Preparation of (Pentafluorosulfanyl)benzenes by Direct Fluorination of Diaryldisulfides: Synthetic Approach and Mechanistic Aspects

Article abstract of DOI:10.1002/chem.201902651

Direct fluorination of ortho-, meta- and para-substituted aromatic thiols and disulfides using elemental fluorine afforded substituted (pentafluorosulfanyl)benzenes. This work thus represents the first study of the scope and limitation of direct fluorination for the synthesis of new SF₅-containing building blocks. Fluorinations in batch and flow modes were compared. A comprehensive computational study was carried out employing density functional and wave function methods to elucidate the reaction mechanism of the transformation of ArSF₃ into ArSF₅. Eliminating various nonradical pathways, it has been shown that the reaction proceeds by a radical mechanism, initiated by the attack of the F^. on the ArSF₃ moiety, propagated via an almost barrierless F₂+ArSF₄ ^.→ArSF₅+F^. step and terminated by the ArSF₄ ^.+F^.→ArSF₅. Most of the calculated data are in very good agreement with experimental observations concerning the ortho-substituent effect on the reaction rates and yields.

Full text of DOI:10.1002/chem.201902651

A Journal of
Accepted Article
Title: Preparation of (Pentafluorosulfanyl)benzenes by Direct
Fluorination of Diaryldisulfides: Synthetic Approach and
Mechanistic Aspects
Authors: Javier Ajenjo, Blanka Klepetářová, Martin Greenhall, Daniel
Bím, Martin Culka, Lubomír Rulíšek, and Petr Beier
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: Chem. Eur. J. 10.1002/chem.201902651
Link to VoR: http://dx.doi.org/10.1002/chem.201902651
Supported by

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
Preparation of (Pentafluorosulfanyl)benzenes by Direct
Fluorination of Diaryldisulfides: Synthetic Approach and
Mechanistic Aspects
Javier Ajenjo,[a][b] Blanka Klepetářová,[a] Martin Greenhall,[c] Daniel Bím,[a] Martin Culka,[a] Lubomír
Rulíšek,*[a] and Petr Beier*[a]
[8,12–14]
Abstract: Direct fluorination of ortho-, meta- and para-substituted
aromatic thiols and disulfides using elemental fluorine afforded
substituted (pentafluorosulfanyl)benzenes. This work thus represents
the first study of the scope and limitation of direct fluorination for the
synthesis of new SF5-containing building blocks. Fluorinations in
disulfide using AgF2
or XeF2;[15] however, these methods
generally suffer from low product yields and the need to use
expensive reagents. Recently, some improvement has been
achieved using AgF2 and thin films of coinage metals (Cu, Ag,
Au).[12] A practical method is also represented by the Umemoto’s
two-step synthesis of arylsulfur pentafluorides from aromatic
disulfides or thiols (Scheme 1A).[16] This reaction requires large
excess of chlorine gas (which can be substituted by
trichloroisocyanuric acid) and potassium fluoride and displays a
relatively good substrate scope. The intermediates ArSF4Cl are
rather sensitive compounds and the Cl/F exchange reaction
necessitates the use of anhydrous HF, IF5, Ag2CO3 or can be
accomplished by heating with ZnF2 or AgF.[16–22] The second
practical method for the synthesis of arylsulfur pentafluorides is
the direct fluorination of diaryl disulfides.[9,12,23] It has been
reported only for nitro(pentafluorosulfanyl)benzenes (Scheme 1B)
which are produced in this way in multikilogram quantities. Almost
batch and flow modes were compared.
A
comprehensive
computational study has been carried out employing density
functional and wave functions methods to elucidate the reaction
mechanism of the transformation of ArSF3 to ArSF5. Eliminating
various non-radical pathways, it has been shown that the reaction
proceeds by a radical mechanism, initiated by the attack of the F· on
the ArSF3 moiety, propagated via almost barrier-less F2 + ArSF4· →
ArSF5 + F· step and terminated by the ArSF4· + F· → ArSF5. Most of
the calculated data are in a very good agreement with experimental
observations concerning the ortho-substituent effect on the reaction
rates and yields.
all
heteroaromatics
nitro(pentafluorosulfanyl)benzenes
other
commercially
available
SF5-aromatics
derived
and
from
are
Introduction
by
either
nucleophilic
aromatic substitution chemistry or reduction to anilines and
diazotization chemistry.[5,9,24–26]
The pentafluorosulfanyl (SF5) functionality is
a group of
remarkable properties such as high electronegativity,[1] large
dipole moment,[2] significant steric hindrance,[2] good cell
penetration[3] and high lipophilicity.[3,4] Organic compounds
containing this group are widely exploited in agrochemical and
pharmaceutical science as well as in new materials.[5–7] Aromatic
SF5 compounds display extraordinary thermal and chemical
stability.[5,8,9] They tolerate strong Brønsted acids and bases[8] and
conditions for transition metal-catalyzed reactions.[9,10] At the
same time, pentafluorosulfanyl derivatives are somewhat
sensitive towards strong Lewis acids,[11] alkali metals and
organolithiums.[10]
Importantly, the scope of direct fluorination of diaryl
disulfides has not been reported yet. It has been argued that
strongly electron-withdrawing groups on the aromatic ring of the
starting diaryl disulfide are necessary to avoid ring fluorination. In
addition, ortho-nitro-substituted starting disulfide was shown to
fail in the reaction. Unstable arylsulfur trifluoride intermediate was
formed and the reaction did not proceed further.[9]
In this work, we set out to investigate the scope of the direct
fluorination of diaryl disulfides using elemental fluorine affording
arylsulfur pentafluorides in one step (Scheme 1C). In particular,
we were interested in functional group compatibility and
substitution tolerance. Elemental fluorine is an inexpensive
reagent on industrial scale (cost of production is 5-8 $/kg) and
although its use requires special facility and safety precautions,
the direct fluorination could become a viable strategy for the
production of basic SF5-containing aromatic building blocks. The
synthetic efforts are complemented by quantum chemical
calculations to investigate mechanistic features of the reported
reaction and deepen the understanding of fluorination reactions.
Only a few synthetic methods for the preparation of
arylsulfur pentafluorides have been developed to date.
(Pentafluorosul-fanyl)benzene can be prepared from diphenyl
[a]
J. Ajenno, Dr. B. Klepetářová, Dr. D. Bím, Dr. M. Culka, Dr. L.
Rulíšek, Dr. P. Beier
Institute of Organic Chemistry and Biochemistry of the Czech
Academy of Sciences, Flemingovo nám. 2, 166 10 Prague 6, Czech
Republic.
E-mail: rulisek@uochb.cas.cz; beier@uochb.cas.cz
J. Ajenjo
Deparment of Organic Chemistry, Faculty of Science, Charles
University, Hlavova 2030/8, 128 43 Prague 2, Czech Republic.
Dr. Greenhall
[b]
[c]
Results and Discussion
Synthesis. Owing to the differences in technical setup (reactor
size and shape, mode of stirring and introduction of diluted F2 into
the reaction mixture) of each individual F2 facility, it was important
F2 Chemicals Ltd, Lea Lane, Lea Town, Preston, PR4 0RZ, UK.
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
Scheme 3. Synthesis of SF5-nitrobenzene 3b.
Scheme 1. Practical methods for the synthesis of arylsulfur pentafluorides.
(also for the synthesis of SF5-benzoic acid), could be prepared in
one simple synthetic step in synthetically viable yield. Substituting
acetonitrile solvent for anhydrous HF or 98% H2SO4 did not afford
any traces of 3b. Therefore, all other experiments were performed
in acetonitrile with 10% F2/N2 mixture. The scope of direct
fluorination of aromatic thiol and disulfides to afford SF5-aromatics
is shown in Table 1.
to “calibrate” our facility by comparing it to the previously reported
results.[9] To this aim, 1,2-bis(4-nitrophenyl)disulfane (1a)
underwent a reaction with diluted F2 to form SF5-subsituted
nitrobenzene 3a. The limitations of our system are the size of the
reactor (100 mL, translating to maximum 30 mL of the reaction
mixture), F2 concentration (up to 20%), gas mixture flow of
maximum 30 L/h and the reaction time of maximum several hours.
With an F2 concentration of 10%, compound 3a was prepared in
Aromatic disulfides substituted with nitro, cyano, trifluoromethyl,
pentafluorosulfanyl,
chlorosulfonyl,
fluorosulfonyl
and
fluorocarbonyl groups in para-position underwent smooth
fluorination to afford SF5-aromatics 3 in moderate to good yields
(in the context of F2 chemistry). Due to high volatility the yields of
products 3c, 3i, 3j, 3m were determined by 19F NMR
spectroscopy using internal standard. Acid fluoride 3g was
observed in the crude mixture in good NMR yield; however,
isolation attempts led to hydrolysis to benzoic acid 3h. The same
product was not formed by the fluorination of 1h. Presumably, this
is due to sensitivity of arylsulfur trifluoride intermediate towards
hydrolysis. The degree of over-fluorination to side-products 4 was
found to be substrate dependent and also depended on the
stoichiometry of the used fluorine gas. Therefore, we have
routinely monitored progress of the reaction by 19F NMR
spectroscopy (using an NMR tube insert made of PFA) and
stopped the reaction after consumption of arylsulfur trifluoride
intermediate. Typical NMR spectra of the reaction progress are
shown in Figure 1. Arylsulfur trifluoride intermediate is clearly
visible in 19F NMR spectra. Two fluorine atoms appear at around
61 ppm (2JF-F ≈ 70 Hz) and one at around -50 ppm (not visible on
Figure 1). This corresponds to a T-shaped arrangement of the
three fluorine atoms with all three S-F bonds perpendicular to the
S-C1(aryl) “molecular axis”. As reaction proceeds, ArSF3 is
converted to 3, which results in two sets of signals – doublet at 62
ppm (2JF-F ≈ 150 Hz) for four equatorial fluorine atoms and pentet
at 82 ppm for one axial fluorine atom. Simultaneously with the
formation of 3, over-fluorination impurities 4 are also formed.
The suggested mechanism of direct fluorination of 1 or 2 is shown
in Scheme 4. Elemental fluorine is a strong oxidant capable of
oxidizing the aromatic thiol 2 to aromatic disulfide 1 and producing
anhydrous hydrofluoric acid. Interestingly, the formed HF did not
have a detrimental effect on further reaction steps. Next, the ArSF
intermediate is postulated (never observed by NMR), and further
fluorination takes place to yield ArSF3 and finally 3. Arylsulfur
trifluoride intermediate formed much faster than the follow-up
fluorination to 3. In the presence of protic impurities
45% isolated yield on 1.6 mmol scale (Scheme 2). GCMS (or 19
F
NMR) analysis of the crude reaction mixture revealed that a small
amount of ring fluorinated impurities 4a was formed (3a:4a ratio
of 98:2); however, they were successfully separated from 3a by
column chromatography. Although the yield of 3a was higher,
limitations of our system in terms of minimum reaction volume and
maximum reaction time has led to the use the lower substrate
concentration. A higher excess of F2 had to be used probably due
to inefficient mass transfer of F2 in our system. Nevertheless, this
study focuses on the scope of direct fluorination for the
preparation of SF5 aromatics and further engineering issues
associated with scale-up were not investigated in detail.
Scheme 2. Synthesis of SF5-nitrobenzene 3a.
In the next experiment, 4,4’-disulfanediyldibenzonitrile (1b)
was fluorinated with F2/N2 mixture of varying concentrations,
ranging from 2.5% to 10% (Scheme 3). Although the yield of 3b
and 3b:4b ratio slightly decreased with increasing F2
concentration, we concluded based on reaction time differences
that the most practical dilution of F2 for the preparative scale direct
fluorination was 10%. 3b was isolated from the product mixture in
pure form by chromatography. This experiment also
demonstrated that benzonitrile 3b, a highly useful building block
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
Table 1. Substrate scope of direct fluorination of aromatic thiols and disulfides to SF5-aromatics.[a]
[a] Reaction conditions: 1 or 2 (1-3 mmol), 10% F2 in nitrogen (30 L/h, excess), dry MeCN (25-30 mL), –5 °C, 1-4.5 h. Yields of pure products 3 are shown. Crude 3:4
product ratio is shown in parentheses. [b] 19F NMR yield using 3a as an internal standard.
(such as wet solvent) the arylsulfur trifluoride intermediate
undergoes hydrolysis, followed by fluorination and hydrolysis to
arylsulfonyl fluoride side product.
Halogen-substituted as well as methoxycarbonyl substrates
afforded low product yields. Ortho-substituted substrates showed
an interesting behavior. Fluorination of 1o stopped at the ArSF3
intermediate and no SF5 product was formed. However, this does
not mean that the fluorination of ortho-substituted aromatic
disulfides is not possible. Compound 1p underwent fluorination to
o-(pentafluorosulfanyl)benzonitrile 3p in low yield with significant
over-fluorination. Attachment of electron-acceptor nitro group
suppressed these side reactions and compounds 3q and 3r were
produced in high yields and excellent selectivities from the
corresponding aromatic thiols. Products 3q, 3r and 3t were
synthesized from the corresponding thiols 2. Here, the first step is
presumably the oxidation with F2 to afford disulfides 1 and HF.
The formed equimolar amount of HF does not seem to have a
detrimental effect on the fluorination reaction. 19F NMR of
Figure 1. 19F NMR monitoring of the reaction mixture of fluorination of 3b.
compound 3q showed two sets of signals, one doublet (2JFF
=
149.6 Hz) corresponding to the four equatorial fluorine atoms of
SF5 and one pentet corresponding to one axial fluorine (Figure 2).
In compound 3r containing the CF3 group in ortho-position to SF5
an additional 5JFF coupling was visible between fluorine atoms of
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
CF3 and equatorial fluorine atoms of SF5 (5JFF = 18.5 Hz) (Figure
3). 2-Fluoro-4-nitro-1-(pentafluorosulfanyl)benzene (3y) and 2,6-
optimizations were carried out using the hybrid PBE0 functional[31]
with empirical dispersion correction (D3-BJ),[32] the def2-SVP
basis set and COSMO implicit solvation model[33] with r = 37.8,
corresponding to acetonitrile; all methods as implemented in
TURBOMOLE 7.2 program package.[34] Single-point energies
were computed at the PBE0-D3/def2-TZVP//COSMO-RS[35]
(solvent = acetonitrile) level of calculations. Several pathways
were also investigated by employing correlated ab initio (WFT)
difluoro-4-nitro-1-(pentafluorosulfanyl)benzene
(3z)
were
synthetized by Kirsch and co-workers by direct fluorination of the
corresponding disulfides in 9% and 4.6% yields, respectively.[23]
Additionally, perfluoro-diphenyl disulfide was not
a viable
substrate as decomposition to unidentified products occurred and
3u was formed only in trace amounts. Ortho-fluoro substitution
seems to be problematic and we speculated that the
decomposition pathway might involve
substrates for the
formation of bis and tris(pentafluorosulfanyl)benzenes. Although
the yields of 3v, 3n, 3w and 3d from the corresponding
dimercaptobenzenes were moderate to low, two SF5 groups were
formed in one step, which makes the direct fluorination a viable
route due to straightforward synthesis from readily available and
inexpensive starting compounds. We presume that the formation
of cyclic (or linear in the case of 2d) oligomeric aromatic disulfides
took place initially, followed by fluorination with F2. SF3 is a strong
electron-acceptor group; therefore, once the bis-SF3 aromatic
was formed, further fluorination took place. Aryl bromide 3w has
been
widely
used
for
the
introduction
of
3,5-
bis(pentafluorosulfanyl)phenyl moiety in the synthesis of Pd
complexes,[27] in medicinal chemistry,[28] in dye synthesis,[29] and
in catalyst design.[30]
Translating the fluorination from batch setup to a flow
process might lead to increased yields and better selectivity.
Toward this goal, fluorination of 1b in flow was investigated. Initial
experiments using a solution of 1b in acetonitrile pumped through
a PTFE tubing using a syringe pump revealed that this setup does
not deliver enough fluorine for the formation of arylsulfur trifluoride
intermediate. Instead, Ar-SF3 was formed in batch by introducing
10% F2/N2 mixture to the solution of 1b and then the resulting
mixture was pumped using a syringe pump through a PTFE tubing
and 20% F2 was introduced via a connection made of passivated
stainless steel T-joint resulting in a liquid-gas plug flow. The
product 3b was formed in significantly higher yield and better
3b:4b selectivity than in batch-only process (Figure 4). For scale-
up of the synthesis of arylsulfur pentafluorides the hybrid batch-
flow process is a promising method to increase yields and
selectivity in direct fluorination.
Figure 2. 19F NMR (377 MHz, CDCl3) of 3q.
Figure 3. 19F NMR (377 MHz, CDCl3) of 3r.
Scheme 4. Synthesis of SF5-nitrobenzene 3a.
Computational Investigation of the Reaction Mechanism of
the ArSF3 → ArSF5 Conversion. To probe the mechanism of the
conversion of arylsulfur trifluorides (5) to arylsulfur pentafluorides
(3) (the slowest step) and to explain the lack of a reactivity of
ortho-nitro substrates, a computational study was performed
employing both density functional theory (DFT) and wave function
methods (WFT, wave function theory). At the DFT level, geometry
Figure 4. A hybrid bath-flow process for improved synthesis of 3b.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
methods: SCS-MP2/aug-cc-pVDZ//COSMO and DLPNO-
CCSD(T)/aug-cc-pVDZ//COSMO.[36] For the latter, we used
ORCA 4.0 program.[37] Both WFT methods (SCS-MP2 and
DLPNO-CCSD(T)) were used for resolving non-negligible
discrepancies between the pure and hybrid functionals found
during initial tests of methodology. Considering the DLPNO-
CCSD(T) as the benchmark, it has been observed that many
hybrid functionals (e.g., PBE0 and B3-LYP) as well as the SCS-
MP2 method[38] yield quite consistent and accurate results,
especially when it comes to relative
Scheme 5. Studied reaction pathways for the conversion of arylsulfur trifluorides (5) to arylsulfur pentafluorides (3).
Table 2. The computed energetics along the reaction coordinates for studied reaction mechanisms (in kcal.mol–1).[a]
5 (X)
TS1
–
5 + F2
0.0
TS2
9.5
9.0
9.8
8.7
8.7
A
TS3
n.a.
3
5 + F2
0.0
TS4
19.0
14.7
17.4
20.1
10.2
B
TS5
-7.3
-10.2
n.a.
3
5 + 1/2F2
0.0
C
TS6
9.1
8.4
8.6
8.2
8.8
3
5o (NO2)
5y (F)
5.1
3.7
4.0
4.3
2.9
-83.4
-84.6
-84.2
-83.3
-84.6
-37.8
-42.0
n.a.
-83.4
-84.6
-84.2
-83.3
-84.6
-12.9
-21.0
-21.0
-14.9
-22.3
-80.0
-87.7
-87.7
-82.1
-88.9
–
0.0
30.2
n.a.
0.0
0.0
5q (CN)
5r (CF3)
5a (H)
–
0.0
0.0
0.0
–
0.0
37.8
28.9
0.0
-30.0
-33.5
-0.8
-4.9
0.0
>50
0.0
0.0
0.0
[a] PBE0-D3/def2-TZVPD//COSMO-RS values, in italic SCS-MP2/aug-cc-pVDZ//COSMO values. n.a. Calculation diverged into different reaction channels
(see text).
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
energies between the reactants (vide infra), which further
strengthens the conclusions. Gas-phase thermodynamics was
taken into account by employing standard formulas of statistical
mechanics to obtain the Gibbs free (reaction and activation)
energies in solvent, after addition of Gsolv(COSMO-RS) term. In
the following text, we decided to use mostly SCS-MP2/aug-
ccpVDZ//COSMO values for argumentation concerning non-
radical mechanisms since we assume this WFT method might be
slightly more reliable for many “non-typical” bonding
arrangements and transition states than DFT functionals. On the
contrary, we prefer hybrid DFT methods for the mechanistically
more simple radical mechanism to avoid potential difficulties with
the unrestricted formulation of the SCS-MP2 method. Wherever
possible, the comparison has been made and the results were
found quite robust (primary energetic data as well as all
equilibrium structures are deposited in the SI).
Four plausible pathways for five model reactants with
varying reactivities (X = H, F, CN, CF3, NO2 in the ortho-position)
were considered (Scheme 5): (a) oxidative addition of F2 molecule
to 5 (side-on attack of F2), (b) polar or addition of F2 on the sulfur
of 5, (c) nonpolar addition involving fluorination of ortho-carbon in
the intermediate step and (d) radical addition mechanism. Correct
prediction of the experimentally observed reactivity represents a
stringent test of a plausible reaction mechanism.
Concerning the structure of the reactants 5, it has been
clearly shown by the calculations that energy minima (equilibrium
structures) always correspond to the experimentally observed
arrangement of F atoms in the SF3 moiety (cf. Scheme 5), that is
all three F atoms occupy equatorial positions with respect to the
main molecular S-C1 axis. The structures with a fluorine atom in
axial position with respect to the axis (i.e., trans to aryl moiety) are
often not even minima on the potential energy surface.
The computed data are summarized in Table 2. It can be
seen that concerning the overall thermodynamics, the
calculations have convincingly shown that the reaction of 5 to 3
(more precisely, 5 + F2 → 3) in acetonitrile is highly exergonic
(G0 = –80 to –90 kcal.mol–1) for all five studied model reactants
(Table 2). Moreover, we may already qualitatively perceive origins
of the different reactivity when taking into account PBE-D3/def2-
TZVP//COSMO-RS free energy values (right part of Table 2),
though these trends are only qualitatively reproduced by the SCS-
MP2 calculations. The highest (least exergonic) values
correspond to 5o and 5r, that is for systems with sluggish
reactivity observed experimentally. The most favorable overall
thermodynamics was computed for 5a.
turned out to be energetically unfeasible. Even at the distance of
R(S–FA) = R(S–FB) = 1.9 Å (close to the S–F equilibrium distance
of 1.64 Å in the ArSF5) the F–F bond does not break and the
energy rises well above 50 kcal.mol–1 (c.f. Scheme 5).
The second alternative of a non-radical mechanism is the
polar addition of the F2 molecule on the sulfur atoms of the SF3
group. This seems to be more feasible than the side-on attack and
it is associated with the initial barrier (activation energy) of 8.7 to
9.8 kcal.mol–1, computed at the SCS-MP2/aug-cc-pVDZ//COSMO
level. This leads to a stationary point at the potential energy
surface – corresponding to the ArSF4+ + F– ion pair (intermediate
A), which is 2.9–5.1 kcal.mol–1 less stable than the two reactants,
ArSF3 + F2. However, the second step of the reaction, an attack
of the fluoride onto the SF4+ group turned out to be energetically
unfeasible with computed energy barriers of 28.9–37.8 kcal.mol–1
(Table 2) and additional entropic terms of at least 10 kcal.mol–1
which are typically associated with {A… B} → AB reactions.
Moreover, no correlation with the experimentally observed order
of reactivity has been found and in two cases (X = NO2, CN) this
reaction channel lead to fluorination of the aromatic ring – a
thermodynamic trap (vide infra). We assume that these rather
+
large energy barriers are reflecting the steric demands of the SF4
group (all four F atoms occupy equatorial positions with respect
to the main Ar-C1-S axis of the molecule) and no pathway has
been found for the F– atom to get close to the axial position
(including the unlikely attack of the second F2 molecule at the Ar
SF4+ + F– ion pair which turned out to be unfeasible as well).
As the last alternative for the non-radical mechanism
pointed to by the calculations while examining all possible
pathways we considered the end-on attack of F2 on the ortho-
carbon (nonpolar addition). While admitting that it is rather
counter-intuitive and unusual, computations predict that the
addition of fluorine to the ortho-position and concomitant
attachment of the second fluorine to the SF3 group, leading to a
dearomatization of the benzene ring (cf. Scheme 5) is associated
with activation energies of 10.2–17.8 kcal.mol–1. The
intermediates B are energetically very stable, –35.3 to –42.8
kcal.mol–1 with respect to the ArSF3 + F2 state. The final step on
this reaction pathway is the attack of the fluorine from the ortho-
carbon on the sulfur, through the TS5 (Scheme 5). The energy
difference between the ArSF3 + F2 and TS5 is 0.4 to –10.2
kcal.mol–1, which translates into ~30 kcal.mol–1 barriers with
respect to the intermediates B. Again, in two cases the reaction
channel does not lead to the desired (and experimentally
observed) products. At this point, we can mention that one of the
thermodynamic traps on a non-radical pathway is dissociation
(rearrangement) of the ArSF5 into ArF and SF4. This has been
shown by all computational methods employed in this study to be
thermodynamically favored by 10–15 kcal.mol–1. In analogy with
polar addition pathway, no correlation between the computed
activation energies along the non-polar reaction coordinate and
Next, we may safely exclude the oxidative addition as a
plausible reaction pathway. A side-on attack of the F2 moiety onto
the two vacant sites around sulfur in the SF3 group,
computationally implemented as the relaxed scan of the two S–
FA and S–FB distances (FA, and FB are two atoms in the incoming
F2 molecule) with the boundary condition R(S–FA) = R(S–FB),
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
the experimentally observed reactivity has been found which is,
in our opinion, a strong argument to rule out this reaction pathway.
After an exhaustive computational investigation of the non-
radical pathways with no consistent reaction mechanism found,
we focused on radical mechanism(s). The simplest pathway
assumes homolytic cleavage of F2 into two F· atoms. The
experimental value of the F2 bond dissociation enthalpy in
vacuum is 36.9 kcal.mol–1 [39] at 298.15 K. Corresponding Gibbs
free energy in vacuum is 29.6 kcal.mol–1. We assume that the
value in acetonitrile would be similar since the computed
COSMO-RS solvation energy for the F2 molecule in acetonitrile is
+1.3 kcal.mol–1, whereas the F· solvation free energy has been
computed as +0.6 kcal.mol–1, thus pointing to the overall change
in solvation free energy for the F2 → 2 F· reaction close to 0
kcal.mol–1. F2 dissociation is then presumably the highest barrier
along the radical reaction pathway. The F· radical then attacks
the SF3 group to form radical C (Scheme 5) in a barrier-less
process. We postulate that it is the stability of C, calculated as Ar
SF3 + 1/2 F2 → Ar-SF4· (free) energy change, that determines the
overall kinetics of the reaction. The experimentally observed order
of reactivity (inferred from the yields and selectivities listed in
Table 2) is nicely reproduced for 5a (X=H) > 5q (X= CN) > 5o (X
= NO2). For the 5r and 5y (5y has not studied in this work, but
reported in the literature),[23] the agreement is worse. Calculations
predict that 5r should be less reactive than 5q and on the contrary,
5y almost as reactive as 5q. However, this does not seem to be
the case in experiments. At the moment, we do not have an
explanation for the computed discrepancy for 5y and 5r. The
propagation step (TS6) is then associated with negligible
activation energy of 0.1–0.6 kcal.mol–1 with respect to the radical
intermediate C, which translates into 8.2–9.1 kcal.mol–1 in
activation Gibbs free energy due to the entropic terms. No clear
correlation between the computed barriers and the experimentally
observed reactivity is seen for this step and we do not consider it
as the rate determining. The termination is expectedly barrier-less.
In analogy with the structure of the closed-shell 5 and of the
blocks, such as SF5-substituted benzonitrile or benzoic acid,
ortho-substituted aromatics with rather bulky groups (previously
considered to be unreactive), and bis(pentafluorosulfa-
nyl)benzenes were prepared using elemental fluorine in one step.
The hybrid batch-flow process was developed and in terms of
isolated yield and selectivity it compared favorably to the
traditional batch process. A detailed computational study using
various density functional and wave function methods revealed
the reaction mechanism of the final and decisive step of the
studied fluorinations: conversion of the ArSF3 into ArSF5.
Excluding three non-radical pathways we postulate that the
reaction proceeds via the radical mechanism after homolytic
cleavage of the F-F. The remaining (propagation, termination)
steps are (almost) barrier-less and the overall kinetics is,
according to the computed data, determined by the stability of the
ArSF4· species. Experimentally observed order of reactivities has
been reproduced for three compounds studied computationally
while questions arise concerning the systems with the fluor-
containing X group in ortho- position. It must be emphasized that
we see a lot of room for further computational modeling of other
potential reaction channels that might be accompanied by the
elaborate experimental identification of all side-products, and may
include kinetic modeling. However, such analysis is beyond the
scope of this study.
Acknowledgements
This work was supported by the Czech Academy of Sciences
(Research Plan RVO: 61388963), by the Initial Training Network,
FLUOR21, funded by the FP7 Marie Curie Actions of the
European Commission (FP7-PEOPLE-2013-ITN-607787) and by
the Czech Science Foundation (18-00215J). We would like to
thank Professor Josef Michl (IOCB Prague) and Professor
Graham Sandford (Durham University, UK) for allowing J.A. to
use their F2 facility.
+
ArSF4 intermediates, in the equilibrium structures of radicals C
(molecular geometries deposited in the SI), fluorine atoms prefer
four equatorial positions with respect to the main molecular axis.
This leads, in the case of 5o (and slightly in the case of 5r) to a
steric clash, which is resolved, in the case of 5o – by
deplanarization of the NO2 group which comes at the cost of ~5 8
kcal.mol–1. This is, in qualitative terms, an explanation of the
sluggish reactivity observed for the 5o experimentally. We
postulate, while admitting the absence of a tangible computational
(or experimental) proof, that the reaction of 5y (and perhaps also
5r) may not follow the above-postulated reaction mechanism and
the origin of the sluggish reactivity towards the ArSF5 products is
the different reaction channel that leads to other products.
Keywords: pentafluorosulfanyl • sulfur pentafluoride • radical
fluorination • elemental fluorine • mechanism
[1]
L. J. Sæthre, N. Berrah, J. D. Bozek, K. J. Børve, T. X. Carroll, E. Kukk,
G. L. Gard, R. Winter, T. D. Thomas, J. Am. Chem. Soc. 2001, 123,
10729–10737.
[2]
[3]
W. A. Sheppard, J. Am. Chem. Soc. 1962, 84, 3072–3076.
M. V. Westphal, B. T. Wolfstädter, J.-M. Plancher, J. Gatfield, E. M.
Carreira, ChemMedChem 2015, 10, 461–469.
[4]
C. Hansch, R. M. Muir, T. Fujita, P. P. Maloney, F. Geiger, M. Streich, J.
Am. Chem. Soc. 1963, 85, 2817–2824.
[5]
[6]
P. R. Savoie, J. T. Welch, Chem. Rev. 2015, 115, 1130–1190.
P. J. Crowley, G. Mitchell, R. Salmon, P. A. Worthington, Chimia 2004,
58, 138–142.
[7]
[8]
[9]
S. Altomonte, M. Zanda, J. Fluorine Chem. 2012, 143, 57–93.
W. A. Sheppard, J. Am. Chem. Soc. 1962, 84, 3064–3072.
R. D. Bowden, P. J. Comina, M. P. Greenhall, B. M. Kariuki, A. Loveday,
D. Philp, Tetrahedron 2000, 56, 3399–3408.
Conclusions
In conclusions, the scope of the direct fluorination of aromatic
disulfides and thiols for the preparation of SF5-benzenes was
investigated.
[10] P. Kirsch, M. Bremer, M. Heckmeier, K. Tarumi, Angew. Chem., Int. Ed.
1999, 38, 1989–1992.
Over
twenty
new
examples
of
[11] J. Wessel, G. Kleemann, K. Seppelt, Chem. Ber. 1983, 116, 2399–2407.
(pentafluorosulfanyl)benzenes ranging from simple building
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
[12] A. M. Sipyagin, C. P. Bateman, A. V. Matsev, A. Waterfeld, R. E. Jilek,
C. D. Key, G. J. Szulczewski, J. S. Thrasher, J. Fluorine Chem. 2014,
167, 203–210.
[26] J. Ajenjo, M. Greenhall, C. Zarantonello, P. Beier, Beilstein J. Org. Chem.
2016, 12, 192–197.
[27] C. Berg, T. Braun, R. Laubenstein, B. Braun, Chem. Commun. 2016, 52,
3931–3934.
[13] A. M. Sipyagin, C. P. Bateman, Y.-T. Tan, J. S. Thrasher, J. Fluorine
Chem. 2001, 112, 287–295.
[28] Y.-D. Yang, E. Tokunaga, H. Akiyama, N. Saito, N. Shibata,
ChemMedChem 2014, 9, 913–917.
[14] A. M. Sipyagin, V. S. Enshov, S. A. Kashtanov, C. P. Bateman, B. D.
Mullen, Y.-T. Tan, J. S. Thrasher, J. Fluorine Chem. 2004, 125, 1305–
1316.
[29] N. Iida, E. Tokunaga, N. Saito, N. Shibata, J. Fluorine Chem. 2014, 168,
93–98.
[15] X. Ou, A. F. Janzen, J. Fluorine Chem. 2000, 101, 279–283.
[16] T. Umemoto, L. M. Garrick, N. Saito, Beilstein J. Org. Chem. 2012, 8,
461–471.
[30] Y.-D. Yang, X. Lu, E. Tokunaga, N. Shibata, J. Fluorine Chem. 2012, 143,
204–209.
[31] C. Adamo, V. Barone, J. Chem. Phys. 1999, 110, 6158–6170.
[32] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104.
[17] O. S. Kanishchev, W. R. Dolbier, Angew. Chem. Int. Ed. 2015, 54, 280–
284.
[18] M. Kosobokov, B. Cui, A. Balia, K. Matsuzaki, E. Tokunaga, N. Saito, N.
Shibata, Angew. Chem., Int. Ed. 2016, 55, 10781–10785.
[19] B. Cui, M. Kosobokov, K. Matsuzaki, E. Tokunaga, N. Shibata, Chem.
Commun. 2017, 53, 5997–6000.
[33] A. Klamt, G. Schüürmann, J. Chem. Soc., Perkin Trans. 2, 1993, 799–
805.
[34] TURBOMOLE V7.2 2017, a development of University of Karlsruhe and
Forschungszentrum Karlsruhe GmbH, 1989–2007, TURBOMOLE
GmbH, since 2007, available from http://www.turbomole.com.
[35] A. Klamt, J. Phys. Chem. 1995, 99, 2224–2235.
[36] M. Sparta, F. Neese, Chem. Soc. Rev. 2014, 43, 5032–5041.
[37] F. Neese, Wiley Interdiscip. Rev.-Comput. Mol. Sci. 2012, 2, 73–78.
[38] Grimme S. J. Chem. Phys. 2003, 118, 9095–9102.
[39] J. J. DeCorpo, R P. Steiger, J. L. Franklin, J. L. Margrave, J. Chem. Phys.
1970, 53, 936.
[20] B. Cui, S. Jia, E. Tokunaga, N. Saito, N. Shibata, Chem. Commun. 2017,
53, 12738–12741.
[21] C. R. Pitts, D. Bornemann, P. Liebing, N. Santschi, A. Togni, Angew.
Chem., Int. Ed. 2019, 58, 1950–1954.
[22] K. Lummer, M. V. Ponomarenko, G. V. Röschenthaler, M. Bremer, P.
Beier, J. Fluorine Chem. 2014, 157, 79–83.
[23] P. Kirsch, A. Hahn, Eur. J. Org. Chem. 2005, 2005, 3095–3100.
[24] O. S. Kanishchev, W. R. Dolbier, Adv. Het. Chem. 2016, 120, 1–42.
[25] P. Das, E. Tokunaga, N. Shibata, Tetrahedron Lett. 2017, 58, 4803–4815.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902651
Chemistry - A European Journal
FULL PAPER
Layout 2:
FULL PAPER
Javier Ajenjo, Blanka Klepetářová,
Martin Greenhall, Daniel Bím, Martin
Culka, Lubomír Rulíšek*, Petr Beier*
Page No. – Page No.
Preparation of (Pentafluorosulfanyl)-
benzenes by Direct Fluorination of
Diaryldisulfides: Synthetic Approach
and Mechanistic Aspects
Direct fluorination of substituted aromatic disulfides or thiols provided arylsulfur
pentafluorides. The mechanism of the second step was investigated
computationally.
This article is protected by copyright. All rights reserved.