Intramolecular OH...Fluorine Hydrogen Bonding in Saturated, Acyclic Fluorohydrins: The γ-Fluoropropanol Motif

Article abstract of DOI:10.1002/chem.201503253

Fluorination is commonly exercised in compound property optimization. However, the influence of fluorination on hydrogen-bond (HB) properties of adjacent functional groups, as well as the HB-accepting capacity of fluorine itself, is still not completely understood. Although the formation of OH...F intramolecular HBs (IMHBs) has been established for conformationally restricted fluorohydrins, such interaction in flexible compounds remained questionable. Herein is demonstrated for the first time - and in contrast to earlier reports - the occurrence of OH...F IMHBs in acyclic saturated γ-fluorohydrins, even for the parent 3-fluoropropan-1-ol. The relative stereochemistry is shown to have a crucial influence on the corresponding ^h1J_OH...F values, as illustrated by syn- and anti-4-fluoropentan-2-ol (6.6 and 1.9 Hz). The magnitude of OH...F IMHBs and their strong dependence on the overall molecular conformational profile, fluorination motif, and alkyl substitution level, is rationalized by quantum chemical calculations. For a given alkyl chain, the "rule of shielding" applies to OH...F IMHB energies. Surprisingly, the predicted OH...F IMHB energies are only moderately weaker than these of the corresponding OH...OMe. These results provide new insights of the impact of fluorination of aliphatic alcohols, with attractive perspectives for rational drug design.

Full text of DOI:10.1002/chem.201503253

DOI: 10.1002/chem.201503253
Full Paper
&
Hydrogen Bonds
Intramolecular OH···Fluorine Hydrogen Bonding in Saturated,
Acyclic Fluorohydrins: The g-Fluoropropanol Motif
Bruno Linclau,*[a] Florent Peron,[a] Elena Bogdan,[b] Neil Wells,[a] Zhong Wang,[a]
Guillaume Compain,[a] Clement Q. Fontenelle,[a] Nicolas Galland,[b] Jean-Yves Le Questel,[b]
and JØrôme Graton*[b]
h1
Abstract: Fluorination is commonly exercised in compound
property optimization. However, the influence of fluorination
on hydrogen-bond (HB) properties of adjacent functional
groups, as well as the HB-accepting capacity of fluorine
itself, is still not completely understood. Although the forma-
tion of OH···F intramolecular HBs (IMHBs) has been estab-
lished for conformationally restricted fluorohydrins, such
interaction in flexible compounds remained questionable.
Herein is demonstrated for the first time—and in contrast to
earlier reports—the occurrence of OH···F IMHBs in acyclic sa-
turated g-fluorohydrins, even for the parent 3-fluoropropan-
1-ol. The relative stereochemistry is shown to have a crucial
influence on the corresponding
J
values, as illustrated
OH···F
by syn- and anti-4-fluoropentan-2-ol (6.6 and 1.9 Hz). The
magnitude of OH···F IMHBs and their strong dependence on
the overall molecular conformational profile, fluorination
motif, and alkyl substitution level, is rationalized by quantum
chemical calculations. For a given alkyl chain, the “rule of
shielding” applies to OH···F IMHB energies. Surprisingly, the
predicted OH···F IMHB energies are only moderately weaker
than these of the corresponding OH···OMe. These results
provide new insights of the impact of fluorination of
aliphatic alcohols, with attractive perspectives for rational
drug design.
Introduction
perturbation theory) also support the occurrence of OH···F
H-bonding.[1d,4]
The question of whether organofluorines are effective hydro-
gen-bond (H-bond, HB) acceptors or not has been a heavily
debated topic over the years.[1] Key experimental evidence for
intermolecular OH···F H-bonding includes IR[2] and NMR[3]-based
measurements between 4-fluorophenol and fluoroalkanes in
solution. The conclusion of these studies is that organofluor-
ines are able to act as HB acceptors, albeit with a weaker affini-
ty than the usual oxygen- and nitrogen-based HB acceptors.
Computational studies using various theoretical approaches
(e.g. quantum theory of atoms in molecules, intermolecular
In the case of intramolecular HBs (IMHBs), ambiguities about
observed contacts being true HBs or forced consequences of
the molecular structure are complicating factors.[5] Only
a modest number of examples have described IM OH···F inter-
actions in the solution phase, typically with the observation of
h1
a “through-space”
J
coupling.[6] In all of these cases, there
OH···F
is a significant degree of conformational restriction, promoting
or fixing the proximity between the OH and F groups.[7]
Examples featuring g-fluorohydrin motifs (Figure 1) are restrict-
ed to monocyclic carbohydrates (e.g., 1),[8] conformationally
3
restricted cyclohexanes, (e.g., 2,[9] [10]), and bicyclic levogluco-
san derivatives (e.g., 4),[11] all of which display 1,3-coaxial CÀO/
[a] Prof.Dr. B. Linclau, Dr. F. Peron, Dr. N. Wells, Z. Wang, Dr. G. Compain,
Dr. C. Q. Fontenelle
Department of Chemistry, University of Southampton
Highfield, Southampton SO17 1BJ (UK)
Fax: (+44)23-8059-6805
E-mail: bruno.linclau@soton.ac.uk
[b] Dr. E. Bogdan, Dr. N. Galland, Prof. Dr. J.-Y. Le Questel, Dr. J. Graton
CEISAM UMR CNRS 6230, FacultØ des Sciences et des Techniques
UniversitØ de Nantes, 2, rue de la Houssini›re - BP 92208
44322 NANTES Cedex 3 (France)
Fax: (+3)2-51-12-54-02
E-mail: jerome.graton@univ-nantes.fr
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201503253.
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons At-
tribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
Figure 1. g-Fluorohydrin (3-fluoroalkanol)-containing structures with their
h1
J
coupling constants (CDCl3 or CD2Cl2).
OH···F
Chem. Eur. J. 2015, 21, 17808 – 17816
17808 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
CÀF bonds. The peri-substituted naphthalene derivative 5[12]
Table 1. g-Fluorohydrins investigated, with major populated conforma-
tions at 258C/À508C.
and a,a-diphenyl-o-fluorobenzyl alcohol 6 also show a signifi-
h1
cant
J
coupling, whereas it was not detected for an o-flu-
OH···F
Fluorohydrin Major conformations[a] in CHCl3
[b]
orobenzyl alcohol motif.[13] Where applicable, the JOHÀH value
gives further positional information of the O-H proton relative
to the fluorine atom. A nice recent illustration involves 7 and
3
h1
3
8, in which the reduced
J
and JOHÀH values for 8 indicate
OH···F
CFÀF as a weaker HB acceptor than CHÀF, and thus less able
to compete with the ring oxygen.[14]
In contrast, to our knowledge, there are no examples of ex-
perimentally demonstrated OH···F IMHB as part of a flexible,
fully saturated, acyclic 1,3-fluorohydrin in the solution state.
Through gas-phase electron diffraction, the IMHB conformer of
3-fluoropropan-1-ol was identified as a secondary conformer,
with rather low relative populations.[15] Despite the presence of
a bond critical point (BCP) demonstrated through atoms in
molecules (AIM)[16] analysis, 3-fluoropropan-1-ol was recently
h1
reported not to feature a
J
coupling, either in CD2Cl2 or in
OH···F
[D12]cyclohexane.[17] It was attributed to the low calculated
population of the IMHB conformation (11 % in CH2Cl2). In
h1
CDCl3, the 3-fluoropropan-1-ol
J
coupling was also not
OH···F
observed.[18] Similarly, for 3-fluoro-1,2-propanediol, no
coupling could be observed by 1H NMR spectroscopy.[19]
J
OH···F
h1
Herein we describe an extensive combined NMR spectro-
scopic and computational analysis of a range of g-fluorinated
alcohols (Figure 2). We provide evidence of IMHB between flu-
orine and alcohol groups as part of an acyclic chain in solution.
h1
For the first time, NMR
J
couplings have been experimen-
OH···F
tally evidenced and quantified for flexible fluorohydrins, even
for 3-fluoro-, 3,3-difluoro- and 3,3,3-trifluoropropanol. A thor-
ough analysis of the fluorohydrin conformational profile and
the OH···F IMHB energies is provided, as well as an assessment
of relative IMHB strengths with corresponding OH···OMe
interactions.
[a] Unless shown, all OH(y) rotamers are grouped together and figures
represent combined populations; [b] calculated at MP2/6-311+ +
G(2d,p)//MPWB1K/6-31+G(d,p)) level of theory.
Figure 2. List of (racemic) g-fluorohydrins under study.
Results and Discussion
MPWB1K-calculated populations and therefore, only the MP2
results are given. Computed IMHB conformation properties are
summarized in Table 2.
Conformational analysis reveals very different conformer
populations for the investigated fluorohydrins
For syn-4-fluoropentan-2-ol (syn-A), the most stable confor-
mer, gÀ g+ (g+), showing an OH···F IMHB, is stabilized by
2.7 kJmolÀ1 towards the first secondary minimum at 258C, and
represents 39% of the whole population. The dOH···F distance is
2.00 Š, which is well below the sum of the van der Waals radii
(2.57 Š).[20] The next stable conformations are tg+ (rotation
around the C2ÀC3 bond) and gÀt (rotation around the C3ÀC4
The main minimum-energy conformers of the monofluorinated
fluorohydrins are shown in Table 1, where the various C-C-O-H
rotamers have generally been grouped together for the sake
of clarity. Dihedral angle definitions and detailed results are
provided in the Supporting Information (SI1, Tables S1–S10).
No significant differences were observed between MP2- and
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17809 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. Computed features of the IM H-bonded conformations of all
fluorohydrins.
Table 3. g,g-Difluoro- and g,g,g-trifluorohydrins investigated, with major
populated conformations at 258C/À508C.
Fluorohydrin Major conformations[a] in CHCl3
ð2Þ
[b]
Compound Conformation pi dOH···F 1bcp
[%] [Š][a] [ebohrÀ3
EHB
[kJmolÀ1
E
n!s
*
[b]
[c]
[d]
]
]
[kJmolÀ1
]
syn-A
gÀ g+ (g+) 39 2.000 0.0206
24.4
23.7
21.6
22.0
21.4
20.4
21.7
21.7
18.4
19.8
16.6
16.1
17.1
15.4
25.1
24.7
20.1
22.3
19.2
20.0
17.8
19.4
13.8
16.9
9.0
gÀ g+ (g+)
6
3
2.008 0.0202
2.056 0.0186
anti-A
g+ gÀ(g+)
B
C
D
gÀ g+ (g+) 10 2.037 0.0190
gÀ g+ (g+)
gÀ g+ (g+)
8
8
2.065 0.0184
2.074 0.0178
gÀ g+ gÀ (g+) 31 2.062 0.0185
gÀ g+ t(g+) 13 2.050 0.0187
E
F
gÀ g+ gÀ (g+)
gÀ g+ t(g+)
gÀ g+ t(g+)
gÀ g+ gÀ (g+)
gÀ (g+)
6
5
5
4
2.133 0.0161
2.090 0.0173
2.175 0.0148
2.201 0.0142
G
10.5
10.9
8.3
H
I
47 2.172 0.0151
23 2.234 0.0136
gÀ (g+)
[a] IMHB OH···F distance; [b] electron density at the bond critical points
from AIM analysis; [c] HB energy at the MP2/6-311+ +G(2d,p) level; [d] In-
teraction energies from the nF fluorine lone pair to the s*OH antibonding
orbital at the MPWB1K/6-31+G(d,p) level.
bond), their combined populations (49%) exceeding that of
gÀ g+ (g+). The conformational profile of anti-A is very differ-
ent: the dominance of the major conformers (gÀ gÀ) is even
more pronounced, but they do not exhibit any IMHB. Indeed,
the conformations with the linear (zigzag) pentyl chain repre-
sent 68% of the population, and the conformers featuring an
IMHB are only slightly populated (6% and 3%). Hence, the
extent of IM H-bonding significantly depends on relative
fluorohydrin stereochemistry. Compared to anti-A, the
gÀ gÀconformer remains the absolute minimum for 4-fluorobu-
tan-2-ol (B), but is much less populated (44%). In contrast, the
amount represented by the gÀ g+ (g+) IMHB conformer is
raised, resulting in a 10% population. The conformational pro-
file for the 3-fluoropropan-1-ols C and D shows that the gemi-
nal dimethyl group does not have a large effect. In both cases,
the gÀ gÀconformer appears consistently as the largest popu-
lated one, with the IMHB conformation representing only 8%
with a lengthening of the OH···F distances (2.07 Š) with respect
to compounds A (2.00 Š) and B (2.04 Š). For comparison,
Cormanich et al.[17] calculated a population of 11 % for 3-fluoro-
propan-1-ol (D) in CH2Cl2 at the MP2/aug-cc-pVDZ level, and
Badawi and co-workers[18] predicted 13% (B3LYP/6-311 +
G(d,p)) and 10% (MP2/6-311+G(d,p)) of such IMHB conformers
in the gas phase.
[a] Unless shown, all OH (y) rotamers are grouped together, and figures
represent combined populations; [b] calculated at MP2/6-311+ +G(2d,p)//
MPWB1K/6-31+G(d,p)) level of theory.
5 times more populated in E. In addition, the syn-fluorine of E
appears significantly more chelated than the anti-fluorine,
whereas the populations of the IM H-bonded conformations of
the diastereotopic fluorines are very similar for F. For 3,3-di-
fluoropropanol G, the gÀ gÀ t conformation remains the major
conformer. Both its IM H-bonded conformers are populated in
small amounts and the OH···F distances are significantly longer
than in the previous compounds.
The three lowest energy minima of the trifluorinated H
show a gÀ 2-butanol chain, with the major conformer featuring
an IMHB. Finally, in 3,3,3-trifluoropropanol I, the population of
the IM H-bonded conformer is now significantly lower, and the
OH···F distance observed (2.23 Š) is the longest of the series.
The main conformer now has a trans dihedral angle between
the OH and the fluorinated group, a feature that is not ob-
served with such a high population in any of the other motifs
(except for the tgÀ t conformation in G).
Within the whole substrate series, from CCl4 to CH2Cl2 (for
full data, see the Supporting Information, SI1), it is worth
noting that our analysis shows that an increase in the solvent
polarity favors the non-chelated conformers.
The difluorinated alcohols E and F, both featuring diastereo-
topic fluorine atoms, display the same set of conformational
minima, but the presence of the C5 methyl group significantly
affects their relative populations (Table 3). For E, gÀ gÀ t is the
main conformer. This is the only instance of a significant stable
conformation with a syn OH/CH3 relationship. Consistently, the
gÀ gÀ t conformation is also the major conformer for F and is
much more populated without the C5 methyl substituent. The
IMHB conformers for E represent 44% of the whole popula-
tion, which is notably greater than the corresponding value
calculated for F. For instance, the gÀ g+ gÀ (g+) conformation is
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17810 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
The change of temperature is predicted to strongly affect
the fluorohydrin conformational profile
shown in Figures 3 and 4 (for detailed relevant spectra, see the
Supporting Information, SI2). The results are compared with
those obtained for the rigid 3-fluorocyclohexanol 2 (h1JOH···F
=
The influence of temperature on the conformational distribu-
tion was also studied (Table 1, Table 3). For the compounds
with a significantly populated IMHB conformation (syn-A, E,
and H), a population increase is calculated upon temperature
decrease. This is expected since lowering the temperature in-
creases the probability to populate the lowest energy con-
formers. But, the change is more subtle for the IMHB conform-
ers weakly populated at 258C. For B, C, D, and I, the popula-
tion of the IMHB conformations slightly increases at À508C,
whereas, for anti-A, F, and G, a population decrease is predict-
ed. Interestingly, the two chelated conformers of 4,4-difluoro-
pentan-2-ol E have opposite behaviors: the gÀ g+ gÀ(g+) con-
former, chelated with the syn-fluorine, is more populated at
low temperature, whereas the gÀ g+ t(g+) conformer, chelated
with the anti-fluorine, is slightly less populated. This is consis-
tent with the observations for syn-A and anti-A, respectively.
12.1 Hz, 93% IMHB conformer population in CHCl3).[9]
For syn-A, OH appeared as a doublet of doublets, with
1H{19F} analysis proving a 6.6 Hz coupling to fluorine (Figure 3).
Its magnitude is in the expected range, based on the
calculated population of 39%. The computed value (À7.9 Hz),
weighted according to the Boltzmann distribution of the
various conformers in chloroform, agrees well with the experi-
1
mental coupling. At À508C, the H NMR spectrum shows both
h1
a significant increase in
J
value (9.9 Hz, Figure 3) and a
OH···F
chemical shift change (Dd=0.48 ppm), attributed to a popula-
tion increase of the IMHB conformer. This is supported by
theoretical calculations with the higher predicted population
of such conformers (20%), and the large increase of the
computed weighted coupling constants (about 4 Hz). The
h1
observed
J
value is much smaller, but clearly observable
OH···F
for anti-A (see the Supporting Information, SI2, 5.3.2), with
a computed combined population of IMHB conformers of 9%.
No increase in coupling constant magnitude is then observed
at À508C, in line with the computational predictions, and
a weaker chemical shift change (Dd=0.22 ppm) is measured.
Similar coupling constants were found for B (2.2 Hz) and C
(1.7 Hz), in reasonable agreement with the corresponding
theoretical h1JOH···F values.
NMR experiments reveal OH···F coupling constants for all
investigated substrates
h1
NMR analysis focused on the multiplicity,
J
value, and
OH···F
chemical shift of the alcohol protons at 258C and À508C,
taking into account that the observed values are averaged
over the conformer populations. The results are given in
Table 4 and spectral details for syn-4-fluoropentan-2-ol (syn-A),
3-fluoropropan-1-ol (D), and 4,4-difluoropentan-2-ol (E) are
The 3-fluoropropan-1-ol (D) deserves special attention, given
h1
previous communications reporting that no
J
value could
OH···F
be observed for this compound in CD2Cl2, CDCl3, or
[D12]cyclohexane.[17,18] With an exceptionally resolved NMR
spectrum (Figure 3), we have succeeded to detect an OH
signal clearly appearing as a triplet of doublets in CDCl3, and
1
the H{19F} analysis proved that there is a 1.4 Hz coupling to
Table 4. Experimental and computed NMR data obtained in CDCl3.
fluorine. This coupling constant remained essentially the same
h1
h1
Compound
J
@258C [Hz]
J
@À508C [Hz]
Dd [ppm][a]
exp.
OH···F
OH···F
upon cooling to À508C (1.7 Hz). These observations are sup-
exp.[b]
calc.[c]
exp.[b]
calc.[c]
h1
ported by the weighted theoretical
J
values of À1.2 and
OH···F
syn-A
6.6
À7.9
À1.5
À1.8
À1.2
À1.2
9.9
À11.9
À1.2
À2.4
À1.3
À1.4
À1.4 Hz, computed at 25 and À508C, respectively. Further-
0.48
0.22
1
more, the H NMR spectrum of D in the more polar CD2Cl2 also
anti-A
1.9
2.2
1.7
1.4
1.8
shows the presence of an IMHB at 258C (see the Supporting
h1
Information, SI2, 5.6.8), with a decreased
J
value (1.0 Hz).
OH···F
B
C
D
The computed value is similarly slightly weaker in CD2Cl2
(1.1 Hz) owing to the slightly less populated IMHB conformer.
It is moreover in good agreement with the weighted value
computed by Cormanich et al (1.68 Hz).[17]
1.7
0.26
0.32
3.5
1.4
À4.9
À1.2
4.7
1.7
À5.9
À1.0
For the di- and trifluorinated derivatives, multiple couplings
of the alcohol hydrogen atom with the fluorine atoms were ex-
pected. The conformational analysis of 4,4-difluoropentan-2-ol
E
F
0.6
0.6
À0.2
À0.4
À0.3
À0.2
(E) revealed this compound as an interesting case with the
h1
prediction of a significant difference in
J
values (À4.9 and
OH···F
À0.3
À0.1
À0.2
À1.2 Hz) for its two diastereotopic fluorine atoms. We were de-
G
0.4(t)
0.1
lighted to be able to observe such a distinction experimentally
(Figure 4): a triplet of doublets was seen for the OH group,
H
I
0.7(q)
0.3(q)
À1.3
À0.1
À1.7
À0.2
h1
with
J
values of 3.5 and 1.4 Hz. In addition, the predicted
OH···F
increase of the former coupling at À508C (+1.0 Hz) was also
h1
observed experimentally (+1.2 Hz, with
J
value of 4.7 Hz).
OH···F
[a] Chemical shift difference upon cooling to À508C; [b] sign not
For F, an apparent triplet was observed with a very small cou-
pling constant (0.6 Hz). Equally, for G a triplet was observed
determined; [c] calculated at the B97-2/pcJ-2 level of theory.
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17811 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1H NMR coupling constant analysis confirms the
chair-like conformation of syn-A
The chair-like IMHB conformation for syn-A was con-
firmed by further NMR spectroscopic analysis
(Figure 5). An “axial” H3 and “equatorial” H3’ can clear-
ly be identified; the former with two large ax–ax cou-
pling constants (9.0, 7.9 Hz), and the latter with two
smaller eq–ax coupling constants (4.1, 4.1 Hz), to H2
and H4. Upon cooling to À508C, these values in-
crease or decrease in accordance with the population
increase in the IMHB conformation compared to tg+
3
and gÀ t. The increase in JH3’ÀF of approximately 5 Hz
3
upon cooling is interpreted similarly, and the JOHÀH2
coupling value corresponds to a gauche dihedral
angle. Its calculated value is in remarkable coherence
with experiment (both 3.4 Hz).
AIM analyses provide evidence for IMHB in all
cases and allow quantification of their energies
For all the compounds under study, BCPs between
the H(O) and F atoms were systematically found
through AIM analyses on their IMHB conformations
(Table 2), validating the presence of IMHB interac-
tions. Beyond the difficulties to estimate the strength
of intramolecular interactions, less properly defined
than intermolecular interactions through the super
molecule approach, the estimation of the HB energy
(EHB), based on the potential energy density Vb at the
BCP, is however informative, with the caveat that it
overestimates the actual HB strength.[21] For this
reason, the following comparison will focus on the
relative trends calculated rather than on the absolute
values. The highest value of the series is calculated
for syn-A (24.4 kJmolÀ1), which corresponds to the
structure exhibiting the shortest IMHB (2.00 Š). Con-
versely, 3,3,3-trifluoropropan-1-ol (I) shows the weak-
est HB energy (15.4 kJmolÀ1) and the longest IMHB
(2.23 Š), these two energetic and structural parame-
ters being strongly correlated (r2 =0.983). It is inter-
esting to note that the IMHB strengths are of the
same order of magnitude in anti-A and syn-A, despite
an IMHB conformation 6 times less populated for the
former. In the same vein, the OH···F distances (2.06
and 2.05 Š) and the HB energies (21.6 and
21.7 kJmolÀ1) are very similar for the gÀ g+ gÀ (g+)
and gÀ g+ t(g+) conformers in E, whereas their rela-
tive populations differ significantly (31 and 13%, re-
spectively). It is worth noting that the electron densi-
ty values at the BCP, also commonly used as HB strength de-
scriptor, are strongly correlated to EHB (r2 =0.998).
1
Figure 3. Details of the H NMR spectra of syn-A and D at 258C and À508C, showing the
h1
changes in d(OH) and
J
.
OH···F
with a likewise low J value. Finally, it was surprising to observe
a doublet of quartets for the OH groups in both trifluorinated
fluorohydrins H and I (see the Supporting Information, SI2,
5.10–11), even if the magnitude of the coupling constants is
very small (0.3–0.7 Hz).
The effective strength of the OH···F IMHB deserves compari-
son with conventional OH···O IMHB energies. In b-diketones,
strong OH···O IMHBs involve a conjugated system between the
carbonyl and the hydroxy groups with a calculated EHB of
around 100 kJmolÀ1 [22]
The OH···F HB strengths found in the
.
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17812 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the EHB values of syn-A and anti-A (24.4 and
23.7 kJmolÀ1). These relative energies correlate with
the intramolecular distances, which are shorter for
OH···OMe (1.92 Š for both J diastereomers) than for
OH···F (2.00 Š). The impact of the OH···OMe IMHB on
the populations of the corresponding structures is
consistent with the increased EHB values: the IMHB
structures of anti-J represent almost 60% of the total
population conformers (9% for anti-A), whereas
those of syn-J are as high as 95% (39% for syn-A).
Conversely, comparisons with other weak interac-
tions demonstrate the stronger fluorohydrin OH···F
interactions in compounds A–I. The CH···O IMHBs of
adenosine derivatives were computed to range from
7 to 16 kJmolÀ1 using EHB descriptor.[23] Similarly, in
short intermolecular CH···F H-bonds identified in crys-
talline organic fluorine structures EHB reached
12 kJmolÀ1 and the complexation energies of small
organofluorine molecules were calculated at high
levels of theory to be lower than 10 kJmolÀ1.[4,24]
Further analyses give insight into the various stabi-
lizing and destabilizing interactions
Both noncovalent interaction (NCI) and natural bond
orbital (NBO) analyses corroborate the AIM results for
the chelated conformers. With NCI, an attractive con-
tribution relative to the interaction between the CÀF
and CÀO(H) groups systematically outweighs the re-
pulsive counterpart associated with the parallel orien-
tation of the two corresponding dipoles (see the Sup-
porting Information, Tables S14–S22), hence corre-
sponding to an effective IMHB, as illustrated by the
1
Figure 4. Details of the H NMR spectra of E at 258C and À508C, showing the changes in
blue isosurfaces shown in Figure 7. The NBO interac-
h1
d(OH) and
J
.
OH···F
ð2Þ
tion energies E
from the nF fluorine lone pairs to
n!s
*
the s*OH antibonding orbital, describing the charge
transfer component of the interaction, range from 25 to
8 kJmolÀ1 (Table 2), with variations in agreement with those
observed with the EHB descriptor.
Figure 6. Reference compounds included for comparison with the
fluorohydrins.
Figure 5. Conformer populations, with relevant coupling constants of syn-A
at 258C and À508C
current work are indeed 4–6 times weaker. In contrast, compar-
ison of syn-A and anti-A with the corresponding 4-methoxy-
pentan-2-ol diastereomers syn-J and anti-J (Figure 6 and
Table S11 in the Supporting Information, SI1) showed a different
picture. The calculated EHB for both diastereomers of J are 30.7
and 29.7 kJmolÀ1, which is only 25% higher than
Figure 7. NCI isosurface plots of gÀ g+ (g+) conformers of syn-A and anti-A
compounds drawn with a reduced density gradient (RDG) value of 0.6 and
the blue-green-red values ranging from À0.02 to 0.01 a.u.
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17813 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
In addition to the OH···F IMHB, other secondary interactions
contribute to the stabilization or destabilization of the various
conformers. For example, the gÀ gÀ conformations of anti-A,
representing almost 70% of the whole population, show simul-
taneously a CH···O and a CH···F 5-membered interaction, yield-
ing conformers more stabilized than the gÀ g+ (g+) IM H-
bonded conformer. Conversely, these secondary interactions
cannot operate simultaneously in syn-A, with only CH···F in the
tg+ or CH···O in the gÀ t conformers, resulting in the IMHB con-
former becoming the global energetic minimum. These fea-
tures, which are not detectable through AIM analysis, are cor-
roborated by the NBO calculations. Indeed, weak interaction
energies, from 1.0 to 2.5 kJmolÀ1, are found for syn-A and anti-
A between either the nF fluorine or nO oxygen lone pairs and
the s*CH antibonding orbitals. It may rationalize why, despite
similar IMHB characteristics (dOH···F and EHB), the corresponding
relative populations differ significantly in syn-A and anti-A.
Moreover, the H-bonded conformers of anti-A undergo an ad-
ditional CH···CH3 destabilizing interaction (a classic gauche–
butane interaction), whereas such repulsive contributions
rather occur in tg+ and gÀ t conformers of syn-A. The higher
population of non-chelated conformers for anti-J compared to
syn-J is similarly explained. With 4,4-difluoropentan-2-ol E, the
higher stabilization of the gÀ g+ gÀ (g+) vs. gÀ g+ t(g+) conform-
ers is explained by the attractive CH···F contribution in the
former replaced by a repulsive CH···CH3 interaction in the
latter.
with a CHF motif. This effect was also invoked by Suhm and
co-workers to explain the hydrogen-bonding properties of
progressively fluorinated ethanol molecules with water.[27]
This trend is consistent with the computed EHB energies for
A–I, which range from 24.4 to 20.4 kJmolÀ1 for the monofluori-
nated compounds, from 21.7 to 16.1 kJmolÀ1 for the difluori-
nated compounds, and from 17.1 to 15.4 kJmolÀ1 for the tri-
fluorinated compounds (Table 2), provided the alkyl chain is
strictly conserved. Indeed, variations as subtle as methylation
can lead to an overlap of these three energetic ranges: the di-
fluorinated E has larger EHB values than the monofluorinated C
and D, and the same is true when comparing the trifluorinated
H to the difluorinated G. Similar conclusions can be drawn by
ð2Þ
considering the NBO interaction energies E
. These behav-
n!s
*
iors are not in line with the “rule of shielding” proposed by
Dalvit and Vulpetti[26a] (e.g., E: dF =À90.3, À89.5 ppm with
E
HB =À21.7 kJmolÀ1
;
vs. D: dF =À221.8 ppm with EHB
=
À20.4 kJmolÀ1).
These observations are confirmed by the calculation of rele-
vant electrostatic potential descriptors. The Vmin descriptor is
related to HB acceptor ability, and was thus shown by Dalvit
and Vulpetti to reflect the reduced HB-accepting capacity from
mono- to trifluoro derivatives.[3] On the other hand, the Va(r)
descriptor is related to HB donor ability.[28] As it is not possible
to calculate these descriptors for the chelated conformations
due to the perturbation of the IMHB, the descriptors calculated
for the tt(t) conformer of compounds anti-A, B, and D–I were
used for illustration, as shown in Table 5.
Considering the whole data set of conformations, it appears
that the structures that are more stable than the IMHB con-
formers contain systematically at least two CH···X stabilizing in-
teractions. This is consistent with the conformational profiles
of syn- and anti-2-fluoro-4-methoxypentanes (syn-K and anti-K;
see the Supporting Information, Table S12, SI1). The linear
(zigzag) pentane conformation is clearly preferred for anti-K
(7 kJmolÀ1, 87%), as found for anti-A for which the three first
linear conformers represented 68% of the whole population.
These anti-K conformations have short CH3···F or CH3···O
intramolecular distances (2.34–2.43 Š). With syn-K, the linear
conformation represents only 2% of the whole population due
to the significant repulsion between the fluorine and
methoxy groups, whereas it reaches 40% of the population in
syn-A, due to its ability to establish an IMHB. This analysis is
also consistent with the conformational profile of other 2,4-di-
substituted pentanes, as reported by Hoffmann et al.[25]
Table 5. Electrostatic potential values [kJmolÀ1] calculated for the tt(t)
conformers on the OH [Va(r)] and F [Vmin] sites at the MPWB1K/6-31+
G(d,p) level of theory.[a]
Monofluoro[b]
Difluoro[c,d]
Trifluoro[d,e]
Va(r)
Vmin
Va(r)
Vmin
Va(r)
Vmin
pentanol
butanol
propanol
860.4 À143.0
863.8 À135.5
870.4 À132.7
868.5 À120.0
872.2 À106.7
880.9 À104.5
882.2 À68.5
891.4 À66.0
[a] The higher Va(r), the better the HB-donating capacity; the lower the
Vmin the better the HB-accepting capacity; [b] anti-A9, B17, and
,
F
C10/D10; [c] E7, F5, G3; [d] Vmin values given for the trans fluorine in the
polyfluorinated derivatives; [e] H6, I1.
It can be clearly seen that, although additional fluorination
indeed reduces the fluorine HB-accepting capacity, there is
a concomitant, but weaker, increase in OH HB donating capaci-
ty. Furthermore, it is evident from Table 5 that both fluorine
and OH HB properties depend on the alkane chain length, in
a significant way. As a consequence, the OH···F IMHB strength
encountered in the current series significantly depends on in-
terlinked competing electronic factors, leading to small energy
changes, and resulting in an overlapping energy range be-
tween mono- and difluorinated fluorohydrins, and between di-
and trifluorinated fluorohydrins depending on the alkyl chain
length. Interestingly, in the difluorinated structures, compared
to the Vmin value of the trans-fluorine atom shown in Table 5,
Comparison of IMHB between the monofluorinated and
di-/trifluorinated derivatives
Dalvit and Vulpetti[3,26a] demonstrated the impact of the fluo-
rine environment on its HB-accepting capacity in the context
of intermolecular OH···F interactions, with CHF>CF2 >CF3 as
a general HB-accepting trend. This ranking is consistent with
the evolution of the electron density on fluorine, as displayed
by their respective fluorine chemical shift values [dF(CF)<
dF(CF2)<dF(CF3)] and referred to as the “rule of shielding”. Fur-
thermore, Bernet and Gouverneur[14] demonstrated that the
OH···F IMHB is weaker when a CF2 motif is involved than it is
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17814 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the Vmin values of the second fluorine atom are significantly de-
creased (DVꢀ20 kJmolÀ1; not shown) suggesting a much
weaker HB-accepting ability. In contrast, for the trifluorinated
alcohols, the difference between the H-bond basicity of the
trans fluorine and the two other fluorine atoms found is much
less pronounced (DVꢀ3 kJmolÀ1; not shown).
Experimental Section
Computational details
All DFT calculations were performed by using version D.01 of the
Gaussian 09 program.[30] The conformational landscape of the fluo-
rohydrins was exhaustively investigated at the MPWB1K/6–31+
G(d,p) level in CCl4 medium through, in a first step, simultaneous
rigid scans of their two f(C-C-C-X) dihedral angles from 08 to 3608
in steps of 308, by considering in addition three different orienta-
tions of the f(H-O-C-H) dihedral angle (1808, 608, and À608). Sol-
vent effects were systematically introduced by means of the polar-
izable continuum model (PCM) within the integral equation formal-
ism. The geometry optimization and the frequency calculation of
the various energetic minima were then carried out at the same
level of theory. Eventually, single-point calculations at the MP2/6–
311+ +G(2d,p) level were carried out in CCl4, CHCl3, and CH2Cl2
solvents. The electronic energies were then converted into Gibbs
free energies by using standard thermodynamic corrections from
the MPWB1K/6–31+G(d,p) frequency calculations. The high flexi-
bility of the investigated compounds generates significant
amounts of secondary conformers. In the Tables S1–S12 (see the
Supporting Information, SI1), the relative energies and Boltzmann
populations are given for all conformations within 12 kJmolÀ1
from the global energy minimum, for each fluorohydrin, together
with a detailed explanation of the used nomenclature.
Conclusion
This work introduces compelling experimental evidence of the
occurrence of OH···F intramolecular hydrogen bonding in fully
saturated acyclic compounds containing a 1,3-fluorohydrin
h1
motif. The presence of
J
coupling constants, sometimes
OH···F
of considerable magnitude (up to 6.6 Hz at 258C; 9.9 Hz at
À508C), was demonstrated. The experimental NMR data were
fully consistent with DFT calculations. The comparison between
the 4-fluoropentan-2-ol and 4-methoxypentan-2-ol systems
highlights the significance of the OH···F interaction, indicating
that the IMHB energy of the former reaches almost 80% of the
latter. Following the “rule of shielding” reported by Dalvit and
Vulpetti,[26a] as well as the findings of Bernet and Gouver-
neur,[14] di- and trifluorination leads to a reduction in H-bond
strength. However, the effect was found to be moderate and
could be easily overcompensated by other electronic effects,
such as the concomitant increase in alcohol HB donating ca-
pacity. Finally, the rule appears to be limited to a given alkyl
chain. Significantly, the absence of conformational rigidity re-
moves any ambiguity about the OH···F interaction being the
result of a forced contact, allowing for an unbiased study of
the multitude of often opposing effects that determine the
extent of IMHB. Fluorination of alkanols at the g-position re-
sulted in a complex conformational profile, with the influence
of the fluorination operating simultaneously through OH···F
IMHB, attractive CÀH···F interactions, and steric considerations
such as repulsive contributions of CÀMe with CÀH and CÀF, or
CÀO/CÀF dipole-mediated interactions. As an illustration, the
very low population (2%) of the linear (zigzag) alkyl chain in
syn-4-fluoro-2-methoxypentane is raised to 39% by
introducing OH···F IMHB, as in syn-4-fluoropentan-2-ol, and to
95% for syn-4-methoxy-pentan-2-ol.
The spin–spin coupling constants (J) were estimated from the pre-
vious optimized geometries by using the gauge-invariant atomic
orbital (GIAO) method. The hybrid B97–2 functional[31] and the pcJ-
2 basis set, specifically designed for the calculation of these NMR
parameters,[23] were used. Again, solvent (CHCl3) effects were intro-
duced through the PCM model. Calculated J values were averaged
over all conformers according to their relative populations in CHCl3
at 298 K and 223 K.
To gain more insights on the IMHB interactions at work in relevant
conformers of the various compounds, AIM topological analy-
ses[16,32] of the PCM/MP2/6–311+ +G(2d,p) wave functions were
carried out using the AIM2000 program.[33] Electron density values,
1bcp, are computed at the BCP, and the corresponding HB energies
EHB are estimated from the potential energy densities Vb.[34] In addi-
tion, NCI[35] analyses of the same wavefunctions were also per-
formed by using the NCIPLOT 3.0 program,[36] to detect additional
secondary interactions and to estimate their contributions. Finally,
the NBO[37] method was applied at the PCM/MPWB1K/6–31+
G(d,p) level to provide a complementary description of the IMHB.
Its strength is related to the charge transfer between the nF fluo-
rine lone pairs and the s* HB-donor antibonding orbitals by using
ð2Þ
the corresponding E
interaction energies computed from the
n!s
*
This work thus provides significant new insights on OH···F IM
H-bonding in fluoroalkanols, and shows that these are much
more important than previously assumed. The advances re-
ported herein will not only contribute to a better understand-
ing of the impact of aliphatic fluorination, currently increasing-
ly exercised in property optimization of organic materials and
bioactive compounds, but will also be of interest to the many
areas where hydrogen bonding is of importance, for example
rational drug design, where the existence of OH···F IMHB in
non-aqueous environments can be exploited: the formation of
IMHB has a pronounced effect on important ligand molecular
properties, including membrane permeability.[26b,29] Further
investigations about the influence of the fluorination on the
intermolecular hydrogen-bonding properties of acyclic
1,3-fluorohydrins are in progress.
second-order perturbation theory.
NMR spectroscopy
1
1
For all substrates, the H, 19F, and H{19F} NMR spectra were collect-
ed after rigorous drying of the solutions (9–15 mm) with activated
molecular sieves, which is required to suppress water–solute inter-
actions that would interfere with the OH···F IMHB. A detailed
procedure is provided in the Supporting Information (SI2).
Fluorohydrin synthesis
The synthesis of the novel compounds syn- and anti-A, B, C, E, F,
and G is detailed in the Supporting Information (SI3 and SI4). The
other compounds were commercially available and used without
purification.
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17815 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[18] H. M. Badawi, W. Forner, S. A. Ali, Spectrochim. Acta A 2008, 69, 263–
271.
Acknowledgements
[19] L. A. F. Andrade, J. M. Silla, C. J. Duarte, R. Rittner, M. P. Freitas, Org.
Biomol. Chem. 2013, 11, 6766–6771.
[20] A. Bondi, J. Phys. Chem. 1964, 68, 441–445.
[21] I. Mata, I. Alkorta, E. Espinosa, E. Molins, Chem. Phys. Lett. 2011, 507,
185–189.
[22] a) M. Zahedi-Tabrizi, B. Gerivani, S. F. Tayyari, Spectrochim. Acta Part A
2015, 136, 731–742; b) H. S. Delarami, A. Ebrahimi, S. Bazzi, S. M. H.
Khorassani, Struct. Chem. 2015, 26, 411 –419.
[23] F. Jensen, J. Chem. Theory Comput. 2006, 2, 1360–1369.
[24] I. Hyla-Kryspin, G. Haufe, S. Grimme, Chem. Eur. J. 2004, 10, 3411–3422.
[25] R. W. Hoffmann, D. Stenkamp, T. Trieselmann, R. Gottlich, Eur. J. Org.
Chem. 1999, 2915–2927.
[26] a) C. Dalvit, A. Vulpetti, ChemMedChem 2011, 6, 104–114; b) C. Dalvit, A.
Vulpetti, ChemMedChem 2012, 7, 262–272.
The EPSRC (grant EP/K016938/1), the European Community
(INTERREG IVa, AI-Chem, project 4494/4196), and the ANR
(JCJC “ProOFE” grant (ANR-13-JS08-0007-01) are gratefully ac-
knowledged for their financial support. The current work was
granted access to the HPC resources of (CCRT/CINES/IDRIS)
under the allocation c2014085117 made by GENCI, and HPC re-
sources from ArronaxPlus, grant ANR-11-EQPX-0004 funded by
the ANR. We thank the CCIPL for grants of computer time. We
are grateful for EPSRC Core Capability Funding (EP/K039466/1).
Apollo Scientific is gratefully acknowledged for the donation of
chemicals.
[27] M. Heger, T. Scharge, M. A. Suhm, Phys. Chem. Chem. Phys. 2013, 15,
16065–16073.
[28] a) P. W. Kenny, J. Chem. Inf. Model. 2009, 49, 1234–1244; b) J. Graton, F.
Besseau, A.-M. Brossard, E. Charpentier, A. Deroche, J.-Y. Le Questel, J.
Phys. Chem. A 2013, 117, 13184–13193.
Keywords: fluorination
hydrogen bonds · NMR spectroscopy · quantum calculations
·
intramolecular interactions
·
[29] B. Kuhn, P. Mohr, M. Stahl, J. Med. Chem. 2010, 53, 2601–2611.
[30] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Na-
katsuji, 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. Ha-
segawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. Heyd,
E. N. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K.
Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, N. J. 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. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox,
Revision D.01 ed., Gaussian, Inc., Wallingford, CT, USA, 2009.
[31] P. J. Wilson, T. J. Bradley, D. J. Tozer, J. Chem. Phys. 2001, 115, 9233–
9242.
[1] a) J.-F. Paquin, P. Champagne, J. Desroches, Synthesis 2014, 47, 306–
322; b) H.-J. Schneider, Chem. Sci. 2012, 3, 1381–1394; c) D. O’Hagan,
Chem. Soc. Rev. 2008, 37, 308–319; d) J. D. Dunitz, R. Taylor, Chem. Eur.
J. 1997, 3, 89–98; e) J. A. K. Howard, V. J. Hoy, D. O’Hagan, G. T. Smith,
Tetrahedron 1996, 52, 12613–12622.
[2] a) W. Adcock, J. Graton, C. Laurence, M. Lucon, N. Trout, J. Phys. Org.
Chem. 2005, 18, 227–234; b) C. Ouvrard, M. Berthelot, C. Laurence, J.
Chem. Soc. Perkin Trans. 2 1999, 1357–1362.
[3] C. Dalvit, C. Invernizzi, A. Vulpetti, Chem. Eur. J. 2014, 20, 11058–11068.
[4] P. Panini, D. Chopra, Cryst. Growth Des. 2014, 14, 3155–3168.
[5] G. Desiraju, T. Steiner, The Weak Hydrogen Bond: Applications to Structur-
al Chemistry and Biology, Oxford University Press, Oxford, UK 1999.
[6] Reviews: a) I. Alkorta, J. Elguero, G. S. Denisov, Magn. Reson. Chem.
2008, 46, 599–624; b) J.-C. Hierso, Chem. Rev. 2014, 114, 4838–4867.
[7] a) M. D. Struble, C. Kelly, M. A. Siegler, T. Lectka, Angew. Chem. Int. Ed.
2014, 53, 8924–8928; b) H. Takemura, M. Kotoku, M. Yasutake, T. Shin-
myozu, Eur. J. Org. Chem. 2004, 2019–2024.
[8] K. Nakai, Y. Takagi, T. Tsuchiya, Carbohydr. Res. 1999, 316, 47–57.
[9] J. Graton, Z. Wang, A.-M. Brossard, D. GonÅalves Monteiro, J.-Y. Le Ques-
tel, B. Linclau, Angew. Chem. Int. Ed. 2012, 51, 6176–6180; Angew.
Chem. 2012, 124, 6280–6284.
[10] A. Zapata, B. Bernet, A. Vasella, Helv. Chim. Acta 1996, 79, 1169–1191.
[11] B. Bernet, A. Vasella, Helv. Chim. Acta 2007, 90, 1874–1888.
[12] H. Takemura, R. Ueda, T. Iwanaga, J. Fluorine Chem. 2009, 130, 684–688.
[13] H. Takemura, M. Kaneko, K. Sako, T. Iwanaga, New J. Chem. 2009, 33,
2004–2006.
[14] G. T. Giuffredi, V. Gouverneur, B. Bernet, Angew. Chem. Int. Ed. 2013, 52,
10524–10528; Angew. Chem. 2013, 125, 10718–10722.
[15] A. D. Richardson, K. Hedberg, J. Mol. Struct. 2001, 567, 187–196.
[16] R. F. W. Bader, Chem. Rev. 1991, 91, 893–928.
[32] R. F. W. Bader, Atoms in Molecules:
Oxford, UK, 1994.
A Quantum Theory, Clarendon,
[33] F. Biegler-Konig, J. Schonbohm, D. Bayles, J. Comput. Chem. 2001, 22,
545–559.
[34] E. Espinosa, E. Molins, C. Lecomte, Chem. Phys. Lett. 1998, 285, 170–
173.
[35] E. R. Johnson, S. Keinan, P. Mori-Sanchez, J. Contreras-Garcia, A. J.
Cohen, W. Yang, J. Am. Chem. Soc. 2010, 132, 6498–6506.
[36] J. Contreras-García, E. R. Johnson, S. Keinan, R. Chaudret, J.-P. Piquemal,
D. N. Beratan, W. Yang, J. Chem. Theory Comput. 2011, 7, 625–632.
[37] F. Weinhold, C. R. Landis, Valency and Bonding: A Natural Bond Orbital
Donor–Acceptor Perspective, Cambridge University Press, Cambridge,
UK, 2005.
[17] R. A. Cormanich, R. Rittner, M. P. Freitas, M. Buhl, Phys. Chem. Chem.
Phys. 2014, 16, 19212–19217.
Received: August 17, 2015
Published online on October 23, 2015
Chem. Eur. J. 2015, 21, 17808 – 17816
www.chemeurj.org
17816 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim