Mechanistic Perspectives on Organic Photoredox ... - ACS Publications

Article pubs.acs.org/accounts

Mechanistic Perspectives on Organic Photoredox Catalysis for Aromatic Substitutions Published as part of the Accounts of Chemical Research special issue “Photoredox Catalysis in Organic Chemistry”. Michal Majek and Axel Jacobi von Wangelin* Institute of Organic Chemistry, University of Regensburg, 93040 Regensburg, Germany CONSPECTUS: Photoredox catalysis has emerged as a powerful tool for the utilization of visible light to drive chemical reactions between organic molecules that exhibit two rather ubiquitous properties: colorlessness and redox-activity. The photocatalyst, however, requires significant absorption in the visible spectrum and reversible redox activity. This very general framework has led to the development of several new modes of reactivity based on electron and energy transfer steps between photoexcited catalyst states and various organic molecules. In the past years, major effort has been devoted to photoredox-catalytic aromatic substitutions involving an initial reductive activation of various aryl electrophiles by the photocatalyst, which opens a new entry into selective arene functionalizations within organic synthesis endeavors. This, however, has led to a unilateral emphasis of synthetic developments including catalyst modifications, substrate scope studies, and combinations with other chemical processes. This Account summarizes recent reports of new protocols for the synthesis of aromatic esters, thioethers, boronates, sulfonates, heterobiaryls, deuteroarenes, and other functionalized arenes under mild photoredox conditions with organic dyes. On the other hand, mechanistic studies were largely neglected. This Account emphasizes the most relevant experiments and techniques, which can greatly assist in the exploration of the mechanistic foundation of aromatic photoredox substitutions and the design of new chemical reactivities. The nature and physicochemical properties of the employed organic dyes, the control of its acid−base chemistry, the choice of the irradiation sources, and the concentrations of substrates and dyes are demonstrated to decisively affect the activity of organic photocatalysts, the chemo- and regioselectivities of reactions, and the operating mechanisms. Several methods of distinction between photocatalytic and radical chain processes are being discussed such as the determination of quantum yields by conventional actinometric studies or modern photon counter devices. Careful analyses of key thermodynamic and kinetic data of the single electron transfer steps involved in aromatic photoredox substitutions by experimental and theoretical techniques are being exemplified with recent examples from the literature including the determination of redox potentials by DFT and CV, fluorescence quenching studies, and transient absorption/emission spectroscopy. This Account provides the uninitiated reader with an overview of the potential of organic photoredox catalysis for aromatic substitution reactions and encourages the practitioners to consult highly instructive synthetic, mechanistic, theoretical, and spectroscopic tools that are available in research laboratories.

1. INTRODUCTION

undergo direct substitution to give ArZ (Scheme 1, left). This pathway is favored with nitrogen/chalcogen-containing LGs, which stabilize the unpaired electron. Eventually, LG• (or follow-up products) undergoes SET oxidation with the catalyst so that the overall process is redox-neutral. An alternative mechanism operatives if a H atom bearing nucleophile [Z]−H stabilizes the radical adduct Ar[Z]H• (Scheme 1, right). Backelectron transfer to the photocatalyst and deprotonation give the arylation product Ar[Z]. This mechanism is mostly observed with π-donors, which exert a stabilizing effect onto Ar[Z]H•. It is important to note that both mechanistic

Radical arylations with electrophilic ArX substrates have been known for almost 100 years, with the radical nucleophilic aromatic substitutions (SRN1)1 and Meerwein arylations being the most prominent examples.2 These reactions have been complemented with new photoredox-catalytic mechanisms, which operate under milder conditions. The photocatalyst is generally acting as single-electron reductant, which generates the key aryl radical intermediate by reductive cleavage of Ar−X such as aromatic diazonium salts, halides, and sulfonyl chlorides (Scheme 1). The electrophilic nature of aryl radicals allows rapid reactions with various π-nucleophile, electron-transfer, or atom-transfer reagents according to two mechanistic scenarios: Electron donors containing good radical leaving groups (LG•) © 2016 American Chemical Society

Received: June 13, 2016 Published: September 26, 2016 2316

DOI: 10.1021/acs.accounts.6b00293 Acc. Chem. Res. 2016, 49, 2316−2327

Article

Accounts of Chemical Research

2. OVERVIEW OF REACTIONS The first example of aromatic Meerwein-type substitutions by a photoredox mechanism with an organic dye was reported by König et al. in 2012.6 Arenediazonium salts were coupled with electron-rich heteroarenes to form biaryls at short reaction times and low catalyst loading (Scheme 4). Electrophiles with electron-donating and electron-withdrawing substituents were tolerated; furan, thiophenes, and N-Boc pyrroles were competent nucleophilic partners when employed in excess. The reaction was later expanded to involve in situ preparation of arenediazonium salts from anilines and t-butylnitrite, which allowed the conversion of unstable pyridine- and thiazolediazonium salts.7 Highly electrophilic perfluorinated aryl bromides were employed by König (Scheme 5).8 This reaction showed a common feature of many radical arylations: the formation of regioisomers. Regioselective arylations were possible with fluorinated heterarenes where the intermediate biaryl radical adduct is considerably stabilized. Aromatic photoredox substitutions can be embedded within benzanellation strategies (Scheme 6). 2-Methylthioarenediazonium ions can be annulated with a variety of alkynes to form substituted thiophenes (Scheme 6A).9 Photocatalytic annulation of alkynes with 2-phenyldiazonium salts afforded phenanthrenes (Scheme 6B).10 Arenesulfonyl chlorides are convenient precursors to similar radical annulations. Gu et al. reported a photocatalytic synthesis of phenanthridines from 2isocyanobiphenyls (Scheme 6C).11 Irradiation with blue LED of unspecified emission properties was applied, and it is also unclear whether protonated eosin Y or its salt was used. So far, all discussed examples of aromatic photoredox substitutions exploited the nucleophilicity of π-electron donors to trap the intermediate aryl radical. On the other hand, there are numerous available reagents that contain σ- and n-electron donor moieties such as heteroatoms bearing electron lone pairs and weak heteroatom−heteroatom bonds. Our group has reported the first example of photocatalytic carbonylation, which uses gaseous CO as n-electron donor (Scheme 7).12 Interestingly, tert-butanol was tolerated as coupling partner, which is in contrast to “dark” carbonylations with Ni or Pd catalysts. It was also shown that similar carbonylations proceed with catalytic fluorescein under 80 bar CO.13 The required elevated CO pressure is a consequence of the high reactivity of the radical intermediates with other partners. Carbonylations with electron-rich arenes to diarylketones were reported.14−16 The use of σ-electron donors in aromatic photoredox substitutions was reported by Yan et al. in a photocatalytic borylation protocol with arenediazonium salts and bis(pinacolato)diboron (Scheme 8).17 A wide range of substituents were tolerated in this protocol under UV−vis irradiation with a broad-band compact fluorescent lamp (CFL). The authors did not provide details of the acidobasic form of the employed eosin Y. In a similar manner, chalcogen−chalcogen σ-bonds can be cleaved. We developed a photocatalytic thiolation under mild conditions (Scheme 9),18 which also allowed the conjugation with functional thiol derivatives such as cysteine. A combination of the general procedure with the in situ generation of arenediazonium salts from anilines and tertbutyl nitrite was reported.19 C−H bonds can serve as formal σ-electron donors and transfer a hydrogen atom. This was embedded in an overall

Scheme 1. Mechanisms of Aromatic Photoredox Substitutions with Initial Reductive SET to Electrophilic ArX and Terminating Back-ET to the Catalyst (Oxidation of Intermediates LG• or Ar[Z]H•)

pathways involve sequential SET events but are overall redoxneutral. The photocatalyst acts as a 1e-redox shuttle. The most common catalysts applied to photoredox reactions are organic dyes and metal complexes. The first example of a photoredox-catalyzed aromatic substitution was reported by Cano-Yelo and Deronzier in 1984 with catalytic tris(bipyridine)−ruthenium(II).3 [Ru(bpy)3]2+ and related Ru/Ir complexes are the most widely used photoredox catalysts.4 However, metal-free alternatives are appreciated for their lower price, wider structural variation, and better environmental profile. Among them, benzanellated heteroarenes have emerged as most versatile photocatalysts for aromatic redox substitutions (Scheme 2). Scheme 2. Organocatalysts of Aromatic Photoredox Substitutions

Most recent literature reports utilize the xanthene dyes eosin Y (2′,4′,5′,7′-tetrabromofluorescein) and eosin B (4′,5′dibromo-2′,7′-dinitrofluorescein). Only recently, application of the near-UV active phenothiazine PTH (10-phenylphenothiazine) was reported. The key characteristics and photo and redox properties of eosin Y, [Ru(bpy)3]2+, and Ir(ppy)3 as prototypical photocatalysts are summarized in Scheme 3.5 Ir(ppy)3 is the strongest reductant, while [Ru(bpy)3]2+ and eosin Y show similar reductive power. [Ru(bpy)3]2+ is a stronger oxidant than eosin Y. Most striking is the price difference between eosin Y and the metal catalysts in the order of 1/100 (vs the cheapest Ru dye) and 1/1300 (vs Ir(ppy)3). This limitation will certainly attract major research efforts toward the use of cheaper 3d-metal complexes and organic dyes as photocatalysts. 2317


Article

Accounts of Chemical Research Scheme 3. Comparison of Photocatalystsa

a

The highest reduction potentials are given in red, the highest oxidation potentials in blue. Redox potentials were calculated by the Rehm−Weller equation and are given as Ered (vs SCE).

Scheme 4. Regioselective 2-Heterobiaryl Synthesis from Arenediazonium Salts

Scheme 5. Photocatalytic Synthesis of Fluorinated Biaryls

hydrodefunctionalization of aniline derivatives. A key step is a photocatalytic hydrodediazonation of arenediazonium ions with DMF as inexpensive H-donor (Scheme 10).20 This frequently observed side reaction in photocatalyses with poor electron-

donor reagents was never studied in full detail. This eosin Bcatalyzed hydrodefunctionalization under green light irradiation is tolerant of functional groups such as esters and azides. The 2318


Article

Accounts of Chemical Research Scheme 9. Dichalcogenides as Formal σ-Donors in Thiolations/Selenations

Scheme 6. Photocatalytic Benzanellations

Scheme 10. DMF as σ-Donor in Deutero/ Hydrodediazotations

use of d7-DMF as solvent provides a selective aromatic deuteration method. The phenothiazine catalyst PTH enables a related hydrodehalogenation of the less electrophilic aryl iodides and bromides in the presence of trialkylamine/formic acid as H atom donor under blue/near UV irradiation.21 The photocatalytic activation of unreactive substrates such as aryl halides remains a major limitation of visible light-driven redoxsubstitutions. The problem resides within the insufficient reductive power of most organic dyes when employing electrophiles with Ered < −1.5 eV. Recently, two innovative approaches were developed that combine the photonic activation of two photons in an overall photoredox process. König et al. postulated two consecutive photoinduced electrontransfer steps (conPET) with perylene diimide (PDI) as photocatalyst. The reaction is initiated by dye excitation and SET from a sacrificial electron donor to form the relatively stable, colored radical anion PDI−• (Scheme 11A). Further excitation presumably generates a [PDI−•]* species that harbors sufficient reduction potential to cleave nonactivated aryl iodides and bromides.22 The original PDI protocol had several drawbacks: high loading of a noncommercial, insoluble catalyst, elevated temperature, high excess of coupling partner (Scheme 11B). A second generation protocol with the inexpensive commercial, soluble rhodamine 6G (Rh6G) photocatalyst was reported (Scheme 11C).23 Only very electron-rich N-heterocycles can be used as nucleophilic reactants, which compete with H atom abstraction from the solvent. The second formal two-photon approach to the activation of unreactive aryl halides was developed in our group based on a triplet−triplet annihilation (TTA, Scheme 12).24 This photon upconversion was first described in 1962 and involves energy transfer between a sensitizer and an annihilator and ultimately leads to anti-Stokes fluorescence.25 In our case, biacetyl (2,3butanedione, BD) was excited to its triplet state, which undergoes energy transfer to PPO (2,5-diphenyloxazole). The bimolecular TTA event of 3PPO populates the excited singlet

Scheme 7. CO as n-Donor in Alkoxycarbonylations

Scheme 8. Diborane as σ-Donor in Borylations

2319


Article

Accounts of Chemical Research Scheme 11. Postulated Consecutive Photonic−Electronic− Photonic Activation (A) and Synthetic Applications with Catalytic PDI (B) and Rhodamine 6G (C)

Scheme 12. Photon Upconversion (TTA) for ArBr Activation

highlighted that assist practitioners in the careful evaluation of photocatalytic protocols and their associated mechanisms.

3. ACID−BASE CHEMISTRY A major difference between organic and metallic photocatalysts is the pronounced acid−base chemistry of the former due to the availability of electron lone pairs at heteroatoms. For example, eosin Y and other fluorescein dyes exist as an equilibrating mixture of four components:27 two neutral forms Y (e.g., spirocyclic eosinYH2spiro and ring-opened eosinYH2) and upon sequential deprotonations the monoanionic eosinYHNa and dianionic eosinYNa2. The negative charge at the long-wavelength absorbing xanthene core exerts a significant effect on the photophysical properties. pKa values of 2.0 and 3.8 were derived.28 The neutral forms of fluoresceins adopt spirocyclic structures27 in which the xanthenoid π-system is disrupted and visible absorption and photocatalytic activity are extinguished (Scheme 13).35 Unfortunately, the recent literature has not entirely appreciated the relevance of acid− base behavior in photocatalysis. The mere designation “eosin Y” Scheme 13. Acid−Base Behaviour of Eosin Y

state 1PPO, which provides sufficient reductive power to cleave aryl bromides. The proof-of-concept was shown and led to moderately effective hydrodebromination of aryl bromides, but the general method required high light intensities and suffered from low photocatalyst stability.26 The recent developments of photo-organocatalytic aromatic substitutions by reduction−oxidation mechanisms have enabled new modes of reactivity to access functional organic molecules under mild conditions. However, only a deep understanding of the mechanistic minutiae will allow the control of reaction outcome and design of new reactivities. In the following sections, basic synthetic, technical, and mechanistic tools will be 2320


Article

Accounts of Chemical Research in experimental and mechanistic sections is a shortcoming.11,17 In other reports, a similarly ambiguous “spirit soluble” eosin Y was used,6,9 which, according to Sigma-Aldrich could mean eosinYH2spiro or eosinYH2, both being photoinactive in the visible region. Even more surprising, the majority of publications that do specify the acido-basic form of eosin Y employed the inactive eosinYH2 but do not discuss any followup base-mediated reactions in their mechanistic proposal. The first reports of eosin Y photocatalysis involved α-amine oxidations, which proceeded in the presence of stoichiometric amines to ensure sufficient formation of the dibasic eosin Y.29 In some occasions, weakly basic reactants such as sulfinates can enable conversion to the photoactive forms of eosin Y.30 However, the picture changes under the conditions of aromatic photoredox substitutions with arenediazonium salts or arenesulfonyl chlorides, which are neutral at best. Here, one must rely on self-equilibration to generate minor amounts of the photoactive catalyst, which strongly limits the observed rate of the catalytic reaction. Unlike aqueous solutions, organic solvents are poor buffers so that minor impurities easily affect acid−base equilibria. For example, the Brønsted base properties of organic solvents vary considerably, with DMSO being much more basic than acetonitrile.31 Consequently, different compatibilities with (acidic) additives/impurities were observed in different solvents: Eosin Y-catalyzed arylations are more robust in (relatively basic) DMSO while being capricious in acetonitrile. The addition of strong acids to an acetonitrile solution of eosinYNa2 completely shuts down visible absorption and fluorescence (Figure 1).

Scheme 14. Thermal Decomposition of Arenediazonium Salts

Figure 2. Absorption spectra of eosinYNa2 in acetonitrile (A), after addition of 2 equiv of fresh 4-TolN2BF4 (B), and after addition of 2 equiv of aged 4-TolN2BF4 (stored for weeks in refrigerator, C).

to determine the visible absorption of the respective dye under reaction conditions by a simple UV−vis spectrometer before running (unknown) photocatalyses.

Figure 1. Absorption spectra of eosinYNa2 in acetonitrile (A) and after (B) addition of 4 equiv of p-TSA.

4. IRRADIATION There is no photocatalysis without proper irradiation. The majority of reports use high-power, narrow-band commercial LEDs as light sources. In every case, the crucial overlap of LED emission and catalyst absorption should be checked. For example, eosinYNa2 exhibits an absorption maximum at 525 nm, which coincides with commercial green LEDs (530 nm). The use of nonmatching light sources11 leads to low reaction rates, caused by the small portion of light that interacts with the catalyst. When compact fluorescent lamps (CFLs) are used,10,17 the significant emission in the UV and intense bands below 450 nm should be considered in any mechanistic rationalization (Figure 3). Despite the fluorescent coating, the mercury emission line at 365 nm is clearly visible. This translates to 3.4 eV of energy, which is sufficient for direct, noncatalytic C−I

We suggest that the poor performance of eosin Y and similar photocatalysts in many cases can be attributed to interfering acid−base chemistry. This can be especially pronounced with arenediazonium tetrafluoroborates despite their alleged thermal stability: the Schiemann reaction32 proceeds only at high temperatures. Contrary to anecdotal reports of virtually unlimited longevity at 5 °C, we have observed slow decomposition over days to months, which appeared to be highly dependent on the nature of substituents. BF3 is generated and upon hydrolysis forms a strong acid (Scheme 14), which has severe ramifications for the use of aged arenediazonium salt batches (Figure 2). These and other reports indicated the crucial acid−base properties of organic photocatalysts, especially with trace impurities and in the absence of bases. It is therefore prudent

Figure 3. Emission of a Sylvania Daylight CFL.33 2321


Article

Accounts of Chemical Research and C−Br photolysis. Therefore, it is likely that CFLs enhance unwanted side reactions and catalyst degradation. Further care must be taken with regard to light attenuation by photoactive substrates. Generally, substrate concentrations are orders of magnitude higher than catalyst concentrations (mol %) and well beyond the regime of standard spectroscopic studies at (∼10−5 mol L−1). With concentrations approaching 1 M, even compounds with very low extinction coefficients (0.1− 10) efficiently attenuate the incident light (Lambert−Beer law). The electrophilic substrates of photoredox substitutions can undergo direct photolysis with near-UV or blue light. Arenediazonium salts have appreciable absorption below 450 nm and form weak charge transfer complexes with solvent molecules whose absorption tails into the visible range.34 Direct irradiation of such complexes can lead to heterolytic cleavage of Ar−X bonds, even in the absence of catalyst. We have demonstrated the alteration of reaction outcome by different light sources.10,17,35 The borylation protocol by Yu et al. with CFL irradiation (Scheme 8)17 involves significant photolysis to give aryl cations and product formation in the absence of catalyst (Scheme 15).35,36 This underlines the importance of assessing the rate of background reactions in the absence of photocatalysts under otherwise identical conditions.

Scheme 16. Modulation of Reactivity by the Light Source

breaking steps. However, a direct mechanistic interpretation based on Φ is only straightforward for monomolecular processes. Photocatalytic reactions per se are not monomolecular because they involve bimolecular energy/electron transfer between catalyst and substrate. However, simplifications are feasible in the regime of total light absorption by solutions of high catalyst concentrations. Then, the calculated Φ values represent the lower limits of the real quantum yields. An improved model also accounts for the rate of excited catalyst quenching by the substrate.37 The determination of Φ requires the quantification of the photon flux from the light source, which is problematic in the visible spectrum. Only recently, solar cell devices for direct photon flux measurements became available (Figure 4).36 On the other hand, chemical

Scheme 15. Photocatalysis vs Direct Photolysis in Borylations

Figure 4. Custom-made photon flux counter in our laboratory.39

The clear mechanistic distinction between photolysis and photocatalysis can become manifest in the formation of different products. The formal [4 + 2] annulation of visible light-absorbing 2-phenyl-arenediazonium salts with alkynes is inhibited with high-energy light sources due to the operation of a Meerwein addition−elimination pathway to give acetanilides in acetonitrile solution.10,35 The use of a broad-band CFL in the original publication gave 54% yield of the desired phenanthrene, whereas selective catalyst excitation at 525 nm enhanced the yield (Scheme 16). In light of these observations, we emphasize the importance of selective narrow-band excitation of the corresponding dyes for the collection of reproducible and interpretable synthetic and mechanistic data. Further, it should be noted that any incident light that does not contribute to the population of excited photocatalyst states is a waste of energy and lowers the overall efficiency of the photocatalytic process.

actinometry remains the prevalent indirect method. Potassium ferrioxalate is the most widely used actinometer but its utility at wavelengths >500 nm is poor. None of the standard chemical actinometers for this region are commercial. Potassium Reineckate, a robust >500 nm-actinometer, can be prepared in three steps.35 Recently, a visible-range actinometer based on Ru(bpy)32+ was also proposed.38 For a simple photocatalytic process, the theoretical quantum yield is expected to be between 0 < Φ ≤ 1, approaching unity as the efficiency of photocatalysis increases. In reality, Φ > 1 are observed if photocatalytic intermediates (i.e., sufficiently electron-rich species in photoredox substitutions of electrophilic ArX) can trigger “dark” radical chain reactions by SET (Scheme 17). Contrary to this, a pure photocatalysis shuttles electrons between catalyst and chemical reactants and only experiences catalyst turnover by the absorption of a photon.40 The unambiguous distinction between photocatalysis and radical chain processes is an important prerequisite for the design of new catalysts and reactions. The former mechanism requires a stable catalyst under reaction conditions that undergoes facile SET with the substrate and back-ET with

5. QUANTUM YIELDS The quantum yield, Φ, the number of specific reaction events per absorbed photon, provides a solid measure of the efficiency of radiative energy utilization for the chemical bond making/ 2322


Article


6. THERMODYNAMICS One of the most important indirect methods to probe the likelihood of a proposed mechanism is a simple thermodynamic analysis. Most of the key steps of aromatic photoredox substitutions are electron transfers. Their feasibility can be judged upon the redox potentials of the half-reactions involved. For most starting materials (arenediazonium, arenesulfonyl chlorides) such values are readily available by electrochemical methods, for example, cyclic voltammetry (CV). This allows the determination of the ground-state redox potentials of all reagents and catalysts. The prediction of catalyst activity in a new photoredox reaction, however, requires knowledge of the redox potential of the excited state catalyst as only this determines the substrate scope, which is susceptible to SET. This unfortunately cannot be measured directly due to the short-lived, transient nature of the excited catalyst (Scheme 19). A good estimation of the redox potential of the excited catalyst can be derived from the Rehm−Weller theory (vide inf ra).41

Scheme 17. Photocatalysis vs Radical Chain Mechanisms

the product. The latter mechanism, on the other hand, relies on a radical initiator, which can be consumed upon chain initiation. In an effort to categorize photoredox-induced aromatic substitutions operating by a reduction−oxidation mechanism, we determined the quantum yields of several protocols (Scheme 18).20,35 Scheme 18. Quantum Yields of Photosubstitutions35

Scheme 19. Thermodynamics of Photoredox Catalysis

We demonstrate this approach at a model SET from A to B upon excitation. The excited state energy of A* (from emission spectrum) and the redox potentials of A and B (from CV) can be obtained. The experimentally unavailable energy of electron transfer (blue arrow) can be calculated from the Born−Haber cycle (Scheme 20). Coulomb attraction between the generated charge-separated pair must be considered to obtain the proper free energy ΔG°: ΔG° = ΔEred − E0 − 0 + C

The Coulomb term is negligible in comparison with the other two, especially in polar solvents. Moreover, if we are interested in the hypothetical electrochemical half-reaction A•+ → A*, there is explicitly no Coulomb interaction, and the modified Rehm−Weller equation can be used for the calculation of (experimentally unavailable) redox potentials. All excited state potentials reported herein (Scheme 3) were obtained by this method.

Two cases that exhibited a substantial participation of radical chain processes stand out: the heteroarylation reported by König et al. and our hydrodediazotation method. In the other cases (Φ < 1), radical chain processes in principle are not disproved. If the photoredox step is inefficient and followed by a short radical chain, quantum yields can indeed be below 1, yet radical chains would be operative. In comparison with the photoinduced enamine activations studied by Cismesia and Yoon, generally lower quantum yields were observed.37

Ered(A•+/A*) = Ered(A•+/A) − E0 − 0

Beyond this, the determination of redox potentials of transient intermediates that interact with catalyst species that are inaccessible for experimental studies is still impossible. An example is the thermodynamic analysis of the back-ET from radical products to the oxidized photocatalyst. We have used 2323


Article

Accounts of Chemical Research Scheme 20. Thermodynamics of Photo-induced Electron Transfer (PET)

Scheme 22. Acyl Radical/Acylium Ion Potentials and Experimental Yields

DFT calculations to estimate such redox potentials.12,20,24 Initially, the equilibrium geometries of the reactants before and after electron transfer are calculated using the polarized continuum model to approximate the solvent. The obtained ΔE correlates well with the redox potential. After subtraction of 4.19 V (which compensates for the reference saturated calomel electrode (SCE) vs vacuum),42 this can be directly compared to the values obtained from CV with an SCE. The errors of DFTderived energies prevent precise redox potential determinations, but the accuracy is sufficient for the prediction of redox reactivities. Recently, Hansen et al. applied a similar approach to photoredox systems with metallic photocatalysts.43 The utility of such DFT methods was meaningfully demonstrated in the mechanistic rationalization of the photoredox carbonylation developed in our group (Schemes 7 and 21).12

Another example where DFT-supported thermodynamic analysis assisted the mechanistic rationalization is the photoredox-catalyzed hydrodediazotation from our laboratories.20 The determined Φ of ∼2.5 indicates the participation of radical chain processes. Further support of the proposed mechanism was derived from thermodynamic analyses of the relevant SET steps (Scheme 23). Scheme 23. Proposed Hydrodediazotation Mechanism

Scheme 21. Mechanism of the Photoredox Carbonylation12

The redox potentials of arenediazonium salts and excited state eosin Y (from Rehm−Weller theory) supported the feasibility of SET reduction of the former with EY*. The reaction of aryl radicals (II) with CO to an acyl radical (III) is known.44 On the other hand, the oxidation potentials of III → IV are not experimentally available and were calculated by DFT. For most substrates, this potential was lower than the oxidation potential of the eosin Y radical anion EY−• (EY−•/ EY2−) meaning that the oxidation of acyl radical III to acylium ion IV is thermodynamically downhill (Scheme 22). Indeed, the ester products were isolated in good yields for these cases. The 2-nitro substituent renders the oxidation of III highly unfavorable; this theoretical finding was experimentally corroborated by the absence of any product formation. Instead, TEMPO addition afforded the acyl radical adduct supporting the formation of III.

The redox potentials of eosin B and arenediazonium salts are experimentally available; the redox potential of excited EB* was obtained from the Rehm−Weller equation.41 DFT calculations furnished the oxidation potentials of the DMF-derived radical species X. Both arenediazonium reduction by EB* and catalyst regeneration by SET from X are energetically feasible (Schemes 23, 24). On the other hand, the reduction of arenediazonium (VI) by X is at best energetically neutral and unfavorable by as much as 0.75 V for some substrates (e.g., 2-biphenyldiazonium). While this finding renders a radical chain mechanism mediated by the DMF-radical X unlikely, SET-redox reactions are still possible to operate against a moderate potential gradient of approximately +500 mV if the onward-reaction following the electron transfer is irreversible (such as the loss of 2324


Article


from the longer lived triplet excited state. Then, transient spectroscopy can be used to access the kinetics of the 3cat* decay. With the observed lifetimes, Stern−Volmer analyses can also be executed,

Scheme 24. Redox Potentials of Photo-hydrodediazotation Steps

1 1 = + k[Q] τ τ0

with τ0 and τ being the 3cat* lifetimes in the absence and presence of Q.47 Such spectroscopic techniques can be especially useful when the initial photoinduced ET of the overall mechanism is considered. With a sacrificial electron donor, two distinct photoredox mechanisms are possible where oxidative or reductive quenching are the initiating steps (Scheme 26).

dinitrogen after SET reduction of arenediazonium).45 The predicted inertness of 4-biphenyldiazonium correlates well with the published results of “dark” hydrodediazotations in DMF with Fe(II) as reductants.46 Under such conditions, most arenediazonium salts were reduced with substoichiometric amounts of Fe(II), which also suggests the operation of a DMF-mediated radical chain process. Both carbonylative and hydrodediazotative aromatic substitutions developed in our group demonstrated the great utility of the combination of DFT calculations with cyclovoltammetry (CV) to obtain the potentials of the redox pairs involved in SET steps.

Scheme 26. Oxidative/Reductive Quenching in Eosin Y Photocatalysis

7. SPECTROSCOPY The toolbox of meaningful mechanistic experiments is greatly expanded if thermodynamic data are complemented with an understanding of reaction kinetics. The reaction kinetics are the key phenomena that define catalytic reactions. The efficiency of all elemental steps is mainly determined by their rate constants, a property that is not easily accessible by electrochemistry or quantum mechanics. Moreover, many photoredox catalyst intermediates are short-lived. The kinetic fate of such intermediates can only be monitored by spectroscopic techniques (Scheme 25). Mechanistic studies by transient spectroscopy were recently performed by Slanina and Kö n ig in the arylation of perfluoroarenes (Scheme 27).8 Laser-flash spectroscopy showed a decreased lifetime of 3EY* on addition of the sacrificial electron donor (triethylamine, TEA), while addition of aryl halides showed no effect. This study indicated the

Scheme 25. Spectroscopy for Kinetic Analyses

Scheme 27. Quenching of Eosin-Catalyzed Arylations8

Stern−Volmer analysis is a useful tool to access the rates of SET in photoredox catalysis and can be used to determine the rates of deactivation of both singlet and triplet excited states. The rate of photoinduced SET with the singlet excited state catalyst is derived from fluorescence quenching experiments with a set of suitable substrates by the Stern−Volmer equation,47 I0 = 1 + kτS[Q] I where I0 and I are the integral intensities of fluorescence in the absence and presence of the quencher, k is the rate constant of the SET, τs is the 1cat* lifetime, and [Q] the quencher concentration.47 However, photoredox catalysts often react 2325


Article

Accounts of Chemical Research operation of reductive catalyst quenching. The resulting 3EY•− radical anion underwent effective quenching with the aryl halide, which suggests the second redox step being the SET reduction of ArBr. The final example is devoted to the observation of excited state fluorescence and lifetime quenching by transient emission spectroscopy in a triplet−triplet annihilation (TTA) experiment with a dual dye system.24,25 The observed fluorescence from the excited sensitizer 2,3-butanedione, 1BD*, and the delayed anti-Stokes fluorescence from the annihilator 2,5diphenyloxazole, 1PPO*, testify to the operation of a TTA cascade (Scheme 28, Figure 5). Following the photophysical

However, the narrow redox potential window requires relatively activated aryl electrophiles. Only recently have methods become available that extend the potential window by exploiting combinations of photonic and electronic activations with single or dual catalyst systems (conPET, TTA) or operating at the near-UV edge (PTH). We firmly believe that only a detailed understanding of the underlying principles and mechanisms warrants the knowledge to design more active catalysts and discover new reactivities. This Account discussed several key parameters and highlighted simple experimental and theoretical techniques that can provide quick yet meaningful answers to the true nature of a presumable photoredox process.

■

Scheme 28. Photophysics of Triplet−Triplet Annihilation (TTA)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest. Biographies Michal Majek hails from Slovakia and studied Chemistry at the Institute of Chemical Technology, Prague. He joined the University of Regensburg where he graduated with highest honors in 2015. Michal is currently a postdoctoral fellow with A.J.v.W. Axel Jacobi von Wangelin grew up in Berlin and studied Chemistry at the Universities of Erlangen, Utah, and Rostock. He worked with John A. Gladysz (Erlangen), Matthias Beller (Rostock), Kingsley J. Cavell (Cardiff), and Barry M. Trost (Stanford). Axel started his independent career at the University of Cologne and since 2011 is Professor of Organic Chemistry at the University of Regensburg. His research interests involve metal, organo-, and photocatalysis. Axel is a 2016 ERC Consolidator grantee for work on iron catalysis.

■

ACKNOWLEDGMENTS We acknowledge financial and technical support from the Graduate School on Photocatalysis of the DFG. We thank the reviewers for helpful comments.

■

REFERENCES

(1) Bunnett, J. F. Aromatic Substitution by the SRN1 Mechanism. Acc. Chem. Res. 1978, 11, 413−420. (2) Hari, D. P.; König, B. The Photocatalyzed Meerwein Arylation: Classic Reaction of Aryl Diazonium Salts in a New Light. Angew. Chem., Int. Ed. 2013, 52, 4734−4743. (3) Cano-Yelo, H.; Deronzier, A. Photocatalysis of the Pschorr reaction by Tris-(2,2′-bipyridyl)ruthenium(II) in the Phenanthrene Series. J. Chem. Soc., Perkin Trans. 2 1984, 1093−1098. (4) (a) Prier, C. K.; Rankic, D. A.; MacMillan, D. W. C. Visible Light Photoredox Catalysis with Transition Metal Complexes: Applications in Organic Synthesis. Chem. Rev. 2013, 113, 5322−5363. (b) Teplý, F. Photoredox Catalysis by [Ru(bpy)3]2+ to trigger transformations of organic molecules. Collect. Czech. Chem. Commun. 2011, 76, 859−917. (5) Properties of [Ru(bpy)3]2+ and Ir(ppy)3 were taken from ref 4a. Properties of eosin Y from: (a) Lazarides, T.; McCormick, T.; Du, P.; Luo, G.; Lindley, B.; Eisenberg, R. Making Hydrogen from Water Using a Homogeneous System Without Noble Metals. J. Am. Chem. Soc. 2009, 131, 9192−9194. (b) Xu, J.; Shanmugam, S.; Duong, H. T.; Boyer, C. Organo-photocatalysts for photoinduced electron transferreversible addition−fragmentation chain transfer (PET-RAFT) polymerization. Polym. Chem. 2015, 6, 5615−5624. (c) Penzkofer, A.; Beidoun, A.; Daiber, M. Intersystem-crossing and excited-state absorption in eosin Y solutions determined by picosecond double

Figure 5. Transient emission spectroscopy of TTA with 2,3butanedione (BD)/2,5-diphenyloxazole (PPO).24

steps of the TTA, a similar setup allowed the determination of lifetime quenching of 1PPO* by the addition of aryl bromides, which appeared to be the initial redox-chemical step (see Scheme 12).24

8. CONCLUSIONS Aromatic substitutions by photocatalytic reduction−oxidation mechanisms with organic dyes have become a powerful alternative not only to traditional methods but also as a successful rival of metal-based photocatalysts. These photoorganocatalytic methods combine mild conditions with inexpensive and nontoxic catalysts and wide substrate scope. 2326


Article

Accounts of Chemical Research pulse transient absorption measurements. J. Lumin. 1992, 51, 297− 314. Prices: http://www.sigmaaldrich.com/catalog/product/sial/ e6003 (accessed Jun 13, 2016). (6) Hari, D. P.; Schroll, P.; König, B. Metal-Free, Visible-LightMediated Direct C−H Arylation of Heteroarenes with Aryl Diazonium Salts. J. Am. Chem. Soc. 2012, 134, 2958−2961. (7) Maity, P.; Kundu, D.; Ranu, B. C. Visible-Light-Photocatalyzed Metal-Free C−H Heteroarylation of Heteroarenes at Room Temperature: A Sustainable Synthesis of Biheteroaryls. Eur. J. Org. Chem. 2015, 2015, 1727−1734. (8) Meyer, A. U.; Slanina, T.; Yao, C.-J.; König, B. Metal-Free Perfluoroarylation by Visible Light Photoredox Catalysis. ACS Catal. 2016, 6, 369−375. (9) Hari, D. P.; Hering, T.; König, B. Visible Light Photocatalytic Synthesis of Benzothiophenes. Org. Lett. 2012, 14, 5334−5337. (10) Xiao, T.; Dong, X.; Tang, Y.; Zhou, L. Phenanthrene Synthesis by Eosin Y-Catalyzed, Visible Light-Induced [4 + 2]-Benzannulation of Biaryldiazonium Salts with Alkynes. Adv. Synth. Catal. 2012, 354, 3195−3199. (11) Gu, L.; Jin, C.; Liu, J.; Ding, H.; Fan, B. Transition-metal-free, visible-light induced cyclization of arylsulfonyl chlorides with 2isocyanobiphenyls to produce phenanthridines. Chem. Commun. 2014, 50, 4643−4945. (12) Májek, M.; Jacobi von Wangelin, A. Metal-Free Carbonylations by Photoredox Catalysis. Angew. Chem., Int. Ed. 2015, 54, 2270−2274. (13) Guo, W.; Lu, L.-Q.; Wang, Y.; Wang, Y.-N.; Chen, J.-R.; Xiao, W.-J. Metal-Free, Room-Temperature, Radical Alkoxycarbonylation of Aryldiazonium Salts through Visible-Light Photoredox Catalysis. Angew. Chem., Int. Ed. 2015, 54, 2265−2269. (14) Gu, L.; Jin, C.; Liu, J. Metal-free, visible-light-mediated transformation of aryl diazonium salts and (hetero)arenes: an efficient route to aryl ketones. Green Chem. 2015, 17, 3733−3736. (15) Zhang, H.-T.; Gu, L.-J.; Huang, X.-Z.; Wang, R.; Jin, C.; Li, G.-P. Synthesis of indol-3-yl aryl ketones through visible-light-mediated carbonylation. Chin. Chem. Lett. 2016, 27, 256−260. (16) Albar, H. A.; Shawali, A. S.; Abdaliah, M. A. Substituent effects on azo coupling of indoles. Can. J. Chem. 1993, 71, 2144−2149. (17) Yu, J.; Zhang, L.; Yan, G. Metal-Free, Visible Light-Induced Borylation of Aryldiazonium Salts: A Simple and Green Synthetic Route to Arylboronates. Adv. Synth. Catal. 2012, 354, 2625−2628. (18) Májek, M.; Jacobi von Wangelin, A. Organocatalytic visible light mediated synthesis of arylsulfides. Chem. Commun. 2013, 49, 5507− 5509. (19) Kundu, D.; Ahammed, S.; Ranu, B. C. Visible Light Photocatalyzed Direct Conversion of Aryl/Heteroarylamines to Selenides at Room Temperature. Org. Lett. 2014, 16, 1814−1817. (20) Májek, M.; Filace, F.; Jacobi von Wangelin, A. Visible Light Driven Hydro/Deuterodefunctionalization of Anilines. Chem. - Eur. J. 2015, 21, 4518−4522. (21) Discekici, E. H.; Treat, N. J.; Poelma, S. O.; Mattson, K. M.; Hudson, Z. M.; Luo, Y.; Hawker, C. J.; Read de Alaniz, J. A highly reducing metal-free photoredox catalyst: design and application in radical dehalogenations. Chem. Commun. 2015, 51, 11705−11078. (22) Ghosh, I.; Ghosh, T.; Bardagi, J. I.; König, B. Reduction of aryl halides by consecutive visible light-induced electron transfer processes. Science 2014, 346, 725−728. (23) Ghosh, I.; König, B. Chromoselective Photocatalysis: Controlled Bond Activation through Light-Colour Regulation of Redox Potentials. Angew. Chem., Int. Ed. 2016, 55, 7676−7679. (24) Májek, M.; Faltermeier, U.; Dick, B.; Pérez-Ruiz, R.; Jacobi von Wangelin, A. Application of Visible-to-UV Photon Upconversion to Photoredox Catalysis: The Activation of Aryl Bromides. Chem. - Eur. J. 2015, 21, 15496−15501. (25) Singh-Rachford, T. N.; Castellano, F. N. Photon upconversion based on sensitized triplet-triplet annihilation. Coord. Chem. Rev. 2010, 254, 2560−273. (26) Häring, M.; Perez-Ruiz, R.; Jacobi von Wangelin, A.; Diaz-Diaz, D. Intragel Photoreduction of Aryl Halides by Green-to-Blue

Upconversion under Aerobic Conditions. Chem. Commun. 2015, 51, 16848−16851. (27) Klonis, N.; Sawyer, W. H. Spectral Properties of the Prototropic Forms of Fluorescein in Aqueous Solution. J. Fluoresc. 1996, 6, 147− 157. (28) Batistela, V. R.; Pellosi, D. S.; de Souza, F. D.; da Costa, W. F.; de Oliveira Santin, S. M.; de Souza, V. R.; Caetano, W.; de Oliveira, H. P. M; Scarminio, I. S.; Hioka, N. pKa determinations of xanthene derivates in aqueous solutions by multivariate analysis applied to UVVis spectrophotometric data. Spectrochim. Acta, Part A 2011, 79, 889− 897. (29) (a) Hari, D. P.; König, B. Eosin Y Catalyzed Visible Light Oxidative C-C and C-P bond Formation. Org. Lett. 2011, 13, 3852− 3855. (b) Neumann, M.; Füldner, S.; König, B.; Zeitler, K. Metal-Free, Cooperative Asymmetric Organophotoredox Catalysis with Visible Light. Angew. Chem., Int. Ed. 2011, 50, 951−954. (30) Meyer, A. U.; Straková, K.; Slanina, T.; König, B. Eosin Y Photoredox-Catalyzed Sulfonylation of Alkenes: Scope and Mechanism. Chem. - Eur. J. 2016, 22, 8694−8699. (31) Himmel, D.; Goll, S. K.; Leito, I.; Krossing, I. A Unified pH Scale for All Phases. Angew. Chem., Int. Ed. 2010, 49, 6885−6888. (32) Sushitzky, H. The Balz-Schiemann reaction. In Advances in Fluorine Chemistry; Stacey, M., Tatlow, J. C., Sharpe, A. G., Eds.; Butterworths: London, 1965; Vol. 4; p 1. (33) Technical Information Bulletin, Spectral Power Distributions of SYLVANIA Fluorescent Lamps, Osram Sylvania, 2000. (34) Hirose, Y.; Wahl, G. H.; Zollinger, H. Dediazoniation of arenediazonium ions in homogeneous solution. Part VIII. Reaction kinetics and products in dimethyl sulfoxide. Helv. Chim. Acta 1976, 59, 1427−1437. (35) Májek, M.; Filace, F.; Jacobi von Wangelin, A. On the mechanism of photocatalytic reactions with eosin Y. Beilstein J. Org. Chem. 2014, 10, 981−989. (36) Zhu, C.; Yamane, M. Transition-Metal-Free Borylation of Aryltriazene Mediated by BF3OEt2. Org. Lett. 2012, 14, 4560−4563. (37) Cismesia, M. A.; Yoon, T. P. Characterizing chain processes in visible light photoredox catalysis. Chem. Sci. 2015, 6, 5426−5463. (38) Pitre, S. P.; McTiernan, C. D.; Vine, W.; DiPucchio, R.; Grenier, M.; Scaiano, J. C. Visible-Light Actinometry and Intermittent Illumination as Convenient Tools to Study Ru(bpy)3Cl2 Mediated Photoredox Transformations. Sci. Rep. 2015, 5, 16397. (39) Megerle, U.; Lechner, R.; König, B.; Riedle, E. Laboratory apparatus for the accurate, facile and rapid determination of visible light photoreaction quantum yields. Photochem. Photobiol. Sci. 2010, 9, 1400−1406. (40) (a) Studer, A.; Curran, D. P. Catalysis of Radical Reactions: A Radical Chemistry Perspective. Angew. Chem., Int. Ed. 2016, 55, 58− 102. (b) Studer, A.; Curran, D. P. The Electron is a Catalyst. Nat. Chem. 2014, 6, 765−773. (41) Knibbe, H.; Rehm, D.; Weller, A. Intermediates and Kinetics of Fluorescence Quenching by Electron Transfer. Ber. Bunsen-Ges. Phys. Chem. 1968, 72, 257−263. (42) Reiss, H.; Heller, A. The absolute potential of the standard hydrogen electrode: a new estimate. J. Phys. Chem. 1985, 89, 4207− 4213. (43) Demissie, T. B.; Ruud, K.; Hansen, J. H. DFT as a Powerful Predictive Tool in Photoredox Catalysis: Redox Potentials and Mechanistic Analysis. Organometallics 2015, 34, 4218−4228. (44) Ryu, I.; Sonoda, N. Free-Radical Carbonylations: Then and Now. Angew. Chem., Int. Ed. Engl. 1996, 35, 1050−1066. (45) Steckhan, E. Indirect Electroorganic Syntheses - A Modern Chapter of Organic Electrochemistry. Angew. Chem., Int. Ed. Engl. 1986, 25, 683−701. (46) Wassmundt, F. W.; Kiesman, W. F. Efficient Catalysis of Hydrodediazoniations in Dimethylformamide. J. Org. Chem. 1995, 60, 1713−1719. (47) Klán, P.; Wirz, J. Photochemistry of Organic Compounds; Wiley: Chichester, 2009.

2327


Mechanistic Perspectives on Organic Photoredox ... - ACS Publications

Recommend Documents