Lawsone Derivatives as Efficient Photopolymerizable Initiators for Free-Radical, Cationic Photopolymerizations, and Thiol—Ene Reactions

Two new photopolymerizable vinyl (2-(allyloxy) 1,4-naphthoquinone, HNQA) and epoxy (2-(oxiran-2yl methoxy) 1,4-naphthoquinone, HNQE) photoinitiators derived from lawsone were designed in this paper. These new photoinitiators can be used as one-component photoinitiating systems for the free-radical photopolymerization of acrylate bio-based monomer without the addition of any co-initiators. As highlighted by the electron paramagnetic resonance (EPR) spin-trapping results, the formation of carbon-centered radicals from an intermolecular H abstraction reaction was evidenced and can act as initiating species. Interestingly, the introduction of iodonium salt (Iod) used as a co-initiator has led to (1) the cationic photopolymerization of epoxy monomer with high final conversions and (2) an increase of the rates of free-radical polymerization of the acrylate bio-based monomer; we also demonstrated the concomitant thiol–ene reaction and cationic photopolymerizations of a limonene 1,2 epoxide/thiol blend mixture with the HNQA/Iod photoinitiating system.


Introduction
In the past few years, photopolymerization gained an increasing interest and became the cornerstone of many investigations related to microelectronics, adhesives inks, dentistry, and the elaboration of (3D) biomaterials [1]. The success of photochemistry is undoubtedly based on undeniable advantages [2] over the thermal process, e.g., polymerizations under low-temperature conditions without the addition of hazardous solvent, reduced reaction times, the spatial control of the light penetration inside the photosensitive resins, and the capability to adjust overall properties of the photoinduced materials. However, the use of UV-light irradiation still remains an issue of concern in photopolymerization, and alternatives must be found.
In light of these considerations, huge efforts have been made to propose efficient visible-light photoinitiating systems [3][4][5]. These systems can be classified into two distinct categories according to how radicals are generated. Norrish I and II type photoinitiators, acting in the visible range, are well-known to perform free-radical and cationic photopolymerizations. Even if type I photoinitiators, which can cleave under light activation, are extensively used in industries, the irreversible consumption of these kinds of photoinitiators during polymerization is clearly the main drawback. Therefore, scientists have turned their efforts towards developing Norrish type II photoinitiators. In particular, relevant reviews [3,4,6] have described the recent progress photoinitiating strategies envisaged to perform polymerization under light irradiation. Notably, new photocatalytic systems the proposed mechanisms were fully supported by electron paramagnetic resonance (EPR) spin-trapping.

Experimental Section Materials
Metachloroperbenzoic acid (m-CPBA), tetraethylammonium bromide (TEAB), 2-hydroxy-1,4-naphthoquinone (lawsone, HNQ), allyl bromide, potassium carbonate (K 2 CO 3 ), soybean oil acrylate (SOA), 3,4-epoxycyclohexane)methyl-3,4-epoxycyclohexylcarboxylate (EPOX), limonene 1,2-epoxide (>97%), trimethylolpropane tris(3-mercaptopropionate) (trithiol, >95%), 5,5-dimethyl-pyrroline N-oxide (DMPO), and bis(4-methylphenyl) iodonium hexafluorophosphate (Iod) were purchased from Sigma-Aldrich. Cyclohexane, ethyl acetate, and silica gel were obtained from commercial suppliers. Table 1 displays the chemical structures of the main compounds used in this investigation. Table 1. Chemical structures of the monomers and photoinitiating systems used in this study.  and transferred to a 30 mL capacity microwave tube, where 6 mL of allyl bromide was added. Reaction was carried out in a microwave reactor under pressure at 75 • C for 1 h. The mixture obtained after reaction in the microwave reactor was transferred to a separatory funnel containing 100 mL of water. The product of interest is extracted three times with 50 mL of diethyl ether. The organic phases are washed with water and dried with magnesium sulfate (MgSO 4 ). The solution was filtered into a flask and evaporated. The crude product was purified by column chromatography (silica gel) with the eluent cyclohexane/ethyl acetate in 80/20 proportion. The separation of the compounds was followed by thin-layer chromatography (TLC), and the tubes containing the product of interest were collected in a tared flask, and the contents evaporated. A 1 H NMR and a 13 C NMR spectrum were performed to check the structure of the product obtained. The vinylic photoinitiator HNQA is a yellow powder. Yield = 48%. 1 Figure S2).
Synthesis of the epoxide derivative (HNQE). Lawsone vinylic derivative HNQA (0.3 g-1.4 mmol-1 equivalent) was dissolved in degassed dichloromethane. m-CPBA (2.408 g-14 mmol-10 equivalents) was added four times over 10 h. Each addition must be made after cooling the solution to 0 • C. The solution was then slowly brought to room temperature. After the complete addition of m-CPBA, the solution was evaporated. The epoxidation was monitored by thin-layer chromatography (TLC) by regular sampling to observe the disappearance of the HNQA stain. The crude product was purified by column chromatography (silica gel) with cyclohexane/ethyl acetate as eluent in a 60/40 ratio. The epoxidized photoinitiator HNQE is a yellow powder. Yield = 10%. Photopolymerization experiments. The photopolymerization experiments were carried out using real-time Fourier transform infrared spectroscopy (RT-FTIR). The kinetics of photopolymerization were carried out with a time-resolved Fourier transform infrared spectrometer (JASCO 4100 series) to evaluate the absorbance decrease of acrylate, epoxy, vinyl, and thiol functional groups at 1636 cm −1 , 842 cm −1 , 1645 cm −1 , and 2560 cm −1 , respectively. Formulations were deposited on a BaF 2 pellet and were irradiated in laminate (with the addition of a transparent polypropylene film on the top of the photosensitive layer) or under air upon LED@385 and @405nm irradiation at room temperature.
Fluorescence Experiments. Fluorescence experiments were performed and collected at room temperature with a FluoroMax+ spectrofluorometer from HORIBA. For fluorescence measurements, ACN solution of HNQ, HNQA, or HNQE were excited at λ exc = 330 nm. The fluorescence quenching of the lawsone derivatives was evaluated after the addition of a gradual concentration of Iod.
Cyclic Voltammetry. Cyclic voltammetry measurements of HNQ, HNQA, and HNQE were performed at room temperature under argon atmosphere according to a previously described method 18 . Glassy carbon, saturated calomel, and gold wire electrodes were used as the working electrode, reference electrode, and counter electrode, respectively. For cyclic voltammetry measurements, the three electrodes were immersed in an ACN solution of HNQ, HNQA, or HNQE containing a supporting electrolyte solution of tetraethylammonium tetrafluoroborate at 10 −3 mol. L −1 . The free energy change ∆G eT for an electron transfer reaction between HNQA, or HNQE and Iod can be evaluated from the classical Rehm−Weller equation [29].
Electron paramagnetic resonance (EPR) spin-trapping. The X-band cw-EPR spectra (modulation frequency of 100 kHz) were monitored with the EMXplus spectrometer (Bruker) equipped with the High Sensitivity Probe-Head (Bruker) in a small quartz flat cell (Wilmad-LabGlass, WG 808-Q). The solutions prepared in acetonitrile and saturated with argon were irradiated at 295 K directly in the EPR resonator utilizing a LED@400 nm source (λ max = 400 nm; Bluepoint LED, Hönle UV Technology). The EPR spectra were recorded in situ during/after a defined exposure as described previously in Reference [18]. The g-factors were determined with an uncertainty of ± 0.0001 exploiting a nuclear magnetic resonance teslameter (ER 036TM, Bruker) and integrated frequency counter. The experimental EPR spectra were analyzed by the WinEPR software (Bruker), and the calculations of simulated spectra were performed with the EasySpin toolbox working on the MatLab ® platform [30]. The standard EPR spectrometer settings were as follows: microwave frequency,~9.43 GHz; microwave power,~10.5 mW; center field,~336.0 mT; sweep width, 2-8 mT; gain, 2.00 × 10 5 -5.02 × 10 4 ; modulation amplitude, 0.025 mT-0.10 mT; sweep time, 45 s; time constant, 10.24 ms or 20.48 ms; the number of scans, 5 or 10. Sample compositions comprising the reagents are discussed in the text.
Computational method. Standard MP2 method [31,32] and cc-pVTZ basis sets for all atoms [33] were used for geometry optimizations of the anionic compounds under study in acetonitrile in doublet ground spin states. Solvent effects were approximated by the Solvation Model based on the Density [34] (SMD) continuum model. The stability of the optimized structures was confirmed by vibrational analysis (no imaginary vibrations). Subsequently, their isotropic hyperfine coupling constants in acetonitrile were calculated by DFT method using standard B3LYP hybrid functional [35], EPR-III basis sets for all atoms [36], and SMD solvent model [34]. All calculations were performed using the Gaussian16 [37] program package.

Results and Discussion
Synthesis of the lawsone derivatives. Allylated (HNQA) and epoxidized (HNQE) lawsone were synthesized from lawsone in a one or two-step procedure, following the procedure described in Scheme 1. The synthesized molecules have been fully characterized by NMR spectroscopy (see supporting information). Allylation of lawsone (HNQ) is expected according to a previously reported synthesis [38] using allyl bromide (co-reactant) with NaOH/K 2 CO 3 as a base catalyst. Prior to the addition of the allyl bromide, a deprotonation of the alcohol group of HNQ occurs and leads to an alcoholate formation, which ensues a nucleophilic SN2 type reaction with allyl bromide. The purification of the crude on silica gel chromatography led to the expected allylated lawsone with a yield of 48%. Interestingly, new proton signals attributed to the vinyl function [38,39] of HNQA appear accordingly at 6.1 ppm (H 2 C=CH-) and at 5.5 ppm (H 2 C=CH-). The oxidation of HNQA was performed in degassed dichloromethane at room temperature with m-CPBA to convert the allyl bonds into epoxides [39]. After purification of the crude, HNQE was obtained with a 10% yield. A complete disappearance of the allyl protons of HNQA is observed, leading to the apparition of new signals, which are attributed to the protons of the HNQE epoxy group between 2.8 and 3.5 ppm [39]. Polymers 2021, 13, x FOR PEER REVIEW 6 of 20 Scheme 1. Synthesis of the lawsone derivatives.
Steady-state photolysis and fluorescence. The UV-vis absorption spectra and the molar extinction coefficient of the lawsone derivatives are displayed in Figure S5. The photophysical properties of the lawsone derivatives are gathered in Table 2. Lawsone derivatives exhibit high molar absorption coefficients in comparison with common visiblelight photoinitiators [40] (e.g., camphorquinone or benzophenone) and display a low absorption band in the visible range. This makes lawsone derivatives probable visible photoinitiators in the UV-vis range. The overall photoreactivity of the lawsone derivatives with oxidative compounds was first studied by steady-state photolysis and fluorescence experiments.
Photolysis of HNQ, HNQA, and HNQE upon LED@385 nm exposure, without and with Iod, are given in Figure 1A−F. Under light irradiation, the absorbance of HNQ, HNQA, and HNQE alone ( Figure 1A,B,C) decreases, and an isosbestic point is clearly observed for each photoinitiating system, suggesting the formation of a new species. Radical species resulting from the irradiation of the naphthoquinone derivatives are likely to appear as EPR spin-trapping could further demonstrate. New carbon-centered radicals are expected to be formed through an H-abstraction reaction as previously described with thioxanthone derivatives [41]. Surprisingly, the absorbance decrease of the lawsone derivatives was not accelerated when adding Iod ( Figure 1D Figure S8 for Eox(HNQA) and Figure S9 for Eox(HNQE)) demonstrate that a thermodynamically favorable electron transfer reaction between the excited singlet states of HNQA or HNQE and Iod (Table 2) may occur. Steady-state photolysis and fluorescence. The UV-vis absorption spectra and the molar extinction coefficient of the lawsone derivatives are displayed in Figure S5. The photophysical properties of the lawsone derivatives are gathered in Table 2. Lawsone derivatives exhibit high molar absorption coefficients in comparison with common visible-light photoinitiators [40] (e.g., camphorquinone or benzophenone) and display a low absorption band in the visible range. This makes lawsone derivatives probable visible photoinitiators in the UV-vis range. The overall photoreactivity of the lawsone derivatives with oxidative compounds was first studied by steady-state photolysis and fluorescence experiments. Photolysis of HNQ, HNQA, and HNQE upon LED@385 nm exposure, without and with Iod, are given in Figure 1A−F. Under light irradiation, the absorbance of HNQ, HNQA, and HNQE alone ( Figure 1A,B,C) decreases, and an isosbestic point is clearly observed for each photoinitiating system, suggesting the formation of a new species. Radical species resulting from the irradiation of the naphthoquinone derivatives are likely to appear as EPR spintrapping could further demonstrate. New carbon-centered radicals are expected to be formed through an H-abstraction reaction as previously described with thioxanthone derivatives [41]. Surprisingly, the absorbance decrease of the lawsone derivatives was not accelerated when adding Iod ( Figure 1D-F). Quenching of fluorescence was performed to learn about the reactivity of HNQA, and HNQE excited singlet states toward the addition of Iod. Fluorescence spectra of HNQA and HNQE were recorded in ACN (see Figures S6 and S7, respectively), and the maximum emission wavelength was observed at 410 and 411 nm, respectively. A quenching rate of fluorescence was observed for HNQA and HNQE in the presence of Iod ( Figures S6 and S7), with K SV Iod (HNQA) = 4215 M −1 and K SV Iod (HNQE) = 110 M −1 . The negative ∆G eT (S) values calculated from the Rehm-Weller equation 29 and the potentials of oxidation (see Figure S8 for E ox (HNQA) and Figure S9 for E ox (HNQE)) demonstrate that a thermodynamically favorable electron transfer reaction between the excited singlet states of HNQA or HNQE and Iod (Table 2) may occur. EPR experiments. The photoexcitation of HNQ in the deoxygenated acetonitrile solution under given experimental conditions resulted in the generation of EPR signal (g = 2.0044) shown in Figure 2A, assigned to the corresponding radical anion [42] (HNQ -). Due to its low intensity, it was not possible to acquire EPR spectra with the lower modulation amplitude, which prevented us from obtaining the better resolved hyperfine structure. Consequently, the simulation interpretation of the experimental spectrum of HNQ represents the hyperfine interaction of four dominant hydrogen nuclei with the unpaired EPR experiments. The photoexcitation of HNQ in the deoxygenated acetonitrile solution under given experimental conditions resulted in the generation of EPR signal (g = 2.0044) shown in Figure 2A, assigned to the corresponding radical anion [42] (HNQ •− ). Due to its low intensity, it was not possible to acquire EPR spectra with the lower modulation amplitude, which prevented us from obtaining the better resolved hyperfine structure. Consequently, the simulation interpretation of the experimental spectrum of HNQ •− represents the hyperfine interaction of four dominant hydrogen nuclei with the unpaired electron. As the EPR spectra of HNQ •− found in the literature are significantly affected by the experimental conditions [43], the assignment of hyperfine couplings of individual hydrogen atoms is based on the theoretical data evaluated by the B3LYP/EPR-III/SMD (acetonitrile) calculations (Table 3). HNQ •− contains three oxygen sites where, in principle, the hydroxyl H atom can be located. MP2/cc-pVTZ/SMD (acetonitrile) geometry optimizations of the corresponding structures confirmed the highest stability of the 2-hydroxy isomer, whereas the concentrations of the 1-and 4-hydroxy isomers are vanishing because their Gibbs free energies at 298 K are higher by 109.1 and 33.1 kJ/mol, respectively. While a low-intensity EPR spectrum was obtained during irradiation of HNQA in acetonitrile under given experimental conditions (g = 2.0047; Figure 2B), the photoactivation of HNQE in acetonitrile upon LED@400 nm is analogous, but a better-resolved EPR spectrum is achieved (g = 2.0047; Figure 2C); the signals are attributed to the corresponding radical anions of HNQA and HNQE, respectively (Table 3). Table 3. The hyperfine coupling constants of 2-substituted naphthoquinone radical anions calculated at B3LYP/EPR-III/SMD (acetonitrile) level of theory in comparison with those elucidated from the simulations of experimental EPR spectra (in italic).
Polymers 2021, 13, x FOR PEER REVIEW electron. As the EPR spectra of HNQ found in the literature are signif the experimental conditions [43], the assignment of hyperfine coupling drogen atoms is based on the theoretical data evaluated by the B3LYP/ tonitrile) calculations (Table 3). HNQ contains three oxygen sites wher hydroxyl H atom can be located. MP2/cc-pVTZ/SMD (acetonitrile) ge tions of the corresponding structures confirmed the highest stability of mer, whereas the concentrations of the 1-and 4-hydroxy isomers are their Gibbs free energies at 298 K are higher by 109.1 and 33.1 kJ/mol, r a low-intensity EPR spectrum was obtained during irradiation of HN under given experimental conditions (g = 2.0047; Figure 2B), the photoac in acetonitrile upon LED@400 nm is analogous, but a better-resolved achieved (g = 2.0047; Figure 2C); the signals are attributed to the corr anions of HNQA and HNQE, respectively (Table 3). The continuous in situ LED@400 nm irradiation of HNQ/aceton presence of DMPO spin trap resulted in the immediate generation of s ure 2D) with the spin-Hamiltonian parameters characteristic for the D a ketyl radical [44] (  C(OH)R1R2, Table 4). On the other hand, the irradia ated acetonitrile solutions of HNQA and HNQE containing DMPO le of  DMPO-CR adduct ( Figure 2E,F), which correspond to the carbon-c α-position of the vinyl or epoxy group of HNQA or HNQE, respectively an intermolecular H-abstraction reaction.

a (mT) H(2) H(3) H(5) H(6) H(7) H(8)
HNQ The continuous in situ LED@400 nm irradiation of HNQ/acetonitrile/argon in the presence of DMPO spin trap resulted in the immediate generation of six-line signal ( Figure 2D) with the spin-Hamiltonian parameters characteristic for the DMPO-adduct with a ketyl radical [44] ( • C(OH)R 1 R 2 , Table 4). On the other hand, the irradiation of deoxygenated acetonitrile solutions of HNQA and HNQE containing DMPO led to the formation of • DMPO-CR adduct ( Figure 2E,F), which correspond to the carbon-centered radical in α-position of the vinyl or epoxy group of HNQA or HNQE, respectively, and formed after an intermolecular H-abstraction reaction. Magnetic field (mT) The effective generation of DMPO-adducts with ketyl and phenyl(4-methyl) radicals is well evidenced upon irradiation of HNQ/Iod/DMPO/acetonitrile/argon solutions, also reflecting the efficient electron transfer from the photoexcited HNQ to Iod ( Figure 3A). The electron transfer reaction from the photoactivated HNQA and HNQE molecules to Iod producing phenyl(4-methyl) radicals upon LED@400 nm exposure is demonstrated in Figure 3B,C: a well-resolved EPR signal of the DMPO-phenyl(4-methyl) adduct [45] is only observed. The EPR spectra measured upon in situ LED@400 nm irradiation of 2-substituted naphthoquinone derivatives in deoxygenated acetonitrile solutions containing DMPO/Iod/TT are shown in Figure 3D-F. The EPR spectrum obtained under given experimental conditions for HNQ represents a superposition of three individual DMPOadducts: (i) • DMPO-C(OH)R 1 R 2 ; (ii) • DMPO-phenyl(4-methyl); and (iii) • DMPO-SR; the spin-Hamiltonian parameters are gathered in Table 4. Surprisingly, EPR spectra measured for the photoactivated HNQA and HNQE under identical experimental conditions (DMPO/Iod/TT/acetonitrile/argon) demonstrated the presence of two latter signals only ( • DMPO-phenyl(4-methyl) 17 and • DMPO-SR 17 ).
According to the EPR results, various active radical species are produced under light irradiation of the lawsone derivatives photoinitiating systems, i.e., HNQA (or HNQE)/Iod and HNQA (or HNQE)/Iod/trithiol, and a proposed mechanism is described in Scheme 2 with HNQA (Equations (a)-(c)). The effective generation of DMPO-adducts with ketyl and phenyl(4-methyl) radicals is well evidenced upon irradiation of HNQ/Iod/DMPO/acetonitrile/argon solutions, also reflecting the efficient electron transfer from the photoexcited HNQ to Iod ( Figure 3A). The electron transfer reaction from the photoactivated HNQA and HNQE molecules to Iod producing phenyl(4-methyl) radicals upon LED@400 nm exposure is demonstrated in Figure 3B,C: a well-resolved EPR signal of the DMPO-phenyl(4-methyl) adduct [45] is only observed. The EPR spectra measured upon in situ LED@400 nm irradiation of 2-substituted naphthoquinone derivatives in deoxygenated acetonitrile solutions containing DMPO/Iod/TT are shown in Figure 3D-F. The EPR spectrum obtained under given experimental conditions for HNQ represents a superposition of three individual DMPO-adducts: (i)  DMPO-C(OH)R1R2; (ii)  DMPO-phenyl(4-methyl); and (iii)  DMPO-SR; the spin-Hamiltonian parameters are gathered in Table 4. Surprisingly, EPR spectra measured for the photoactivated HNQA and HNQE under identical experimental conditions (DMPO/Iod/TT/acetonitrile/argon) demonstrated the presence of two latter signals only (  DMPO-phenyl(4-methyl) 17 and  DMPO-SR 17 ).
According to the EPR results, various active radical species are produced under light irradiation of the lawsone derivatives photoinitiating systems, i.e., HNQA (or HNQE)/Iod and HNQA (or HNQE)/Iod/trithiol, and a proposed mechanism is described in Scheme 2 with HNQA (Equations (1)- (3)). Free-radical polymerization of SOA. HNQ, HNQA, and HNQE have been first used as a one-component photoinitiating system without the addition of any co-initiators ( Figure 4A,B). The free-radical polymerizations of SOA were followed in laminate by RT-FTIR under LEDs@405 nm and 385 nm irradiation. The final acrylate conversions of SOA are displayed in Table 5. Interestingly, HNQA and HNQE alone can initiate the FRP of SOA with interesting acrylate conversions. These results confirm the previous EPR results and highlight the capability of the resulting carbon-centered radicals ( Figure 2E,F) to initiate FRP of acrylate monomer. This is particularly interesting in terms of the green chemistry approach as only one component photoinitiating system is used, reducing the carbon atoms consumption. Not surprisingly, when HNQA (or HNQE) is combined with Iod, the final acrylate conversions and the rates of polymerization (Rp/[M0]) strongly increase whatever the LEDs used, as described in Table 5 and Figure 4C,D. This confirms the formation of phenyl-derived radicals as highlighted by EPR spin-trapping ( Figure  3B,C); such radicals can also initiate the FRP of SOA. Unfortunately, HNQ and HNQ/Iod systems cannot initiate FRP as 1) the anion radical HNQ •− is not an effective initiation of radical species, and 2) even if phenyl-derived radicals are generated from the photolysis of HNQ/Iod ( Figure 3A), ketyl radicals, also produced by the photolysis of HNQ/Iod, can act as a terminating agent of the polymerization reaction [46]. Magnetic field (mT) oms consumption. Not surprisingly, when HNQA (or HNQE) is combined with Iod, the final acrylate conversions and the rates of polymerization (Rp/[M0]) strongly increase whatever the LEDs used, as described in Table 5 and Figure 4C,D. This confirms the formation of phenyl-derived radicals as highlighted by EPR spin-trapping ( Figure 3B,C); such radicals can also initiate the FRP of SOA. Unfortunately, HNQ and HNQ/Iod systems cannot initiate FRP as 1) the anion radical HNQ is not an effective initiation of radical species, and 2) even if phenyl-derived radicals are generated from the photolysis of HNQ/Iod ( Figure 3A), ketyl radicals, also produced by the photolysis of HNQ/Iod, can act as a terminating agent of the polymerization reaction [46].  Cationic photopolymerization of EPOX. The kinetics of EPOX polymerization (epoxy conversion vs. irradiation time) in the presence of HNQA (or HNQE)/Iod photoinitiating systems under different LEDs irradiation under air are described in Figure 5. HNQ/Iod has been used as a reference photoinitiating system for comparison. When Iod is used alone, no photopolymerization of EPOX occurs under LED irradiation (385 and 405 nm) because Iod shows a maximum absorption band around 250 nm. Cationic photopolymerizations of EPOX successfully take place ( Figure 5) when using both photoinitiating systems (HNQA (or HNQE)/Iod) under LEDs@385 and 405 nm exposure. Accordingly, the final conversions of epoxy groups reach 50% after 600 s of irradiation without any inhibition period ( Figure 5). These results are in accordance with the production of photoacids (according to the rhodamine B method 21 , Figure S10), responsible for the cationic photopolymerization of EPOX. In comparison, the HNQ/Iod photoinitiating system was also tested, and no cationic photopolymerization was observed: hydroxyl group of HNQ may be responsible for the termination of the growing polymer chain during photopolymerization, thus inhibiting the cationic photopolymerization of EPOX. As a comparison, and under the same irradiation conditions, two well-known type II photoinitiators (i.e., BP and CQ) fail to initiate the photopolymerization of EPOX when coupled with Iod, therefore demonstrating the superior initiating properties of HNQA (or HNQE)/Iod systems. Therefore, only HNQA (or HNQE)/Iod systems will be considered in the following study.
itiating systems under different LEDs irradiation under air are described in Figure 5. HNQ/Iod has been used as a reference photoinitiating system for comparison. When Iod is used alone, no photopolymerization of EPOX occurs under LED irradiation (385 and 405 nm) because Iod shows a maximum absorption band around 250 nm. Cationic photopolymerizations of EPOX successfully take place ( Figure 5) when using both photoinitiating systems (HNQA (or HNQE)/Iod) under LEDs@385 and 405 nm exposure. Accordingly, the final conversions of epoxy groups reach 50% after 600 s of irradiation without any inhibition period ( Figure 5). These results are in accordance with the production of photoacids (according to the rhodamine B method 21 , Figure S10), responsible for the cationic photopolymerization of EPOX. In comparison, the HNQ/Iod photoinitiating system was also tested, and no cationic photopolymerization was observed: hydroxyl group of HNQ may be responsible for the termination of the growing polymer chain during photopolymerization, thus inhibiting the cationic photopolymerization of EPOX. As a comparison, and under the same irradiation conditions, two well-known type II photoinitiators (i.e., BP and CQ) fail to initiate the photopolymerization of EPOX when coupled with Iod, therefore demonstrating the superior initiating properties of HNQA (or HNQE)/Iod systems. Therefore, only HNQA (or HNQE)/Iod systems will be considered in the following study. Concomitant cationic photopolymerization and thiol-ene reactions of a limonene 1,2 epoxide/trithiol blend with HNQA/Iod ( Figure 6) and HNQE/Iod (Figure 7) photoinitiating systems. The vinyl, thiol, and epoxy groups were followed under air and in laminate under LEDs@385 and 405 nm. Effective kinetic profiles for the concomitant thiol−ene process and cationic photopolymerization of the limonene 1,2 epoxide/trithiol blend mixture are observed under LEDs@385 and 405 nm exposure with HNQA/Iod ( Figure 6) and HNQE/Iod systems ( Figure 7). Interestingly, the final vinyl bond conversions reach 100% after 600 s of LED exposure in laminate or under air (Table 6), and still remain higher than those of thiol, thus demonstrating that the limonene 1,2 epoxide polymerization occurs through a thiol-ene reaction (Equation (1)) and albeit to a lesser extent in a radical process. Indeed, according to EPR results, the phenyl(4-methyl) radicals that are produced from the photolysis of HNQA (or HNQE)/Iod systems ( Figure 3B,C) may add to the vinyl group of limonene 1,2 epoxide or favor a H-abstraction reaction to generate thiyl radicals according to Equation (c). Under air, the thiyl and phenyl(4-methyl) radicals react with dissolved oxygen to form peroxide radicals (ROO•, see Equation (2)) responsible for the slowing down of the limonene 1,2 epoxide polymerization. Therefore, the rates of polymerization are twice as high in laminate than under air. The following proton transfer reaction between ROO• and trithiol (Equation (2)) regenerates thiyl radicals that can add to the vinyl group, explaining that the thiol-ene reaction occurs under air whatever the LED used. Surprisingly, the cationic photopolymerization of limonene 1,2 epoxide occurs only with the HNQA/Iod photoinitiating system ( Figure 6C). This could be easily explained by the fact that the polymerization of limonene 1,2 epoxide/trithiol blend mixture with HNQE/Iod system is faster than the one involving HNQA/Iod, increasing thus the viscosity of the formulation, reducing the diffusion of cationic species in the blended mixture, therefore inhibiting the cationic ring-opening polymerization of limonene 1,2 epoxide.
ing to Equation (3). Under air, the thiyl and phenyl(4-methyl) radicals react with dissolved oxygen to form peroxide radicals (ROO•, see Equation (5)) responsible for the slowing down of the limonene 1,2 epoxide polymerization. Therefore, the rates of polymerization are twice as high in laminate than under air. The following proton transfer reaction between ROO• and trithiol (Equation (5)) regenerates thiyl radicals that can add to the vinyl group, explaining that the thiol-ene reaction occurs under air whatever the LED used. Surprisingly, the cationic photopolymerization of limonene 1,2 epoxide occurs only with the HNQA/Iod photoinitiating system ( Figure 6C). This could be easily explained by the fact that the polymerization of limonene 1,2 epoxide/trithiol blend mixture with HNQE/Iod system is faster than the one involving HNQA/Iod, increasing thus the viscosity of the formulation, reducing the diffusion of cationic species in the blended mixture, therefore inhibiting the cationic ring-opening polymerization of limonene 1,2 epoxide.

Conclusions
Two new photopolymerizable lawsone derivatives bearing vinyl (HNQA) or epoxy (HNQE) groups have demonstrated effective initiating properties for the free-radical photopolymerization of a bio-based acrylate (SOA) when used alone. According to EPR investigations, the carbon-centered radicals in α-position of the vinyl or epoxy group of HNQA or HNQE, respectively, and formed through an intermolecular H-abstraction reaction, can initiate the FRP of SOA under LED@385 and 405 nm irradiation. Furthermore, the addition of Iod, an electron acceptor and co-initiator of the lawsone derivatives, leads to effective free-radical and cationic photopolymerizations of acrylate and epoxy monomers, respectively, under visible-light irradiation. Notably, the rates of free-radical polymerization with Iod strongly increase whatever the LEDs are used. We also demonstrated that the HNQA/Iod system could promote both the concomitant thiol-ene reaction and the cationic photopolymerization of a limonene 1,2 epoxide/trithiol blend mixture upon LED@385 and 405 nm exposure with final vinyl bond conversions of around 100% even under air. On the contrary, the thiol-ene reaction toward limonene 1,2 epoxide double bonds is so effective using the HNQE/Iod system that the diffusion of cationic species is reduced in the blended mixture, thus inhibiting the cationic ring-opening polymerization of limonene 1,2 epoxide.