Triphenylphosphine-catalyzed Alkylative Iododecarboxylation with Lithium Iodide under Visible Light

Photoactivation of an electron donor–acceptor encounter complex in an organic solvent cage, a phenomenon that has been described in Mulliken theory, has been known for decades, but it has not been employed as a photoactivation step in the design of photocatalysis for organic synthesis until recent years. We report herein an iododecarboxylation reaction that applies this concept for photoactivation by using a catalyst to facilitate electron transfer and to suppress back electron transfer in the photoexcited state. Under irradiation of 456 nm blue light-emitting diodes, PPh3 catalyzes the iododecarboxylation of aliphatic carboxylic acid-derived N-(acyloxy)phthalimide with lithium iodide as iodine source. The reaction delivers primary, secondary, and bridgehead tertiary alkyl iodides in acetone solvent, and the alkyl iodide products were easily used to generate C–N, C–O, C–F, and C–S bonds to allow various decarboxylative transformations without using transition-metal or organic dye-based photocatalysts. This protocol is applicable to redox-active esters derived from various natural products and pharmaceuticals.


Background
Decarboxylative transformations 1-4 that convert a carboxylate group into a functionality that is a versatile handle for various further transformations, such as the recent development of alkylative decarboxylative borylation 5,6 , are of great signi cance in organic synthesis. Alternatively, a low-cost alkylative decarboxylative iodination reaction has similarly high merits for use with aliphatic carboxylic acids in organic synthesis. The importance of decarboxylative halogenation of aliphatic carboxylates in organic synthesis is re ected in several named reactions, such as the Hunsdiecker reaction 7 , the Kochi reaction 8 , and Barton halogenative decarboxylation 9 (Fig. 1A). Decarboxylative halogenation generates organohalides, which are among the most versatile building-blocks in modern organic synthesis [10][11][12] . Alkyl iodide represents the most reactive electrophile among the various alkyl halides; thus, e cient decarboxylative iodination offers a platform for a variety of decarboxylative transformations 13 from easily available carboxylic acids. Decarboxylative iodination of aromatic carboxylic acids and their derivatives has been extensively reported 14,15 . Notably, a recent report by Larrosa and co-workers 16 delineated an e cient transition-metal-free decarboxylative iodination of arene carboxylic acid using molecular iodine. However, reports on e cient decarboxylative iodination of aliphatic carboxylic acid derivatives remain relatively rare because alkyl iodides are susceptible to various metal catalysts and nucleophiles. Reported examples of decarboxylative iodination of aliphatic carboxylic acid derivatives include halogenative decarboxylation of Barton esters with iodoform 17 , oxidative methods using iodobenzene diacetate and molecular iodine under light 18 , decarboxylation of bridgehead carboxylic acid with t-BuOCl and HgI 2 19 , and a recent photoredox method using iridium photoredox catalyst and Niodosuccinimide that proceeds with moderate yields 20 . The propensity of iodine cations to undergo electrophilic substitution with arenes presents another challenge in chemoselective decarboxylative iodination for complex molecules containing electron-rich arene moieties 21 . Decarboxylative iodination using an iodide salt under mild redox neutral conditions would thus be useful for direct functionalization of complex substrates. We posited that our recently reported strategy of photoactivation of transiently assembled chromophores composed of a redox-active ester (RAE), an iodide salt, and a triarylphosphine for alkyl radical generation, provides an expedient method for alkylative decarboxylative iodination 22 . The principle of radical generation is based on the photoactivation of an electron donor-acceptor (EDA) encounter complex in an organic solvent cage to generate free-radical ions after diffusion 23 (Fig. 1B).
Solvation and noncovalent interactions between substrates play crucial roles in determining the productive photoactivation and subsequent diffusion process [24][25][26][27] . This principle of photoactivation can be utilized to design a catalytic cycle for bond formation with a catalyst that facilitates electron transfer from a donor moiety to an acceptor moiety and to suppress undesired back electron transfer to induce further fragmentation of radical ion species 28,29 (Fig. 1B). Herein, we implement this hypothesis to design a decarboxylative iodination of aliphatic carboxylates.
As depicted in Fig. 1C, an iodide salt, triphenylphosphine, and an RAE transiently assemble to form a chromophore as indicated by the light-yellow appearance of the solution and absorbs up to blue visible light in the UV-Vis spectrum (see Supplementary Information for details). The EDA encounter complex can be a transiently assembled species in a solvent cage that is held through weak, noncovalent interactions [30][31][32][33] , and hence is not isolable. Photoactivation of the EDA encounter complex results in an electron transfer process that forms an • I-PPh 3 radical and a phthalimide radical anion, which triggers subsequent decarboxylation to deliver an alkyl radical. It is worth mentioning that, although similar UV-Vis absorption spectra were observed in the absence of PPh 3 (see Supplementary Information for more details), the presence of PPh 3 is crucial to suppress back electron transfer from the phthalimide radical anion to • I by forming thermodynamically stable • I-PPh 3 , and to prevent formation of I 2 , which was found to be detrimental. The alkyl radical reacts with • I-PPh 3 to produce alkyl iodides and regenerate PPh 3 ; hence, the reaction is catalytic in PPh 3 . We document herein a PPh 3 Fig. 2, showed that as the cation radius of the alkali metal increases (from Li to Cs), the yield of 2 gradually decreases. Protonation by-product 3 was detected in signi cant amounts when RbI or CsI was used. The conversion of 1 dramatically decreased when CaI 2 was used, resulting in a low yield (46%) of 2. The use of either ZnI 2 or n-Bu 4 NI as iodine source was entirely ineffective. These results revealed a signi cant cation effect and suggest that the cations affect assembly of the chromophore in a solvent cage, and hence affect the rate of electron transfer and subsequent radical decarboxylation. Solvation heavily in uences the extent to which a transiently assembled EDA encounter complex 18 can affect the reaction outcome (second row of Fig. 2). Amide solvents such as DMF and DMA primarily resulted in the formation of decarboxylative protonation products. Acetonitrile (MeCN), ethyl acetate (EtOAc), dichloromethane (DCM), and tri uorotoluene (PhCF 3 ) were all unsuitable solvents. Tetrahydrofuran (THF) was a suitable solvent, whereas the use of dioxane as solvent gave no product. A mixed THF/acetone solvent system appeared to be optimal. The remarkable and subtle solvent effect was also demonstrated by testing ketone derivatives as solvent.
Acetone is an optimal solvent, and using butan-2-one gave reduced yield. In sharp contrast, the use of nonan-5-one or 3-methylbutan-2-one as the solvent resulted in no decarboxylative transformation.
The mild redox-neutral conditions of the protocol encouraged us to test synthetic modi cations of a series of RAEs derived from natural products and pharmaceuticals. As shown in Fig. 4, RAEs derived from linoleic acid (28), oleic acid (29), erucic acid (30), and undecenoic acid (31) smoothly underwent decarboxylative iodination with the stereochemical integrity of the alkene moieties remaining intact. RAEs derived from medicinal compounds and complex natural products, such as pregabalin (32,33), mycophenolic acid (34), gabapentin (35,36), dehydrocholic acid (37), chloroambucil (38), baclofen (39), estrone (40), and lithocolic acid (41) also reacted smoothly to deliver the corresponding iodides. The relatively low yield of chloroambucil (38) could be partially explained by a competitive Finkelstein reaction. It is worth noting that the unprotected phenolic hydroxyl in mycophenolic acid (34) is compatible. For estrone analogue 40, the electron-rich phenyl ring, which is susceptible to electrophilic halogenation, remained unaffected. The alkyl iodides derived from these natural products and pharmaceuticals are suitable for introduction into bioactive structure motifs to construct complex molecules or for further diversi cation.
The reaction did not work well for obtaining ordinary tertiary iodide, probably because of the low bonddissociation energy of the tertiary alkyl-I bond and because of its tendency to generate a tertiary carbon cation 34 . Testing RAE derived from gem brozil resulted in the formation of a mixture of alkene regioisomers (equation 1) with no product of decarboxylative iodination detected.
Radical cyclization experiments unequivocally proved that the reactions proceed through a process involving free alkyl radicals (Scheme 1). Reactions using 1.5 equivalent of LiCl or LiBr in the presence of 10 mol% of LiI did not produce any decarboxylative chlorination or bromination product (equation 2), suggesting that a carbon cation is not oxidatively generated with • I-PPh 3 .

Scheme 1 Radical cyclization experiments.
Simply treating the obtained cyclic secondary alkyl iodides with 1,8-diazabicyclo [5.4.0]undec-7-ene (DBU) at 60 °C in one pot generated the alkene products in high yields (Fig. 5, 49-53). Decarboxylative elimination of these carboxylic acids was recently reported using either photoredox/synergistic catalysis 35,36 or palladium catalysis under light 37 . Facile access to alkenes expands the synthetic utility of this low-cost and mild decarboxylative iodination protocol.
The products of decarboxylative iodination can be further used to construct C-O, C-N, C-F, and C-SCN bonds, allowing their subsequent use without requiring the expensive transition metals [38][39][40][41] . As exempli ed by the reaction of oleic acid, shown in Scheme 2, a gram-scale decarboxylative iodination reaction produced the corresponding iodide in 80% yield. The primary iodide underwent S N 2-type reaction with oxygen, uorine, nitrogen, and sulfur nucleophiles to generate ethers (54, 55), uoride (56), amine (57, 59), and thiocyanate (58). Subsequent formation of C-N and C-SCN bonds can also be achieved in one pot in good yields without isolation of alkyl iodides (

Conclusions
In summary, a PPh 3 -catalyzed iododecarboxylation protocol for use with aliphatic carboxylates and lithium iodide under irradiation with blue light has been developed. The reaction uses lithium iodide as iodine source, proceeds under mild, redox-neutral conditions, and hence is suitable for modi cation of complex natural products and pharmaceuticals. The activation principle of this protocol is based on the photoactivation of an EDA encounter complex in a solvent cage, and a catalyst for this process facilitates electron transfer and suppresses back electron transfer. The protocol has the advantages of low cost and simplicity, and allows versatile follow-up transformations to be applied, thereby expanding the use of aliphatic carboxylic acids in organic synthesis. Competing nancial interests.

Methods
The authors declare no competing nancial interests.

Figure 1
Iododecarboxylation using PPh3 and MI. A, Traditional decarboxylative halogenation reactions. B, Concept of the photoactivation of an electron donor-acceptor encounter complex in a solvent cage for radical generation that can be used in the design of photocatalysis systems. D, electron donor substrate or moiety; A, electron acceptor substrate or moiety; CT, charge transfer; ET, electron transfer. C, Working hypothesis for triphenylphosphine-catalysed decarboxylative iodination.

Figure 2
Key reaction-controlling parameters for decarboxylative iodination. The yield was determined by 1H NMR spectroscopic analysis using diphenylmethane as internal standard.