DBFOX-Ph/metal complexes: Evaluation as catalysts for enantioselective fluorination of 3-(2-arylacetyl)-2-thiazolidinones

We examined the catalytic enantioselective fluorination of 3-(2-arylacetyl)-2-thiazolidinones 1 with N-fluorobenzenesulfonimide (NFSI) by DBFOX-Ph/metal complexes under a variety of conditions. After optimization of the metal salts, solvents and additives, we found that the fluoro-2-thiazolidinones 2 were obtained in good to high yields with moderate to good enantioselectivities (up to 78% ee) when the reaction was carried out in the presence of DBFOX-Ph (11 mol%), Ni(ClO4)2·6H2O (10 mol%) and 2,6-lutidine (0 or 1.0 equiv) in CH2Cl2.


Results and Discussion
Our previous studies of the DBFOX-Ph/Ni(II)-catalyzed enantioselective fluorination of β-keto esters have shown that the optimal reaction conditions require NFSI as the fluorine source and a catalytic amount of Ni(ClO 4 ) 2 ·6H 2 O in CH 2 Cl 2 at room temperature. Therefore, we first attempted the reaction of 1a with the same conditions and found that the desired fluorinated product 2a was obtained in 42% yield with 69% ee (Table 1, entry 1). The reaction at higher temperature (40 °C) improved the yield to 62% with slightly lower enantioselectivity (63% ee, entry 2). The reaction time in these experiments was shortened by the addition of 1 equiv of 2,6-lutidine and 2a was obtained in 87% yield with 66% ee at room temperature (entry 3). Both yield and selectivity were improved to 90% and 74% ee when the reaction was performed at 0 °C (entry 4). The highest ee value of 2a was obtained at −20 °C, but resulted in a decrease in yield (24%, 79% ee, entry 5). Changing the metal salts did not improve the results (entries 6 and 7). The absolute stereochemistry of 2a was determined by comparing the optical rotation and HPLC analysis with the literature values [57]. Although the enantioselectivities are moderate to good in these examples (63-79% ee), the results are quite impressive because the fluorination proceeds even in the absence of base (entries 1 and 2). That is, both Ni(ClO 4 ) 2 -DBFOX-Ph (unary system, entries 1 and 2) and Ni(ClO 4 ) 2 -DBFOX-Ph/lutidine (binary system, entries 3-6) are moderately effective in the enantioselective fluorination of 1a. According to the report by Sodeoka using their NiCl 2 -BINAP/R 3 SiOTf-lutidine (trinary system, up to 88% ee obtained), the reaction requires both R 3 SiOTf and lutidine [57]. They mentioned in the paper that a binary system consisting of Ni(OTf) 2 -binap complex and 2,6-lutidine failed to promote asymmetric fluorination. We also briefly attempted the fluorination of 1a using the (S,S)-Box-Ph ligand instead of DBFOX-Ph. While the Box-Ph/Cu(OTf) 2 catalyst was not effective (run 8), the Box-Ph/Ni(ClO 4 ) 2 ·6H 2 O catalyst gave the desired product 2a in 33% yield with low enantioselectivity (15% ee, entry 9).
The DBFOX-Ph/Ni(ClO 4 ) 2 ·6H 2 O catalysis for fluorination showed high generality for various 3-(2-arylacetyl)-2thiazolidinones 1a-k in good to high yields with moderate to good enantioselectivities. The results are summarized in Table  2. The fluorination reaction was not very sensitive to substitu-tion in the position of the phenyl group and the desired products with methoxy or methyl groups at the o-, m-, or p-position of the benzene ring were obtained in 65-78% ee (entries 2-7). The reactions of fluoro or bromo-substituted 1h, i and bulky-substituted 1j, k afforded the desired products 2h-k in good yields with slightly lower enantioselectivities (56-62% ee, entries 8-11).
The R-enantioselection of 2 can be explained by assuming an octahedral complex coordinated with a water molecule for DBFOX-Ph/Ni(II)/1 as shown in Scheme 1. In the complex, the Si face is shielded by one of the phenyl groups of DBFOX-Ph so that NFSI approaches from the Re face of the substrates (Scheme 1). Since a major difference in ee values of 2 was not observed for the fluorination reaction of 1 with NFSI in the presence or absence of 2,6-lutidine (entries 1-3, Table 1), 2,6lutidine presumably just accelerates the tautomerization of 1 to its enol form.

Conclusion
This research has demonstrated that DBFOX-Ph/Ni(II) catalysis can be used for the catalytic enantioselective fluorination of 3-(2-arylacetyl)-2-thiazolidinones with or without 2,6lutidine to afford chiral 2-fluoro-2-arylacetate derivatives in good to high yields with moderate to good enantioselectivities of up to 78% ee. The Box-Ph ligand was not effective for this reaction. Our best ee value is slightly lower than that of Sodeoka's report [57]; this is presumably due to the low activity of our catalyst system which requires higher reaction temperature conditions (0 °C vs. −20 °C [57]). Racemization of the products 2 during the fluorination reaction was ruled out since no racemization was observed when 2a was stirred overnight under the same fluorination conditions. Further studies to improve the enantioselectivity of DBFOX-Ph/metal catalysis in enantioselective fluorination are under way.

Supporting Information
Supporting Information File 1 Experimental methods. General methods, general procedure for the enantioselective catalytic fluorination, spectral data of 2, copies of 1 H, 13