Enantioselective Protonation of Radical Anion Intermediates in Photoallylation and Photoreduction Reactions of 3,3-Diaryl-1,1-dicyano-2-methylprop-1-ene with Allyltrimethylsilane

Photoreactions of acetonitrile solutions of 3,3-diaryl-1,1-dicyano-2-methylprop-1-enes (1a–c) with allyltrimethylsilane (2) in the presence of phenanthrene as a photoredox catalyst and acetic acid as a proton source formed photoallylation (3) and photoreduction (4) products via photoinduced electron transfer pathways. When (S)-mandelic acid was used as the proton source, the reactions proceeded with 3.4 and 4.8 %ee for formation of 3 and 4, respectively. The results of studies of the effect of aryl ring substituents and several chiral carboxylic acids suggested that the enantioselectivities of the reactions are governed by steric controlled proton transfer in intermediate complexes formed by π-π and OH-π interactions of anion radicals derived from 1a–c and chiral carboxylic acids.


Introduction
Coupling reactions proceeding through photoinduced electron transfer (PET) pathways have been extensively studied from both a synthetic as well as a mechanistic viewpoint [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. Because radical ions that serve as intermediates in these processes are short-lived and highly reactive, control of the stereochemistry of these reactions is often difficult [15][16][17][18][19][20][21]. We have previously developed photoallylation and photoreduction reactions of electron deficient alkenes with allyltrimethylsilane that occur via PET pathways [22][23][24]. In addition, we also demonstrated that diastereoselectivity of this process can be achieved by steric control of allyl radical or proton addition to radical anions that are generated from electron deficient alkenes (Scheme 1) [25][26][27]. The current study was aimed at the development of enantioselective PET promoted coupling reactions, and specifically, at assessing the effect of chiral carboxylic acids on the stereochemical outcomes of photoallylation and photoreduction reactions of prochiral electron deficient alkenes. The results showed that these processes took place with maximum 3.4-4.8 %ee when (S)-mandelic acid was used as the chiral proton source.

Results and Discussion
Irradiation of an acetonitrile solution containing 1,1-dicyano-2-methyl-3,3-diphenylprop-1-ene (1a), 3 equiv of allyltrimethylsilane (2), a catalytic amount of phenanthrene (Phen) as a photoredox catalyst in a Pyrex vessel using a 300 W high-pressure mercury lamp was found to produce reduction product 4a in 53% yield (Scheme 2, Table 1, entry 1, supplementary). Photoreaction of 1a with 2 in the presence of acetic acid produced the allylated product 3a in addition to 4a in 34 and 31% yields, respectively (entry 2). The corresponding products 3b-c and 4b-c were produced in photoreactions of bis(p-methoxyphenyl) (1b) and bis(p-chlorophenyl) (1c) derivatives conducted under the same conditions (entries [3][4][5][6]. The irradiation times used for these processes are those required for complete consumption of 1a-c. The observed efficiencies of the reactions, based on the required irradiation times, decreased in the order 1c > 1a > 1b. Scheme 2. Photoallylation and photoreduction of 1a-c by using allyltrimethylsilane (2). Structures of photoproducts 3a-c and 4a-c were determined by using spectroscopic methods. In 1 H-NMR spectra of CDCl3 solutions of 3a and 4a (Figure 1), the chemical shifts of resonances for protons that are bonded to the asymmetric carbons, i.e., Hb in 3a and Hh in 4a, were 2.97 (qd) and 3.02 (qt) ppm, respectively. Authentic samples of the photoproducts were prepared by hydrogenation of 1a using Pd/C to form 4a and ensuing allylation of 4a using allyl chloride to form 3a (Scheme 3). The spectral data for the synthesized compounds were identical to those of photoproduced 3a and 4a. Scheme 1. Our previous works.

Results and Discussion
Irradiation of an acetonitrile solution containing 1,1-dicyano-2-methyl-3,3-diphenylprop-1-ene (1a), 3 equiv of allyltrimethylsilane (2), a catalytic amount of phenanthrene (Phen) as a photoredox catalyst in a Pyrex vessel using a 300 W high-pressure mercury lamp was found to produce reduction product 4a in 53% yield (Scheme 2, Table 1, entry 1, supplementary). Photoreaction of 1a with 2 in the presence of acetic acid produced the allylated product 3a in addition to 4a in 34 and 31% yields, respectively (entry 2). The corresponding products 3b-c and 4b-c were produced in photoreactions of bis(p-methoxyphenyl) (1b) and bis(p-chlorophenyl) (1c) derivatives conducted under the same conditions (entries [3][4][5][6]. The irradiation times used for these processes are those required for complete consumption of 1a-c. The observed efficiencies of the reactions, based on the required irradiation times, decreased in the order 1c > 1a > 1b.

Results and Discussion
Irradiation of an acetonitrile solution containing 1,1-dicyano-2-methyl-3,3-diphenylprop-1-ene (1a), 3 equiv of allyltrimethylsilane (2), a catalytic amount of phenanthrene (Phen) as a photoredox catalyst in a Pyrex vessel using a 300 W high-pressure mercury lamp was found to produce reduction product 4a in 53% yield (Scheme 2, Table 1, entry 1, supplementary). Photoreaction of 1a with 2 in the presence of acetic acid produced the allylated product 3a in addition to 4a in 34 and 31% yields, respectively (entry 2). The corresponding products 3b-c and 4b-c were produced in photoreactions of bis(p-methoxyphenyl) (1b) and bis(p-chlorophenyl) (1c) derivatives conducted under the same conditions (entries [3][4][5][6]. The irradiation times used for these processes are those required for complete consumption of 1a-c. The observed efficiencies of the reactions, based on the required irradiation times, decreased in the order 1c > 1a > 1b. Scheme 2. Photoallylation and photoreduction of 1a-c by using allyltrimethylsilane (2). Structures of photoproducts 3a-c and 4a-c were determined by using spectroscopic methods. In 1 H-NMR spectra of CDCl3 solutions of 3a and 4a (Figure 1), the chemical shifts of resonances for protons that are bonded to the asymmetric carbons, i.e., Hb in 3a and Hh in 4a, were 2.97 (qd) and 3.02 (qt) ppm, respectively. Authentic samples of the photoproducts were prepared by hydrogenation of 1a using Pd/C to form 4a and ensuing allylation of 4a using allyl chloride to form 3a (Scheme 3). The spectral data for the synthesized compounds were identical to those of photoproduced 3a and 4a. Scheme 2. Photoallylation and photoreduction of 1a-c by using allyltrimethylsilane (2). Structures of photoproducts 3a-c and 4a-c were determined by using spectroscopic methods. In 1 H-NMR spectra of CDCl 3 solutions of 3a and 4a (Figure 1), the chemical shifts of resonances for protons that are bonded to the asymmetric carbons, i.e., H b in 3a and H h in 4a, were 2.97 (qd) and 3.02 (qt) ppm, respectively. Authentic samples of the photoproducts were prepared by hydrogenation of 1a using Pd/C to form 4a and ensuing allylation of 4a using allyl chloride to form 3a (Scheme 3). The spectral data for the synthesized compounds were identical to those of photoproduced 3a and 4a. Molecules 2019, 24, x 3 of 11 In order to explore the enantioselectivities of these photoreactions, samples of 3a and 4a were subjected to HPLC using a chiral stationary phase with the effluents being monitored by using UV and CD detectors (Figure 2a-d). The results showed that two peaks in the HPLC trace for the enantiomers of 3a and 4a were completely resolved. Unfortunately, HPLC conditions could not be found for resolution of the enantiomers of 3b and 4b. Moreover, the enantiomers of 3c and 4c can be separated by using GC with a chiral capillary column (Figure 2e).  In order to explore the enantioselectivities of these photoreactions, samples of 3a and 4a were subjected to HPLC using a chiral stationary phase with the effluents being monitored by using UV and CD detectors (Figure 2a-d). The results showed that two peaks in the HPLC trace for the enantiomers of 3a and 4a were completely resolved. Unfortunately, HPLC conditions could not be found for resolution of the enantiomers of 3b and 4b. Moreover, the enantiomers of 3c and 4c can be separated by using GC with a chiral capillary column (Figure 2e).  In order to explore the enantioselectivities of these photoreactions, samples of 3a and 4a were subjected to HPLC using a chiral stationary phase with the effluents being monitored by using UV and CD detectors (Figure 2a-d). The results showed that two peaks in the HPLC trace for the enantiomers of 3a and 4a were completely resolved. Unfortunately, HPLC conditions could not be found for resolution of the enantiomers of 3b and 4b. Moreover, the enantiomers of 3c and 4c can be separated by using GC with a chiral capillary column (Figure 2e). In order to explore the enantioselectivities of these photoreactions, samples of 3a and 4a were subjected to HPLC using a chiral stationary phase with the effluents being monitored by using UV and CD detectors (Figure 2a-d). The results showed that two peaks in the HPLC trace for the enantiomers of 3a and 4a were completely resolved. Unfortunately, HPLC conditions could not be found for resolution of the enantiomers of 3b and 4b. Moreover, the enantiomers of 3c and 4c can be separated by using GC with a chiral capillary column (Figure 2e).  In order to prove that these separation techniques led to the individual enantiomers, the effluents of peaks A-D in Figure 2a-d were collected, concentrated in vacuo and the residues in ethanol were subjected to UV-vis absorption and CD spectroscopic analysis (Figures 3 and 4). The UV-vis absorption spectra of substances in effluents corresponding to peaks A and C were identical to those from peaks B and D, respectively. In addition, 1 H-NMR and mass spectra of the respective substances in peaks A and C were also identical to those in peaks B and D, respectively. Moreover, CD spectral traces of substances comprising peaks A and B, and peaks C and D, respectively, were mirror images relative to the horizontal base line. The combined results indicated that the enantiomers of these substances can be resolved by using chromatographic methods. In order to prove that these separation techniques led to the individual enantiomers, the effluents of peaks A-D in Figure 2a-d were collected, concentrated in vacuo and the residues in ethanol were subjected to UV-vis absorption and CD spectroscopic analysis (Figures 3 and 4). The UV-vis absorption spectra of substances in effluents corresponding to peaks A and C were identical to those from peaks B and D, respectively. In addition, 1 H-NMR and mass spectra of the respective substances in peaks A and C were also identical to those in peaks B and D, respectively. Moreover, CD spectral traces of substances comprising peaks A and B, and peaks C and D, respectively, were mirror images relative to the horizontal base line. The combined results indicated that the enantiomers of these substances can be resolved by using chromatographic methods.  To assess the potential of introducing enantioselectivity into the photoreactions described above, irradiations were carried out on solutions of 1a-c and allyltrimethylsilane (2) containing chiral carboxylic acids. The yields and percent enantiomeric excesses (%ee) of products formed in these processes are listed in Table 2. The %ee in each case was calculated using the ratio of areas under the chiral HPLC or GC peaks corresponding to the enantiomers as %ee when acetic acid was used becoming zero. A positive %ee value corresponds to a situation in which the major isomer is the second peak, while a negative value shows that the major isomer is the first peak. The absolute structures could not be decided. The data arising from photoreactions in the absence or presence of achiral acetic acid are also included in Table 2 for comparison purposes.
Use of 1 equiv of (R)-mandelic acid in photoreaction of 1a with 2 led to formation of 3a and 4a with respective +1.5 and +4.1 %ee values (entry 3). A reversal in major enantiomers of the products arose from the reaction of 1a with 2 conducted in the presence of (S)-mandelic acid (entry 4), which supports the reaction proceeding in an enantioselective manner. Also, when L-lactic acid was used in this photoreaction, the major enantiomers were the reverse of those formed in reactions in the  In order to prove that these separation techniques led to the individual enantiomers, the effluents of peaks A-D in Figure 2a-d were collected, concentrated in vacuo and the residues in ethanol were subjected to UV-vis absorption and CD spectroscopic analysis (Figures 3 and 4). The UV-vis absorption spectra of substances in effluents corresponding to peaks A and C were identical to those from peaks B and D, respectively. In addition, 1 H-NMR and mass spectra of the respective substances in peaks A and C were also identical to those in peaks B and D, respectively. Moreover, CD spectral traces of substances comprising peaks A and B, and peaks C and D, respectively, were mirror images relative to the horizontal base line. The combined results indicated that the enantiomers of these substances can be resolved by using chromatographic methods.  To assess the potential of introducing enantioselectivity into the photoreactions described above, irradiations were carried out on solutions of 1a-c and allyltrimethylsilane (2) containing chiral carboxylic acids. The yields and percent enantiomeric excesses (%ee) of products formed in these processes are listed in Table 2. The %ee in each case was calculated using the ratio of areas under the chiral HPLC or GC peaks corresponding to the enantiomers as %ee when acetic acid was used becoming zero. A positive %ee value corresponds to a situation in which the major isomer is the second peak, while a negative value shows that the major isomer is the first peak. The absolute structures could not be decided. The data arising from photoreactions in the absence or presence of achiral acetic acid are also included in Table 2 for comparison purposes.
Use of 1 equiv of (R)-mandelic acid in photoreaction of 1a with 2 led to formation of 3a and 4a with respective +1.5 and +4.1 %ee values (entry 3). A reversal in major enantiomers of the products arose from the reaction of 1a with 2 conducted in the presence of (S)-mandelic acid (entry 4), which supports the reaction proceeding in an enantioselective manner. Also, when L-lactic acid was used in this photoreaction, the major enantiomers were the reverse of those formed in reactions in the To assess the potential of introducing enantioselectivity into the photoreactions described above, irradiations were carried out on solutions of 1a-c and allyltrimethylsilane (2) containing chiral carboxylic acids. The yields and percent enantiomeric excesses (%ee) of products formed in these processes are listed in Table 2. The %ee in each case was calculated using the ratio of areas under the chiral HPLC or GC peaks corresponding to the enantiomers as %ee when acetic acid was used becoming zero. A positive %ee value corresponds to a situation in which the major isomer is the second peak, while a negative value shows that the major isomer is the first peak. The absolute structures could not be decided. The data arising from photoreactions in the absence or presence of achiral acetic acid are also included in Table 2 for comparison purposes.
Use of 1 equiv of (R)-mandelic acid in photoreaction of 1a with 2 led to formation of 3a and 4a with respective +1.5 and +4.1 %ee values (entry 3). A reversal in major enantiomers of the products arose from the reaction of 1a with 2 conducted in the presence of (S)-mandelic acid (entry 4), which supports the reaction proceeding in an enantioselective manner. Also, when l-lactic acid was used in this photoreaction, the major enantiomers were the reverse of those formed in reactions in the presence of (S)-mandelic acid (entry 5). The use of C 2 symmetric dibenzoyl l-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6-methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15). presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).
a Conditions: 1a-c (0.14 mmol), 2 (0.42 mmol), phenanthrene (0.07 mmol), CH3CN (8 mL), additive (0.14 mmol), 300 W high-pressure mercury lamp, Pyrex, r.t. b Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart. f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).
a Conditions: 1a-c (0.14 mmol), 2 (0.42 mmol), phenanthrene (0.07 mmol), CH3CN (8 mL), additive (0.14 mmol), 300 W high-pressure mercury lamp, Pyrex, r.t. b Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart. f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).
a Conditions: 1a-c (0.14 mmol), 2 (0.42 mmol), phenanthrene (0.07 mmol), CH3CN (8 mL), additive (0.14 mmol), 300 W high-pressure mercury lamp, Pyrex, r.t. b Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart. f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).  , additive (0.14 mmol), 300 W high-pressure mercury lamp, Pyrex, r.t. b Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart. f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).  , additive (0.14 mmol), 300 W high-pressure mercury lamp, Pyrex, r.t. b Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart. f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15). Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15). Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15). Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are (S)-mandelic acid 24  presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).  Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart.
f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).  Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart.
f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15). Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or presence of (S)-mandelic acid (entry 5). The use of C2 symmetric dibenzoyl L-tartaric acid did not promote an increase of %ee of either product (entry 6). The photoreaction of 1b with 2 also occurred when (R)-and (S)-mandelic acids were used, however the %ee of either product could not be determined (entries 9 and 10). Like in the case of 1a, photoreaction of 1c produced products 3c and 4c in which the major enantiomers were reversed when (R)-and (S)-mandelic acids were utilized (entries 13 and 14). Moreover, the results showed that the %ee improved up to 3.5 when (S)-2-(6methoxy-2-naphthyl)propionic acid was used as a chiral acid (entry 15).  f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are (S)-2-(6-methoxy-2naphthyl)propionic acid 2 0 g 68 g (−3.5 i,k ) a Conditions: 1a-c (0.14 mmol), 2 (0.42 mmol), phenanthrene (0.07 mmol), CH 3 CN (8 mL), additive (0.14 mmol), 300 W high-pressure mercury lamp, Pyrex, r.t. b Determined by using GC. c 1 mL. d Determined by using chiral HPLC. e Major isomer corresponds to the second peak in the HPLC chart. f Major isomer corresponds to the first peak in the HPLC chart. g Determined by using 1 H-NMR. h Ee could not be determined. i Determined by using chiral GC. j Major isomer corresponds to the second peak in the GC chart. k Major isomer corresponds to the first peak in the GC chart.
Each photoreaction described above takes place through a process termed photoredox sensitization by phenanthrene (Phen) (Scheme 4) [22][23][24][25][26][27]. In the pathway, the excited singlet state of Phen, generated by light absorption, transfers one electron (SET (single electron transfer)) to the electron-deficient alkene 1 to form the phenanthrene radical cation (Phen •+ ) and the alkene radical anion 1 •− . The subsequent SET from allyltrimethylsilane (2) to Phen •+ generates recovered Phen and the radical cation 2 •+ , which undergoes nucleophile-assisted Si-C bond cleavage [28][29][30] to form the allyl radical. Also, radical anion 1 •− is protonated by the carboxylic acid to produce radical 5, which upon coupling with the allyl radical generates the allylation product 3. In a competitive pathway, radical 5 undergoes hydrogen abstraction or one-electron reduction followed by protonation or disproportionation to form reduction product 4 [23]. The inefficiency of the photoreaction of the MeO-substituted substrate 1b and high efficiency of the reaction of Cl-substituted reactant 1c are likely consequences of the stabilities of the corresponding radical anions 1b •and 1c •which governs their rates of formation by SET from relative to unproductive decay of the excited singlet state of Phen.
Based on the results of molecular orbital calculations with related compounds, it is estimated that the radical contribution to radical anion 1 •is large at the dicyano substituted carbon (α) and that negative charge density is large at the dialkyl substituted carbon (β) [23,24,26,27]. In accord with this conclusion, the photoreaction of 1a with 2 using CH 3 COOD as the additive produced mainly mono-deuteriated forms of 3a and 4a in which deuterium is present at the stereogenic carbons marked with * in Scheme 4. Therefore, enantioselectivity is governed at the step where protonation of the radical anion takes place.
Molecules 2019, 24, x 6 of 11 their rates of formation by SET from relative to unproductive decay of the excited singlet state of Phen.
Based on the results of molecular orbital calculations with related compounds, it is estimated that the radical contribution to radical anion 1 •-is large at the dicyano substituted carbon (α) and that negative charge density is large at the dialkyl substituted carbon (β) [23,24,26,27]. In accord with this conclusion, the photoreaction of 1a with 2 using CH3COOD as the additive produced mainly monodeuteriated forms of 3a and 4a in which deuterium is present at the stereogenic carbons marked with * in Scheme 4. Therefore, enantioselectivity is governed at the step where protonation of the radical anion takes place. The stereochemistry of protonation of the radical anion 1 •-can be discussed using a Felkin-Anh model (Scheme 5) [31][32][33]. Specifically, in reaction of 1a in the presence of (S)-mandelic acid, proton transfer to the Re face of 1a •-should be preferred in a complex in which a π-π stabilizing interaction occurs between the phenyl groups and the OH group of the acid is located in a sterically less hindered position. Proton transfer to the Re face of 1a leads to the eventual formation of (S)-3a and (S)-4a. On the other hand, in the reaction of 1a in the presence of L-lactic acid, an OH-π interaction between the OH group of the acid and the phenyl group of 1a •-takes place to form a complex in which proton transfer from the carboxylic acid group occurs preferentially to the Si face to minimize steric repulsion of methyl group. This process then gives rise to formation of (R)-3a and (R)-4a. In photoreaction of 1c in the presence of (S)-2-(6-methoxy-2-naphthyl)propionic acid, the main enantiomers produced were the same as those generated in reaction of 1c in the presence of (S)-mandelic acid, and %ee increased. This outcome might be a consequence of a strong π-π interaction between the chlorophenyl and the methoxynaphthyl groups. The stereochemistry of protonation of the radical anion 1 •can be discussed using a Felkin-Anh model (Scheme 5) [31][32][33]. Specifically, in reaction of 1a in the presence of (S)-mandelic acid, proton transfer to the Re face of 1a •should be preferred in a complex in which a π-π stabilizing interaction occurs between the phenyl groups and the OH group of the acid is located in a sterically less hindered position. Proton transfer to the Re face of 1a leads to the eventual formation of (S)-3a and (S)-4a. On the other hand, in the reaction of 1a in the presence of l-lactic acid, an OH-π interaction between the OH group of the acid and the phenyl group of 1a •takes place to form a complex in which proton transfer from the carboxylic acid group occurs preferentially to the Si face to minimize steric repulsion of methyl group. This process then gives rise to formation of (R)-3a and (R)-4a. In photoreaction of 1c in the presence of (S)-2-(6-methoxy-2-naphthyl)propionic acid, the main enantiomers produced were the same as those generated in reaction of 1c in the presence of (S)-mandelic acid, and %ee increased. This outcome might be a consequence of a strong π-π interaction between the chlorophenyl and the methoxynaphthyl groups.

Preparation of 1b
A THF (50 mL) solution of 4-bromoanisole (17.53 mL, 140.0 mmol) was added dropwise to stirred Mg turnings (3.889 g, 160.0 mmol). A small amount of I 2 was added to facilitate the reaction. A THF (20 mL) solution of ethyl l-lactate (4.587 mL, 40.0 mmol) was added dropwise to the solution, and the resulting mixture was stirred at reflux, cooled, and extracted with Et 2 O and NH 4 Cl aq [36]. The organic layer was dried over Na 2 SO 4 , filtered, and concentrated in vacuo, giving a residue that was subjected to silica gel column chromatography to give 1,1-bis(4-methoxyphenyl)propane-1,2-diol (5.76 g, 20.0 mmol, 50% yield, including inpurity).

Preparation of 1c
A THF (12 mL) solution of 4-bromochlorobenzene (6.647 g, 34.7 mmol) was added dropwise to stirred Mg turnings (0.729 g, 30.0 mmol). A small amount of I 2 was added to facilitate the reaction. A THF (5 mL) solution of ethyl l-lactate (1.247 mL, 10.9 mmol) was added dropwise to the solution, and the resulting solution was stirred at reflux, cooled, and extracted with Et 2 O and NH 4 Cl aq [36]. The organic layer was dried over Na 2 SO 4 , filtered, and concentrated in vacuo, giving a residue that was subjected to silica gel column chromatography to give 1,1-bis(4-chlorophenyl)propane-1,2-diol (4.023 g, including inpurity).

Conclusions
In summary, we found that photoreactions of prochiral 3,3-diaryl-1,1-dicyano-2-methylprop-1-enes 1a-c with allyltrimethylsilane, carried in the presence of enantiomerically pure chiral carboxylic acids, generates photoallylation and photoreduction products with low but finite levels of enantioselectivity. The percent enantiomeric excesses in the products of the process was highest (4.8 %ee) when (S)-mandelic acid was used. Enantioselectivities in these reactions are a consequence of sterically governed asymmetric proton transfer in intermediate complexes formed by π-π and OH-π interactions between radical anions of the prochiral alkenes and the chiral carboxylic acids.