Cross-coupling of organic fluorides with allenes: a silyl-radical-relay pathway for the construction of α-alkynyl-substituted all-carbon quaternary centres

Controlling the transformation of versatile and reactive allenes is a considerable challenge. Herein, we report an efficient silylboronate-mediated cross-coupling reaction of organic fluorides with allenes to construct a series of sterically demanding α-ethynyl-containing all-carbon quaternary centers (ACQCs), using catalyst-free silyl-radical-relay reactions to selectively functionalize highly inert C–F bonds in organic fluorides. The key to the success of this transformation lies in the radical rearrangement of an in situ-generated allenyl radical to form a bulky tertiary propargyl radical; however, the transformation does not show efficiency when using the propargyl isomer directly. This unique reaction enables the cross-coupling of a tertiary carbon radical center with a C(sp2)–F bond or a benzylic C(sp3)–F bond. α-Ethynyl-containing ACQCs with (hetero)aromatic substituents and benzyl were efficiently synthesized in a single step using electronically and sterically diverse organic fluorides and allenes. The practical utility of this protocol is showcased by the late-stage functionalization of bioactive molecules and the modification of a liquid crystalline material.


Introduction
All-carbon quaternary centers (ACQCs) exhibit rigidity and structural diversity and are key structural units that occur frequently in many natural products, pharmaceuticals, and bioactive molecules. 1Moreover, at least 12% of the 200 topselling prescription drugs in the US since 2011 contain a stereo-quaternary carbon center. 2 Therefore, the construction of ACQCs presents a quite attractive challenge for organic synthetic chemists. 3In particular, ACQCs that contain an alkyne moiety serve as versatile intermediates and basic functionalized groups in organic transformations. 4Generally, the synthetic methods for alkyne-containing ACQCs involve either the direct introduction of alkynyl moieties into target molecules 5 or the transformation from halogenated allenes in the presence of Knochel reagents. 6owever, the aforementioned methodologies are associated with major drawbacks, including the use of transition-metal (TM) catalysts, dimerization of terminal alkynes, b-H elimination of branched tertiary alkyl units, and/or the reliance on special functionalized precursors.Therefore, the development of efficient strategies to overcome these limitations and the extremely strong steric effect to realize alkynyl-substituted ACQCs through C-C bond coupling reactions remains challenging.
C-C bond formation is a perpetual subject of interest in organic chemistry and represents one of the most important transformations for the manufacture of products used in daily life. 7The majority of C-C bond coupling reactions require TM catalysts and/or organic (pseudo)halides (Ar/alkyl-X; X = e.g., I, Br, Cl, OTf, OMs). 8However, organic uorides are rarely employed as coupling partners because the C-F bond is rather inert and has a higher bond dissociation energy (BDE) (e.g., uorobenzene: 126 kcal mol −1 ; 1-uoropropane: 114 kcal mol −1 ) compared with the corresponding C-I/Br/Cl bonds. 9dditionally, with the rapid development of synthetic methodologies, 10 the abundance and ready accessibility of organo-uorine compounds 11 make them attractive functional moieties as well as building blocks for further organic transformations.
Yet, the activation of robust C-F bonds remains a major challenge in contemporary chemistry.In this context, TM catalysis has proven a promising strategy for the direct functionalization of otherwise unreactive C-F bonds via the oxidative addition of C-F bonds to TMs followed by selective functionalization, providing access to the desired deuorinated molecules. 12However, the selective activation of C-F bonds usually either suffers from high oxidative-addition barriers, thus the employment of highly elaborate TM catalysts 13 and/or forcing conditions 14 is oen indispensable, or conned to multi/poly-uorinated arenes 15 (Fig. 1A).Therefore, an alternative method that efficiently bypasses the high barriers required for the progress of the oxidative addition and that promotes the smooth coupling of C-F and C-H bonds under mild conditions would make a great contribution to this eld.
Specically, we envisioned that in situ-generated silyl radicals would quickly abstract a proton from targeted C-H reactants to form the corresponding C-centered radicals, thus enabling subsequent catalyst-free transformations with organic uorides via the activation of an inert C-F bond to realize C-C coupling products (Fig. 1B).Such silyl-radical-relay reactions have already been demonstrated by our recently reported TMfree silylboronate (R 3 SiBpin)-mediated cross-couplings of organic uorides with alkenes/arylmethanes (Fig. 1C). 16ecently, the unique structural features and versatile reactivity of allenes enabled them to be not only a versatile functional group that is incorporated in natural products, pharmaceuticals, and organic materials, 17 but also serve as an ideal platform for the development of new methodologies in synthetic transformations, chiral ligands or even catalysts. 18In this context, reactions between silylboronates and allenes have been pioneered by the Suginome group 19 and further explored by Stratakis; 20 however, most hitherto reported protocols require TM catalysts and many provide silaboration products with one double bond retained.To date, only the Chen group 21 has reported the successful transformation of vinylidene cyclopropanes into propargylic silanes in the presence of copper(I) chloride and NaO t Bu (Fig. 1D).Nevertheless, the silyl moiety is involved in the nal products.
Based on these previous results, we designed a new silylboronate-mediated cross-coupling of organic uorides with allenes to access ACQCs that feature an a-ethynyl group via C-F bond activation and radical rearrangement at room temperature (Fig. 1E). 16,22Initially, R 3 SiBpin and potassium tert-butoxide (KO t Bu) in an ether-based solvent smoothly generate intermediate A. Owing to the radical-initiation properties of KO t Bu 23 and the steric demand of intermediate A, A splits into the bulky frustrated radical pair B, 24 which consists of a trialkylsilyl radical (cSiR 3 ) and a boron-radical species (Bc), via homolytic cleavage of the Si-B bond.The silyl radical in B then directly abstracts a hydrogen atom from the allene (3) to form allenyl radical-containing frustrated radical pair C, which could easily isomerize to the sterically highly demanding propargylic radical-containing frustrated radical pair D. 25 Subsequently, D would attract organic uoride 1 or 2 by preferential interaction between the F atom and the B center to afford TS-I.Finally, the desired ethynyl-containing product with a quaternary carbon center (4 or 5) would be obtained via C-C bond coupling, accompanied by the release of E ([Bpin(O t Bu)F]K), which would promptly react with another equivalent of KO t Bu to provide a stable [Bpin(O t Bu) 2 ] species and KF.

Silylboronate-mediated cross-coupling reactions of aryl uorides and aryl allenes
As depicted in our mechanistic hypothesis for the proposed C-C coupling (Fig. 1E), we expected that a silyl radical generated from the silylboronate and KO t Bu could abstract a hydrogen atom from the allene terminal.Therefore, we initiated the crosscoupling reactions by using 4-uorobiphenyl (1a) and penta-1,2dien-3-ylbenzene (3a) as model substrates.The desired product, 4-(3-phenylpent-1-yn-3-yl)biphenyl (4aa), which possesses a quaternary carbon center with an a-ethynyl moiety, was obtained in 34% yield under the standard reaction conditions [Et 3 SiBpin (2.0 equiv.),KO t Bu (4.0 equiv.),THF, room temperature; entry 1, Table 1].Furthermore, among the silylboronates tested under the same conditions, the Suginome reagent (PhMe 2 SiBpin) improved the yield of 4aa to 41% (entry 2).Subsequent optimization focused on screening the number of reactant equivalents, solvent, and reaction time in the presence of PhMe 2 SiBpin and nally afforded 4aa in 94% yield under the optimized conditions (entry 3; for details, see the ESI; Tables S1-S4 †).Control experiments showed that the reaction did not proceed in the absence of silylboronate or KO t Bu (entries 4).Moreover, the desired product was not obtained using other bases such as KOMe, NaO t Bu, LiO t Bu, or KHMDS (entries 5 and 6).This indicated that the countercations in MO t Bu exhibited superior ability for K + over Na + or Li + in facilitating this transformation may lie in its established capacity to function as a single-electron reductant. 23Decreasing the amount of PhMe 2 SiBpin or KO t Bu resulted in lower yields of 4aa (86% and 51%, respectively; entries 7 and 8).Replacing PhMe 2 SiBpin with Et 3 SiBpin under otherwise identical optimal conditions gave 4aa in only 75% yield (71% isolated; entry 9).Additionally, the use of an inadequate amount of 3a (2.0 equiv.)had a negative effect on the reaction yield (entry 10).
It should also be noted here that the Suginome reagent usually generates the undesired side product 1,2-di-tert-butoxy-1,1,2,2-tetramethyldisilane 26 (same polarity as 4aa), which is formed by the dimerization of the tert-butoxydimethylsilyl radical ( t BuOMe 2 Sic), which renders the purication of 4aa difficult via column chromatography on silica gel.Instead, the use of Et 3 SiBpin afforded pure 4aa, albeit in a lower yield.Furthermore, the slightly elevated reaction temperature (50 °C) could signicantly reduce the consumption of both the Suginome reagent, KO t Bu, and allene 3a to half the amount, with 79% yield (entry 11).However, only a 38% yield was observed when PhMe 2 SiBpin was replaced by Et 3 SiBpin under the same conditions as in entry 11 (entry 12).
Fig. 3 Further scopes and limitations.a Unless otherwise noted, all reactions were conducted using 1 or 2 (0.2 mmol), 3a (43.2 mg, 1.5 equiv.),PhMe 2 SiBpin (78.7 mg, 1.5 equiv.),KO t Bu (67.5 mg, 3.0 equiv.), and diglyme (1.5 mL) at 50 °C for 12 h.Synthetic applications.To highlight the synthetic applications of this silylboronate-mediated deuorinative coupling reaction, some easily accessible functional molecules with aalkynylated quaternary centers were obtained aer several drug derivatives and liquid-crystalline materials were evaluated under the standard conditions (Fig. 4A).Adapalene derivative 4adc with two substituents at the b-position was successfully obtained in 71% yield by coupling b-uoronaphthyl-containing adapalene derivative 1ad with allene 3c.Steroid derivative 4aei was synthesized in 42% yield via the deuorinative coupling of uoro-incorporated estrone derivative 1ae and 3i.Blonanserinderived uoroarene 1af underwent the coupling reaction with 3g to generate Blonanserin derivative 4afg in 89% yield.Moreover, the liquid-crystalline material 1ag was successfully functionalized using this transformation with 3g to give 4agg in 67% yield.The presence of the a-ethynyl group in these derivatives could potentially allow further late-stage functionalization.
Synthetic transformations.It is noteworthy that quenching with different reagents efficiently afforded a variety of products with functional moieties.Specically, the cross-coupling of 1a with 3a under the standard conditions afforded 6 (62% yield) and d-4aa (70% yield) when quenched using p-anisaldehyde and deuterium oxide, respectively.Bromoethynyl derivative 7 (89% yield) was also successfully synthesized when the coupling reaction of 1a with 3g was quenched with N-bromosuccinimide (NBS) in the presence of a catalytic amount of silver(I) nitrate (Fig. 4B, le).Additionally, 4aa can serve as a versatile precursor for further synthetic transformations.First, a gram-scale reaction of 1a and 3a proceeded smoothly to afford 4aa in 68% yield under the standard conditions in the presence of Et 3 SiBpin.Thereaer, a dual catalyst (Pd(OAc) 2 /PPh 3 and CuI) enabled the cross-coupling reaction of 4aa with a substituted arylbromide, 27 which furnished the phenyl-coupling product 8 (95% yield) in the presence of Et 3 N in THF at 80 °C.Treatment of 4aa with trimethylsilyl chloride (TMS-Cl) in the presence of butyllithium furnished TMS-acetylene product 9 (95% yield). 28A Pd/C-catalyzed reduction of 4aa employing acetic acid/NaBH 4 afforded hydrocarbon product 10 in 88% yield (Fig. 4B, right). 29hese straightforward functionalization reactions signicantly expand the scope and utility of these silyl-radical-relay crosscoupling reactions between aryl uorides and allenes.

Chemical Science
Edge Article isomer of 3a, i.e., 3-phenyl-1-pentyne (12).While 3a and 12 are isomers of each other, their pK a values differ signicantly, and 12 should be more reactive under the basic conditions (pK a of 12: H = 28.0;H 1 = 25.7;pK a of 3a: H 2 = 27.0). 30Interestingly, the use of 1a and 12 under the standard conditions (for details, see the ESI †) only afforded 4aa in 28% 1 H NMR yield, and alkynyl adduct 13 was not detected; instead, the corresponding deuorosilylation product, i.e., 4-biphenyltriethylsilane (4-Ph-C 6 H 4 -SiEt 3 ) was detected (Fig. 5B).This result excludes both the anionic pre-allenyl/propargyl-isomerization pathway from 3a to 12 and the anionic S N Ar pathway.In contrast, the predicted bond dissociation energy (BDE) 9 of C-H 2 in 3a (85.4 kcal mol −1 ) was higher than that of C-H (80.6 kcal mol −1 ) but much smaller than that of C-H 1 in 12 (129.3kcal mol −1 ).Thus, radical cleavage of C-H 2 in 3a is preferable to that of C-H 1 in 12, and the low yield of 4aa can be explained by the tertiary propargyl C-H bond in 12 which could be radically removed owing to its low BDE value; however, owing to the high acidity of C-H 1 in 12, C-H 1 is promptly deprotonated by KO t Bu to provide potassium acetylide 12 0 (BDE of C-H in 12 0 : 82.4 kcal mol −1 ).Thus, the generation of tertiary propargylic radicals is slow because of the instability of the generated radical anion isomers.When allene 3a was treated with the Suginome reagent and KO t Bu at 50 °C in diglyme, only 12 was obtained (53% isolated yield).However, self-coupling dimer product 14 was not observed, which might be due to high steric repulsion.Similarly, allene 3g gave the same result using Et 3 SiBpin regardless of whether it reacted at room temperature or 50 °C, that is, 15 (65% isolated yield) without the observation of 16 (Fig. 5C).In addition, a radicalclock experiment employing 1a and a-cyclopropyl substituted allene (3o) was conducted.The conversion of 1a was only 40% (determined by 19 F NMR; for details, see the ESI †), whereas 3o was fully consumed to give a complex mixture (Fig. 5D).
Fortunately, the desired product (4ao) was isolated in 24% yield, which agrees with our 1 H NMR analysis of the crude reaction mixture (for details, see the ESI †).The low yield of 4ao can be explained by the low reactivity of the primary carbon radical 3o 00 generated via the ring-opening process.We then evaluated the effect of (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) on the coupling reaction between 1a and 3a under the optimal conditions in the presence of Et 3 SiBpin.Although 4aa was obtained in a yield of 75% ( 1 H NMR yield) under standard conditions, the yield decreased signicantly when the amount of TEMPO was increased (1.0 equiv. of TEMPO: 33%; 2.0 equiv. of TEMPO: 13%; 4.0 equiv. of TEMPO: 0%; Fig. 5E-i).It should be noted that the premixed silylboronate and KO t Bu were used to prevent TEMPO from being consumed by the reducing reagent silylboronate, followed by the addition of TEMPO and other materials.No desired product 4aa was detected, which indicated that the in situ generated silyl radical should be fully trapped by TMEPO (for details, see the ESI †).ESR experiments were also performed.Since we have already demonstrated the generation of silyl radicals from Et 3 SiBpin and KO t Bu under the same reaction conditions, 16b we tried to nd the radical species derived from allene 3.As expected, the ESR spectrum (tripletriplet) was detected for the reaction of PhMe 2 SiBpin, KO t Bu, and allene 3a in diglyme at room temperature, which was assigned to the spin-adduct of the tert-propargyl radical trapped by 2,4,6-tri-tert-butylnitrosobenzene (TTBNB) (Fig. 5E-ii). 31amely, the hyperne splitting (hfs) constant due to nitrogen (A N ; spin quantum number I = 1) was 1.86 mT, and the small splitting constant due to the two hydrogens at the meta-position of the TTBNB benzene ring (A Hm ; I = 1/2) was 0.089 mT.The gvalue of 2.006 was assigned to a nitroxide-type radical.Although further studies are required to show clear evidence, this observation strongly supports the formation of the propargyl radical (see the ESI † for details).
The kinetic-isotope effect of the C-H/C-D cleavage under ionic conditions is more substantially observed than that by the radical reactions. 32Thus, we evaluated 1a in combination with deuterated allenes (d 2 -3g and d-3g) in several parallel reactions under the standard conditions in the presence of PhMe 2 SiBpin (Fig. 5F).However, independent of the existence of deuterium in allene 3g, the formation of D-or H-4ag depends on the quenching method, i.e., on using H 2 O or D 2 O, even when quenching with ice water.The formed products 4/5 should exist as potassium acetylides in the reaction mixture, as the excess of t BuOK could easily result in a further deprotonation process.Therefore, the acetylide can be captured by D + , H + , Br + , RCHO, etc. (Fig. 5G).Besides, due to the high acidity of the C(sp)-H/D moiety in the terminal alkynyl position of 4ag, the H/D exchange occurs easily during the workup steps. 32,33Interestingly, the reaction time and yield were almost identical independent of the use of 3g, d 2 -3g, or d-3g.Therefore, we concluded that the C(sp 2 )-H/D bond cleavage should occur prior to the C-F bond cleavage, and thus the C(sp 2 )-H/D bond cleavage should not be the rate-limiting step.All observations in the mechanistic study led us to conclude that this deuorinative cross-coupling reaction proceeds via a single-electron transfer (SET)/radical process, in accordance with our mechanistic hypothesis shown in Fig. 1E.

Conclusions
In summary, we have realized the rst cross-coupling reaction of organic uorides with allenes to construct a library of aethynyl-containing all-carbon quaternary centers via C-F bond and C(sp 2 )-H bond radical cross-couplings using a silyl-radicalrelay strategy.The C-F bond cleavage occurs concomitant with the formation of an isomerized propargylic radical, which takes place through cleavage of a C(sp 2 )-H bond, to cooperatively form a new C-C(sp 3 ) bond.A notable feature of this crosscoupling reaction is that the in situ-generated silyl radical is able to directly abstract a proton from a C(sp 2 )-H bond of the allene to form an allenyl radical, which then easily isomerizes to form a propargylic radical that exerts a more profound steric inuence.Signicantly, in this transformation it is not possible to use the corresponding propargyl isomers directly instead of the allenes.This method proceeds under very mild conditions and efficiently obviates the use of TM catalysis or light irradiation to allow a range of para-, meta-, and even ortho-substituted (hetero)aryl uoride, benzyl uoride and allene substrates to undergo the normally challenging deuorinative coupling process to afford all-carbon quaternary centers in moderate-toexcellent yield, with good functional-group compatibility and C-F bond selectivity.This radical-coupling system was further extended to the late-stage functionalization of several biologically active molecules.

Fig. 1
Fig. 1 Cross-couplings of C-F and C-H bonds.(A) Transition-metal-catalyzed coupling reactions of C-F and C-H bonds.(B) A low-barrier cross-coupling of C-F and C-H bonds enabled by silylboronate and KO t Bu. (C) Our previous defluoronative functionalization work involving Ccentered radicals using silylboronates and organic fluorides.(D) Related representative work on silylboronates and allenes.(E) Mechanistic design for the cross-coupling of organic fluorides and allenes under transition-metal-free conditions (this work).

Fig. 5
Fig. 5 Mechanistic study.(A) Chemoselectivites of organic (pseudo)halides Ar-X.(B) An attempt to use an alternative process.(C) Homocoupling attempt of generated propargyl radicals.(D) Radical ring-opening reaction.(E) Effect of TEMPO on this silylboronate-mediated coupling reaction (i) and ESR experiment in the presence of spin trapping reagent TTBNB (ii).(F) Kinetic-isotope-effect experiments.(G) Possible mechanism for the quenching process.