Structural assignment of fluorocyclobutenes by19F NMR spectroscopy – Comparison of calculated19F NMR shielding constants with experimental19F NMR shifts

Article abstract of DOI:10.1002/ejoc.201800482

Although the optimized reduction of perfluorocyclo-butene with LiAlH₄ gave a quantitative yield of the target 3,3,4,4-tetrafluorocyclobut-1-ene, unoptimized reductions led to complex inseparable mixtures of fluorocyclobutenes. These mixtures showed highly complex¹⁹F NMR spectra, the assignment of which was quite tedious. Hence, we accomplished a series of single-reference computations of the¹⁹F NMR magnetic shieldings of the corresponding fluorine atoms. Surprisingly, various DFT approaches, including both traditional and advanced functionals, gave highly diverse results with poor correlations between the experimental and computed¹⁹F chemical shifts, and the individual fluorocyclobutenes could not be identified. In contrast, the domain-based local pair natural orbital coupled clusters (DLPNO-CCSD) method, developed recently as a part of the ORCA computational package, gave shielding values that enabled the assignment of all structures observed, albeit with some systematic errors. Slightly better magnetic shielding values were obtained by a simple Hartree–Fock (HF) method with a specially tailored IGLO-III basis set. The method developed was successfully employed for the assignment of the¹⁹F NMR shifts of unknown fluorocyclobutenes.

Full text of DOI:10.1002/ejoc.201800482

A Journal of
Accepted Article
Title: Structural assignment of fluorocyclobutenes by 19F NMR
spectroscopy: comparison of calculated 19F NMR shielding
constants with experimental 19F NMR shifts
Authors: Kateřina Kučnirová, Ondřej Šimůnek, Markéta Rybáčková,
and Jaroslav Kvicala
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.201800482
Link to VoR: http://dx.doi.org/10.1002/ejoc.201800482
Supported by

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
Structural assignment of fluorocyclobutenes by 19F NMR
spectroscopy: comparison of calculated 19F NMR shielding
constants with experimental 19F NMR shifts
Kateřina Kučnirová,[a] Ondřej Šimůnek,[a] Markéta Rybáčková,[a] and Jaroslav Kvíčala*[a]
Abstract: While optimized reduction of perfluorocyclobutene with
LiAlH4 gave quantitative yield of the target 3,3,4,4-tetrafluoro-
cyclobut-1-ene, unoptimized reductions led to complex unseparable
mixtures of fluorocyclobutenes. These mixtures showed highly
complex 19F NMR spectra, the assignment of which proved to be
quite tedious. We hence accomplished a series of single reference
computations of 19F NMR magnetic shielding of the corresponding
fluorine atoms. Suprisingly, various DFT approaches, including both
traditional and advanced functionals, gave highly diverse results with
poor correlation of experimental and computed 19F chemical shifts,
from which individual fluorocyclobutenes could not be identified.
Contrary to that, DLPNO-CCSD method, developed recently as a
part of ORCA computational package, gave the shielding values
which, although bearing some systematic error, enabled to assign all
structures observed. Even slightly better values of the magnetic
shielding were obtained by simple HF method using specially
tailored IGLO-III basis set. The method developed was successfully
employed for assigning the 19F NMR shifts to yet unknown
fluorocyclobutenes.
Alkene and enyne metathesis of fluoroalkenes with one or more
fluorines in the side chain proceeds successfully.[8] On the other
hand, 2-fluoroalkenes can[9] or need not[10] to be productive in
alkene metathesis and the substitution on the other side of the
double bond with an appropriate substituent, e.g. benzyl or
phenyl group can be highly productive.[11] Reactions with vinyl-
or vinylidenefluoride gave monofluoro- or difluoromethylene
substituted ruthenium complexes, catalytic activity of which was
low presumably due to their high stability.[12] Recently,
successful cross metathesis of polyfluoroalkenes with vinyl
ethers[13] has been published indicating critical role of the other
metathesis partner. This was in excellent agreement with our
computational analysis pointing at the key role of productive and
non-productive cycles in fluoroalkene metathesis.[14]
Surprisingly, no ROM of fluorocycloalkenes have yet been
reported with the exception of ROM polymerization of fluorinated
norbornenes[8a,15] (Scheme 1) and this, in connection with our
recent activity in the field of fluorinated alkene metathesis
precatalysts,[16] raised our interest in fluorocycloalkenes.
Introduction
Fluorine atom due to its high electronegativity, small atomic
volume and low polarizability occupies extraordinary place as a
modifier of biologically active compounds. It is estimated that
about 20% of recently selled pharmaceuticals contain one or
more fluorine atoms. A review from 2010 claims that three of top
ten blockbuster drugs in 2008 contained fluorine.[1] With
increasing popularity of biological drugs, only five small organic
compounds fell into twelve blockbuster drugs in 2015, but three
of them bear fluorine atoms.[2] The interest in new biologically
active fluorinated compounds inevitably results in the demand
for new fluorine-containing building blocks and, consequently, in
the development of new synthetic approaches leading to them.
Although fluoroalkenes are materials traditionally employed in
organic and polymer chemistry as building blocks and
monomers, they receive increasing attention in the last years as
environmentally friendly substitutes for CFC’s and HFC’s,[3] as
ligands or substrates in transition metal chemistry[4] or as
peptide isosteres.[5] For analysis of the properties of fluorinated
compounds, 19F NMR spectroscopy can be employed with
advantage. Thus, it was recently adopted for structural studies of
enamines and ylides formed by reaction of fluoroalkenes and
cycloalkenes with NHC ligands.[6] Fluoroalkenes were also
recently studied as ligands for Ni complexes and NMR
spectroscopy again chosen as the key analytical method.[7]
Scheme 1. ROM polymerization of fluorinated norbornenes
Among potential substrates for ROM of fluorinated cycloalkenes,
fluorocyclobutenes should lead to building blocks containing
two-carbon fluorinated units enveloped by double bonds. While
hexafluorocyclobut-1-ene (1) is easily available by thermal
dimerization of chlorotrifluorethene followed by dehalogenation
with zinc,[17] 3,3,4,4-tetrafluorocyclobut-1-ene (3a) can be
obtained by thermal dimerization of tetrafluoroethene with
chloroethene, which is potentially dangerous due to concurrent
exothermic polymerization, and by subsequent reductive
removal of chlorine.[18] Attempts to obtain fluorocyclobutene 3a
by reduction of fluorocyclobutene 1 with LiAlH4 led to mixtures
of pentafluorocyclobutenes (2), tetrafluorocyclobutenes (3) and
trifluorocyclobutenes (4) depending on reaction conditions, from
which the target fluorocyclobutene 3a had to be isolated by
preparative GC.[19] Similarly, reduction of fluorocyclobutene 1
[20]
with Cp*2ZrH2
or in situ formed titanocene hydrides[21] gave
mixtures of fluorocyclobutenes 2-4.
For the identification of such complex mixtures, 19F NMR
spectroscopy represents probably the best analytical method.
However, long distance fluorine-fluorine and fluorine-hydrogen
coupling
constants
between
individual
atoms
in
fluorocyclobutenes lead often to higher order spectra[22] and
hence identification of individual compouds in the mixtures can
be quite difficult.[21]
[a]
Department of Organic Chemistry
University of Chemistry and Technology, Prague
Technická 5, 166 28 Prague 6, Czech Republic
E-mail: kvicalaj@vscht.cz
The success of computational approaches in predicting NMR
shifts,[23] our previous positive experience in predicting
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
unexpected NMR shifts of fluorinated carbanions,[24] as well as
recent rapid progress in new computational methods including
electron correlation[25] initiated our computational study of 19F
NMR shieldings of a series of fluorocyclobutenes. This should
not only show which computational approach is the most
suitable for this type of calculations, but could also help in the
assignment of complex NMR spectra of the mixtures of
fluorocyclobutenes, obtained by various reductions of
hexafluorocyclobutenes.
shielding in the expected products. Its main aim was not only to
find out which method or functional gives best fitting agreement
with experiment, but preferably to depict the relative shifts in
individual molecules. We employed two computational packages,
Gaussian 09[26] and ORCA.[27]
Table 1. Optimization of LiAlH4 reduction of hexafluorocyclobut-1-ene (1).
NMR yield
Addition
time
(min)
Reaction
Solvent temp.
(°C)
Reag.
equiv.
Conversion
(%)
Entry
2a 3a 3b 4a 4b
Results and Discussion
1
2
3
4
0.5
0.65
1.1
5
5
Et2O
Et2O
Et2O
Et2O
-78
-40
-78
-78
91
22 58
<2
<2 94
<2 >98 <2 <2 <2
5
3
3
Reduction of hexafluorocyclobut-1-ene (1)
For ongoing experimental and theoretical study of the ROM of
100
100
100
6
<2 82 12
<2 <2
fluorocycloalkanes,
3,3,4,4-tetrafluorocyclobut-1-ene
(3a)
10
15
6
belongs to the most suitable substrates. Due to safety issues,
we decided to prefer the synthesis starting from
chlorotrifluoroethene[19] over that from tetrafluoroethene.[18] To
prevent overreduction, we, contrary to the original procedure,[19]
employed inverse addition of LiAlH4 solution in diethyl ether to
hexafluorocyclobut-1-ene (1) solution in diethyl ether. Using a
50% molar amount of the reduction reagent and 5 minutes
addition time at -78 °C, a complex mixture of products with
incomplete conversion was obtained (Fig. 2, Table 1 Entry 1).
1.1
Computations of 19NMR shielding using DFT approaches in
Gaussian 09
For the computations employing Gaussian 09, we chose five
functionals. Three of them belong to the most advanced recent
functionals, which include dispersion effects and hence are
expected to give the best molecular descriptions. While hybrid
ωB97X-D functional of Head-Gordon and coworkers contains
empirical long range dispersion correction,[28] MN15 hybrid
functional of Truhlar and coworkers includes non-empirical
medium ranged dispersion correction.[25] As a representative of
advanced pure functionals, we chose MN15-L of the same
family.[29] Finally, we included for comparison two older hybrid
functionals, traditional but still widely used B3LYP functional[30]
and also popular PBE0 functional with no empirical
parameters.[31]
Because 19F chemical shift values are obtained from chemical
shieldings by substraction of the chemical shielding of the
standard, trichlorofluoromethane, we started the computations
with the calculation of chemical shielding of the standard.
Optimized geometry was obtained at standard DFT level using
the PBE0 functional and a double zeta def2-SVP basis set.[32]
The most principle and up to date approach to computing
chemical shielding of flexible molecules requires careful
implementation of all significantly stable geometries or, even
better, snapshots from molecular dynamics calculations.[33]
Fortunately for us, both CCl3F and fluorocyclobutenes are rigid
molecules, for which static description based on minimum
energy geometries was sufficient.
Scheme 2. Products of LiAlH4 reduction of hexafluorocyclobut-1-ene (1)
Analogous experiment with somewhat higher amount of reagent,
the same reaction time and higher temperature (-40 °C) gave
only little target product 3a and mainly 1,4,4-trifluorocyclobut-1-
ene [4a] (Table 1 Entry 2]. However, the use of 1.1 equiv. of
LiAlH4, -78 °C and slower addition (10 min) gave nearly pure
target product 3a with only small amount of the isomeric side-
product 3b (Table 1 Entry 3). Finally, using the solution of LiAlH4
in diethyl ether/THF mixture and even longer addition time (15
min) gave the target product 3a without any by-products
detectable by 19F NMR spectroscopy (Table 1 Entry 4).
Under the conditions of Entry 1, a complex mixture of products
was formed. Although all products have been individually
characterized, some of them contain stereogenic centres with
high anisochronicity of CF2 groups and other display higher
order spectra. This made complete analysis of this mixture
rather difficult. With the development of new DFT functionals
and novel approaches to the description of electron correlation,
we decided to start a short computational study of 19F NMR
For the computations of chemical shielding, a triple zeta def2-
TZVP basis set of the same family[32] was used. Surprisingly, the
use of the above described selection of functionals led to highly
diverse values of shielding (Table 2 Entry 1) with three hybrid
functionals (B3LYP, PBE0 and MN15) giving closest values of
shielding between 173 and 179. On the other hand, inclusion of
empirical dispersion in the ωB97X-D functional gave significantly
higher values of the shielding. Omitting the exchange integral in
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
the pure MN15-L functional resulted in even higher value of the
shielding. The most probable reason causing the large
differences in the shielding is the choice of the gauge origin,[23]
emphasizing the neccessity of computing the shielding of the
standard at the same level as the shielding of the studied
substances. Only minor changes in the shielding values were
observed when diffuse functions[34] were included into the basis
set (Table 2 Entry 2). Similarly, only small shielding differences
were found when the role of solvent was simulated using the
SCRF approach (Table 2 Entry 3), or when both modifications
were employed simultaneously (Table 2 Entry 4). For the solvent
simulation, we employed integrated electron field-polarization
continuum model approach (IEF-PCM), which is the default
approach in Gaussian 09. The differences in the chemical
shielding did not exceed 2 ppm. The IEF-PCM method has been
reported to describe well optical rotation of chiral compounds in
polar solvents as acetone or methanol, while the agreement with
experiments in less polar solvents as benzene or chloroform
was significantly worse. The main reason can be that the IEF-
PCM method desciption of solvent-solute interaction is
exclusively electrostatic.[35] Because the difference between the
least and most sophisticated approach (Table 2) was negligible,
def2-TZVP basis set was used further on.
Table 3. DFT computed values of magnetic shielding and chemical shift for
3,3,4,4-tetrafluorocyclobut-1-ene (3a).
Functional
Experiment PBE0 B3LYP MN15 MN15-L ωB97X-D
Shielding
302.4 299.0 307.8
-123.8 -125.4 -130.1 -110.0
-14.4 -16.0 -20.7 -0.6
326.9
306.2
-117.5
-8.1
Chemical shift
Shift difference
-109.4
For the analysis of fluorocyclobutenes, the accuracy of absolute
values of chemical shifts is not essential and some offset error
can be tolerated, providing it is systematic over the range of all
studied compounds. Especially desirable is then to provide good
relative values of chemical shift in one molecule. With that in
mind, we chose hexafluorocyclobut-1-ene (1) as the second
studied compound. Due to symmetry of the molecule, only two
signals are observed in experimental 19F NMR spectra, the
signal of CF2 group resonating at -118.6 ppm and the signal of
=CF group observed at -128.0 ppm. The computations of the
CF2 group gave analogous results and errors pattern as those of
the CF2 groups of 3,3,4,4-tetrafluorocyclobut-1-ene (3a), i.e. best
fit using pure MN15-L functional, followed by ωB97X-D with
largest deviation for MN15 hybrid functional (Table 4).
Table 2. Chemical shielding of CCl3F computed by various functionals and
basis sets.
Functional→
Basis set↓
Entry
PBE0 B3LYP MN15 MN15-L ωB97X-D
Table 4. DFT computed values of magnetic shielding and chemical shift for
hexafluorocyclobut-1-ene (1).
1
2
3
4
def2-TZVP
def2-TZVPD
178.6 173.6
177.2 171.8
180.5 175.7
177.6
177.6
179.7
179.6
216.9
215.6
218.1
216.8
188.7
188.1
190.5
189.8
Group
Functional
Shielding
Exp. PBE0 B3LYP MN15 MN15-L ωB97X-D
PCM-def2-TZVP
310.8 310.0 315.3 334.6
-132.2 -136.4 -137.6 -117.7
315.5
-126.8
1.2
PCM-def2-TZVPD 179.0 173.9
=CF Chemical shift -128.0
Shift difference
-4.2
-8.4
-9.6
10.3
Shielding
311.3 308.1 316.7 336.3
314.1
-125.4
-6.8
The
first
fluorocycloalkene
studied
was
3,3,4,4-
tetrafluorocyclobutene (3a). As all fluorine atoms are chemically
equivalent, only one signal (although of a higher order) was
observed in the 19F NMR spectrum at -109.4 ppm. While the
differences in computed shielding of individual fluorine atoms did
not exceed 0.1 ppm for each functional, the values across the
functionals varied quite significantly (Table 3). Thus,
„traditional“ functionals gave lowest values of shielding (299.0
for B3LYP and 302.4 for PBE0), while the use of
„modern“ functionals including dispersion gave reasonably
higher values of shielding (306.2 for ωB97X-D and 307.8 for
MN15). Strikingly high shielding (326.9) was obtained with pure
MN15-L functionals. Similar trends could be observed in the
chemical shift values. Quite surprisingly, the only method which
gave nearly exact values of chemical shift was MN15-L
functional. For all hybrid functionals, significant upfield error in
the shift was observed ranging from 8.1 ppm for ωB97X-D to
20.7 for MN15 (Table 3).
CF2 Chemical shift -118.6 -132.7 -134.5 -139.0 -119.4
Shift difference
-14.1 -15.9 -20.4 -0.8
However, the shielding and the chemical shift of the =CF groups
displayed completely different picture. The upfield shift error
typical for the CF2 groups in 1 and 3a was reduced quite
systematically for all functionals by about 10 ppm, which led to
the best fit for the ωB97X-D functional, lower errors for all other
hybrid functionals and, on the other hand, to a large downfield
error for pure MN15-L functional. What is worse, while in the
experimental spectrum the =CF groups resonate at significantly
upper field by nearly 10 ppm, computational description shows
opposite or nearly equal arrangement of the shifts. The only
exception is the B3LYP functional, but still the shift difference is
much lower. This was the first hint that the DFT methods are
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
probably not optimal for the evaluation of the NMR shifts of
fluorocyclobutenes.
Again, in analogy to the previous two calculations, hybrid
functionals gave acceptable shift for the =CF group, but
underestimated the shielding of both CF2 groups. Also in
agreement with previous calculations, the shift of both =CF
groups was described well by the pure M06L functional, which
however failed to describe correctly the CF2 group shifts.Linear
regression showed for all applied DFT very poor correlation with
correlation coefficient R2 not excceding 0.7. Among them,
relatively best results gave the traditional B3LYP functional (R2 =
0.67), while worst behave quite surprisingly one of the most
As the second compound bearing both CF2 and =CF group, we
chose 1,4,4-trifluorocyclobut-1-ene (4a), which is formed as a
major product by the reduction of hexafluorocyclobut-1-ene (1)
with LiAlH4 at -40 °C (Table 1). While the experimental shifts of
both groups differ by only 6 ppm, all DFT method tested gave
differences in shielding higher than 20 ppm (Table 5).
advanced dispersion including hybrid functionals, MN15 (R2
0.51) (Fig. 1).
=
Table 5. DFT computed values of magnetic shielding and chemical shift for
1,4,4-trifluorocyclobut-1-ene (4a).
Group
Functional
Shielding
Exp. PBE0 B3LYP MN15 MN15-L ωB97X-D
280.8 280.7 286.6 305.7
287.0
-98.3
7.6
=CF Chemical shift -105.9 -102.2 -107.1 -108.9 -88.8
Shift difference
Shielding
3.7
-1.2
-3.0
17.1
306.0 302.7 312.7 330.3
309.7
-121.0
-8.8
CF2 Chemical shift -112.2 -127.4 -129.1 -135.0 -113.3
Shift difference -15.2 -16.9 -22.8 -1.1
Figure 1. Correlation of experimental and DFT calculated shifts of
cycloalkenes 1, 2a, 3a and 4a.
Similarly to the computations of hexafluorocyclobut-1-ene (1), all
hybrid functionals gave acceptable computed shifts for the =CF
group, but failed to reproduce experimental shift for the CF2
group. Interestingly, for the pure functional, the situation was
exactly the opposite.
Computations of 19NMR shielding using DLPNO-CCSD and HF
methods in ORCA
Disappointed by the DFT method, we turned our attention to
other computationally feasible methods including electron
correlation. Among them, single reference multiconfiguration
DLPNO-CCSD (domain-based local pair natural orbital coupled
clusters) method[36], invented as a part of ORCA computational
program,[27] appeared to be the most attractive due to high
computational efficiency. Indeed, computations of the chemical
shifts of the starting fluorocycloalkene set 1, 2a, 3, 4a, using the
basis set, resulted in much better agreement between calculated
and experimental shift. Although the average offset error
reached nearly 20 ppm, the R2 coefficient of the correlation
between the calculated and experimental values was
significantly higher (0.8861, Fig. 2). For comparison, we also
compared the coupled cluster calculations with a simple Hartree-
Fock (HF) approach. To our surprise, the HF approach with
limited electron correlation gave even better agreement between
computed and experimental values giving R2 = 0.9456 (Fig. 4).
Both approaches overestimated the magnetic shielding by 11-27
ppm for the DLPNO method and by 12-24 ppm for the HF
method.
Our final studied compound was 1,3,3,4,4-pentafluorocyclobut-
1-ene (2a, Table 6), formed as a side product in the reduction of
cyclobutene 1 with lower amount of LiAlH4 (Table 1 Entry 1).
Table 6. DFT computed values of magnetic shielding and chemical shift for
1,3,3,4,4-pentafluorocyclobut-1-ene (2a).
Group
Functional
Shielding
Exp. PBE0 B3LYP MN15 MN15-L ωB97X-D
280.5 280.1 283.7 306.2
286.7
-97.9
8.0
=CF Chemical shift -105.3 -101.9 -106.5 -106.1 -89.3
Shift difference
Shielding
4.0
-0.6
-0.2
16.6
306.5 303.1 312.7 330.8
310.3
-121.6
-9.4
3-CF2 Chemical shift -112.2 -127.9 -129.5 -135.1 -113.9
Shift difference
Shielding
-15.7
-17.3
-22.9
-1.7
310.8 307.5 318.0 333.7
314.8
-126.1
-13.9
In the attempt to further improve the results of DLPNO and HF
computations, we first decided to improve the geometry of
starting fluorocyclobutenes by using more advanced M062X
functional[37] and more flexible def2-TZVPPD[34] basis set. For
the computations of the chemical shifts, we furthermore
employed an IGLO-III basis set[38] with improved description of
4-CF2 Chemical shift -117.2 -132.2 -133.9 -140.4 -116.7
Shift difference
-20.0 -21.7 -28.2 -4.5
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
the 1s orbital, specially tailored for NMR computations. The
improved approach resulted in the improved correlation both for
the DPLNO-CCSD (R2 = 0.9416) and the HF (R2 = 0.9685)
method (Fig. 2, Table 7). Moreover, the overestimation of the
chemical shifts sank to 6-14 ppm for the DLPNO method and to
7-13 ppm for the HF method.
Table 7. Computed values of magnetic shieldings and chemical shifts for
fluorocyclobutenes 1, 2a, 3a and 4a using DLPNO and HF methods.
DLPNO
TZVP
HF
TZVP
DLPNO
IGLO-III IGLO-III
HF
Group
Functional
Exp.
3,3,4,4-Tetrafluorocyclobut-1-ene (3a)
Shielding
339.4
-92.7
16.7
338.4
-92.7
16.7
341.0
-102.9
6.5
340.6
-102.3
7.1
CF2
Chemical shift -118.6
Shift difference
Hexafluorocyclobut-1-ene (1)
Shielding
280.5
-101.9
4.0
280.1
-106.5
-0.6
283.7
-106.1
-0.2
306.2
-89.3
16.6
=CF
CF2
Chemical shift -105.3
Shift difference
Shielding
Figure 2. Correlation of experimental and DLPNO or HF calculated shifts of
cycloalkenes 1, 2a, 3a and 4a.
306.5
-127.9
-15.7
303.1
-129.5
-17.3
312.7
-135.1
-22.9
330.8
-113.9
-1.7
Chemical shift -112.2
Shift difference
We employed the developed approaches for the computation of
the 19F NMR shifts of three other fluorocyclobutenes, namely of
1,3,4,4-tetrafluorocyclobut-1-ene (3b)
and 3,3,4-trifluoro-
1,4,4-Trifluorocyclobut-1-ene (4a)
cyclobut-1-ene (4b), which we observed as the side-products in
the reduction of hexafluorocyclobut-1-ene (1), as well as of
1,2,3,3,4-pentafluorocyclobut-1-ene (2b), reported by Lentz et
al.[21] Both DLPNO and HF calculations of these cycloalkenes
fitted well into the previous set of calculations, improving the
correlation coefficients R2 to 0.9907 for the DLPNO and 0.9917
for the HF approach (Fig. 2, Table 8).
Predictive power of our approach can be documented by the
computation of 19F chemical shielding of 1,3,4,4-tetrafluorocyclo-
but-1-ene (2b), reported also by Lentz et al.[21] The structure of
alkene 2b was assigned mainly based on coupling constants,
but these can be quite similar for isomeric 1,3,3,4-
tetrafluorocyclobut-1-ene (2c). Using correlation equation from
Fig. 3, the maximum difference between corrected computed
and experimental shift in alkene 2b did not exceed 2.5 ppm,
while the maximum differences of the experimental shifts from
the corrected computed shift of hypothetical fluoroalkene 2c
exceeded 10 ppm, confirming thus the assignment from Ref.[21]
Shielding
334.8
-88.0
17.9
331.7
-86.1
19.8
336.5
-98.4
7.5
333.6
-95.2
10.7
=CF
CF2
Chemical shift -105.9
Shift difference
Shielding
340.2
-93.5
18.7
340.8
-95.1
17.1
343.1
-104.9
7.3
339.8
-101.5
10.7
Chemical shift -112.2
Shift difference
1,3,3,4,4-Pentafluorocyclobut-1-ene (2a)
Shielding
325.7
-79.0
26.9
328.0
-82.3
23.6
329.3
-91.2
14.1
331.0
-92.7
12.6
=CF
Chemical shift -105.3
Shift difference
Shielding
338.7
-91.9
20.3
338.7
-93.1
19.1
340.3
-102.2
10.0
340.6
-102.2
10.0
3-CF2
4-CF2
Chemical shift -112.2
Shift difference
Shielding
347.3
-100.6
11.6
345.2
-99.5
12.7
349.2
-111.0
6.0
348.4
-110.0
7.0
Chemical shift -117.2
Shift difference
Figure 3. Correlation of experimental and DLPNO or HF calculated shifts of all
considered cycloalkenes 1 - 4.
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
cis-CF
trans-CF
CHF
Chemical shift
Shift difference
Shielding
-119.6
-113.1
-187.1
-113.3
6.3
-111.5
8.1
Table 8. Computed values of magnetic shieldings and chemical shifts for
fluorocyclobutenes 2b, 3b and 4b using DLPNO and HF methods.
341.9
-103.8
9.3
341.3
-102.9
10.2
DLPNO
TZVP
HF
TZVP
Group
Functional
Exp.
Chemical shift
Shift difference
1,3,4,4-tetrafluorocyclobutene (3b)
Shielding
414.4
-176.3
10.8
415.8
-177.4
9.7
Shielding
333.4
-95.2
8.2
332.7
-94.3
9.1
Chemical shift
Shift difference
=CF
cis-CF
trans-CF
CHF
Chemical shift
Shift difference
Shielding
-103.4
-118.7
-109.9
-182.0
351.0
-112.8
5.9
351.2
-112.8
5.9
Chemical shift
Shift difference
Conclusions
Shielding
Chemical shift
Shift difference
Shielding
340.8
-102.7
7.2
342.4
-104.0
5.9
We developed short and efficient synthesis of 3,3,4,4-
tetrafluorocyclobut-1-ene (3a) as a potential substrate for ring-
opening metathesis. The key reduction of intermediary
hexafluorocyclobut-1-ene (1) is highly sensitive to reaction
conditions and their inaccurate choice can result in the formation
of a mixture of fluorocyclobutenes 1-4. Quite suprisingly, both
classical and newly introduced dispersion including DFT
functionals failed to predict correctly 19F chemical shieldings of
fluorocyclobutenes 1-4. On the other hand, both coupled cluster
DLPNO method and simple HF approach gave the calculated
values of 19F chemical shieldings, which fitted well with the
experimental values. Both methods thus can be used as a tool
for the assignment of yet unknown fluorocyclobutenes.
408.9
-170.7
11.3
409.1
-170.8
11.2
Chemical shift
Shift difference
3,4,4-tetrafluorocyclobutene (4b)
Shielding
346.5
-108.3
5.1
345.1
-106.7
6.7
cis-CF
trans-CF
CHF
Chemical shift
Shift difference
Shielding
-113.4
-98.3
329.3
-91.2
7.1
328.8
-90.5
7.8
Experimental Section
Chemical shift
Shift difference
Shielding
General description of methods and materials. Temperature data
were uncorrected. NMR spectra were recorded with
a Varian
MercuryPlus spectrometer, 1H NMR spectra at 299.97 MHz and 19F
NMR spectra at 282.23, or with a Agilent 400-MR DDR2, 1H NMR
spectra at 399.94 MHz and 19F NMR spectra at 376.29 MHz, using
residual deuterated solvent signals as the internal standards for the 1H
NMR spectra or CCl3F as the internal standard for the 19F NMR spectra.
Chemical shifts are given in parts per million, coupling constants in hertz.
All reactions were performed in a dry inert atmosphere (Ar) in an oven-
dried flasks. Chlorotrifluoroethene was kindly gifted by Spolek pro
chemickou a hutní výrobu, Ústi nad Labem. Diethyl ether and THF were
distilled from the solution of diphenyl ketyl radical, methanol was pre-
dried with sodium and finally distilled from CaH2.
405.6
-167.5
8.7
406.5
-168.1
8.1
Chemical shift
Shift difference
-176.2
1,2,3,3,4-pentafluorocyclobutene (2b)
Shielding
Chemical shift
Shift difference
Shielding
350.7
351.0
-112.7
9.8
1-CF=
2-CF=
-122.5
-127.4
-112.6
9.9
Preparation of 1,2-dichloro-1,2,3,3,4,4-hexafluorocyclobutane. The
synthesis of the title cyclobutane was performed according to Ref.[17]
Thus, a 200 mL steel autoclave was cooled to -78 °C, charged with
chlorotrifluoroethene (120 g, 1.03 mol), sealed and heated for 5 days
while slowly raising the temperature from 125 °C to 250 °C. The heating
was continued for one more day, during which the pressure rose up to 6
MPa and then slowly sank to 2 MPa. After cooling, the liquid product was
363.0
-124.9
2.5
363.7
-125.3
2.1
Chemical shift
Shift difference
Shielding
351.4
349.8
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
fractionally distilled yielding the target fluorocyclobutane as a mixture of
cis- and trans-isomers in a 55:45 ratio (107 g, 91%, b.p. 54-59 °C. 19F
NMR of the product corresponded to the literature data.[39] cis-1,2-
This work was supported by specific university research (MSMT
No. 21-SVV/2018). Computational resources were supplied by
the Ministry of Education, Youth and Sports of the Czech
Republic under the Projects CESNET (Project No. LM2015042)
and CERIT-Scientific Cloud (Project No. LM2015085) provided
within the program Projects of Large Research, Development
and Innovations Infrastructures.
Dichloro-1,2,3,3,4,4-hexafluorocyclobutane,
19F NMR
(282.23 MHz,
2
2
CDCl3): δ -117.9 (dm, JF-F = 218 Hz, 2F, FCF-FCF); -129.9 (dm, JF-F
218 Hz, 2F, FCF-FCF); -137.2 (m, 2F, CFCl-CFCl) ppm. trans-1,2-
Dichloro-1,2,3,3,4,4-hexafluorocyclobutane, (282.23 MHz,
19F NMR
=
2
2
CDCl3): δ -121.8 (dm, JF-F = 213 Hz, 2F, FCF-FCF); -125.7 (dm, JF-F
213 Hz, 2F, FCF-FCF); -127.7 (m, 2F, CFCl-CFCl) ppm.
=
Keywords: fluorocyclobutene • 19F NMR spectroscopy •
computational chemistry • DFT • DLPNO-CCSD
Preparation of hexafluorocyclobut-1-ene (1). The original procedure[40]
was slightly modified. Zinc powder (56.0 g, 0.86 mol), activated by
grinding with few drops of acetic acid and acetic anhydride, was placed
into a 3-necked flask equipped with addition funnel and Vigreux column
topped by low-temperature fraction distillation head, followed by anh.
methanol (50 mL). 1,2-Dichloro-1,2,3,3,4,4-hexafluorocyclobutane was
slowly added to the mixture at a rate enabling gentle reflux of the mixture.
After 6 h, pure target product 1 was collected in the distillate (21.0 g,
76.5%, b.p. 0-2 °C). 19F NMR (282.23 MHz, CDCl3): δ -118.6 (m, 4F,
CF2); -128.0 (m, 2F, CF) ppm.
[1]
[2]
D. O’Hagan, J. Fluorine Chem. 2010, 131, 1071-1081.
http://www.pharmacompass.com/radio-compass-blog/top-drugs-by-
sales-revenue-in-2015-who-sold-the-biggest-blockbuster-drugs.
Retrieved 1 April 2017.
[3]
[4]
C. Hogue, Chem. Eng. News 2011, 31-32.
a) M. Ohashi, S. Ogoshi, Catalytic Transformations of Fluorinated
Olefins in Organometallic Fluorine Chemistry (Eds.: T. Braun, R. P.
Hughes), Topics in Organometallic Chemistry 52, Springer, Dordrecht,
2015, pp. 197-216; b) D.L. Orsi, R. A. Altman, Chem. Commun. 2017,
53, 7168-7181.
Preparation of 3,3,4,4-tetrafluorocyclobut-1-ene (3a). General
procedure: LiAlH4 was dissolved at -78 °C in diethyl ether (20 mL). The
solution was slowly cannulated to a cooled flask, charged with diethyl
ether or 1:1 diethyl ether/THF mixture (15 mL) and fluorocyclobutene 1
(3.1. g, 19 mmol). The reaction was stirred at low temperature for 15 min
and then quenched with aq. NH4Cl (10 mL). Aqueous layer was
separated and extracted with diethyl ether (3 × 10 ml). Combined organic
layers were washed with brine (10 mL) and anh. MgSO4. Diethyl ether
was removed by fractional distillation using a rectification column filled
with steel spirals, leaving the target product 3a as a distillation residue.
Run 1 (Table 1): Reaction with 0.5 eq. of LiAlH4 (360 mg, 9.5 mmol) in
diethyl ether, using addition time 5 min at -78 °C, gave a 91% conversion
[5]
[6]
H. Sakaguchi, Y. Uetake, M. Ohashi, T. Niwa, S. Ogoshi, T. Hosoya, J.
Am. Chem. Soc. 2017, 139, 12855−12862.
a) M. C. Leclerc, Serge I. Gorelsky, B. M. Gabidullin, I. Korobkov, R. T.
Baker, Chem. Eur. J. 2016, 22, 8063 – 8067; b) A. J. Arduengo III, J. C.
Calabrese, H. V. Rasika Dias, F. Davidson, J. R. Goerlich, A. Jockisch,
M. Kline, W. J. Marshall, J. W. Runyon, Phosphorus Sulfur 2016, 191,
527-534.
[7]
[8]
W. Xu, H. Sun, Z. Xiong, X. Li, Organometallics 2013, 32, 7122−7132.
a) S. Fustero, A. Simꢀn-Fuentes, P. Barrio, G. Haufe, Chem. Rev. 2015,
115, 871−930; b) S. Imhof, S. Randl, S. Blechert, Chem. Commun.
2001, 1692–1693; c) A. K. Chatterjee, J. P. Morgan, M. Scholl, R. H.
Grubbs, J. Am. Chem. Soc. 2000, 122, 3783-3784.
of starting cyclobutene 1, yielding
a
22:58:5:3:3 mixture of
fluorocycloalkenes 2a, 3a, 3b, 4a and 4b. 1,3,3,4,4-Pentafluorocyclobut-
1-ene (2a), 19F NMR (282.23 MHz, CDCl3):[21] δ -105.3 (m, 1F, CF=CH); -
111.2 (m, 2F, CF2-CF=CH); -117.0 (m, 2F, CF2-CH=CF) ppm. 3,3,4,4-
Tetrafluorocyclobut-1-ene (3a), 19F NMR (282.23 MHz, CDCl3):[21]
δ -109.4 (m, 4F, CF2) ppm. 1,3,4,4-Tetrafluorocyclobut-1-ene (3b),
19F NMR (282.23 MHz, CDCl3):[21] δ -103.4 (m, 1F, CF=CH); -109.9 (dm,
2JF-F = 205 Hz, 1F, FCF-CFH); -118.7 (dm, 2JF-F = 205 Hz, 1F, FCF-CFH);
[9]
a) S. S. Salim, R. K. Bellingham, V. Satcharoen, R. C. D. Brown, Org.
Lett. 2003, 5, 3403-3406; b) V. De Matteis, F. L. van Delft, J. Tiebes, F.
P. J. T. Rutjes, Eur. J. Org. Chem. 2006, 1166–1176.
[10] M. Marhold, A. Buer, H. Hiemstra, J. H. van Maarseveenb, G. Haufe,
Tetrahedron Lett. 2004, 45, 57-60.
[11] D. Guérin, A.-C. Gaumont, I.Dez, M. Mauduit, S. Couve-Bonnaire, X.
Pannecoucke, ACS Catalysis 2014, 4, 2374-2378.
2
-182.0 (dm, JH-F = 58 Hz, 1F, CFH) ppm. 1,4,4-Trifluorocyclobut-1-ene
(4a), 19F NMR (282.23 MHz, CDCl3):[21] δ -105.9 (m, 1F, CF=CH); -112.2
(m, 2F, CF2) ppm. 3,3,4-Trifluorocyclobut-1-ene (4b), 19F NMR
(282.23 MHz, CDCl3): δ -98.3 (dm, 2JF-F = 205 Hz, 1F, FCF-CFH); -113.4
[12] a) M. L. Macnaughtan, M. J. A. Johnson, J. W. Kampf, Organometallics
26, 2007, 780-782; b) T. M. Trnka, M. W. Day, R. H. Grubbs, Angew.
Chem. Int. Ed. 2001, 40, 3441-3444.
2
2
(dm, JF-F = 205 Hz, 1F, FCF-CFH); -176.2 (dm, JH-F = 62 Hz, 1F, CFH)
ppm. Run 2 (Table 1): Reaction with 0.65 eq. of LiAlH4 (470 mg, 12.4
mmol) in diethyl ether, using addition time 5 min at -40 °C, gave a 100%
conversion of starting cyclobutene 1, yielding a 6:82:12 mixture of
fluorocycloalkenes 3a, 4a and 4b. Run 3 (Table 1): Reaction with 1.1 eq.
of LiAlH4 (790 mg, 20.9 mmol) in diethyl ether, using addition time 10 min
at -78 °C, gave a 100% conversion of starting cyclobutene 1, yielding a
94:6 mixture of fluorocycloalkenes 3a and 3b. Run 4 (Table 1): Reaction
with 1.1 eq. of LiAlH4 (790 mg, 20.9 mmol) in a mixture of diethyl ether
and THF, using addition time 15 min at -78 °C, gave a 100% conversion
of starting cyclobutene 1, yielding pure fluorocycloalkene 3a.
[13] Y. Takahira, Y. Morizawa, J. Am. Chem. Soc. 2015, 137, 7031−7034.
[14] M. Rybáčková, J. Hošek, O. Šimůnek, V. Kolaříková, J. Kvíčala,
Beilstein J. Org. Chem. 2015, 11, 2150-2157.
[15] W. J. Feast, M. Gimeno, E. Khosravi, Polymer 2003, 44, 6111-6121..
[16] a) J. Kvíčala, M. Schindler, V. Kelbichová, M. Babuněk, M. Rybáčková,
M. Kvíčalová, J. Cvačka, A. Březinová, J. Fluorine Chem. 2013, 153,
12-25; b) M. Babuněk, O. Šimůnek, J. Hošek, M. Rybáčková, J. Cvačka,
A. Březinová, J. Kvíčala, J. Fluorine Chem. 2014, 161, 66-75; c)
J.Hošek, M. Rybáčková, J. Čejka, J. Cvačka, J. Kvíčala,
Organometallics 2015, 34, 3327-3334; d) P. Lipovská, L. Rathouská, O.
Šimůnek, J. Hošek, V. Kolaříková, M. Rybáčková, J. Cvačka, M.
Svoboda, J. Kvíčala, J. Fluorine Chem. 2016, 191, 14-22; e) J. Hošek,
O. Šimůnek, V. Kolaříková, P. Lipovská, K. Kučnirová, A. Edr, N.
Štěpánková, M. Rybáčková, J. Cvačka, J. Kvíčala, ACS Sustainable
Chem. Eng. 2018, 6, 7026−7034.
Computational details. DFT calculations were performed using
Gaussian 09W program suite[26] or the ORCA computational program.[27]
Vibrational frequencies were calculated for all species to characterize
them as minima.
[17] A. T. Blomquist, E. A. LaLancette, J. Am. Chem. Soc. 1961, 83, 1387-
1391.
ACKNOWLEDGMENTS
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
[18] a) D.D. Coffman, P.L. Barrick, R.D. Cramer, M.S. Raasch, J. Am. Chem.
Soc. 1949, 71, 490–496; b) J. Mizukado, Y. Matsukawa, H.-D. Quan, M.
Tamura, A. Sekiya, J. Fluorine Chem. 2006, 127, 79–84.
[28] J.-D. Chai, M. Head-Gordon, Phys. Chem. Chem. Phys. 2008, 10,
6615–6620.
[29] H. S. Yu, X. He, D. G. Truhlar, J. Chem. Theory Comput. 2016, 12,
1280-1293.
[19] G. Fuller, J. C. Tatlow, J. Chem. Soc. 1961, 3198-3203.
[20] B. M. Kraft, E. Clot, O. Eisenstein, W. W. Brennessel, W. D. Jones, J.
Fluorine Chem. 2010, 131, 1122–1132.
[30] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648-5652; b) P. J. Stephens,
F. J. Devlin, C. F. Chablowski, M. J. Frisch, J. Phys. Chem 1994, 98,
11623-11627.
[21] M. F. Kuehnel, P. Holstein, M. Kliche, J. Krüger, S. Matthies, D. Nitsch,
J. Schutt, M. Sparenberg, D. Lentz, Chem. Eur. J. 2012, 18, 10701-
10714.
[31] C. Adamo, V. Barone, J. Chem. Phys. 1999, 110, 6158-6170.
[32] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297-3305.
[33] a) S. C. Silva, S. M. M. Rodrigues, V. Nardini, A. L. L. Vaz, V. Palaretti,
G. V. José da Silva, R. Vessecchi, G. C. Clososki, J. Mol. Struct. 2018,
1163, 280-286; b) D. P. Demarque, D. R. Pinho, N. P. Lopes, C. Merten,
Chirality 2018, 30, 432–438; c) A. Bröhl, B. Albrecht, Y. Zhang, E.
Maginn, R. Giernoth, J. Phys. Chem. B 2017, 121, 2062−2072; d) P. A.
G. Soares, I. N. L. Queiroz, V. H. Pomin, J. Biomol. Struct. Dyn. 2017,
35, 1069-1084; e) M. Dračínský, H. M. Möller, T. E. Exner, J. Chem.
Theory Comput. 2013, 9, 3806−3815.
[22] R. K. Harris, V. J. Robinson, J. Magn. Res. 1969, 1, 362-377.
[23] a) A. C. de Dios, J. Progr. Nucl. Magn. Spectr. 1996, 29, 229-278; b) A.
Navarro-Vázquez, Magn. Reson. Chem. 2017, 55, 29-32; c) J. C.
Facelli, M. B. Ferraro, Concepts Magn. Reson. 2013, 42, 261–289; D.
Casarini, L. Lunazzi, A. Mazzanti, Eur. J. Org. Chem. 2010, 2035–2056.
[24] a) J. Kvíčala, R. Hrabal, I. Bartošová, O. Paleta, A. Pelter, J. Fluorine
Chem. 2002, 113, 211-218; b) J. Kvíčala, J. Czernek, S. Böhm, O.
Paleta, J. Fluorine Chem. 2002, 116, 121-127.
[25] R. Peverati, D. G. Truhlar, Phil. Trans. R. Soc. A 2012, 372, 20120476.
[26] Gaussian 09, Revision D.01, M.J. Frisch, G.W. Trucks, H.B. Schlegel,
G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B.
[34] D. Rappoport, F. Furche, J. Chem. Phys. 2010, 133, 134105.
[35] B. Mennucci, J. Tomasi, R. Cammi, J. R. Cheeseman, M. J. Frisch, F. J.
Devlin, S. Gabriel, P. J. Stephens, J. Phys. Chem. A 2002, 106, 6102-
6113.
Mennucci, G.A. Petersson, H.
Nakatsuji, M. Caricato, X. Li, H.P.
Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery,
Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N.
Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A.
Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M.
Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J.
Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R.
Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G.
Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D.
[36] W. B. Schneider, G. Bistoni, M. Sparta, M. Saitow, C. Riplinger, A. A.
Auer, F. Neese, J. Chem. Theory Comput. 2016, 12, 4778-4792; b) C.
Riplinger, P. Pinski, U. Becker, E. F. Valeev, F. Neese, J. Chem. Phys.
2016, 144, 024109.
[37] Y. Zhao, D. G. Truhlar, Theor. Chem. Accounts 2008, 120, 215–241.
[38] W. Kutzelnigg, U. Fleischer, M. Schindler in NMR Basic Principles and
Progress, Vol. 23 (Ed.: J. Seelig), Springer-Verlag, Heidelberg, 1990,
pp. 165-262.
[39] V. J. Gazzard, R. K. Harris, Org. Magn. Reson. 1974, 6, 404-406.
[40] A. L. Henne, W. J. Zimmerschmied, J. Am. Chem. Soc. 1947, 49, 281-
283.
Daniels, Ö. Farkas, J.B.
Foresman, J.V. Ortiz, J. Cioslowski, D.J.
Fox, Gaussian, Inc., Wallingford CT, 2013.
[27]
F. Neese, Wiley interdiscipl. Rev. – Comp. Mol. Sci. 2012, 2, 73-78.
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201800482
European Journal of Organic Chemistry
FULL PAPER
FULL PAPER
3,3,4,4-Tetrafluorocyclobut-1-ene, a
promising substrate for ring-opening
metathesis, was synthesized.
Fluorocyclobutenes
Kateřina Kučnirová,Ondřej Šimůnek,
Markéta Rybáčková, Jaroslav Kvíčala*
Surprisingly, while DFT calculations
using both traditional and modern
functionals failed to describe correctly
19F NMR shielding constants of
fluorocyclobutenes, correct relative
values were obtained by DLPNO-
CCSD or HF calculations, enabling
thus reliable assignment of 19F NMR
signals of both unknown and known
fluorocyclobutenes.
Page No. – Page No.
Structural assignment of
fluorocyclobutenes by 19F NMR
spectroscopy: comparison of
calculated 19F NMR shielding
constants with experimental 19F NMR
shifts
This article is protected by copyright. All rights reserved.