Electrophilic fluorination of cationic Pt-aryl complexes

Abstract

The electrophilic fluorination of several (triphos)Pt-aryl⁺ establishes the first example of aryl–F coupling from a Pt center.

The demand for organofluorine compounds has stimulated much recent effort to develop metal mediated fluorination reactions.¹ Despite the versatility of available C–X (X = C, N, O, S, Cl, Br, I, etc.) coupling methodologies,² C–F couplings via reductive elimination remain challenging.^1d,f Metal catalyzed C–F couplings that utilize fluoride sources encounter additional challenges due to the intrinsically low polarizability and nucleophilicity, pronounced hydration power, and high basicity of F⁻. Nevertheless, several notable Pd^0/II catalyzed nucleophilic fluorinations have been recently reported.³ More fruitful have been recent metal-catalyzed electrophilic fluorination reactions,^5–8 wherein high-valent metal fluoro intermediates (e.g. Pd(IV),⁴ Ag(II)···Ag(II),⁵ Au(III),⁶ etc.) are more prone to productive reactivity, including C–H activation, cross-coupling, and C–F reductive elimination.^1f,7

To explore Pt analogues of these electrophilic reactions, we recently demonstrated a system that efficiently fluorinates Pt–C_sp³ bonds.^8b As illustrated in eqn (1), wherein PPP = bis(2-diphenylphosphinoethyl)phenylphosphine (i.e., triphos), the C–F coupling proved to be stereoretentive and was proposed to occur by concerted reductive elimination of a putative dicationic Pt(IV)–F intermediate (A). The reaction was accelerated by increased steric congestion around Pt,^8b however, information on the short-lived Pt(IV)–F species was lacking.

(1)

Sp²-carbon–halogen bond forming reactions from Pt(IV) centers are rare,^1g,9,10 with the few known examples restricted to C–I and C–Br couplings.¹¹ Extending our efforts on Pt–C bond fluorination reactions, we have examined the electrophilic fluorination of (triphos)Pt-aryl⁺ complexes. Herein, we report these reactions and provide evidence that supports the intermediacy of Pt(IV)–F complexes in the C–F reductive coupling reaction.

Complexes 1–4 were synthesized by ligand displacement of (COD)PtAr(X) (COD = cycloocta-1,5-diene, X = Cl, or I) with triphos, followed by salt metathesis with NaBF₄.^12,13 Complex 5 was prepared by treating chloro(2-phenylpyridine)[2-(2-pyridyl)- phenyl-C,N]Pt¹⁴ with triphos, while its dicationic isostere 6 was obtained by reacting [(triphos)Pt(NCC₆F₅)](BF₄)₂¹⁵ with 2-phenylpyridine.¹²

These compounds were characterized by NMR and HRMS, with the molecular structure of 4^§ being verified by X-ray analysis (Fig. 1).¹² Consistent with the solid state structure of 4, NOESY analysis suggested that the ortho-substituent in 2, 4–6 preferentially oriented syn to the central P-Ph group of the triphos ligand. While 2, 4, 5 and 6 exist exclusively in this syn-rotamer, both syn- and anti-forms (2.7: 1) were observed for 3.¹² The preference for the syn- over the anti-form suggests that the face of the square plane containing the apical P-Ph group is less congested and may be more kinetically accessible.

Fig. 1.

Left: complexes 1–6; right: X-ray structure of 4 (H atoms and BF⁴⁻ anion are omitted for clarity).

When subjected to electrophilic fluorination conditions, these (triphos)Pt-aryl⁺ complexes were found to be much less reactive than their Pt-alkyl⁺ analogs.^8b When screening common “F⁺” sources including N-fluorobenzenesulfonimide, several N-fluoropyridinium salts, Selectfluor^® and XeF₂, only the latter two exhibited reasonable reactivity with 1, for which the optimal solvent was identified to be acetonitrile. ³¹P and ¹⁹F NMR spectroscopy proved most advantageous for in situ monitoring of these reactions and Selectfluor^® proved to be cleaner and more productive than XeF₂.

With 1, a complex mixture of phenyl Pt(IV)–F species was obtained upon reacting with XeF₂ (RT,<20 min). By contrast, Selectfluor^® provided one main phenyl Pt(IV)–F complex (RT, ~2 h) (δ_F = −360.3 ppm, J_Pt–F = 1453 Hz).¹² These Pt(IV)–F species, however, failed to reductively eliminate PhF even after prolonged heating (80 °C, >30 h).

The ortho-substituents considerably slowed down the reactions of 2 and 3 with XeF₂ and Selectfluor^®, however, their presence proved beneficial for achieving the desired sp² C–F coupling. In the case of 2, XeF₂ provided one major Pt(IV)–F complex (δ = −352.8 ppm, J_Pt–F = 1442 Hz) in ~75% NMR yield (RT, 12 h).¹² However, the precise structure of this product remains unclear, as all attempts to crystallize it failed and spectroscopic data were not conclusive. Heating a freshly prepared reaction mixture containing this Pt(IV)–F complex at 80 °C led to only traces of the aryl–F coupling product (<5% GC-MS yield). Similar results were obtained when directly reacting 2 with XeF₂ at 80 °C. In contrast, reactions of 3 with XeF₂ (RT, 15 h) directly generated a substantial amount of the aryl–F coupling product 1-fluoro-2,4-dimethylbenzene (~55% NMR yield), along with the corresponding [(triphos)Pt-NCMe]²⁺ by-product. The formation of a Pt(IV)–F complex (δ_F=−351.9 ppm, J_Pt–F=1146 Hz) in~25%NMR yield and other unidentified Pt species was also observed.¹² To our knowledge, this reaction represents the first example of aryl–F coupling from a Pt center.

Despite being unreactive at RT, Selectfluor^® readily fluorinated 2 and 3 at 80 °C to produce the aryl fluoride;¹² no Pt(IV)–F species was observable during in situ monitoring of these reactions. These results are summarized in Table 1.

Table 1.

Fluorination of complexes 1–3 with Selectfluor^®a

Complex	Product	Time	NMR yieldb (%)
1	[(PPP)PtIV(Ph)(F)]2+	<20 min	60–70
2		1 h	91
3		2 h	>95

^aConditions: complexes 1–3 (0.02 mmol), 1.5 equiv. of Selectfluor^®, dry CD³CN (0.5 mL), 80 °C.

^bMass balance: structurally unidentified organometallic Pt species.

Surprisingly, the reaction of 4 with XeF₂ preferentially yielded the ortho-cyclometalated complex 7 (Scheme 1). NMR monitoring of the reaction revealed its gradual conversion to the Pt(IV)–F complex, 7^§, which was characterized byNMR, HRMSand X-ray diffraction.¹² In contrast to the aforementioned Pt(IV)–F species, this complex exhibits a ¹⁹F NMR resonance at δ = −299.9 ppm with a considerably diminished ¹⁹⁵Pt–¹⁹F coupling (~173 Hz).

Scheme 1.

Generation of complex 7; inset: X-ray structure of 7 (H atoms and anion are omitted for clarity).

As shown in Scheme 1, the Pt center in 7 adopts an octahedral coordination geometry, with the Pt–F bond (2.099(2) Å) oriented anti to the central P-Ph group of the triphos ligand, and the biphenyl moiety adopting a C,C′-chelating mode. Similar cyclometalation of an ortho sp²-C–H bond was previously noted upon fluorinating (triphos)Pt-CH₂Ph⁺ with XeF₂ in melting acetonitrile.^8b This reactivity mode apparently reflects the intermediacy of Pt(IV) fluorides in both cases.^8b The propensity of Pt(IV) and Pd(IV) centers in metalating aromatic C–H bonds has been demonstrated and exploited recently in several coupling strategies.^7,8a

Heating an acetonitrile solution of 7 at 80 °C resulted in slow F⁻ extrusion and the concomitant formation of a dicationic Pt(IV)–MeCN adduct, 8 (eqn (2)). No C–F reductive elimination was observed during the process, and X-ray diffraction^§ revealed that the MeCN ligand coordinates syn to the triphos ligand’s central P-Ph group (eqn (2)).¹² Consistent with the increase in the net charge of the Pt(IV) center are large downfield shifts of the ³¹P NMR signals as compared to 7 (e.g., Δδ = +22.7 ppm for the central P) and ¹H NMR signals of the biphenyl moiety.

(2)

In addition to 7, reactions of 4 with XeF₂ at RT (Scheme 1) also yielded traces of 8 (<5%).¹² By contrast, reactions of 4 with Selectfluor^® directly provided 8 (85%,~5 h), along with 15% of 2-fluorobiphenyl and the corresponding [(triphos)Pt-NCMe]²⁺ (Scheme 2).¹² We reason that the formation of both 7 and 8 implies the presence of Pt(IV) intermediates.

Scheme 2.

Competitive cyclometalation and C–F coupling pathways.

The contrasting outcomes for reactions of 2–4 with XeF₂ and Selectfluor^® presumably stem from the presence of a basic fluoride anion in the former case, though a size difference in the “F⁺” source is also conceivable.¹⁶ Shown in Scheme 1 is one way wherein F⁻ could accelerate ortho-metalation vs. reductive elimination. Since the two Pt(II) faces were shown to be sterically different, it is also possible that these reactions evolve differently based on which face F⁺ attacks.¹⁶ Recently, Vigalok and co-workers have also reported an F⁺ reagent-dependent reaction behavior when fluorinating Pt-aryl complexes.^8a

To gain more insights into the Pt(IV)–F species proposed in Schemes 1 and 2, the fluorination of 5 and 6 by XeF₂ was examined. In particular, we hoped that the ortho-pyridyl group in 5 could trap the coordinatively unsaturated Pt(IV)–F intermediate. Instead, XeF₂ converted 5 into the Pt(II) complex 9 (eqn (3)), whose configuration was deduced from ³¹P NMR data (e.g., δ_F = −37.9 ppm, J_P1-F = 652 Hz; J_Pt-P3 = 3745 Hz vs. J_Pt-P2 = 1863 Hz).¹² The formation of this complex presumably occurred via associative displacement of one triphos phosphine arm (P₁, eqn (3)) in 5 by the pyridyl ligand, followed by oxidation of the unligated phosphine ligand. We have previously shown that phosphine fluorination by XeF₂ is rapid.^8b Despite its structural analogy to 4 and 5, complex 6 failed to react with XeF₂, indicating that a dicationic Pt(II) center may be too electron deficient to generate a tricationic Pt(IV) structure.

(3)

In summary, we report the first sp² C–F coupling from a Pt center. Like Pt–C_sp³ bonds, steric congestion is a key factor, as is F⁺ source. We have also demonstrated that ortho-metalation may be competitive with C–F reductive elimination. The intermediacy of Pt(IV)–F complexes, the product of direct F⁺ addition to Pt(II), is supported by the direct spectroscopic observation of several Pt(IV)–F species and the isolation of ortho-metalation products.