Electron donor-free photoredox catalysis via an electron transfer cascade by cooperative organic photocatalysts†

Electron-donating sacrificial reagents are highly important for certain photo-redox reactions. However, the use of excessive amounts of sacrificial reagents, mostly amines, often leads to undesired side products and is especially troublesome for product purification. Herein, we take the light-induced electron transfer cascade process of natural photosystems as a role model and assemble organic photocatalysts into cooperative photocatalyst couples. The cooperative photocatalyst couples could undergo intermolecular electron transfer to facilitate the charge separation process and overcome the need for an extra electron donor. Time-resolved photoluminescence spectroscopy was conducted to precisely characterize the photoexcited dynamics within the cooperative photocatalyst couples. As a model photoredox reaction, the carbon–carbon formation reaction between heteroarenes and malonates, which usually requires electrondonating sacrificial reagents such as amines, was conducted to demonstrate the feasibility of the cooperative photocatalyst couples under visible light irradiation. A significant reaction conversion improvement from trace conversion for a single photocatalyst system to over 90% by cooperative photocatalyst couples was achieved.


Introduction
Electron-donating sacrificial reagents are of great importance for photocatalytic redox reactions including chemical transformation and water decomposition into hydrogen and oxygen. [1][2][3] Especially for semiconductor-catalyzed reactions, the catalytic efficiency can be largely improved while the photo-generated holes can be scavenged by the electrondonating reagent molecules and the charge recombination can be reduced. 2 Examples of organic photoredox reactions without the employment of sacrificial reagents were conducted using reactive substrates such as aryl diazonium 4 or trifluoromethanesulfonyl salts, 5 as well as when the solvent molecule could also act as an electron donor. 6 However, for specific photoredox reactions involving less reactive substrates, the use of electron donating reagents is necessary. And the use of usually excessive amounts of donor reagents during the catalytic cycle, mostly amines, 7-10 often leads to undesired oxidized side products and is especially troublesome for product purification. 11 To overcome the abovementioned challenge and maintain reduced charge recombi-nation within the semiconductor-based photocatalyst, an enhanced photo-generated electron transfer pathway should be developed.
Natural photosystems offer a promising model to overcome the charge recombination issue, where a series of stepwise electron transfer processes are involved to enhance the photogenerated charge separation and transfer process. 12 So far, chemists have designed different photocatalytic systems for enhancing the photocatalytic efficiency. For example, bridged transition metal complexes [13][14][15] have been reported to greatly promote the photogenerated charge separation and transfer.
Organic semiconductor (OS)-containing photocatalysts have emerged as a promising alternative to traditional metalcontaining photocatalysts due to their highly tunable electronical properties via flexible structural design. Recent research activities have demonstrated a vast number of structural design methods of small molecular 8,10,[16][17][18][19][20] or macromolecular [21][22][23][24][25][26][27][28][29][30] OS systems as efficient photocatalysts for visible light-promoted photoredox reactions. Nevertheless, most developed OS photocatalysts, similar to traditional transition metal complexes, [31][32][33][34][35][36][37][38] are single photocatalytic systems. As illustrated in Fig. 1a, the photocatalytic ability of the OS system originates from the photo-generated electron/hole pair, which can function as the reductive and oxidative species. To maintain the efficient electron transfer from the ground to the exited state of the OS, extra sacrificial agents are usually needed due to direct recombination of the electron/hole pair. To overcome the necessary use of electrondonating sacrificial reagents, the single photocatalyst system is not sufficient; a new concept of catalyst design is needed.
Herein, we aim to take the mechanism of the lightinduced electron transfer cascade process of natural photosystems as a role model and assemble organic photocatalysts into cooperative photocatalyst couples. As illustrated in Fig. 1b, the two molecular OS photocatalysts within the cooperative couple consisted of different organic photocatalysts possessing different highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) levels. By coupling the OSs, an extra intermolecular electron transfer could occur between the OSs, leading to enhanced photo-generated electron/hole separation under light irradiation and thereby more stable reductive and oxidative species. The intermolecular electron transfer within the OS couple can decay the direct recombination of the electron/hole pair and overcome the mandatory requirement of extra electron donors during the photoredox cycle. As a model reaction, the photocatalytic carbon-carbon formation reaction between electron-rich heteroarenes and malonates, which usually required electron-donating sacrificial reagents, 9,39 was chosen to demonstrate the feasibility and photocatalytic activity of the cooperative photocatalyst couple. No extra sacrificial reagents were needed. The light-induced electron transfer between the OSs within the cooperative photocatalyst couple and the substrates could be revealed by mechanistic studies.
Steady-state fluorescence quenching experiments showed that by only exciting Th-BT-Th (at 500 nm) or TA-BT-TA (at 450 nm) in both couples, the fluorescence intensity of couple 2 exhibited a larger decrease in PL intensity than that of couple 1, indicating an enhanced electron transfer within couple 2, which was likely caused by the larger distance between the two HOMO levels (Fig. S2 †).
To further study the photo-induced charge transfer within the couples, time-resolved photoluminescence (TRPL) spectroscopy was conducted to characterize the photo-excited dynamics within the OS couples. The PL measurements of the pristine materials and a 1 : 1 (mol%) blend were all performed in DMF solution with excitation at 400 nm ( Fig. 2c and d). The selected PL decays correspond to the emission maximum for each component OS (Fig. 2c). The fluorescence features of the different component materials overlap with each other. Therefore, the decay associated spectra (DAS) were derived by a global fitting analysis 41 of the TRPL data (Fig. S2 †). This allows us to characterize the spectral distribution and lifetime of each emitting species, particularly for the measurements of couple 1 and couple 2. In couple 1 (Fig. S2a †), one can see that DAS 1, centered at 460 nm, matches exactly with the fluorescence spectrum of Th-BTz-Th and decays in 0.6 ns, which is quenched with respect to the pristine donor (2.6 ns). DAS 2, which is centered at 590 nm, matches the fluorescence spectrum of Th-BT-Th and decays with the same ∼12 ns lifetime as the pristine Th-BT-Th sample, exhibiting no quenching. This indicates that in couple 1 the excitation energy transfer (EET) from Th-BTz-Th to Th-BT-Th likely dominates the excited state dynamics, as illustrated in Fig. 2e. In couple 2 ( Fig. S2b †), the DAS 1, centered at 460 nm, originates again from Th-BTz-Th and decays in 0.97 ns, which is quenched with respect to the pristine Th-BTz-Th (2.6 ns). DAS 2, centered at 530 nm, is assigned to the emission of TA-BT-TA and has a decay time of 10.6 ns. Here, the fluorescence of both Th-BTz-Th and TA-BT-TA is quenched with respect to the pristine materials. This indicates an electron transfer from Th-BTz-Th to TA-BT-TA and an enhanced charge transfer within couple 2 (Fig. 2f).
We chose the photocatalytic carbon-carbon formation reaction between electron-rich heteroarenes and malonates as a model reaction. 9,10 Previous studies demonstrated the mandatory use of sacrificial reagents, mostly amines, which ought not to only act as electron donors, but also as electron mediators between the radical intermediate and the photogenerated hole of the photocatalyst (Fig. S4 †). We first tested the three designed OSs as single photocatalyst systems without using sacrificial reagents. As expected, by using 3-methylbenzofuran and diethyl bromomalonate in the model reaction, no product was obtained using Th-BT-Th and TA-BT-TA as photocatalysts (entries 1 and 2 in Table 1). Interestingly, using Th-BTz-Th as a photocatalyst led to a slight conversion of 9%. This indicates that a minimal electron transfer could still occur between the photocatalyst and the substrate, which corresponds to the highest LUMO level of Th-BTz-Th among the three OSs and thereby the largest overpotential for the reduction of diethyl bromomalonate.
In contrast, by using both cooperative photocatalyst couples, the product could be obtained in significantly higher conversions, in particular, 43% for couple 1 and 93% for couple 2 (entries 4 and 5). Both the apparent quantum yield (Φ AQY ) and catalytic turnover number (TON) for the model reaction (entry 5) using coupling 2 as a photocatalyst were calculated to be 0.04% and 8.7, respectively (ESI †). A kinetic study confirmed the superior catalytic activity of couple 2 compared to couple 1 via a fluorescence quenching experiment by the substrates (Fig.  S6 †). This is in agreement with the photophysical study ( Fig. 2), indicating that the enhanced charge transfer between Th-BTz-Th and TA-BT-TA in couple 2 led to a better catalytic efficiency than that of couple 1 containing Th-BTz-Th and Th-BT-Th. The reason could lie on the larger HOMO energy gap (ΔE HOMO = 0.59 V) between Th-BTz-Th and TA-BT-TA than that (ΔE HOMO = 0.17 V) in couple 1 between Th-BTz-Th and Th-BT-Th. 42,43 By using a single wavelength light source (blue LED, 450 nm), where only TA-BT-TA of couple 2 could be excited, the model reaction exhibited a high conversion of 98%. This indicates a hole transfer from TA-BT-TA to Th-BTz-Th, leading to a similar effect of an electron transfer from Th-BTz-Th back to TA-BT-TA, as illustrated in Fig. 2f.
In comparison with the use of Th-BT-Th as a single photocatalyst with the employment of an extra sacrificial electron donor, in particular, triphenylamine from our previous report for the model reaction, 44 where the apparent quantum yield was calculated to be 0.96%, the apparent quantum yield of couple 2 was lower due to the longer reaction time. However, the avoidance of the extra sacrificial electron donor using the molecular photocatalyst couples is still advantageous. It can efficiently prevent the post-purification procedures caused by the addition of usually excess amounts of amine donors.
Screening experiments using different substrates confirmed the feasibility of the OS couple as a photocatalyst (entries 7-9). Other control experiments either without the photocatalyst or without light led to no product formation (entries 12 and 13), indicating the mandatory roles of both components.
To gain more mechanistic insight into the interaction between the photocatalyst and substrate, we conducted further fluorescence quenching experiments. It was shown that the fluorescence of the OSs as single photocatalysts could be only minimally quenched by adding diethyl bromomalonate (Fig.  S5 †). In contrast, the fluorescence of both photocatalyst couples could be gradually quenched by adding diethyl bromomalonate ( Fig. S6 †), confirming the possible electron transfer between the photocatalyst couple and the substrate.
To precisely study the actual electron transfer process within the catalytic cycle, we then employed ethyl bromoacetate as a substrate, which possessed a reduction potential In comparison, diethyl bromomalonate (E red. = −1.00 V vs. SCE) could be reduced via electron transfer originating from the LUMO of TA-BT-TA, leading to the formation of the malonate radical, which further reacts with 3-methylbenzofuran to form the final product in the following step of the catalytic cycle. An additional model reaction using Th-BTz-Th as the single photocatalyst system and triphenylamine as the sacrificial reagent led to the formation of the final product, confirming that a direct electron transfer from Th-BTz-Th to ethyl bromoacetate was only possible in the absence of TA-BT-TA (Fig. S12 †). It is also worth noting that further scope reactions with other substrates led to high product yields using couple 2 as a photocatalyst, demonstrating the general feasibility of the cooperative photocatalyst design. It is worth noting that a possible radical chain mechanism as introduced by Yoon et al. 45 could be clearly ruled out due to our previous study on the influence of the oxidation potential of the photocatalysts on the reaction rate. 10 The reason for the lower efficiency using the cooperative photocatalyst couple in comparison with the one catalyzed in the presence of sacrificial reagents might be caused by different factors. The multi-step electron transfer within the photocatalyst couple and further to the substrate might not be efficient due to the low concentration of the catalysts. Another important factor might be the electron transfer steps between the intermediates and the photogenerated hole. For this point, an additional study was conducted as described below.
The photo-induced electron transfer within the cooperative photocatalyst couple usually occurs based on the rather random contact between the two OSs in the liquid reaction medium. To create a non-random contact and thereby a more defined electron transfer process, we then connected Th-BTz-Th and TA-BT-TA via the C6 alkyl chain as a molecular bridge as a control photocatalyst. As displayed in Fig. 4c, a large molecule of Th-BTz-Th@TA-BT-TA containing both OSs in couple 2 was synthesized. Remarkably, the fluorescence intensity of Th-BTz-Th/TA-BT-TA strongly decreased after connection compared to their unconnected form as couple 2 (Fig. S8 †). This result suggests a strong interaction between the two OSs inside the large molecule. A study on the impact of Th-BTz-Th/TA-BT-TA in couple 2 on their photocatalytic efficiency showed that the highest reaction rate was achieved with a ratio of 2 : 1 of Th-BTz-Th/TA-BT-TA (Fig. 4a), which in fact corresponds to the connected molecule of Th-BTz-Th@TA-BT-TA. However, when comparing the catalytic efficiency between the connected catalyst Th-BTz-Th@TA-BT-TA and the unconnected couple 2 with the same ratio between both OSs, only a decreased conversion was observed for the connected catalyst (Fig. 4b).
In spite of the better electron transfer occurring within connected system, the reduced catalytic efficiency could be likely caused by the steric hindrance within the large molecule or the competing charge recombination process, 46 which limits the contact between the substrate and the individual OSs in the large molecule and efficient electron transfer between the catalyst and substrate. To overcome this limitation, new molecular bridges other than the C6 alkyl chain could be needed.

Conclusions
In conclusion, we have taken the light-induced electron transfer cascade process in natural photosystems as a role model and developed a new sacrificial reagent-free photoredox pathway by employing molecular organic semiconductors as cooperative photocatalyst couples. The photocatalyst couples consists of two molecular organic semiconductors with complementary HOMO/LUMO band positions, leading to enhanced photoinduced intermolecular charge transfer between the organic photocatalysts within the couple. By the cooperative photocatalyst design, no extra electron-donating sacrificial reagents are needed and a new paradigm in the photoredox reaction without undesired side reactions under the oxidation of the sacrificial reagents can be established. We believe that the cooperative photocatalyst design ought not to be limited in chemical transformation reactions; it can also be applied in the field of photocatalytic water splitting or CO 2 reduction.

Materials and methods
Chemicals were purchased from commercial sources and used as-received without further purification. 4,7-Dibromobenzoĳc]ĳ1,2,5]thiadiazole (97%) was bought from Combi-Blocks. Sodium tert-butoxide (97%), tert-butanol (99.5%), sodium borohydride (96%), 1-bromobutane (99%), 2-(tributylstannyl)thiophene (97%), ethynyltrimethylsilane (98%), sodium ascorbate (98%), PdĲPPh 3 )Cl 2 (98%), diethyl bromomalonate (92%), 3-methylbenzofuran (97%), 1-methylindole (97%) and 3-methylindole (98%) were bought from Sigma-Aldrich. Other commonly used solvents were bought from Fisher Scientific and VWR International. Glassware was dried under heating by a heating gun at 600°C and cooled under vacuum prior to use if necessary. All photocatalyst reactions were conducted using common dry, inert atmosphere techniques via a Schlenk tube. The white LED (silent air  Gmbh with an intensity of 0.07 Wcm −2 from a distance of 5 cm. UV-vis absorption and emission spectra were recorded on a Perkin Elmer Lambda 100 spectrophotometer and a J&M TIDAS spectrofluorometer at ambient temperature, respectively. Cyclic voltammetry measurement was performed with an Autolab PGSTAT204 potentiostat/galvanostat of Metrohm (a three electrode cell system): a glassy carbon electrode as the working electrode, a Hg/HgCl 2 electrode as the reference electrode, a platinum wire as the counter electrode, and Bu 4 NPF 6 (0.1 M Acetonitrile) as the supporting electrolyte with a scan rate of 100 mV s −1 in the range of −2 eV to 2.5 eV. 1 H and 13 C spectra were recorded on a Bruker Avance 300 spectrometer, and d-CHCl 3 was used as the solvent unless otherwise noted. GC-MS spectra were recorded on a SHIMADZU GCMS-QP2010 gas chromatograph mass spectrometer. Time-resolved photoluminescence (TRPL) on a nanosecond timescale was taken with a streak camera system (Hamamatsu C4742) in slow sweep mode. The excitation wavelength of 400 nm was provided using the frequencydoubled output of a commercial titanium:sapphire amplifier (Coherent LIBRA-HE, 3.5 mJ, 1 kHz, 100 fs).

Synthesis of the molecular organic semiconductor molecules
Synthesis of Th-BT-Th. 44 A 100 ml Schlenk tube equipped with a stirring bar and stopper was heated under vacuum then cooled three times before 4,7-dibromo-2,1,3-benzothiadiazole (1.76 g, 6.00 mmol), 2-(tributylstannyl)thiophene (4.76 ml, 15 mmol) and PdĲPPh 3 ) 2 Cl 2 (201.60 mg, 5 mol%) were added. 30 ml anhydrous THF was added via a syringe and the reactor was degassed under vacuum and then back-filled with nitrogen three times. The mixture was heated at 90°C overnight, then poured into 200 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a short column then crystallization from methanol give a needlelike reddish solid (1.14 g, 63%). 1  Synthesis of Th-BTz-Th.

Synthesis of 1a.
The suspension of 4,7-dibromobenzoĳc]-1,2,5-thiadiazole (2.94 g, 10 mmol) in 100 ml methanol was previously cooled to −5°C in a water-ice bath. Then sodium borohydride (8 g, 21.7 mmol) was added portion-wise to the suspension. The mixture was warmed up to room temperature and stirred overnight. Then, 500 ml water was added, the mixture was extracted with CH 2 Cl 2 , the organic layer was washed with water twice and dried over anhydrous MgSO 4 , and the solvent was removed under vacuum to give the final product as a pale yellow solid in 80% yield. The H-NMR data was consistent with the literature and the product used without further purification. 47 Synthesis of 1b. Dibromobenzene-1,2-diamine (2 g, 7.5 mmol) (1a) was added to a flask containing 15 ml glacial acetic acid. Sodium nitrite (1.03 g, 15 mmol) was dissolved in 10 ml water and added dropwise to the solution. The mixture was further reacted for another 3 hours before it was cooled with ice to 0°C. The suspension was filtered and washed with water until the pH reached 7. The crude product was dried in an oven at 70°C under vacuum overnight. The H-NMR data was consistent with the literature 1 and the product used without further purification.
Synthesis of 1c. To a 100 ml Schlenk tube was added 4,7dibromo-2H-benzoĳd]ĳ1,2,3]triazole (2.4 g, 8.6 mmol) (1b) and Nat-OBu (1.24 g, 12.9 mmol), then 20 ml methanol and 1-bromobutyl (4.3 ml, 40 mmol) were added via a syringe. The mixture was degassed under vacuum slightly then backfilled with N 2 three times. After being stirred at 90°C for 24 h, the mixture was poured into 200 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /hexane = 1 : 1 as the eluent gave the final product as a white solid (1.28 g, 45%). 1  Synthesis of Th-BTz-Th. A 100 ml Schlenk tube equipped with a stirring bar and stopper was heated under vacuum then cooled three times before 4,7-dibromo-2-(n-butyl)-2Hbenzoĳd]-[1,2,3]triazole (1 g, 3.00 mmol) (1c), 2-(tributylstannyl)thiophene (2.38 ml, 7.5 mmol) and PdĲPPh 3 ) 2 Cl 2 (100.8 mg, 5 mol%) were added. 15 ml anhydrous THF was added via a syringe and the reactor was degassed under vacuum and then back-filled with nitrogen three times. The mixture was heated at 90°C overnight, then poured into 200 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /hexane = 1 : 1 as the eluent gave the final product as a light red solid (713 mg, 70%). 1

View Article Online
Synthesis of TA-BT-TA.

Synthesis of 2a.
A 100 ml Schlenk tube equipped with a stirring bar and stopper was heated under vacuum then cooled three times before 4,7-dibromobenzothiadiazole (3 g, 10.2 mmol), PdĲPPh 3 ) 2 Cl 2 (143.2 mg, 2%), CuI (77 mg, 4%) were added. 30 ml THF and 10 ml NEt 3 were added via a syringe. The mixture was cooled to freeze in liquid N 2 , then trimethylsilyl acetylene (4 ml, 30.6 mmol) was added. The reactor was degassed via the freeze-pump-thaw method and the reaction mixture was stirred at 50°C overnight. Then the mixture was poured into 100 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /hexane = 1 : 1 as the eluent gave the final product as a light yellow solid (2.78 g, 83%). The H-NMR data was consistent with the literature reported. 48 Synthesis of 2b. In a 100 ml two neck flask was added KOH methanol solution (1 mol L −1 , 50 ml) and 4,7-bisĳ2-(trimethylsilyl)ethynyl]benzothiadiazole (1 g, 3 mmol) (2a). The solution was stirred at room temperature for 2 h, and then it was poured into 300 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /hexane = 1 : 1 as the eluent gave the final product as a light brown solid (420 mg, 876%). The H-NMR data was consistent with the literature reported. 48 TA-BT-TA. 4,7-Diethynylbenzothiadiazole (400 mg, 2.17 mmol) (2b), 1-azidobutane (synthesis of 1-azidobutane: a mixture of 10 mmol 1-bromobutane and 20 mmol sodium azide in 100 ml DMSO was stirred overnight at RT, then the reaction mixture was poured into 500 ml water and extracted with CH 2 Cl 2 then dried over anhydrous MgSO 4 , the solvent was removed under vacuum, and the mixture was used assynthesized without column purification. Caution: the excessive amount of sodium azide in the water phase must be slowly added to a solution of NaOCl before deposition). CuSO 4 (51.4 mg, 0.32 mmol), Na-ascorbate (87 mg, 0.44 mmol), 20 ml t-BuOH, 2 ml H 2 O were added into a 100 ml two neck flask. The reactor was degassed under vacuum and then back-filled with nitrogen three times. The mixture was heated at 60°C overnight, then poured into 200 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /acetone = 10:1 as the eluent then recrystallization from methanol gave the final product as a sheet-like yellow solid (747 mg, 90%). 1  Synthesis of the connected catalyst Th-BT-Th@TA-BT-TA.
Synthesis of 3a. To a 100 ml Schlenk tube was added 4,7dibromo-2H-benzoĳd]ĳ1,2,3]triazole (3 g, 10.8 mmol) and Nat-OBu (1.56 g, 16.2 mmol), then 40 ml methanol and 1,6dibromohexane (8.3 ml, 54 mmol) were added via a syringe. The mixture was degassed under vacuum slightly then backfilled with N 2 three times. After being stirred at 90°C for 24 h, the mixture was poured into 200 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /hexane = 1 : 1 as the eluent gave the final product as a white solid (1.82 g, 41%). 1  Synthesis of 3b. 4,7-Dibromo-2-(bromohexane)benzoĳd]ĳ1,2,3]triazole (1.8 g, 2.27 mmol) (3a) and sodium azide (295 mg, 4.54 mmol) were dissolved in 20 ml DMSO and stirred under room temperature overnight. Then the mixture was poured into 200 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. The product was used directly for the next step without any purification.
Synthesis of 3c. In a 100 ml two neck flask was added 4,7diethynylbenzothiadiazole (184 mg, 1 mmol) (3b), 4,7dibromo-2-(bromohexane)-benzoĳd]ĳ1,2,3]triazole, CuSO 4 (24 mg, 0.15 mmol), Na-ascorbate (41 mg, 0.21 mmol), 10 ml t-BuOH, and 1 ml H 2 O. The reactor was degassed under vacuum and then back-filled with nitrogen three times. The suspension was stirred under heating at 60°C for two days, then it was poured into 100 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 / acetone = 20 : 1 as the eluent gave the final product as a yellow solid (750 mg, 76%). 1  Th-BT-Th@TA-BT-TA. In a capped vial, 3c (750 mg, 0.76 mmol), 2-(tributylstannyl)thiophene (1.64 ml, 4.56 mmol) and PdĲPPh 3 ) 2 Cl 2 (53 mg, 10 mol%) were added. 10 ml anhydrous THF was added via a syringe and the reactor was degassed via bubbling nitrogen for 5 min. The vial was put in a microwave reactor (120°C, 0.8 T) for 1 h, then poured into 100 ml water and extracted with CH 2 Cl 2 three times. The organic layer was washed with water twice and dried over anhydrous MgSO 4 ; the solvent was removed under vacuum. Purification with a column using CH 2 Cl 2 /acetone = 20 : 1 as the eluent gave the final product as a yellow-red solid (403 mg, 53%). 1  General procedure for the electron-donating sacrificial reagent-free and photocatalytic C-C bond formation reactions using cooperative photocatalyst systems A 25 ml Schlenk tube equipped with a stirring bar and stopper was heated under vacuum then cooled for one hour and then back-filled with argon before the electron-rich heteroaromates (3-methyl benzofuran, etc.) (0.38 mmol, 1 equiv.), cooperative photocatalyst couple system (declared amount and ratio as described in Table 1), diethyl bromomalonate (0.76 mmol, 2.0 equiv.), 2.5 ml DMF were added. The reactor was degassed via the freeze-pump-thaw method and irradiated under a white LED lamp (0.07 W cm −2 ) in a water bath (see the photograph in the ESI †) for 24 h. The conversion was determined via GC-MS and the pure product was obtained via chromatography.
Comparison control experiment for the C-C bond formation reactions between 3-methylbenzofuran and ethyl bromoacetate using Th-BTz-Th as a catalyst and 4-methoxyltriphenylamine as an electron-donating sacrificial reagent A 25 ml Schlenk tube equipped with a stirring bar and stopper was heated under vacuum then cooled for one hour and then back-filled with argon before 3-mehtylbenzofuran (0.38 mmol, 1 equiv.), 4-methoxyltriphenylamine (0.76 mmol, 2 equiv.), ethyl bromomalonate (0.76 mmol, 2.0 equiv.), Th-BTz-Th (1 mol%), and 2.5 ml DMF were added. The reactor was degassed via the freeze-pump-thaw method and irradiated under a white LED. The reaction time was determined by GC-MS when the signal of the starting material totally disappeared. After the reaction was completed, the mixture was poured into 20 ml water and extracted with CH 2 Cl 2 , the organic layer was dried over anhydrous MgSO 4 , and the solvent was removed under vacuum. The crude product was purified on silica gel using the indicated solvent system to obtain the desired product. 1

Conflicts of interest
The authors declare no conflict of interest.