Proton-Catalyzed, Silane-Fueled Friedel-Crafts Coupling of Fluoroarenes

See allHide authors and affiliations

Science  29 Apr 2011:
Vol. 332, Issue 6029, pp. 574-577
DOI: 10.1126/science.1202432

Putting the F in Friedel-Crafts

The Friedel-Crafts class of reactions, among the oldest and most broadly applied in organic chemistry, form carbon-carbon bonds between aromatic rings and a variety of non-aromatic substituents, such as alkyl groups. Generally, a metal complex is used to activate chlorinated or brominated precursors of these substituents, but by using silicon-based reagents to activate a fluorinated precursor, Allemann et al. (p. 574) extend the reaction to coupling of two different aromatic sites, leading to efficient formation of elaborate polycyclic structures. The method relies on the unusual strength of silicon-fluorine bonds as a driving force.


The venerable Friedel-Crafts reaction appends alkyl or acyl groups to aromatic rings through alkyl or acyl cation equivalents typically generated by Lewis acids. We show that phenyl cation equivalents, generated from otherwise unreactive aryl fluorides, allow extension of the Friedel-Crafts reaction to intramolecular aryl couplings. The enabling feature of this reaction is the exchange of carbon-fluorine for silicon-fluorine bond enthalpies; the reaction is activated by an intermediate silyl cation. Catalytic quantities of protons or silyl cations paired with weakly coordinating carborane counterions initiate the reactions, after which proton transfer in the final aromatization step regenerates the active silyl cation species by protodesilylation of a quaternary silane. The methodology allows the high-yield formation of a range of tailored polycyclic aromatic hydrocarbons and graphene fragments.

Formation of carbon-carbon bonds via Friedel-Crafts methodology is one of synthetic organic chemistry’s charter reactions, involving the attack by electron-rich π-clouds of aromatic rings onto carbocation-like substrates (1, 2). This reaction is of fundamental importance to industry (3), and its mechanistic principles also pertain to many organometallic transformations (4). With respect to aryl-aryl couplings, Friedel-Crafts chemistry is limited by the thermodynamic instability of the phenyl cation (5); however, the importance of such couplings has mothered the invention of a plethora of noble and coinage metal–mediated methodologies (6, 7). These highly useful aryl-aryl couplings typically involve an activated nucleophilic aryl partner, an aryl bromide or iodide, and a transition metal catalyst (810). Omission of the transition metal, selection of an aryl fluoride, and use of an unactivated aryl nucleophile are not contemplated by any reactive scheme currently in the chemist’s repertoire (11, 12). Therefore, the efficient intramolecular coupling of an arene with an aryl fluoride by carbon-fluoride bond activation constitutes an avenue of synthetically useful Friedel-Crafts reactivity ripe for exploration.

Carbon-fluorine bond activation by silyl cations has been applied to assist hydrodefluorination of benzyl and alkyl fluorides (13, 14). In addition, silyl cations are instrumental in supporting aliphatic carbon-carbon bond formation (15). The mechanism is ascribed to the formation of well-known alkyl cation equivalents, which can be trapped in classical Friedel-Crafts reactions. In contrast, aryl carbon-fluorine bonds appeared sufficiently inert in these studies such that aryl-fluorides could be used as solvents without interference; formation of phenyl cation was not anticipated (13, 14).

The driving force for the reaction comes from the substantial Si–F bond strength relative to that of the C(aliphatic)–F bond. Notably, the strength of the Si–F bond exceeds even that of the C(aryl)–F bond by about 120 kJ mol–1 (16, 17), and fluoride abstraction from fluoroarenes by silyl Lewis acids was recently demonstrated (18). Given these findings, a silylium-promoted Friedel-Crafts arylation appears feasible in terms of bond enthalpies and reaction kinetics.

In a brute-force approach, one would combine the organic substrates with a stoichiometric amount of a silyl cation to effect fluoride abstraction, and a Brønsted base to neutralize the Wheland intermediate (B, Fig. 1), which arises from attack by the nucleophilic arene on the incipient aryl cation. This strategy necessitates the availability of a full equivalent of a sufficiently reactive silylium-like species. Indeed, addition of triisopropylsilyl carborane (1) (19, 20) to a substrate such as 1-(2-fluorophenyl)naphthalene (2) effects the activation of the carbon-fluorine bond and formation of a new arene-arene bond, leading to the product fluoranthene (3).

Fig. 1

Suggested catalytic cycle of the fluoride abstraction and the subsequent intramolecular attack, using the example of transformation of 1-(2-fluorophenyl)naphthalene to fluoranthene.

An equivalent of one proton is generated during the reaction; the neutralization of this charge by a sterically hindered Brønsted base prevents undesired polymerization reactions. A moment’s reflection reveals that this proton need not be simply scavenged, but rather can be channeled into a productive part of the reaction. Protodesilylation of arylsilanes to yield an equivalent of a silylium ion and an arene is a well-known phenomenon relying on the greater bond enthalpies of C–H bonds over those of Si–C bonds (2123). Given that the silyl cation–initiated Friedel-Crafts arylation formally generates one equivalent of HF, the proton formed after the first C(aryl)–C(aryl) coupling can serve as a catalyst to generate a silyl cationic species from a neutral silane precursor. Overall, the exchange of covalent bonds [C(ar)–H, C(ar)–C(ar), and Si–F for C–(ar)–H, Si–C(ar), and C(ar)–F] is thermodynamically highly favorable. Thus, an elegant self-sustaining reaction system comprises the arene or aryl fluoride (as substrate) and an arylsilane (as fuel that consumes protons and releases silyl cations), plus a substoichiometric amount of a silyl cation (as initiator).

Substoichiometric use of the preformed silyl cation species described above only provides a shunt into this reaction manifold. The proton generated during the reaction (for example, as [fluoranthene-H]+[carborane]) is the actual catalyst. Arenium carboranes—protonated arenes paired with carborane counterions—are a young class of superacids that can be isolated in pure form (24). Thus, direct addition of an arenium carborane, instead of the silylium initiator, should effect a reaction wherein a proton serves as the catalyst to transform a neutral silane into an intermediate silyl cation. This cation then activates the carbon-fluorine bond, thereby inducing arene-arene coupling and release of a proton to recommence the process.

The substrate 1-(2-fluorophenyl)naphthalene (2) served as a model substrate for developing suitable Friedel-Crafts reaction conditions. Upon activation of the C–F bond, 2 undergoes ring closure to furnish fluoranthene (3). Initially, conditions involving stoichiometric amounts of silyl cation and a sterically hindered Brønsted base were investigated (25). Performing the reaction with iPr3Si–CB11H6Cl6 and either P(o-tol)3 or 2,6-di-tert-butylpyridine at 110°C for 8 to 15 hours led to high conversions and isolated yields of 3 (Table 1). Chlorobenzene proved to be a suitable solvent for this transformation and afforded better results than toluene. The more Lewis-basic toluene may deactivate the silylium carborane by π coordination (26). Intermolecular aryl-aryl coupling between toluene and the incipient phenyl cation was not observed.

Table 1

Conditions for C–F activation applied on 1-(2-fluorophenyl)naphthalene. Reactions were all conducted with 0.1 mmol of 1-(2-fluorophenyl)naphthalene under permanent inert atmosphere. The product was purified by flash column chromatography. The conversion was measured by GC-MS; the isolated yields are reported only for conditions that led to complete conversion. Mes = 2,4,6-trimethylbenzene.

View this table:

The ring closure also proceeded effectively with substoichiometric silyl cation as initiator in combination with a stoichiometric amount of silane. This approach led to isolated yields of up to 93% using iPr3Si–CB11H6Cl6 (1) and Me2SiMes2 (4). It is possible to carry out the reaction either with preformed or in situ generated cations (Table 1), the former giving slightly better yields. An attempt to run the reaction at room temperature yielded only small quantities of product.

Triggering the reaction with mesitylenium carborane (5) (a very strong Brønsted acid) resulted in complete consumption of starting material and high isolated yield of product. Triisopropylsilylium (1) and mesitylenium ion (5) act as initiators of the catalytic cycle (Fig. 1). They start the catalytic cycle at a different state (B and C) but eventually lead to the same active species, the mesityldimethylsilyl cation (D).

Counterions play a significant role in reactive cation chemistry; chemical stability and low nucleophilicity are preferred characteristics for clean cation reactivity. Two classes of anions, perfluorinated tetraphenylborates and carboranes (27, 28), are very weakly coordinating anions with widely distributed negative charge (29). In the present case, the carborane [CHB11H5Cl6] afforded superior results to those with [B(C6F5)4], consistent with earlier reports showing a higher robustness of carboranes toward high temperature and Brønsted acidity relative to the perfluorinated phenylborates (24, 30). High acidity is an inherent component of the acid-catalyzed reaction mechanism, and heating is necessary for the product to form within an adequate period of time.

It is likely that C–F activation does not take place by a completely dissociative mechanism, because C6H4R+ would have a very unfavorable electronic configuration (31). Irrespective of the mechanistic details, after ring closure, the resultant protonated aryl moiety B is evidently acidic enough to protonate the electron-rich mesityl ring of dimethyldimesitylsilane. Elimination of mesitylene reforms the silyl cation (D) and completes the catalytic cycle.

An early assessment of the scope of the reaction shows its applicability to the synthesis of a variety of polynuclear aromatic hydrocarbons (Table 2). It is effective in couplings involving fullerene fragments and could be an alternative to Scholl chemistry toward the formation of graphene and other carbon-based mesogens (32, 33). Reductive activation of carbon-halogen bonds, other than fluorine, has been used toward this goal of forming graphene subunits, but often steric crowding prohibits the formation of the necessary precursors. The small steric profile of the fluorine atom and the need to have only one derivatized partner adds substantial value to this methodology for preparing tailored graphene models.

Table 2

C–F activation reaction on various substrates; conditions from entry 5 of Table 1 were used for these transformations.

View this table:

For the formation of less strained five- or six-membered ring systems, conditions were adapted to obtain yields between 49 and 99%. In contrast, it has not yet been possible to generate the four-membered ring of biphenylene (7) from substrate 6 by applying this method. Terphenyl educts 8, 10a, 10b, and 12 showed full conversion to products 9, 11a, 11b, 13a, and 13b using optimized conditions (Table 1, entry 5). The yields were between 97 and 99%. Substrates 14 and 16 were more difficult to transform to the desired products 15 and 17. Additional silylium carborane and silane, plus extra hours of stirring, were required to achieve a higher yield. Treatment of substrate 18 afforded a yield of 51%; the lower yield was probably caused by decomposition of the product (19).

The terphenyl substrates containing a methyl group on the nonfluorinated phenyl ring (10b and 12) showed a mixture of ortho and para products. The distribution was slightly in favor of the para position, which should be more accessible; however, the small preference for the para methyl indicates a very reactive intermediate that does not distinguish strongly between the two positions.

Additional studies enabled us to make a qualitative statement on the reaction rate. The terphenyl substrates (8, 10a, 10b, and 12) were used in competitive C–F activation reactions. Two substrates were added to the same reaction mixture with only half an equivalent of reagent. Gas chromatography–mass spectrometry (GC-MS) consistently showed that substrates containing methyl groups (at any of the two rings) reacted faster than did the electron-poorer rings without substituents. The tendency of reaction rates for different substrates was found to be 8 < 10a10b < 12 (25). This observation matches with the proposed transition state A. Its positive charge would be stabilized by an increased electron density of the fluorophenyl ring, provided by an additional substituent such as the methyl group. Methylation of the donor arene had the same effect on the reaction rate.

For the transformation of substrate 2, quantum mechanical calculations predicted a transition state for the fluoride abstraction (Fig. 2) that nearby aryl moieties can stabilize (18, 25). The structural and activation parameters predicted by the quantum chemical model (calculated Ea = 21.6 kcal mol–1; transition-state interatomic lengths, C–F = 2.55 Å, C–C = 2.86 Å) fit well with the experimental findings (34). The model further supports the observation that a more electron-rich arene accelerates the reaction.

Fig. 2

B98/DZ+(2df,pd) calculated transition state (TS), including effects of chlorobenzene solvent, for fluorine abstraction from 1-(2-fluorophenyl)-naphthalene (calculated Ea = 21.6 kcal mol–1; transition-state interatomic lengths, C–F = 2.55 Å, C–C = 2.86 Å).

The proton is likely the universe’s oldest catalyst, and it is now available as a crystalline arenium carborane (25). Coupled to a neutral silane as fuel, arenium acids are competent catalysts for the long-sought phenyl cation–based Friedel-Crafts reactivity. Such C–F activation for the formation of arene-arene bonds complements transition metal–based arene-arene couplings, particularly for the formation of designed graphenes and higher-order polynuclear aromatic hydrocarbon–based materials. Understanding the science of silyl cations and phenyl cation–like intermediates at play here, along with the design of tailored precursors, will certainly lead to a useful expansion of the synthetic chemist’s tool box and the material chemist’s objects of investigation.

Supporting Online Material

Materials and Methods

Table S1

References 37 to 45

References and Notes

  1. The action of SbF5 on perfluorinated naphthalene forms radical cations capable of coupling with pentafluorinated benzene, albeit with limited substrate scope; see (35).
  2. Flash vacuum pyrolysis appears to facilitate the lysis of C(aryl)–F bonds and thereby promote arene-arene coupling; see (36).
  3. See supporting material on Science Online.
  4. Specific coordination of solvent or counterion is estimated to have an effect of <5 kcal/mol on activation parameters, such that the model is still consistent with a reaction at 80°C over several hours.
  5. Acknowledgments: We thank the Swiss National Science Foundation for financial support.
View Abstract

Navigate This Article