Chemistry of the Highly Strained Alkene Perfluorobicyclo[2.2.0]hex-1(4)-ene

Article abstract of DOI:10.1002/ejoc.201800058

Several varieties of cycloaddition as well as nucleophilic and free radical addition to the highly reactive central bond of the title fluorocarbon (1) are described. [2 + 2] Cycloaddition to 1 has made available four of the five known representatives of the [2.2.2]propellane ring system, which are rare because of their extreme strain (in the absence of fluorine substitution) and high reactivity. Two of these, a propellanone and a propellene, are first reported here. Quantum mechanical calculations have shed light on many of these transformations.

Full text of DOI:10.1002/ejoc.201800058

A Journal of
Accepted Article
Title: Chemistry of the Highly Strained Alkene
Perfluorobicyclo[2.2.0]hex-1(4)-ene
Authors: Christopher P Junk, Yigang He, Yin Zhang, Joshua R Smith,
David A Dixon, Monica Vasiliu, and David M Lemal
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Org. Chem. 10.1002/ejoc.201800058
Link to VoR: http://dx.doi.org/10.1002/ejoc.201800058
Supported by

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Chemistry of the Highly Strained Alkene
Perfluorobicyclo[2.2.0]hex-1(4)-ene
Christopher P. Junk,‡ Yigang He,‡ Yin Zhang,‡ Joshua R. Smith,‡ David A. Dixon, ‡‡
Monica Vasiliu,‡‡ and David M. Lemal‡*
Abstract
Several varieties of cycloaddition as well as nucleophilic and free radical addition
to the highly reactive central bond of the title fluorocarbon (1) are described. [2 + 2]
Cycloaddition to 1 has made available four of the five known representatives of the
[2.2.2] propellane ring system, rare because of their extreme strain (in the absence of
fluorine substitution) and high reactivity. Two of these, a propellanone and a propellene,
are first reported here. Quantum mechanical calculations have shed light on many of
these transformations.
______________________________________________________________________
‡ Department of Chemistry, Dartmouth College, Hanover, NH 03755 (USA).
https://home.dartmoutth.edu
‡‡ Department of Chemistry, University of Alabama, Tuscaloosa, AL 35487 (USA)
*Corresponding author e-mail address: david.m.lemal@dartmouth.edu
Introduction
1
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Previous reports have presented syntheses, properties, and aspects of the chemistry
of the title alkene 1.1-6 The compound is a very reactive and versatile
electrophile, yet it is also strikingly robust.7 That stability opened a wide temperature
window for exploration of its reactivity, making possible, for example, Diels-Alder
addition to benzene.3 Its behavior is the more remarkable when compared with that of
the parent hydrocarbon, which dimerizes and polymerizes at temperatures below zero.8
The present account describes further transformations of 1, all of which were carried out
near r.t. or below. Reactions were run either with isolated alkene prepared by flash
vacuum pyrolysis of its Diels-Alder adduct with N-benzylpyrrole, or in situ by reduction
of a dihalide adduct.6
Results and Discussion
Quantum mechanical calculations have been performed to enhance understanding
of both newly and previously reported reactions of the alkene. All were carried out
initially at the B3LYP/cc-pVDZ+ level of theory,9 but it became apparent that some of
the computed energies were grossly inconsistent with experimental findings. The earlier-
described3 Diels-Alder addition of alkene 1 with benzene to yield adduct 2 was an
2
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
egregious case, as the reaction that afforded adduct in 65% yield (120 oC, 6 h) was
predicted to be endoergic by nearly 20 kcal/mol. Additional calculations for this
cycloaddition were therefore performed with an array of theoretical methods (Table 1),
optimizations starting from the B3LYP/cc-pVDZ+ results. The only reasonable results
were obtained with the MO6/DZVP2 and composite correlated molecular orbital theory
G3MP2 levels (see Experimental Section for discussion of the computational details).
These are the methods used later in this account, and energies obtained with the B3LYP
functional are also reported for the sake of comparison.
Table 1. Benzene adduct formation energies in kcal/mol
Level of Theory
B3LYP/DZVP2
BP86/DZVP2
PW91/DZVP2
PBE/DZVP2
M06/DZVP2
SVWN/DZVP2
G3MP2
∆H0K
-6.8
∆H298K
-8.0
∆G393K
14.7
12.7
6.9
-8.8
-9.9
-14.6
-13.9
-23.7
-43.0
-25.8
-15.8
-15.1
-24.8
-44.2
-27.0
7.7
-2.5
-21.5
-4.4
Non-cyclic Additions. All additions to alkene 1 occur in syn fashion; anti addition
would create a very strained trans 4/4 ring fusion. When the alkene was generated in
methanolic acetonitrile by treatment of bromoiodo precursor 3 with mercury in an
ultrasonic bath, methanol adduct 4 was formed in high yield (Scheme 1).1 A similar
experiment with aqueous acetonitrile afforded monocyclic gem-diol 6,1 the consequence
3
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
of ring opening of the initial water adduct 5 with loss of HF followed by hydration of the
resulting ketone.
Scheme 1
Seeking to understand the mechanism by which 5 (and 4) were formed, we
considered the possibility of concerted cycloaddition by a water molecule and by water
dimer. Calculations revealed pseudopericyclic transition states represented by 7 and 8
(Figure 1).10-12 Activation enthalpies and free
these processes are
presented in Table 2, where it is clear that concerted addition of a single water molecule
at r.t. is out of the question. On the other hand, reaction with water dimer to yield adduct
5 plus a molecule of water is very facile as evaluated at the G3MP2 level of theory, and
we propose that this is how the addition occurs.13
4
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Figure 1. Two depictions of the transition states for addition of water (left) and of water
dimer (right) to alkene 1 (B3LYP/DZVP2). A single negative frequency was found for
each transition state.
Table 2. Energetics of water and water dimer addition to alkene 1 in kcal/mol.
B3LYP/DZVP2
∆H‡ ∆G‡
M06/DZVP2
∆H‡ ∆G‡
G3MP2
∆G‡
Transition State
∆H‡
298K
298K
298K
298K
298K
298K
5
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Water TS 7
35.2
11.6
47.4
27.4
30.2
7.5
42.9
22.5
36.6
-5.0
48.7
10.5
Water dimer TS 8
When dibromide 9 was treated in an ultrasonic bath with zinc dust in acetonitrile
containing a little water, adduct 6 was formed again via 1, but the dominant product with
this more electropositive metal was dihydro compound 10 (Scheme 2).14,15 Probably
Scheme 2
generation of 1 was followed by reduction to radical anion 11 and/or organozinc anion
12, then protonation by water.
As described below, 1,3-cyclohexadiene reacts readily with alkene 1, but the
reaction fails completely when 1 is generated in situ from dibromide 9 and zinc in
acetonitrile. If the solvent is wet, 10 and 6 are formed again,16 and if anhydrous no
product is obtained because the alkene is destroyed by zinc. Much of the chemistry of 1
explored using this method was carried out before the alkene was isolated, and the
method was only effective when the desired reaction competed successfully with attack
on the alkene by zinc.
6
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Ethyl formate added spontaneously to isolated 1 in ether at r.t. Ester 13 was the
sole product by 19F NMR and was isolated by preparative GC (Scheme 3). When the
reaction was carried out in CDCl3, it stopped after low conversion and deuterium
Scheme 3
replaced the bridgehead hydrogen in 13. The same thing happened in THF-d8, but now
the reaction was much faster and stopped after about 60% conversion. Apparently there
was a trace of a reactive initiator present that set off a radical chain process involving the
solvent. Presumed chain propagation steps are shown in Scheme 4.17
Scheme 4
In the Paterno-Büchi reaction, ultraviolet irradiation of an aldehyde or ketone
results in [2 + 2] cycloaddition to an alkene, yielding an oxetane.18 Attempts to carry out
such an addition to isolated 1 with hexafluoroacetone and with benzophenone resulted in
no reaction, but benzaldehyde did react, albeit to give an unexpected product. Addition
7
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
of the aldehyde C-H bond occurred across the alkene double bond to afford ketone 15,
presumably the result of a light-initiated radical chain process (Scheme 5). No reaction
took place in the absence of UV irradiation.
Scheme 5
Addition of methyllithium to 1 and its consequences have been described
elsewhere.6
[4 + 2 ] Cycloadditions. When dibromide 9, zinc dust, isobutene, and acetonitrile
were combined in a sealed tube and subjected to ultrasonic irradiation in a water-cooled
bath, ene product 16 was formed in high yield by NMR (Scheme 6). Similarly, isolated
Scheme 6
1 combined with 1,4-cyclohexadiene in ether at r.t. in another ene reaction to afford 17
(Scheme 7). Because this adduct has a chiral center, all eight fluorines are inequivalent,
Scheme 7
8
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
with the result that the 19F NMR spectrum is a very complicated combination of two
narrow AB quartets and two AX patterns. While alkene 1 failed to react with 1,3-
cyclohexadiene when generated from dibromide 9 with zinc, as noted earlier, isolated 1
added to the diene in ether at rt, quantitatively by NMR (Scheme 8). Both ene reaction to
Scheme 8
give 18 and Diels-Alder addition to afford 19 occurred, with the former predominant.
Steric hindrance by the flanking fluorines in the Diels-Alder transition state may explain
the preference for the ene reaction. In contrast to its isomer 17, the 19F NMR spectrum of
achiral 18 comprised a single AA'BB' quartet and AA'XX' pattern. Derivatives of 1 with
a bridgehead hydrogen such as 16-18 characteristically fragment in the mass
spectrometer by splitting down the middle to produce a prominent peak at m/z 113. This
+
allylic cation, C3HF4 , is usually the base peak.
Bromine reacted readily with 18 at -45 oC in the dark to give 20 (Scheme 9).
Scheme 9
9
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Adduct 19 was inert under these conditions, presumably because the backside of the
double bond was shielded from attack. Vacuum transfer at rt separated the Diels-Alder
adduct from the nonvolatile tetrabromide, from which 18 could be regenerated with zinc.
In the presence of light, bromination of 19 did occur at rt, affording 21 (Scheme 9). This
was a radical chain reaction, as it failed in the dark. Catalytic hydrogenation of 19
yielded symmetrical (C2v) molecule 22, the crystal structure of which is presented in the
Supporting Information (Scheme 10).
Scheme 10
Diels-Alder reactions of 1 with furan, pyrroles, and aromatic hydrocarbons have
been reported elsewhere.1,3,6 Analogous addition to cyclopentadiene at rt to give 23 was
rapid and quantitative by NMR (Scheme 11). Butadiene reacted at r.t. with 1 generated
Scheme 11
10
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
in situ from dibromide 9 and zinc, affording Diels-Alder adduct 24 (Scheme 11).
Because butadiene exists ~97% in the s-trans form at rt,19 if reaction had occurred
stepwise, the principal diradical that arose would have been 25 (Scheme 12). Bond
rotation in the allylic moiety of 25 required to form 24 would have faced a
Scheme 12
barrier of ~14 kcal/mol,20 too high to compete with ring closure to the
[2.2.2]propellane 26.21 Thus, formation of 24 instead of 26 constitutes strong evidence
that the Diels-Alder reaction occurred concertedly.
Diazomethane added instantaneously to 1 to give 1-pyrazoline 27, and 2-
diazopropane reacted similarly, affording adduct 28 (Scheme 13). Attempted purification
of 27 by gas chromatography resulted in isomerization to 2-pyrazoline 29. Both 27 and
28 were labile compounds that decomposed upon standing at rt.
Scheme 13
11
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
We hoped that photolysis of pyrazoline 27 would produce the extremely strained
[2.2.1]propellane 31 via 1,3-diradical 30 (Scheme 14).22 Though 31 might not
be observed directly under our reaction conditions, low temperature photolysis or
trapping experiments should reveal it if formed. Nature decided otherwise, as strain-
Scheme 14
releasing ring scission instead of ring closure occurred to give methylenecyclohexane 33,
and 28 reacted analogously (Scheme 15). This might have proceeded via ring opening
of diradical 30 to carbene 32a, which underwent sequential abstraction of hydrogen
atoms from the ether solvent. Reviewers have proposed that the electrophilic tertiary
radical center in diradical 30 abstracted a hydrogen atom from ether to give 32b; ring
opening to 32c and a second hydrogen transfer followed. Though perfluoroalkylcarbenes
are capable of both C-H insertion and hydrogen abstraction,23 the absence of insertion
product here argues for the pathway that bypasses the carbene.
Scheme 15
12
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
[2 + 2] Cycloadditions. Arguably, the most interesting transformations of alkene 1
are [2 + 2] cycloadditions because they create the highly strained [2.2.2]propellane ring
system. The literature offers only a single well authenticated example of a
[2.2.2]propellane other than those derived from 1. In 1973, Eaton and Temme
synthesized 34, a very labile compound with a half-life for ring-opening to 35 and 36 of
28 min at 25 oC (Scheme 16).24
Scheme 16
Such ring-opening is orbital topology-forbidden to occur concertedly.10 Hoffmann25 and
Newton26 showed that the process begins with rate-determining stretching of the very
strained central C-C bond, illustrated with the unknown parent [2.2.2]propellane 37,
which results in an electronic state crossing to antisymmetric singlet 38 (Scheme 17). As
the bond lengthens, through-space interaction between the bridgehead carbons loses out
to through-bond interaction with the three bridges that both stabilizes the antisymmetric
and destabilizes the symmetric state of the developing diradical. Subsequent scission of
an external C-C bond in 38 to give 39 is orbital topology-allowed and highly exothermic.
Scheme 17
13
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Ethyl vinyl ether reacted with alkene 1 to give [2.2.2] propellane 40.1 The reaction
was first performed with alkene generated in situ from a diiodo precursor and mercury in
acetonitrile in an ultrasonic bath. A much better yield of 40, 77% after purification, was
obtained using dibromide 9 and zinc under similar conditions to produce the alkene
(Scheme 18). In light of the fact that formation of propellane 40 is an orbital topology-
forbidden reaction, it is striking that it takes place so readily.
Scheme 18
The propellane suffered ring opening at r.t. to give dimethylenecyclohexanes 41 and
42 in the ratio 2.5 :1 (Scheme 19). The first order decomposition had a half-life of 40.1 h
at 21 oC, signifying a considerable increase in thermal stability relative to Eaton's
propellane 34. Since the enthalpy of 42 is 13.2 kcal/mol more positive than that of 41
(G3MP2), the fact that they were formed in comparable amounts indicates that, in accord
with the Hammond postulate, the second transition state comes very early with a low
barrier. Formation of 41 from 40 was exothermic by 46.1 kcal/mol (G3MP2).
14
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Scheme19
Whereas Eaton's propellane 34 added bromine across the central bond
instantaneously at -78 oC,24 propellane 40 resisted attack by bromine in CH2Cl2 at rt, even
with irradiation. The same was true with iodine in CCl4. An attempt to attack 40 with
electrophilic iodine using iodine and silver acetate in acetonitrile also resulted in no
reaction. Impressively. trifluoroacetic and even concentrated sulfuric acid failed to react
with the propellane. So 40 is immune to assault by both electron-deficient free radicals
and powerful electrophiles.
The story with nucleophiles is very different. Treatment of propellane 40 with
tetrabutylammonium iodide and iodine in acetonitrile at r.t. produced diiodide 43
(Scheme 20). Attack by iodide ion opened the central bond, and the resulting anion was
captured
Scheme 20
by iodine. When iodine was omitted and the solvent was not dry, the product was HI
15
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
adduct 44 (X = I), as the intermediate anion was protonated by water (Scheme 21). The
reaction was regiospecific, and a NOESY experiment revealed that halide had attacked
distal to the oxygen. Thus, stabilization of the intermediate anion by oxygen's
electronegativity was more important in determining the course of reaction than the
relative electron density at the bridgehead carbons in the starting material.
Tetrabutylammonium bromide, chloride, and fluoride gave 44 (X = Br, Cl, and F,
respectively) under similar conditions.27
Scheme 21
The ability of fluorine substituents to stabilize a [2.2.2]propellane suggested that
other electron-withdrawing groups such as carbonyl might have a similar effect.
Dimethylketene, generated by reduction of -bromoisobutyryl bromide with zinc,28
reacted in situ with alkene 1 below r.t. to yield [2.2.2]propellanone 45 (Scheme 22). The
X-ray crystal structure and data for this ketone are presented in the Supporting
Scheme 22
Information. We were surprised to find that a solution of 45 in CDCl3 was unchanged
after > 4 months at rt. Then, when a neat sample was heated at 120 oC for 5 h in a sealed
16
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
tube, 38% of the propellanone survived.
Initial efforts to combine even more electron-deficient addends with 1 were
unsuccessful. An attempt to add the elusive difluoroketene, formed in situ from
bromodifluoroacetyl chloride and zinc in acetonitrile, produced a complex mixture, and
the highly electron-deficient bis(trifluoromethyl)ketene failed to react with 1 even after
10 h at 80 oC.29 Addition to 1 of tetrafluoroethylene would yield a particularly interesting
compound, perfluoro[2.2.2]propellane, but the two alkenes failed to react even after 4 h
at 120 oC. Irradiation at wavelengths down to 210 nm gave the same result.
We surmised that substitution of one or more fluorines of tetrafluoroethylene by
oxygen, the second most electronegative element, could create a promising addend
because of oxygen's -donor ability. While that ability would facilitate the addition to 1,
in the resulting propellane only oxygen's electron-withdrawing power would be expected
to influence the thermal stability. So ketene acetal 47 was chosen as a target for
synthesis. Treatment of trifluoroacetaldehyde with 2-chloroethanol afforded a hemiacetal
which was cyclized with potassium carbonate to give the known ketal 46 (Scheme 23).30
Scheme 23
Dehydrofluorination of 46 to obtain 47 proved challenging because a variety of bases and
conditions resulted in polymerization of the very sensitive alkene. Success was finally
achieved with butyllithium in ether at -78 oC followed by low temperature vacuum
17
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
transfer at 60 mTorr. The resulting ketene acetal 47 in ether solution (50-70% yield) was
quickly destroyed by moisture, and it polymerized in the presence of both acids and
bases. It reacted immediately and quantitatively with 1 to produce propellane ketal 48
(Scheme 24).2
Scheme 24
The X-ray crystal structures of propellanes 45 and 48 reveal that the length of the
central C-C bond is 1.558 and 1.573 Å, respectively, very similar to that of the C-C bonds
in cyclobutane (1.548 Å).31 Ketal 48 is even more stable thermally than 45. A solution
of 48 in CDCl3 suffered no decomposition after several months at rt. When a sample was
heated in a sealed tube at 100 oC for 10 h, no decomposition was detected; after 5 h at
140 oC, 45% of the propellane remained unchanged. Thermal stability data for the four
known [2.2.2] propellanes are summarized in Table 3.
Fluorine substituents are known to stabilize cyclobutane rings, and the phenomenon
has been attributed to relief of nonbonded repulsion within the rings.32 The distances
between opposite carbons in the three four-membered rings in 48 are all ~2.2 Å, far less
than the sum of the van der Waals radii (3.4 Å). Electron-withdrawing substituents
diminish the repulsion by decreasing electron density inside the rings. Thus, the ability
of fluorine to stabilize [2.2.2] propellanes may stem principally from this effect.
18
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Despite its thermal robustness, 48 is extremely susceptible to attack by
nucleophiles, electron-rich radicals, and electron donors, all of which cleave the central
bond.2 Interestingly, neither the LUMO nor the LUMO + 1, in which the central bond
lies along a node, is suited for nucleophilic attack on that bond (Figure 2). However, the
LUMO + 2, which lies 10.2 kcal/mol above the LUMO, has large lobes projecting from
the central carbons. The larger of the two controls the direction of nucleophilic attack,
consistent with the observed regiochemistry that places developing negative charge at the
more electron-deficient end of the central bond. For comparison, the corresponding
19
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
orbital (LUMO + 2) for the yet unknown [2.2.2] perfluoropropellane is also shown in
Figure 2.
Figure 2. LUMO+2 of 48 (left) and LUMO+2 of the unknown [2.2.2]
perfluoropropellane (right) at the B3LYP/cc-pVDZ level of theory. Isocontour = 0.006.
When 1 was generated from dibromo precursor 9 in the presence of
ethoxyacetylene, the major product was bicyclic alkene 50, accompanied by bromo
analogue 51 (Scheme 25). [2 + 2] Cycloaddition must have occurred to create the very
strained [2.2.2] propellene 49, which was reduced by zinc to produce 50 via proton and
perhaps hydrogen atom abstraction from the hexane/acetonitrile solvent. Opening of the
central bond of the propellene by bromide ion followed by proton abstraction explains the
formation of 51.
Scheme 25
20
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Isolated 1 was found to react readily with ethoxyacetylene at temperatures far
below 0 oC to yield propellene 49 (Scheme 26). Its 19F NMR spectrum, measured at -70
Scheme 26
oC, comprised two AA'XX' patterns. Warming of the reaction mixture to r.t. resulted in
polymerization of 49, perhaps as shown in Scheme 27, but the propellene could be
trapped in the cold as HI adduct 52 by addition of a slightly moist solution of
Scheme 27
tetrabutylammonium iodide in acetonitrile (Scheme 28). The regiochemistry of the
reaction was confirmed by an NOE measurement that revealed close proximity of the
bridgehead hydrogen to ethyl hydrogens. That formation of 49 is another orbital
21
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
Scheme 28
topology-forbidden process10 makes its occurrence at very low temperatures all the more
remarkable. We did find a closed shell intermediate that lies below the reactants: H298K
= -4.1 kcal/mol, G298K = 10.1 kcal/mol (G3MP2B3, Figure 3). For the corresponding
triplet diradical, H298K = 7.0 kcal/mol, G298K = 19.8 kcal/mol. (An open shell singlet
diradical at the triplet geometry is even higher in energy.)
Figure 3. Closed shell singlet intermediate in the reaction of ethoxyacetylene with alkene
1, and the corresponding triplet.
Heats and free energies of reaction with alkene 1 for a selection of addends are
compiled in Table 4. They are arranged in order of increasing exothermicity as computed
22
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
at the G3MP2 level of theory. All are strongly exothermic, including the five leading to
the [2.2.2] propellane skeleton that fall in the range -40 to -65 kcal/mol.
Conclusions
Though very stable thermally, perfluoroalkene 1 is highly reactive and boasts a rich
chemistry. Water adds spontaneously across the double bond, and our calculations
indicate that it does so via a pseudopericylic cycloaddition of the water dimer. Radical
chain addition to that alkene double bond occurs readily; e.g., with benzaldehyde
promoted by light. Ene reaction with alkenes and dienes takes place at r.t., and with 1,3-
cyclohexadiene ene reaction dominates over Diels-Alder addition. The fact that reaction
of 1 with butadiene results in Diels-Alder addition and not [2+2] cycloaddition
constitutes strong evidence that the reaction occurs concertedly. Diazo compounds
undergo 1,3-dipolar cycloaddition reactions with 1, yielding pyrazolines. Their
photolysis with elimination of nitrogen produces 1,3-diradicals that, instead of ring
23
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
closing to create extremely strained [2.2.1] propellanes, opt for ring opening to generate
carbenes. [2 + 2] Cycloadditons of 1 have been successful at or below r.t. with both
electron-donor alkenes and a ketene. All three of the [2.2.2] propellanes (including the
propellanone) derived from 1 are much more stable thermally than the sole [2.2.2]
propellane in the literature that has a hydrocarbon skeleton, and the stability increases
with the degree of electron withdrawal from the skeleton. This effect is probably
attributable primarily to a decrease in C-C nonbonded repulsion within the interior of the
molecules. Like alkene 1, the fluorinated propellanes display both of the commonly
antithetical qualities of robustness and high reactivity. Reaction of 1 with
ethoxyacetylene afforded the first [2.2.2] propellene, a fragile compound that formed
readily at temperatures far below 0 oC in an orbital topology-forbidden reaction.
Calculations carried out for many of the addition reactions of alkene 1 reveal them all to
be strongly exothermic, some extremely so.
Experimental Section
The 19F NMR spectra were recorded at 282.2 or 470.3 MHz on a 300 or 500 MHz
spectrometer. Chemical shifts are reported on the  scale (ppm from internal standard
1
trichlorofluoromethane, upfield negative). H NMR spectra were recorded at 300 or 500
MHz on the same spectrometers with tetramethylsilane as an internal standard. 13C NMR
spectra were recorded at 75.4 MHz on the 300 MHz spectrometer. Melting points were
measured in open capillary tubes and were not corrected. Ultrasonic irradiation was
performed in a Bronsicator Ultrasonic B-22-4 cleaning bath. Preparative gas
chromatography was performed with a thermal conductivity detector. The column was
24
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
10 ft. x 1/8 in., 10% OV 101 on Chromosorb-P acid washed, and the standard program
was: det. 200 oC, inj. 200 oC, column temperature as noted in text. All analytical GC
was done using a 25 m methyl silicone capillary column and flame ionization detector.
The standard program was: carrier pressure 25 p.s.i.; detector 200 ˚C; injector 150 ˚C;
temp1= 50 ˚C; temp2 = 250 ˚C; time1= 5 min. @ 50 ˚C then 15 ˚/min to 250 ˚C. Solvents
used in this work were reagent grade. Acetonitrile was distilled from calcium hydride,
and tetrahydrofuran from potassium/benzophenone. Zinc dust was purified according to
Shriner and Neumann.33
The initial geometry optimization calculations for all structures were performed
using Jaguar9 at the level of density functional theory34 with the B3LYP exchange-
correlation functional35,36 and the cc-pVDZ+ basis set.9b Subsequently, the structures
were reoptimized with the B3LYP, BP86,37,38 PW91,39,40,41 PBE,42,43 M06,44 and
SVWN45,46 functionals with the DZVP247 basis set for the benzene adduct formation from
1. In addition, calculations for that reaction were done at the G3MP2 level48 starting
from the B3LYP/DZVP2 geometries. Nearly all other calculations were performed only
at the DFT B3LYP//DZVP2 and MO6/DZVP2 levels and at the G3MP2 level. All but
the initial geometry optimizations were done with Gaussian 0949 by D.A.D. and M.V.
2,2,3,3,5,5,6,6-Octafluorobicyclo[2.2.0]hexane (10) and 2,2,3,5,5,6,6-
Heptafluorocyclohex-3-ene-1,1-diol (6). 1,4-Dibromoperfluorobicyclo[2.2.0]hexane
(70 mg, 0.18 mmol) was dissolved in 3 mL of acetonitrile in a 25 mL round-bottom flask.
After nitrogen had been bubbled in for 5 min, zinc dust (0.2 g, 3 mmol) and water (1
drop) wre added to the flask, which was then stoppered and rocked in an ultrasound bath
for 4 h. 19F NMR revealed two products, dihydro compound 10 and diol 6 the ratio 2 : 1.
25
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
1
For 10:  -111.3, -112.7 (AB q, J = 225 Hz, 8F). H NMR:  2.0, 2H). MS: m/z 207
+
(M+-F), 157 (base, C5F5H2 ). For 6, 19F NMR:  -105.7 (2F), -120.4 (2F), 121.8 (1F), -
+
133.3 (2F), 1H NMR:  5.9 (vinyl, 1H), 4.0 (2H). MS(CI): m/z 258 (MNH4 ), 220 (M+-
HF), 175 (base). IR (cm-1): 1627 (C=C), 3500 (br, OH).
Ethyl 2,2,3,3,5,5,6,6-octafluorobicyclo[2.2.0]hexane-1-carboxylate (13). N-
Benzylpyrrole adduct (350 mg, 0.92 mmol) was subjected to FVP to generate alkene 1, to
which was added ether (3 mL) and ethyl formate (130 mg, 1.8 mmol, 2 eq). The solution
was stirred 4 h at r.t. to generate adduct 13 cleanly by 19F NMR. Purification was
achieved by preparative GC at 150 oC to afford 13 as a colorless oil. 19F NMR:  -112.0,
1
-115.7 (AB q, J = 226 Hz, 4F); -114.1, -124.2 (AX, J = 216 Hz, 4F). H NMR:  3.83 (q,
J= 7.2 Hz, 2H), 3.69 (m, bridgehead, 1H), 1.31 (t, J = 7.2 Hz, 3H). MS: m/z 269 (M+ -
+
C2H5), 194, 175, 144, 113, 100 69, 29 (base, C2H5 ).
(2,2,3,3,5,5,6,6-Octafluorobicyclo[2.2.0]hexan-1-yl)(phenyl)methanone (15).
To a Pyrex tube (1 x 10 cm) was added alkene 1 (0.16 g, 0.71 mmol), benzaldehyde
(0.228 g, 2.15 mmol) and 3 mL of dry acetonitrile. The solution was irradiated with a
450 W medium-pressure mercury lamp at 15 oC under nitrogen for 1 h. After solvent
removal, the residue was purified by flash chromatography on neutral alumina
(pentane/CH2Cl2, 25:1) to afford 15 as colorless crystals (0.19 g, 0.58 mmol, 82%). Mp
97-98 oC. 19F NMR:  -111.4, -112.9 (AB q, J = 222 Hz, 4F); -112.0, 112.7 (AB q, J =
222 Hz, 4F). 1H NMR:  7.99 (d, J = 7.5 Hz, 2H), 7.74 (t, J = 7.5 Hz, 1H), 7.60 (t, J = 7.5
Hz, 2H), 4.34 (m, 1H). 13C NMR:  183.5, 135.8, 134.3, 129.5, 129.1, 115.4 (t, J = 295
Hz), 113.9 (t, J = 269 Hz), 63.9 (m), 49.4 (m). IR (film, cm-1): 1688, 1598, 1453, 1345,
26
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
1264, 1225, 1160, 1103, 1063, 995. Anal. Calc. for C13H6F8O: C, 47.29; H, 1.83; F,
46.03. Found: C, 47.02; H, 1.76; F, 46.30.
2,2,3,3,5,5,6,6-Octafluoro-1-(2-methylallyl)bicyclo[2.2.0]hexane (16). To a
flame-dried, heavy-walled glass tube (1 x 20 cm, 4 mm ID) containing zinc dust (0.50 g,
7.7 mmol) was added dibromide 9 (0.40 g, 1.1 mmol) and 4 mL of dry acetonitrile.
Isobutene (0.5 mL, ~6 mmol) was condensed into a graduated trap at -78 oC, and its
volume was measured after warming to 0 oC. It was vacuum transferred into the reaction
tube cooled to 77 K. The tube was sealed with a flame under vacuum, allowed to warm
to rt, and placed in a water-cooled ultrasonic bath for 24 h. It was cooled again to 77 K,
opened, and allowed to thaw. Water (20 mL) was added to the light brown solution, and
the mixture was extracted with 3 x 10 mL of pentane. The extract was dried over
MgSO4, then concentrated to 1 mL. Ene product 16 was present in 81% yield by 19F
NMR. Preparative GC at 150 oC gave 16 as a colorless liquid. 19F NMR:  -111.4, -
1
113.6 (AB q, J = 227 Hz, 4F); -111.7, -118.0 (AX, J = 220 Hz, 4F). H NMR:  4.96 (s,
vinyl. 1H), 4.81 (s, vinyl, 1H), 3.38 (bs, bridgehead, 1H), 2.60 (s, methylene, 2H), 1.81
(s, methyl, 3H). 13C NMR:  138.7, 116.2, 114.3, 113.1, 57.0 (both bridgeheads), 32.0,
+
23.4. MS: m/z 280 (M+), 167, 147, 127, 113 (base, C3HF4 ), 69, 41, 27. IR (CCl4, cm-1):
1653 (C=C), 1435, 1350. 1289, 1249, 1187, 1145, 1118, 1068, 973, 938, 901.
1-(Cyclohexa-2,4-dien-1-yl)-2,2,3,3,5,5,6,6-octafluorobicyclo[2.2.0]hexane
(17). Alkene 1 was obtained by FVP of the N-benzylpyrrole adduct (300 mg, 0.79
mmol). A solution of 1,4-cyclohexadiene (60 mg, 0.75 mmol) in ether (2 mL) was
added, and the reaction mixture was allowed to stand at r.t. for 10 h. By 19F NMR, the
yield of 17 based on 1 was quantitative. 19F NMR:  -108.2, -114.0 (AX, J = 236 Hz, 2F);
27
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
-111.5, -117.7 (AX, J = 217 Hz, 2F); -112.2 to -113.3 (m, 2 overlapping AB q's, 4F). 1H
NMR:  6.02 (m, vinyl, 1H); 5.98 (m, vinyl, 1H); 5.86 (m, vinyl, 1H); 5.51 (m, vinyl,
1H); 3.64 (m, bridgehead, 1H); 3.14 (t, J = 14 Hz, 1H); 2.31, 2.28 (AB q, J = 4.5 Hz,
+
+
2H). MS m/z 304 (M+), 253, 151, 113 (C3HF4 ), 79 (base, C6H7 ), 69, 40. Anal. Calcd.
For C12H8F8: C, 47.37; H, 2.63. Found: C, 47.25; H, 2.60.
1-(Cyclohexa-2,5-dien-1-yl)-2,2,3,3,5,5,6,6-octafluorobicyclo[2.2.0]hexane
(18) and 7,7,8,8,9,9,10,10-Octafluorotetracyclo[4.2.2.22,5.01,6]dodec-3-ene (19).
To a solution of alkene 1 (0.80 g, 3.6 mmol) in 20 mL of dry ether was added 1,3-
cyclohexadiene (2.68 mL, 28.7 mmol). The solution was allowed to stand under N2 at r.t.
for 4 h. Solvent was evaporated and the residue was dissolved in 15 mL of dry CH2Cl2.
The solution was cooled to -45 oC, Br2 was added dropwise with stirring in the dark until
its color persisted, and stirring was continued for 2 h. After removal of the solvent and
excess Br2, the volatile portion of the residue was vacuum transferred to afford adduct 19
(0.23 g, 21% yield). Mp 72-74 oC. 19F NMR:  19F NMR:  -107.9, -110.8 (subsplit AB
1
q, J = 228 Hz, 4F); -108.7, -113.4 (subsplit AB q, J = 222 Hz, 4F). H NMR:  6.31 (m,
vinyl, 2H); 3.13 (bs, bridgeheads, 2H); 2.17 (d, J = 10.5 Hz, 2H), 1.26 (d, J = 10.5 Hz,
2H). HRMS Calcd. for C12H8F8: 304.0498, found: 304.0495.
The residue after vacuum transfer containing tetrabromide 20 was dissolved in 20
mL of dry THF and zinc dust (8.0 g, 0.12 mol) was added. The mixture was flushed with
N2 for 2 min, a small crystal of iodine was added, and the slurry was stirred for 2 h. After
removal of zinc by filtration, solvent was evaporated and the residue was flash
chromatographed on silica gel to give adduct 18 as a colorless liquid (0.30 g, 28% yield).
19F NMR:  -109.8, -119.0 (AX, J = 220 Hz, 4F); -112.7, -133.5 (AB q, J = 227 Hz, 4F).
28
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800058
European Journal of Organic Chemistry
1H NMR:  5.88 (m, vinyl, 2H), 5.48 (m, vinyl, 2H), 3.57 (m, bridgehead, 1H), 3.40 (bs,
1H), 2.65 (m, 2H). HRMS Calcd. for C12H8F8: 304.0498, found: 304.0496.
cis-3,4-Dibromo-7,7,8,8,9,9,10,10-octafluorotetracyclo[4.2.2.22,5.01,6]dodecane
(21). Alkene 19 (0.12 g, 0.40 mmol) was dissolved in 15 mL of dry CH2Cl2, and Br2 (31
L, 0.6 mmol) in 1 mL of dry CH2Cl2 was added. The solution was stirred at r.t.
overnight exposed to ambient light. After removal of solvent, the residue was purified by
flash chromatography on silica gel (pentane/CH2Cl2, 10 : 1) to obtain dibromide 21 as
colorless crystals (0.93 g, 93%). Mp 120-122 oC. 19F NMR:  -105.1, -109.4 (AB q, J =
231 Hz, 4F); -105.2, -108.4 (AB q, J = 229 Hz, 4F). 1H NMR:  5.07 (s, 2H), 2.73 (s,
2H), 2.47 (d, J = 10.0 Hz, 2H), 2.05 (d, J = 10.0 Hz, 2H). IR (film, cm-1): 2965, 1486,
1346, 1223, 1155, 1059, 982, 902. Anal. Calcd. for C12H8F8Br2: C, 31.06; H, 1.74; Br.
34.44; F, 32.76. Found: C, 31.11; H, 1.73; Br, 34.56; F, 32.63.
7,7,8,8,9,9,10,10-Octafluorotetracyclo[4.2.2.22,5.01,6]dodecane (22). To alkene
19 (0.105 g, 0.345 mmol) in 20 mL of methanol was added 10% Pd/C (100mg). The
mixture was cooled to -78 oC and purged by evacuating and filling with N2 three times.
A balloon filled with H2 was attached to the apparatus, and the slurry was stirred at r.t. for
6 h. The catalyst was removed by filtration and washed with CH2Cl2. After evaporation
of the solvent, the residue was purified by chromatography on silica gel with pentane to
afford hydrofluorocarbon 22 as colorless crystals (87.5 mg, 83% yield). Mp 173-175 oC.
19F NMR:  -106.1, -111.0 (subsplit AB q, J = 230 Hz, 8F). 1H NMR:  2.31 (s, 2H), 2.12
(d, J = 9.0 Hz, 4H), 1.58 (d, J = 9.0 Hz, 4H). IR (film, cm-1): 2963, 1303, 1260, 1103,
982, 799. Anal. Calcd. for C12H10F8: C, 47.07; H, 3.29; F, 49.64. Found: C, 47.09; H,
29
This article is protected by copyright. All rights reserved.