Fluorinated Radicals in Divergent Synthesis via Photoredox Catalysis

-By Derek Lowe

Fluorine’s unique features, including high electronegativity, low polarizability, and small size, have a profound impact on the physical, chemical, and biological properties of molecules. , As a result, fluorinated compounds are widely employed to enhance metabolic stability and lipophilicity, while also finding extensive applications in ¹⁸F imaging and advanced materials. Due to these properties, fluorine is an important component of pharmaceutical development and is often considered as “magic element”. A Reaxys database analysis shows that trifluoromethyl (R–CF₃), where R can be any element, and difluoromethyl (R–CF₂H) compounds are the most common fluorinated building blocks, owing to their synthetic versatility and commercial availability. In contrast, structures such as R–CBr₂F or R–CI₂F are much less common due to synthetic difficulties or instability. Fluorine’s broader impact is evident by its presence in over 20% of pharmaceuticals and 30% of agrochemicals (FigureA). − However, due to the strength of the C–F bond, these compounds are highly resistant to degradation, which also contributes to their toxicity and environmental persistence.

Since naturally occurring organofluorine compounds are rare, recent efforts have focused on developing selective, safe, and controlled fluorination methods to improve their accessibility for academic research and industrial applications. ,, Conventional fluorination often relies on harsh reagents, leading to low selectivity and significant handling challenges, although recent breakthroughs by the Gouverneur group are reshaping the field. − To circumvent these challenges, researchers have developed bench-stable fluorinated functional group transfer reagents (FGTRs) as safer and more attractive alternatives. − Despite their efficiency and versatility, FGTRs are limited by step-intensive preparation. To meet the growing demand in organofluorine chemistry, the focus has shifted toward commercially viable feedstocks (including anhydrides, acids, etc.) to facilitate previously inaccessible transformations while ensuring atom economy and cost-effective production. − The activation of these laboratory commodities using green methodologies, such as photoredox catalysis, has proven beneficial for sustainable transformations, generating reactive radicals under mild conditions to facilitate selective, safe, and controlled fluoroalkylation reactions (FigureB).

Radical reactions are decisive for progress in chemical synthesis as they enable transformations that are difficult or inaccessible with traditional two-electron ionic mechanisms. For example, paradigms such as hydrogen atom transfer (HAT), radical addition to alkenes, and radical-polar crossover (RPC) have opened up new horizons in the selective functionalization of unsaturated hydrocarbons, which are of utmost importance in synthetic organic chemistry. A crucial realization in this context is that controlling reaction intermediates in radical processes allows for selective modulation of product outcomes. By leveraging this principle, precise adjustments of reaction conditions such as solvent, concentration, additives, temperature, light, and catalyst can steer the reaction pathway, leading to the production of a variety of different fluorinated compounds from the same starting materials (FigureC). Such an approach, coined switchable divergent synthesis, is of great value as it takes full advantage of the polyvalent reactivity of key radical intermediates, minimizing the cost and effort to expand the chemical space and making it an attractive tool for the synthesis of multiple products through simple adjustment of reaction parameters.

Over the past five years, we have focused our efforts on developing new methodologies for the direct activation of commercially viable, low-cost sources as redox active reagents, including trifluoroacetic anhydride (TFAA), chlorodifluoroacetic anhydride (CDFAA), α-halocarboxylic acids, and α-halotrifluoroacetones to generate fluorinated radicals. These radicals facilitate fluoroalkyl incorporation onto alkenes, constructing C–C bonds while simultaneously forming C–O, C–N, and C–H bonds. In some instances, they act as bifunctional reagents, enhancing atom economy, which is crucial for the synthesis of complex organic molecules and late-stage diversification. We have showcased that these switchable divergent reactions are scalable and demonstrate excellent chemo- and regioselectivity with the applications in concise synthesis of natural products. Extensive studies combining computational modeling, spectroscopy, and experimental analysis provide deep insights into the reaction mechanism and reactivity of fluorinated radicals.

Given the significance of this field, applied theoretical chemistry provides a rational understanding of the reactivity and nature of radicals, and it is an essential component in enhancing the chemical intuition of experimental chemists. , In this context, the electrophilicity concept introduced by Parr et al. has proven particularly useful in the description and prediction of radical reactivity (e.g., rate enhancement through polarity match). − We have been involved in computational studies to better understand the reactivity of fluorinated radicals , through calculations of their philicity parameters (electrophilicity and nucleophilicity indices) based on De Proft’s method. Shortly thereafter, this concept was extended to various commonly encountered radicals. While the electrophilicity parameter is a useful and straightforward approach to predict chemical reactivity, it remains a relatively simplistic descriptor that can overlook key mechanistic complexities. In contrast, the calculation of the reaction barrier provides a more accurate perspective. Therefore, we elaborated a reactivity scale based on high-accuracy DLPNO–CCSD­(T) calculations of reaction barriers for the addition of fluorinated radicals to benzene, providing a complete picture of their overall reactivity. ,

In this account, we take the opportunity to unify our efforts in switchable divergent synthesis and quantify the reactivity of fluorinated radicals involved through the elaboration of a reactivity scale. We have calculated Gibbs-free energy barriers for the addition of various radicalsincluding those explored experimentally by our group and key radicals as referenceto styrene and compiled the results in FigureD, alongside their electrophilicity (ω) (see SI for computational details↗). As observed in our previous study on radical addition to benzene and consistent with Dolbier’s report, highly fluorinated species rank among the most reactive of the series. The activation energies span approximately 6 kcal·mol^–1, ranging from 9.1 kcal·mol^–1 for ^•CF­(CF₃) radical to 15 kcal·mol^–1 for the ^•CMe₂CO₂H radical. The reactivity window of carboxyl-substituted radicals is relatively broad, highlighting the notable influence of the substitution on reactivity. Among these, ^•CF₂CO₂H is the most reactive toward styrene (11.2 kcal·mol^–1). The reactivity trend with respect to fluorine substitution in these acids follows ^•CF₂CO₂H > ^•CH₂CO₂H > ^•CHFCO₂H (11.2, 13.0, 13.4 kcal·mol^–1, respectively). The −CO₂H moiety has a pronounced effect on radicals electrophilicity, increasing it by approximately 0.5 eV compared to a hydrogen-substituted analogue ω­(^•CF₂H) = 1.20 versus ω(^•CF₂CO₂H) = 1.71). The impact of deprotonation of the −CO₂H moiety on philicity was also assessed for ^•CF₂CO₂M, ^•CHF₂CO₂M, and ^•CH₂CO₂M radicals (M = Li, Na, and K). A decrease in electrophilicity of 0.3–0.5 eV was observed for the Li salts, while a more pronounced drop of 0.6–0.9 eV occurred for the K salts, consistent with the weaker coordination of the larger K⁺ ion with the negatively charged −CO₂^– group (see SI↗). Interestingly, electrophilicity decreases with increasing fluorination in the ^•Rf-CO₂H series: ω­(^•CH₂CO₂H) = 2.01 > ω­(^•CHFCO₂H) = 1.81 > ω­(^•CF₂CO₂H) = 1.71. In contrast, the electrophilicity of bromo- and chloro-substituted radicals (^•CHBrCO₂H, ^•CHClCO₂H and ^•CCl₂CO₂H) remains unchanged (ω = 1.97–2.03), closely resembling ^•CH₂CO₂H. This suggests a significant stabilization of fluorinated radicals via n_F→p_SOMO bonding interactions (see SI↗), a stabilization effect that does not translate efficiently to larger halogens due to their more diffuse orbitals. As expected, steric effects also influence radical behavior, with bulkier substituents leading to reduced reactivity (CH­(ⁱPr)­CO₂H 14.9 and CMe₂CO₂H 15.0 kcal·mol^–1). Finally, a direct correlation between computed reaction barriers and electrophilicity values yielded no meaningful results (r² = 0.09). However, applying a modified Roberts-Steel relationship , similar to our previous work, incorporating radical electrophilicityresulted in a significantly improved correlation (r² = 0.79, see SI↗). This underscores the importance of electrophilicity in the description of radicals’ reactivity, although the electrophilicity value alone is insufficient, particularly for structurally diverse radicals.

An outline of the mechanism of this process is illustrated in FigureF. Our approach relied on photoredox catalysis, specifically leveraging the redox properties of tris­(2-phenylpyridine) iridium­(III) fac-[Ir­(ppy)₃]. Upon excitation by visible light at 440 nmwavelengths where most organic molecules do not absorbfac-[Ir­(ppy)₃] undergoes single-electron transfer (SET) with anhydride, acid, or ketone reagents (1–4, FigureE). Radical I forms via outer-sphere electron transfer, enabled by the strongly reducing excited-state photocatalyst (E_1/2 [Ir^IV/Ir^III*] = – 1.73 V vs SCE) compared to the reagents (E_red = – 1.2 to 0 V vs SCE). Radical intermediate I undergoes a mesolytic cleavage to yield the corresponding radical II, which adds to the olefin forming alkyl radical III. From III, the reaction outcomes can diverge either through direct trapping by HAT reagents to afford hydro-functionalization (radical domain), or subsequent SET oxidation by Ir^IV (E_1/2 [Ir^IV/Ir^III] = +0.77 V vs SCE) to form carbenium ion IV (ionic domain) and regenerates the photocatalyst via a RPC pathway. Afterward, solvent and concentration adjustments allow control over the reactivity of species IV, facilitating difunctionalization through intra- or intermolecular nucleophilic attack.

We now revisit the reactions and reactivity investigated in our previous studies, with particular emphasis on the concept of divergence and the fundamental mechanisms that drive these transformations.

Our journey on fluorinated radicals started with inspiration from prior photocatalytic decarboxylative methodologies, such as those reported by Stephenson and Sodeoka, who demonstrated the generation of fluoroalkyl radicals from perfluoroalkyl anhydrides using an external oxidant (e.g., pyridine N-oxide or urea·H₂O₂) , We hypothesized that trifluoroacetic anhydride (TFAA, FigureA), a cost-effective and widely available reagent, could be directly activated under photoredox conditions and serve as a precursor for trifluoroacetyl radicals (FigureB).

Cyclic voltammetry analysis revealed that TFAA exhibits a reduction onset at approximately – 1.2 V (vs SCE), suggesting its susceptibility to a SET reduction when paired with a suitable photocatalyst. We anticipated that this reductive event would trigger an irreversible C–O bond fragmentation, forming a transient electrophilic ^•COCF₃ radical (ω = 1.75 eV) and a trifluoroacetate anion. Theoretical calculations supported this hypothesis, indicating an exergonic process (ΔG = – 4.8 kcal·mol^–1) favoring the formation of the trifluoroacetyl radical over alternative pathways.

To evaluate this trifluoroacetylation strategy, Ir­(ppy)₃ was chosen as the photocatalyst to reduce TFAA and promote the proposed SET. In the presence of an olefin, this approach efficiently delivered the trifluoroacetylated product 5 in an impressive 79% yield, demonstrating excellent chemo- and regioselectivity. Concentration adjustments further refined the protocol; reducing the substrate concentration to 0.05 M exclusively produced the trifluoromethylated adduct 6 in 85% yield, suggesting that lower concentration favors decarbonylation. This equilibrium between the trifluoroacetyl and trifluoromethyl radicals was further corroborated by high-pressure experiments, where increasing CO pressure (1–10 bar) suppressed trifluoromethylation by forcing the equilibrium toward ^•COCF₃ radical, resulting in nearly exclusive formation of 5 (FigureC). To further elucidate the competition between the two pathways, we have now calculated the reaction barrier for the decarbonylation using the same computational method as in the reactivity scale (FigureD) to enable reactivity comparison (see SI↗). The barrier for decarbonylation is found to be 12.0 kcal·mol^–1 in acetonitrile (11.8 kcal·mol^–1 in EtOAc), which is 0.4 kcal·mol^–1 higher than the barrier for radical addition of ^•COCF₃ to styrene (11.4 kcal·mol^–1 in EtOAc). These results indicate that radical addition is favored over decarbonylation in the case of styrene, yet the small energetic difference allows one to efficiently steer the reaction pathway by varying the reaction concentration. With optimized conditions established, we evaluated the substrate scope across a diverse array of alkenes (Figure). Aryl-substituted olefins bearing electron-donating and electron-neutral groups at various positions underwent trifluoroacetylation in moderate to excellent isolated yields. Scaling the reaction to 15.6 mmol with 4-tert-butylstyrene maintained a 67% yield, affirming its preparative potential. Notably, trifluoromethylated products became predominant when low concentrations (6, 14) or electron-deficient olefins were used (15), resulting from Giese addition in the latter case. This is consistent with a slower radical addition due to the polarity mismatch between the electrophilic ^•COCF₃ radical (ω = 1.75 eV) and electron-deficient alkenes, collaterally favoring decarbonylation. Remarkably, functional groups such as halogens, esters, amides, and acetal remained untouched under the reaction conditions (7-10), offering handles for subsequent synthetic elaboration. The reaction also proved to be compatible with substrates bearing α-substituents and heterocycles, highlighting the method’s versatility. The protocol’s utility can be extended to pharmacologically relevant scaffolds. Key precursors to anti-inflammatory drugs, such as those bearing trifluoromethylated pyrazole moieties (e.g., Mavacoxib analog). Structurally intricate substrates like cholesterol (12) and nitogenin (13) exhibited exclusive functionalization at the less-hindered olefin, affording the corresponding CF₃-enones. Finally, derivatization of these latter enabled the synthesis of trifluoromethylated heterocycles (S-, O-, N-containing, varying ring sizes), challenging to access via conventional routes (16-18). These results collectively demonstrate the method’s robustness and relevance to medicinal chemistry.

The mechanistic hypothesis (FigureA) was then investigated, the irreversible reduction of TFAA at – 1.2 V (vs SCE) confirmed by cyclic voltammetry analysis, Stern–Volmer studies revealed efficient quenching of ³[Ir­(ppy)₃] triplet state by TFAA (k = 2 × 10⁹ mol^–1 s^–1), but not by styrene (FigureB). This excluded energy or electron transfer to the alkene and indicated TFAA activation. Irradiation of Ir­(ppy)₃ in the presence of TFAA produced a broad absorption (500–700 nm) and emission (470–520 nm), consistent with Ir^IV formation, corroborated by spectroelectrochemical data (FigureB). Cyclic voltammetry analysis under irradiation showed depletion of the Ir^III/Ir^IV feature, with a new oxidation peak at 0.6 V vs Fc⁺/Fc (0.98 V vs SCE), indicating irreversible catalyst oxidation. Adding 4-tert-butylstyrene preserved the Ir^III/Ir^IV couple, suggesting regeneration of the Ir^III photocatalyst through oxidation of a radical intermediate.

The fleeting nature of the ^•COCF₃ radical (II), prone to decarbonylation to form ^•CF₃ (VI) radical and CO, was further validated by DFT calculation. These revealed a Gibbs free-energy barrier of ΔG^‡ = 9.5 kcal·mol^–1 (t_1/2 ≈ 0.98 μs, based on Eyring’s theory) and a reaction Gibbs free-energy of ΔG = −3.7 kcal·mol^–1, consistent with a facile and reversible process. While the reduction potential of the benzylic radical III formed after ^•COCF₃ radical addition was calculated to be E_ox,calc = 0.76 V vs SCE, suggesting an essentially isoergonic back electron transfer (BET) to Ir^IV (E_1/2 = 0.77 V vs SCE), the enol tautomerism (IV) which was confirmed through a deuterium labeling experiment, lowered this value to E_calc = 0.43–0.47 V vs SCE, rendering the exergonic process (ΔG = – 7.8 kcal·mol^–1). Subsequent oxidation of intermediate III, followed by deprotonation of the enol carbocation V, resulted in the selective formation of (E)-alkene 5. To better understand the stereoselectivity of this event, the carbenium ion conformation was studied by DFT. Calculations revealed that the minimum energy conformation of this species is indeed that of the (E)-conformer, with an electronic energy about 4.5 kcal·mol^–1 lower than that of the (Z)-conformer (VIII to VII). Additionally, the control experiment supported that photoisomerization occurs during the reaction since subjecting (Z)-enone to blue light irradiation yielded a 2:1 (E/Z)-mixture, suggesting thermodynamic control. The CF₃-product 6, on the other hand, is formed through the addition of VI onto the olefin, followed by a sequence of SET and nucleophilic addition.

In our pursuit to investigate the chemical space of fluorinated anhydrides, with the concept of divergent synthesis at the core of our strategy, we decided to explore chlorodifluoroacetic anhydride (CDFAA) as a compelling bifunctional reagent. This compound is commonly employed as a protecting group and an electrophilic acyl donor for O-, N-, and S-centers (FigureA). The Stephenson and Sodeoka groups have demonstrated that the difluoromethylene functionality (−CF₂-) can be effectively introduced into olefins and heteroarenes through the indirect activation of CDFAA. This strategy utilizes oxidants such as pyridinium N-oxide and perfluorodiacyl peroxides for efficient radical chlorodifluoromethylation.

These findings inspired us to explore the possibility of direct activation of chlorodifluoroacetic anhydride (CDFAA) without relying on oxidants. Nonetheless, we quickly recognized the inherent challenges, as SET reduction could lead to either an acyl radical (^•COCF₂Cl) via C–O cleavage or a gem-difluoroalkyl radical II via C–Cl cleavage. The weak C–O bond poses a risk of decarbonylation to ^•CF₂Cl. However, DFT calculations showed C–Cl cleavage is highly exergonic and irreversible, releasing the gem-difluoro radical and chloride ion in both solvent models (MeCN and DMF). Consistent with this, no acyl product was observed experimentally.

Notably, alkyl radical II has two reactive centers: a radical site and a nucleophilic oxygen atom. With this in mind, we envisioned leveraging the dual reactivity of this bifunctional synthon and hypothesized that in the presence of olefins, it could enable a new atom- and step-efficient synthetic route, leading to the formation of cyclic gem-difluoro compounds. After carefully designing and optimizing the reaction conditions, we screened key parameters such as photocatalyst, concentration, solvents, and other factors. This led to the establishment of conditions enabling the divergent synthesis of three different products. In MeCN, a three-component Ritter-type addition selectively formed lactam (19). In DMF, lactone formation (20) was exclusively occurred. Meanwhile, toluene was found to promote the oxy-perfluoroalkylation reaction (21). This demonstrates that tuning reaction conditions offers control over the selectivity of the product outcomes. The reaction demonstrated broad versatility, which highlights the applicability of the protocol (Figure). The reaction with α,α-disubstituted olefins yielded products 22 and 23. Using CD₃CN or a deuterated olefin gave products 24 and 25. The method proved versatile, producing lactones from disubstituted styrenes (26, 28, 29) while alkynes remained unreacted (27). In toluene, product 30 was exclusively obtained in the reaction with 4-fluorostyrene. The protocol was equally successful for late-stage functionalization, as shown by L-camphanic acids yielding products 31–33 in different solvents. The D-fructopyranose derivative gave 34 and 35 in MeCN and DMF, while non-fluorinated anhydrides afforded 36 and 37. Lactam 19 and lactone 20 were also transformed into valuable compounds 38–41 in a single step, which are otherwise difficult to access.

The strategic reaction mechanism is illustrated in FigureA. At the beginning of the photoredox catalytic cycle, Ir­(III) is the only species capable of absorbing the 440 nm blue LED light, as shown by the UV–vis spectrum (FigureB). Upon excitation, Ir­(III)* undergoes a SET event to form the radical anion intermediate I. This species then undergoes mesolytic cleavage, yielding the gem-difluoro carboxyalkyl radical II and a chloride ion. Computational data indicate this cleavage is highly exergonic in all solvent environments (ΔG_DMF = – 71.5, ΔG_MeCN = – 71.6, and ΔG_Toluene = – 42.3 in kcal·mol^–1). The generated radical II adds to the olefin, yielding the stabilized alkyl radical intermediate III. The SET oxidation of this species by Ir­(IV) generates the corresponding carbocation IV. The slightly negative Hammett constants of – 0.52 for the lactam and – 0.42 for the lactone indicate that substituent effects exert only a modest influence on the overall reaction rate, making it unlikely that the SET oxidationleading to a carbocationic intermediateis rate-determining. Instead, the modest correlation suggests that the radical addition step is rate-determining. This interpretation is further supported by the high electrophilicity of the radical species involved (ω = 2.38 eV), which induces significant charge polarization in the transition state. At this critical point, the intermediate IV is intercepted by the solvent, influencing the selective formation of different products. In the case of DMF, coordination of the solvent results in intermediate V, while in MeCN, the selective Ritter-type addition to the carbocation IV takes place. The polar coordinating solvent DMF was found to be more favorable than the addition of MeCN by nearly 10 kcal·mol^–1. To further validate this exergonic step, we conducted a competition experiment using a solvent mixture of DMF and MeCN, where even with just 10% DMF in MeCN, the lactone product predominated. The intramolecular cyclization of intermediates V and VI leads to the formation of products 20 and 19, respectively, through subsequent reaction steps.

A final series of experiments was conducted to elucidate the mechanism leading to the formation of product 21. Experimental evidence suggests the process likely involves a HAT from intermediate II to generate a tolyl radical (VIII). This radical predominantly reacts with the anhydride to form a carboxyl radical IX, which then undergoes decarboxylation to generate the ^•CF₂Cl radical (X). The recombination of VIII and X was observed by GC-MS, and this process occurs only in solvents capable of undergoing HAT, supporting the proposed pathway. Furthermore, radical X adds to the olefin, followed by an RPC process with Ir­(IV), ultimately leading to the formation of the difunctionalized product 21. Although spontaneous C–O bond fragmentation of I was not observed during our DFT studies, it was computed to be more exergonic than the C–Cl bond cleavage in toluene. Therefore, the straightforward formation of radical X through C–O mesolytic cleavage cannot be excluded.

As seen so far, anhydrides have proven to be excellent precursors for the synthesis of gem-difluoro cycles, making them valuable reagents in this transformation. However, their synthesis originates from corresponding acids or acyl chlorides, reducing efficiency in terms of atom- and step economy. Furthermore, anhydrides are moisture sensitive, which can lead to synthetic complications. To overcome this limitation, we employed α-halodifluoroacetic acids, particularly CDFA, as a cost-effective, widely available redox-active reagent for gem-difluoro synthons. We hypothesized that upon activation of the C–Cl bond, it could generate a gem-difluoro carboxy alkyl radical (FigureA), which could also serve as a bifunctional species in the presence of an olefin molecule. Additionally, it also holds potential for delivering divergent products upon solvent and additive switch (FigureB).

After extensive optimization, we identified a straightforward condition for synthesizing lactone 20. However, lactam formation proved challenging, as the reaction predominantly yielded lactone adduct. We attributed this to the preferential intramolecular cyclization of the acid’s oxygen rather than MeCN addition to the carbocation. To suppress the acid’s nucleophilicity, we introduced Boc-anhydride to form a mixed anhydride in situ, which successfully facilitated lactam formation 19. Encouraged by this result, we aimed to gain further control over the alkyl radical intermediate and found that sodium ascorbate (NaAsc) effectively serves as a HAT reagent, while MeOH engages in esterification of the acid group, thereby facilitating the formation of linear difluoro esters in a single step (42). After optimizing the formation of γ-lactones, γ-lactams, and difluoromethyl esters, we applied the method to a variety of alkenes. The representative examples are summarized in Figure. Lactone formation proved to be highly productive. For instance, styrene with benzylic chlorine yielded 43, while indene afforded product 44. For compound 45, the reaction occurred exclusively at the exocyclic olefin motif and was confirmed by X-ray. Although lactam synthesis was less efficient, we successfully obtained the products from indene and estrone derivatives, 46 and 47, respectively.

One of the primary challenges in our previous studies involving anhydrides and acids under Ir-photocatalyzed conditions was the lack of reactivity with nonstyrenyl olefins. It is generally encountered in the literature that this photocatalyst is unable to oxidize alkyl radicals, − despite the oxidation potential of secondary alkyl radicals being only slightly higher than benzylic radicals (E_1/2 (iPr^•) = 0.47 V > E_1/2 (PhCHMe^•) = 0.37 V, vs SCE), a value that remains lower than that for the reduction of Ir^IV(ppy)₃ (E_1/2 [Ir^IV/Ir^III] = 0.77 V vs SCE). One can assume that the absence of an aromatic group in alkyl derivatives disfavors the formation of the complex necessary for the SET event to occur between Ir^IV(ppy)₃ and the alkyl radical, therefore inhibiting the RPC pathway. To overcome this challenge, we relied on an alternative approach consisting in closing the Ir^IV/Ir^III catalytic cycle with an external reductant. After a thorough analysis of literature reports, we found that sodium ascorbate (NaAsc) could not only reduce the photocatalyst but also serve as a HAT reagent, playing a dual role in the reaction. , Building on these results, we broadened the scope of α,α-difluoroester synthesis beyond styrenes, successfully reacting a variety of unactivated olefins and achieving good yields. Benzyl ethers, cyclic olefins, and pregnenolone acetate all participated in the reaction, yielding products 48–50 in good yields. Notably, when d₃-MeOH was employed, the reaction resulted in the formation of the CD₃-incorporated product (51). Our methodology thus provides a straightforward platform for accessing various gem-difluoro esters from a single acid precursor by simply varying the alcohol additive, unlike traditional methods that rely on preformed esters. We also examined different halogen substitutions at the α-position of the acid group. Our findings revealed that fluorine remained untouched under the reaction conditions, while chlorine exhibited slightly less reactivity than bromine (55–57), with the reactivity order being (F ≪ Cl < Br).

This developed protocol was also applied to the concise synthesis of bioactive compounds. We efficiently synthesized (±)-Boivinianin A (58) and its fluorinated analogue (59) in a single step from all commercially available starting materials. Additionally, we achieved the most step-efficient synthesis reported for a mixture of (±)-3,4-epi- and 4-epi-eupomatilone-6 (61). The mechanism for the formation of lactam 19 and lactone 20 begins with an irreversible SET of gem-difluoroacetic acids (3), followed by the subsequent reaction with olefins, achieving divergent synthesis in a similar manner in MeCN and DMF, as described in FigureA.

Moreover, a mechanism for the synthesis of the hydrofunctionalized product 42 in MeOH:MeCN via a NaAsc-mediated HAT mechanism is outlined in Figure, highlighting the key steps in the reaction pathway. The process begins with the conversion of CDFA to its ester under acidic conditions, leading to the formation of gem-chlorodifluoroester I at room temperature. Upon photoexcitation, the Ir photocatalyst facilitates the SET reduction of I, generating radical anion II. Subsequent mesolytic cleavage of II furnishes radical III, which undergoes radical addition to the unactivated olefin, forming the alkyl radical IV. This intermediate then participates in a HAT with NaAsc, ultimately yielding product 42 and ascorbate radical anion. NaAsc plays a crucial additional role in this reaction, as it is also required to reduce Ir­(IV) back to Ir­(III), thereby closing the catalytic cycle, while simultaneously producing radical VI. In line with our proposed mechanism, the facile oxidation of benzylic radicals produced from styrene substrates predominantly leads to lactone as the major product through an RPC mechanism. In contrast, the challenging oxidation of nonstabilized radicals generated from unactivated olefins favors the HAT pathway. This observation clearly indicates that the SET process is significantly more efficient and faster than HAT under these conditions.

During our investigation on the activation of TFAA for the generation of trifluoroacetyl radical, we encountered a significant challenge related to the limited lifetime of ^•C­(O)­CF₃ radical due to its dissociation into CO and ^•CF₃ radical. To address this challenge, we envisioned that direct activation of chlorotrifluoroacetone could enable the synthesis of the homologated trifluoromethylketones through the generation of a stable ^•CH₂COCF₃ moiety (FigureA).

By employing an Ir-catalyzed photoredox strategy, we successfully developed a protocol enabling selective and divergent access to hydro- and halo-trifluoromethyl ketones through precise control of reaction intermediates. We revealed that in chloroform, the reaction follows an RPC mechanism, yielding halo-trifluoromethylated products. Alongside, switching the solvent system to MeOH: MeCN and employing sodium ascorbate (NaAsc) as the HAT reagent allowed for the selective formation of hydro-trifluoromethyl ketones (FigureB). After establishing the protocols, we first explored the application of chloro-trifluoromethyl ketones and found that this process is only limited to styrenes (62, 64).

The reaction proceeds via a similar SET reduction of reagent 4, generating the trifluoroacetonyl radical, which adds to the olefin to form alkyl radical intermediate I (FigureB). The presence of species I was confirmed using TEMPO as a radical trap (FigureB). Following a subsequent SET event through RPC mechanism, the ion pair collapse between carbenium ion II and the chloride anion yields the halo-trifluoromethyl ketone product 62 (ΔG = −60.0 kcal·mol^–1). The RPC mechanistic pathway is followed when the reaction is carried out with a Cl-based reagent, whereas the XAT mechanism is favored when a Br-based reagent is used. Therefore, the reactivity can be extended beyond styrenes to include less activated unsaturated hydrocarbons (65, 66). This difference is likely attributed to a radical chain process via a halogen atom transfer (XAT) mechanism, where the alkyl radical abstracts a bromine atom from reagent 4.

In the case of styrene substrate, DFT calculations revealed that the radical chain pathway involving a benzylic radical was unfavorable (ΔG^‡ = 20.6, ΔG = +2.0 in kcal·mol^–1) compared to the exergonic RPC process (ΔG = −3.5 kcal·mol^–1) (FigureB). We demonstrated again the dual role of NaAsc in this study, enabling the divergent synthesis of products 63, 67 and 68. In the presence of NaAsc, exclusive formation of the hydro-functionalization product was observed following a HAT event (ΔG^‡ = 17.2, ΔG = −19.7 in kcal·mol^–1).

This account highlights our journey in advancing organofluorine and radical chemistry, demonstrating how photoredox catalysis serves as a powerful tool for the direct activation of bulk prepared chemicals. The unique reactivity of fluorinated radicals has enabled highly selective bond formations under mild conditions, unlocking new pathways for constructing divergent products from common reactive intermediates. Our development of switchable divergent synthesis, in which reaction conditions dictate distinct product outcomes from a common radical intermediate, represents a significant achievement. By harnessing electronic effects and tunable reaction environments, we have demonstrated that fluorinated radicals can be selectively directed toward different functionalization pathways with high selectivity. By capitalizing on the power of fluorinated radicals and the strategic implementation of switchable reaction manifolds, we anticipate continued innovation at the interface of radical chemistry and modern organic synthesis, shaping the next generation of functional molecules.

Moving forward, the development of novel redox-active reagents capable of generating elusive radicals for the selective functionalization of hydrocarbons within the framework of switchable divergent synthesis represents an exciting research direction for advancing synthetic methodology and expanding the scope of radical chemistry. Several promising routes for further exploration include developing asymmetric strategies for constructing cyclic structures such as lactams and lactones, as well as activating widely available feedstock materials, particularly those without carbonyl functional groups, such as alkyl chlorides, to enable olefin functionalization and broaden the chemical space of organofluorine compounds. Additionally, we foresee that insights from computational studies will further enhance our understanding of applied theoretical chemistry, enabling more precise reaction design and a deeper comprehension of underlying mechanisms involving fluorinated species.

Although concerns over fluorinated compounds, known as ″forever chemicals,″ have led to legal actions, production discontinuation, and calls for stricter regulations, these issues are crucial and deserve focused attention. Importantly, recent breakthroughs in the degradation of PFAS instil a lot of promises and inspiring opportunities for the field. − Nevertheless, the smaller organofluorine compounds discussed in this account are less toxic than PFAS and therefore lie outside the scope of the present discussion. Here, we focus on the beneficial role of fluorine in synthetic organic chemistry, where fluorinated motifs continue to drive innovation in medicinal chemistry, materials science, and other fields, leveraging fluorine’s unique electronic and steric properties to enhance molecular performance.

D.K. and A.J.F. acknowledge the Swiss National Science Foundation (PCEFP2_186964 and TMPFP2_209846, respectively) for the financial support of this research. A.J.F. acknowledges the Holcim Foundation for generous financial support. Generous and continuous support by the University of Bern is acknowledged.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.accounts.5c00239↗.

R.G. and A.J.F. contributed equally. CRediT: Rahul Giriconceptualization, data curation, writing - original draft, writing - review and editing; Anthony J. Fernandesinvestigation, writing - original draft, writing - review and editing, funding acquisition; Dmitry Katayev conceptualization, project administration, supervision, funding acquisition, writing - original draft, writing - review and editing.

The authors declare no competing financial interest.