DEAROMATIVE ALKYLATION OF HETEROARENIUM SALTS VIA BLUE LIGHT ENABLED HOMOLYTIC C–C BOND CLEAVAGE

Abstract

Alkyl substituted nitrogen heterocycles are important building blocks in drug molecules, agrochemicals, and materials. Herein we report a simple and efficient way to prepare such compounds by coupling alkyl halides with heteroarenium salts. Detailed experimental and computational studies indicate that this reaction goes through a radical coupling and avoids formation of undesired homocoupling of heteroarenium salts via recycling such dimers through C-C bond cleavage under blue light.

Partially and fully saturated heterocycles such as piperidines, dihydropyridines and tetrahydroisoquinolines are frequently present in natural products and pharmaceuticals.^1–3Such abundance of these six membered nitrogen heterocycles in pharmaceuticals and agrochemicals makes them an appealing topic of study. Traditional methods of accessing partially and fully saturated heterocycles involve cyclization or cycloaddition approaches such as Pictet-Spengler, Bischler-Napieralski, and hetero Diels-Alder reactions.^4–10These methods often have specific stereoelectronic requirements that limit their application.

Cross-coupling reactions have revolutionized organic synthesis and pharmaceutical chemistry.^11–14Traditional coupling reactions require a nucleophilic and an electrophilic partner. The scope of C–C bond forming cross-couplings is limited by the availability of carbon nucleophiles such as organometallic or organoboron reagents. These nucleophiles may require prior synthesis and can be sensitive to reaction conditions.¹⁵ Cross-electrophile coupling presents a practical and step-economical alternative that avoids the preparation and handling of sensitive nucleophiles.^{11, 16–19}However, as both reaction partners are electrophiles, often irreversible formation of homocoupling products is observed.^{15, 20–23}Thus, a significant effort is spent to design reactions which avoid homodimerization by taking advantage of differential reactivity of the two electrophiles that are being coupled (Scheme 1A). In this regard, scope of these reactions could be improved if the homodimerization products could be recycled back into the cross-coupling reaction path.

Scheme 1.

A.Cross-coupling and homocoupling of electro-philes; B. Ni catalyzed cross coupling of aryl iodides with heteroarenium salts.

Recently, our group reported a nickel catalyzed dearomative coupling of heteroarenium salts with aryl iodides (Scheme 1B).²⁴ The reaction proceeds through heteroarene dimers that were converted to the cross-coupled product under the blue light and in the presence of a nickel catalyst. This is a novel reactivity mode as dimerization of one of the reaction partners could be reversed to increase the amount of cross-coupling product. To take advantage of this finding and expand the scope of electrophiles that could be coupled with heteroarenium salts we further investigated the reactivity of these dimers. Herein, we report results of these studies and successful coupling of alkyl halides with heteroarenium salts.

To understand how the dimers derived from heteroarenium salts are recycled into the cross-coupling process, we studied their behavior under the blue light. Phenanthridine derived dimer 1 which was formed via reduction of methyl-phenanthridinium iodide by zinc was chosen for these studies, due to its increased stability over pyridine and (iso)quinoline derived dimers (Scheme 2). Dimer 1 was obtained as a mixture of diastereomers and structures of both diastereomers were confirmed by NMR and X-ray crystallography. When a solution of single diastereomer of the dimer 1 (1a) in deuterated benzene was exposed to blue light, equilibration between the dimer diastereomers 1a to 1b was observed (Scheme 2A). A time-course study of this process was performed at two different temperatures, which showed that the rate of the isomerization was independent of temperature. At both 80 °C and 30 °C diastereomer 1b formed at a similar rate reaching equilibrium after approximately one hour and settling at the ratio of 5:2 (1a:1b). Under dark conditions heating 1a to 80 °C did not result in its conversion to diastereomer 1b. This indicated that under blue light conditions dimer 1 diastereomers interconvert presumably through formation of radical 2. In order to trap radical 2, we exposed dimers 1 to blue light in the presence of AIBN at 80 °C (Scheme 2B). AIBN is known to decompose to 2-cyanoprop-2-yl radical which then reacts with dihydrophenantradinyl radical 2 to give 3 in 60% yield. It is worth noting that without the blue light, reaction of the AIBN with the dimer did not lead to any detectable cross coupling product 3. Computational studies indicate that dimer 1 can undergo initial excitation to corresponding excited singlet state followed by intersystem crossing to a triplet state which then leads to its fragmentation to radical 2 (see Supporting Information for details). These experimental and theoretical evidence support the fact that dimers such as 1 exist in equilibrium with radical 2, thus acting as a radical reservoir for these radicals. Reports of direct homolytic C—C bond cleavage are rare and are achieved by thermal control.^25–27 Recently, similar light mediated reactivity of dihydroacridine dimers have enabled regenerative photocatalysis.²⁸

Scheme 2.

A. Isomerization of diastereomer 1a under blue light; B. Reaction of dimer 1 with AIBN under dark and blue light conditions. CCDC registry number for 1a and 1b is 2377422 and 2377423 respectively.

Based on the results obtained above we envisioned that various radicals generated under reductive conditions can be coupled with dihydroheteroaryl radicals generated under the same conditions. We hypothesized that if the cross-coupling reaction of such dihydroaryl radicals were slow, dimers such as 1 will form, but under blue light conditions these can still generate dihydroheteroaryl radicals and improve the yield of cross-coupled product.

Alkyl halides can react with metallic zinc to generate corresponding alkyl radicals.^29–33 Initial attempts for the coupling of alkyl halides with heteroarenium salts were carried out with presence of zinc. Under these conditions excellent conversion of the starting material was observed, however only 12% of the methyl-isoquinolinium iodide 4 and isopropyl iodide in the product was dihydroisoquinoline (DHI) 5, with the remaining 82% being undesired DHI dimer 6 (Table 1, entry 1). Higher concentrations of isopropyl iodide increased the yield of 5 up to 43% (Table 1, entries 2,3), however significant dimer formation was still observed. To our delight in the presence of blue light, the yield of the DHI 5 was increased to 88% and no homodimerization product 6 was observed (Table 1, entry 4).

Table 1.

Optimization of reaction conditions for cross-coupling isoquinolinium salt 4 and isopropyl iodide.

With the optimal reaction conditions in hand, both alkyl and azaheteroarene substrate scope of the reaction were studied (Scheme 3). Overall, the reaction had good functional group compatibility, tolerating ester, nitrile, ether, Bpin, double and triple C—C bonds, and alcohol functional groups. Coupling of unactivated tertiary, secondary, cyclic secondary, and primary alkyls with phenanthridinium salt was achieved in moderate to high yields (8a-f). These systems require the use of alkyl iodides as reagents except for 8b. Coupling of ethyl bromodifluoroacetate resulted in difluoro substituted product 8b in 93% yield. Unactivated alkyl bromides resulted in trace reactivity. Activated alkyl bromides such allyl (8g and 8h) and propargyl (8i and 8j) bromides gave high yields of coupling products. In the case of propargyl bromides the rearranged allene containing product was observed, the yield of which depended on the structure of the propargyl bromide used. While addition of propargyl bromide resulted in alkyne 8i as the major product, coupling of 1-bromo-2-butene resulted in formation of the allene 8j as the major product in 77% yield.

Scheme 3.

Substrate scope for the coupling of alkyl halides with heteroarenium salts.[a]

Benzyl bromides with various substitution patterns were successfully coupled with the phenanthridinium salt in excellent yields (8k-t). Substrates with electron withdrawing or electron donating groups all coupled efficiently. Piperonyl and 2-methyl naphthalene fragments were also successfully coupled with the salt (8l and 8p). Lastly, benzyl bromide containing a primary alcohol was coupled, although in slightly lower yield (8s). Sterically hindered ortho-substituted and secondary benzyl bromides coupled with high efficiency (8r and 8t respectively). Benzyl bromide was coupled with various azaheterocycles to explore the scope of heterocycles that can be used in this reaction. Dearomative coupling of ester substituted pyridines with various substitution patterns usually resulted in mixtures of regioisomers (8u-8ab). Pyridines containing an ester group in the 3-position resulted in addition to the 4-position selectively (8u, 8v-y, and 8aa) unless the C4-position had a steric hindrance, in which case C6 selective addition was observed(8z). C4-ester substituted pyridinium underwent addition at the C4 position exclusively (8ab). In the case of 4-(trifluoromethyl) pyridine, C2 and C4 substituted regioisomers were obtained in a 5:3 ratio, C2 being the major product (8ac). Isoquinolines and a quinoline were also successfully coupled with benzyl bromide in high yields (8ad-af). Since the immediate dearomatization products were unstable to air, these products were further reduced and were isolated as corresponding tetrahydro(iso)quinoline products Isopropyl iodide was also coupled with methyl isoquinolinium salt and was characterized as a dihydroisoquinoline product (8ag). While unsubstituted N-alkylpyridiniums did not react under the reaction conditions, activation of pyridines with a benzoxazole group resulted in dearomatized products in moderate yields. Addition of allyl bromide into the benzoxazole activated pyridine resulted in 2-substituted dihydropyridine (8ah), while benzyl addition resulted in 4-substituted dihydropyridine (8ai).

Further mechanistic studies were carried out to understand the nature of alkyl species that participate in this reaction (Scheme 4). A series of radical clock experiments were performed to probe for the formation of radicals and reaction kinetics. In the case of substrate cyclopropylmethyl iodide, rearrangement and ring opening of the cyclopropane occurred to give 10 in 54% yield. In contrast, direct coupling without rearrangement was observed with 6-iodo-1-hexene to give 11, suggesting that the rate constant for the coupling is between 1.3 × 10⁸ (s⁻¹) and 1.0 × 10⁵ (s⁻¹).³⁴ The observation of rearrangement for cyclopropylmethyl iodide suggests a radical formation of alkyl halides in the presence of Zn. As expected, the preformed alkyl zinc species Et₂Zn reacted with 7a to give 9. However, the reductive coupling reaction displayed no sensitivity to water or acidic functional groups, suggesting that organometallic species are unlikely to be present. When 5.0 equivalents of water were added to the reaction mixture, no decrease in yield was observed (Scheme 4 formation of 8k) and a hydroxyl group was tolerated under the reaction conditions (Scheme 3, 8s). Subjecting dimer 1 to standard reaction conditions with isopropyl iodide and zinc yielded the coupled product in 90% yield. Reaction of the dimer with Et₂Zn resulted in recovery of the starting material only. Finally, reaction of the dimer with isopropyl iodide in the absence of Zn gave 8c in 26% yield along with salt 7a in 15% yield.

Scheme 4.

Mechanistic Studies.

Based on these results, a proposed mechanism for the reactions is shown in Scheme 5. Reaction of Zn with alkyl halides 12 can produce either an alkyl radical 13 or an alkylzinc reagent 14 (Scheme 5A). Alkyl zinc species 14 can directly react with heteroarenium salt to give alkyl dihydroheteroaryls 15 (Scheme 5B). However, since the reaction tolerates water and acidic functional groups, we believe that the formation of alkyl zinc reagents is unlikely. On the other hand, direct reductive Minisci-type addition of alkyl radicals to the heteroarenium salt 7a could lead to cation radical 16 which is then reduced to give 15. We propose that this reductive Minisci path³⁵ is a possible path for activated alkyl halides as these rapidly generate alkyl radicals. Addition of these radicals does not require blue light, and we never observed formation of dimer 1 for these reactions. Competing reductive dimerization of heteroarenium salt 7a to give 1 is observed with alkyl iodides. Our experiments described above indicate that these dimers under blue light are in equilibrium with radical 2 which can recombine with alkyl radicals 13 to give the cross-coupling product 15. Alternatively, radicals 2 can participate in formation of alkyl radicals from alkyl halides through a halogen atom abstraction process (Scheme 5C). Such reactivity can explain conversion of dimer 1 to product 8c in the absence of zinc. Computational studies indicate that such halogen atom abstraction is feasible with a reaction barrier of 15 kcal/mol, although the process is thermodynamically uphill (Scheme 5D, for details, see Supporting Information).

Scheme 5.

Proposed mechanisms for the cross coupling of heteroarenium salts and alkyl halides.

A method for the coupling of alkyl halides with heteroarenium salts in the presence of blue light was developed. The reversible nature of homodimerization products of heteroarenium salts under blue light enables opportunities for expanding the scope of cross-electrophile couplings for heteroarenium salts. The radical clock experiments and compatibility with water support the hypothesis of radical pathway for the alkyl partner. Additionally, trapping of the heteroarene monomer with AIBN suggests formation of radical intermediate for the heteroarene partner. Generation of another diastereomer from an isolated diastereomer showcases the cleavage of the dimer C—C bond in the absence of any substrate or metal. Furthermore, the applicability of the method to various alkyl and heteroarene substrates was demonstrated.