenecyclo-
pentadiene (8) system, which he calls "calicenes."
For a more recent example see reference 33.
An unsuccessful attempt to prepare a derivative of
8 was briefly mentioned in reference 20,
HGckel calculations on the triafulvalene system (U)
give its delocalization energy as 0.24 0 per t-electron
(32). The small predicted stabilization energy and the
obviously high degree of strain in the molecule suggest
that the system would be highly unstable and would pose a
formidable synthetic problem.
Although the calculated delocalization energy per
t-electron for methylenecyclopropene (6) is also only about
0.24 P per t-electron in the Hickel approximation (32),
some derivatives have been prepared as mentioned above. As
in the analogous heptafulvene series, the stable methylene-
cyclopropenes which have been reported have electron-
withdrawing groups in-the 4-position.
The delocalization energy of cyclopropenylidene-
cyclopentadiene by the HGckel method was found to be 0.37 P
per T-electron (32). Inspection of the dipolar canonical
forms 5b and 8b and awareness of the similarities of the
three- and seven-membered unsaturated ring systems suggest
that mixed fulvalene 8 might bear striking resemblance in
properties to the interesting sesquifulvalene (5). We
therefore undertook to synthesize derivatives of cyclopropenyl-
idenecyclopentadiene (8) and hoped, in the process to develop
a general method for the preparation of methylenecyclopro-
Results and Discussion
Synthesis of a prototype.--After a survey of the
existing preparations of methylenecyclopropenes, we decided
to attempt to find a more general method that might be
applied to the synthesis of cyclopropenylidenecyclopenta-
dienes. A method which appeared especially attractive was
proton removal from an appropriately substituted cyclo-
propenium salt. This method has been successfully applied
several times in the heptafulvene series (,15,40), and was
R1 R1. R2
R RI R3
the method used by Kende (34) to prepare his quinocyclo-
propenes (9) from phenolic substituted cyclopropenium ions.
F6hlisch and Burgle (37) have quite recently reported the
preparation of a dibenzoquinocyclopropene by a route which
presumably involves the same steps.*
These authors found that reaction of diphenylcyclo-
propenyl perchlorate with anthrone in acetic acid and
pyridine yielded iii. In their case, it is believed that,
following initial formation of i, a second molecule of the
cyclopropenyl cation abstracts a hydride to give cation ii
Successful application of this method calls for
satisfaction of three basic requirements. First, the
appropriate carbonium ion must be available. Second, the
rate of proton removal must be fast relative to nucleo-
philic attack on the positively charged ring. Third, the
resulting methylenecyclopropene must be stable to the
conditions of proton removal.
As a prototype system for application of this method,
the known 2,3-diphenyl-4-carbethoxymethylenecyclopropene
(lOa) was selected. This compound has recently been pre-
pared by Battiste (35) through the reaction of the
appropriate phosphorous ylid with diphenylcyclopropenone.
which then loses a proton (presumably to the pyridine).
Q H BF LiCH2CO2C2H5
As a potential source of the necessary precursor
cation 15, 1,2-diphenyl-3-carbethoxymethylcyclopropene (15)
was synthesized in the 27 per cent yield by reaction of cold'
ether solution of ethylc-lithioacetate with a slurry of di-
phenylcyclopropenyl tetrafluoroborate (12) in methylene
chloride at -75. The lithio-compound was prepared from
n-butyllithium and ethyl bromoacetate in ether at the same
temperature. Aqueous work-up of the cation-anion condensa-
tion mixture followed by chromatography gave ester 13 as a
yellow oil. The compound was identified by its analysis,
infrared spectrum (ester carbonyl and a peak at 5.5 u
characteristic of cyclopropenes ( 1,42.)) and ultraviolet
absorptions around 230 mu and 316 mp. Two absorptions in
the lower region and at least three peaks centered between
300 and 325 my appear to be typical of diphenylcycloprop-
penes (41,42,43). The structure was further confirmed by
the n.m.r. spectrum which showed in addition to the phenyl
and ethyl proton resonances, an AB2 heptuplet at = 7.69
and 7.51 due to the cyclopropenyl and the methylene protons.
Hoping for improved yield, we prepared the cyclo-
propene from 1,2-diphenylcyclopropene-3-carboxylic acid (14,
44) by the Arndt-Eistert synthesis as modified by Newman and
Beal (45).* In contrast to our expectations, the yield of
17 per cent was even lower than that by the first route.
However, the identity of the two esters reinforced the
previous structural assignment.
The cyclopropene 13 was quite readily converted to
cyclopropenyl salt 15 by hydride abstraction with triphenyl-
methyl perchlorate in acetonitrile (4,1_5,40,46). Concentra-
tion and precipitation with cold, dry ether gave perchlorate
Masamune (47) recently reported synthesis of the
analogous methyl ester by this method.
15 as a white powder in 78 per cent yield. In addition to
the ester carbonyl absorption, the infrared spectrum showed
a strong peak at 7.02 p characteristic of the cyclopropenyl
cation (c1) and strong absorption at ca. 9.2 p due to the
perchlorate ion (41). The assignment of structure was
corroborated by the very distinctive diphenylcyclopropenyl
cation ultraviolet absorptions at 246, 293, and 308 mn in
acidic acetonitrile (41,42,4_). Final evidence for the
structure of the perchlorate was provided by comparison
with authentic sample prepared by treating the methylene-
cyclopropene lOa, prepared by the method of Battiste (35),
with perchloric acid in acetic anhydride and acetonitrile.
The salt is relatively insoluble in water although the
solid does slowly turn yellowish when wet. It was found
to be very stable, showing no signs of decomposition after
sitting at room temperature under air for seven months.
Interestingly, simple dissolution of the perchlorate
15 in acetonitrile (reagent grade) for an ultraviolet spec-
trum resulted in quantitative proton removal giving
absorptions characteristic of the methylenecyclopropene (35).
On a preparative scale the most satisfactory method of
proton abstraction was found to be with aqueous base.
Typically, vigorous shaking of a suspension of perchlorate
in very dilute aqueous base and ether gave a yellow organic
layer which was found after drying to contain 74 per cent of
the theoretical amount of methylenecyclopropene. No
extraneous peaks were present in the ultraviolet spectrum
although thin layer chromatography (t.l.c.) demonstrated
traces of three impurities. Even though isolation of pure
methylenecyclopropene 1Oa is difficult due to its notable
propensity for polymerization (5), crystalline material
was isolated and characterized as described in Chapter III.
Use of trimethylamine in ether in the proton removal
step (15,34) resulted in production of a spurious compound
in at least 50 per cent yield while the yield of methylene-
cyclopropene in the solution was at most 29 per cent. This
anomalous product absorbed at 257 and 333 mj and was com-
pletely converted to a diphenylcyclopropenyl cationic
species in acidic acetonitrile as indicated by the ultra-
violet spectrum of the solution.
Synthesis of cyclopropenylidenecyclopentadienes.--
Having successfully applied the method to one prototype
system we decided to promptly attack the main problem --
synthesis of mixed fulvalenes 19. Choice of diphenyl-
substituted cyclopropenes was made because of the ready
availability of precursors and the convenience of using
ultraviolet analyses in this series. Anticipating diffi-
culties with the less highly substituted members of the
series and recalling the successful preparations of benzo
0 R 2
R 2 2
BF or C104
a, R1 = R2 = H
b, RI = H, R2, R2 = C4H4
c, R1, R1, = R2, R2 = C4H4
derivatives in the heptafulvene (7) and the sesquifulvalene
(28) series we chose to attack the dibenzo derivative first.
The necessary substituted cyclopropene was prepared
by the low temperature reaction of fluorenyllithium with
diphenylcyclopropenyl perchlorate. Aqueous work-up,
chromatography, and one recrystallization gave fluorenyl-
cyclopropene 17c in 31 per cent yield. Structural assign-
ment followed from elemental analysis, infrared spectrum
(weak peak at 5.53 p and peaks for substituted fluorene (48)),
and the ultraviolet absorptions. The latter peaks were
those characteristic of a diphenylcyclopropene superimposed
on those expected for a substituted fluorene (49). The
n.m.r. spectrum was completely consistent with the assign-
ment, showing two aliphatic doublets at T= 5.63 (broadened)
and 7.12 in addition to the aromatic absorptions. The
broadening of the former absorption is probably due to
interaction with the other fluorenyl protons.
Reaction of fluorenylcyclopropene 17c with triphenyl-
methyl perchlorate or tetrafluoroborate in acetonitrile or
methylene chloride followed by precipitation with dry ether
gave 1,2-diphenyl-3-(9'-fluorenyl)-cyclopropenyl herchlorate
or tetrafluoroborate (18c) and triphenylmethane. The latter
was isolated in yields of ca. 50 per cent. The almost white
salts exhibited expected infrared absorptions due to the
cyclopropenyl cation (7.06 p) and the anionic species
(C104- or BF4-). Ultraviolet spectra in acidic acetonitrile
were those of the diphenylcyclopropenyl cation in addition
to those of the fluorenyl moiety.
Ultraviolet spectra of the salts 18c in neutral
acetonitrile again repeated the behavior of our prototype
system, that is, the proton was removed giving long wave-
length absorptions characteristic of triafulvenes (35,36,
38). Preparatively, proton abstraction was performed by
shaking the fluorenylcyclopropenyl tetrafluoroborate at 00
with a small excess of trimethylamine (15,34) in any one
of several solvents. This procedure very rapidly gave a
deep red solution and white solid. The white solid was
shown to be trimethylammonium tetrafluoroborate by compari-
son with an authentic sample and was isolated in 54 per cent
from starting cyclopropene (17c). Removal of solvent from
the red solution and recrystallization gave 9-(diphenyl-
cyclopropenylidene)-fluorene as deep red needles in 47 per
cent yield. Elemental analysis and n.m.r. spectrum were
both consistent with the chemical evidence (0 CH and
Me NHBF4) for net dehydrogenation from fluorenylcyclopro-
pene 17c. Resonances were present only in the region
between 7 = 1.86 and 2.80. In the infrared region the same
process was observed. Consistent with our observation of a
relatively large extinction coefficient for diphenylcyclo-
propene C-H stretch absorption, removal of the hydride
resulted in a great decrease in the peak at ca. 5.4 p. The
last vestige of aliphatic C-H absorption was removed in
passing from the salt to the cyclopropenylidenefluorene 19c.
Other infrared absorptions were suggestive of a methylene-
cyclopropene structure and will be discussed in a later
section. The ultraviolet spectrum (Table 1) was most
ELECTRONIC SPECTRA OF 19c AND MODEL COMPOUNDS
Compound Solvent log log
19c Cyclohexane 245 4.79 367 4.40
269 4.40 376sh 4.35
285 4.26 478 3.47
19c Isooctanea 238sh 4.79 297 4.11
243 4.85 365 4.48
268 4.45 375sh 4.44
283 4.33 476 5,50
19c Acetonitrile 238sh 4.74 363sh 1-.8
242 4.77 375 4.50
262 4.49 437 3.60
19c Acetonitrile, 252 4.52 296 4.50
HBF 262 4.50 309 4.50
16 Acetonitrile- 246 4.03 293 4.51
10o ethanol, 307 4.53
0.1 N HC104 b
9-Methyl- Ethanolc 228 3.80 273sh 4.11
fluorene 258s 4.20 290 3.77
265 4.23 301 3.95
aContained o4% CH2C12 to dissolve the hydrocarbon.
distinctive in having long wavelength absorption apparently
characteristic of the diphenyltriafulvene system and in
having no peaks in the usual diphenylcyclopropene region
around 300-325 mj.
A simpler preparation of this cyclopropenylidene-
fluorene is by dissolving 18c in methylene chloride and
chromatographing it on acid alumina (activity 1) with the
same solvent. A green band was followed by a red fraction,
which gave 19c in 42 per cent yield after solvent removal
The crystalline material appears to be quite stable --
a sample being unaffected by sitting at room temperature
under air for over one month. In solution it is less stable
to oxygen but still does not react rapidly. Lithium
aluminum hydride and n-butyllithium* both react rapidly and
cleanly with 19c at room temperature to give as yet un-
characterized products, although the reaction with the
organometallic does appear to destroy the three-membered
ring. Reaction with tetracyanoethylene (TCNE) at room
temperature as followed by ultraviolet analysis was somewhat
unclear and products have not been isolated. However, it is
apparent that the reaction is not nearly as fast as that of
our prototype triafulvene 10a (35).
The singly annelated derivative was prepared in an
This type reaction with sesquifulvalenes was des-
cribed in reference 28.
exactly analogous manner. The precursor cyclopropene 17b
was prepared in 73 per cent yield and did not require
chromatography. Its identity was established by analysis,
infrared spectrum, and typical ultraviolet spectrum. In
the n.m.r. region the aromatic protons were normal, the
olefinic protons were in two broadened doublets at T= 3.43
and 3.69, and the two aliphatic protons appeared at = 6.27
(broad) and 7.50 doublett). The presence of two olefinic
protons eliminates the possibility that the cyclopropene
ring might be on the 3-position of indene.
Hydride abstraction under argon gave crude tetra-
fluoroborate 18b (76 per cent) and triphenylmethane (65 per
cent). The infrared spectrum of the tetrafluoroborate
showed the expected peak at 7.07 p and tetraflucrobSJi
absorptions. An acetonitrile solution of tetrafluoiooorabo
18b which had been treated with acid showed a rather smeared
out spectrum with absorptions at 248, 288, and 305 mp.
Considerably more tailing toward the visible was observed
than is seen in the spectrum of 18c or other diphenylcyclo-
propenyl cations. These effects are not unreasonable if
the possibility of isomeric cations is considered. Under
the described conditions the cyclopropenylidenecyclopenta-
diene 19b is probably formed and then protonated. In the
analogous sesquifulvalenes, Prinzbach (21,28) has reported
that protonation results at the 8-position. In our system
an analogous protonation would give large proportions of
20b. The salt as it is prepared and collected probably has
structure 18b since Bertelli (15) found no evidence for
isomerization of the double bond in hydride abstraction
from 7-allylcycloheptatriene. Either isomer would, probably,
be satisfactory in the proton removal step.
In contrast to the stability of the diannelated 19c,
l-(diphenylcyclopropenylidene)-indene (19b) was found to be
very reactive. This reactivity was evidenced by our
inability to isolate the pure compound and the precautions
with which its solutions must be handled. Treatment of
slurries of tetrafluoroborate with trimethylamine resulted
in immediate formation of deep red colors and precipitation
of trimethylammonium tetrafluoroborate (60-80 per cent from
crude tetrafluoroborate 18b). Best results were obtained
when the system was protected by argon from the hydride
abstraction forward, and when the two abstraction steps
were performed at ca. 00. After removal of solid, it was
found that solvent and excess base could not be removed by
standard methods (that is, aspirator or passing argon over
the solution). Distillation was somewhat better. The prob-
lem was that as the solvent evaporated a dark insoluble
residue was formed on the wall at the expense of the desired
product. Bubbling argon through the solution with a hypo-
dermic needle proved to be an effective, if slow, way of
concentrating the solution. For spectral studies small
amounts of carbon tetrachloride as slurring solvent for
proton removal proved to be best. Infrared and n.m.r.
studies again point to net dehydrogenation as evidenced by
loss of aliphatic C-H absorptions on converting 17b and 19b.
As in the diannelated derivative, reaction side products
suggest the same conclusion. Infrared absorptions at ca.
5.4 and 6.4 p were very similar to those observed in the
dibenzo-system. Additional evidence for structure 18b was
the close similarity of the visible spectra of the two
species (see Tables 1 and 2).
ELECTRONIC SPECTRA OF 19ba
Solvent %max log Eb Solvent max log b
Isooctane 267 (4.2) Cyclohexane 263 (4.2)
371 (3.6) 573 (4.0)
494 (3.5) 490 (4.3)
Dioxane 265 (4.2) Benzene -- --
375 (4.0) 375 (4.1)
485sh (3.5) 475 (3.3)
Acetonitrile 265 (4.2) Acetonitrile,
HBF4 253 (4.2)
275 (4.2) 295 (4.3)
aSee Chap. III for the method used for these values.
The log C values are only rough approximations to
the real values and should not be used in any meaningful
comparisons between different solvents (see Chap. III).
There was no apparent inflection in the spectrum
above 385 mn although strong tailing did occur.
In the following experiments cyclopropenylidene-
cyclopentadiene 19b was assumed to be the species absorbing
at 370 my and the difference in the absorption at this wave-
length between neutral and acidic acetonitrile was used as
the analytic criteria for loss of 19b. Due to the sensi-
tivity of the red solutions to oxygen, the best way of
minimizing losses was found to be storage of an argon swept
solution under argon in powdered dry ice. Removal of
solvent left a sticky residue which was shown by its ultra-
violet spectrum to have lost ca. 50 per cent of the
methylenecyclopropene. Cold precipitation with pentane
resulted in pale tan material which contained little material
with the desired absorption ca. 370 my. The material was
also unstable to chromatography through neutral alumina or
silica gel. Reaction with TCNTE at room temperature was
fairly rapid but a small peak around 370 mu remained in the
presence of an excess of TCNE even after ten hours.
Somewhat disturbing and still unexplained is the
isolation from the above TCNE-reaction and from red solu-
tions from various runs small amounts of a high melting,
deep red solid which had ultraviolet and infrared spectra
similar (but not identical) to those of the red solutions.
The ultraviolet spectrum of one of these samples had an
extinction coefficient at 370 mu larger than at 263 mu.
The isolation of this solid was quite fortuitous and the
material was never isolated in quantity and purity suf-
ficient for elemental analysis or n.m.r. spectral data.
Attempts to prepare the non-anellated derivative 19a
have failed. The precursor cyclopropene was prepared in a
manner completely analogous to that used on the anellated
members. Structure 17a was assigned on the basis of its
ultraviolet spectrum and the properties of its reaction
product with maleic anhydride. The latter gave a correct
analysis for a mono-adduct, exhibiting the anhydride ab-
sorptions (50) in the infrared which obscured the cyclo-
propene peaks. The ultraviolet spectrum of the adduct
confirmed the presence of the unsaturated three-membered
ring. The position of the double bonds in cyclopenta-
dienylcyclopropene has not yet been established by n.m.r.
data. Exploratory attempts to prepare the corresponding
cation have failed.
Spectra of cyclopropenylidenecyclopentadienes 19b
and 19c.--Several points are worthy of note regarding the
infrared spectra of the cyclopropenylidenecyclopentadienes.
To date seven compounds possessing the methylenecycloprop-
pene structure have been reported in the literature. Three
of these are quinocyclopropenes (9a,9b (34), and iii (37) on
pages 9 and 10). When the three which we have synthesized*
See below for third one.
are also considered, we find that all eight for which this
region is reported (3.,55,36,38,39) have peaks between 5.32
and 5.57 p and in at least five of the systems (lOa, lOb,
19b, 19c, and 23) these absorptions are split. Furthermore,
the absorptions in this region are quite strong while in
simply substituted cyclopropenes the double bond absorption
at ca. 5.5 ) is weak (41,42). A second characteristic
absorption occurs between 6.2 and 6.6 p in seven of these
compounds (the three quinocyclopropenes have insufficiently
reported infrared data). This absorption is also split in
several cases. Assignment of this very intense peak to
aromatic C = C stretching modes is unsound not only because
they are usually weak (50), but also because triafulvene 11
shows this absorption (39) but has no aromatic groups. It
is quite possible, however, that the small peaks or
shoulders on the lower wavelength side of these absorptions
are due to the aromatic modes. Since these two absorptions
(5.4 and 6.4 p) appear to be characteristic of triafulvenes
it is probable that they are due to the two double bonds
possessed by each example, that is, the endo- and exo-double
bonds. But which absorption is due to which double bond is
an even riskier assignment, since absorption at 5.6-5.8 p
in methylenecyclopropene has been assigned to the exo-double
bond (5l) and Breslow has assigned the peak at 5.4 p to the
carbonyl and the one at 6.1 p. to the double bond in cyclo-
propenones (52). If, however, the 6.4 p absorption is due
to the exo-double bond, the shift from normal position and
the increased intensity are consistent with appreciable
contribution to the ground state from the dipolar structures
corresponding to 6b. These challenging questions must
probably await a sophisticated study by a trained spectro-
The presence of phenyl groups on the cyclopropene
lessens the usefulness of the n.m.r. spectra as probes for
qualitative assessment of the electronic situation in these
molecules. Nevertheless, the downfield shift of the
aromatic resonances (presumably the phenyls') as compared
to those of the dihydro derivatives is consistent with
that of our prototype (lOa) (35), the only other methylene-
cyclopropene for which a characterized exo-dihydro deriva-
tive exists (Table 3). Use of dialkyl substituted cyclo-
propenes would result in allowing a more direct comparison
of the resonances with those reported, and to be reported,
in the sesquifulvalene (28) and isoelectronic heterocyclic
sesquifulvalene analogs (20,21). The authors of the latter
reports have concluded from n.m.r. data that the homopolar
structures are an adequate description of the molecules
(20,21),but Berson (25) has challenged the basis on which
those conclusions were formed.
Probably the most dramatic property of the two
cyclopropenylidenecyclopentadienes reported is the effect
NUCLEAR MAGNETIC RESONANCE SPECTRA OF SOME METHYLENE-
CYCLOPROPENES AND THEIR DIHYDRO DERIVATIVES
Compound Solvent Positions of resonances of
aromatic and olefinic protons
(values of T)
17c CC14 2.30 3.15a
19c CDC13 1.86 2.80a
17b CC14 2.50 4.00
19b CC14 1.84 3.91a
13 CC4 2.40b 2.73b
l0aC C14 1.48b 2.20b 2.70b
aRange of resonance packet.
Center of unresolved multiple.
of solvent on their longest wavelength absorption* (Tables
1 and 2). While only small shifts with solvent change are
observed in the absorptions below 400 my, the long wave-
length bond shows a very strong hypsochromic shift with
increasing solvent polarity. Although the effect is
observed in both 19b and 19c, it is less clear in the
indenyl derivative because of the stronger tailing of the
375 mn peak which obscures the long wavelength maximum in
all but the most non-polar solvents. The shift in Xmax
does not exactly follow; for example, Kosower's Z-values
(54), but there is good qualitative agreement between
polarity and shift throughout the range of solvents
examined, with the greatest shift (41 mp) for 19c being
between cyclohexane and acetonitrile. This type of solvent
effect can only be explained as due to a transition from a
polar ground state to a less polar (or even of reversed
polarity) excited state (55,56). This effect leads to the
conclusion that the dipolar canonical forms of 19b and 19c
make appreciable contributions to the ground states of the
molecules. The extent of this contribution may be evaluated
more completely when dipole moments have been measured.
Similar solvent effects have been observed for lOa
(.5), 10b (53), and lOc (38), but they are smaller. Com-
pound 11 (39), however, shows no solvent effect.
19b, R = R = H
19c, R, R = C4H4
This result is also contrasted with the small solvent
effects observed for all but one of the isoelectronic
heterocyclic analogs of sesquifulvalene (20,22,2_,24).
Prinzbach has not reported solvent effects on the absorptions
of his sesquifulvalenes.
Another point of interest is the effect of anellation
on the long wavelength transition. Prinzbach (28) has re-
ported that successive anellation of benzo groups on the
five-membered ring of sesquifulvalene results in a hypso-
chromic shift in the long wavelength absorption. The
XEtOH 400-420 mp 392 my 380 mp
cyclopropenylidenecyclopentadienes show the same effect.
In cyclohexane the positions of these absorption maxima are
at 490 and 478 my for 19b and 19c, respectively. These
results are parallel to those predicted for fulvenes and
observed for alkylfulvenes, and opposite to those observed
for cyclopentadienylidenedihydropyridines and predicted for
sesquifulvalene.* Prinzbach interprets this shift as
suggesting that sesquifulvalenes are more similar to open,
cross-conjugated fulvenes than to a specially resonance
stabilized system (28).
Attempts to prepare other methylenecyclopropenes.--
After successful synthesis of the three previously described
methylenecyclopropenes (10a, 19b, and 19c), we attempted to
ascertain the generality of this synthetic route. For this
purpose, the methylenecyclopropene (lOb) of Jones (36) was
chosen because of its known properties and stability. Low
temperature reaction of n-butyllithium with dimethyl bromo-
succinate was followed by addition of that mixture to a
cold slurry of diphenylcyclopropenyl tetrafluoroborate.
After one hour, aqueous work-up gave a yellow oil and small
amounts of bis-(diphenylcyclopropenyl) ether. Chroma-
tography of the oil gave among other products, dimethyl
See references 20, 22, 2., 28 and references cited
therein for a more complete discussion of these results.
fumarate and a yellow oil (22) which absorbed at 298, 307,
315, and 333 nu. Rather than purify it further and
4 Products of 00
+1 H LiC4H9 and 22 1) 3C BF4
-H CHBrCO2CH3 + 2) H20-ether 23
> I dimethyl
BF4 CH2 C02CH3 fumarate
characterize it, we subjected the oil to hydride abstraction
conditions and the resulting solution was shaken with
water-ether. A yellow solid (23) was obtained which was
not the expected lOb. It was characterized as a methylene-
cyclopropene by its ultraviolet spectrum (with and without
acid) which was similar to (but not identical to) methylene-
cyclopropenes lOa and lOb. It also had the characteristic
infrared absorptions* at 5.36 and 5.49 u and at 6.35 P. It
had a carbonyl peak at 5.87 p. The n.m.r. showed aromatic
absorptions similar to those described by Jones (36) and
Battiste (35) at the lower field positions in their tria-
fulvenes. Absorptions in the aliphatic region suggest the
loss of a -OCH3 group and gain of ca. two butyl groups in
the molecule. Elemental analysis on a small sample also
implied the gain of alkyl groups although no exactly suitable
See previous discussion of infrared spectrum.
structure could be drawn. This unusual methylenecyclopro-
pene was not investigated further.
The preparation of dicyanomethylenecyclopropene 10c*
was also attempted. Precursor cyclopropene 24 was synthe-
sized and characterized by elemental analysis and infrared
and ultraviolet spectra. All were completely consistent
with the proposed structure. Attempts to remove the
hydride were unsuccessful. Triphenylmethyl salts do not
react with 24 at room temperature and do not appear to
produce much, if any, triafulvene at 70-800. Reaction
with 2,3-dicyano-5,6-dichloroquinone (DDQ) (L7) also failed
to give 10c.
The analogous dicarbethoxymethylcyclopropene 25
was prepared and characterized only by its ultraviolet
Recently prepared by Bergman and Agranat (58).
spectrum. Exploratory attempts to prepare the correspon-
ding cation failed in this case also.
The most obvious reasons for these failures are
steric hinderance to the approach of the trityl group and
inductive destabilization of the desired cation by the
electron withdrawing groups. Since the cyclopropenyl
hydrogens in the cyclopentadienyl series were also rather
crowded, we prefer the latter rationale. Bergmann's
observation that lOc was not protonated in trifluoroacetic
acid is consistent with this explanation (58).
REACTION OF A METHYLENECYCLOOROPENE WITH DIAZOMETHANE
Reports of the synthesis of four methylenecyclopropenes
were in the literature at the inception of this work (34,5_5,
56). Knowledge of the types of reaction which these species
may undergo was, and still is, after six more examples, very
limited. With the finding in these laboratories of a route
to the stable methylenecyclopropene 10b (56), we became
interested in the possibility of studying its reaction with
diazomethane. Although Wiberg (58) reported the addition of
diphenyldiazomethane and ethyl diazoacetate to cyclopropene
itself, we anticipated that at least some addition would
occur at the exo-cyclic double bond. The resulting 1-
pyrazoline, we hoped, would yield a spiropentene upon
thermolysis or photolysis. Consequently, we were greatly
_i Co2C H3C .c:CH o2CH 3
CO2CH3 ( H2C02CH3
-C + CH2N2 I
surprised to find in exploratory experiments that reaction
of yellow methylenecyclopropene with diazomethane gave deep
red solutions instead of the expected colorless solutions.
When the solution was found to contain only one product in
major proportions, we began investigation of this reaction
as an end in itself.
REACTION SCHEME FOR SECTION II
1) CH2N2 / N or
and geometrical isomer
Results and Discussion
Reaction of methylenecyclopropene 10b with diazo-
methane.--Reaction of 10b with diazomethane in ether-
tetrahydrofuran (THF) between -20 and 0 was moderately
fast (2-5 hrs.) in concentrated solution and gave evidence
(t.l.c.) of only one product (red) in major proportions.
After removal of solvents under vacuum at 00 a deep red oil
was obtained. Attempts to effect crystallization of the
oil failed. It was dissolved in carbon tetrachloride and
stored in dry ice to await further study.
It was assigned structure 27 on the basis of the
following observations and of elemental analysis of a sub-
sequent rearrangement product. Two infrared absorptions in
the carbonyl region at 5.71 and 5.85 p indicated the
presence of conjugated and unconjugated ester groups.
Presence of a diazo group was confirmed by the character-
istically intense absorption at 4.85 p ( 0). The deep red
color due to the weak absorption at 485 my was consistent
with assignment as a phenyldiazomethane (59). A terminal
methylene was suggested by the strong, but somewhat low
wavelength absorption at 10.98 p (50). The presence of two
weakly coupled olefinic protons was confirmed by the n.m.r.
spectrum which showed two broadened peaks of equal intensity
at r= 4.46 and 4.74, each of which possessed an area of
ca. 1/10 that of the aromatic absorptions. Also in the
n.m.r. spectrum were several incompletely resolved
absorptions in the aliphatic region between 1 = 6.45 and
6.90. These resonances are not inconsistent with structure
27 for the following reason. Attack of diazomethane on the
cyclopropene would be expected to show little, if any,
preference in position relative to the asymmetric end of
the molecule. The result would be formation of approxi-
mately equal amounts of the two isomeric diazo compounds
differing only in the configuration of the groups about the
internal double bond. A maximum of six peaks would then
be predicted in the region of the methylene and :oreth;l,! .
hydrogens. Any overlapping would reduce the total number
of peaks and increase the intensity of some. Although.
there are six resonances in the observed resonance packet,
it appears likely from approximate areas that some small
ones may be due to impurities. Spurious peaks were not
unexpected since the species could not be purified.
Consideration of the different functional groups
identified in the compound suggested that the most likely
route for their formation is the addition of diazomethane
to the endo-double bond to give intermediates 29 which
could then open to the assigned products.
CNlOb C2N2 ----- -- 27
10b CH2N2 CO2CH3
and other isomer
Precedent for attack on the ring unsaturation can
be found in Wiberg's report (58) that cyclopropene reacts
with diphenyldiazomethane at low temperatures to give a 1:1
adduct possessing properties suggesting 31. This species
is believed to result from olefin attack to give inter-
mediate 30 which would be expected to rearrange quite
+ 42CN2 --> N
easily to the observed product. Until recently, when Izzo
and Kende reported attack on diphenyl- and dipropyl-cyclo-
propenone (60) there was no reference to diazoalkane attack
on substituted cyclopropene rings. Although anticipating
substituted cyclobutenones, they actually obtained 33,
probably resulting from ring opening of 32, again,'the
intermediate expected from attack on the ring unsaturation.
0 NN2 0 -
R = ) or n-C3H7
The analogy with these reactions makes our suggested
scheme (lOb + CH2N2--+29---27) appear quite feasible. In
this case, however, instead of opening to the six-membered
heterocycle, the intermediate apparently opens to the
observed diazo compounds 27.
This, therefore, is another interesting example of
diazo-exchange between alkyl groups as discussed by
Walborsky (61) in relation to the reaction of diazoalkanes
with olefins and as observed by Farnum and Yates (62).*
Thermal rearrangement of 27.--Upon refluxing a carbon
tetrachloride solution of 27, it lost its red color and after
one hour was slightly yellow due to traces of methylene-
cyclopropene lOb. Concentration and chromagraphy of the
In the reaction of i and ii, products were found
which could be attributed to reactions involving iii and iv.
0COCH2N2 0CHC02CH3 OCN2o02CHO 0COCH2
i ii iii iv
solution gave pyrazole 26 as a white solid in 76 per cent
yield (from 10b). Its structural assignment was based on
analysis, spectral properties, and its reaction with base.
N = 02CH3
27 H2C02CH3 C2CO2CH3
Sor i CO2 "I3
Analysis of this product demonstrated that it contained the
elements of triafulvene 10b and diazomethane in a 1:1 ratio.
Carbonyl absorption was present at 5.71 p with only.slight
shouldering at 5.74 a. Intuitively, the ester at the 1-
position was expected to absorb at appreciably longer wave-
lengths than the other ester grouping because the former is
both an aromatic ester and a carbamate. This apparent
anomaly was easily accommodated when it was found that
pyrazoles 34 and 35 (63) absorb as shown. The possible
presence of an N-H bond was eliminated by both the infrared
and n.m.r. spectra. One-proton resonances at T= 4.09 and
4.71 with coupling constant 1.35 c.p.s. (64) established
the continued presence of the terminal methylene group
which was suggested by its fairly strong infrared absorption
CH3 N C2 0co
C 02C2H5 C02C2H5
Xmax (broad) 5.62-5.71p Xmax 5.65 and 5.79p
at 10.97 p. Absorptions at T = 5.96 and 6.35 may be assigned
to the methoxyl protons in the aromatic and aliphatic
esters, respectively. The methylene resonance at 7'= 6.05
is ca. 0.3-0.5 ppm. below that predicted using pertinent
parameters (65). This deshielding could be due to a unique
position of these protons relative to the styryl group.
The ultraviolet spectrum is discussed below.
Basic hydrolysis and alcoholysis of pyrazole 26.--
Further evidence for structure 26 was obtained by exami-
nation of products from its reaction with base in methanol.
Complete hydrolysis was accomplished by refluxing a
methanolic solution of pyrazole 26 and potassium hydroxide
for four and one-half hours. The single isolated product 56
26 KOH or
a N N
(86 per cent yield) was shown by elemental analysis to be
a monobasic acid. Infrared absorptions consistent with the
proposed structure 36 were found at 2.99 (N-H), 4.05 (acid
0-H), 5.85 (unconjugated acid C = 0) and 11.09 p (C = CH2).
A broad absorption in the ultraviolet region at 250 my
was further confirmation of the proposed structure (see.
below). During these hydrolysis studies, we discovered that
diester 26 reacts immediately with base to produce a dif-
ferent neutral species which was then slowly hydrolyzed.
This reaction occurred rapidly even with trace amounts of
sodium methoxide in dry methanol at dilutions ca. 1 2 !.-L
By preparative t.l.c. we have isolated the product as an
oil and have assigned structure 37 to this neutral species.
The assignment is based on spectral and circumstantial
chemical data. In addition to the single species observed
by t.l.c., the reaction gave a product with the same re-
tention time as dimethyl carbonate upon vapor phase
chromatography. Also, reaction of acid with diazomethane
26 trace of \ or
CH30 Na N N- H
gave a compound having the same Rf-value as ester 26.
Absorptions at 3.14 (N-H), 5.71 (unconjugated C = 0), and
11.05 p (C = CH2) are suggestive of key parts of the
structure for the monoester. The n.m.r. spectrum showed
absorptions at T= 1.41 (N-H), 2.61 and 2.90 (phenyl pro-
tons), 4.29 and 4.85 (C = CH2) and 6.57 aliphaticc protons),
with resonance areas in rough accord with the assignment.
The low field position of the N-H absorption is very strong
evidence for the pyrazole core in the molecule (65).
Ultraviolet spectra of pyrazoles 26, 36 and 37.--The
ultraviolet spectrum of esters 26, 37 and acid 36 are also
all quite reasonable for the assigned structures when
available model compounds are examined (Table 4). Models
of the pyrazoles clearly show that because of non-bonding
interactions the phenyl group and the styryl group cannot
both be in the plane of the heterocyclic ring. They also
suggest that a simple resolution of the interference while
maintaining maximum --overlap might be obtained by rotating
the styryl grouping out of the plane of the pyrazole ring.
This configuration would result in a 0-substituted styrene
system and a phenylalkylpyrazole system with little elec-
tronic interaction between the two. It may be seen from
the model compounds 39-43 that as the size of the substitu-
ent on a styrene system is increased there is a hypsochromic
shift in %max and the intensity is reduced. In our
ULTRAVIOLET ABSORPTION DATA FOR SOME STYRENES AND PYRAZOLES
Compound %max log
39 Styrene 248a,b 4.17
40 / -Methyl- 243a,b 4.06
41 f-Ethyl- 240a,b 4.00
42 -5-(n-propyl)- 29a,b 3.97
43 f-Isopropyl- 234a,b 3.89
44 1-Phenyl- 251,d 4.41
45 3-Phenyl- 249,d 4.24
46 3-Phenyl-5-methyl- 251, d 4.28
37 250c 4.33
36 248c 4.29
47 1-Phenyl-3-methyl- 256c d 4.15
48 1-Phenyl-3-methyl-4-carboxy- 266a,e 4.33
26 258c 4.37
CIn CH OH.
system the overall effect of crowding the styryl chroma-
phore could conceivably reduce the intensity of its
absorption, allowing the substituted pyrazole chromaphore
to dominate the spectrum. In fact, the spectrum of ester
36 and acid 37 agree quite well with this suggestion. As
may be seen from the absorption of 47 and 48 the intro-
duction of a carbomethoxy group into the phenylalkylpyrazole
might be expected to produce a bathochromic shift o' ca.
10 mu and an increase in C. In comparing esters 26 with
37 we find increased E and an 8my shift in Xmax. These
correlations lend further credence to the structure assign-
ments of 26, 36, and 37.
In considering the products which have been described
above, we find adequate precedent in the literature for
possible reaction schemes. Van Alphen (69) found that in
the presence of alcoholic base 1-carbethoxy pyrazoles lost
the ester group. At first glance this would appear to be
simply a hydrolysis followed by loss of carbon dioxide, but
in light of our findings the loss of the ester group may
not even require hydroxide bases.
Internal reaction of an a, P-unsaturated diazo com-
pound was first reported in 1935 (70,71) in the formation
of pyrazole from diazopropene. More recently, Closs et al.
(72) have reported the following reaction where R1, R2'
and R3 = H or Me. From some of the reaction mixtures they
R1 CH3 R CH
Diglym pe + when R H
have also isolated corresponding alkenyldiazo species and
Van Alphen (69) has also reported the facile re-
arrangements of pyrazolenines to pyrazoles in the presence
of such mild catalysts as acid anhydrides or acetic acid.
In cases where only one group migrated, he observed re-
arrangement to the pyrazole even on recrystallization from
methanol. Consequently, examination of the structures of
diazo 27 and N-substituted pyrazole 26 might lead one to
anticipate the possible intermediacy of pyrazolenine 28 in
27 2 ---- 26
It should be noted, however, that they feel that
alkenyl diazo species are not necessarily the precursors
of the pyrazolenines in this case since the one isolated
diazo did not rearrange under reaction conditions.
It was not surprising, therefore, to find that
maintaining a solution of diazo 27 for five weeks at freezer
temperatures left very little color but that t.l.c. showed
the presence of little N-carbomethoxypyrazole 26. Major
amounts of a species with Rf-value between 27 and 26 were
observed. Chromatography at 0-50 gave a relatively pure
oil which we have been unable to crystallize. Upon further
heating this oil is converted almost quantitatively to
pyrazole 26. On the basis of spectral evidence and the
identities of subsequent products in the reaction scheme we
have tentatively assigned structure 28 to this product.
Participation of the internal double bond in the rearrange-
ment was apparent from the absence of splitting in the
carbonyl peak at 5.71 ), and disappearance of the diazo
group was suspected from loss of color and absence of
absorption at ca. 4.8 y. The infrared absorption at 10.95 )
and the n.m.r. absorptions at 7 = 4.26 and 4.80 were de-
cisive evidence for the presence of the terminal unsatura-
tion. The resonance for one methoxyl group was at r = 6.55,
whereas the other methoxyl and the methylene group absorp-
tions appeared at T= 6.70. These positions are unusual in
that the methylene group appears at lower fields than normal
and the methoxyl groups are somewhat higher than their usual
F= 6.30 (65). Both of these effects may possibly be due to
certain orientations of these protons relative to the nearby
phenyl and olefinic groups. The ultraviolet spectrum shows
strong absorption at 242 my and a much weaker absorption
at 303 mu. Models suggest that the 1,3-diphenylbutadiene
system in 28 can not be planar because of steric inter-
actions. As before, the most probable effect would be the
rotating of the styryl group out of the plane of the five-
membered ring resulting in absorption at lower wavelengths
than would be expected on the basis of an all planar
system. The primary absorption at 242 mu is not unreasonable
for structure 28 (see Table 4 and earlier discussion).
This paper then describes a rather interesting, but
not at all unreasonable addition-rearrangement sequence for
the reaction of diazomethane with a methylenecyclopropene.
General.--Melting points were taken in a Thomas
Hoover Unimelt apparatus and are uncorrected. The infrared
spectra were recorded with a Perkin-Elmer Infracord unless
otherwise specified, and ultraviolet and visible spectra
were with a Cary 14 spectrophotometer. Elemental analyses
were performed by Galbraith Laboratories, Inc., Knoxville,
Tennessee. Nuclear magnetic resonance spectra were
determined with a Varian 4300-2 high resolution spectro-
meter operating at 56.4 Mc. Chemical shifts are reported
in ppm. from tetramethylsilane.
The n-butyllithium used was a 1.55 M solution in
hexane from the Foote Mineral Company.- Trimethylamine was
from Matheson Scientific Company. Triphenylmethyl tetra-
fluoroborate and perchlorate were prepared as described by
Dauben et al. (75). In all reactions involving organo-
lithium compounds, the system was closed under an atmosphere
of argon or was exposed to a slow but constant flow of
argon through the system.
Analytical thin layer chromatography (t.l.c.) was on
ca. 0.25 mm. silica gel layers and preparative layers were
ca. 1.00 mm. The spots were visualized by 50 per cent
sulfuric acid and heat, iodine vapor, or by quenching of
the fluorescence of an indicator (256 m)) in the silica gel.
phenylcyclopropene-3-carboxylic acid was prepared by a
slight modification of the procedure of Breslow et al. (44).
A mixture of 43.3 g. (0.244 mole) of tolane and 7 g. of
electrolytic copper dust were heated to 1400 and 27.8 g.
(0.244 mole) of freshly distilled ethyl diazoacetate was
added with stirring at such a rate that the temperature was
maintained at 130-1400 without external heating. After the
addition was complete the mixture was refluxed for 16 hours
with 63 g. of KOH in 350 ml. of methanol. After most of
the solvent had been removed under vacuum, water was added
and the mixture filtered. The filtrate was acidifiA .nd
extracted with chloroform. After percolation of the
extracts through Florex, removal of the solvent gave a dark
residue which could be cleaned up by washing with ether.
The yield of white solid, m.p. 203-2090 (dec.) (lit. (44)
m.p. 2050 dec.) was 15.8 g. (31.6 per cent). A second crop
of 0.6 g., m.p. 190-1970 (dec.), of tan crystals was ob-
tained from the ether washings. The acid was used without
further purification for the preparation of cyclopropenyl
DiohenylcyclooroDenyl perchlorate (4.) and tetra-
fluoroborate.--A solution of 9.0 ml. of 70 per cent
perchloric acid in 55 ml. of acetic anhydride was prepared
by slowly adding cold anhydride to the acid while stirring
and cooling in an ice-bath. A slurry of 6.78 g. (28.7
mmoles) of 1,2-diphenylcyclopropene-3-carboxylic acid in
20 ml. of cold acetic anhydride was added in portions to
the ice-cold perchloric acid solution. The mixture was
stirred in an ice-bath for one and one-half hours. Pre-
cipitation and washing with cold dry ether gave 4.4 g.
(53 per cent) of diphenylcyclopropenyl perchlorate as a
light solid, m.p. 160-1620 (dec.) (lit. (4) m.p. 148.5-
150.5 dec.). The tetrafluoroborate salt could be obtained
by dissolving the perchlorate in a quantity of solution of
4b per cent tetrafluoroboric acid in acetic anhydride such
that there was about a 20-fold excess of tetrafluoroboric
acid. Precipitation with ether gave a salt containing
little perchlorate. The salts could be used in this condi-
tion or recrystallized from acetonitrile with ether. An
attempt to prepare the tetrafluoroborate directly from the
acid was unsuccessful.
Dinhenylcyclopropenyl hexafluorophosphate (7_).--
Hexafluorophosphoric acid (65 per cent) (1.92 ml.) was
added dropwise with stirring to acetic anhydride (10.0 ml.)
with ice-bath cooling. The solution was allowed to warm to
140 and finely powdered 1,2-diphenylcyclopropene carboxylic
acid (1.92 g.) was added with stirring during six minutes.
After stirring three minutes longer the temperature was
lowered to 0 and crystallization was induced by scratching
or seeding. After stirring an additional 30 minutes in the
ice-bath, benzene (10 ml.) was added and the mixture chilled
to freezing. Filtration, washing with cold acetic anhy-
dride-benzene (1:1), and then with benzene followed by air
drying gave white needles of diphenylcyclopropenyl hexa-
fluorophosphate, dec. ca. 170-180.
The ultraviolet spectrum in acetonitrile (neutral or
acid) was that expected for a diphenylcyclopropenyl cation
(45). The infrared spectrum (KBr) exhibited peaks at 3.25,
5.57, 6.27, 6.67, 7.10 (broad, cyclopropenyl cation), (C1l)
7.66, 7.86, 8.55, 9.82, 10.07, 10.67, ca. 12 (broad) 15.20,
and 14.83 i.
Successive freezing-filtering operations on the
filtrate raised the yield. The usual yield on these prepa-
rations was between 63 and 72 per cent.*
The following method is the same as that reported by
D. G. Farnum (23) reports consistent yields of 85-
100 per cent by above method except that his entire addition
and reaction were at 00 and one crystallization was suffi-
cient to obtain the product.
Hasamune (47) for the preparation of the methyl ester.
Treatment of 1.0 g. (4.2 mmoles) of 1,2-diphenylcyclopropene
carboxylic acid with 1.5 ml. of thionyl chloride for 45
minutes followed by removal of excess thionyl chloride under
vacuum gave the acid chloride as a pale tan solid (56).
A solution of the acid chloride in 20 ml. of dry
ether was added during 45 minutes to an ice-cooled solution
of 13 mmoles of diazomethane in 20 ml. of ether. The
solution was stirred for one and one-half hours at room
temperature before the solvent and excess diazomethane were
removed by the aspirator. The pale yellow solid was washed
with a pentane-ether mixture to give 0.8 g. (73 per cent
based on acid) of (2,3-diphenyl-A2-cyclopropenyl)-diazo-
methyl ketone (42), m.p. 1220 (dec.) (lit. (42.) .o, 1151
dec.). The infrared spectrum (nujol mull) showed signili-
cant absorption at 4.72 (-C=N2) (24), 5.45 (weak, cyclo-
propene C=C) (41,42), 6.19 (C=0) (74), 7.33, 7.60, 8.80,
8.98, 9.11, 15.00, 13.51, and 14.59 i.
The catalyst solution of Newman and Beal (45) was
used to effect the rearrangement of the diazoketone to the
ethyl ester. A solution of 550 mg. of the diazoketone in
20 ml. of absolute ethanol at room temperature was treated
with 0.5 ml. of catalyst solution (0.5 g. silver benzoate
in 5 ml. of triethylamine). Nitrogen evolution was finally
induced after warming and adding a speck of paraformaldehyde.
A further 0.3 ml. portion of catalyst solution and 30 mg.
of solid silver benzoate were added to help speed the re-
action. A total of nine and one-quarter hours at room
temperature and another 0.1 ml. of catalyst mixture were
required for completion of gas evolution. The mixture was
filtered, washed with dilute hydrochloric acid, brine,
bicarbonate solution, brine again, and dried over magnesium
sulfate. Removal of the ether gave 470 mg. of dark oil
which chromatographed on 30 g. of Woelm acid alumina.
Elution with pentane-ether (19:1) gave 141 mg. (24 per cent
from diazoketone) of 1,2-diphenyl-5-carbethoxymethylcyclo-
propene as a pale yellow oil. The ultraviolet spectrum
(cyclohexane) possessed peaks at 228, (log C = 4.22), 237
(4.13), 302 sh (4.26), 3.07 (4.51), 316 (4.39), 325 sh
(4.30), and 335 my (428). An infrared spectrum (CC14) of
this oil showed significant absorption at 5.25, 3.34, 3.40,
5.52 (weak, cyclopropene C=C), 5.75 (C=O), 6.70, 6.92,
7.32, 7.70, 8.60, 9.60, 9.80, and 14.52 i. The n.m.r.
spectrum (CC14, external acetaldehyde standard) exhibited
resonances at = 2.40 and 2.73 (unresolved aromatic proton
multiplets), 5.96 (methylene of ethyl group, quadruplet),
7.51 and 7.69 (methylene and methine protons in a A2B hep-
tuplet) (64), and 8.83 (methyl triplet). An analytical
sample was obtained by short path distillation at 0.6 mm.
and 1350 (bath temperature).
Anal. Calcd. for C19H1802: C, 81.99; H, 6.52.
Found: C, 81.75 per cent; H, 6.57 per cent.
Alternate Dreoaration of ethyl (1,2-diphenyl-3-
carbethoxymethylcyclopropene (13).--Ethyl bromoacetate
(0.286 ml., 2.5 mmoles) was added during two minutes to a
stirred solution of 1.61 ml. (2.5 mmoles) of a 1.55 M solu-
tion of n-butyllithium in hexane and 8.0 ml. of dry ether
maintained at -750. The mixture was stirred for an addi-
tional eight minutes at this temperature. This solution
was then rapidly transferred with a cold syringe to a
stirred slurry of 278 mg. (1.0 mmoles) of diphenylcyclo-
propenyl tetrafluoroborate in 8.0 ml. of methylene chloride
also at -75. After stirring for 50 minutes under argon
the mixture was poured into a very dilute acid solution
and extracted well with ether. The ether layer was washed
with brine, dried over iMgSO, and filtered through Drierite.
Removal of the ether left an oil which was chromatographed
on 18 g. of Woelm acid alumina (activity 1). Elution with
pentane-ether (9:1) gave 8 mg. of a yellow oil followed by
several cuts which gave the same ultraviolet absorption as
the desired ester 15. A middle cut of this oil exhibited
an infrared spectrum which was identical to that of the
previously characterized 1,2-diphenyl-3-carbethoxymethyl-
cyclopropene. The n.m.r. spectra in carbon tetrachloride
were identical except for a very small peak just upfield
from the methylene quartet in the ester from this route.
All the cuts which had the same well resolved ultraviolet
spectrum were combined to yield 75 mg. (27 per cent) of the
cyclopropenylacetate 13 as a pale yellow oil.
perchlorate (15).--The 75 mg. (0.27 mmole) of ester 13 from
the previous preparation was dissolved in 4.0 ml. of dry
acetonitrile and filtered from a small amount of insoluble
material. After this solution was treated with 97 mg. of
triphenylmethyl perchlorate (46) and stirred for a few
minutes, most of the solvent was removed at reduced pressure
and room temperature. Addition of the remaining solution
to 40 ml. of cold, dry ether precipitated 77 mg. of the
perchlorate 15 as a white solid. Cooling the yello.!- .
trates gave another 2 mg. (total 0.21 mmole, 75 per cent
yield). The solid melted at 158-1630 (dec.);h KBr.22
(aromatic C-H), 3.52 aliphaticc C-H), 5.47 (w), 5.77 (C=0),
6.01 (w), 6.28, 7.01 (cyclopropenyl cation) (41), 7.31, 7.52,
8.20, 8.49, 8.73, 8.98, 9.2 (C104-) (41), 9.77, 13.10, and
14.70 uI; max (log e) 242 (4.36), 250 (4.38), 267 (4.34),
285 sh (4.30), 295 (4.34), 308 (4.26), and 374 (3.95) mp;
Xma (1 drop 48 per cent HBF4) 246, (4.24), 293 (4.50)
and 308 my (4.52).
Alternate DreDaration of 1,2-diphenyl-3-carbethoxy-
methylcyclooroDenyl perchlorate (15).--A solution of 412 mg.
(2.0 mmoles) of diphenylcyclopropenone and carbomethoxy-
methylenetriphenylphosphorane (Aldrich Chemical Co.) in
25 ml. of dry methylene chloride was allowed to sit for
three days according to the method of Battiste (35). The
solution was concentrated and the methylene chloride re-
placed by benzene. Chromatography on Woelm acid alumina
(activity 1) gave several small yellow fractions. One
middle fraction of these was concentrated and treated with
a few drops of an acetonitrile solution of 0.034 ml. of 70
per cent perchloric acid in 0.12 ml. of acetic anhydride.
The intense yellow color disappeared, and treatment with
cold, dry ether precipitated a small amount of white solid.
The infrared spectrum of this solid showed all peaks that
were present in the material from the hydride abstraction
route except for small ones at 6.0 and about 14.5 ). The
ultraviolet spectrum appeared identical to that of the other
preparation both with and without acid.
1 ,2-Dihenyl-4-carbethoxymethylenecyclooropene (lOa).-
A 37 mg. sample (0.098 mmole) of the carbethoxycyclopropenyl
perchlorate 15 was suspended in 3.0 ml. of dry ether. Addi-
tion of 9 ml. of water containing 0.06 mmole of sodium
bicarbonate followed by a period of vigorous shaking pro-
duced a yellow ether layer. Removal from the water and
drying with magnesium sulfate gave 5.0 ml. of a yellow
solution which was indicated by its ultraviolet spectrum
(35) to contain 0.073 mmole (74 per cent yield) of the
methylenecyclopropene 10a. Its spectrum in acid also showed
no extraneous peaks. Thin layer chromatography indicated
trace amounts of three impurities. A 2.0 ml. portion was
concentrated at 0 by passing a stream of argon through the
solution by use of a hypodermic needle. Pentane was added,
a little floc removed by centrifuging, and the solution
concentrated again. This sequence was repeated with con-
centration of the pentane solution of its cloud point.
Cooling at 00 and then at -100 gave bright yellow crystals.
Rapid filtration gave 1-2 mg. of the methylenecyclopropene
lOa, m.p. 71-735 (lit. (35) m.p. 72-750). The infrared
spectrum (KBr) showed significant peaks at 3.23-3.50 (C-H),
5.41, 5.57, 5.96 (conjugated C=O), 6.50, 6.74, 6.88, 7.25,
7.44, 8.70, 8.84, 9.29, 9.49, 12.85, 13.22, and 1>:70 .
Less careful attempts to obtain the solid methylone-
cyclopropene did not succeed. For example, a dried yellow
ether solution resulting from proton abstraction from 82 mg.
of the perchlorate 15 was evaporated to dryness, and the
orange residue was crystallized and triturated with hexane.
The resulting 51 mg. of yellow solid did not possess a dis-
tinct melting point (melt was not clear at 1350), and the
infrared spectrum (CC14) showed a peak at 5.76 p which was
very much larger than that at 5.96 p. The absorption at
5.44 p was also smeared out with no resolution.
An attempt to use carbon tetrachloride as the organic
phase in the proton removal step was unsuccessful, apparently
due to insolubility of the perchlorate in either phase.
Reaction of diphenylcarbethoxymethylcycloDronenyl
perchlorate (15) with trimethylamine.--A mixture of 39 mg.
(0.104 mmole) of perchlorate 15 and 1.0 ml. of dry ether
were cooled in an argon-flushed tube to ice-bath temperature.
To the mixture was added in four portions ca. 0.12 mmole of
trimethylamine (as a 1.1 M solution in ether) with vigorous
shaking between additions. A yellow solution and white
solid resulted. The solid had an appearance different from
that of the original perchlorate salt. After centrifuging,
an aliquot was taken for analysis of its ultraviolet spec-
trum (MeCN). In addition to all the characteristic peaks
of the methylenecyclopropene lOa, two shoulders at 257 and
333 mu were present. The latter absorption was relatively
strong. By use of absorption in the 370 mp-region, a maxi-
mum amount of 0.030 mmole (29 per cent) of methylenecyclo-
propene or starting perchlorate was calculated to be in the
solution. However, in acidic acetonitrile no peaks un-
characteristic of the cyclopropenyl cation were observed.
Furthermore, the yellow solution was calculated to contain
0.078 mmole of cyclopropenyl cation-producing species.
Attempts to obtain filterable crystalline material as before
resulted in oils.
Fluorenyllithium was prepared by adding 6.46 ml. (10.0
mmoles) of butyllithium solution to 1.664 g. (10.0 mmoles)
of fluorene dissolved in 13.6 ml. of ether and eefluxing
(76). The addition took one-half hour and the solution
was refluxed for one and one-half hours. This solution was
then dropped slowly into a well-stirred slurry of 1.162 g.
(4.0 mmoles) of diphenylcyclopropenyl perchlorate in 25.0
ml. of ether which was cooled with dry ice-acetone. The
addition was stopped when an orange color persisted in the
reaction slurry. The mixture was stirred for an additional
one-half hour at this temperature, after which it was shaken
with dilute acid and the brine. Drying over magnesium
sulfate and filtration through Drierite gave a yellowish
solution. Removal of the solvent left an oil which was
chromatographed on 60 g. of Woelm acid alumina (activity 1).
Elution with pentane, pentane-benzene (9:1) and then
pentane-ether (49:1) gave a white solid. One recrystalliza-
tion from benzene-petroleum ether gave 445 mg. (31 per cent)
of 1,2-diphenyl-3-(9'-fluorenyl) cyclopropene, m.p. 168-
1710. A sample of analytical purity, m.p. 170-171.50, was
obtained after two additional recrystallizations. The
infrared spectrum (KBr) of 17c contains significant ab-
sorption at 3.24 (aromatic C-H), 3.39 (primarily the cyclo-
?ropenyl C-H), 5.53 (cyclopropene C=C), 6.28, 6.70, 6.92,
7.37, 7.76, 8.14, 8.23, 8.41, 8.70, 9.37, 9.67, 9.79,
10.62, 11.00, 11.40, 11.97, 13.08, 15.35, 13.67, and 14.64 p.
The ultraviolet spectrum (isooctane) exhibited absorption
at 228 (log = 447), 238 (4.40, 257 (4.31), 264 (4.32),
268 (4.32), 297 sh (4.31), 304 (4.43), 311 (4.35), 321
(4.42), 327 sh (4.36) and 339 m)i (4.30). The n.m.r.
spectrum (CC14) showed absorption by the aromatic protons
between t = 2.30 and 3.15 with the maximum at 1 = 2.73.
The two aliphatic proton peaks were at = 5.63 (broadened
doublet) and 7.12 (sharp doublet).
Anal. Calcd. for C28H20: C, 94.34; H, 5.66. Found:
C, 94.31; H, 5.63.
(100 ng., 0.281 mmole) and 105 mg. (0.306 mmole) of tri-
phenylmethyl perchlorate were dissolved in 8 ml. of aceto-
nitrile and allowed to react at room temperature for 40
minutes. Concentration of the solution and addition to
80 ml. of cold, dry ether gave 50 mg. (39 per cent) of
pinkish solid by filtration. After several washiL: ;
ether the compound melted at 120-1350 (dec.). Infrar
absorptions (KBr, pellet turned deep red) were present at
5.22 (aromatic C-H), 5.40, 5.55 sh, 6.28, 6.47, 6.68, 6.77,
6.92, 7.06 (cyclopropenyl cation), 7.60, ca. 9 (strong,
0104-), 15.02, 13.42, 13.62, 14.50, and 14.70 }. The
ultraviolet spectrum (IeCN) showed peaks at 237 sh, 242 sh,
263 sh, 295 sh, 363 sh, and 375 mu.* In the presence of a
small amount of tetrafluoroboric acid the spectrum exhibited
peaks at 252, 262, 273 sh, 295, and 308 mu.
200 mg. sample (0.563 mmole) of the fluorenylcyclopropene
17c and Y ml. of dry methylene chloride were put into a dry
test tube which then was covered with a septum. *' 9 *
and its contents were flushed by passing argon in through
a hypodermic needle and letting it flow out through another.
The mixture was cooled in an ice-bath and treated with a
cooled solution of 186 mg. (0.563 mmole) of triphenylmethyl
tetrafluoroborate in 2 ml. of methylene chloride. After
25 minutes the resulting dark green solution was added
through a septum to 12 ml. of cold, dry ether by means of
a hypodermic syringe. The precipitated pink solid was
centrifuged and the supernatant solution removed with a
syringe. After one washing with ether the solid was sus-
pended in 8 ml. of ether and M ml. of methylene chloride
and cooled in an ice-bath. Addition in one portion of
0.44 ml. of a 1.56 M solution of trimethylamine in ether
immediately produced a deep red coloration. The tube re-
mained in the ice-bath for 20 minutes with occasional
A spectrum of the visible region was not obtained.
shaking, after which the red supernatant solution was re-
moved and the solid washed with methylene chloride until
white. The ultraviolet spectrum of the solution indicated
a yield of 57 per cent (from the fluorenylcyclopropene
17c) of the methylenecyclopropene 19c. The solvent was
removed under vacuum and petroleum ether was added. Fil-
tration gave 118 mg. of a dark solid, m.p. 155-157.
Recrystallization from benzene-cyclohexane gave 93 mg. (47
per cent) of deep red needles, m.p. 158-158-160. Succes-
sive recrystallization gave an analytical sample, m.p.
160-161. Significant infrared absorptions* (KBr) are at
3.25 (aromatic C-H), 5.39, 5.54, 6.39, 6.43 (strong), 6.73,
6.90, 7.47, 8.30, 8.81, 9.07, 9.79, 9.92, 10.07, 10.86,
13.07, 13.39, 13.65, 14.26, 14.51, and 14.98 p. The n.m.r.
absorption (CDC13) was in three unresolved multiplets in
the aromatic region. Their ranges were 7 = 1.86-2.24,
2.24-2.55, and 2.55-2.80. The ultraviolet and visible
absorptions are reported in Table 1.
Anal. Calcd. for C28H18: C, 94.88; H, 5.12. Found:
C, 94.51; H, 5.16.
When the supernatant solution and washings from the
hydride abstraction step were evaporated, 197 mg. of dark
residue was obtained (theoretical yield of triphenylmethane
is 137 mg.). This residue was dissolved in benzene and
shaken with portions of H2SO4 until all color was gone.
Perkin-Elmer 21 Spectrophotometer.
After washing with water, drying and evaporation, 148 mg.
(108 per cent) of yellow solid was obtained, m.p. 85-89.
A thin layer chromatogram showed a large spot and two small-
er spots. The large spot had an Rf-value the same as that
for triphenylmethane. Recrystallization from EtOH gave
79 mg. of pure triphenylmethane, m.p. 91-920 (lit. m.p.
The white solid (45 mg., 54 per cent based on
03CBF4) isolated after the base treatment was identified
as trimethylammonium tetrafluoroborate by the identity of
its infrared spectrum to that prepared by the reaction of
trimethylamine with tetrafluoroboric acid in ethanol.
Alternate preparation of 9-(diphenylcyclopropenyli-
dene)-fluorene (19c) from 17c.--A suspension of 100 mg.
(0.281 mmole) of cyclopropenylfluorene 17c in 1Y ml. of
methylene chloride was swept with argon as described before
and treated with 105 mg. (0.300 mmole) of triphenylmethyl
perchlorate in 35 ml. of acetonitrile. After 35 minutes
the solution was concentrated and chromatographed on Woelm
acid alumina (activity 1) eluting with methylene chloride.
A greenish band was followed by a red fraction. Removal of
the solvent from the red solution gave 61 mg. of dark solid.
One recrystallization from benzene-cyclohexane gave 42 mg.
(42 per cent yield from 17c) of deep red needles, m.p. 161-
Reactions of 9-(diphenylcycloproenylidene)-fluorene
(19c).--The red needles of methylenecyclopropene 19c are
fairly stable; the melting point is unchanged after sitting
for one month under air at room temperature. A solution of
6.9 mg. of the methylenecyclopropene in 1.0 ml. of benzene
lost, at most, 15 per cent of its absorption at 375 m) when
it had oxygen bubbled through it for one hour at room
When a small amount of 19c in ether-cyclohexane was
treated with excess n-butyllithium at room temperature and
then quenched with ethanol followed by dilute acid, a
colorless organic layer was obtained. Thin layer chroma-
tography on silica gel with benzene gave only one spot.
The ultraviolet spectrum of the organic phase in acetonitrile
showed peaks at 301, 290, and a broad group of absorptions
between 258 and 279 mn. Only a slight shouldering was
observed above 305 mu.
Reaction of the red compound with excess lithium
aluminum hydride in ether-tetrahydrofuran at room tempera-
ture instantaneously removed the color. Treatment with
ethanol followed by dilute acid gave a colorless solution
producing three spots on a thin layer chromatogram developed
with benzene. One spot had the same Rf-value as 1,2-di-
phenyl-3-(9'-fluorenyl)-cyclopropene. Another moved only
slightly, and the third was between the other two. The
ultraviolet spectrum of the solution in cyclohexane showed
strong absorptions at 228, 249, 258 my in addition to
weaker absorptions at 270, 282, 303, and 316 mu.
The methylenecyclopropene 19c (3 mg.) was dissolved
in 2.0 ml. of acetonitrile. Addition of 23/ mg. of tetra-
cyanoethylene at room temperature caused, in five minutes,
a drop in the absorption at 375 mu to 70 per cent of its
original value. After 12 hours it had only 34 per cent of
the original absorption at 375 mp, and 38 per cent of the
remaining absorption was not quenched by acid (the methylene-
cyclopropene 19c has no absorption in acid solution at
375 mu). The spectrum is much more smeared out showing
absorptions at 310, 295, 270, 256, and 247 mu with the
256 mu peak being the largest. The products of the reaction
have not been isolated.
Indenyllithium was prepared by adding 5.80 ml. (9.0 mmoles)
of butyllithium solution to 1.21 ml. (10.0 mmoles) of indene
(97.5 per cent pure) in 13 ml. of ether. This basic solu-
tion was added at room temperature during one-half hour and
the addition was followed by two and one-half hours reflux.
The mixture was then salmon colored and had large amounts
of suspended white solid. A slurry of 1.16 g. (4.00
mmoles) of diphenylcyclopropenyl perchlorate in 25 ml. of
ether was cooled in a dry ice-acetone bath. About 80 per
cent of the suspension of indenyllithium was added to the
stirred slurry in 10 minutes, and the mixture was stirred
in the bath for 20 minutes longer. The almost clear solu-
tion was quenched with dilute acid, washed with brine,
dried over magnesium sulfate, and filtered through Drierite.
The ultraviolet spectrum (MeCN) showed peaks at 312 sh, 320,
and 337 ) in the typical cyclopropene region. Acid had no
effect on the spectrum, suggesting that no symmetrical
diphenylcyclopropenyl ether was present. This species
would result from the aqueous work-up if any cyclopropenyl-
salt remained from the reaction (42,~1). The ether
removed and replaced with pentane and the pentane removed
at reduced pressure. The resulting oily solid was treated
with a little pentane, filtered, and washed one time with
pentane to give 768 mg. (62.6 per cent) of 1,2-diphenyl-3-
(l'-indenyl)-cyclopropene, m.p. 97.5-99.00. Concentration
and cooling of the filtrates gave an additional 129 mg. of
yellowish solid, m.p. 94-990. An analytical sample, m.p.
98-990, resulted from two recrystallizations from ethanol.
Its infrared spectrum (KBr) exhibited significant peaks at
5.24 (aromatic C-H), 5.42 (primarily cyclopropene C-H),
5.50 (cyclopropene C=C), 6.28, 6.70, 6.88, 6.92, 7.52,
7.48, 7.63, 9.32, 9.64, 9.78, 9.91, 10.90, 11.33, 11.87,
12.97, 15.11, 15.51, 15.63, 14.28, and 14.51 p. The
ultraviolet region (isooctane) displayed peaks at 215 sh
(log ( = 4.48), 223 (4.48), 228 sh (4.45), 257 (4.39),
250 sh (4.12), 311 (4.36), 320 (4.43), 326 sh (4.37), and
338 mp (4.30). The n.m.r. absorbances (CC14) were at )'=
2.5 and 2.8 (unresolved aromatic multiplets), 3.43 and 3.69
(broadened doublets of olefinic protons), 6.27 (broad,
aliphatic proton), and 7.51 doublett, aliphatic proton).
The latter four resonances all were of approximately the
Anal. Calcd. for C24H18: C, 94.08; H, 5.92.
Found: C, 94.07; H, 5.71.
Diphenyl(3-or 1-indenyl)cyclopropenyl tetrafluoro-
borate (18b).--A cold solution of 475 mg. (1.44 mmoles) of
triphenylmethyl tetrafluoroborate in 4 ml. of methylene
chloride was added in five minutes to an ice-cold :
of 400 mg. (1.31 mmoles) of indenylcyclopropene 17b in 1l/
ml. of methylene chloride. The reaction mixture was pro-
tected with argon as described before. After remaining in
the ice-bath for 15 minutes, the solution was concentrated
and injected into 25 ml. of cold, dry ether. The mixture
was centrifuged, the red supernatant solution removed, and
the precipitate washed one time with ether to give 395 mg.
(76 per cent) of tan solid after drying with argon. The
solid's infrared spectrum (KBr pellet was deep red) ex-
hibited significant peaks at 3.23 (aromatic C-H), 3.43
(weak shoulder, aliphatic C-H), 5.42, 5.56 sh, 6.28, 6.40,
6.73, 6.94, 7.07 (cyclopropenyl cation), 7.47, 7.67, ca.
9-10 (broad, -BF4) (77), 13.3 (broad), 14.3, and 14.6 )
(sh). The ultraviolet spectrum (MeCN) was obtained from
another similar run and showed absorptions at 264 (log
( = 4.25), 273 sh (4.23), 371 (4.07), and 385 mu sh (4.02).*
Addition of one drop of 48 per cent tetrafluoroboric acid
gave peaks at 248 (4.19), 288 (4.17), and 305 my sh (4.13).
These extinction coefficients are based on the assumption
that the cyclopropenyl salt 18b was pure.
The supernatant ether solution and washings from the
hydride abstraction were evaporated to give 426 mg. of dark
residue. Chromatography on Woelm basic alumina (activity
1) with cyclohexane gave 207 mg. (64.7 per cent) of tri-
phenylmethane, m.p. 89-920. The infrared spectrum was
identical to that of authentic material and the purity of
the sample was confirmed by t.l.c.
teen preparations of this compound were attempted with
varying amounts, solvents, concentrations, work-ups and
degrees of success. Although it did not give the highest
yield of desired species, the following run gave the most
concentrated and probably the purest solution of the
methylenecyclopropene. Triphenylmethyl tetrafluoroborate
A spectrum of the visible region was not obtained.
(220 mg., 0.67 mmole) in 2.0 ml. of methylene chloride was
cooled to 0 and added to an ice-cold, argon-swept solution
of 200 mg. (0.65 mmole) of the indenylcyclopropene 17b in
0.6 ml. of methylene chloride. After reacting at 0 for
about 20 minutes the mixture was injected into 12 ml. of
cold, dry ether under argon. After centrifuging, the red
supernatant solution was removed and the solid washed one
time with 2 ml. of petroleum ether. Drying with argon
followed by vacuum drying gave 192 mg. (0.49 mmole) of the
salt. A slurry of this solid in 2h ml. of carbon tetra-
chloride was prepared under argon, cooled to 00, and
allowed to react with 0.25 ml. of a 1.94 M solution of tri-
methylamine in carbon tetrachloride for several minutes.
The resulting mixture was filtered through a fritted glass
funnel under an argon atmosphere, but the residue was not
washed. This filtration was accomplished through a 30 ml.
suction funnel which had been capped with a sept' ,-
thoroughly flushed with argon using a hypodermic n:eeo.
The mixture was injected into the funnel with a needle-
equipped syringe and filtered under a positive pressure of
argon. The deep red filtrate (1.75 ml.) was shown by its
infrared spectrum to contain a little trimethylamine. Most
of this was removed by bubbling argon through the solution
for ca. one-half hour while maintaining approximately the
same volume by carbon tetrachloride addition. The resulting
1.65 ml. of solution was sampled for infrared and nuclear
magnetic resonance spectra. The infrared spectrum of the
solution showed absorptions at 3.24 (aromatic C-H), 3.58-
3.62 (weak multiple probably due to remaining trimethyl-
amine), 5.41, 5.46, 6.32, 6.47 (strong), 6.70, 6.75, 6.90,
6.95, 7.34, 7.50, 7.66, 9.01, 9.39, 9.96, 10.07, 14.30 and
14.58 p. The n.m.r. spectrum displayed only strong unre-
solved absorption between T = 1.84 and = 3.91, a weak but
sharp absorption at 7'= 7.87 (trimethylamine) (64), and
very weak absorption consisting of a quartet at 7= 6.60
and a triplet at 7 = 8.85. The ultraviolet and visible
spectral data in Table 2 were obtained by diluting aliquots
to the desired concentration. In runs where this data were
gathered, more care was taken to thoroughly wash the red
color into the product solution before diluting it.
Approximate extinction coefficients were obtained by assuming
the indenylcyclopropenyl salt precursor to be pure and the
yield in the proton removal step to be quantitative.
The white solid obtained in the filtration after the
proton removal step was shown to be trimethylammonium tetra-
fluoroborate by its infrared spectrum. In the runs where
the yield of this product was checked, it was usually be-
tween 60 and.80 per cent based on crude cyclopropenyl
The many attempts to isolate pure cyclopropenylidene-
indene 19b were all failures. The difference in the
absorption (AE) at 370 my between neutral and acidic solu-
tions of acetonitrile was used as an analytical tool to
follow loss of the methylene cyclopropene in the following
studies. Removal of the solvent with a stream of argon,
followed by vacuum drying and redissolving the residue
resulted in a loss of over 50 per cent in AE at 370 mp.
The product's infrared spectrum (CC14) exhibited losses in
the peaks at ca. 5.4 and 14.6 p among others while areas at
3.4, 3.5, and 6.03 p were building up in absorption. The
methylenecyclopropene was found to be sensitive to oxygen
also. A dilute solution of 19b in carbon tetrachloride
which was stored in a freezer for 18 hours under an atmos-
phere of air was shown to have lost 62 per cent of the
methylenecyclopropene, as evidenced by its ultraviolet
spectrum. However, even concentrated solutions in carbon
tetrachloride could be maintained with negligible losses
if the solutions were degassed with argon and stored under
an argon atmosphere in powdered dry ice. The solutions
were also unstable to chromatography. Chromatographed
rather rapidly through Woelm neutral alumina with benzene,
the red solution lost 87 per cent of its AE at 370 mu.
Chromatography through silica gel resulted in a loss of ca.
60 per cent. Injections of red solutions into cold petroleum
ether resulted in precipitation of a flocculent, pale tan
solid which defied attempts at purification. The ultra-
violet data indicated that although the solid contained
some of the 570 mp-absorbing material, the overall result
of the precipitation process was loss of large proportions
of the methylenecyclopropene.
Reaction with tetracyanoethylene (TCNE) in aceto-
nitrile at room temperature resulted in a loss of 63 per
cent in AE37o within 15 minutes. There were concurrently
slight, but distinct shifts in peak positions in the
370 my region. The peak at 568 mu shifted to 371 my and
there was the development-of a very distinct shoulder at
ca. 387 mp. There was very little change thereafter, even
after 10 hours, although the presence of excess TCNE was
suggested by the spectral data. There also was no peak in
the visible region in cyclohexane after reaction with TONE.
Chromatography of the reaction mixture on silica gel with
ether-cyclohexane gave an early red cut which yielded a
very small amount of red solid, m.p. 215-2320. It had
infrared absorptions (KBr) at 3.24 (aromatic C-H), 5.41,
6.42, 6.58, 6.71, 6.78, 6.97, 7.50, 7.70, 8.90, 9.24, 9.45,
9.71, 9.84, 10.10, 13.12, 13.40, 13.70, 14.30 and 14.71 p.
Its ultraviolet spectrum showed absorption at 263 (6 = 89
ml./mg.-cm.), 271 sh (89), 370 (60) and 387 mu sh (52).*
Spectra in the visible region were not obtained.
Addition of tetrafluoroboric acid gave peaks at 250 (94),
287 sh (77), and 305 m)i sh (65). Very small amounts of
solids having similar properties were obtained at various
times from various red solutions. For example, elution of
the indenylcyclopropenyl tetrafluoroborate 18b through acid
alumina with methylene chloride and acetonitrile, followed
by concentration and a second chromatography of the red
fraction, eluting with methylene chloride, gave a red solu-
tion. Concentration and redilution with hexane gave ca.
2 mg. of a deep red solid, m.p. 230-2330. This solid had
the same infrared peaks as the aforementioned compound. In
the ultraviolet region the peak at 371 m) was higher than
that at 263 mu. The species was almost impossible to
obtain intentionally and was never isolated in quantity
and purity sufficient for elemental analysis or n.m.r.
1,2-Diohenyl-3- [5' (?)-cyclopentadienyl -eve-'. *-.. ; : -
(17a).--A solution of 82.0 ml. (1.0 mmoles) of freshly
distilled cyclopentadiene in l ml. of dry ether was added
in 10 minutes to a stirred solution of 1.0 mmoles of n-
butyllithium in 3 ml. of ether and ca. Z ml. of hexane.
The mixture was at room temperature during the addition and
during subsequent one-half hour reaction periods. A slurry
of 145 mg. (0.50 mmole) of diphenylcyclopropenyl perchlorate
in 3 ml. of ether was cooled in a dry ice-acetone bath and
treated with ca. 55 per cent of the slurry of cyclopenta-
dienyllithium during 40 minutes. When the mixture was
sampled after one and one-third hours and analyzed to t.l.c.
on silica gel with benzene, one large spot was observed at
Rf- 0.70 and traces of two spots at Rf < 0.10. At t
point the mixture was worked-up with dilute acid and the
organic phase dried with magnesium sulfate and Drierite.
If this solution was concentrated to near dryness, the
material of high Rf-value disappeared leaving only non-
moving spots. The material also reacted very rapidly with
TCIO in acetone. Purification by preparative t.l.c. and
extraction of the front band with a minimum of methylene
chloride gave a solution showing peaks in the ultraviolet
region (MeCN) at 226, 236, 503 sh, 313, and 330 my (rela-
tive E = 1.19, 1.00, 1.45, 1.76, and 1.15). However,
addition of one drop of 48 per cent tetrafluoroboric acid
to the acetonitrile solution produced a spectrum showing
increased absorption at 255 mp and a broad peak with its
maximum at 296 mu and a shoulder at 309 mg. None of the
longer wavelength peaks were present.
Maleic anhydride adduct of 17a.--A dried ether solu-
tion of the cyclopentadienylcyclopropene 17a resulting from
reaction of 0.50 mmole of diphenylcyclopropenyl salt was
treated with 62 mg. (0.64 mmole) of maleic anhydride. The
solvent was removed and the residue triturated with
benzene-pentane to give 131 mg. of white solid. One
recrystallization from benzene-petroleum ether gave 102 mg.
(58 per cent from perchlorate) of a mono adduct of 17a,
m.p. 147-1510. One further recrystallization gave an
analytical sample, m.p. 148.5-149.5. The adduct had
significant absorptions in the infrared region (KBr) at
3.23, .3.39, 3.45 sh, 5.40 and 5.57 (anhydride C=0) (50),
6.29, 6.69, 6.93, 7.39, 7.51, 7.72, 7.81, 8.20, 9.20,
9.50, 10.61, 10.74, 10.97, 13.28, 13.41, and 14.66 p.
Absorptions in the ultraviolet region (MeCN) were at 228
(log C = 4.28), 237 (4.17), 307 sh (4.37), 316 (4.47), and
333 mr (4.35).
Anal. Calcd. for C24H1803: C, 81.34; H, 5.12.
Found: C, 81.12; H, 5.36.
Reaction of 1,2-diphenyl-3-[5' (?)-cyclopentadienyl]-
cyclopropene with triphenylmethyl tetrafluoroborate.--A
solution of the cyclopropene 17a (purified by preparative
t.l.c.) in methylene chloride was treated with triphenyl-
methyl tetrafluoroborate in the same solvent at room
temperature. The appearance of the ultraviolet spectrum
(MeCN) was similar to that of an acid treated sample, that
is, a loss in intensity of absorption in the 290-340 mp
region and a hypsochromic shift. A rather broad peak
centered at 303 my, was the only significant absorption.
In the presence of acid the spectrum appeared about the
same except that the 355 mj peak was larger. Addition of
trimethylamine produces a spectrum having less absorption
at 255 mu and a shoulder at 290 mp, a broad maximum at
297 mp and a shoulder at 318 mp.
When a dark methylene chloride solution of cyclo-
pentadienyl cyclopropene 17a and triphenylmethyl tetra-
fluoroborate was injected into cold dry ether, a greenish-
yellow, gelatinous solid was obtained. After centrifuging
and the removal of the solvent the material was dried with
a flow of argon. The solid darkened rather rapidly after
it was dried. A spectrum of the solid in acetonitrile
showed absorption at 257 (weak), 262 (weak), 300 sh, and
310 mu. Addition of acid did not affect the positions but
reduced the intensity of all absorptions except the one
at 256 mp. When the solid was shaken with trimethylamine
in carbon tetrachloride, a solution was obtained which gave
absorptions at 292 sh, 302, and 317 mp sh. An infrared
spectrum (KBr) on the solid before it became black showed
significant absorptions at 3.23, 3.40, 5.48 (very weak),
6.27, 6.68, 7.00, 7.10, 8.50, 9.5 (strong, broad) (BF4-),
11.91, 13.10, 14.40, and 14.80 u.
Attempted preparation of 2,3-diphenyl-4-carbo-
A solution of 0.750 ml. (5.0 mmoles) of dimethyl bromo-
succinate in 2.0 ml. of ether was added slowly to a stirred
solution of 5.22 ml. (5.0 mmoles) of n-butyllithium solution
(hexane) in 10.0 ml. of ether. The flask was in a dry-ice
acetone bath during this addition and remained there for
five more minutes while the cloudy yellow mixture was
stirred. The mixture was then added to a stirred slurry of
556 mg. (2.0 mmoles) of diphenylcyclopropenyl tetrafluoro-
borate in 8.0 ml. of methylene chloride at the same
temperature. After a reaction time of one hour the mixture
was worked-up with dilute acid solution. Drying the
solution and removal of the solvent left a yellowish oil.
Trituration with ether-pentane gave 48 mg. of tan solid,
m.p. 158-1680 (m.p. of the symmetrical diphenylcyclopropenyl
ether is 163-165 dec.). The ultraviolet spectrum of the
solid also agreed with that expected for the ether. The
filtrates were concentrated and chromatographed on 50 g. of
Woelm acid alumina (activity 1). Elution with 9:1 pentane-
ether gave ca. 20 mg. of unidentified oil, followed by 91 mg.
of white solid. The solid was washed with pentane to give
39 mg. of dimethyl fumarate (no m.p. depression with
authentic material). Elution with ether gave 376 mg. of
yellow oil containing material absorbing at 332 mp, followed
by ca. 220 mg. of unidentified oil. The cuts absorbing at
332 mp, were combined and rechromatographed on 20 g. of
Woelm acid alumina (activity 1). Elution with 3:1 pentane-
ether gave early cuts absorbing at 298 sh, 307 sh, 515 (max.),
and 333 m)i in the typical diphenylcyclopropene region. They
were combined to give 87 mg. of oil (22) which was dis-
solved in acetonitrile, filtered, and reacted with 89 mg.
of triphenylmethyl perchlorate. After a few minutes the
yellow solution was concentrated and diluted with water and
ether. The organic layer was dried and the solvent removed
under vacuum. Multiple washing of the residue with pentane
gave 57 mg. of yellow solid (23). T.l.c. on silica gel
with ether hexane (9:1) showed one large yellow spot, and
lesser amounts of materials at larger and smaller Rf-values
which were visualized with 50 per cent sulfuric acid and
heat. The Rf-value was appreciably larger than that of
the methylenecyclopropene of Jones (56) and about the same
as that of Battiste (u5). The yellow spot was removed and
extracted for an ultraviolet spectrum in acetonitrile. It
showed peaks at 247, 255, 269, 300 sh, 311, 323, and 392 my
(relative C = 3.81, 3.92, 2.98, 2.77, 3.27, 2.59, an.d 1,00).
Addition of one drop of tetrafluoroboric acid gave a spec-
trum with peaks at 250, 295, and 308 mn (relative E = 1.00,
1.47, 1.64). The n.m.r. spectrum (CDC13) of the yellow
solid showed absorption in the aromatic region at 7 = 1.42,
2.11, and 2.45. A singlet was present at ''= 6.91 and a
broad absorption at 7 = 8.50 was smeared into a sharper one
at 7= 9.06. The areas of the absorbances of these three
regions were ca. 30:8.3:66. Recrystallization from methanol
gave 23 as yellow needles, m.p. 174-1770. The infrared
spectrum of the needles in methylene chloride had absorp-
tions at 5.36, 5.49, 5.87 (C=0), 6.35 (strong), 6.72, 6.89,
7.43, 9.62, and 10.16 P.
Anal. Found: C, 79.58; H, 7.97.
ether (8.0 ml.) was added to 1.94 ml. (5.0 mmoles) of n-
butyllithium solution (1.55 M in hexane) and was stirred at
room temperature. To this basic solution was added 0.190 ml.
(5.0 mmoles) of freshly distilled malononitrile in 4.0 ml.
of dry ether. The addition was completed in six minutes and
stirring was continued for 40 minutes longer. About 80 per
cent of the resulting white slurry was added during 40
minutes to a stirred slurry of 772 mg. (2.3 mmoles).of di-
phenylcyclopropenyl hexafluorophosphate in 4.0 ml. of dry
ether. The latter slurry had been pre-cooled in an acetone-
dry ice bath and the resulting temperature was maintained
during the addition and subsequent reaction time. The
mixture was allowed to react for two and one-third hours
after the addition was complete, and then was worked-up in
the usual manner with dilute acid. The solvent was removed
to give a yellowish residue which was chromatographed on
30 g. of silica gel. Elution with benzene gave early cuts
which were indicated by t.l.c. to contain small amounts of
two species which were not investigated. The next several
cuts were combined and the solvent removed to give 441 mg.
of sticky yellow solid, m.p. 107-1220. One recrystalliza-
tion from 95 per cent ethanol returned 332 mg., m.p. 113-
1240. A second recrystallization from benzene-cyclohexane
gave 202 mg. (29 per cent yield) of off-white solid (24)
which was indicated by t.l.c. to be essentially pure. One
further recrystallization from the latter solvent pair gave
a sample of analytical purity, m.p. 126-127.50. The ultra-
violet spectrum (MeCN) had peaks at 225 (log 6 = 4.30), 233
(4.22), 294 sh (4.33), 300 (4.38), 308 (4.49), 315 sh
(4.37), and 325 mu (4.40). Significant absorptions in the
infrared region (KBr) were at 3.23 (aromatic C-H), 3.40
aliphaticc C-H), 4.43 (C=N), 5.11, 5.51 (cyclopropene C=C),
6.75, 6.93, 7.47, 7.60, 9.10, 9.39, 9.62, 9.79, 10.36,
10.92, 11.19, 15.20, 15.42, and 14.60 u.
Anal. Calcd. for C18H12N2: C, 84.35; H, 4.90; N,
10.93. Found: C, 84.53; H, 4.90; N, 11.02.
When the cold reaction mixture containing lithio-
malononitrile and the cyclopropenyl perchlorate 16 was
sampled directly into acetonitrile with a syringe, a broad
peak at 353 mu was observed in addition to those characteri-
zing the cyclopropene 24. The spectrum was not appreciably
affected by addition of tetrafluoroboric acid. When the
mixture was sampled directly into methylene chloride or
acidic acetonitrile no anomalous peaks appeared.
Reaction of the lithio derivative of l,2-diphenyl-3-
dicyanomethylcyclopropene in acetonitrile.--A solution of
50 mg. (0.20 mmole) of dicyanomethylcyclopropene 24 in 1.0
ml. of dry ether and 0.13 ml. of benzene was treated slowly
with 0.22 mmoles of n-butyllithium in hexane while cooling
in an ice-bath. A deep yellow solution with a little floc
resulted. An aliquot of this solution in acetonitrile
gave peaks at 239 (weak), 245 (weak), and 353 mu. Assuming
the complete conversion of starting material to this species
log (353 mp) = 4.2. Sampling into dioxane gave the
typical spectrum of the cyclopropene 24. T.l.c. gave simi-
lar results, that is, sampling of the mixture directly gave
a spot corresponding to 24 and spotting after reaction with
acetonitrile showed no spot at that Rf-value. After three
hours the mixture was poured into 3 ml. of acetonitrile,
concentrated, diluted with methylene chloride and prepara-
tively chromatographed on a thin layer plate. At least six
bands were observed. The third band from the front gave
18 mg. of yellow solid after solution and solvent removal.
After trituration with benzene-pentane a yellow solid, m.p.,
160-1700 was obtained. Its ultraviolet absorptions were at
239, 245, and 353 m, in acetonitrile and 240, 245 sh, and
350 mu in dioxane. Significant infrared absorptions (KBr)
were at 3.24, 3.40 sh, 4.48 (C=N), 6.25, 6.48 (strong),
6.95, 7.10, 7.40, 7.59, 7.66, 7.75, 8.20, 8.49, 8.65, 9.12,
9.35, 9.70, 10.05, 10.48, 11.16, 11.81, 13.10, 13.66, 14.28,
and 14.58 1.
Attempts to prepare diohenyldicyanomethylcyclopropenyl
salts.--As evidenced by ultraviolet analysis, no reaction
occurred when dicyanomethylcyclopropene 24 and triphenyl-
methyl tetrafluoroborate were dissolved in acetonitrile or
methylene chloride at room temperatures. Cyclopropene 24
(10 mg.) and 14 mg. of triphenylmethyl tetrafluoroborate
were dissolved in 0.05 ml. of acetonitrile, covered with
argon, and hated to 70-800 in a water bath. The reaction
was followed by t.l.c. and spectral analysis. After ca.
one week the ultraviolet spectrum was essentially that of
the starting cyclopropene except for a low broad shoulder
at ca. 350 my. A thin layer chromatogram showed starting
material, a small spot at an Rf-value corresponding to
triphenylmethane, and a small yellow spot superimposed on
starting material spot, in addition to at least four spots
of smaller Rf-value. Addition of more hydride abstractor
seemed to have little if any effect.
Reaction of the dicyanomethylcyclopropene 24 with
2,3-dicyano-5,6-dichloroquinone (57) in benzene at room
temperature showed no loss of starting material and no build-
up of long-wavelength absorbing species. In toluene at
70-800 starting material was consumed in one hour. The thin
layer chromatogram and ultraviolet spectrum were rather
The latter reaction in acetic acid (57) (heating for
ca. five minutes) also destroys starting material but pro-
duces no encouraging data.
In a manner analogous to the method used for the dicyano
derivative, diethyl lithiomalonate was prepared by the
reaction of 0.75 mmole of diethyl bromomalonate with 0.75
mmole of n-butyllithium in 3 ml. ether and ca. 0.4 ml. of
hexane. After reacting for 30 minutes at room temperature
the mixture was added during 15 minutes to a stirred slurry
of 145 mg. (0.50 mmole) of diphenylcyclopropenyl perchlorate
in 3.0 ml. of ether (dry ice-acetone bath). The reaction
was quenched after one and one-half hours with dilute acid,
the organic phase dried, and the solution concentrated.
T.l.c. of the solution demonstrated at least six components
with one definitely in major proportions. Preparative t.l.c.
on silica gel with methylene chloride and extraction of the
band at the second highest Rf-value gave 145 mg. of oil
(25). A thin layer chromatogram developed with benzene
showed only one spot that quenched the fluorescent indicator.
Visualization with iodine vapors showed two small spots at
lower Rf-value. The ultraviolet spectrum of the oil in
acetonitrile showed peaks at 226, 233 sh, 295 sh, 303 sh,
310, and 327 mp (relative E = 1.24, 1.00, 1.18, 1.31, 1.59,
and 1.18). This compound was not completely characterized.
No loss of the ultraviolet peaks or buildup of any
new ones was observed when the oil and triphenylmethyl
tetrafluoroborate were dissolved in acetonitrile at room
temperature. There also was no apparent reaction between
2,3-dichloro-5,6-dicyanoquinone (DDQ) in benzene at room
temperature when the solution was analyzed in the same
Reaction of diazomethane and 1,2-diphenyl-4-carbo-
A 0.252 gm. (0.75 mmole) portion of methylenecyclopropene
10b (prepared by the method of Jones and Denham) (56) was
dissolved in 2.5 ml. of tetrahydrofuran, which had been
dried over 4A sieves and passed through Woelm basic alumina
(activity 1). This solution was in a small flask equipped
with a drying tube, stirring bar, and a side arm septum.
The yellow solution was cooled in an ice-salt bath and a
10 per cent excess of diazomethane in ether (ca. 0.4 M
solution) was added. During the course of the reaction,
the color changed from bright yellow to deep red. The
progress of the reaction was followed by the loss of
methylenecyclopropene peaks in the ultraviolet spectrum and
by thin layer chromatography (t.l.c.) on silica gel. After
three hours t.l.c. indicated that yellow starting material
was still present, so another small portion of diazomethane
was added. After four and three-fourths hours, when the
methylenecyclopropene was indicated to be all consumed, the
mixture was evacuated with stirring and the solvent .nd
excess diazomethane were removed. A second evacuation was
performed after addition of 0.7 ml. of carbon tetrachloride.
The flask remained in the bath during both evacuations.
The red oil was transferred in 1l ml. of carbon tetra-
chloride which was then divided into two portions and was
frozen in dry ice-acetone to await further examinations.
T.l.c. of the red solution with ether-hexane (3:2)
as developer showed a fast-moving deep red spot as the
major produce and a trace of yellow methylenecyclopropene
10b with a smaller Rf-value. Development with 50 per cent
H2SO4 produced a small spot just behind the main spot. An
n.m.r. spectrum of the solution was obtained in carbon
tetrachloride at -130 and was run promptly after preparation.
It showed unresolved absorptions at T = 2.67 and 2.82
(aromatic protons) and two smaller ones at 7 = 4.47 and
4.74 (terminal olefin protons) (rel. areas = 10:0.87:0.87).
At Y= 4.23 and 4.59 were even smaller broad resonances of
unequal intensity. Between 7= 6.43 and 6.90 aliphaticc
region) were six absorbances of varying intensity with their
total area about the same as that in the aromatic region.
There was a quintet at T= 8.29 of ca. 1/8 of the area of
the aromatic proton absorptions. Significant peaks were
present in the infrared region (CC14) at 4.85 (-C=N2) (74,
7D), 5.71 (unconjugated C=O), 5.85 (conjugated C=0), 6.35
(broad), 6.70, 6.99, 7.49, 7.80, 8.11, 8.37, 8.54, 8.80,
9.37, 9.99, 10.30, 10.98, and 14.40 p (broad). Absorption
in the ultraviolet and visible region (MeCN) was present
at 226, 241 sh, 286, 330 sh, and 485 my (relative E = 1.49,
1.42, 1.00, 0.46, and 0.01).
A red solution of this product (27) in carbon tetra-
chloride was essentially decolorized in one-half hour at
reflux temperature, overnight at room temperature, or in a
month in a freezer. All attempts to crystallize the red
material met with failure.
methylpyrazolenine (?) (28).--A red carbon tetrachloride
solution of the diazomethane 27 was allowed to sit in a
freezer for 38 days. The faintly colored solution was
then chromatographed on 20 g. of silica gel in an ice-water
jacketed column. Elution with pentane-ether (9:1, followed
by 4:1) gave some small oily fractions followed by several
cuts which were indicated by t.l.c. to contain only one
component in appreciable amounts. This species has been
tentatively assigned structure 28. The colorless thin layer
spot of 28 became deep red when treated with 50 per cent
H2S04 and heated. The Rf-value of the compound was slightly
less than that of 27 when developed with hexane-ether.
This product was obtainable only as a moderately stable
oil which could be maintained with little loss only at
freezer temperatures or below.
The purest chromatography cuts were concentrated and
the solvents replaced with carbon tetrachloride for spectral
data. T.l.c. indicated the presence of two minor impurities
in the resulting solution. The n.m.r. spectrum of the
solution at -50 showed absorptions at T = 1.99 and 2.83
(aromatic protons) and at T = 4.26 and 4.80 (terminal ole-
finic protons) with areas in the ratio of 2:8:1:1. Two
resonances of unequal area at T = 6.55 (methoxyl protons)
and 6.70 (methoxyl and methylenic protons) had a total
area of 7.6 relative to 10 for the aromatic protons. In
addition there were small unassigned absorptions at r =
6.35, 7.17, 7.4b, 8.02, and 8.83 (total relative area ca.
2.7). The infrared spectrum (CC14) exhibited significant
peaks at 3.24, 3.34, 3.48, 5.71 (sharp) (unconjugated C=O),
6.29, 6.38, 6.70, 6.98, 7.15, 7.40, 8.15, 8.15, 8.30, 9.20
(broad), 9.75, 9.91, 10.61, 10.95, 11.53, 14.30, and 14.45 p.
Absorption in the ultraviolet region (MeCN) appears at 242
and 303 my with relative extinction coefficients ca.
Chromatographic cuts subsequent to those containing
28 in the separation above were found to contain an in-
creasing amount of a material which gave a second red spot
(after H2SO4 treatment) with an even smaller Rf-value on
t.l.c. (see below).
styryl)-5(or 3)-phenylpyrazole (26).--Methylenecyclopropene
10b was dissolved in a minimum of tetrahydrofuran and cooled
in an ice-bath. A 10-20 per cent excess of a 0.36 M solu-
tion of diazomethane in ether was added with stirring.
T.l.c. and ultraviolet spectra indicated that the reaction
was essentially completed at 00 after two hours. The
solvent and excess diazomethane were removed in an ice-bath
by evacuating with an aspirator. The residual red oil was
dissolved in 30 ml. of carbon tetrachloride and brought
rapidly to reflux temperature in a dark flask with a
septum-covered side arm. Periodic samples were cooled
rapidly and stored in a freezer prior to analysis by t.l.c.
The latter technique definitely demonstrated that the
initial red product 27 is converted successively to a second
compound 28 and then to a third compound 26 which was stable
at this temperature. The conversion to pyrazole 26 was
complete in one hour leaving a yellow solution due to traces
of methylenecyclopropene 10b. The solution was concentrated
and chromatographed on 20 g. of silica gel, eluting with
pentane-ether (9:1, 4:1, and then 2:1). The latter solvent
mixture gave several yellow cuts and washing the residue
with pentane-ether gave 171 mg. (76 per cent) of off-white
solid, m.p. 100-105. Recrystallization from 95 per cent
ethanol gave 160 mg. of pyrazole 26, m.p. 103.5-104.5.
Some other apparently identical preparations melted at 112-
1140. The lower melting form could be converted to the
higher melting form by finely pulverizing it. Its n.m.r.
spectrum (CDC13) showed absorption at T= 2.41 and 2.79
(10 aromatic protons in unresolved multiplets), T = 4.09 and
4.72 (one olefinic proton in each doublet, J = 1.35 c.p.s.),
Z= 5.96 and 6.05 (total of five protons -- one methoxyl
group and the methylene group), and T = 6.55 (three protons
in methoxyl group). Significant peaks in the infrared
region (KBr) were at 3.26, 3.34, 5.71 (with slight"
shouldering) (C=0), 6.24 (broad), 6.36, 6.69, 6.81, 6.90,
7.13, 7.32, 7.74, 8.24, 8.55, 9.00, 9.11, 9.66, 9.79, 9.97,
10.45, 10.85, 10.97, 11.17, 11.76, 12.21, 12.60, 12.73,
12.90, 13.00, 13.47, 14.16, 14.30, and 14.43 p. However,
in nujol mull the carbonyl peak was distinctly shouldered
absorbing at 5.71 and 5.74 u (sh). In methylene chloride
it was sharp at 5.71 p. The ultraviolet spectrum showed
one rather broad peak, MeCN = 257 my (log E = 4.40),
one rather broad peak, max
keOH = 258 mm (log = 4.37).
Anal. Calcd. for C22H20N204: C, 70.20; H, 5.56;
N, 7.44. Found: C, 69.92; H, 5.57; N, 7.29.
Hydrolysis of pyrazole 26.--A solution of 89 mg.
(0.25 mmole) of pyrazole 26 and 0.58 mmole of KOH in 2.6
ml. of methanol was refluxed for four and one-half hours.
After the solvent was removed under vacuum, the residue was
dissolved in 6 ml. of water and extracted two times with
ether. The aqueous solution was acidified with 2 I y::o-
chloric acid and extracted with methylene chloride. After
drying the solution, evaporation of the solvent and washing
the residue with pentane-ether gave 62 mg. (86 per cent
yield) of 5(or 5)-carboxymethyl-4-(3-styryl)-5(or 5)-
phenylpyrazole (36). The melting point was 166-1680 and
two recrystallizations gave an analytically pure sample,
m.p. 167-168.5. Significant infrared absorptions (KBr)
were at 2.99 (N-H), 3.27 (aromatic C-H), 3.41 aliphaticc
C-H), 4.05 (broad) (acid 0-H), 5.51 (broad), 5.85 (acid
C=0), 6.19, 6.49, 6.68, 6.96, 7.15, 7.25, 7.42, 7.58, 7.91,
8.15, 8.59, 8.77, 9.10, 9.38, 9.77, 10.00, 10.16, 10.62,
11.09, 12.06, 12.85, 13.00, 13.6 (broad), 13.70, 14.30,
and 14.47 p. The ultraviolet spectrum (MeOH) shows
absorption at 248 my (log f = 4.29).
Anal. Calcd. for C19H10N202: C, 74.98; H, 5.30;
N, 9.20. Found: C, 75.10; H, 5.44; N, 8.97.
Base-catalyzed reaction of pyrazole 26 in methanol.--
The progress of the hydrolysis of 26 with equivalent amounts
of KOH in methanol was followed by t.l.c., with visualization
by quenching of the fluorescent indicator in the absorbent.
Development of the spots with ether-acetone (1:1) on silica
gel showed compound 37 having an Rf-vale of 0.80 and the
hydrolysis product having an Rf-value < 0.05. However,
when the basic solution was spotted immediately after dis-
solution of 26, a spot was present at 0.75 which disappeared
during the course of the hydrolysis with concurrent increases
in the low Rf-value spot. A solution of 26 and KOH in
methanol in an equivalent ratio of 20:1 gave similar results
when analyzed by t.l.c., that is, only one spot was visible
under ultraviolet light and that was of lower Rf-value than
that of 26.
The 1-carbomethoxypyrazole 26 (75 mg., 0.199 mmole)
was pulverized and covered with 1! ml. of dry methanol. A
solution of ca. 1 mg. of sodium methoxide in 1 ml. of metha-
nol was added. The solid all dissolved rapidly and t.l.c.
of a sample of the solution showed one spot at Rf-value
0.62 (pyrazole 26 Rf-value = 0.70). The solution was
concentrated and diluted with carbon tetrachloride prior to
preparative t.l.c. on silica gel (ca. 1.0 mm. layer) with
acetone-ether (1:1). Extraction of the front band with
acetone gave ca. 55 mg. of a viscous oil (37). The oil
could not be made to crystallize. Significant peaks in the
infrared region (C001) were at 3.14 (N-H), 3.29, 5.40,
5.71 (0=0), 6.23, 6.39, 6.69, 6.96, 7.58, 7.74, 7.98, 8.34,
8.59, 9.06, 9.41, 9.77, 9.95, 10.20, 11.05, 14.21, and
14.43 )u. The nuclear magnetic resonances (CC14) were at
7= 1.41 (N-H singlet), 2.61 and 2.90 (unresolved aromatic
multiplets), 4.29 and 4.85 (slightly split olefinic
absorptions), 6.57 (methoxyl and methylenic protons) and
8.78 (complex absorptions. The relative areas were 1.2,
10 (total aromatic), 0.8, 0.8, 4.3, and 0.6. The ultra-
violet spectrum was obtained by dissolving a known amount
of pyrazole 26 in methanol, measuring the absorbances and
then adding a trace of sodium methoxide to the solution.
After shaking the solution for a few seconds, another
spectrum was obtained. maxH = 250 m (log C = 4.33).
This peak was unchanged in position or height after several
In one case, a small-scale base-catalyzed reaction
was run in a methanol-ether mixture. T.l.c. of the solu-
tion again showed only one spot. However, upon vapor phase
chromatography (5' x 1/8" column of 6 per cent apiezon L on
Chromsorb Z at 300) of the solution a peak appeared at a
retention time that was the same as that of dimethyl
LIST OF REFERENCES
1. D. Ginsberg, Ed., "Non-benzenoid Aromatic Compounds,"
Interscience, New York, 1959.
2. M. E. Vol'pin, Russian Chem. Reviews,.29, 129(1960).
3. W. Baker and J. F. W. McOrmie in "Progress in Organic
Chemistry," Vol. 111, J. W. Cook, Ed., Academic Press,
New York, 1955, P. 44.
4. K. Hafner, Angew. Chem., internal. Ed., 1, 165(1964).
5. K. Hafner, et al., ibid., 2, 123(1963).
6. T. Nakajima and S. Katagiri, Bull. Chem. Soc. Janan,
7. E. D. Bergmann in "Progress in Organic Chemistry,"
Vol. 111, J. W. Cook, Ed., Academic Press, New York,
1955, p. 44.
8. P. Frangois and A. Julg, J. chim. phys., 57, 490(1960).
9. A. J. Silvestri, L. Goodman, and J. A. Dixon, J. Chem.
Phys., 36, 148(1962).
10. E. Sturm and K. Hafner, Angew. Chem., internal. Ed.,
11. G. Berthier and B. Pullman, Trans. Far. Soc., 45,
12. A. Pacault, Bull. soc. chim. France, 16, D371(1949).
15. W. von Doering and D. W. Wiley, Tetrahedron, 11,
14. M. Yakamawa, et al., J. Am. Chem. Soc., 82, 5665(1960).
15. D. J. Bertelli, C. Golino, and D. L. Dreyer, ibid.,
16. W. von Doering in "Theoretical Organic Chemistry
(Kekule Symposium Butterworths, London, 1959, P. 35.
17. E. C. Schreiber and E. I. Beker, J. Am. Chem. Soc.,
18. P. L. Pauson and W. J. William, J. Chem. Soc., 4153
19. J. F. Tinker, J. Chem. Phys., 19, 981(1951).
20. G. W. Boyd and I. M. Jackman, J. Chem. Soc., 548
21. A. LUttringhaus, E. Futterer, and H. Prinzbach,
Tetrahedron Letters, 1209 (1965).
22. J. A. Berson, E. M. Eveleth, and F. Hamlet, J. Am.
Chem. Soc., 82, 3795(1960).
23. J. A. Berson and E. M. Eveleth, Chem. Ind., 1362
24. J. A. Berson and E. M. Eveleth, ibid., 901 (1959).
25. E. M. Eveleth, Jr., J. A. Berson, and S. L. Manatt,
Tetrahedron Letters, 5087 (1964).
26. H. Prinzbach and W. Rosswog, Angew. Chem., 73, 545(1961).
27. H. Prinzbach, ibid., 73, 169(1961).
28. H. Prinzbach and W. Rosswog, Tetrahedron Letters, 1217
29. H. Prinzbach, Angew. Chem., internal. Ed., I, 319(1964).
30. A. Julg, J. chim. phys., 50, 652(1955).
31. G. Berthier and B. Pullman, Bull. soc. chim. France,
16, D457 (1949).
32. J. D. Roberts, A. Streitwieser, Jr., D. M. Regan,
J. Am. Chem. Soc., 74, 4579(1952).
33. 0. Chalvet, R. Daudel, and J. J. Kaufman, J. Phys.
Chem., 68, 490(1964).
34. A. S. Kende, J. Am. Chem. Soc., 85, 1882(1963).
35. M. A. Battiste, ibid., 86, 942(1964).