Key words
fluorine - phosphorus - sulfur - pentafluorosulfanylation - isosteres
Nearly a quarter of the pharmaceuticals in the market contain at least one fluorine
atom in their structure.[1] The impact of fluorine in drug discovery campaigns has been remarkable and methods
to create X-F bonds in organic molecules are highly coveted.[2] In this context, chemists have identified groups of fluorinated functionalities
which have had a dramatic impact on the ADME (i.e. Absorption, Distribution, Metabolism
and Excretion) properties of certain biologically active compounds.[3] For example, CF3,[4] OCF3,[5] SCF3,[6] CF2H,[7] CFH2,[8] and trifluorocyclopropyl[9] have all been studied as bioisosteres of CH3, OCH3, and
t
Bu groups. In recent years, a related fluorinated functionality has also been identified
as a bioisostere of the CF3 and
t
Bu groups: the pentafluorosulfanyl group (SF5).[10] This hypervalent sulfur moiety is characterized by an octahedral arrangement of
the F atoms around the S(VI) atom, resulting in high electronegativity[11] (Scheme [1]A). Compared to its CF3 analog, the SF5 group is more hydrophobic and is robust when confronted to harsh acidic or basic
conditions (for Ph-SF5).[12]
Scheme 1 (A) Physical and chemical properties of Ar-SF5; (B) Current strategies to access Ar-SF4Cl; (C) Use of aryl phosphorothiolates in oxidative fluorination
With the volume comparable to a
t
Bu group and the electronegativity resembling a NO2 group, the introduction of SF5 into lead compounds has captivated the interest of medicinal chemists. For example,
analogues of mefloquine (antimalarial) or bosentan (pulmonary arterial hypertension)
bearing an SF5 group have shown to be more potent than its CF3 analog, highlighting some of the potential applications of the pentafluorosulfanyl
group (Scheme [1]A).[13] Despite the interesting chemical properties of this group, its synthesis and strategies
for its straightforward introduction are still somewhat limited. Although early examples
using Cl2, F2, or XeF2 are known, limitations in functional group tolerance, scope, and practicality have
prevented their adoption by organic chemists.[14] In groundbreaking work, Pitts, Santschi, Togni, and co-workers reported a practical
variation of the Umemoto process,[14d] which enabled the synthesis of a wide variety of Ar-SF5 compounds in a simple and straightforward manner.[15] The strategy consists of the oxidation of aryl disulfides (from thiols) with TCICA
(tetrachloroisocyanuric acid) in the presence of an excess of KF, to forge the key
intermediate Ar-SF4Cl (Scheme [1]B). Simple Cl-F exchange then leads to Ar-SF5. It is important to mention that such a strategy has also been used in the oxidation
of other chalcogens and even organophosphorus compounds.[16] Our group has recently contributed to this area reporting the possibility to use
aryl sulfenylphthalimides as precursors, which can be obtained from the parent Ar-ZnX
compounds (Scheme [1]B).[17] Despite these advances, the current methodologies are still restricted to thiophenols
and organozinc reagents as precursors. With the aim of expanding the palette of opportunities
in terms of a wider spectrum of precursors, we focused our attention on the oxidative
fluorination of aryl phosphorothiolates, en route to valuable Ar-SF4Cl (Scheme [1]C).
Indeed, aryl phosphorothiolates can be accessed through a variety of different starting
materials and their synthesis has been widely explored. For example, Ar-S-P(O)(OR)2 can be easily accessed in one step from thiophenols by the simple reaction with H-P(O)(OR)2 or Cl-P(O)(OR)2.[18] Schoenebeck has recently shown that Ar-S-P(O)(OR)2 can also be easily accessed via palladium catalysis from the parent aryl iodides.[19] Additionally, Gooβen and Tang have developed protocols which enable the synthesis
of aryl phosphorothiolates through SEAr from electron-rich arenes or via Cu-catalyzed/mediated cross-coupling from the
corresponding diazonium salt or boronic acid.[20] The possibility to access these compounds from a myriad of diverse starting materials
further supports the consideration of Ar-S-P(O)(OR)2 as convergent linchpin reagents. In this work, we demonstrate that oxidation of Ar-S-P(O)(OR)2 with TCICA in the presence of KF delivers Ar-SF4Cl in good yields with a variety of substitution patterns at the aryl moiety (Scheme
[1]C). Additionally, we also demonstrate that Ar-SF4Cl can be converted into the corresponding Ar-SF5 via the use of AgBF4.
Optimization of the oxidative fluorination started by the testing of various phenyl
phosphorothiolates. As shown in Scheme [2], when a mixture of TCICA and KF is used in MeCN at room temperature, phenyl phosphorothiolates
bearing OMe (1) or OEt (2) led to good yields of 6a (entries 1 and 2). However, side products were also observed, which were identified
to be mainly the S(IV) product Ph-SF3 and the partially hydrolyzed S(VI) product Ph-SOF3 (shown together as 7). When the alkoxy groups in the phosphorus ester are replaced by phenoxy (3), the yields dramatically decreased, resulting in only 25% overall yield with almost
no selectivity for Ar-SF4Cl (entry 3). When the P(O)(OR)2 group is replaced by P(O)Ph2 (4), good yields were also obtained in high selectivity (entry 4). Finally, the replacement
of P=O by P=S reduced the yield of the desired product 6a, presumably through the undesired oxidation of the terminal sulfide group (entry
5). Although slightly better yields were obtained for compound 4, we selected compound 2 as our phosphorothiolate of choice, because of the wider availability of methods
to access this particular moiety. Although several protocols can lead to Ar-S-P(O)(OEt)2, we utilized the methods reported by Gooβen and Zhao to access our starting materials
2b–m.[20]
Scheme 2 Optimization of the reaction conditions. a Combined integration for Ph-SF3 and Ph-SOF3. b Yields determined by 19F NMR spectroscopy with α,α,α-trifluorotoluene as internal standard.
With this optimization in hand, we scrutinized the scope of this transformation. As
depicted in Scheme [3], the method functioned well in the presence of halogens. For example, aryl groups
substituted with p-Br (6b), p-Cl (6c), p-F (6d), or multiple halogens (6e) afforded good yields of the corresponding Ar-SF4Cl 6. Product 6f with an alkyl groups in the meta position of the ring was smoothly obtained in 51% yield. Aromatic substituents were
also tolerated, as exemplified by 6g (52%). The presence of other electron-withdrawing substituents such as CF3 both in the meta (6h) and para (6i) positions did not present any hurdles in the oxidative fluorination. When CN or
NO2 groups are attached to the aryl ring, good yields of the desired Ar-SF4Cl products 6j and 6k were also obtained. Interestingly, the presence of an ester did not pose any hurdle,
affording 6l in excellent yield. Unfortunately, the methodology met its limitations when alkyl
aryl ketones are present, affording only traces of 6m. This is probably due to side reactions on the α-carbon. As noticed in previous works,
Ar-SF4Cl compounds are highly reactive and their isolation is extremely challenging. Therefore,
the yields were calculated by 19F NMR spectroscopy by using an internal standard.
Scheme 3 Scope of the reaction for aryl tetrafluorosulfanyl chlorides. Reagents and conditions: Ar-S-P(O)(OEt)2 (0.2 mmol, 1.0 equiv), trichloroisocyanuric acid (TCICA, 18 equiv), rigorously dried
KF (32 equiv), 0.1 M TFA in MeCN (0.2 mL), MeCN (2 mL), in a PTFE vessel, 25 °C, under
argon, 24 h; yields calculated by 19F NMR using α,α,α-trifluorotoluene as reference. a Reaction performed on 1.0 mmol scale.
Having shown the viability of aryl phosphorothiolates as precursors to access Ar-SF4Cl, we decided to explore whether the protocol was suited for the formation of Ar-SF5. It is important to mention that the byproducts formed during oxidative fluorination
are P(V) fluoro compounds, which could potentially affect the Cl-F exchange and generate
side reactions through the additional fluorides required.[21] With these potential issues in mind, we developed a proof-of-concept protocol for
the Cl-F exchange using BF4 anions. To exemplify this possibility, compound 6b was mixed with AgBF4 (2.0 equiv) in DCM at 100 °C; 8 was smoothly formed in 49% yield from 2b, which is obtained from (4-bromophenyl)boronic acid (Scheme [4]). Although this process resembles the silver-induced self-immolative protocol using
Ag2CO3, more information is required to provide a full mechanistic rationale by which this
last Cl-F exchange occurs.[22]
Scheme 4 Cl-F exchange protocol based on AgBF4 to access Ar-SF5
In conclusion, we have developed an oxidative fluorination protocol that converts
Ar-S-P(O)(OR)2 to Ar-SF4Cl in a practical manner. This new protocol broadens the palette of starting materials
to access Ar-SF5 compounds, whose practical synthesis is still highly coveted by practitioners in
medicinal and agrochemical sciences. Although these results are still far from ideal,
we believe that they provide a step forward in the field and will be an incentive
for the development of even more practical methods, finally leading to routine investigations
of Ar-SF5 compounds in drug discovery campaigns.
Unless stated otherwise, all manipulations were performed using standard Schlenk techniques
under anhydrous argon in flame-dried glassware. Anhydrous solvents were distilled
from appropriate drying agents and were transferred under argon: MeCN (CaH2), DCM (CaH2), hexane (Na/K), Et3N (MS). Unless stated otherwise, all chemicals were purchased from Sigma-Aldrich,
Alpha Aesar, and TCI and used without prior drying or purification. Cu(OTf)2 (34946-82-2) and Cs2CO3 (534-17-8) were purchased from Sigma-Aldrich and stored under argon. Trichloroisocyanuric
acid (TCICA, powder, 87-90-1) was purchased from Alpha Aesar and transferred in an
argon-filled glovebox before usage. Potassium fluoride (KF, powder, 7789-23-3) was
purchased from Sigma-Aldrich, rigorously dried under high vacuum (10–6 mbar) at 120 °C for 24 h, and stored under argon. Flash chromatography was performed
on Merck silica gel 60 (40–63 μm). GC-MS (FID) was carried out on a GC-MS-QP2010 equipped
instrument (Shimadzu Europe Analytical Instruments). NMR spectra were recorded using
a Bruker Avance VIII-300 spectrometer. 1H NMR spectra (300.13 MHz) were referenced to the residual protons of the deuterated
solvent used. All 19F NMR spectra were acquired on a 300 MHz spectrometer. For 19F NMR yield determination, α,α,α-trifluorotoluene was used as internal standard (19F, δ = –63.10 in CD3CN). 13C{1H} NMR spectra (75.47 MHz) were referenced internally to the D-coupled 13C resonances of the NMR solvent. The 12 mL PTFE vials were purchased from AHF Analysentechnik
in Tübingen, Germany.
Ar-SF4Cl 6; General Procedure
In a glovebox under argon, the appropriate Ar-S-P(O)(OEt)2 precursor 2 (0.2 mmol, 1 equiv), TCICA (840 mg, 3.6 mmol, 18 equiv), and rigorously dried KF
(372 mg, 6.4 mmol, 32 equiv) were added to an oven-dried 12 mL PTFE reaction vial
equipped with a stir bar. Under vigorous stirring, anhydrous and degassed MeCN (2.0
mL) was added to the mixture, followed by the addition of a 0.1 M solution of TFA
in MeCN (0.2 mL). Then the vial was sealed with a septum-pad cap and the reaction
was stirred at rt in the glovebox for 24 h. After this time, the atmosphere in the
vial was vented carefully and the internal standard α,α,α-trifluorotoluene was added
into the mixture. After 10 min of stirring, an aliquot of the resulting solution was
filtered under argon. The NMR sample was prepared with the filtered aliquot (0.4 mL)
and CD3CN (0.1 mL) for 19F NMR yield determination. Please note that, although Ar-SF4Cl is not too sensitive to moisture, the use of dry reaction vials and anhydrous solvent,
as well as carrying out the experiment and workup under argon benefited the reaction.
See Supporting Information for NMR analysis.
4-BrC6H4-SF5 (8)
4-BrC6H4-SF4Cl (6b) was synthesized according the general procedure described above. Upon the completion
of the reaction, the atmosphere in the vial was vented carefully and the suspension
was transferred to a flame-dried Schlenk tube under argon. Then the solvent and other
volatile constituents were evaporated carefully under vacuum at 0 °C. To the residue,
an anhydrous and degassed mixture of hexane/DCM (9:1) was added to extract the Ar-SOF3 compound (3 × 4 mL). The resulting solution was filtered in two batches of ca. 6
mL under argon followed by concentration of the filtrate under vacuum. The crude product
of Ar-SF4Cl was used immediately for the next step. (Please note, some kinds of Ar-SF4Cl are very volatile. Even at low temperature, the concentration led to significant
loss of Ar-SF4Cl). The crude product of Ar-SF4Cl (ca. 0.1 mmol, 1 equiv) was dissolved in 1 mL of dry and degassed DCM followed
by the addition of AgBF4 (0.2 mmol, 2 equiv) under argon. The mixture was stirred at 100 °C for 48 h. Upon
completion of the reaction, the resulting solution was then concentrated and further
purified by column chromatography (silica gel, pentane/EtOAc, 20:1); this gave 8; yield: 19 mg (49%).
1H NMR (298 K, 300 MHz, CDCl3): δ = 7.55–7.48 (m, 4 H).
19F{1H} NMR (298 K, 282 MHz, CDCl3): δ = 83.5 (m, 1 F), 63.0 (d, J = 150.3 Hz, 4 F).
13C NMR (298 K, 75 MHz, CDCl3): δ = 152.6 (m), 131.9, 127.6 (quin, J = 4.7 Hz), 126.1.
