Synthesis of Bis(amino acids) Containing the Styryl-cyclobutane Core by Photosensitized [2+2]-Cross-cycloaddition of Allylidene-5(4H)-oxazolones

Abstract

The irradiation of 2-aryl-4-(E-3′-aryl-allylidene)-5(4H)-oxazolones 1 with blue light (456 nm) in the presence of [Ru(bpy)₃](BF₄)₂ (bpy = 2,2′-bipyridine, 5% mol) gives the unstable cyclobutane-bis(oxazolones) 2 by [2+2]-photocycloaddition of two oxazolones 1. Each oxazolone contributes to the formation of 2 with a different C=C bond, one of them reacting through the exocyclic C=C bond, while the other does so through the styryl group. Treatment of unstable cyclobutanes 2 with NaOMe/MeOH produces the oxazolone ring opening reaction, affording stable styryl-cyclobutane bis(amino acids) 3. The reaction starts with formation of the T₁ excited state of the photosensitizer ³[Ru*(bpy)₃]²⁺, which reacts with S₀ of oxazolones 1 through energy transfer to give the oxazolone T₁ state ³(oxa*)-1, which is the reactive species and was characterized by transient absorption spectroscopy. Measurement of the half-life of ³(oxa*)-1 for 1a, 1b and 1d shows large values for 1a and 1b (10–12 μs), while that of 1d is shorter (726 ns). Density functional theory (DFT) modeling displays strong structural differences in the T₁ states of the three oxazolones. Moreover, study of the spin density of T₁ state ³(oxa*)-1 provides clues to understanding the different reactivity of 4-allylidene-oxazolones described here with respect to the previously reported 4-arylidene-oxazolones.

1. Introduction

The role of amino acids as key building blocks in the synthesis of peptides and proteins is critical [1,2]. Each parameter of their three-dimensional structure, as well as their electronic and steric characteristics, must be carefully designed to produce the necessary effect within the structure of the protein that hosts them. Given their relevance, it is clear that having versatile and selective synthetic methods for the preparation of amino acid analogues of natural ones is highly desirable. In fact, interest in new methods for the synthesis of unnatural amino acids derives from the possibility of introducing small controlled changes that lead to unusual structural situations and that allow modulation of the activity of the proteins [3,4,5,6,7,8,9,10].

The synthesis of truxillic and truxinic acids, originally isolated from plant extracts [11,12], has experienced growing interest in recent years due to their pharmacological properties (Figure 1a). In this respect, they show strong anti-inflammatory and antinociceptive activities through inhibition of different targets, such as fatty acid binding protein (FABP), anoctamin-1, NO, and radicals [13,14,15,16,17,18,19,20]. They also show antioxidant and anti-hypoglycaemic activities [21] and cytotoxic activity [22]. In addition, they are involved in glucose metabolism through activation of peroxisome proliferator-activated receptors (PPAR) [23,24]. Beyond biological activities, they also show interesting properties as internal donors in Ziegler–Natta catalysts [25].

Closely related to the truxillic and truxinic acids are the unnatural bis-amino acids 1,3-diaminotruxillic and 1,2-diaminotruxinic derivatives (Figure 1b), which also show antinociceptive activity, and, moreover, are promising substrates in the treatment of diabetes type-II because they are the only non-peptidic agonists of the Glucagon Like-Peptide-1 (GLP-1) receptor [26,27,28]. The basic scaffold of 1,3-diaminotruxillic and 1,2-diaminotruxinic amino acids is a cyclobutane; therefore, the simplest synthetic approximation for their preparation could be the [2+2]-photocycloaddition of two olefins [29], more specifically, two 4-aryliden-5(4H)-oxazolones (Figure 1b). Despite its apparent simplicity, this synthetic method has been scarcely investigated, and, as far as we know, only four reports have appeared in the literature (Figure 2).

Between 2007 and 2012, Wang and coworkers reported two examples of [2+2]-photocycloaddition of oxazolones, involving irradiating oxazolone for 3 days with a high-power Hg lamp (500 W); they obtained the corresponding 1,3-diaminotruxillic cyclobutanes as four different isomers with a global yield of 10% [26,27,28]. In 2017, our group improved both the scope and the yield of the reaction using low-power (20 W) LED blue light (465 nm), but still a mixture of diaminotruxillic derivatives was obtained [30] (Figure 2a). Further contributions from Amarante [31] and our group [32] have included the synthesis of 1,2-diaminotruxinic derivatives by irradiation of aryliden-oxazolones with blue light in the presence of a photocatalyst ([Ru(bpy)₃](BF₄)₂, versus Eosin Y, Figure 2b,c). Remarkably, the reaction in the presence of the photocatalyst takes place with a different orientation with respect to that in its absence (1,2- instead of 1,3-coupling) and with complete stereoselectivity, since the corresponding 1,2-diaminotruxinic derivatives were obtained as single isomers. This fact is quite noteworthy because this methodology allows control of the regio- and diastereoselectivity, so the method is simple, cheap, and efficient. The geometric isomers obtained were different as a function of the photocatalyst since the μ-isomer was obtained in the case of the Ru-photosensitized reaction, while the zeta-isomer was obtained for the eosin Y photocatalyst. All these observations give an idea of the potential of this known methodology to be applied to the synthesis of bis-amino acids using oxazolones as starting materials.

However, as mentioned previously, these processes have only been studied using 4-aryliden-5(4H)-oxazolones as precursors. These substrates contain only one C=C bond able to be activated, so the number of isomers of the bis-amino acid is limited. The introduction of additional activatable C=C bonds in the molecule would give access to new bis-amino acids not achievable by other synthetic methods. In addition, the possibility to control regio- and stereoselectivity, as the number of C=C double bonds increases, is highly challenging.

Here, we report the results obtained from the irradiation of 2-aryl-4-(E-3′-aryl-allylidene)-5(4H)-oxazolones 1, in solution and in the presence of the [Ru(bpy)₃](BF₄)₂ photocatalyst. These oxazolones contain an additional C=C bond conjugated with the exocyclic C=C bond, resulting in different reactivity. As well as the structural characterization of the resulting bis-amino acids, the determination of the photochemical active species was accomplished by transient absorption spectroscopy and density functional theory (DFT) modeling. The theoretical studies also provide clues for the understanding of the different reactivity found here.

2. Results and Discussion

2.1. Synthesis of the 2-Aryl-4-(E-3′-aryl-allylidene)-5(4H)-oxazolones 1a–1h

The oxazolones 1a-1h, shown in Figure 3, were prepared following the classical Erlenmeyer–Plöch method [33,34,35,36,37,38,39,40]. The cinnamaldehydes and the hippuric acids employed in the synthesis were selected in order to cover a wide set of possibilities, with electron-releasing and electron-withdrawing substituents in the two aryl rings and in the allyl group. Oxazolones 1a, 1c, and 1d appear on SciFinder, but 1c and 1d have not been characterized. Therefore, they are fully characterized here. For oxazolone 1a, data were given [41] but in different solvents (CDCl₃, dmso-d₆) than that reported here (toluene-d₈). Oxazolones 1b, 1e-1h are new compounds and were fully characterized by NMR, MS, and X-ray diffraction data. All the oxazolones were obtained in moderate to good yields, except oxazolone 1g which is highly soluble in ethanol (see Materials and Methods for details).

NMR data (Figures S1–S80) showed that oxazolones 1d and 1e (Figures S17–S30) were obtained as single isomers. The configuration of the C=C double bonds of the C(H)=C(Me)-C(H)=C fragment can be clearly inferred in the ¹H-selective NOESY of 1d. The perturbation of the signal at 7.16 ppm (H_3’, see Figure 3 and Figure S23 of Supplementary Materials) produced a strong NOE in the signal at 7.07 ppm (H_1’), as well as in the signal assigned to the ortho H of the 3′-Ph ring (see also Materials and Methods), showing their proximity. This means that the configuration of the C3’=C2’ double bond is E in 1d. The same conclusions can be derived from the observation in 1e of a NOE between the peaks at 8.44 ppm (H_3’) and 7.08 ppm (H_1’), although in this case, due to the presence of the Br atom, the configuration is Z. The configuration of the exocyclic C1’=C4 double bond was determined, in turn, by measurement of the ³J_CH coupling constant of the carbonyl signal with H_1’ in the ¹³C NMR spectrum of 1d. The value determined (5.2 Hz) shows that this C=C bond has a Z-configuration by comparison with reported data [42,43]. Therefore, 1d and 1e show EZ and ZZ configurations, respectively (Figure 3).

However, analysis of the NMR data in toluene-d₈ of 1a–1c (Figures S1–S16) and 1f–1g (Figures S31–S45) showed the presence of a mixture of two isomers in 3.3/1 to 1.7/1 molar ratios, depending on the oxazolone. Two sets of signals assigned to the protons of the -C3’(H)=C2’(H)-C1’(H)=C4 fragment were observed, with ³J_H3’H2’ values identical in the two isomers (15.7 Hz), suggesting an E configuration of the C3’=C2’ double bond in both isomers. As expected, the large ³J_H1’H2’ value (11.6 Hz) suggests the formation of the s-trans rotamer of the diene system. Thus, the source of isomerism in these oxazolones must be located in the relative disposition of the oxazolone ring with respect to the C1’=C4 bond. The proton-coupled ¹³C NMR spectrum of 1a showed the presence of two doublet peaks assigned to the CO group. The major species showed a value of the coupling constant ³J_CH = 4.2 Hz. According to previous reports [42,43], this suggests that the C1’=C4 bond has a Z-configuration; therefore, the major isomer of the oxazolone is EZ. As expected, in the minor isomer, this value was ³J_CH = 11.6 Hz, suggesting an E-configuration for this bond and that the oxazolone is EE.

The determination of the crystal structure of 1a provides additional information. A molecular drawing is shown in Figure 4, while relevant crystallographic data and selected bond distances and angles are given as Supplementary Materials (Tables S2–S6).

The structure shows that the two C=C double bonds (C2-C10 and C11-C12) are trans, thus corresponding to the minor EE-isomer. The allylidene fragment C13-C12-C11-C10-C2 shows the expected long-short-long-short pattern of bond distances, corresponding to a C-C=C-C=C- molecular skeleton. Despite this pattern, the individual values of the bond distances show electronic delocalization throughout this system, as both the long bond distances (C12-C13 = 1.455(4) Å and C10-C11 = 1.427(4) Å), as well as the short ones (C11-C12 = 1.358(4) Å and C2-C10 = 1.361(4) Å) are in the respective range of distances found in the literature for this type of conjugated bonds [44]. Other internal parameters of the oxazolone and Ph rings do not show deviations with respect to similar compounds found in the literature [45,46,47,48].

2.2. Synthesis of the Cyclobutane-bis(oxazolones) 2 and 1,2-Diaminotruxinic Bis-amino Acids 3 by [2+2]-Photocycloaddition of Oxazolones 1

The irradiation of oxazolones 1a–1h in CD₂Cl₂ at room temperature with the blue light provided by a Kessil lamp (456 nm) was monitored by ¹H NMR and showed the appearance of many peaks in the 4–6 ppm region. This fact suggests the formation of the expected cyclobutanes, but the large number of peaks observed suggests the formation of many isomers and that the reaction takes place without selectivity. We previously observed that the introduction of a photosensitizer improved the selectivity of the reaction [32]. In this respect, the irradiation of oxazolones 1a, 1b, 1d, and 1f with blue light (456 nm, Kessil lamp) in deoxygenated CD₂Cl₂ at room temperature, under Ar atmosphere and in the presence of [Ru(bpy)₃](BF₄)₂ (5% mol), took place with complete conversion of the oxazolone 1 in 18 h and formation of the corresponding cyclobutanes 2a, 2b, 2d and 2f (Figure 5). The characterization of these cyclobutanes by NMR spectroscopy (Figures S46–S66) showed that each one was formed by a mixture of two isomers, one of them clearly as a major species with respect to the other (molar ratio 10.1/1 for 2a, 3.6/1 for 2b, 2.9/1 for 2d, 4.8/1 for 2f). For 2a and 2f this mixture could be separated by flash column chromatography. The structure determined from NMR data for isolated 2a and 2f is shown in Figure 5. For 2b and 2d, the mixture could not be separated, and the structure represented in Figure 5 was assigned to the major isomer. The structure of the minor isomers could not be determined due to their low molar ratio.

The formation of only two isomers for cyclobutanes 2a, 2b, 2d, and 2f, one of them in a clear majority, is remarkable (i.e., more than 80 possible isomers) and shows that the photosensitized reactivity of allyliden-5(4H)-oxazolones 1 also has a high degree of selectivity. Despite this, the scope of this reaction was more limited, since not all attempted oxazolones showed clear reactivity. The monitoring of the irradiation of 1c, 1e, 1g, and 1h under the same reaction conditions (456 nm, Ru 5%, CH₂Cl₂, Ar) by ¹H NMR showed complete conversion of the oxazolones 1, but also the presence of many peaks in the 4–6 ppm region, suggesting the formation of many different isomers of the respective cyclobutanes 2c, 2e, 2g and 2h. These mixtures proved to be intractable and were not further analyzed.

The cyclobutanes 2a, 2b, 2d, and 2f were stable once isolated as solids, but, in solution at room temperature, they underwent a thermal retro [2+2] reaction giving back the starting oxazolones 1a, 1b, 1d, and 1f. This low stability is related to intramolecular steric strains. It was also observed in μ-truxinic derivatives [32], and it seems to be responsible for the low yields observed after chromatographic purification (Figure 5). The simplest strategy to eliminate this steric constraint and to obtain stable cyclobutane derivatives is the transformation of the oxazolone ring in compound 2 into the corresponding ester 3 (Figure 6). The reaction can be performed in a one-pot, two-steps way, without isolating the cyclobutane intermediate 2, therefore minimizing the erosion of yield during isolation of 2.

This strategy was exemplified in the cases of 1a and 1b, as represented in Figure 6. The irradiation of 1a and 1b in deoxygenated CH₂Cl₂ in the presence of the Ru-photosensitizer was performed as reported above. Once the full conversion of 1a,b was observed (18 h), the reaction solvent was evaporated to dryness, while irradiation was kept to minimize the retro [2+2] reaction in intermediates 2. The dry residues were suspended in methanol and subjected to ring opening reaction by treatment with NaOMe and heating, giving the 1,2-diaminotruxinic acids 3a and 3b in good to excellent yields (Figure 6). As is evident from Figure 6, the improvement in the reaction yields was more than notable. As expected, both 3a and 3b were stable in solid and in solution and could be purified, crystallized, and characterized in solution without observing retro-[2+2] or other side-reactions. Therefore, this is a very convenient method for the synthesis of this kind of new bis-amino acid.

The characterization of compounds 2 and 3 was carried out on the basis of their MS and NMR data (Figures S46–S80). The ESI⁺ spectra of 2 and 3 showed perfect agreement with the structures shown in Figure 5 and Figure 6. The analysis of the NMR spectra of 2 (Figures S46–S66) enabled inference of the cyclobutane scaffold, but the absence of significative NOE cross-peaks precluded knowledge of the spatial distribution of the cyclobutane substituents, and hence, their full structural characterization. Fortunately, the NMR data of 3a and 3b (Figures S67–S80 of Supplementary Materials) allowed an unambiguous structural determination, showing that photodimerization occurred between the exocyclic C_1’=C₄ bond attached to the oxazolone ring of one molecule and the styryl C_3’=C_2’ bond of the other molecule (Figure 5 and Figure 6). The simultaneous presence in the NMR spectra of signals due to the vinyl oxazolone proton (C_1’(H)=C) and the styryl fragment (Ar-C_3’(H)=C_2’(H)-) supports this hypothesis. The ¹H COSY spectrum of 3a (Figure S71) showed a clear correlation between the vinylic protons and the peak at 3.76 ppm, which was then assigned to H-C1’. The COSY spectrum of 3a also showed a correlation between the proton at 3.30 ppm with those appearing at 3.76 ppm and 4.68 ppm, but no correlation was observed between the latter, showing that the signal at 3.30 ppm was due to the H at C3’. The multiplicity of the signals due to the three chemically inequivalent protons at C1’, C2’ and C3’ showed that only head-to-head coupling was possible. In turn, the peak at 4.68 ppm (H at C2’) correlated with the peak at 6.50 ppm, which was then assigned to the proton at the N-C=CH group. This structural information of 3a is presented in Figure 7a, and similar conclusions can be obtained from the analysis of the ¹H COSY of 3b.

To determine the relative spatial arrangement of the cyclobutane substituents, the ¹H NOESY spectrum of 3a (Figure S72) was measured, as well as the selective ¹H 1D-SELNOESY spectra of 3b (Figures S78–S80). The ¹H NOESY spectrum of 3a showed clear NOE correlations between the H at 3.30 ppm (H_3’) and the ortho H of the styryl fragment, as well as with the H at 6.50 ppm, showing that all these groups were close in space and pointing to the same side of the molecular plane defined by the cyclobutane (upwards of the plane in Figure 7b). The vinylic H at 6.50 ppm also showed an intense NOE with one of the NH protons (7.72 ppm); this indicated that the configuration of this alkene fragment was E. In addition, the signal at 4.68 ppm (H-C_2’) showed strong NOEs with the H_ortho of the Ph ring at C_3’ and with the NH at 8.25 ppm, allowing their full assignation and showing that all these groups were also on the same side of the molecular plane (downwards of the plane in Figure 7b). Once the structure of 3 was established and taking into account that the ring opening reaction of the oxazolone did not alter the configuration of the cyclobutane carbons [32], we assume the structures shown in Figure 5 for the cyclobutane-bis(oxazolones) 2.

2.3. Characterization of the Oxazolone Reactive Excited State by Transient Absorption Spectroscopy

The remarkable difference among the cyclobutanes 2a, 2b, 2d, 2f and the truxinic derivatives 3a, 3b with respect to examples previously reported is that, in the cyclobutane rings 2 and 3, the exocyclic C=C bond attached to the oxazolone has reacted with a C=C bond quite differently, instead of with another exocyclic C=C bond. This fact is of considerable interest because it opens the door for new heterodimerizations involving oxazolones and other alkenes as sources of new cyclobutane-amino acids. A key point for understanding this different reactivity is the characterization of the reactive species in the excited state; this task was accomplished using transient absorption spectroscopy.

A comparison of the absorption spectrum of [Ru(bpy)₃](BF₄)₂, which showed strong absorption at 460 nm [49], with the spectra of the oxazolones 1 (having broad absorption around 380–400 nm; see Figures S81–S84 of Supplementary Materials) indicated that the absorption of the incident light of the Kessil lamp (456 nm) was mostly produced by the ruthenium species. This excitation resulted in the formation of its singlet excited state ¹[Ru(bpy)₃²⁺]*, which was followed by fast ISC (intersystem crossing) to give its reactive triplet excited state ³[Ru(bpy)₃²⁺]*. Hence, there are two main possibilities for the reaction between ³[Ru(bpy)₃²⁺]* and the oxazolone: either an electron transfer (a redox process) or an energy transfer. Taking into account that the redox potentials reported for the Ru species are [Ru(bpy)₃]³⁺/[Ru(bpy)₃]²⁺ and [Ru(bpy)₃]²⁺/[Ru(bpy)₃]⁺ (+1.29 V and −1.33 V vs. SCE, respectively) [49,50,51], and that the measured redox potentials for oxazolone 1a (Figures S92 and S93 of Supplementary Materials; CH₂Cl₂; NBu₄PF₆ 1M) are E⁰(oxazolone^•+/oxazolone) = +1.65 V and E⁰(oxazolone/oxazolone^•−) = −1.15 V, it is possible to see that both redox processes are thermodynamically unfavorable by ΔG⁰_et = +0.08 eV and + 0.60 eV, respectively. Therefore, we assume at this point that the reaction takes place through an energy transfer process from the ruthenium to the oxazolone, as will be shown later (Section 2.4).

The energy transfer from ³[Ru(bpy)₃²⁺]* to oxazolone promoted efficient quenching of the emission of the ruthenium sensitizer when increasing amounts of oxazolone 1a were added. As previously reported, the transient absorption spectrum of the ³[Ru(bpy)₃²⁺]* (Figure S85) after laser excitation at 532 nm showed a stimulated emission at 620 nm, a ground state bleaching at 450 nm, and a transient absorption band at 360 nm [32]. The three bands showed the same kinetic behavior, confirming that all of them belonged to the same ruthenium species. The effect of the addition of increasing the amounts of oxazolone 1a to a [Ru(bpy)₃]²⁺ solution was to produce effective quenching of the emission, as is clearly shown in Figure 8a. Analysis of the decay using the Stern–Volmer equation produced a value of 4.25 × 10¹⁰ M⁻¹s⁻¹ for the deactivation constant, showing that the reaction was diffusion-controlled [52]. Similar values of this constant were found for the oxazolones 1b and 1d (Figures S86–S91 and Table S1 in Supplementary Materials).

Moreover, the measurement of the transient absorption spectra of deoxygenated solutions of [Ru(bpy)₃]²⁺ in CH₂Cl₂ and in the presence of oxazolones 1 at different times after the laser pulse enabled visualization of the formation of a new transient species. This is shown in Figure 8b, where it is clear that the spectrum of the triplet excited state ³[Ru(bpy)₃²⁺]* vanished with time, and a new species with an absorption band at 450 nm emerged and disappeared in a few microseconds. At a short interval after the pulse (12 or 44 ns, black or red lines, respectively) the main species was the triplet of the Ru, while at long intervals (552 or 924 ns, grey and ochre lines), the triplet of the Ru had almost disappeared due to energy transfer to the oxazolone and the subsequent formation of a transient species, which was visible in the emerging absorption at 450 nm. Our proposal is that this transient species is the oxazolone excited state, which appears as the photosensitizer disappears. Figure 9 shows the decay curve (black line) measured at 470 nm of the new species detected in deoxygenated solutions by the reaction of 1a with ³[Ru(bpy)₃²⁺]*. The half-life of this intermediate, determined from the exponential decay, was 12.15 μs in deoxygenated solution. Both the half-life of this intermediate, in the scale of microseconds, as well as the fast deactivation of this species in the presence of O₂ (red line in Figure 9; a value of 1 × 10⁹ M⁻¹s⁻¹ was determined for the deactivation rate constant) points to the triplet nature of this excited state. A similar value for the half-life was measured for 1b (10.72 μs), while for 1d, the value was shorter (726.1 ns), although this still suggests a triplet nature of the excited state. From all these data, we can conclude that the ruthenium species behaves as a photosensitizer of the oxazolones 1, whose reactive state ³[oxa-1]* is a triplet excited state generated, presumably, by an energy transfer process from ³[Ru(bpy)₃²⁺]* to the oxazolone oxa-1, as shown in Figure 10. Due to the importance of the correct characterization of the reactive species, it was further studied using DFT methods.

2.4. Characterization of the Reactive Excited State by DFT Methods

We employed DFT (ωB97X-D/def2-QZVPP//ωB97X-D/6-31+G(d), SMD = dichloromethane in all calculations) [53,54,55,56,57,58,59,60,61] to gain insights regarding the T₁ excited state of 1a, ³[oxa-1a]*, which could reveal the reasons behind the reactivity differences shown by the allyliden-5(4H)-oxazolone 1a and aryliden-5(4H)-oxazolones 4 (Figure 11). In this study, we focused on the major isomer E,Z-1a observed experimentally.

There were significant variations in the excited state properties of the two oxazolone derivatives considered. The Gibbs free energy (G) gap was considerably lower in 1a (29.8 vs. 35.7 kcal·mol⁻¹), which was probably due to the higher delocalization of the two unpaired electrons from T₁ of 1a with the longer conjugated system (Figure 11A). Moreover, the atoms with higher spin density differed: in 1a, the α and δ C atoms of the diene system showed the highest spin density (0.58 and 0.35 unpaired electrons), while in 4, the α and β positions were predominant (0.70 and 0.42 unpaired electrons). The unpaired electrons accumulated in the reactive positions in each species, explaining the switch in reactivity observed.

The T₁ geometries of the two oxazolones also differed significantly. While in 1a, the conjugated system was planar (θ = 178°), the Ph ring was perpendicular to the oxazolone ring in 4 (θ = 96°) (Figure 11A). Interestingly, we found that 1a and 4 showed rotation transition states (TSs) to interconvert from E to Z structures with low barriers (2.6 and 2.5 kcal·mol⁻¹, respectively, Figure S94). The low calculated barriers were consistent with previous photophysical experimental results, which indicated that the same triplet state was rapidly formed when starting from the E and Z forms of compound 4 [32]. These rapid isomerization processes may have been an important contributing factor to the low yields obtained experimentally since multiple competitive isomers of 2 can be formed. We studied how these T₁ geometries distorted over time using molecular dynamics (MD, Figure 11B) [62]. The population distribution of 1a was relatively narrow and rested around the planar system (average θ = 168°, approximate range of θ = 130-180°). This result suggested that the perpendicular rotation TS (θ = 90°) with low energy was not immediately reached. In contrast, the geometry of 4 can easily switch from the planar to the perpendicular geometry (average θ = 110°, approximate range of θ = 60–180°).

Figure 11. (A) G difference between optimized geometries of 1a and 4 in S₀ and T₁, along with natural spin populations [63] of their conjugated systems and representations of T₁ geometries. (B) Distribution of the θ dihedral angle from MD simulations of 1a and 4 in T₁. Since rotations were symmetrical in both directions, any values above 180° were converted to their equivalent degrees between 0 and 180° for clarity (i.e., 190° would be equivalent to 170°).

Figure 11. (A) G difference between optimized geometries of 1a and 4 in S₀ and T₁, along with natural spin populations [63] of their conjugated systems and representations of T₁ geometries. (B) Distribution of the θ dihedral angle from MD simulations of 1a and 4 in T₁. Since rotations were symmetrical in both directions, any values above 180° were converted to their equivalent degrees between 0 and 180° for clarity (i.e., 190° would be equivalent to 170°).

3. Materials and Methods

3.1. General Procedures

The [2+2] photocycloadditions were performed under Ar atmosphere, using deoxygenated CH₂Cl₂. Other reactions were carried out in reagent-grade solvents in open air. Flash column liquid chromatographies were carried out on silica gel (70−230 μm). ¹H, ¹³C, and ¹⁹F NMR spectra were measured in CDCl₃, CD₂Cl₂, or toluene-d₈ solutions at 25 °C on Bruker AV300 or Bruker AV500 spectrometers (δ in ppm, J in Hz) at ¹H operating frequencies of 300.13 MHz and 500.13 MHz, respectively. ¹H and ¹³C NMR spectra were referenced using the solvent signal as the internal standard, while ¹⁹F NMR spectra were referenced to CFCl₃. The assignment of ¹H NMR peaks was performed with the help of 2D ¹H−COSY, 2D ¹H-NOESY and 1D ¹H SELNOE experiments (mixing times of 1.2−1.8 s, as a function of the irradiated signal), while ¹³C NMR peaks were identified using ¹H−¹³C edited HSQC and ¹H−¹³C HMBC 2D experiments. HRMS and ESI (ESI+) mass spectra were recorded using a MicroToF Q, API-Q-ToF ESI instrument with a mass range from m/z 20 to 3000 and a mass resolution of 15,000 (full width at half-maximum). The absorption spectra of 1a−1h were measured in an Evolution 600 UV−Vis spectrophotometer in CH₂Cl₂ solutions (10⁻⁵ M), the excitation and emission spectra were recorded on a Horiba Jobin Yvon Fluoromax-P spectrophotometer, and the quenching of the phosphorescence of [Ru(bpy)₃](BF₄)₂ by 1a was measured on a Horiba Jobin Yvon Fluorolog FL-3.11 spectrophotometer. All spectra were recorded at 25 °C using 10 mm quartz cuvettes. The cyclic voltammetry of 1a (5 · 10⁻⁴ M) was carried out using a Voltalab50 potentiostat/galvanostat, equipped with a glass electrochemical cell with the typical configuration of three electrodes: a Pt working electrode, another Pt counter electrode, and the SCE electrode. The solution of the pure electrolyte (NBu₄PF₆, 0.1 M) was measured over the whole window of the solvent (CH₂Cl₂) to check the absence of electroactive impurities. The oxazolones 1a–1h were prepared following published methods [33,34,35,36,37,38,39,40]. The photosensitizer [Ru(bpy)₃](BF₄)₂ was synthesized following methods found in [64,65] and stored under Ar at 4 °C. The melting points (degrees Celsius) were determined on a Gallenkamp apparatus and were uncorrected.

3.2. Irradiation Setup

The solution of oxazolone 1a–1h in a Schlenck flask was irradiated at 456 nm by a Kessil PR160L LED lamp (maximal power of 50W, but this intensity can be tuned). The lamp and the flask were placed 5 cm apart to avoid overheating the solution. To maximize the light of the lamp received by the solution, a mirror was placed in front of the lamp.

3.3. X-ray Crystallography

The crystals of oxazolone 1a were obtained by slow evaporation at room temperature of a CDCl₃ solution of 1a. The crystals were assembled at low temperature (100 K) using a specific commercial system called MiTeGen micromunts Cryoloop. Diffraction data were obtained on a Bruker D8 Venture diffractometer using Mo-Kα radiation (λ = 0.71073 Å), filtered through a graphite monochromator and multilayer optics, at low temperature (100 K). The diffraction frames were integrated using SAINT [66] and the integrated intensities were corrected for absorption with SADABS [67]. The structure was determined and developed by Fourier methods [68]. All non-hydrogen atoms were refined with anisotropic thermal parameters. The H atoms were placed at idealized positions and treated as riding atoms. Both structure resolution and structure refinement were carried out using a commercial Bruker package (Bruker APEX3 software package) [69]. CCDC-1972175 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif (accessed on 9 December 2020).

3.4. Photophysical Experiments

The laser flash photolysis (LFP) experiments were performed using a pulsed Nd:YAG SL404G-10 Spectron Laser Systems laser at an excitation wavelength of 532 nm. The energy of the single pulses (around 10 ns duration) was smaller than 15 mJ pulse⁻¹. The LFP system is formed by the following elements: a pulsed laser, a pulsed Lo255 Oriel Xe-lamp, an Oriel monochromator model 77,200, an Oriel photomultiplier tube (PMT) housing, a power supply model 70,705 PMT, and a Tektronix oscilloscope TDS-640A. Quenching rate constants (k_q) were determined according to the Stern−Volmer equation 1/τ = 1/τ_ο + k_q[Q]. In this equation τ_ο is the triplet lifetime of Ru(bpy)₃²⁺ in the absence of oxazolone (Q), τ is the lifetime of ³[Ru(bpy)₃²⁺]* in the presence of a given concentration of oxazolone, and [Q] is the oxazolone concentration. The quenching rate constants (k_q, M⁻¹ s⁻¹) were the corresponding slopes of the linear fittings of the Stern−Volmer plots.

3.5. General Procedure for the Synthesis of 2-Aryl-4-(E-3′-aryl-allylidene)-5(4H)-oxazolones 1a–1h

The synthesis of 1a–1h was carried out following published procedures [33,34,35,36,37,38,39,40], which are detailed here for 1a. All other oxazolones were prepared using the same procedure.

3.5.1. Synthesis of 2-Phenyl-4-(E-3′-phenylallylidene)-5(4H)-oxazolone 1a (EZ and EE Isomers)

A Schlenk flask was charged with cinnamaldehyde (0.776 g, 5.88 mmol), hippuric acid (1.054 g, 5.88 mmol), sodium acetate (0.467 g, 5.69 mmol), and acetic anhydride (4 mL). This mixture was stirred while heating at 100 °C for 2 h. Then, the resulting suspension was allowed to cool, giving oxazolone 1a as a viscous waxy solid. This solid was vigorously stirred with 25 mL of ethanol. The solid thus obtained was filtered, washed with additional ethanol (60 mL), dried by suction, and characterized as the mixture of EZ and EE isomers of oxazolone 1a. Obtained: 1.002 g (62% yield). Ratio 2.5/1 (EZ/EE). HRMS (ESI⁺) [m/z]: calculated for [C₁₈H₁₃NO₂Na]⁺ = 298.0844; found 298.0833. ¹H NMR major isomer (EZ) (toluene-d8, 300.13 MHz, 25 °C): δ = 7.99 (dm, 2H, H_o, NCOPh, ³J_HH = 7.7 Hz), 7.66 (dd, 1H, H_2’, ³J_HH3 = 15.7 Hz, ³J_HH1 = 11.6 Hz), 7.17 (m, 2H, H_o, Ph), 7.10–6.95 (m, 6H, H_m + H_p, NCOPh + H_m + H_p, Ph), 6.88 (dd, 1H, H_1’, ³J_HH = 11.6 Hz, ⁴J_HH = 0.8 Hz), 6.51 (d, 1H, H_3’, ³J_HH = 15.7 Hz). ¹³C{¹H} NMR major isomer (EZ) (toluene-d8, 300.13 MHz, 25 °C): δ = 166.12 (CO), 162.62 (CN), 143.03 (C_3’), 132.74 (C_p, NCOPh), 132.37 (C_1’), 136.53, 134.76, 126.49 (C_i, NCOPh + C_i, Ph + C_q, oxa, C₄), 129.67(C_p, Ph), 128.97, 128.85 (C_m, Ph + C_m, NCOPh), 128.33 (C_o, Ph), 128.04 (C_o, NCOPh), 123.74 (C_2’).

3.5.2. Synthesis of 2-Phenyl-4-(E-3′-(4-chlorophenyl)allylidene)-5(4H)-oxazolone 1b (EZ and EE Isomers)

1b, yellow solid. 4-chlorocinnamaldehyde (1.042 g, 6.28 mmol), hippuric acid (1.084 g, 6.05 mmol), sodium acetate (0.528 g, 6.43 mmol) and acetic anhydride (2 mL). Obtained: 1.368 g (74% yield). Ratio 1.7/1 (EZ/EE). HRMS (ESI⁺) [m/z]: calculated for [C₁₈H₁₂ClNO₂Na]⁺ = 332.0454; found 332.0441. ¹H NMR major isomer (EZ) (CDCl₃, 500.13 MHz, 25 °C): δ = 8.14 (m, 2H, H_o, NCOPh), 7.65 (dd, 1H, H_2’, ³J_HH3 = 15.6 Hz, ³J_HH1 = 11.6 Hz), 7.61–7.49 (m, 5H, H_m, H_p, NCOPh + H_o, C₆H₄), 7.37 (m, 2H, H_m, C₆H₄), 7.11 (dd, 1H, H_1’, ³J_HH = 11.6 Hz, ⁴J_HH = 0.8 Hz), 7.06 (d broad, 1H, H_3’, ³J_HH = 15.6 Hz). ¹³C{¹H} NMR major isomer (EZ) (CDCl₃, 125.7 MHz, 25 °C): δ = 166.68 (CO), 162.45 (CN), 135.77, 125.62, 125.61 (C_i, C₆H₄ + C_i, NCOPh + C_p, C₆H₄), 142.12 (C_3’), 134.48 (C_q, oxa, C₄), 132.32 (C_1’), 133.22, 129.24, 129.04, 128.98 (C_m, C_p, NCOPh + C_o, C_m, C₆H₄), 128.20 (C_o, NCOPh), 123.88 (C_2’).

3.5.3. Synthesis of 2-Phenyl-4-(E-3′-(2-nitrophenyl)allylidene)-5(4H)-oxazolone 1c (EZ and EE Isomers)

1c, yellow solid. 2-nitrocinnamaldehyde (1.122 g, 6.33 mmol), hippuric acid (1.128 g, 6.30 mmol), sodium acetate (0.556 g, 6.77 mmol) and acetic anhydride (2 mL). Obtained: 1.865 g (92% yield). Ratio 1.7/1 (EZ/EE). HRMS (ESI⁺) [m/z]: calculated for [C₁₈H₁₂N₂O₄Na]⁺ = 343.0695; found 343.0696. ¹H NMR major isomer (EZ) (toluene-d8, 500.13 MHz, 25 °C): δ = 7.96 (m, 2H, H_o, NCOPh, ³J_HH = 7.0 Hz), 7.46 (dd, 1H, H_2’, ³J_HH3 = 15.6 Hz, ³J_HH1 = 11.4 Hz), 7.39 (m, 1H, H₃, C₆H₄), 7.10–6.95 (5H, H_m, H_p, NCOPh, H_6, C₆H₄, H_1’), 6.78 (m, 1H, H₄, C₆H₄), 6.74 (dd, 1H, H_3’, ³J_HH2 = 11.4 Hz, ⁴J_HH3 = 1.0 Hz), 6.64 (m, 1H, H₅, C₆H₄). ¹³C{¹H} NMR major isomer (EZ) (CDCl₃, 125.7 MHz, 25 °C): δ = 166.35 (CO), 163.34 (CN), 148.31 (C_q-NO₂, C₆H₄), 137.19, 133.61, 133.38, 129.86, 129.12, 128.72, 128.46, 127.86 (C_m, C_p, NCOPh + C₂, C₃, C₄, C₅, C₆, C₆H₄), 136.09, 131.52, 125.51 (C_i, C₆H₄ + C_i, NCOPh + C_q, Oxazolone, C₄), 131.31 (C₁), 125.19 (C₃, Ph).

3.5.4. Synthesis of (Z)-2-Phenyl-4-(E-2′-methyl-3′-phenylallyliden)-5(4H)-oxazolone 1d

1d, yellow solid. α-methyl-cinnamaldehyde (1.93 mL, 13.82 mmol), hippuric acid (2.485 g, 13.87 mmol), sodium acetate (1.355 g, 16.51 mmol) and acetic anhydride (4 mL). Obtained: 3.484 g (87% yield). HRMS (ESI⁺) [m/z]: calculated for [C₁₉H₁₅NO₂Na]⁺ = 312.1000; found 312.0991. ¹H NMR (CDCl₃, 500.13 MHz, 25 °C): δ = 8.12 (dm, 2H, H_o, NCOPh, ³J_HH = 7.0 Hz), 7.59 (tt, 1H, H_p, NCOPh, ³J_HH = 7.0 Hz, ⁴J_HH = 2.2 Hz), 7.51 (m, 2H, H_m, NCOPh), 7.46 (m, 2H, H_o, Ph), 7.42 (m, 2H, H_m, Ph), 7.34 (tt, 1H, H_p, Ph, ³J_HH = 7.0 Hz, ⁴J_HH = 2.2 Hz), 7.16 (s broad, 1H, H_3’), 7.07 (d, 1H, H_1’, ⁴J_HH = 1.0 Hz), 2.59 (d, 3H, 2-Me, ⁴J_HH = 1.0 Hz). ¹³C{¹H} NMR (CDCl₃, 125.7 MHz, 25 °C): δ = 168.32 (CO), 161.86 (CN), 143.93 (C_3’), 137.74 (C_1’), 136.64 (C_i, Ph), 135.83 (C_q, oxazolone, C₄), 133.09 (C_p, NCOPh), 132.67 (C_2’), 130.05 (C_o, Ph), 129.03 (C_m, Ph), 128.62 (2C overlapped, C_m, NCOPh + C_p-Ph), 128.21 (C_o, NCOPh), 126.01 (C_i, NCOPh), 16.81 (Me).

3.5.5. Synthesis of (Z)-2-Phenyl-4-(Z-2′-bromo-3′-phenylallylidene)-5(4H)-oxazolone 1e

1e, yellow solid. α-bromo-cinnamaldehyde (2.461 g, 11.72 mmol), hippuric acid (2.110 g, 11.78 mmol), sodium acetate (0.962 g, 11.69 mmol) and acetic anhydride (4 mL). Obtained: 1.282 g (32% yield). HRMS (ESI⁺) [m/z]: calculated for [C₁₈H₁₂BrNO₂Na]⁺ = 375.9949; found 375.9954. ¹H NMR (CDCl₃, 500.13 MHz, 25 °C): δ = 8.44 (s, 1H, H_3’), 8.15 (m, 2H, H_o, NCOPh, ³J_HH = 7.2 Hz), 7.89 (m, 2H, H_o, Ph, ³J_HH = 7.0 Hz), 7.63 (tt, 1H, H_p, NCOPh, ³J_HH = 7.4 Hz, ⁴J_HH = 1.1 Hz), 7.53 (m, 2H, H_m, NCOPh), 7.47–7.41 (m, 3H, H_p + H_m, Ph), 7.08 (s, 1H, H_1’).¹³C{¹H} NMR (CDCl₃, 125.7 MHz, 25 °C): δ = 166.59 (CO), 164.81 (CN), 141.86 (C_3’), 135.30 (C_i, Ph), 133.86 (C_p, NCOPh), 133.38 (C_q, oxazolone, C₄), 132.59 (C_1’), 130.12 (C_o, Ph), 129.93 (C_p, Ph), 129.06 (C_m, NCOPh), 128.62 (C_m, Ph), 128.43 (C_o, NCOPh), 125.20 (C_i, NCOPh), 114.75 (C_2’).

3.5.6. Synthesis of 2-(4-Cyanophenyl)-4-(E-3′-phenylallylidene)-5(4H)-oxazolone 1f (EZ and EE Isomers)

1f, yellow solid. Cinnamaldehyde (2.100 g, 15.89 mmol), 4-cyanohippuric acid (3.089 g, 15.12 mmol), sodium acetate (1.381 g, 16.82 mmol) and acetic anhydride (4.5 mL). Obtained: 1.990 g (65% yield). Ratio 3.3/1 (EZ/EE). HRMS (ESI⁺) [m/z]: calculated for [C₁₉H₁₃N₂O₂]⁺ = [M + H]⁺ = 301.0977; found 301.0960. ¹H NMR major isomer (EZ) (CDCl₃, 300.13 MHz, 25 °C): δ = 8.23 (dm, 2H, H_o, NCOC₆H₄CN), 7.80 (tm, 2H, H_m, NCOC₆H₄CN), 7.68 (dd, 1H, H_2’, ³J_H2H3 = 15.7 Hz, ³J_H2H1 = 11.6 Hz), 7.62 (m, 2H, H_o, Ph), 7.44–7.38 (m, 3H, H_m + H_p, Ph), 7.23 (dd, 1H, H_1’, ³J_H1H2 = 11.6 Hz, ⁴J_H1H3 = 0.9 Hz), 7.19 (d broad, 1H, H_3’, ³J_H3H2 = 15.6 Hz). ¹³C{¹H} NMR Major isomer (EZ) (CDCl₃, 75.5 MHz, 25 °C): δ = 166.08 (CO), 160.41 (CN), 145.63 (C_3’), 135.85 (C_i, Ph), 135.32 (C_1’), 133.54 (C_q, oxazolone, C₄), 132.75 (C_m, NCOC₆H₄), 130.57 (C_p, Ph), 129.86 (CN), 129.19 (C_m, Ph), 128.53 (C_o, NCOC₆H₄), 128.32 (C_o, Ph), 123.30 (C_2’), 118.05 (C_i, NCOC₆H₄), 116.21 (C_p, NCOC₆H₄).

3.5.7. Synthesis of 2-(E-Styryl)-4-(E-3′-phenylallylidene)-5(4H)-oxazolone 1g (EZ and EE Isomers)

1g, yellow solid. Cinnamaldehyde (0.735 g, 5.56 mmol), cinnamoylglycine (1.063 g, 5.18 mmol), sodium acetate (0.4222 g, 5.14 mmol) and acetic anhydride (1.5 mL). Obtained: 0.180 g (17% yield). Ratio 1.7/1 (EZ/EE). HRMS (ESI⁺) [m/z]: calculated for [C₂₀H₁₅NO₂]⁺ = 301.1103; found 301.1164. ¹H NMR major isomer (EZ) (CDCl₃, 300.13 MHz, 25 °C): δ = 8.10 (dd, 1H, H_2’, ³J_HH3 = 15.6 Hz, ³J_HH1 = 12.1 Hz), 7.70–7.35 (m, 11H, H_o, H_m, H_p, NCOPh + H_o, H_m, H_p, Ph + = CH_5”), 7.20 (dd, 1H, H_1’, ³J_HH = 12.1 Hz, ⁴J_HH = 0.7 Hz), 7.08 (d broad, 1H, H_3’, ³J_HH = 15.6 Hz), 6.72 (d, 1H, =CH_4”, ³J_HH = 16.2 Hz). ¹³C{¹H} NMR major isomer (EZ) (CDCl₃, 75.5 MHz, 25 °C): δ = 166.33 (CO), 161.69 (CN), 144.91 (C_3’), 142.70 (C_5”), 137.24 (C_1’), 135.99 (C_i, Ph), 134.69 (C_i, NCOPh), 133.82 (C_q, oxazolone, C₄), 130.57, 129.99, 129.08, 128.95, 128.04 (2C overlapped) (C_o, C_m, C_p, Ph + C_o, C_m, C_p, NCOPh), 123.45 (C_2’), 113.17 (C_4”).

3.5.8. Synthesis of (Z)-2-Phenyl-4-(3’,3′-diphenylallylidene)-5(4H)-oxazolone 1h

1h, yellow solid. β-phenyl-cinnamaldehyde (0.502 g, 2.41 mmol), hippuric acid (0.432 g, 2.41 mmol), sodium acetate (0.202 g, 2.46 mmol) and acetic anhydride (1 mL). Obtained: 0.341 g (40% yield). HRMS (ESI⁺) [m/z]: calculated for [C₂₄H₁₇NO₂Na]⁺ = 374.1157; found 374.1124. ¹H NMR (CDCl₃, 300.13 MHz, 25 °C): δ = 8.14 (m, 2H, H_o, NCOPh, ³J_HH = 7.0 Hz), 7.68 (d, 1H, H_2’, ³J_HH = 12.1 Hz), 7.59 (m, 1H, H_p, NCOPh), 7.53 (m, 2H, H_m, NCOPh), 7.49–7.26 (m, 10H, Ph), 7.08 (d, 1H, H_1’, ³J_HH = 12.1 Hz). ¹³C{¹H} NMR (CDCl₃, 75.5 MHz, 25 °C): δ = 167.02 (CO), 162.08 (CN), 154.92 (C_3’), 141.15, 138.36 (C_i, Ph), 134.74 (C_q, oxazolone, C₄), 133.16 (C_p, NCOPh), 131.09 (C_1’), 130.78, 129.62, 129.06 (2C overlaped), 128.86, 128.67, 128.60 (C_m, NCOPh + C_o, C_m, C_p, 2 Ph), 128.25 (C_o, NCOPh), 125.89 (C_i, NCOPh), 122.52 (C_2’).

3.6. Synthesis and Characterization of Cyclobutane-bis(oxazolone) Intermediates 2a, 2b, 2d and 2f

The synthesis of cyclobutanes 2 was carried out by [2+2]-photocycloaddition of oxazolones 1, photosensitized by [Ru(bpy)₃](BF₄)₂. This procedure is detailed here for oxazolone 1a. All other cyclobutanes were prepared using the same procedure.

3.6.1. Synthesis of Cyclobutane 2a

The oxazolone 1a (0.138 g, 0.502 mmol) and the photocatalyst [Ru(bpy)₃](BF₄)₂ (0.0186 g, 0.025 mmol) were dissolved in deoxygenated CH₂Cl₂ (5 mL) under Ar atmosphere. This red solution was irradiated with the blue light (456 nm) provided by a Kessil LED lamp (50 W) for 18 h. Then the solvent was evaporated to dryness and the residue characterized as a mixture of isomers of cyclobutane 2a (10.1/1 molar ratio): The major isomer in this mixture was separated and purified by flash chromatography on silica gel using hexane/ethyl acetate as eluent (8/2 ratio). The orange band collected was evaporated to dryness to give cyclobutane 2a as an orange-yellowish solid. Obtained: 0.060 g (43% yield). HRMS (ESI⁺) [m/z]: calculated for [C₃₆H₂₆N₂NaO₄]⁺ = 573.1793 [M+Na]⁺; found: 573.1778. ¹H NMR (CDCl₃, 300.13 MHz, 25 °C): δ = 7.97 (m, 2H, H_o, Ph-oxa), 7.82 (m, 2H, H_o, Ph-oxa), 7.50 (m, 2H, H_p, Ph-oxa), 7.39 (m, 2H, H_m, Ph-oxa), 7.23–7.15 (m, 6H, 4H_o + 2H_m, Ph), 7.11–6.98 (m, 6H, 4H_m + 2H_p, Ph), 6.26 (dd, 1H, H_2′-vinyl, ³J_HH = 10 Hz, ⁴J_HH = 1.8 Hz), 6.16 (d, 1H, =CH_1′-oxa, ³J_HH = 12 Hz), 5.69 (dd, 1H, H_3′-vinyl, ³J_HH = 10 Hz, ⁴J_HH = 1.8 Hz), 4.97 (m, 1H, H_2’, cyclo), 3.78–3.69 (m, 2H, H_3’ + H_1’, cyclo). ¹³C{¹H} NMR (CDCl₃, 75.5 MHz, 25 °C): δ = 177.10 (C(O)O-cyclo), 164.64 (C(O)O-vinyl), 162.46, 161.93 (C=N), 141.68, 140.25 (C_i, Ph), 138.69 (C_q, oxa), 137.48 (CH vinyl, C_2’), 138.86 (C_1’H-oxa), 133.44, 133.15 (C_p, Ph-oxa), 128.95, 128.92, 128.63, 128.55, 128.46, 128.21, 128.18 (C_m, C_o, Ph + Ph-oxa), 127.25, 126.96 (C_p, Ph), 125.39, 125.03 (C_i, Ph-oxa), 123.45 (CH vinyl, C_3’), 73.42 (C₄, cycle), 50.45 (CH cyclo, C_3’), 49.03 (CH cyclo, C_1’), 44.06 (CH cyclo, C_2’).

3.6.2. Synthesis of Cyclobutane 2b

The synthesis of the cyclobutane 2b was carried out following the same experimental procedure described for 2a. Therefore, 1b (0.154 g, 0.498 mmol) and the photocatalyst (0.0186 g, 0.025 mmol) were irradiated in CH₂Cl₂ (5 mL) with blue light (456 nm) for 18 h to give a mixture of two cyclobutanes in 3.6/1 molar ratio, which could not be separated by column chromatography. The major isomer in this mixture was characterized spectroscopically as 2b. Obtained: 0.030 g (19% yield). HRMS (ESI⁺) [m/z]: calculated for [C₃₆H₂₄Cl₂N₂NaO₄]⁺ = 641.1013 [M+Na]⁺; found: 641.0977. ¹H NMR (CDCl₃, 300.13 MHz, 25 °C): δ = 7.95 (m, 2H, H_o, Ph-oxa), 7.83 (m, 2H, H_o, Ph-oxa), 7.56–7.51 (m, 4H, H_p, Ph-oxa + H_m, Ph-Cl), 7.44–7.38 (m, 4H, H_m, Ph-oxa), 7.19 (m, 2H, H_o, Ph-Cl), 7.10 (m, 2H, H_m, Ph-Cl), 6.93 (m, 2H, H_o, Ph-Cl), 6.19 (dd, 1H, H_2′- vinyl, ³J_HH = 9.8 Hz, ⁴J_HH = 1.9 Hz), 6.10 (d, 1H, =CH_1′-oxa, ³J_HH = 12 Hz), 5.72 (dd, 1H, H_3′-vinyl, ³J_HH = 9.8 Hz, ⁴J_HH = 2.3 Hz), 4.94 (m, 1H, H_2’, cyclo), 3.64–3.69 (m, 2H, H_3’ + H_1’, cyclo). ¹³C{¹H} NMR (CDCl₃, 75.5 MHz, 25 °C): δ = 177.05 (C(O)O-cyclo), 164.68 (C(O)O-vinyl), 162.84, 162.12 (C=N), 139.89 (C_i, Ph-vinyl), 139.05 (C_q, oxa), 138.58 (C_i, Ph-vinyl), 136.52 (CH vinylic, C_2’), 135.82 (CH-oxa), 133.63, 133.28 (C_p, Ph-oxa), 133.16, 133.00 (C₄, Ph-Cl), 129.90, 129.01, 128.98, 128.96, 128.81, 128.30, 128.24 (C_m, C_o, Ph-Cl + Ph-oxa), 125.25, 124.90 (C_i, Ph-oxa), 124.11 (CH vinylic, C_3’), 73.24 (C₄, cyclo), 49.94 (CH cyclo, C_3’), 48.47 (CH cyclo, C_1’), 43.74 (CH cyclo, C_2’).

3.6.3. Synthesis of Cyclobutane 2d

The synthesis of the cyclobutane 2d was carried out following the same experimental procedure described for 2a. Therefore, 1d (0.154 g, 0.533 mmol) and the photocatalyst (0.0186 mg, 0.025 mmol) were irradiated in CH₂Cl₂ (5 mL) with blue light (456 nm) for 17 h to give a mixture of two cyclobutanes in 2.9/1 molar ratio, which could not be separated by column chromatography. The major isomer in this mixture was characterized spectroscopically as 2d. Obtained: 0.035 g (22% yield). HRMS (ESI⁺) [m/z]: calculated for [C₃₈H₃₀N₂NaO₄]⁺ = 601.2106 [M+Na]⁺; found: 601.2081. ¹H NMR (CDCl₃, 300.13 MHz, 25 °C): δ = 8.00 (m, 2H, H_o, Ph-oxa), 7.84 (m, 2H, H_o, Ph-oxa), 7.62–7.04 (overlapped aromatics, 16H), 5.40 (s, 1H, H_3′-vinyl), 4.06 (m, 1H, H_3’, cyclo), 3.80 (m, 1H, H_1’, cyclo), 1.62 (s, 3H, CH₃-C_2’ vinyl), 1.30 (s, 3H, CH₃-C_2’ cyclo). ¹³C{¹H} NMR (CDCl₃, 75.5 MHz, 25 °C): δ = 177.68 (C(O)O-cyclo), 167.55 (C(O)O-vinyl), 162.68, 161.73 (C=N), 143.87 (C vinylic, C₁), 141.51 (C_i, Ph-vinyl), 138.38 (C_i, Ph-vinyl), 138.20 (C_q, oxa), 137.20 (CH-oxa), 133.29, 133.04, 127.03, 126.64 (C_p, Ph-oxa + Ph), 128.88 (br), 128.84 (br), 128.40 (br), 128.32, 127.96 (C_m, C_o, Ph + C_m, Ph-oxa), 125.58, 125.69 (C_i, Ph-oxa), 119.83 (CH vinyl, C_2’), 77.40 (C₄, cyclo), 54.39 (CH cyclo, C_3’), 51.50 (CH cyclo, C_1’), 48.51 (C cyclo, C_2’), 22.30 (CH₃-C vinyl), 19.92 (CH₃-C cyclo).

3.6.4. Synthesis of Cyclobutane 2f

The synthesis of the cyclobutane 2f was carried out following the same experimental procedure described for 2a. Therefore, the oxazolone 1f (0.145 g, 0.482 mmol) and the photocatalyst (0.0186 g, 0.025 mmol) were irradiated in CH₂Cl₂ (5 mL) with blue light (456 nm) for 18 h to give cyclobutane 2f as a yellow solid after separation and purification by column chromatography (silica gel, hexane/ethyl acetate = 8/2 as eluent). Obtained: 0.026 g (18% yield). HRMS (ESI⁺) [m/z]: calculated for [C₃₈H₂₄N₄NaO₄]⁺ = 623.1698 [M+Na]⁺; found: 623.1656. ¹H NMR (CDCl₃, 300.13 MHz, 25 °C): δ = 8.16 (m, 2H, H_o, Ph-CN), 8.03 (m, 2H, H_o, Ph-CN), 7.83 (m, 2H, H_m, Ph-CN), 7.65 (m, 2H, H_m, Ph-CN), 7.42–7.31 (m, 4H, H_o, Ph), 7.21–7.18 (m, 2H, H_m, Ph), 7.04–6.97 (m, 4H, H_p, Ph + H_m, Ph), 6.33 (dd, 1H, H_2′-vinyl, ³J_HH = 10 Hz, ⁴J_HH = 1.7 Hz), 6.09 (d, 1H, =CH_1′-oxa, ³J_HH = 11 Hz), 5.68 (dd, 1H, H_3′-vinyl, ³J_HH = 10 Hz, ⁴J_HH = 1.7 Hz), 4.52 (m, 1H, H_2’, cyclo), 3.79–3.70 (m, 2H, H_3’ + H_1’, cyclo). ¹³C{¹H} NMR (CDCl₃, 75.5 MHz, 25 °C): δ = 175.95 (C(O)O-cyclo), 163.82 (C(O)O-vinyl), 162.38, 160.59 (C=N), 141.15 (C_i, Ph-vinyl), 140.34 (C_q, oxa), 139.32 (C_i, Ph-cyclo), 138.20 (CH vinyl, C_2’), 134.27 (CH-oxa), 132.80, 132.67 (C_o, Ph), 129.18, 129.15 (CN-Ph), 128.83, 128.68, 128.58, 128.54 (br), 128.42 (C_m, C_o, Ph-CN + C_m, Ph), 127.36, 127.19 (C_p, Ph), 122.50 (CH vinylic, C_3’), 117.80, 117.68 (C_i, Ph), 117.00, 116.71 (C₄-CN, Ph-CN) 73.37 (C₄, cyclo), 50.49 (CH cyclo, C_3’), 48.40 (CH cyclo, C_1’), 46.39 (CH cyclo, C_2’).

3.7. Synthesis and Characterization of Cyclobutanes 3a and 3b

3.7.1. Synthesis of Methyl-1-benzamido-2-((E)-2-benzamido-2-methoxycarbonylprop-1-en-1-yl)-3-phenyl-4-((E)-styryl)cyclobutane-1-carboxylate 3a

Oxazolone 1a (0.2982 g, 1.084 mmol) and [Ru(bpy)₃](BF₄)₂ (0.0391 g, 0.053 mmol) (5% mol) were dissolved in deoxygenated CH₂Cl₂ under Ar atmosphere and irradiated for 18 h (456 nm, Kessil lamp, 50 W). Then the solution was evaporated to dryness (irradiation was maintained during evaporation to prevent thermal retro-[2+2]) and the oily residue treated with methanol (10 mL) and NaOMe (9 mg, 0.167 mmol). The resulting suspension was heated in an oil bath to the reflux temperature for 45 min. The resulting solution was cooled and evaporated to dryness. The residue was taken in the minimal amount of CHCl₃ (2 mL) and subjected to flash chromatography on silica gel using CHCl₃ as eluent. The colorless fraction containing 3a was evaporated to dryness and treated with n-pentane (15 mL), giving 3a as a white solid. The white solid was crystallized in CH₂Cl₂/n-pentane at −18 °C to give 3a as a white crystalline solid. Obtained: 0.2401 g (80% yield). Mp: 145–146 °C. HRMS (ESI⁺) [m/z]: calculated for [C₃₈H₃₄N₂NaO₆]⁺ = 637.2309 [M+Na]⁺; found: 637.2322. ¹H NMR (CDCl₃, 500.13 MHz, 25 °C): δ = 8.25 (s, 1H, NH-C₄ cyclo), 7.92 (m, 2H, H_o, C₆H₅-CONHC₄ cyclo), 7.72 (s, 1H, NH-C₄=C_1’ (vinyl)), 7.66 (m, 2H, H_o, C₆H₅-CONHC_q (vinyl)), 7.56–7.33 (m, 6H, 2H_p+ 4H_m, COC₆H₅), 7.23–7.05 (m, 6H, 2H_p+ 4H_m, C₆H₅), 6.99 (m, 2H, H_o, =C(H)-C₆H₅), 6.92 (m, 2H, H_o, C_3′-C₆H₅), 6.50 (d, 1H, =C_1’H, ³J_HH = 10.8 Hz), 6.22 (dd, 1H, C₆H₅-CH=CH, ³J_HH = 10.2 Hz, ⁴J_HH = 2.7 Hz), 6.10 (dd, 1H, C₆H₅-CH=CH, ³J_HH = 10.2 Hz, ⁴J_HH = 2.1 Hz), 4.68 (t, 1H, H-C_2’ cyclo, ³J_HH4 = ³J_HH3 = 12 Hz), 4.01 (s, 3H, C₁-C(O)OCH₃), 3.76–3.70 (m, 4H, =C-C(O)OCH₃ + H-C_1’ cyclo), 3.30 (dd, 1H, H-C_3’ cyclo, ³J_HH = 10.20 Hz, ³J_HH = 12 Hz). ¹³C{¹H} NMR (CDCl₃, 125.75 MHz, 25 °C): δ = 171.51 (C(O)OCH₃), 166.19 (CONH), 165.33 (2 C(O), CONH + C(O)OCH₃), 142.96 (C_q, C₆H₅), 141.25 (C_q, C₆H₅), 134.32 (C_q, C₆H₅), 134.05 (C_q, C₆H₅), 133.52 (CH, C₆H₅), 131.86 (CH, C₆H₅), 131.52 (-CH=CH-), 128.59 (CH, C₆H₅), 128.51 (CH, C₆H₅), 128.38 (2CH, C_o, C₆H₅), 128.25 (C_q=CH-), 128.13 (CH, C₆H₅), 128.08 (CH, C₆H₅), 127.86 (C_q=CH), 127.18 (C_o, C₆H₅), 126.91 (C_o, C₆H₅), 126.71 (-CH=CH-), 126.60 (CH, C₆H₅), 126.43 (CH, C₆H₅), 64.20 (C₄), 53.30 (CH₃, C(O)OCH₃), 52.81 (CH₃, C(O)OCH₃), 51.96 (CH cyclo, C_3’), 50.33 (CH cyclo, C_1’), 44.23 (CH cyclo, C_2’).

3.7.2. Synthesis of Methyl-1-benzamido-2-((E)-2-benzamido-2-methoxycarbonylprop-1-en-1-yl)-3-(4-chlorophenyl)-4-((E)-chlorostyryl)cyclobutane-1-carboxylate 3b

Compound 3b was obtained following the experimental procedure described for 3a. Therefore, oxazolone 1b (0.2991 mg, 0.968 mmol) and [Ru(bpy)₃](BF₄)₂ (0.036 mg, 0.048 mmol) (5% mol ratio) were irradiated for 18 h in CH₂Cl₂ (5 mL) and then reacted with NaOMe in refluxing MeOH (5 mL) for 45 min to give 3b as a white solid after chromatographic purification and crystallization in CH₂Cl₂/n-pentane. Obtained: 230.0 mg (78% yield). Mp: 151–152 °C. HRMS (ESI⁺) [m/z]: calculated for [C₃₈H₃₂Cl₂N₂NaO₆]⁺ = 705.1535 [M+Na]⁺; found: 705.1524. ¹H NMR (CDCl₃, 500.13 MHz, 25 °C): δ = 8.09 (s, 1H, NH, NH-C₄ cyclo), 7.86 (m, 2H, H_o, C₆H₅-CONHC₄ cyclo), 7.81 (s, 1H, NH, NH-C₄=C_1’ (vinyl)), 7.67 (m, 2H, H_o, C₆H₅-CONHC₄), 7.52–7.38 (m, 6H, 2H_p+ 4H_m, COC₆H₅), 7.16 (dm, 2H, H_m, C₆H₄Cl, ³J_HH = 8.4 Hz), 7.09 (dm, 2H, H_m, C₆H₄Cl, ³J_HH = 8.1 Hz), 6.90 (dm, 2H, H_o, C₆H₄Cl, ³J_HH = 8.4 Hz), 6.85 (dm, 2H, H_o, C₆H₄Cl, ³J_HH = 8.1 Hz), 6.64 (d, 1H, =C_1’H oxazolone, ³J_HH= 11.9 Hz), 6.13 (dd, 1H, CH=CH-C₄, ³J_HH= 10.1 Hz, ⁴J_HH= 2.7 Hz), 6.10 (dd, 1H, -CH=CH-C₄, ³J_HH= 10.1 Hz, ⁴J_HH= 2.1 Hz), 4.75 (t, 1H, H-C_2’ cyclo, ³J_HH = ³J_HH = 11.9 Hz), 4.00 (s, 3H, C₁-C(O)OCH₃), 3.69–3.64 (m, 4H, =C-C(O)OCH₃ + H-C_1’ cyclo), 3.27 (dd, 1H, H-C_3’ cyclo, ³J_HH =10.4 Hz, ³J_HH = 11.9 Hz). ¹³C{¹H} NMR (CDCl₃, 125.7 MHz, 25 °C): δ = 171.53 (C(O), C(O)OCH₃), 166.23 (C(O), CONH), 165.57 (C(O), CONH), 165.16 (C(O)OCH₃), 141.42 (C_i, C₆H₄Cl), 139.83 (C_i, C₆H₄Cl), 134.38, 134.30 (C_i, C₆H₅ + C_i, C₆H₅), 133.17 (-CH=CH-), 132.57, 132.54 (C_p, C₆H₄Cl + C_p, C₆H₄Cl), 132.13 (C_p, C₆H₅), 131.83 (C_p, C₆H₅), 129.84 (C_o, C₆H₄Cl), 129.79 (C_o, C₆H₄Cl), 128.83, 128.74 (C_m, C₆H₅ + C_m, C₆H₅), 128.63 (C_m, C₆H₄Cl), 128.59 (C_m, C₆H₄Cl), 128.40 (C_q=CH), 127.27 (2C, C_o, C₆H₅ + C_q=CH), 127.17 (-CH=CH-), 127.06 (C_o, C₆H₅), 64.39 (C₄, cyclo), 53.69 (CH₃,C(O)OCH₃), 53.09 (CH₃, C(O)OCH₃), 51.71 (CH cyclo, C_3’), 49.78 (CH cyclo, C_1’), 43.88 (CH cyclo, C_2’).

4. Conclusions

The Ru-sensitized irradiation of 4-(3-aryl-allyliden)-5(4H)-oxazolones with blue light (456 nm) takes place with the formation of a series of styryl-cyclobutanes by [2+2]-photocycloaddition of two oxazolones. The coupling is produced using the exocyclic C=C bond of one oxazolone and the styryl C=C bond of the other one and occurs with a high degree of selectivity, because one isomer is mostly formed. This coupling is very different from that observed in 4-aryliden-5(4H)-oxazolones, which involves only the exocyclic C=C bond. The reactive species was characterized as a triplet excited state T₁ by transient absorption spectroscopy. In addition, DFT methods showed that this T₁ state is planar and that the spin distribution is concentrated in the α and δ carbons of the diene system, instead of in the α and β carbons as observed in aryliden-oxazolones, providing a reasonable explanation for the different reactivity observed.