Surface-doped graphitic carbon nitride catalyzed photooxidation of olefins and dienes Chemical evidence for electron transfer and singlet oxygen mechanisms

: A new photocatalytic reactivity of carbon-nanodot-doped graphitic carbon nitride (CD-C 3 N 4 ) with alkenes and dienes, has been disclosed. We have shown that CD-C 3 N 4 photosensitizes the oxidation of unsaturated substrates in a variety of solvents according to two competing mechanisms: the energy transfer via singlet oxygen ( 1 O 2 ) and / or the electron transfer via superoxide (O · − 2 ). The singlet oxygen, derived by the CD-C 3 N 4 photosensitized process, reacts with alkenes to form allylic hydroperoxides (ene products) whereas with dienes, endoperoxides. When the electron transfer mechanism operates, cleavage products are formed, derived from the corresponding dioxetanes. Which of the two mechanisms will prevail depends on solvent polarity and the particular substrate. The photocatalyst remains stable under the photooxidation conditions, unlike the most conventional photosensitizers, while the heterogeneous nature of CD-C 3 N 4 overcomes usual solubility problems. utilizes carbon-nanodot-doped g-C 3 N 4 (CD-C 3 N 4 ) as a visible light photocatalyst for the direct photooxidation of alkenes via the intermediacy of singlet oxygen to allylic hydroperoxides as well as the {4 + 2} cycloaddition of singlet oxygen to hexacyclodiene. We also demonstrate chemically that in CD-C 3 N 4 photosensitized reactions, electron transfer and singlet oxygen mechanisms are in competition, depending mostly on the nature of the substrate and the polarity of the solvent. In addition, the CD-C 3 N 4 mediated photooxygenation results were compared with those derived using conventional singlet oxygen photosensitizers.


Introduction
Graphitic carbon nitride (g-C 3 N 4 ) is recently one of the most studied heterogeneous, metal-free photocatalysts. g-C 3 N 4 is a polymeric material, composed of highly abundant elements such as carbon and nitrogen, thermally stable and nontoxic. Ever since the synthesis of this material in bulk quantities [1][2][3], a large number of reports concerning the photochemical properties and photocatalytic activity were published. Next, the demand to optimize the photocatalytic properties of g-C 3 N 4 was inevitable. The relatively low band gap energy Eg of 2.7 eV and high conduction and valence bond positions of the photocatalyst [4] prompted many research groups to improve the photocatalytic activity of g-C 3 N 4 . This was achieved by doping the surface of the catalyst with a variety of metallic elements [3,[5][6][7][8][9][10][11], oxides [12,13], sulfides [14] even graphene [15] and carbon nanotubes [16].
It is important to note here that structure-controlled g-C 3 N 4 is a highly efficient photocatalyst for water splitting to produce hydrogen, in an internal quantum yield of 26.5% under visible light [17]. In a more recent article [18], the metal-free carbon-nanodot-doped graphitic carbon nitride (CD-C 3 N 4 ) impressively photocatalyzed the water splitting via a two-electron mechanism. Although many studies concern the photochemical activity and spectroscopic data analysis of g-C 3 N 4 as well as modification of its surface [19], applications to organic transformations are limited. For example, graphitic C 3 N 4 as well as its surface-doped-materials, catalyze a few reactions; a) photoacetalization of aldehydes/ketones [20] b) the Friedel-Crafts reaction of benzene [21] c) oxidation of alcohols using transition metal doped

Results and Discussion
In the present work, the photocatalyst g-C 3 N 4 was prepared according to the procedure described by Tang and coworkers [45], while the carbon nanodots via thermal decomposition of sucrose [46]. The so produced nanodots were deposited by heating on the g-C 3 N 4 surface to provide the metal-free carbon-nanodot-doped g-C 3 N 4 (CD-C 3 N 4 ). Details on the experimental conditions as well as the spectroscopic characterization are described in our previous work [30].
Photosensitized oxidations were performed in a 4 mL pyrex cell containing 0.09 M solution of 2,3-dimethylbut-2-ene (1) in a variety of solvents with 4 mg/mL of the insoluble CD-C 3 N 4 photocatalyst. While bubbling oxygen at 20 • C, the solution was irradiated for a period of 30 min. A 300 W xenon lamp was utilized as the visible light source. The photooxidation reaction was followed by GC and/or 1 H NMR analysis after filtration from celite followed by reduction of the allylic hydroperoxides with triphenylphosphine to the corresponding allylic alcohols.
We report here the unprecedented photooxygenation of alkenes utilizing CD-C 3 N 4 as the photosensitizer. The electron rich tetramethylethylene (1) was chosen as the singlet oxygen acceptor, Scheme 1. Within a few minutes of irradiation, the allylic hydroperoxide 1a (see GC chromatograms S16, S17 and S18) was formed followed by reduction with Ph 3 P to the corresponding allylic alcohol 1b (GC chromatogram S19).
Catalysts 2019, 9, x FOR PEER REVIEW 2 of 12 of aldehydes/ketones [20] b) the Friedel-Crafts reaction of benzene [21] c) oxidation of alcohols using transition metal doped g-C3N4 [22][23][24][25] d) oxidation of sulfides to sulfoxides using oxidized g-C3N4 (CNO). In addition, selective oxidations of benzylic C-H bonds utilizing mesoporous g-C3N4 with N-OH co catalysts have been reported [26][27][28][29]. Herein, following our recent work in this field [30], we report a new approach that utilizes carbon-nanodot-doped g-C3N4 (CD-C3N4) as a visible light photocatalyst for the direct photooxidation of alkenes via the intermediacy of singlet oxygen to allylic hydroperoxides as well as the {4 + 2} cycloaddition of singlet oxygen to hexacyclodiene. We also demonstrate chemically that in CD-C3N4 photosensitized reactions, electron transfer and singlet oxygen mechanisms are in competition, depending mostly on the nature of the substrate and the polarity of the solvent. In addition, the CD-C3N4 mediated photooxygenation results were compared with those derived using conventional singlet oxygen photosensitizers.
Singlet molecular oxygen ( 1 O2, 1 Δg) plays an important role in chemical [31], biological [32,33] and therapeutic processes [34,35], as well as in the degradation of food and materials [36][37][38]. Historically, although 1 O2 was discovered more than 80 years ago [39,40], and until the early 1960s was considered to be a molecular species of passing astrophysical interest and rather limited as a research subject. The pioneering work by Foote and Wexler [41,42], provided strong evidence for the formation of 1 O2 as the reactive intermediate in photosensitized processes in solution. Today, the use of 1 O2 as a reagent in organic synthesis has received remarkable attention [43,44].

Results and Discussion
In the present work, the photocatalyst g-C3N4 was prepared according to the procedure described by Tang and coworkers [45], while the carbon nanodots via thermal decomposition of sucrose [46]. The so produced nanodots were deposited by heating on the g-C3N4 surface to provide the metal-free carbon-nanodot-doped g-C3N4 (CD-C3N4). Details on the experimental conditions as well as the spectroscopic characterization are described in our previous work [30].
Photosensitized oxidations were performed in a 4 mL pyrex cell containing 0.09 M solution of 2,3-dimethylbut-2-ene (1) in a variety of solvents with 4 mg/mL of the insoluble CD-C3N4 photocatalyst. While bubbling oxygen at 20 °C, the solution was irradiated for a period of 30 min. A 300 W xenon lamp was utilized as the visible light source. The photooxidation reaction was followed by GC and/or 1 H NMR analysis after filtration from celite followed by reduction of the allylic hydroperoxides with triphenylphosphine to the corresponding allylic alcohols.
We report here the unprecedented photooxygenation of alkenes utilizing CD-C3N4 as the photosensitizer. The electron rich tetramethylethylene (1) was chosen as the singlet oxygen acceptor, Scheme 1. Within a few minutes of irradiation, the allylic hydroperoxide 1a (see GC chromatograms S16,S17 and S18) was formed followed by reduction with Ph3P to the corresponding allylic alcohol 1b (GC chromatogram S19). The GC and/or 1 H NMR analysis of product 1b, as can be seen in S2, showed identical results with those when typical singlet oxygen photosensitizers, such as rose Bengal (RB) or tetraphenylporphyrine (TPP), were utilized. The production of allylic hydroperoxides by the sensitized photooxygenation of alkenes, the so called "ene or Schenck" reaction, was studied several years ago, for its synthetic [43,44], mechanistic [47,48] and theoretical [49,50] point of view. Furthermore, addition of a small amount of 1,4-diazabicyclo[2.2.2]octane (DABCO), a well-known 1 O2 quencher [51,52], into the reaction mixture, Scheme 1, retarded completely the production of allylic Scheme 1. CD-C 3 N 4 photosensitized oxidation of tetramethylethylene (1). The GC and/or 1 H NMR analysis of product 1b, as can be seen in S2, showed identical results with those when typical singlet oxygen photosensitizers, such as rose Bengal (RB) or tetraphenyl-porphyrine (TPP), were utilized. The production of allylic hydroperoxides by the sensitized photooxygenation of alkenes, the so called "ene or Schenck" reaction, was studied several years ago, for its synthetic [43,44], mechanistic [47,48] and theoretical [49,50] [51,52], into the reaction mixture, Scheme 1, retarded completely the production of allylic hydroperoxides. This is an additional indication of the 1 O 2 involvement as the reactive intermediate in the above-mentioned reaction.
The kinetics of the CD-C 3 N 4 photosensitized oxygenation of 1, Scheme 1, was examined in a variety of solvents. Reaction progress at several conversions was easily monitored by GC analysis. Diglyme was used as the internal standard. First, assuming that during the photooxygenation, the oxygen concentration remains constant while there is a first-order dependence of the reaction rate on the tetramethylethylene (1) concentration, the following equation can be applied: kt = −ln (1 − x) where k, t, and x define the rate constant, the irradiation time, and the conversion of tetramethylethylene, respectively. However, an adequate linear correlation was not found. Therefore, we simply plot the conversion of tetramethylethylene (1) vs the illumination time, as shown in Figure 1. The plot ( Figure 1) shows that the rate of product formation depends on solvent polarity, e.g., in EtOAc, is favored compared to CCl 4 by roughly a factor of two. hydroperoxides. This is an additional indication of the 1 O2 involvement as the reactive intermediate in the above-mentioned reaction. The kinetics of the CD-C3N4 photosensitized oxygenation of 1, Scheme 1, was examined in a variety of solvents. Reaction progress at several conversions was easily monitored by GC analysis. Diglyme was used as the internal standard. First, assuming that during the photooxygenation, the oxygen concentration remains constant while there is a first-order dependence of the reaction rate on the tetramethylethylene (1) concentration, the following equation can be applied: kt = −ln (1 − x) where k, t, and x define the rate constant, the irradiation time, and the conversion of tetramethylethylene, respectively. However, an adequate linear correlation was not found. Therefore, we simply plot the conversion of tetramethylethylene (1) vs the illumination time, as shown in Figure 1. The plot ( Figure  1) shows that the rate of product formation depends on solvent polarity, e.g., in EtOAc, is favored compared to CCl4 by roughly a factor of two. Next, we examined the regioselectivity of the present oxidation system with an alkyl trisubstituted alkene, Scheme 2. For this purpose, the regiospecifically deuterium labeled alkene 2 was synthesized (see S3, S4, S5 and S6) and was used as an appropriate substrate for sensitized photooxygenation using two different catalysts, one conventional photosensitizer (TPP) and the photocatalyst CD-C3N4. It is well established that double bond formation via singlet oxygen ene reaction will occur preferentially in the most substituted side of the alkyl substituted double bond [53]. Indeed, as seen in Table 1 (entries 1 and 2), both conventional (TPP) and CD-C3N4 photocatalysts upon visible light irradiation show, within experimental error, identical regioselectivity. For example, 2a and 2b which can be seen in S7 and S8, are the main products (cis effect) [53], while the E-methyl group is highly unreactive. This result is again a strong evidence of 1 O2 involvement, excluding the involvement of other reactive oxygen species (ROS).

Entry
Photocatalyst Regioselectivities (%) b Next, we examined the regioselectivity of the present oxidation system with an alkyl trisubstituted alkene, Scheme 2. For this purpose, the regiospecifically deuterium labeled alkene 2 was synthesized (see S3, S4, S5 and S6) and was used as an appropriate substrate for sensitized photooxygenation using two different catalysts, one conventional photosensitizer (TPP) and the photocatalyst CD-C 3 N 4 . It is well established that double bond formation via singlet oxygen ene reaction will occur preferentially in the most substituted side of the alkyl substituted double bond [53]. Indeed, as seen in Table 1 (entries 1 and 2), both conventional (TPP) and CD-C 3 N 4 photocatalysts upon visible light irradiation show, within experimental error, identical regioselectivity. For example, 2a and 2b which can be seen in S7 and S8, are the main products (cis effect) [53], while the E-methyl group is highly unreactive. This result is again a strong evidence of 1 O 2 involvement, excluding the involvement of other reactive oxygen species (ROS). hydroperoxides. This is an additional indication of the 1 O2 involvement as the reactive intermediate in the above-mentioned reaction. The kinetics of the CD-C3N4 photosensitized oxygenation of 1, Scheme 1, was examined in a variety of solvents. Reaction progress at several conversions was easily monitored by GC analysis. Diglyme was used as the internal standard. First, assuming that during the photooxygenation, the oxygen concentration remains constant while there is a first-order dependence of the reaction rate on the tetramethylethylene (1) concentration, the following equation can be applied: kt = −ln (1 − x) where k, t, and x define the rate constant, the irradiation time, and the conversion of tetramethylethylene, respectively. However, an adequate linear correlation was not found. Therefore, we simply plot the conversion of tetramethylethylene (1) vs the illumination time, as shown in Figure 1. The plot ( Figure  1) shows that the rate of product formation depends on solvent polarity, e.g., in EtOAc, is favored compared to CCl4 by roughly a factor of two. Next, we examined the regioselectivity of the present oxidation system with an alkyl trisubstituted alkene, Scheme 2. For this purpose, the regiospecifically deuterium labeled alkene 2 was synthesized (see S3, S4, S5 and S6) and was used as an appropriate substrate for sensitized photooxygenation using two different catalysts, one conventional photosensitizer (TPP) and the photocatalyst CD-C3N4. It is well established that double bond formation via singlet oxygen ene reaction will occur preferentially in the most substituted side of the alkyl substituted double bond [53]. Indeed, as seen in Table 1 (entries 1 and 2), both conventional (TPP) and CD-C3N4 photocatalysts upon visible light irradiation show, within experimental error, identical regioselectivity. For example, 2a and 2b which can be seen in S7 and S8, are the main products (cis effect) [53], while the E-methyl group is highly unreactive. This result is again a strong evidence of 1 O2 involvement, excluding the involvement of other reactive oxygen species (ROS).

Entry
Photocatalyst The second mode of 1 O 2 reactivity is the {4 + 2} cycloaddition to conjugated dienes to yield endoperoxides. We demonstrate here, Scheme 3, that CD-C 3 N 4 photosensitizes the addition of singlet oxygen to 1,3-cyclohexadiene (3) to produce efficiently endoperoxide 3a (Figure 2, S9) in ambient conditions. The second mode of 1 O2 reactivity is the {4 + 2} cycloaddition to conjugated dienes to yield endoperoxides. We demonstrate here, Scheme 3, that CD-C3N4 photosensitizes the addition of singlet oxygen to 1,3-cyclohexadiene (3)   Furthermore, apart from oxygen photoactivation, the CD-C3N4 catalyst was tested for promoting an electron transfer process. The proposed two mechanisms for the sensitized photooxygenation of organic substrates were initially classified by Gollnick [54] as Type I and Type II as shown below in Scheme 4.  The two mechanisms, Type I and Type II as well as singlet oxygen ( 1 O2) and superoxide (O·2) formation, are always in competition. Which of them will prevail depends on the nature of the  The second mode of 1 O2 reactivity is the {4 + 2} cycloaddition to conjugated dienes to yield endoperoxides. We demonstrate here, Scheme 3, that CD-C3N4 photosensitizes the addition of singlet oxygen to 1,3-cyclohexadiene (3)   Furthermore, apart from oxygen photoactivation, the CD-C3N4 catalyst was tested for promoting an electron transfer process. The proposed two mechanisms for the sensitized photooxygenation of organic substrates were initially classified by Gollnick [54] as Type I and Type II as shown below in Scheme 4.  The two mechanisms, Type I and Type II as well as singlet oxygen ( 1 O2) and superoxide (O·2) formation, are always in competition. Which of them will prevail depends on the nature of the Furthermore, apart from oxygen photoactivation, the CD-C 3 N 4 catalyst was tested for promoting an electron transfer process. The proposed two mechanisms for the sensitized photooxygenation of organic substrates were initially classified by Gollnick [54] as Type I and Type II as shown below in Scheme 4. The second mode of 1 O2 reactivity is the {4 + 2} cycloaddition to conjugated dienes to yield endoperoxides. We demonstrate here, Scheme 3, that CD-C3N4 photosensitizes the addition of singlet oxygen to 1,3-cyclohexadiene (3)   Furthermore, apart from oxygen photoactivation, the CD-C3N4 catalyst was tested for promoting an electron transfer process. The proposed two mechanisms for the sensitized photooxygenation of organic substrates were initially classified by Gollnick [54] as Type I and Type II as shown below in Scheme 4.  The two mechanisms, Type I and Type II as well as singlet oxygen ( 1 O2) and superoxide (O·2) formation, are always in competition. Which of them will prevail depends on the nature of the  The two mechanisms, Type I and Type II as well as singlet oxygen ( 1 O 2 ) and superoxide (O· − 2 ) formation, are always in competition. Which of them will prevail depends on the nature of the photocatalyst, the solvent, the substrate's nature and the concentration. In Type II mechanism singlet oxygen is produced by energy transfer whereas superoxide anion (O· − 2 ) by electron transfer mechanism. To check the electron transfer ability of CD-C 3 N 4 catalyst, the electron rich 2-(4-methoxyphenyl)-3-methylbut-2-ene (4) was prepared (see S11). This substrate apart from 1 O 2 acceptor is also a good electron donor. The results from the irradiation of aryl alkene 4 in a variety of solvents and sensitizers, Scheme 5, are shown in Table 2.
Catalysts 2019, 9, x FOR PEER REVIEW 5 of 12 photocatalyst, the solvent, the substrate's nature and the concentration. In Type II mechanism singlet oxygen is produced by energy transfer whereas superoxide anion (O·2) by electron transfer mechanism. To check the electron transfer ability of CD-C3N4 catalyst, the electron rich 2-(4methoxyphenyl)-3-methylbut-2-ene (4) was prepared (see S11). This substrate apart from 1 O2 acceptor is also a good electron donor. The results from the irradiation of aryl alkene 4 in a variety of solvents and sensitizers, Scheme 5, are shown in Table 2.  For comparison reasons, apart from CD-C3N4, four sensitizers were utilized in the above reaction (entries 1-4). Rose Bengal (entry 2, RB), TPP (entry 1) and methylene blue (entry 3, MB) are well studied singlet oxygen photosensitizers. However, 9,10-dicyano-anthracene (entry 4, DCA), an electron poor compound, sensitizes the photooxygenation by both singlet oxygen and electron transfer mechanisms depending on the particular substrate and the polarity of the solvent [55].
Irradiation of 4 in the presence of TPP, RB or MB as photosensitizers afforded the allylic hydroperoxides 4a and 4b in a ratio of approximately 7:3. These hydroperoxides are typical singlet oxygen products whose ratio 7:3 is independent of solvent polarity [47,56] or the para-substitution of the phenyl ring. However, the DCA-photosensitized oxygenation of 4 in acetonitrile (entry 4), afforded apart from the ene products 4a and 4b considerable amount of p-methoxy-acetophenone (4c). In this case, the results are supported by both electron transfer and singlet oxygen mechanisms. Similarly, in the CD-C3N4 catalyzed photooxidation of 4 in non protic solvents, apart from the allylic hydroperoxides 4a and 4b, considerable amounts of 4c in the range of 15-18% were obtained. These results are also similar with those derived earlier from DCA photosensitized oxygenation of diphenyl-ethylene [55]. p-Methoxy-acetophenone 4c is most probably produced from the cleavage of dioxetane 4d via the intermediacy of superoxide anion (O·2) through an electron transfer mechanism, as shown in Scheme 6. The formation of superoxide anion O·2and not a hydroxyl radical has been also recently documented [20]. We must note here that control experiments either with the CD-C3N4 in dark (entry 10) or without the photocatalyst under irradiation conditions (entry 11) did not show any detectable oxidation products (entry 10) or small traces observed by 1 H NMR and GC analysis  For comparison reasons, apart from CD-C 3 N 4 , four sensitizers were utilized in the above reaction (entries 1-4). Rose Bengal (entry 2, RB), TPP (entry 1) and methylene blue (entry 3, MB) are well studied singlet oxygen photosensitizers. However, 9,10-dicyano-anthracene (entry 4, DCA), an electron poor compound, sensitizes the photooxygenation by both singlet oxygen and electron transfer mechanisms depending on the particular substrate and the polarity of the solvent [55].
Irradiation of 4 in the presence of TPP, RB or MB as photosensitizers afforded the allylic hydroperoxides 4a and 4b in a ratio of approximately 7:3. These hydroperoxides are typical singlet oxygen products whose ratio 7:3 is independent of solvent polarity [47,56] or the para-substitution of the phenyl ring. However, the DCA-photosensitized oxygenation of 4 in acetonitrile (entry 4), afforded apart from the ene products 4a and 4b considerable amount of p-methoxy-acetophenone (4c). In this case, the results are supported by both electron transfer and singlet oxygen mechanisms. Similarly, in the CD-C 3 N 4 catalyzed photooxidation of 4 in non protic solvents, apart from the allylic hydroperoxides 4a and 4b, considerable amounts of 4c in the range of 15-18% were obtained. These results are also similar with those derived earlier from DCA photosensitized oxygenation of diphenyl-ethylene [55]. p-Methoxy-acetophenone 4c is most probably produced from the cleavage of dioxetane 4d via the intermediacy of superoxide anion (O· − 2 ) through an electron transfer mechanism, as shown in Scheme 6. The formation of superoxide anion O· − 2 and not a hydroxyl radical has been also recently documented [20]. We must note here that control experiments either with the CD-C 3 N 4 in dark (entry 10) or without the photocatalyst under irradiation conditions (entry 11) did not show any detectable oxidation products (entry 10) or small traces observed by 1 H NMR and GC analysis (entry 11). It is interesting to note here that when this reaction was run in EtOH as the solvent (entry 9, Table 2), the cleavage product 4c was exclusively obtained in 100% conversion. This result indicates that under these conditions, electron transfer is the only operating mechanism (Scheme 6). However, it is not clear at the present why the irradiation of CD-C 3 N 4 surface in EtOH, promotes the electron transfer mechanism leading exclusively to cleavage products 4c, (entry 9, Table 2). Most probably, among some other unknown at the time reasons, the polar protic solvent ethanol stabilizes favorably the pair of radical ions (anion + cation) favoring dramatically the electron transfer path producing exclusively the superoxide anion (O· − 2 ) as the reactive intermediate.
Catalysts 2019, 9, x FOR PEER REVIEW 6 of 12 (entry 11). It is interesting to note here that when this reaction was run in EtOH as the solvent (entry 9, Table 2), the cleavage product 4c was exclusively obtained in 100% conversion. This result indicates that under these conditions, electron transfer is the only operating mechanism (Scheme 6). However, it is not clear at the present why the irradiation of CD-C3N4 surface in EtOH, promotes the electron transfer mechanism leading exclusively to cleavage products 4c, (entry 9, Table 2). Most probably, among some other unknown at the time reasons, the polar protic solvent ethanol stabilizes favorably the pair of radical ions (anion + cation) favoring dramatically the electron transfer path producing exclusively the superoxide anion (O·2) as the reactive intermediate. To assess further the extent of electron transfer efficiency of CD-C3N4 surface in a polar and non protic solvent, the photooxidation of 1,1-di(p-anisyl)ethylene (5) was performed (Scheme 7). An electron transfer test can be probed by utilizing substrate 5; an electron rich alkene, not reactive with singlet oxygen. Scheme 7 shows the CD-C3N4 sensitized photooxygenation of 5 in acetonitrile. This reaction afforded exclusively and quantitatively the cyclic peroxide 5a (Figure 3, S15), by cycloaddition of photochemically produced two radical cations of 5 and one super oxide anion. Similar results under the same conditions have been reported previously [57,58], when DCA was the photosensitizer. Furthermore, upon addition of a small quantity of 1,3,5-trimethoxybenzene, the above reaction, Scheme 7, was retarded. This result supports an electron transfer mechanism considering that the lower oxidation potential of a donor molecule (1,3,5-trimethoxybenzene) than the corresponding potential of the competing electron donor alkene, [55], quenches the electron transfer pathway from the alkene to the sensitizer. Similar mechanism was published earlier by Ericksen and Foote [55], when in their case, electron deficient DCA was used as the photocatalyst and 1,1-diphenylethylene or tetramethylethylene were used as the substrates. To assess further the extent of electron transfer efficiency of CD-C 3 N 4 surface in a polar and non protic solvent, the photooxidation of 1,1-di(p-anisyl)ethylene (5) was performed (Scheme 7). An electron transfer test can be probed by utilizing substrate 5; an electron rich alkene, not reactive with singlet oxygen. Scheme 7 shows the CD-C 3 N 4 sensitized photooxygenation of 5 in acetonitrile.
Catalysts 2019, 9, x FOR PEER REVIEW 6 of 12 (entry 11). It is interesting to note here that when this reaction was run in EtOH as the solvent (entry 9, Table 2), the cleavage product 4c was exclusively obtained in 100% conversion. This result indicates that under these conditions, electron transfer is the only operating mechanism (Scheme 6). However, it is not clear at the present why the irradiation of CD-C3N4 surface in EtOH, promotes the electron transfer mechanism leading exclusively to cleavage products 4c, (entry 9, Table 2). Most probably, among some other unknown at the time reasons, the polar protic solvent ethanol stabilizes favorably the pair of radical ions (anion + cation) favoring dramatically the electron transfer path producing exclusively the superoxide anion (O·2) as the reactive intermediate. To assess further the extent of electron transfer efficiency of CD-C3N4 surface in a polar and non protic solvent, the photooxidation of 1,1-di(p-anisyl)ethylene (5) was performed (Scheme 7). An electron transfer test can be probed by utilizing substrate 5; an electron rich alkene, not reactive with singlet oxygen. Scheme 7 shows the CD-C3N4 sensitized photooxygenation of 5 in acetonitrile. This reaction afforded exclusively and quantitatively the cyclic peroxide 5a (Figure 3, S15), by cycloaddition of photochemically produced two radical cations of 5 and one super oxide anion. Similar results under the same conditions have been reported previously [57,58], when DCA was the photosensitizer. Furthermore, upon addition of a small quantity of 1,3,5-trimethoxybenzene, the above reaction, Scheme 7, was retarded. This result supports an electron transfer mechanism considering that the lower oxidation potential of a donor molecule (1,3,5-trimethoxybenzene) than the corresponding potential of the competing electron donor alkene, [55], quenches the electron transfer pathway from the alkene to the sensitizer. Similar mechanism was published earlier by Ericksen and Foote [55], when in their case, electron deficient DCA was used as the photocatalyst and 1,1-diphenylethylene or tetramethylethylene were used as the substrates. This reaction afforded exclusively and quantitatively the cyclic peroxide 5a (Figure 3, S15), by cycloaddition of photochemically produced two radical cations of 5 and one super oxide anion. Similar results under the same conditions have been reported previously [57,58], when DCA was the photosensitizer. Furthermore, upon addition of a small quantity of 1,3,5-trimethoxybenzene, the above reaction, Scheme 7, was retarded. This result supports an electron transfer mechanism considering that the lower oxidation potential of a donor molecule (1,3,5-trimethoxybenzene) than the corresponding potential of the competing electron donor alkene, [55], quenches the electron transfer pathway from the alkene to the sensitizer. Similar mechanism was published earlier by Ericksen and Foote [55], when in their case, electron deficient DCA was used as the photocatalyst and 1,1-diphenylethylene or tetramethylethylene were used as the substrates.
Column chromatographic separations were carried out by a flash chromatography system using silica gel and hexane/ethyl acetate or petroleum ether/ethyl acetate solvent mixtures. For thin layer chromatography (TLC), Merck silica gel (grade 60 F245, Merck & Co., Kenilworth, NJ, USA) was used.
Column chromatographic separations were carried out by a flash chromatography system using silica gel and hexane/ethyl acetate or petroleum ether/ethyl acetate solvent mixtures. For thin layer chromatography (TLC), Merck silica gel (grade 60 F 245 , Merck & Co., Kenilworth, NJ, USA) was used.

General Photooxidation Method
In a suitable 4 mL pyrex cell were initially added 0.5-2.0 mL of the solvent and the CD-g-C 3 N 4 catalyst in concentration 4 mg mL −1 . Next, the compound, which will be oxidized was added, using a 25 µL syringe. The corresponding amounts were: 93 mM of tetramethylethylene (1), 12 × 10 −3 mM of (E)-2-(1,1,1-trideuteromethyl)-5-methylhex-2-ene (alkene 2), 130 mM of 1,3-cyclohexadiene (3) and 110 mM of 1-methoxy-4-(3-methylbut-2-en-2-yl)benzene (aryl alkene 4). Also, the photosensitizers TTP, MB, DCA or RB were used for control experiments in catalytic amounts (0.5-1.0 mg). In the experiments which include the quenchers; 1,4-diazabicyclo[2.2.2]octane (DABCO) or 1,3,5-trimethoxybenzene, they were added in the amount of 1% mole in proportion to the oxidizing substance. The test tube was then placed into an ice bath in a pyrex receptacle, at a distance of approximately 10 cm from the visible light source (variable intensity Xenon bulb, Cermax 300 W). Oxygen was transferred inside the test tube by a syringe under gentle and continuous flow (gentle bubbling). The reaction was usually left for 30 min., or 60 min. At the end, the carbon nitride catalyst was removed by simple filtration through cotton cloth and celite. In the cases where deuterated solvents were used, they were not evaporated, and the conversion was determined by 1 H NMR analysis. In all the other cases with non-deuterated solvents, either they were evaporated under reduced pressure and the conversion was determined by 1 H NMR analysis using CDCl 3 or with aprotic solvent the reaction progress was monitored by gas chromatography.

Conclusions
In conclusion, we have shown that CD-C 3 N 4 photocatalyst sensitizes the photooxidation of unsaturated compounds. With simple alkenes and dienes, the singlet oxygen mechanism operates efficiently producing the allylic hydroperoxides and endoperoxides respectively. In the case of aryl alkenes, efficient electron donor molecules, both 1 O 2 and/or electron transfer mechanisms can be achieved. Which of these competing mechanisms will prevail depends on the nature of substrate and the protic-non protic solvents. Finally, the easy modification of graphitic carbon nitride to bulk amounts of CD-C 3 N 4 shows the following advantages: a) photocatalyzes the oxidation of a variety of alkenes and dienes b) the reaction is heterogeneous, a simple celite filtration removes the catalyst which can be reused at least three times c) stability of the catalyst under the photooxidation conditions unlike most of the conventional photosensitizers and d) the heterogeneous nature of the present photocatalyst overcomes solubility problems.