Pd-Catalyzed Synthesis of Ar–SCF3 Compounds Under Mild Conditions

The unique chemical properties of aryl trifluoromethyl sulfides (ArSCF₃) have been known for over 60 years.^[1] The capacity of SCF₃ to act as a lipophilic electron-withdrawing group has resulted in the incorporation of ArSCF₃ components into a number of pharmaceutical and agrochemical agents.^[2] Unfortunately, direct access to this important class of compounds is complicated by a lack of efficient, safe and general methods.^{[1a, 3]}

Significant advances in Pd-catalyzed cross-coupling processes have allowed for efficient access to a diverse array of functionalized aromatic products, including aryl sulfides.^[4] While the coupling of many aromatic or aliphatic thiols with aryl halides has been achieved with very high efficiency,^[5] the analogous transformation to form aryl trifluoromethyl sulfides has not been reported. As gaseous CF₃SH (b.p. = -36 °C)^[6] can be difficult to handle in a laboratory setting, several SCF₃ salts have been developed, however, most of these decompose under standard cross-coupling conditions.^[3c]

It has been postulated that reductive elimination of Ar–SR from a palladium center is initiated via a nucleophilic attack on the electrophilic hydrocarbyl group by the metal-bound thiolate.^[7] Thus, metal-catalyzed Ar–SCF₃ coupling might be complicated by the reduced nucleophilicity of the SCF₃ anion^[2b] as compared to a standard thiolate.

Recent reports from our group regarding novel ligands including BrettPhos (1), t-BuBrettPhos (2), XPhos (3) and 3,4,5,6-tetramethyl(t-Bu)XPhos (4) (Scheme 1), have allowed for the successful coupling of weak nucleophiles traditionally thought to be reluctant participants in the transmetalation or reductive elimination steps of a typical Pd(0)/Pd(II) catalytic cycle. Specifically, using these catalyst systems has allowed for the direct formation of diaryl ether,^[8] aryl fluoride,^[9] aryl trifluoromethyl,^[10] and aryl nitro compounds^[11] from their corresponding aryl halides or pseudo halides. In light of these results, we hypothesized that a similar Pd-based system might allow for the formation of an aromatic C–SCF₃ bond.

Scheme 1.

Various ligands used in Pd-catalyzed cross-coupling reactions.

As we suspected that reductive elimination from putative intermediate 11 would be rate limiting in any catalytic process, we began our investigation by attempting its preparation from oxidative addition complex 10 via treatment with AgSCF₃ (Scheme 2). We were surprised when this procedure did not provide the expected transmetalation complex but instead led directly to the Ar–SCF₃ product 12 (presumably via 11).

Scheme 2.

Formation of ArSCF₃ via transmetalation and reductive elimination from an isolated LPdAr(Br) complex.

Given this finding, we attempted to convert 4-(4-bromophenyl)morpholine to the corresponding trifluoromethyl sulfide using AgSCF₃ and a catalytic quantity of 1 and (COD)Pd(CH₂TMS)₂ (Table 1). However, under these conditions, none of 13 was observed. We surmised that failure to observe the coupled product might be due to the inefficient transfer of ^⊖SCF₃ to 10 under catalytic conditions. Thus, we elected to examine the use of a number of alternative previously reported ^⊖SCF₃ sources (Table 1).^{[3c, e]}

Table 1.

Examination of different SCF₃ sources.^[a]

Entry	SCF3 Source	Additive	Yield [%]
1	CsSCF3	none	10
2	AgSCF3	none	0
3	(Me)4NSCF3	None	20
4	AgSCF3	(Me)4NI	0
5	AgSCF3	(Bu)4NCl	0
6	AgSCF3	(Bu)4NBr	56
7	AgSCF3	(Bu)4NI	55
8	AgSCF3	Ph(Me)3NI	80
9	AgSCF3	Ph(Et)3NI	>99

^[a](COD)Pd(CH₂TMS)₂ (2.5 mol %), PhMe (4 mL); all reactions are run on 0.2 mmol scale and all reported yields are based on GC data

Clark’s^[3d] work on the use of (Bu)₄NI and AgSCF₃ for S_NAr reactions with aryl halides indicated to us that the addition of a quaternary ammonium salt might be beneficial. Consistent with this hypothesis, the addition of 1 equivalent of (Bu)₄NI to the reaction mixture increased the yield of 13 from 0 % to 55 % (Table 1). Further examination of different ammonium salts revealed that Ph(Me)₃NI was more effective than (Bu)₄NI and that switching to a more soluble ammonium salt, Ph(Et)₃NI, provided a nearly quantitative yield of the desired product (Table 1). Based on work done by Clark, it is presumed that the iodide anion binds to AgSCF₃ in order generate an anionic “ate” complex. We hypothesize that a large diffuse cation further aids in the solubility of this complex. It is worth noting that while the use of quaternary ammonium iodides and bromides allowed for catalytic turnover, the corresponding chloride analogs were ineffective.

With the optimal combination of Ph(Et)₃NI and AgSCF₃ realized, we re-examined various other previously reported ligands, which have enjoyed a measure of success in Pd-catalyzed cross-coupling reactions.^[12] Our survey revealed that only dialkylbiarylphosphine based ligands were successful carrying out this transformation, while other ligands such as 5 or 6 did not perform well even with higher catalyst loadings.

Accordingly, we were successful in converting electron-rich, -neutral and -deficient aryl bromides to their respective aryl trifluoromethyl sulfides in 2 hours at 80 °C using 1.5 - 3.5 mol % of Pd and 1.65 – 3.85 mol % of 1. Electron-neutral and electron-rich substrates were coupled more efficiently than their electron-poor analogs. This effect has previously been noted in the coupling of aryl halides with NaNO₂.^[11] Substrates containing acid-sensitive functional groups, such as BOC-protected anilines and nitriles, were tolerated and coupled in high yield along with substrates containing ketones, esters, and free NH groups of anilines (Table 3). Aryl bromides containing bulky ortho-groups, e.g., o-cyclohexyl and o-phenyl groups, could also be coupled successfully, although they required the use of the smaller ligand XPhos (3) (Table 3).

Table 3.

Pd-catalyzed coupling of aryl bromides.^[a]

Heteroaryl bromides such as those containing indoles, pyridines, quinolines, thiophenes and furans, were also viable substrates (Table 4). Unfortunately, attempts to extend this methodology to the coupling of aryl chlorides or aryl triflates were unsuccessful. We are currently working to understand and overcome these limitations.

Table 4.

Pd-catalyzed formation of heteroaryl–SCF₃ compounds.^[a]

Finally, to demonstrate the utility of this method, we prepared an intermediate in the reported synthesis of Toltrazuril,^[13] an antiprotozoal agent. Intermediate 14 can be assembled from readily available starting materials in an overall yield of 88%. The key C–SCF₃ bond-forming process proceeded in 95% yield (Scheme 3).

Scheme 3.

Synthesis of Toltrazuril intermediate.

In summary, we have developed a general method for the Pd-catalyzed Ar–SCF₃ bond-forming reaction. Using this method, a wide range of aryl bromides were converted into their corresponding aryl trifluoromethyl sulfides. Additionally, we have been successful in generating a variety of heterocyclic aryl trifluoromethyl sulfides from heteroaryl bromide precursors. Due to the utility of Ar–SCF₃ compounds as biologically active agents, and the mild reaction conditions employed, we expect this method to be immediately implemented in the discovery of novel compounds with pharmaceutical and agrochemical applications.

Table 2.

Examination of various ligands commonly employed in Pd-catalyzed reactions.^[a]

Entry	Ligand	Time (h)	Yield (%)
1	1	1	>99[b]
2	2	1.5	60[b]
3	3	1	84[b]
4	7	2	36[c]
5	8	2	3[c]
6	5	2	29[d]
7	6	2	0[d]
8	9	2	0[d]
9	4	2	<1[c]

^[a]PhMe (4 mL); all reactions are run on 0.2 mmol scale and all reported yields are based on GC data.

^[b]1.15 mol % Pd, 1.27 mol % L.

^[c]1.5 mol % Pd, 1.65 mol % L.

^[d]2.5 mol % Pd, 2.75 mol % L