1,3,2-Diazaphospholene-Catalyzed Reductive Cyclizations of Organohalides.

1,3,2-diazaphospholenes hydrides (DAP-Hs) are highly nucleophilic organic hydrides serving as main-group catalysts for a range of attractive transformations. DAP hydrides can act as stoichiometric hydrogen atom transfer agents in radical reactions. Herein, we report a DAP-catalyzed reductive radical cyclization of a broad range of aryl and alkyl halides under mild conditions. The pivotal DAP catalyst turnover was achieved by a DBU-assisted σ-bond metathesis between the formed DAP halide and HBpin, which rapidly regenerates DAP-H. The transformation is significantly accelerated by irradiation with visible light. Mechanistic investigations indicate that visible light irradiation leads to the formation of DAP dimers, which are in equilibrium with DAP radicals and accelerate the cyclization. The direct use of (DAP)2 enabled a catalytic protocol in the absence of light.

Discovered in the 1980s, 1,3,2‐diazaphospholenes (DAPs) represent a class of electron‐rich heterocycles with a great application potential as versatile main‐group catalysts. For, instance, diazaphospholene hydrides (DAP‐Hs) are powerful nucleophiles and weakly basic organic hydride donors. They efficiently reduce carbonyls and, in conjugate fashion, α,β‐unsaturated amides, esters,[ , ] ketones, and acids (Scheme 1). A σ‐bond metathesis between the alkoxy‐substituted DAPs and pinacol borane (HBpin) regenerating DAP‐H rendered these processes catalytic. Reports by Kinjo and Speed exploited σ‐bond metathesis between the P−N and B−H bonds of ammonia borane and HBpin, respectively to enable catalytic reductions of azobenzenes and imines.[ , ] Cheng demonstrated the cleavage of the P−F bond in DAP fluorides by employing phenylsilane. This enables hydrodefluorination of trifluoromethylalkenes and polyfluoroarenes. The reactivity of DAP‐Hs is not limited to the two‐electron transfer processes. Reductions of alkyl and aryl halides and α‐carboxy ketones via radical pathways have been reported. Radical processes, where DAPs are involved catalytically are scarce. Only a single methodology describing a DAP‐catalyzed deoxygenation of α‐carboxy ketones is reported. It capitalizes on the σ‐bond metathesis between the P−O and B−H bonds for the catalyst's regeneration. In contrast to the DAP‐F, the related halide bearing DAP‐X (X=I, Br, Cl) does not undergo a σ‐bond metathesis with borane or silane reagents. This reactivity gap hampers the closure of catalytic cycles. To exploit the full potential of DAPs in radical chemistry with organohalides, it is essential to render the transformations catalytic.

Scheme 1

a) DAPs in catalysis: Regeneration of DAP‐H to close catalytic cycles. b) Landscape of reduction options for DAP‐Y into DAP‐H and the underlying catalytic processes c) DAP‐catalyzed reductive radical cyclization of organohalides.

Radical cyclizations of organohalides across olefins are highly useful transformations to access diverse cyclic skeletons. Typically, the reductive cyclization of organohalides operating through a radical mechanism requires stoichiometric amounts of toxic organostannanes and radical initiators such as AIBN or Et3B/O2. Advances in photochemical methods provided some sustainable alternatives including the use of metal‐ and organic photocatalysts as well as electron‐donor reagents. The high affinity of DAPs towards organohalides makes them very attractive for activating this broad class of compounds. Speed demonstrated the functionalization of organo‐ iodides and bromides. However, catalytic variants of these processes remain so far elusive. The rapid and efficient regeneration of the DAP‐H from DAP‐X under mild conditions would be an essential gateway to catalytic transformations with organohalides. Herein, we report an efficient Lewis base‐promoted regeneration of DAP‐H from DAP‐X and its application in the catalytic reductive radical cyclization of aryl and alkyl halides.

To close the catalytic cycle, we first investigated the regeneration of the pivotal DAP‐H from the formed DAP‐I of the stoichiometric process. Mixing DAP‐I and HBpin in CD3CN does not generate DAP‐H after 16 h (Scheme 2). We hypothesized that the activation of the borane by a suitable Lewis base could facilitate the σ‐bond metathesis between DAP‐I and HBpin. To our delight, adding DBU to the DAP‐I/HBpin mixture in CD3CN triggered a fast and quantitative conversion to DAP‐H. The same reactivity was maintained with DAP‐Br.

Scheme 2

Fast regeneration of DAP‐H with the DBU/HBpin system.

With the rapid regeneration of DAP‐H from DAP‐X, enabled by the DBU/HBpin system, we turned our efforts towards a fully catalytic transformation for the cyclization of aryl iodide 1 a and aryl bromide 2 a (Table 1). Employing 5 mol% of the robust and conveniently usable secondary phosphine oxide SPO as the pre‐catalyst with HBpin/DBU in MeCN, 1 a afforded 4 a in 23 % yield (entry 1). We noticed that the reaction rate and progress are highly sensitive to light. While still taking place under the exclusion of light (entry 2), the yield of 4 a almost quadrupled when irradiating the reaction with white LEDs (entry 3). This effect was even more pronounced with aryl bromide 2 a. However, no reaction took place in the absence of light (entry 4). The use of white LEDs restored some reactivity giving 4 a in 12 % yield (entry 5). Switching to a more powerful Kessil lamp (427 nm) increased the yield of 4 a to 91 % (entry 6). Elevated temperatures also promoted the cyclization to 4 a, albeit in low efficiency (27 % yield, see Supporting Information). Control experiments omitting SPO, DBU and HBpin showed their indispensable role (entries 7–9). Aryl chloride 3 a did not cyclize (entry 10).

Table 1

Optimization of the DAP‐catalyzed cyclization of 1–2 a.[a

entry	substrate	variation	yield [%][b]
1	1 a	ambient light	23
2	1 a	exclusion of light	19
3	1 a	white LEDs	89
4	2 a	exclusion of light	0
5	2 a	white LEDs	12
6	2 a	Kessil lamp (427 nm)	91
7	2 a	Kessil lamp (427 nm), no SPO	0
8	2 a	Kessil lamp (427 nm), no DBU	5
9	2 a	Kessil lamp (427 nm), no HBpin	0
10	3 a	Kessil lamp (427 nm)	0

[a] Conditions: 0.1 mmol 1 a, 5 μmol SPO, 0.11 mmol HBpin, 0.1 mmol DBU, 0.1 M (0.8 M for 2 a and 3 a) in MeCN at 26 °C for 16 h, [b] Yields determined by 1H NMR with an internal standard.

Next, we evaluated the scope of the DAP‐catalyzed reductive cyclization for a broad set of substituted aryl halides (Scheme 3). Dihydrobenzofurane 4 a was isolated in excellent yields (92 % from 1 a and 96 % from 2 a). The transformation is tolerant to potentially reactive functional groups like methyl esters, free phenols, and carboxylic acids and reliably delivered products 4 b–4 g. Notably, a switch from the Kessil lamp to white LEDs allowed for a selective cyclization originating from the aryl iodide moiety of 1 j in the presence of the adjacent bromide substituent, in 97 % yield. Along the same lines, the reaction of aryl bromide 2 h cleanly proceeded under standard conditions in the presence of an adjacent chloride moiety. The reaction of the bromopyridine 2 i cyclized to product 4 i without DAP‐catalyzed reduction of its pyridine core. Aryl bromides bearing different alkenyl tethers reacted well. For instance, styrene 2 k underwent cyclization in 86 % yield. The nature of the olefin acceptor portion was as well modifiable (4 l–4 n). Substrates, that pass through 1° or 2° alkyl radical intermediates, reacted in reduced yields (4 l and 4 m). Besides the formation of the five‐membered rings, the process enabled the 6‐exo‐trig cyclizations as demonstrated for 4‐isopropylchromane 4 o. Aniline substrates were readily converted into functionalized indolines 4 p–4 r. Next, we explored the potency of alkyl halide substrates to undergo cyclization. In this respect, iodo‐ 1 s and bromo‐acetals 2 s smoothly reacted to bicyclic product 4 s. Primary alkyl bromides 2t and 2 u cyclized to the corresponding cyclopentane 4 t and tetrahydrofuran 4 u in moderate yields. The cyclization of substituted alkenyl amines led to pyrrolidines 4 v–4 w in good yield. A secondary alkyl bromide analogue formed disubstituted pyrrolidine 4 y in 91 % yield and 65 : 35 dr. Noteworthy, substrates bearing a trichloroalkyl group engaged in the transformation providing dichloro‐substituted pyrrolidine 4 z and piperidine 4 za in 82 % and 69 % yield, respectively.

Scheme 3

Scope of the DAP‐catalyzed cyclization of organohalides. Conditions: 0.20 mmol 1  –3  , 10 μmol SPO, 0.22 mmol HBpin, 0.20 mmol DBU in MeCN (1  =0.2 M, 2  –3  =0.8 M) in MeCN at 26 °C for 16 h. [a] white LEDs instead of the Kessil lamp (427 nm). [b] 2.5 equiv HBpin. [c] 10 mol % SPO, 2.2 equiv. HBpin and 2.0 equiv. DBU.

Mechanistic investigations were conducted to better understand the effect of light in this transformation. The stoichiometric reactions indicate that the cyclizations of 1 a and 2 a take place in the dark as well. However, irradiation by the Kessil lamp accelerates the transformation (Scheme 4). In contrast, the catalytic reaction with substrate 2 a occurs only under irradiation. Under catalytic conditions, aryl iodide 1 a still cyclized to 4 a in the dark, albeit with poor efficiency. The direct reaction of DAP‐H and CCl4 does not require activation by light affording DAP‐Cl and a mixture of chloromethanes CH Cl4− (m=0–3). This P−H/C−Cl bond metathesis proceeds eventually through a radical mechanism. Substrate 3 z was exposed to one equivalent of DAP‐H under the exclusion of light yielding 75 % of 4 z. To gain mechanistic insight into the light‐enhanced reaction, we first identified which species absorb light in the visible spectrum. The absorption spectra of DAPs and 2 a indicate that only DAP‐H and DAP‐Br absorb light at wavelengths above 400 nm, an emission tail of the Kessil lamp (427 nm). No ground state associations between the DAPs and 2 a were found (see Supporting Information). Based on this information, two scenarios in which visible light accelerates the cyclization of 2 a are plausible. In the scenario I, a photoexcitation of DAP‐H triggers a SET from the excited DAP‐H* to 2 a. Scenario II involves a photoexcitation of DAP‐H leading to a DAP radical (DAP⋅). Based on electrochemical and spectroscopic measurements, we estimated the redox potential of the excited DAP‐H* is −3.14 V (vs. Fc+/Fc in MeCN), indicating that a SET between DAP‐H* and 2 a (E red 2 a/2 a .−=−3.36 V) is endergonic (see Supporting Information). Then, we examined the generation of DAP⋅ radical species by excitation of DAP‐H. We hypothesized that visible light from the Kessil lamp could trigger a reductive dimerization of DAP‐H. This was reported by Gudat using UV light. The resulting (DAP) species, described as a weakly σ‐bonded dimer, would dissociate in solution to give the persistent radical DAP⋅. Indeed, 31P NMR analysis confirmed the formation of (DAP) from DAP‐H upon irradiation by the Kessil lamp (Scheme 5). This evidence supports the proposed initiation of a radical chain process through (DAP).

Scheme 4

Evaluation of the light influence on the cyclization efficiency and absorption spectra of the reaction components.

Scheme 5

Illustration of the role of (DAP) in the DAP‐catalyzed cyclizations.

Next, we tested the influence of (DAP) on the cyclization reaction rate of 2 a with one equivalent of DAP‐H in the dark. Notably, already 2 mol% of (DAP) remarkably accelerates the cyclization. This observation supports the outlined scenario II. Both findings led to the hypothesis that an exchange of the SPO catalyst by (DAP) would enable a catalytic process without visible light‐promoted activation. Remarkably, 5 mol% (DAP) promoted a catalytic reaction in the absence of light forming 4 a in 70 % yield. Under irradiation, equimolar amounts of DAP‐H afforded the cyclized product 4 a in comparably high yield. Under the exclusion of light, 10 mol% DAP‐H was capable to generate 4 a in 27 % yield. Since the formation of (DAP) is not observed under these conditions, the involvement of an alternative ionic pathway can be conceived. The conversion of 1 a to 4 a with stoichiometric amounts of DAP‐H in the dark eventually supports this.

Initial‐rate kinetic studies inferred a first‐order dependence for DAP‐H and a half order for substrate 2 a, supporting a radical chain mechanism (Scheme 6, see Supporting Information for details). The zeroth‐order dependence of HBpin and DBU indicates that the regeneration of the DAP‐H is not a turnover‐limiting step. The reaction rates increase with the light intensity (see Supporting Information). Deuterium labelling studies with 1 a using DAP‐D in CH3CN or DAP‐H in CD3CN confirmed the origin of the hydrogen atom of 4 a from the catalyst.

Scheme 6

Initial‐rate kinetics for the DAP‐catalyzed cyclization and deuterium labelling studies.

Taking all mechanistic experiments into account, we propose the following catalytic cycle (Scheme 7). With SPO as the pre‐catalyst, the process is initiated by the reduction with HBpin forming DAP‐H. Visible light irradiation converts DAP‐H to (DAP). The dissociation equilibrium of (DAP) into two molecules of DAP⋅ initiates a radical chain process by bromine atom abstraction from 2 a. The resulting aryl radical I adds across the C=C bond in a 5‐exo‐trig fashion forming radical species II. In turn, II abstracts the hydrogen atom from DAP‐H delivering product 4 a and DAP⋅. Enabled by DBU, DAP‐Br is converted back to DAP‐H with HBpin. The use of (DAP) as catalyst allows entering the catalytic cycle bypassing the light activation step. However, the reduced yield of the (DAP)‐catalyzed reaction in the dark indicates that light can heal the catalytic cycle by regenerating DAP⋅ after radical chain terminations.

Scheme 7

Proposed mechanism of the DAP‐catalyzed cyclization of organohalides.

In summary, we developed a DAP‐catalyzed reductive radical cyclization of organohalides. The DAP catalyst turnover was achieved by a DBU‐assisted σ‐bond metathesis between DAP‐X (X=I, Br, Cl) and HBpin, which provided a fast regeneration of DAP‐H. The transformation is significantly accelerated by the irradiation with visible light. The developed process allowed the efficient reductive cyclizations of a broad range of aryl and alkyl halides under mild and convenient conditions. Detailed mechanistic investigations revealed that visible light leads to the formation of (DAP) which is in equilibrium with DAP⋅ and accelerates the cyclization. The direct use of (DAP) enabled a catalytic protocol in the absence of light. These findings will serve as blueprint and accelerator for the further developments of DAP‐catalyzed radical processes.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.