Final Report Summary - BENFLU (A Transition Metal Approach to Benzylic Fluorination)
The number of active molecules that contain fluorine-substituted moieties has increased over the past 30 years. Thirty to forty percent of agrochemicals [1] and twenty percent of pharmaceuticals [2] currently on the market contain at least one fluorine atom. It has been recognized that the introduction of fluorine into organic molecules significantly improves properties such as solubility, bioavailability and metabolic stability.[3] In addition, radioactive 18F-labeled imaging agents, such as 2-[18F]fluoro- 2-deoxyglucose and [18F]-DOPA, are widely used for positron emission tomography (PET) in oncology and imaging the brain.[4]
Despite this potential, methodologies for C(sp3)-F bond formation are still limited. Most of known methodologies require highly reactive and dangerous fluorinating reagents and harsh conditions.[5] Very recently, radical sources of fluorine, such as Selecfluor, have emerged as promising reagents for the direct incorporation of fluorine onto benzylic systems by a formal C-H functionalisation, through a radical addition process. The strategy developed by Lectka and co-workers takes advantage the oxidative properties of Fe(II) to cleave the C-H bond homolytically. ).[6] The benzyl radical is intermolecularly trapped by Selectfluor, resulting in installation of fluorine atom in good yields (Scheme 1a). Analogously, Inoue and co-workers have described a metal-free methodology, employing N,N-dihydroxypyromellitimide (NDHPI) as an N-Oxyl radical catalyst, for the fluorination of benzylic and alkylic compounds.[7] An elegant method have been reported by Groves et al. using Mn(salen) as catalyst in conjunction with a nucleophilic fluoride reagent and iodosylbenzene as oxidant.[8]
However, despite all of these methodologies have been successfully developed, they suffer from a number of limitations. Among them, the common radical mechanism prevents the development of enantioselective variant. In 2006 and then in 2012, Sanford and co-workers reported the palladium-catalysed fluorination of quinoline using an electrophilic source of fluorine [9] and silver fluoride combined with an external oxidant, [10] respectively. In the mechanism postulated, the oxidant species (F+ or F- in presence of an external oxidant) converts the Pd(II) catalyst to Pd(IV)-F intermediate that ultimately participates in C-F bond-forming by reductive elimination. Despite this potential, this approach is restricted to the presence of a directing group on the arene ring able to coordinate the metal.
Given such limitations there are still present in the benzylic fluorination, the BENFLU project aimed to define a new strategy that can overcome the problem presented above. Herein, we developed a new cis hydrofluorination of alkenylarenes using the Pd(II)/Pd(IV) methodology (Scheme 1c). We started our investigations by exploring the efficiency of different fluorine sources in the reaction of 4-vinylbiphenyl with iPr3SiH. The model styrene 1a was selected in basis that its corresponding fluorinated product is enough stable to be isolated and analysed. All the reactions were performed in presence of 10 mol% of Pd(PPh3)4, 2 equiv. of iPr3SiH, 2 equiv. of F+ source in MeCN, at 0°C (Table 1). Whereas the 2,3,4,5,6-pentachloro-1-fluropyridinium triflate reacts rapidly with the styrene, only the related alcohol and the corresponding saturated product 3 were obtained after 1.5 h (entry 1). Differently, although the lower reactivity, a small amount of fluorinated product 2a was observed using N-Fluorodibenzenesulfonimide (NFSI) (entry 2). A higher conversion to the product was observed when XeF2 was used as a fluorine source to the detriment of regioselectivity, probably due to the decomposition of XeF2 in situ (entry 3). To our delight, using Selectfluor, the desired fluorinated product 2a was obtained in 38% yield and surprisingly, in these conditions no product 3 was observed (entry 4). Control experiments were carried out, establishing the effectiveness of the Pd(PPh3)4 as catalyst and the silane in the reaction (entries 5 and 6).
Encouraged by these promising results we next addressed reaction optimization using Selectfluor as F+ source. The screening of different hydride sources highlighted their influence on the reactivity and selectivity of the system. In order to increase the yield, the reaction was performed using PhSiH3 and NaBH4. Although these hydride sources are already reported to be suitable reagents in the cobalt-catalysed hydrofunctionalisation [11] and Fe-catalysed hydrofluorination [12] of olephines, in our case, only a small amount of product was observed with NaBH4. Tin-based compounds such as Bu3SnH have shown low reactivity, giving only traces of product after 24h (entry 9). It worth noting that olephines undergoes dimerisation and polymerization in presence of palladium catalyst.[13] Indeed, more diluted conditions and a larger excess of Selectfluor were used, affording 2a in a better yields (entries 10 and 13). Finally, after testing several trisubstituted silanes, Et3SiH gave the best yields (entries 12 and 13) proving a match between the hydride-donor ability of silane and the overall reaction rate.[14]
The optimized reaction conditions were applied to a series of alkenylarenes (Table 2). Through the investigation of functional group tolerance, electronic effect on the aromatic ring was found crucial for the success of the reaction. Electron-withdrawing substituents such as esters, cyano, nitro, fluoride, trifluoromethyl and aldehyde, have affording the desired products in good yields (from 54 to 68%). Similar results were obtained in presence of weakly activating groups (entries 1 and 2). Furthermore, high stable benzyl fluorides such as 2m, 2n and 2o were successfully synthesized in quantitative yield. Remarkable result had been obtained with the 4-vinylquinoline 1s. It is well known indeed, that compounds such as pyridine and quinolines are able to bind to the metal, deactivating the catalyst. Moreover, the method permits the installation of a fluorine substituent at a quaternary benzylic position as illustrated by the successful synthesis of 2t (65%). Overwise, electron-rich substrates underwent the corresponding hydration and polymerization, afforded the attended products in very low yields (entries 15, 16). Finally, the efficacy of this methodology has also been adopted with 2-substituted alkylenylbenzenes 1u, 1v and 1w giving the corresponding fluorinated products in good yield and exclusive Markovnikov regioselectivity. Pleasingly, (E)- and (Z)-2-benzylidene-3-methylbutan-1-ols (entries 23, 24) gave Anti-2x and Syn-2y, respectively, in good yields and with a d.r. value of greater than 20:1. This result implies that the hydrofluorination is cis specific.
The proposed reaction mechanism for the Pd-catalysed formation of benzyl fluorides is reported in Scheme 2. The reaction proceeded by initial oxidation of the precatalyst, yielding the Pd(II) species A which activates the silane, thus giving access to the Pd(II) hydride species B, which in turn effects reversible syn hydropalladation to the alkenylarene to afford the η3-benzyl complex C. [15] Electrophilic fluorination of C with Selectfluor affords the Pd(IV)-F dication D (or its η3- complex). Reductive elimination then forms generic product E, and regenerates A.
This mechanism is supported by the various experiments, part of them described in Scheme 3. The hydrofluorination of 1b with the deuthero-labeled silane (Et3SiD) leads to (2-2H)-2b and confirms that this reagent serves as the hydrogen atom source. The formation of the labeled alkene (2-2H)-1b (<1%) and dideutero product (2,2-2H2)-2b indicates that the hydropalladation event is reversible during the course of the reaction (Scheme 3a). No deutherated product was observed when the reaction was performed in deutherated MeCN, hence the solvent is not serving as a hydrogen donor. For stereochemical assignment stilbenes (E)- and (Z)-1y were subjected to hydrofluorination using Et3SiD because both syn- and anti-(2-2H)-2y could be prepared through independent synthesis. (Z)-Stilbene led to 21% of anti-(2-2H)-2y along with trace amounts of syn-(2-2H)-2y (Scheme 3b). The stereochemical course of hydrofluorination under palladium catalysis is consistent with the studies of Gagné and co-workers [16] who reported C(sp3)-F bond formation with XeF2 proceeding with retention of stereochemistry from putative Pt(IV)-F intermediates.
Additional experiments were carried out with isolated η3-benzyl complexes (Scheme 5). We could not isolate either C or D (Scheme 2), or its η3-analogue, but found that the η3-benzyl complex C1 ligated with dppp (1,3-bis(diphenylphosphino) propane) led to 2a in 50% yield upon treatment with Selectfluor in MeCN at room temperature (Scheme 4a). Moreover, the enantioenriched naphtylethyl complex C2, prepared from (R)-T-binap according to the protocol of Hartwig et al.,[17] afforded (S)-2-(1-fluoroethyl)naphthalene (2c) in 80% ee at low temperature (82% yield) in the presence of XeF2 (Scheme 4b). By using 1.2 equivalents of Selectfluor at room temperature, 2c was formed in 36% yield (19F NMR spectroscopy) and a depleted ee value of 36%. The sense of enantiocontrol for the major enantiomer remained S (Scheme 4c). These data indicate that the stereochemical outcome for the fluorination is consistent with coordination of F+ leading to the Pd(IV)-F complex and C-F bond-forming reductive elimination. The stereoselctivity observed under palladium catalysis is consistent with the formation of a Pd(II) complex resulting from syn hydropalladation and subsequent oxidative fluorination to generate the Pd(IV)-F complex D. Retention of configuration in the C-F coupling step can then afford a product resulting from net syn hydrofluorination. Notably, the reaction of 2-vinylnaphtalene with 10 mol% of C2, 3.0 equivalents of XeF2, and 1.5 equivalents of Et3SiH in CH2Cl2 at -20 °C delivered, within one hour, enantioenriched (S)-2c (34% yield, 36% ee). This transformation is the first catalytic enantioselective hydrofluorination (Scheme 4d).
In summary, with this project we have disclosed a new palladium-catalysed HF addition to alkenes. The completely regioselectivity and cis specificity, make this reaction an attractive method for the hydrofluorination of alkenylarenes. The mechanism of this reaction involves an ionic pathway and is distinct from known hydrofluorinations involving radical intermediates. The first catalytic enantioselective hydrofluorination is also disclosed.
Despite this potential, methodologies for C(sp3)-F bond formation are still limited. Most of known methodologies require highly reactive and dangerous fluorinating reagents and harsh conditions.[5] Very recently, radical sources of fluorine, such as Selecfluor, have emerged as promising reagents for the direct incorporation of fluorine onto benzylic systems by a formal C-H functionalisation, through a radical addition process. The strategy developed by Lectka and co-workers takes advantage the oxidative properties of Fe(II) to cleave the C-H bond homolytically. ).[6] The benzyl radical is intermolecularly trapped by Selectfluor, resulting in installation of fluorine atom in good yields (Scheme 1a). Analogously, Inoue and co-workers have described a metal-free methodology, employing N,N-dihydroxypyromellitimide (NDHPI) as an N-Oxyl radical catalyst, for the fluorination of benzylic and alkylic compounds.[7] An elegant method have been reported by Groves et al. using Mn(salen) as catalyst in conjunction with a nucleophilic fluoride reagent and iodosylbenzene as oxidant.[8]
However, despite all of these methodologies have been successfully developed, they suffer from a number of limitations. Among them, the common radical mechanism prevents the development of enantioselective variant. In 2006 and then in 2012, Sanford and co-workers reported the palladium-catalysed fluorination of quinoline using an electrophilic source of fluorine [9] and silver fluoride combined with an external oxidant, [10] respectively. In the mechanism postulated, the oxidant species (F+ or F- in presence of an external oxidant) converts the Pd(II) catalyst to Pd(IV)-F intermediate that ultimately participates in C-F bond-forming by reductive elimination. Despite this potential, this approach is restricted to the presence of a directing group on the arene ring able to coordinate the metal.
Given such limitations there are still present in the benzylic fluorination, the BENFLU project aimed to define a new strategy that can overcome the problem presented above. Herein, we developed a new cis hydrofluorination of alkenylarenes using the Pd(II)/Pd(IV) methodology (Scheme 1c). We started our investigations by exploring the efficiency of different fluorine sources in the reaction of 4-vinylbiphenyl with iPr3SiH. The model styrene 1a was selected in basis that its corresponding fluorinated product is enough stable to be isolated and analysed. All the reactions were performed in presence of 10 mol% of Pd(PPh3)4, 2 equiv. of iPr3SiH, 2 equiv. of F+ source in MeCN, at 0°C (Table 1). Whereas the 2,3,4,5,6-pentachloro-1-fluropyridinium triflate reacts rapidly with the styrene, only the related alcohol and the corresponding saturated product 3 were obtained after 1.5 h (entry 1). Differently, although the lower reactivity, a small amount of fluorinated product 2a was observed using N-Fluorodibenzenesulfonimide (NFSI) (entry 2). A higher conversion to the product was observed when XeF2 was used as a fluorine source to the detriment of regioselectivity, probably due to the decomposition of XeF2 in situ (entry 3). To our delight, using Selectfluor, the desired fluorinated product 2a was obtained in 38% yield and surprisingly, in these conditions no product 3 was observed (entry 4). Control experiments were carried out, establishing the effectiveness of the Pd(PPh3)4 as catalyst and the silane in the reaction (entries 5 and 6).
Encouraged by these promising results we next addressed reaction optimization using Selectfluor as F+ source. The screening of different hydride sources highlighted their influence on the reactivity and selectivity of the system. In order to increase the yield, the reaction was performed using PhSiH3 and NaBH4. Although these hydride sources are already reported to be suitable reagents in the cobalt-catalysed hydrofunctionalisation [11] and Fe-catalysed hydrofluorination [12] of olephines, in our case, only a small amount of product was observed with NaBH4. Tin-based compounds such as Bu3SnH have shown low reactivity, giving only traces of product after 24h (entry 9). It worth noting that olephines undergoes dimerisation and polymerization in presence of palladium catalyst.[13] Indeed, more diluted conditions and a larger excess of Selectfluor were used, affording 2a in a better yields (entries 10 and 13). Finally, after testing several trisubstituted silanes, Et3SiH gave the best yields (entries 12 and 13) proving a match between the hydride-donor ability of silane and the overall reaction rate.[14]
The optimized reaction conditions were applied to a series of alkenylarenes (Table 2). Through the investigation of functional group tolerance, electronic effect on the aromatic ring was found crucial for the success of the reaction. Electron-withdrawing substituents such as esters, cyano, nitro, fluoride, trifluoromethyl and aldehyde, have affording the desired products in good yields (from 54 to 68%). Similar results were obtained in presence of weakly activating groups (entries 1 and 2). Furthermore, high stable benzyl fluorides such as 2m, 2n and 2o were successfully synthesized in quantitative yield. Remarkable result had been obtained with the 4-vinylquinoline 1s. It is well known indeed, that compounds such as pyridine and quinolines are able to bind to the metal, deactivating the catalyst. Moreover, the method permits the installation of a fluorine substituent at a quaternary benzylic position as illustrated by the successful synthesis of 2t (65%). Overwise, electron-rich substrates underwent the corresponding hydration and polymerization, afforded the attended products in very low yields (entries 15, 16). Finally, the efficacy of this methodology has also been adopted with 2-substituted alkylenylbenzenes 1u, 1v and 1w giving the corresponding fluorinated products in good yield and exclusive Markovnikov regioselectivity. Pleasingly, (E)- and (Z)-2-benzylidene-3-methylbutan-1-ols (entries 23, 24) gave Anti-2x and Syn-2y, respectively, in good yields and with a d.r. value of greater than 20:1. This result implies that the hydrofluorination is cis specific.
The proposed reaction mechanism for the Pd-catalysed formation of benzyl fluorides is reported in Scheme 2. The reaction proceeded by initial oxidation of the precatalyst, yielding the Pd(II) species A which activates the silane, thus giving access to the Pd(II) hydride species B, which in turn effects reversible syn hydropalladation to the alkenylarene to afford the η3-benzyl complex C. [15] Electrophilic fluorination of C with Selectfluor affords the Pd(IV)-F dication D (or its η3- complex). Reductive elimination then forms generic product E, and regenerates A.
This mechanism is supported by the various experiments, part of them described in Scheme 3. The hydrofluorination of 1b with the deuthero-labeled silane (Et3SiD) leads to (2-2H)-2b and confirms that this reagent serves as the hydrogen atom source. The formation of the labeled alkene (2-2H)-1b (<1%) and dideutero product (2,2-2H2)-2b indicates that the hydropalladation event is reversible during the course of the reaction (Scheme 3a). No deutherated product was observed when the reaction was performed in deutherated MeCN, hence the solvent is not serving as a hydrogen donor. For stereochemical assignment stilbenes (E)- and (Z)-1y were subjected to hydrofluorination using Et3SiD because both syn- and anti-(2-2H)-2y could be prepared through independent synthesis. (Z)-Stilbene led to 21% of anti-(2-2H)-2y along with trace amounts of syn-(2-2H)-2y (Scheme 3b). The stereochemical course of hydrofluorination under palladium catalysis is consistent with the studies of Gagné and co-workers [16] who reported C(sp3)-F bond formation with XeF2 proceeding with retention of stereochemistry from putative Pt(IV)-F intermediates.
Additional experiments were carried out with isolated η3-benzyl complexes (Scheme 5). We could not isolate either C or D (Scheme 2), or its η3-analogue, but found that the η3-benzyl complex C1 ligated with dppp (1,3-bis(diphenylphosphino) propane) led to 2a in 50% yield upon treatment with Selectfluor in MeCN at room temperature (Scheme 4a). Moreover, the enantioenriched naphtylethyl complex C2, prepared from (R)-T-binap according to the protocol of Hartwig et al.,[17] afforded (S)-2-(1-fluoroethyl)naphthalene (2c) in 80% ee at low temperature (82% yield) in the presence of XeF2 (Scheme 4b). By using 1.2 equivalents of Selectfluor at room temperature, 2c was formed in 36% yield (19F NMR spectroscopy) and a depleted ee value of 36%. The sense of enantiocontrol for the major enantiomer remained S (Scheme 4c). These data indicate that the stereochemical outcome for the fluorination is consistent with coordination of F+ leading to the Pd(IV)-F complex and C-F bond-forming reductive elimination. The stereoselctivity observed under palladium catalysis is consistent with the formation of a Pd(II) complex resulting from syn hydropalladation and subsequent oxidative fluorination to generate the Pd(IV)-F complex D. Retention of configuration in the C-F coupling step can then afford a product resulting from net syn hydrofluorination. Notably, the reaction of 2-vinylnaphtalene with 10 mol% of C2, 3.0 equivalents of XeF2, and 1.5 equivalents of Et3SiH in CH2Cl2 at -20 °C delivered, within one hour, enantioenriched (S)-2c (34% yield, 36% ee). This transformation is the first catalytic enantioselective hydrofluorination (Scheme 4d).
In summary, with this project we have disclosed a new palladium-catalysed HF addition to alkenes. The completely regioselectivity and cis specificity, make this reaction an attractive method for the hydrofluorination of alkenylarenes. The mechanism of this reaction involves an ionic pathway and is distinct from known hydrofluorinations involving radical intermediates. The first catalytic enantioselective hydrofluorination is also disclosed.
