Red Light-Based Dual Photoredox Strategy Resembling the Z-Scheme of Natural Photosynthesis

Photoredox catalysis typically relies on the use of single chromophores, whereas strategies, in which two different light absorbers are combined, are rare. In photosystems I and II of green plants, the two separate chromophores P680 and P700 both absorb light independently of one another, and then their excitation energy is combined in the so-called Z-scheme, to drive an overall reaction that is thermodynamically very demanding. Here, we adapt this concept to perform photoredox reactions on organic substrates with the combined energy input of two red photons instead of blue or UV light. Specifically, a CuI bis(α-diimine) complex in combination with in situ formed 9,10-dicyanoanthracenyl radical anion in the presence of excess diisopropylethylamine catalyzes ca. 50 dehalogenation and detosylation reactions. This dual photoredox approach seems useful because red light is less damaging and has a greater penetration depth than blue or UV radiation. UV–vis transient absorption spectroscopy reveals that the subtle change in solvent from acetonitrile to acetone induces a changeover in the reaction mechanism, involving either a dominant photoinduced electron transfer or a dominant triplet–triplet energy transfer pathway. Our study illustrates the mechanistic complexity in systems operating under multiphotonic excitation conditions, and it provides insights into how the competition between desirable and unwanted reaction steps can become more controllable.

All deuterated solvents were purchased from Cambridge Isotope Laboratories Inc or Apollo Scientific. All reagents were purchased from Fluorochem, Alfar Aesar, Acros Organics or Sigma-Aldrich/Merck in "reagent grade" purity or better and were used as received. [Cu(dap)2]Cl and DCA were purchased from Sigma-Aldrich/Merck. Solvents for spectroscopic measurements were purchased "extra dry" in 99.8% purity from Acros Organics.
NMR spectra were recorded on a Bruker Avance III instrument operating at 400 MHz proton frequency. All samples were recorded at 295 K in 5 mm diameter tubes. Chemical shifts were referenced internally to residual solvent peaks using d values as reported by GOTTLIEB et al. [ Absorption spectra were recorded on a Cary 5000 UV-Vis-NIR spectrometer from Varian. Photoluminescence spectra were recorded on a Fluorolog-322 instrument from Horiba Jobin-Yvon. For laser flash photolysis, an LP920-KS apparatus from Edinburgh Instruments was used. A frequency-tripled Nd:YAG laser (Quantel Brilliant, ca. 10 ns pulse width) equipped with an OPO from Opotek and a beam expander (GBE02-A from Thorlabs) in the beam path were used for excitation with visible light. The direct output of another frequency-doubled Nd:YAG laser (Quantel Q-smart 450 mJ, ca. 10 ns pulse width) with a beam expander (GBE02-A from Thorlabs) in the beam path was used for excitation at 532 nm. The excitation energies were varied by the Q-switch delays and measured with a pyroelectric detector. Typically, pulse energies of ~50 mJ were used for the measurements with 532 nm. Detection of transient absorption and time-resolved emission spectra was performed with an iCCD camera (Andor). Kinetics at single wavelengths were recorded using a photomultiplier tube.

S4
Cyclic voltammetry was performed using a Versastat3-200 potentiostat from Princeton Applied Research. A saturated calomel electrode (SCE) served as reference electrode, a glassy carbon disk electrode was employed as working electrode, and a silver wire was used as counter electrode. Measurements were performed with potential sweep rates of 100 mV/s in dry de-aerated solvent with 0.1 M TBAPF6 (tetra-n-butylammonium hexafluorophosphate) as supporting electrolyte. Sample concentrations were adjusted to values between 1 mM and 5 mM of analyte.
Spectro-electrochemical measurements were performed in a quartz cuvette using the abovementioned potentiostat and the UV-Vis-NIR spectrometer. A platinum grid electrode served as working electrode, a platinum wire was used as counter electrode and an SCE was employed as reference electrode.
As light source for cw-laser experiments in photocatalysis, a 635 nm (optical output up to 500 mW) continuous wave (cw) laser (Roithner Lasertechnik) with precisely adjustable radiative power and high output stability (< 1%) was used. For measurements with a 623 nm LED, a high power LED (Thorlabs Solis-623C, 3.8 W) was used.
Output spectra of the LED and cw laser are provided ( Figure S5).

General procedure 1 for tosylation of phenols
In a round-bottom flask the corresponding phenol (1 eq.) was dissolved in dichloromethane and triethylamine (1.5 eq.) and tosyl chloride (1.2 eq.) were added. The reaction mixture was stirred at room temperature and the reaction progress was monitored by TLC. After complete consumption of the substrate (typically 5-20 h), the reaction was quenched by the addition of water and extracted with dichloromethane (3 ´). The combined organic phases were dried over Na2SO4, the solvent was removed under reduced pressure and the crude mixture was purified by flash column chromatography to obtain the desired tosylated product.

General procedure 2 for tosylation of anilines and amines
In a round-bottom flask the corresponding aniline (1 eq.) was dissolved in THF (10 mL for 1-2.5 mmol substrate) and NaH (60% in mineral oil, 2.5 eq.) and tosyl chloride (1.5 eq.) were slowly added. The reaction mixture was stirred at room temperature and the reaction progress was monitored by TLC. After complete consumption of the substrate (typically 5-20 h), the reaction was quenched by the addition of water and extracted with dichloromethane (3 ´). The combined organic phases were dried over Na2SO4, the solvent was removed under reduced pressure and the crude mixture was purified by flash column chromatography to obtain the desired tosylated product. These analytical data are in agreement with previously reported characterization data for this compound. 7
These analytical data are in agreement with previously reported characterization data for this compound.
The combined organic phases were dried over Na2SO4 and the solvent was removed under reduced pressure. The crude mixture was purified by flash column chromatography (SiO2, cyclohexane / dichloromethane 1/0 à 4/1) to obtain the desired 9-tosyl-9H-carbazole (19, 1.60 g, 4.98 mmol, 82%) as a white solid. These analytical data are in agreement with previously reported characterization data for this compound. 9

S16
These analytical data are in agreement with previously reported characterization data for this compound. 14
The reaction mixture was stirred at room temperature over night, quenched by the addition of water (25 mL) and
The NMR spectra are in line with previously reported data.  These analytical data are in agreement with previously reported characterization data for this compound. 29
The reaction was quenched by the addition of water (50 mL) and the two layers were separated. The aqueous layer was extracted with dichloromethane (3 × 50 mL), the organic phases were combined and dried over Na2SO4. The solvent was removed under reduced pressure and the crude mixture was purified by flash column chromatography (SiO2, cyclohexane / ethyl acetate 1/0 à 7/1) to obtain a white solid (450 mg, 1.72 mmol, 85%). These analytical data are in agreement with previously reported characterization data for this compound. 30
The organic phase was dried over Na2SO4 and the solvent was removed. The crude product was purified by flash column chromatography (SiO2, cyclohexane / ethyl acetate 1/0 à 9/1) to obtain a white solid (48,  These analytical data are in agreement with previously reported characterization data for this compound. 32
The conversion and the yield were determined against the internal standard.

Isolated yield: 3,6-di-tert-butyl-9H-carbazole (20-P)
A round-bottom flask with septum cap was equipped with a magnetic stir bar and charged with 3,6-di-tert-butyl- These analytical data are in agreement with previously reported characterization data for this compound. 35      The reaction progress over time ( Figure S1) was determined for substrate 1 with the reaction conditions given in section 2.2.1. The reaction scale given in the general procedure was triplicated and the reaction was performed in a round-bottom flask with a rubber septum under an Argon atmosphere instead of a cuvette with a silicon septum.

Conversion and yield over time
The reaction progress was monitored for the reaction in the absence (red data points in Figure S1) and in the presence (blue data points) of 0.5 eq. of Cs2CO3 as additive. Conversion and yield were determined against 1fluorotoluene as internal standard.

Effect of caesium carbonate
Conversion and yield for substrate 1 in the absence and in the presence of caesium carbonate were determined by 19 F{ 1 H}-NMR measurements against an internal standard, and the data is presented in the Figure S1.  The experimental setup used for photoredox reactions with a cw-laser as irradiation source has been published recently. 43 In this study, the light source was changed from a 447 nm cw-laser (as described previously) 43 to a 635 nm laser, but other than that, the setup remained essentially identical. The experiments performed with LED irradiation were performed using the setup shown in Figure S4.

Photophysical characterization of [Cu(dap) 2 ]Cl and DCA
All measurements were performed in de-aerated solvents at room temperature. Unless otherwise noted, acetonitrile was used as solvent for all optical spectroscopic investigations.    Absorption spectra, emission spectra for Cu(dap)2Cl and DCA were similar as in prior studies. [48][49][50] In principle, we would expect a mono-exponential decay of the emissive 3

Cyclic voltammetry and spectro-electrochemical study
Cyclic voltammetry was performed with a saturated calomel electrode (SCE) as reference electrode, a glassy carbon disk electrode as working electrode, and a silver wire as counter electrode. Measurements were performed in dry de-aerated acetonitrile with 0.1 M TBAPF6 as an electrolyte. For oxidations or reductions with irreversible waves, the peak potential is given.        Figure S13. Overview of one-electron reduction peak potentials of all investigated substrates within this study. The respective votammograms are shown in the following figures of this section.

General overview
The two mechanisms from Figure 6 of the main paper are further discussed in this section. For both mechanisms additional spectroscopic data is provided. Figure S17. Two possible mechanisms for red light driven photoredox catalysis with [Cu(dap)2] + and DCA, as in Figure 6 of the main manuscript, but here including the additional steps 1b, IIb and IIb2. Table S4. Rate constants (k) for the individual elementary processes illustrated in Figure 6 and Figure S17. All values were determined for de-aerated acetonitrile at 20 °C, unless otherwise mentioned.

S49
In Figure S17 the two main mechanisms of our catalytic systems are reproduced from Figure 6 of the main manuscript. In the following section additional spectroscopic data is provided and discussed. All considered steps and the determined rate constants are summarized in Table S4. quenching under these conditions. 57 We observed that small amounts of undissolved DCA result in scattered laser excitation light, and this can lead to an (incorrect) apparent change of the relative intensity ratio between short- Spectroscopic evidence for the oxidized copper complex [Cu(dap)2] 2+ , formed as a result of electron transfer to DCA is given in Figure S23 and Figure S26, and discussed in more detail in section 4.3.3.

Formation of DCA
As mentioned briefly in the main paper, triplet-triplet annihilation upconversion (step IIb in TTET mechanism of Figure S17) could potentially occur in our system, and the resulting 1* DCA species could then be quenched by evidence for the formation of 1* DCA via upconversion, in the form of delayed fluorescence. The delayed fluorescence is substantially more intense in acetone than in acetonitrile, in line with the predominant initial TTET step in acetone and a predominant initial PET step in acetonitrile.
In the TTET mechanism of Figure 6b, DCA •is formed by electron transfer from DiPEA to 3* DCA (step II in TTET mechanism of Figure S16). The kinetics for this elementary step is discussed in the following. These measurements were performed in acetone, in which TTET is the dominant initial elementary step ( Figure S25). In the absence of DiPEA, 3* DCA has decays by a combination of a first order decay pathway (corresponding to the natural decay of 3* DCA to its electronic ground state) and a second order decay pathway (due to triplet-triplet annihilation). 43

Reduction of [Cu(dap) 2 ] 2+ by DiPEA
Following PET from 3* [Cu(dap)2] + to DCA according to the mechanism in Figure 6a of the main manuscript, the spectral signature of [Cu(dap)2] 2+ and DCA •should be simultaneously detectable by transient absorption spectroscopy. The data in Figure S26a shows that this is indeed the case. The DCA radical anion has characteristic absorption bands in the red spectral range, which appear both in UV-Vis spectro-electrochemistry (green trace in Figure S26a, reproduced from Figure S11) and in a transient absorption measurement (red trace in Figure S26a).

S57
Based on the spectrum in Figure S26b,

Substrate activation by 2* DCA •-: insights based on reactivity
In the main manuscript we discuss previous reports on excitation of DCA •and its stability. There has been some debate concerning the lifetime of the lowest electronically excited state of DCA •-, and a very recent study indicates that it is only ca. 3 ps in solution at room temperature. 47,64,65 Ultrafast laser spectroscopy might in principle be able to provide insights about pre-association between DCA •and the substrate in the ground state, or about the reactivity of the excited state, [66][67][68] but this is beyond the scope of this study. Indirect methods, such as the thermodynamic considerations made in the following can however provide some indirect insight into the excitedstate reactivity. 69 In our system, the redox potential Ered of 2* DCA •can be estimated based on the equation With this data, an oxidation potential of about -2.6 V vs SCE can be approxiated for 2* DCA •-. 47 Figure S27. Conversion of substrates after light-driven reaction plotted against substrate reduction potential Ered. (Differences in irradiation time are not considered). The specific conversion of each substrate is given in the main manuscript and the respective reduction potentials are summarized in Figure S13.

S58
For all investigated substrates the reduction potential has been determined by cyclic voltammetry (see section 4.2.), and the substrate conversions determined after irradiation for all substrates are plotted against these potentials ( Figure S27). It is clearly visible, that almost all substrates with reduction potentials less negative than -2.3 V achieve conversions above 80% (green area in Figure S27) -and for a large majority this is the case within 5 hours of continuous irradiation. Focusing on the substrates with reduction potentials between -2.3 V and -2.4 V vs SCE (red area in Figure S27), significantly lower conversions are observable and substrates do not reach full conversion even with prolonged irradiation times. As this analysis ignores differences in the intrinsic reactivity and/or backelectron transfer, 71 as well as structural effects leading to possible pre-aggregation (e.g. the lack of an (electrondeficient) aromatic system in substrate 10, 12, 37 or 50), 68,72 and uncertainty or errors in the determination of (irreversible) reduction potentials, 73

Proposed mechanism for detosylation reactions
In Figure 6 of the main manuscript, a mechanistic proposal for the hydrodehalogenation of substrate 1 is given as as an example for a reductive dehalogenation reaction. An analogous mechanistic proposal is made here for the detosylation reaction of substrate 19, as exemplary case. This proposed detosylation mechanism is in accordance with the literature. 16 The doublet excited state of that radical anion is extremely short-lived, 47 and therefore 2* DCA •is set in quotation marks, to emphasize the possibility that the photoreaction could in fact predominantly occur from pre-aggregated DCA •-/substrate encounter complexes, or could even involve some DCA photo-degradation products.
TTET mechanism PET mechanism [ All fluorinated substrates were analysed by 19 F{ 1 H}-NMR spectroscopy. For all other substrates, 1 H-NMR spectroscopy was applied to determine conversion and yield against an internal standard. 27 In most cases the product is easily detectable by 1 H-NMR spectroscopy (e.g. substrate 6) and reference spectra are readily available in the literature; for all other cases a product reference spectrum is included for comparison (e.g. substrate 11). An experimental uncertainty of around 5% is estimated for the determined conversions and yields with our method.
In general, all detected signals corresponding to the staring material (X), the product (X-P) and internal standard (IS) are labelled within the respective figures.
For detosylation reactions, signals corresponding to the cleaved tosyl group furthermore become detectable. While a detailed analysis of decomposition pathways of the tosyl group is beyond the scope of this paper, in some cases In the special case of substrate 18 (naphthalene-2,3-diyl bis(4-methylbenzenesulfonate), a partial deprotonation of the product was observed by DiPEA in the reaction mixture, and therefore the 1 H-NMR shifts of this sample are not matching to a reference sample. The formation of the desired product was confirmed by comparison to a sample of 2,3-naphthalenediol recorded in the presence of DiPEA ( Figure S48).         S66 Figure S40. 19 F{ 1 H}-NMR spectra monitoring the light-driven reaction progress of substrate 10 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA, 0.5 eq Cs2CO3 and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range.              S78 Figure S64. 19 F{ 1 H}-NMR spectra monitoring the light-driven reaction progress of substrate 34 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range. Figure S65. 19 F-NMR spectra monitoring the light-driven reaction progress of substrate 35 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range. Figure S66. 19 F-NMR spectra monitoring the light-driven reaction progress of substrate 36 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range. Figure S67. 19 F-NMR spectra monitoring the light-driven reaction progress of substrate 37 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range. Figure S68. 19 F-NMR spectra monitoring the light-driven reaction progress of substrate 38 over time using a 623 nm high power LED irradiation (Thorlabs). Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range. Figure S69. 19 F-NMR spectra monitoring the light-driven reaction progress of substrate 39 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (4-fluorotoluene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range. Figure S70. 1 H-NMR spectra monitoring the light-driven reaction progress of substrate 40 over time using a 623 nm high power LED irradiation. Reaction conditions: 25 mM substrate, 1 mol% [Cu(dap)2]Cl, 10 mol% DCA and 20 eq. DiPEA in MeCN-d3; IS = internal standard (mesitylene). The inset presents the same spectra (in the same colours) as the main part over a wider ppm range.