A catalytic borylation/dehalogenation route to o-fluoro arylboronates.

A two-step Ir-catalyzed borylation/Pd-catalyzed dehalogenation sequence allows for the net synthesis of fluoroarenes where the boronic ester is ortho to fluorine. Key elements of this approach include the use of a halogen para to the fluorine to block meta Ir-catalyzed borylation and the chemoselective Pd-catalyzed dehalogenation by KF activated polymethylhydrosiloxane (PMHS).

Fluorinated arenes regularly emerge as lead candidates for pharmaceutical,[1,2] agrochemical,[3] and materials applications. Also common to these fields is the use of arylboronic ester building blocks.[4,5] As such, preparations of arenes bearing both fluoro and boronate substituents are highly desirable. Despite this need, to the best of our knowledge and as indicated by a SciFinder Scholar search, fluorinations of arylboronic esters are unknown. In contrast, boronate substituents have been introduced in fluorinated arenes and heterocycles by various methods, including directed deprotonation and metal–halogen exchange.[6] Such reactions typically demand the use of strong lithium bases and/or cryogenic conditions. Activation of a properly positioned halogen by palladium,[7] C–X borylation, represents a milder approach but demands the regioselective installation of the halogen. Given the substrate regioselectivities in aromatic halogenations and/or Sandmeyer–Schiemann protocols, accessing suitable haloaromatic starting materials can be either trivial or prohibitively difficult. Indeed, for many of the parahalofluorobenzenes used in this study (vide infra) the corresponding isomers where the halogen (Br or Cl) is ortho to fluorine are more expensive or commercially unavailable.

Ir-catalyzed C–H borylations can obviate the need for strong bases, cold temperatures, and/or halogen prefunctionalization. They tolerate numerous functional groups, including fluorine, and have thus been used to generate many fluorobenzenes bearing a 4,4,5,5-tetramethyl-1,3,2-dioxaborolane (Bpin) group.

The regiochemical course of these reactions is primarily driven by sterics, making it possible to install Bpin’s ortho to hydrogen, cyano, or fluorine vs larger aryl substituents.[8] This feature is useful but can also create challenges. For example, under standard conditions, borylations of 3-substituted (or 2,3-disubstituted) fluorobenzenes typically afford ∼1:1 mixtures of borylated arenes (Scheme 1).[9]

Scheme 1

Typical C–H Borylation of 3-Substituted Fluorobenzenes

The indiscriminate nature of Ir-catalyzed C–H borylation in Scheme 1 has made C–X borylation via halogenated starting materials the preferred option for selective borylations of 3-substituted (or 2,3-disubstituted) fluorobenzenes. For selective borylations ortho to the fluorine this requires the acquisition of 6-halo-3-substituted fluorobenzenes. As stated earlier, the availability of arenes with such a substitution pattern is highly dependent on the nature of the substituent C-3 (and that at C-2). In fact, in certain instances the availability and/or costs of arenes with a bromo or iodo substituent positioned ortho to the fluorine make them unattractive starting materials, while analogous arenes with halogens para to the fluorine are more readily sourced.

We hypothesized that readily available 3-substituted fluoroarenes with a halogen (X = Cl, Br, I) para to the fluorine could serve as convenient starting materials for the generation of o-borylated products. Owing to the remarkable halogen tolerance of Ir-catalyzed C–H borylations, X would not serve as an activating group for metalation but rather could be a sacrificial blocking group in a C–H borylation. In this way, borylation would only take place ortho to the fluorine and upon removal of X the desired ortho borylated 3-substituted fluoroarene would be generated (Scheme 2).

Scheme 2

Alternate Approach to o-Fluoroarylboronates

To begin testing this hypothesis a variety of haloarenes were reacted with 1 mol % of [Ir(OMe)(COD)]2, 2 mol % of 4,4′-di-tert-butyl-2,2′-dipyridyl ligand (dtbpy), and 0.55 equiv of bis(pinacolato)diboron (B2Pin2) in THF at room temperature (Scheme 3). Except where otherwise noted, all of these arenes selectively afforded the o-borylated products in good yields.

Scheme 3

Borylation of Fluoroarenes

Isolated yields.

Borylation run at 60 °C for 36 h; product contains 3% of the Bpin para to R1 isomer.

Product contains 1% of the Bpin para to R1 isomer.

Product contains 4% of the Bpin para to R1 isomer.

Borylation run with 0.5 mol % of [Ir(OMe)COD]2, 1 mol % of tmp, and 3.0 equiv of HBpin at 80 °C for 16 h.

Borylation run at 80 °C for 14 h after which 0.25 equiv of HBpin was added and the reaction was allowed to continue at 80 °C for 10 h; product contains 9% of the Bpin para to N.

A key goal was to dehalogenate the borylated arenes without compromising the Bpin group. Radical based methods are not suitable for that task, but despite the potential for unwanted Suzuki couplings, a few such Pd-mediated reductions have been reported. Among these, Pd/C-mediated transfer hydrogenation using ammonium formate as an in situ hydrogen donor was attractive owing to the mild and low cost nature of the reagents.[10] Unfortunately, aside from anisoles, such reductions were almost always accompanied by 5–15% loss of the Bpin group as well as other unidentified impurities in our hands (Scheme 4).[11]

Scheme 4

Dehalogenation with Ammonium Formate

We next turned to our own experience with the hydrodehalogenation of 3-chloro-5-methylphenylpinacolborane using fluoride-activated polymethylhydrosiloxane (PMHS)[12] in the presence of catalytic polysiloxane encapsulated Pd(0) nanoclusters.[13] To see if we could build from this lone example, the borylated fluoroarenes in Scheme 3 were subjected to 4 equiv of PMHS, 2 equiv of aqueous KF, and 5 mol % of Pd(OAc)2 in THF (Scheme 5).[11]

Scheme 5

Dehalogenation with PMHS

Isolated yields of arylboronates.

Combined yield of the 2.4/1 borylated/deborylated material was 60%.

Via the chlorinated starting material; product contains 1% of the Bpin para to R1 isomer per the starting material.

Product contains 4% of the Bpin para to R1 isomer per the starting material.

Product contains 1% of an unidentified fluorinated product and 1% starting material by 19F NMR.

Most substrates responded favorable to these reduction conditions, affording the desired products in 60–90% yield after 4–5 h reaction times and with no evidence of deborylation. Electron deficient arenes tended to undergo hydrodehalogenation slightly faster than electron rich arenes. The method was amenable to heterocycles as borylated 5-chloro-2-fluoropyridine underwent hydrodehalogenation in 1 h using only 2 equiv PHMS.

Although rare, the methylbenzoate example in Scheme 5 illustrates that protiodeborylation can intrude on the dehalogenation of some substrates. In an attempted to overcome this problem, 18-crown-6/KF in a water free reaction was explored.[14] This met with limited success as hydrodehalogenation times increased due to low KF solubility and other unidentified products were observed by 19F-NMR.

The dehalogenation shown in Scheme 6 indicates that the electronic influence of the fluorine is what heightens the propensity toward protiodeborylation. Here the diborylated arene partially lost the Bpin group ortho to fluorine, while the meta Bpin remained completely intact.

Scheme 6

Hydrodebromonations were generally more facile than hydrodechloronations. We were able to exploit this differential reactivity and selectively remove Br in the presence of a Cl by reducing the amount of PMHS to 2 equiv, which also resulted in increasing the reaction time (Scheme 7).

Scheme 7

Selective Debromonation

We also screened Pd/C (10 wt %) as a palladium source (Scheme 8). Employing 5 mol % of Pd/C (with respect to Pd weight) gave the corresponding hydrodehalogenated product but required 24 h to reach full conversion vs 4 h with Pd(OAc)2. We attribute this time difference to the proficiency with which Pd(OAc)2 forms polysiloxane-encapsulated Pd(0) nanoclusters, which have higher catalytic activity.[11]

Scheme 8

Hydrodehalogenation with Pd/C

Lastly, we investigated performing the Ir-catalyzed borylation and the Pd-catalyzed hydrodehalogenation in a single pot (Scheme 9). Again, longer reaction times were required to see full conversion during the dehalogenation step. This may be due to formation of the Pd(0) nanoparticles being slowed by the residuals from the borylation reaction. Nonetheless, the one-pot yields for the substrates tested were comparable to the combined yields observed over two steps.[15]

Scheme 9

One-Pot Borylation/Dehalogenation

In summary, we have demonstrated a solution to the problem of selectively generating arylboronic esters ortho to fluorine via Ir-catalyzed C–H borylations when both the ortho and meta positions are sterically accessible. Furthermore, as para-halogenated fluorobenzenes are often more available and/or less expensive than their ortho counterparts, this protocol can be competitive with Pd-catalyzed borylations. Finally, we have demonstrated that telescoping the borylation and hydrodehalogenation into a single reaction flask is viable.