meso-Antracenyl-BODIPY Dyad as a New Photocatalyst in Atom-Transfer Radical Addition Reactions.

We demonstrate that because of the efficient generation of triplet excited state under UV or visible-light irradiation, meso-antracenyl-BODIPY donor-acceptor dyad can catalyze atom-transfer radical addition (ATRA) reactions between bromomalonate and alkenes. This finding paves the way for the design and application of the new type of heavy atom-free organic chromophores for photocatalysis.

Introduction

In the past few decades, the field of photocatalysis attracted much attention among synthetic chemists. Photocatalysis, as a method of activation of organic molecules, relies on the electron or energy transfer between photoexcited catalysts and substrates and allows bond formation and breakage under mild conditions. Though the most used photocatalysts—ruthenium or iridium polypyridyl complexes—have already proved their efficiency in multiple reactions,[1−5] their high cost, laborious synthesis, and generation of heavy-atom wastes prompted chemists toward the search and development of new photocatalytic systems.[6−15] Among them, there are several examples of the application of metal-free organic chromophores as photocatalysts in the reactions of oxidation,[16−20] reductive dehalogenation,[21] C–C bond formation,[22−26] and others.[27−30] For the majority of catalysts, either organometallic complexes or metal-free chromophores, the ability to form long-lived excited states has been postulated as a prerequisite for the efficient photocatalytic activity.[31,32] Within this context, we turned our attention to the recently reported orthogonal electron donor–acceptor (D–A) boron dipyrromethene (BODIPY)-based dyads.[33] In these dyads, BODIPY chromophores are usually substituted at the meso-position with another π-conjugated “dye,” e.g., anthracene (Ant), so that the planes of BODIPY and the “dye” form an angle close to 90° (Figure a). The dyads possess long-lived triplet excited states of BODIPY, when selectively excited at either BODIPY or the “dye” unit of the dyad. Notably, in contrast to many other organic and organometallic compounds, the formation of triplet excited states in “dye”-BODIPY dyads does not require the presence of heavy atoms in the structure of chromophores. In this case, the generation of the triplet excited state of BODIPY (T1) proceeds through the initial formation of the charge-separated state (1CS) within the dyad, where BODIPY unit usually serves as an electron acceptor and the “dye” as a donor (Figure b). Importantly, the spin-forbidden 1CS → T1 transition, termed spin-orbit charge-transfer intersystem crossing (SOCT-ISC), is favored when the molecular orbitals of the donor (“dye”) and acceptor (BODIPY) are orthogonal, creating a large enough orbital angular momentum for the spin flip.[34−36] The quantum yields of the triplet excited-state formation can reach 90% and above in the dyads with the properly matching photophysical parameters of the BODIPY and the “dye” units (e.g., reduction potentials, excitation energies, etc.). This trend of “dye”-BODIPY dyads to the efficient formation of long-lived triplet excited states encouraged us to explore the possibility of their application as photocatalysts in chemical reactions.

Figure 1

(a) Molecular structures and (b) photophysical scheme of catalysts upon photoexcitation of anthracene. We can also selectively photoexcite BODIPY to directly produce S1[BODIPY]*. EnT = energy transfer; CS = charge separation; bCR = back charge recombination; CRS = singlet charge recombination; DS = singlet decay; and DT = triplet decay.

Results and Discussion

Among various types of chemical transformations, the C–C bond formation seemed to us the most attractive, as it constitutes the fundamental basis for the creation of new organic frameworks. When choosing the reaction, we also considered the fact that catalytic cycles should include the oxidation or reduction of the photoexcited catalyst by the substrate. While excited states of ruthenium or iridium complexes can be easily oxidized or reduced, the oxidation/reduction of the triplet excited state of BODIPY might be challenging due to the lower reduction potentials. However, this issue can be overcome if a sacrificial reductant or oxidant is used to convert the excited state of the catalyst. Based on these considerations, we chose atom-transfer radical addition (ATRA) as a model reaction to test the photocatalytic performance of “dye”-BODIPY dyad. Photoredox ATRA reactions between olefins and various halogen-containing reagents have been thoroughly studied by Stephenson group,[37] using Ru(bpy)32+ as a photocatalyst and a wide range of sacrificial reductants from alkyl- and aryl-amino derivatives to sodium ascorbate. Moreover, recently it was shown that diiodo-BODIPY in combination with sodium ascorbate can also catalyze ATRA reactions.[38] Though the last finding seemed reassuring, we anticipated that the “dye”-BODIPY dyads might perform differently in ATRA reactions due to the different mechanism of the triplet state formation. Indeed, it was shown that the quantum yields of the triplet excited-state formation in various “dye”-BODIPY dyads are largely determined by the energy of 1CS, which in turn depends on the polarity of the solvent. Thereby, any factors, which make up the reaction microenvironment, e.g., substrates, solvent mixtures, and additives can potentially affect the efficiency of photocatalysis.

We started the investigation with the reaction between 5-hexen-1-ol (2) and diethyl bromomalonate (Scheme ) and have chosen antracenyl-BODIPY (1a) and 9-methylantracenyl-BODIPY (1b) as the catalysts, due to the high quantum yields of the triplet excited-state formation (ΦT = 0.93 and 0.90 in acetonitrile, AcCN, respectively) and quite high reduction potentials (BODIPY/BODIPY•– = −1.01 V vs saturated calomel electrode (SCE)).[33] All reactions were performed in deoxygenated solutions under Ar, using 350 nm excitation light source, which selectively excited anthracene unit in the dyads 1a and 1b. As the solvents, we have chosen N,N-dimethylformamide (DMF), dimethyl sulfoxide (DMSO), and AcCN, since the triplet excited states in the dyads 1a and 1b form with the highest efficiency in polar solvents.[33] No product 12 formation was observed when the reaction was performed in DMF/H2O, 1:1 mixture, using triethylamine or 4-methoxy-N,N-diphenylaniline as sacrificial reductants; however, the product was obtained with high yield of 87% when sodium ascorbate was used (Table , entry 1). Since we noticed a sufficient aggregation of the catalyst upon addition of water to the reaction mixture, the amount of water was reduced to solvent/water, 3:1 ratio. In DMF/H2O, 3:1 mixture, the product was formed with 86% yield (Table , entry 3); quite good yield of 66% was obtained in DMSO/H2O, 3:1 (Table , entry 4); and almost quantitative yield of 96% was obtained in AcCN/H2O, 3:1 mixture (Table , entry 5). Our attempts to improve the yield in DMSO/H2O, 3:1 by varying the reaction time were unsuccessful (Scheme S1 and Table S1, Supporting Information). In fact, the highest yield of 66% was obtained, when the reaction proceeded for 24 h, and the yield decreased to 57% and 48% in the case of longer (30 h) or shorter (18 h) reaction time, respectively. In principle, these results were not surprising, since we anticipated that the yield of the product should depend on the solvent polarity. To get more insights on our observations, we measured the quantum yields of the triplet excited-state formation for the dyad 1 in DMF/H2O, AcCN/H2O, and DMSO/H2O, 3:1, and for all solvent mixtures, the obtained values were about the same and equal to ΦT > 0.90. These results do not explain why in DMSO/H2O mixture the yield of the product is lower. Probably, other factors play a crucial role in this case, and more mechanistic studies are required to shed light onto this issue. We then performed the reaction in AcCN/H2O, 3:1 without the addition of LiBr to check whether it affects the reaction efficiency (Scheme S2 and Table S2, Supporting Information). The obtained yield of product 12 was c.a. 2 times lower than for the case when LiBr was present in the reaction mixture. This observation is in agreement with Stephenson’s hypothesis that the addition of Lewis acids facilitates the carbon–halogen bond breakage and the formation of radicals in ATRA reactions.[39]

Scheme 1

ATRA Reactions between Diethyl Bromomalonate and Various Olefins

Table 1

ATRA Reactions of Various Olefins under Different Conditions

#	olefin	product	solvent	cat./load	time (h)	yield (%)
1	2	12	DMF/H2O, 1:1	1a, 1 mol %	24	87
2	2	12	DMF/H2O, 1:1	1b, 1 mol %	24	88
3	2	12	DMF/H2O, 3:1	1b, 1 mol %	24	86
4	2	12	DMSO/H2O, 3:1	1a, 1 mol %	24	66
5	2	12	AcCN/H2O, 3:1	1a, 1 mol %	24	96
6	3	13	AcCN/H2O, 3:1	1a, 1 mol %	24	99
7	3	13	AcCN/H2O, 3:1	1b, 1 mol %	24	99
8	4	14	AcCN/H2O, 3:1	1b, 2 mol %	48	94
9	4	14	DMF/H2O, 1:1	1b, 2 mol %	48	90
10	4	14	DMSO/H2O, 3:1	1a, 1 mol %	24	15
11	5	15	AcCN/H2O, 3:1	1a, 2 mol %	48	40
12	5	15	DMF/H2O, 1:1	1b, 2 mol %	48	19
13	5	15	DMSO/H2O, 3:1	1a, 2 mol %	48	0
14	6	16	AcCN/H2O, 3:1	1a, 1 mol %	48	15
15	7	17	AcCN/H2O, 3:1	1a, 2 mol %	48	62
16	8	18	AcCN/H2O, 3:1	1a, 2 mol %	72	56
17	9	19 + 19a	AcCN/H2O, 3:1	1b, 2 mol %	48	46
18	10	20	AcCN/H2O, 3:1	1a, 2 mol %	48	82
19	11	21	AcCN/H2O, 3:1	1a, 2 mol %	48	10

Next, we studied the scope of the reaction in application to various functionalized alkenes (Scheme and Table ). The reaction between bromomalonate and terminal alkenes proceeded well in the presence of hydroxy-, aldehyde, cyano-, alkoxycarbonyl-, and protected amino groups. As regard solvents, the best yields were obtained in AcCN/H2O, 3:1 mixture, as compared to DMF/H2O, 1:1 and DMSO/H2O, 3:1 (Table , entry 9 vs 10 vs 11; entry 12 vs 13 vs 14). In addition, in DMF/H2O, we detected the formation of the oxidized byproducts in some cases (Scheme S1 and Table S1, Supporting Information), and in the case of DMSO/H2O—the formation of a higher amount of debrominated malonic ester. This trend is similar to what we observed initially for the reaction between 5-hexen-1-ol and diethyl bromomalonate. In respect to alkene substrates, the reaction proceeded well with terminal alkenes but resulted in poor yields with secondary alkenes, e.g., only 10% yield was obtained for cyclohexene (Table , entry 19). For terminal alkenes, almost quantitative yields were obtained when the functional groups were distant from the alkene fragment (Table , entries 5 and 6). The presence of functional groups and/or alkyl substituents in the vicinity of the alkene resulted in lower, yet good yields. In addition, a longer reaction time was required in the case of bulky substrates. We assume that in this case, both steric and electronic factors come to interplay. Thereby, the presence of highly electron-withdrawing cyanosubstituent at α-position to alkene group led to a moderate 15% yield of the product (Table , entry 14).

For the substrate 7 with the less electron-withdrawing hydroxy group at α-position, the yield was 62% (Table , entry 15), and for the substrate 5 with electron-withdrawing aryl substituent at α-position, 40% yield was obtained (Table entry 11). Methyl groups at α-position to the alkene fragment in substrates 8 and 9 led to 56% and 46% yields, respectively (Table , entries 16 and 17). Interestingly, a high yield of 82% was obtained for the substrate 10, despite the presence of the electron-withdrawing OH group at α-position to alkene, and a methyl substituent at the double bond. This result can be attributed to the intermediate formation of stable tertiary radical or tertiary carbocation during the reaction and is in a good agreement with the proposed mechanism.

We hypothesized that the mechanism of the ATRA reaction, catalyzed by the dyads 1a and 1b should be similar to the one proposed for diiodo-BODIPY,[38] except that the triplet excited state of BODIPY in the dyads 1a and 1b forms from the 1CS state (Scheme ). The participation of the triplet excited state in the reaction is supported by the fact that the reaction did not proceed in the presence of oxygen, which quenches the triplet states. Thereby, upon excitation of anthracene unit at λex = 350 nm, ultrafast energy transfer from S1[Ant]* to S1[BODIPY]* occurs, which is followed by an electron transfer within the dyad 1a or 1b, resulting in the formation of 1CS state [Ant•+-BODIPY•–]. Then, T1[BODIPY]* forms as a result of the SOCT-ISC process. In principle, T1[Ant]* could also form. However, as it was shown earlier for the dyad 1a, the majority of the triplet excitation energy is on BODIPY, and the contribution of T1[Ant]* is small.[40] Single electron transfer (SET) from the sacrificial reductant—sodium ascorbate—converts the triplet excited-state T1[BODIPY]* into radical-anion BODIPY•–, which has high enough reduction potential (BODIPY/BODIPY•– = −1.01 V vs SCE)[33] to convert bromomalonate into malonic radical. Malonic radical further undergoes the addition of alkenes with the formation of the corresponding radical. The ATRA product could form through the abstraction of halogen from another equivalent of bromomalonate (radical propagation pathway, A), or through single-electron oxidation to carbocation with subsequent trapping by bromine anion (radical-polar crossover pathway, B), as proposed by Stephenson.[37]

Scheme 2

Proposed Mechanism for ATRA Reaction Catalyzed by Ant-BODIPY Dyad

Hypothetically, the malonic radical could form through the SET from the 1CS state [Ant•+-BODIPY•–], followed by the reduction of Ant•+ unit with sodium ascorbate. However, transient absorption spectroscopy studies eliminated this possibility by showing that the lifetime of the 1CS state is too short to undergo intermolecular reactions (more details in the SI), and further confirming that BODIPY•– is produced by the reduction of T1[BODIPY]* by sodium ascorbate (Figure ).

Figure 2

(a) Transient absorption spectra of 1a at 10 μs after the ns laser pulse (λex = 355 nm) in AcCN/H2O = 3:1 in the presence and absence of sodium ascorbate. (b) Decay kinetics of 1a in AcCN/H2O = 3:1 in the presence of sodium ascorbate at the T1[BODIPY]* absorption band (430 nm) and the BODIPY•– absorption band (570 nm).

Finally, we confirmed that the ATRA reactions proceed with the same efficiency under visible-light irradiation. Thereby, the reaction between diethyl bromomalonate and substrates 2 and 3 in AcCN/H2O, 3:1 mixture, and using catalyst 1b and 470 nm excitation light source led to products 12 and 13 with 98 and 99% yields, respectively (Scheme S3 and Table S3, Supporting Information). In this case, a BODIPY unit was selectively excited at λex = 470 nm to produce S1[BODIPY]*. The remaining steps of the mechanism are analogous to the described above for 350 nm excitation.

Conclusions

In conclusion, we demonstrated that BODIPY-based donor–acceptor dyads can efficiently catalyze light-driven ATRA reactions. The dyads can be excited with UV and visible light. Given the ease of the synthesis, availability of synthetic precursors, and readily tunable photophysical and redox properties, these dyads can become a new inexpensive environmentally friendly alternative to classical organometallic photocatalysts.

Experimental Section

General Experimental Procedures

All solvents and reagents were purchased from commercial sources and used as received. Column chromatography was performed on silica gel (pore size 60 Å, particle size 40–63 μm, Sigma-Aldrich). Analytical thin-layer chromatography (TLC) was carried out using aluminum plates covered with silica gel and containing a fluorescent indicator (layer thickness 200 μm, particle size 25 μm, Sigma-Aldrich). NMR spectra were recorded on the AVANCE III HD Bruker 500 spectrometer operating at 500 MHz (1H) and 125 MHz (13C). Photoreactions were carried out using an RPR-100 photochemical reactor (The Southern New England Co. Inc.), equipped with eight lamps (λex = 350 or 470 nm) and a stirring plate. Prior each reaction, the reaction mixtures were degassed by three to five freeze–pump–thaw cycles and filled with Ar, using the standard Schlenk technique.

Optical Measurements

UV–vis absorption spectra were recorded on a Cary 50 Scan UV–vis spectrophotometer (Varian). Fluorescence measurements were performed on an FLS1000 spectrofluorometer (Edinburgh Instruments, U.K.). The femtosecond and nanosecond transient absorption (fsTA and nsTA) setups were described elsewhere.[33] In short, we used 355 nm, a third harmonic of Nd:YAG (adjusted to 1–4 mJ), and 500 nm (output from an optical parametric amplifier, adjusted to 0.5 μJ) as an excitation pump for nsTA and fsTA, respectively. The samples for nsTA measurements were degassed by three to five freeze–pump–thaw cycles. Quantum yields of triplets were measured using a relative actinometry method compared with the formation of triplets of benzophenone in acetonitrile (MeCN) (ΦT = 1.0).[41] Data sets were analyzed using OriginPro 2017 (OriginLab Corporation).

General Procedure for Photocatalytic ATRA Reactions

A 15 mL Schlenk tube was equipped with a Rotaflo stopcock, magnetic stir bar, and was charged with photocatalyst 1a or 1b (1–2 mol %, 1.2–2.5 mg, 0.0028–0.0057 mmol), sodium ascorbate (19.41 mg, 0.098 mmol, 0.35 equiv), LiBr (48.63 mg, 0.56 mmol, 2 equiv), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol, 2 equiv), olefin 2–11 (0.28 mmol, 1 equiv), and water (0.15 mL). The heterogeneous reaction mixture was degassed (freeze–pump–thaw cycle × 3–5), and the vessel was refilled with Ar. The reaction mixture was tightly sealed and positioned approximately (5–7 cm) away from the light source (λ = 350 nm), where it vigorously stirred for 24–72 h. The reaction mixture was diluted with CH2Cl2 (10 mL), washed with brine (1 mL), and dried over Na2SO4. The solvent was removed in vacuum, and the residue was purified by column chromatography on silica gel, using the solvent system indicated, to obtain the desired products in the stated yields.

Experimental Procedures

Diethyl 2-(2-Bromo-6-hydroxyhexyl)malonate (12)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (1.2 mg, 1 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), 5-hexen-1-ol (2) (28.04 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 24 h. The residue was purified by column chromatography (silica gel, hexanes/ethyl acetate (AcOEt), gradient from 2:1 to 1:1) to give the title compound 12 as orange oil. Yield 91 mg (96%). All characterizations are consistent with those reported in literature.[39] NMR 1H (CDCl3), δ, ppm: 1.25 (3H, t, 3J = 7.2 Hz, −CH3), 1.26 (3H, t, 3J = 7.0 Hz, −CH3), 1.45–1.68 (5H, m, −CH2−), 1.83–1.91 (2H, m, −CH2−), 2.23 (1H, ddd, J1 = 14.8, J2 = 10.7, J3 = 4.2 Hz, −CH2−), 2.44 (1H, ddd, J1 = 13.4, J2 = 10.2, J3 = 3.2 Hz, −CH2−), 3.63 (2H, t, 3J = 6.2 Hz, −CH2−), 3.76 (1H, dd, 3J1 = 10.3, 3J2 = 4.2 Hz, −CH−), 3.95–4.03 (1H, m, −CHBr−), and 4.14–4.26 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.97, 14.01, 23.7, 31.9, 37.8, 39.1, 50.5, 54.6, 61.62, 61.69, 168.8, and 169.0.

Diethyl 2-(2-Bromo-6-hydroxyhexyl)malonate (12*)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (1.2 mg, 1 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), DMF (0.3 mL), diethyl bromomalonate (134 mg, 0.56 mmol), 5-hexen-1-ol (2) (28.04 mg, 0.28 mmol), and water (0.3 mL). The reaction mixture was stirred for 24 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt, gradient from 2:1 to 1:1) to give the title compound 12 as orange oil and compound 12a as red oil. 12: Yield 83 mg (87%). NMR 1H (CDCl3), δ, ppm: 1.25 (3H, t, 3J = 7.2 Hz, −CH3), 1.26 (3H, t, 3J = 7.0 Hz, −CH3), 1.45–1.68 (5H, m, −CH2−), 1.83–1.91 (2H, m, −CH2−), 2.23 (1H, ddd, J1 = 14.8, J2 = 10.7, J3 = 4.2 Hz, −CH2−), 2.44 (1H, ddd, J1 = 13.4, J2 = 10.2, J3 = 3.2 Hz, −CH2−), 3.63 (2H, t, 3J = 6.2 Hz, −CH2−), 3.76 (1H, dd, 3J1 = 10.3, 3J2 = 4.2 Hz, −CH−), 3.95–4.03 (1H, m, −CHBr−), and 4.14–4.26 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.97, 14.01, 23.7, 31.9, 37.8, 39.1, 50.5, 54.6, 61.62, 61.69, 168.8, and 169.0. 12a: Yield 4.5 mg (5%). NMR 1H (CDCl3), δ, ppm: 1.23–1.29 (6H, m, −CH2CH3), 1.44–1.96 (5H, m, −CH2−), 2.18–2.29 (1H, m, −CH2−), 2.40–2.50 (2H, m, −CH2−), 3.74–3.79 (1H, m, −CH−), 3.95–4.02 (1H, m, −CHBr−), 4.14–4.30 (4H, m, −CH2CH3), and 9.76–9.78 (1H, m, −CHO).

1,1-Diethyl 6-Methyl 3-bromohexane-1,1,6-tricarboxylate (13)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (1.2 mg, 1 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), methyl 5-hexenoate (3) (35.89 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 24 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt = 3:1), to give the title compound 13 as yellow oil. Yield 101.5 mg (99%). All characterizations are consistent with those reported in literature.[39] NMR 1H (CDCl3), δ, ppm: 1.22–1.29 (6H, m, −CH2CH3), 1.70–1.81 (1H, m, −CH2−), 1.81–1.94 (3H, m, −CH2−), 2.23 (1H, ddd, J1 = 14.8, J2 = 10.7, J3 = 4.1 Hz, −CH2−), 2.28–2.36 (2H, m, −CH2−), 2.43 (1H, ddd, J1 = 13.6, J2 = 10.3, J3 = 3.1 Hz, −CH2−), 3.65 (3H, s, broad, −OCH3), 3.75 (1H, dd, 3J1 = 10.2, 3J2 = 4.1 Hz, −CH−), 3.93–4.01 (1H, m, −CHBr−), and 4.14–4.28 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.87, 13.91, 22.7, 33.0, 37.6, 38.4, 50.4, 51.4, 53.8, 61.5, 61.6, 168.6, 168.7, and 173.3.

Diethyl 2-(2-Bromo-3-((tert-butoxycarbonyl)amino)-propyl)malonate (14)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1b (2.5 mg, 2 mol %), sodium ascorbate (19.02 mg, 0.096 mmol), LiBr (47.76 mg, 0.55 mmol), AcCN (0.45 mL), diethyl bromomalonate (131.48 mg, 0.55 mmol), tert-butyl allylcarbamate (4) (43.08 mg, 0.274 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt = 5:1), to give the title compound 14 as orange oil. Yield 101.5 mg (94%). All characterizations are consistent with those reported in literature.[39] NMR 1H (CDCl3), δ, ppm: 1.26 (3H, t, 3J = 7.2 Hz, −CH3), 1.27 (3H, t, 3J = 7.2 Hz, −CH3), 1.43 (9H, s, broad, −C(CH3)3), 2.21–2.29 (1H, m, −CH2−), 2.46 (1H, ddd, J1 = 13.3, J2 = 9.5, J3 = 3.7 Hz, −CH2−), 3.45–3.57 (2H, m, −CH2−), 3.73 (1H, dd, 3J1 = 9.5, 3J2 = 5.0 Hz, −CH−), 4.07–4.15 (1H, m, −CHBr−), 4.15–4.26 (4H, m, −CH2CH3), and 4.91–4.99 (1H, m, broad, −NH−). NMR 13C (CDCl3), δ, ppm: 13.9, 14.0, 28.2, 34.6, 47.0, 50.1, 53.2, 61.65, 61.71, 79.8, 155.6, 168.4, and 168.8.

Diethyl 2-(2-Bromo-3-(3-formyl-2-hydroxyphenyl)propyl)malonate (15)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (2.5 mg, 2 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), 3-allylsalicyaldehyde (5) (45.41 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt 5:1), to give the title compound 15 as yellow oil. Yield 45 mg (40%). NMR 1H (CDCl3), δ, ppm: 1.23 (3H, t, 3J = 7.0 Hz, −CH3), 1.25 (3H, t, 3J = 7.2 Hz, −CH3), 2.31 (1H, ddd, J1 = 15.0, J2 = 10.7, J3 = 4.4 Hz, −CH2−), 2.52 (1H, ddd, J1 = 13.4, J2 = 10.2, J3 = 3.2 Hz, −CH2−), 3.16 (1H, dd, 3J1 = 14.2, 3J2 = 8.2 Hz, −CH2−), 3.32 (1H, dd, 3J1 = 14.2, 3J2 = 5.7 Hz, −CH2−), 3.79 (1H, dd, 3J1 = 10.1, 3J2 = 4.3 Hz, −CH−), 4.12–4.24 (4H, m, −CH2CH3), 4.37–4.44 (1H, m, −CHBr−), 6.98 (1H, t, 3J = 7.6 Hz, Ar), 7.43–7.50 (2H, m, Ar), 9.87 (1H, s, −CHO), and 11.30 (1H, s, broad, −OH). NMR 13C (CDCl3), δ, ppm: 13.97, 13.99, 37.5, 39.6, 50.7, 52.4, 61.6, 61.7, 119.5, 120.4, 126.5, 132.8, 138.6, 159.7, 168.5, 168.9, and 196.7.

Diethyl 2-(2-Bromo-3-cyanopropyl)malonate (16)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (1.2 mg, 1 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), allyl cyanide (6) (19.35 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt 3:1), to give the title compound 16 as orange oil. Yield 13 mg (15%). NMR 1H (CDCl3), δ, ppm: 1.21–1.31 (6H, m, −CH2CH3), 2.29–2.40 (1H, m, −CH2−), 2.53 (1H, ddd, J1 = 13.6, J2 = 10.2, J3 = 3.2 Hz, −CH2−), 3.01 (2H, dd, 3J1 = 5.8, 3J2 = 3.0 Hz, −CH2−), 3.71 (1H, dd, 3J1 = 10.4, 3J2 = 4.1 Hz, −CH−), and 4.12–4.30 (5H, m, −CHBr–, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.98, 14.0, 28.6, 36.9, 43.7, 50.1, 61.99, 62.04, 116.0, 168.15, and 168.18.

Diethyl 2-(2-Bromo-3-hydroxypropyl)malonate (17)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (2.5 mg, 2 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), allyl alcohol (7) (16.26 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt 2:1), to give the title compound 17 as orange oil. Yield 51.5 mg (62%). All characterizations are consistent with those reported in literature.[42] NMR 1H (CDCl3), δ, ppm: 1.26 (3H, t, 3J = 7.0 Hz, −CH3), 1.27 (3H, t, 3J = 7.0 Hz, −CH3), 2.29–2.37 (2H, m, −CH2−), 2.48 (1H, ddd, J1 = 13.3, J2 = 9.6, J3 = 3.7 Hz, −CH−), 3.72 (1H, dd, 3J1 = 9.5, 3J2 = 5.0 Hz, −CHBr−), 3.74–3.84 (2H, m, −CH2−), and 4.11–4.27 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.96, 14.00, 33.6, 50.0, 55.4, 61.8, 61.9, 66.9, 168.6, and 169.0.

Diethyl 2-(2-Bromo-3-hydroxy-3-methylbutyl)malonate (18)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (2.5 mg, 2 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), 2-methyl-3-buten-2-ol (8) (24.12 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 72 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt 3:1), to give the title compound 18 as orange oil. Yield 51 mg (56%). NMR 1H (CDCl3), δ, ppm: 1.26 (6H, q, 3J = 7.0 Hz, −CH2CH3), 1.35 (3H, s, broad, −CH3), 1.37 (3H, s, broad, −CH3), 2.13 (1H, s, broad, −OH), 2.18 (1H, ddd, J1 = 15.0, J2 = 11.9, J3 = 3.4 Hz, −CH2−), 2.58 (1H, ddd, J1 = 13.3, J2 = 11.2, J3 = 2.0 Hz, −CH2−), 3.76 (1H, dd, 3J1 = 11.2, 3J2 = 3.3 Hz, −CH−), 4.03 (1H, dd, 3J1 = 11.8, 3J2 = 2.0 Hz, −CHBr−), and 4.14–4.28 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 14.00, 14.05, 25.8, 26.9, 33.3, 50.9, 61.7, 61.8, 67.3, 72.4, 168.7, and 168.9.

Diethyl 2-(2-Bromo-5-hydroxy-3-methylpentyl)malonate (19) and Diethyl [(3-Methyltetrahydrofuran-2-yl)methyl]malonate (19a)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1b (2.5 mg, 2 mol %), sodium ascorbate (19.02 mg, 0.096 mmol), LiBr (47.76 mg, 0.55 mmol), AcCN (0.45 mL), diethyl bromomalonate (131.48 mg, 0.55 mmol), 3-methylpent-4-en-1-ol (9) (27.44 mg, 0.274 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt 2:1), to give the mixture of the title compounds 19 and 19a as pink oil. Yield 43 mg (46%). NMR 1H (CDCl3), δ, ppm (mixture of 19 and 19a): 0.98–1.08 (3H, m, −CH3), 1.22–1.29 (6H, m, −CH2CH3), 1.46–1.61 (1H, m, −CH2−), 1.69–2.10 (2H, m, −CH2−), 2.18–2.44 (2H, m, −CH2–, −CH−), 2.98–3.08 (1H, m, broad, −OH), 3.62–3.85 (3H, m, −CH2–, −CH−), 4.03–4.13 (1H, m, −CHBr−), and 4.13–4.28 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.97, 14.01, 15.0, 17.0, 34.4, 35.70, 35.76, 35.83, 36.8, 38.0, 50.56, 50.59, 60.1, 60.5, 61.31, 61.34, 61.63, 61.70, 168.8, 168.9, 168.97, and 168.99.

Diethyl 2-(2-Bromo-3-hydroxy-2-methylpropyl)malonate (20)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (2.5 mg, 2 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), β-methallyl alcohol (10) (20.20 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt 3:1), to give the title compound 20 as orange oil. Yield 71.5 mg (82%). NMR 1H (CDCl3), δ, ppm: 1.21–1.30 (6H, m, −CH2CH3) 1.72 (3H, s, broad, −CH3), 2.35 (1H, dd, 3J1 = 15.6, 3J2 = 4.9 Hz, −CH2−), 2.61 (1H, dd, 3J1 = 15.6, 3J2 = 7.0 Hz, −CH2−), 2.72–2.81 (1H, m, broad, −OH), 3.51–3.61 (2H, m, −CH2−), 3.72 (1H, dd, 3J1 = 6.9, 3J2 = 4.9 Hz, −CH−), and 4.13–4.27 (4H, m, −CH2CH3). NMR 13C (CDCl3), δ, ppm: 13.9, 14.0, 28.7, 39.3, 49.8, 61.9, 62.1, 70.5, 70.6, 169.3, and 170.1.

Diethyl 2-(2-Bromocyclohexyl)malonate (21)

In accordance with the General Procedure, the reaction was performed using photocatalyst 1a (2.5 mg, 2 mol %), sodium ascorbate (19.41 mg, 0.098 mmol), LiBr (48.63 mg, 0.56 mmol), AcCN (0.45 mL), diethyl bromomalonate (134 mg, 0.56 mmol), cyclohexene (11) (23 mg, 0.28 mmol), and water (0.15 mL). The reaction mixture was stirred for 48 h. The residue was purified by column chromatography (silica gel, hexanes/AcOEt, gradient from 5:1 to 10:1), to give the title compound 15 as yellow oil. Yield 9 mg (10%). All characterizations are consistent with those reported in literature.[38] NMR 1H (CDCl3), δ, ppm (mixture of cis- and trans-isomers): 1.23–1.30 (6H, m, −CH2CH3), 1.30–1.58 (4H, m, −CH2−), 1.69–2.01 (3H, m, −CH2–, −CH−), 2.10–2.44 (2H, m, −CH2−), 3.47 (1H, d, 3J = 10.8 Hz, −CH−), 4.14–4.25 (4H, m, −CH2CH3), 4.70–4.75 (1H, m, -CHBr-). NMR 13C (CDCl3), δ, ppm (mixture of cis- and trans-isomers): 14.05, 14.08, 14.13, 20.2, 24.7, 25.3, 25.4, 27.2, 28.5, 34.8, 38.8, 42.0, 46.3, 54.3, 56.4, 57.4, 58.4, 61.1, 61.4, 61.50, 61.53, 167.9, 168.1, 168.4, and 169.1.