A Titanium‐Catalyzed Reductive α‐Desulfonylation

Abstract A titanium(III)‐catalyzed desulfonylation gives access to functionalized alkyl nitrile building blocks from α‐sulfonyl nitriles, circumventing traditional base‐mediated α‐alkylation conditions and strong single electron donors. The reaction tolerates numerous functional groups including free alcohols, esters, amides, and it can be applied also to the α‐desulfonylation of ketones. In addition, a one‐pot desulfonylative alkylation is demonstrated. Preliminary mechanistic studies indicate a catalyst‐dependent mechanism involving a homolytic C−S cleavage.

Reactions involving air sensitive reagents were performed in flame-dried Schlenk tubes or Schlenk flasks under argon atmosphere (argon 5.0) and using absolute solvents unless noticed otherwise. Absolute THF and absolute toluene were dried over potassium under argon atmosphere and freshly distilled prior to use. Dichloromethane, pentane, ethyl acetate and diethyl ether were purchased in p.a. quality. Cyclohexane for column chromatography was purchased in technical quality and purified by destillation with a rotary evaporator. An IKAmag temperature modulator in combination with an oil bath was used to control the reaction temperatures. Thin-layer chromatography (TLC) was performed using E. Merck silica gel 60 F254 precoated plates (0.25 mm) and visualized by UV fluorescence quenching or staining (KMnO4). In general, Macherey-Nagel Silica gel 60 (particle size 0.04-0.063 mm) was used for flash chromatography. 1 H, 19 F and 13 C NMR spectra were recorded on a Bruker DRX 500 ( 1 H: 500 MHz and 13 C: 125 MHz), a Bruker Avance II 400 ( 1 H: 400 MHz and 13 C: 100 MHz), and a Bruker Avance III 300 ( 1 H: 300 MHz and 19 F: 282 MHz) spectrometer and reported to CDCl3 [d( 1 H) = 7.26 ppm and d( 13 C) = 77.16 ppm]. The following abbreviations were used: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet. NMR spectra were recorded at 23 ºC unless noted otherwise. Melting points were determined using a Coesfeld MPM-HV2 melting point apparatus and are uncorrected. IR spectra were recorded on a Thermo Scientific Nicolet iS10 FT-IR spectrometer equipped with a diamond ATR unit and are reported in frequency of absorption. Low-and high-resolution mass analyses were performed by the service department at the Institute for Organic Chemistry and Biochemistry, University of Freiburg using a Thermo Finnigan TSQ 700 for electron impact ionization (EI) at 70 eV, 200 °C. High resolution mass analyses (HRMS) were carried out on a Thermo Exactive with Orbitrap-Analyzer using atmospheric pressure chemical ionization (APCI) or electrospray ionization (ESI). Half-wave reduction potentials were determined from cyclic voltammograms that were recorded with a VersaSTAT 4 Potentiostat Galvanostat under moisture-and oxygen-free conditions. A standard three-electrode setup was employed, with a Pt millielectrode (model G0228, AMETEK) as working electrode. The counter electrode consisted of a platinum wire in an electrolyte (0.2 M Bu4NPF6 in THF), while the reference electrode consisted of a silver wire in a 3 M NaCl/sat. AgCl solution. [1] S4 2-(Phenylsulfonyl)-3-(p-tolyl)propanenitrile. [6] Synthesized from 2-(phenylsulfonyl)acetonitrile (136.4 mg, 0.75 mmol, 1.0 equiv) and 4-methylbenzaldehyde following the representative procedure for 1a. The product was purified using column chromatography (CH2Cl2, Rf = 0.5) and then obtained as a colorless solid in 93% yield (200 mg, 0.70 mmol). The NMR data matched the literature values. 1

Titanium-Catalyzed Desulfonylation Reactions
Extended Screening Results for the Catalytic Reductive Desulfonylation

Additive Influence
Reducing the number of additives (ZnCl2, Coll•HCl, TMSCl) led to a diminished reaction outcome (Table S2) and each was previously shown to have a beneficial effects in low-valent titanium catalysis. [1,14,15,16] In our hands, the absence of ZnCl2 led to a significant drop in yield. The removal of Coll•HCl or TMSCl had a smaller although still significant negative effect, resulting in 15-21% lower yields. However, the reaction was still productive and this observation could be of interest for desulfonylations in presence of functional groups that are incompatible with these reagents. For completion, a reaction without TMSCl and Coll•HCl and a reaction without all three additives were carried out as well, leading to inferior results. Given the inexpensive and benign nature of the additives, we continued using the optimal conditions from Table 1 in the manuscript. The compounds 2a-r were then synthesized according to the following representative desulfonylation procedure on a 0.5 mmol scale (entries in Scheme 2 of the manuscript).

Representative Procedure for the Catalytic Desulfonylation Reaction (1a → 2a)
3-Phenylpropanenitrile. [17] A flame-dried, argon-back-filled Young tube was charged with Cp*2TiCl2 (19.5 mg, 0.05 mmol, 0.1 equiv), ZnCl2 (68.1 mg, 0.5 mmol, 1.0 equiv), Zn (98.0 mg, 1.5 mmol, 3.0 equiv), and 2,4,6-collidine hydrochloride (S2, 158 mg, 1.0 mmol, 2.0 equiv). The tube was evacuated and back-filled with argon three times, before freshly distilled toluene (2 ml) was added. The suspension was stirred for 5 minutes followed by the addition of 3phenyl-2-(phenylsulfonyl)propanenitrile (1a, 136 mg, 0.5 mmol, 1.0 equiv) and TMSCl (190 µl, 1.5 mmol, 3.0 equiv). The reaction vessel was sealed and immersed into a preheated oil bath at 110°C. The reaction mixture was stirred for 24 hours and was then allowed to cool to 23 ºC, resulting in a clear solution with a solid precipitate. CH2Cl2 (2 ml) was added and the mixture was stirred for 5 minutes under air. The solution was filtered and the residue was rinsed three times with a small amount of CH2Cl2 (3 × 2 ml). Here, a spatula was used to suspend the residue in the CH2Cl2. The combined filtrates were then concentrated under reduced pressure.

Michael Addition/Desulfonylation Tandem Reaction
In initial experiments, acrylonitrile was identified as a suitable partner for the one-pot desulfonylative Michael reaction. Other Michael acceptors such as methyl acrylate, phenyl acrylate, cinnamonitrile, and methyl vinyl ketone were tested as well during our optimization studies, but no coupling products were observed. A broad expansion of this tandem reaction will require an additional thorough optimization that is not part of this study.

Final Optimized Procedure
A flame-dried, argon-back-filled Young tube was charged with Cp*2TiCl2 (9.73 mg, 0.025 mmol, 0.1 equiv), ZnCl2 (34.1 mg, 0.25 mmol, 1.0 equiv), Zn (49.0 mg, 0.75 mmol, 3.0 equiv) and 2,4,6-collidine hydrochloride (S2, 78.8 mg, 0.5 mmol, 2.0 equiv). The tube was evacuated and back-filled with argon three times before freshly distilled toluene (1 ml) was added. The resulting suspension was stirred for 5 minutes followed by the addition of 3-phenyl-2-(phenylsulfonyl)propanenitrile (1a, 67.8 mg, 0.25 mmol, 1.0 equiv), acrylonitrile (26.2 µl, 0.40 mmol, 1.6 equiv) and TMSCl (95 µl, 0.75 mmol, 3.0 equiv). The reaction vessel was sealed and immersed into a pre heated oil bath (80 ºC). The reaction mixture was stirred for 72 hours and was then allowed to cool to 23 ºC, resulting in a clear solution with a solid precipitate. CH2Cl2 (2 ml) was added and the mixture was stirred for 5 minutes under air. The solution was filtered and 2 ml acetone was added to partially dissolve the residue. The remaining solid was filtered off and the filtrates were combined followed by concentration under reduced pressure.

Reaction in Absence of Titanium Catalyst
The reaction was carried out following the above procedure without titanium catalyst at 110 ºC oil bath temperature and with 24 h reaction time.

Control Experiments (eqs 2 and 3)
Background Experiment with Thioether 5 Thioether 5 (0.25 mmol) was submitted to the representative desulfonylation procedure without using any titanium catalyst. After 3 h the reaction was quenched and worked up as described. Purification

Desulfonylation with a Stoichiometric Amount of Cp*2TiCl
All glassware was treated in a KOH bath, rinsed with distilled water and stored in a drying oven (120 °C) over night prior to the experiment. [29] Oxygen and water must be rigorously excluded during this experiment (argon atmosphere and absolute solvents are required). In a flame-dried and argon-back-filled Schlenk tube equipped with a magnetic stir bar Cp* 2 TiCl 2 (97.3 mg, 0.25 mmol, 1.0 equiv), Zn (572.0 mg, 8.75 mmol, 35.0 equiv) and 2,4,6collidine hydrochloride (78.8 mg, 0.5 mmol, 2 equiv) were suspended in freshly distilled toluene (2 ml). [30] The mixture was vigorously stirred for 10 minutes and the mixture changed from a brown-red suspension to a dark blue solution, indicating the formation of Cp* 2 TiCl. [31,32] The dark blue solution was filtered using a Schlenk frit into a second flame-dried and argon-backfilled schlenk tube containing a magnetic stir bar. The former reaction flask and Schlenk frit were rinsed with toluene (1 ml and was then allowed to cool to 23 °C, resulting in a clear brown solution with a solid precipitate. CH 2 Cl 2 (2 ml) was added and the mixture was stirred for 5 minutes under air. The solution was filtered and the residue was washed three times with a small amount of CH2Cl2 (3 × 2 ml) followed by filtration. Here, a spatula was used to suspend the residue in the CH2Cl2. The combined filtrates were then concentrated under reduced pressure. 1 H NMR analysis of the crude product mixture indicated 30-34% conversion with the remaining material being unreacted 1a. Product 2a was purified by column chromatography [pentane/Et2O, 20:1→10:1, Rf(pentane/Et2O, 20:1) = 0.5] and obtained as colorless liquid in 30% yield (9.8 mg, 0.075 mmol).

A note on potential byproducts emerging from the phenylsulfonyl radical
After workup of the stoichiometric experiment as described we did not observe byproducts originating from the phenylsulfonyl radical in the crude mixture. However, the literature suggests that phenylsulfonyl radicals undergo a number of reactions including loss of SO2 to give a phenyl radical, [33] and homo coupling and disproportionation reactions that can be promoted by protic additives such as water (Scheme S1). [34] Reduction by titanium(III) is another possible pathway. This renders the identification and quantification of the byproducts difficult and further investigations will be required to elucidate the nature of the byproducts. [35] Scheme S1. Potential sequential reactions of phenylsulfonyl radicals.

Computational Details
The Orca 4.2.0 program package was used for the DFT calculations. [36] The RI-J approximation for Coulomb integrals and the COSX numerical integration for HF exchange (RIJCOSX) were applied. [37,38] Furthermore, the D3 dispersion correction with Becke-Johnson damping, D3(BJ), was applied in all calculations. [39,40] All structure optimizations were finalized using the TPSS functional [41] together with the def2-TZVP basis set [42] and matching auxiliary basis sets. [43] The conductor-like polarizable continuum model (CPCM) [44] was applied for the optimizations and the structures were separately optimized in THF and toluene. The optimizations were carried out with the Grid3 FinalGrid5 TightSCF options. Frequency analyses were carried out numerically (NumFreq). Stationary points (minimum structures) were characterized by the absence of imaginary frequencies. The correction to the Gibbs Free Energy was obtained from the Orca output of the frequency calculation. Single-point calculations were carried out using the PW6B95 [45] functional and the def2-QZVP basis set and matching auxiliary basis sets. [46] The single point calculations were carried out with the Grid4 FinalGrid5 options. The CPCM model together with the correction DG* → ºsolv (= 1.90 kcal mol -1 , see Born-Haber cycle in Scheme S2) were applied to obtain the energy in solution.
Scheme S2. Interconversion Scheme for the calculation of solvation energies for a compound X.