Thermal NMP at high temperatures often causes hydrogen abstraction by the nitroxide mediator from the propagating radical [64,65]. Photopolymerization carried out at room temperature can eliminate this side reaction.

Photoradical polymerization of MMA was performed by azobis(4-methoxy-2,4-dimethylvaleronitrile) (AMDV) as the initiator and 4-methoxy-TEMPO (MTEMPO) as the mediator in the presence of bis(alkylphenyl)iodonium hexafluorophosphate (BAI) as the photo-acid generator [37]. The bulk polymerization was carried out at room temperature by irradiation with a high-pressure mercury lamp. The orange-colored monomer solution turned colorless after polymerization. These results are shown in Table 1. Polymerization in the absence of MTEMPO produced poly(MMA) (PMMA) with a broad MWD. BAI had a slight effect of decreasing the MWD, although the MWD was still broad. It was found that the MWD dramatically decreased in the presence of MTEMPO.

Figure 1 shows the GPC profiles of the PMMA obtained in the presence and absence of MTEMPO. The PMMA in the presence of MTEMPO provided a unimodal GPC with a comparatively narrow MWD, whereas the PMMA prepared in its absence showed a bimodal GPC. BAI also affected the MWD in the presence of MTEMPO. As a result of increasing of the molar ratio of BAI to MTEMPO (BAI/MTEMPO), the polymerization was accelerated, so that BAI served as the accelerator of the polymerization. However, the increase in BAI caused the broadening of the MWD of the resulting polymer. On the other hand, an increase in the molar ratio of MTEMPO to AMDV (MTEMPO/AMDV) decelerated the polymerization.

Figure 2 shows the ¹H NMR spectra of BAI, AMDV, and the PMMA (Mn = 7,030, Mw/Mn = 1.60, conversion = 23%) prepared by the MTEMPO-mediated photopolymerization at MTEMPO/AMDV = 1.1 and BAI/MTEMPO = 0.5. No observed signals originating from BAI in the spectrum of the PMMA indicates that no fragments of BAI were contained in the polymer structure. It is suggested that the BAI fragments did not combine with the growing polymer chain end or MTEMPO, but just interacted with MTEMPO. The polymerization had no effect on the tacticity of PMMA based on the observation of the signals at 0.84 ppm (syndiotactic), 1.01 ppm (atactic), and 1.25 ppm (isotactic) [66]. It was found that the PMMA involved the 1-cyano-1,3-dimethyl-3-methoxybutyl group (CDM) at the chain head and MTEMPO at the chain end. Signals of the methoxy protons originating from CDM were discerned at 3.13–3.22 ppm, while those from MTEMPO were observed at 3.29–3.39 ppm (Figure 3). The molar ratio of MTEMPO to CDM in the PMMA was estimated to be 0.952 based on their respective methoxy protons, indicating that the PMMA had the CDM and MTEMPO at the 1:1 molar ratio. This good agreement in the molar ratio of MTEMPO/CDM indicates that the growing radical generated by the initiation with the CDM radical was completely captured by MTEMPO (Scheme 1). The molecular weight of the PMMA was estimated by ¹H NMR to be Mn = 6500 based on the methoxy protons of CDM and the methyl ester protons of the MMA units. The theoretical molecular weight when the initiator efficiency (IE) determined by UV analysis (IE = 0.378) was taken into account was calculated to be Mn = 6270, being in close agreement with the molecular weight estimated by ¹H NMR.

The first order time-conversion plots for the polymerization demonstrated that the number of polymer chains was constant throughout the course of the polymerization. The ln([M]₀/[M]) linearly increased over time (Figure 4). As can be seen in Figure 5, the plots of the molecular weight of the resulting PMMA vs. the monomer conversion also linearly increased and the molecular weights were in good agreement with the theoretical molecular weight calculated on the basis on the initiator efficiency (IE = 0.378), indicating the livingness of the polymerization. The MWD of the PMMA was maintained at ca. 1.6 throughout the polymerization. The GPC curve was shifted to the higher side of the molecular weight with an increase in the conversion (Figure 6), also supporting the living mechanism.

The polymerization was revealed to have a photoswitching ability based on the investigation of the polymerization in the dark. Figure 7 shows the variation in the monomer conversion and molecular weight of PMMA when the irradiation was interrupted for a time during the polymerization. There were negligible changes in the conversion and molecular weight during the dark reaction. However, the conversion and molecular weight increased again by continued irradiation. The progress of the polymerization can be controlled by switching the irradiation on or off.

The MTEMPO-mediated controlled/living photoradical polymerization was performed using the azoinitiator in the presence of the BAI photo-acid generator. In order to clarify which characteristic of the azoinitiator dominates the MWD of the PMMA, the MTEMPO-mediated photopolymerization of MMA was performed using eight different kinds of azoinitiators in the presence of BAI [42]. The azoinitiators are shown in Scheme 2 and the results are summarized in Table 2.

The concentration of the growing polymer chain ([P]) was estimated on the basis of the conversion and the molecular weight of the resulting PMMA. Based on the [P], the IE was also determined. Racemic-AMDV (r-AMDV), meso-AMDV (m-AMDV), V-601, and VAm-110 provided much narrower MWDs with moderate conversions, although VAm-110 produced only a 25% conversion. In particular, r-AMDV and m-AMDV produced the PMMA with the narrowest MWD. r-AMDV had a higher [P] and IE than m-AMDV, being in good agreement with the fact that r-AMDV has a higher reactivity than m-AMDV [67]. As a result of doubling MTEMPO to the initiator, all the initiators produced PMMA with a MWD below 1.7, and the resulting PMMAs showed sharp GPC curves. There were negligible differences in the [P] and IE values between the MTEMPO/initiator ratios of unity and 2.

The characteristics of the UV absorption spectra of the initiators, coupled with their 10-h-half-life temperatures are listed in Table 3. Initiators with higher ε values tended to more strictly control the molecular weight and provide a higher IE. The reason that r-AMDV, m-AMDV, V-601, and VAm-110 provided much narrower MWDs than the other initiators may partly be due to the fact that these initiators exhibit the n → σ^* transition in addition to the high ε. It was found that the half-lives of the initiators had little effect on the molecular weight control. The [P] and IE were also independent of the half-life temperature.

In order to clarify the mechanism of the MTEMPO-mediated photopolymerization, the polymerization was performed using 1-(cyano-1-methylethoxy)-4-methoxy-2,2,6,6-tetramethylpiperidine (CMTMP), the alkoxyamine adduct as the initiator [39]. CMTMP was prepared by the reaction of MTEMPO and AIBN in methanol. CMTMP had a slight absorption at 257 nm as λ_max (ε = 3.43) and almost no overlap with the illumination of the mercury lamp, implying a low photo activity compared to AIBN. The photopolymerization of MMA was performed using CMTMP as the initiator, at room temperature. The results are shown in Table 4.

The polymerization only slightly occurred using CMTMP itself; however, it smoothly proceeded in the presence of BAI. The monomer conversion increased with the increasing molar ratio of BAI to CMTMP (BAI/CMTMP). On the other hand, the molecular weight of the resulting polymer decreased with the increasing BAI/CMTMP ratio, suggesting that the BAI/CMTMP ratio affected the initiator efficiency. The IE of CMTMP was determined on the basis of the concentration of the [P] calculated by the monomer conversion and the molecular weight of the resulting polymer. The IE increased with the increase in the BAI/CMTMP ratio, although it was not quantitative even at 1.0 of BAI/CMTMP. This nonquantitative IE was probably caused by the cage effect and the deactivation of CMTMP by the disproportionation in the hydroxylamine and methyl methacrylonitrile (Scheme 3). In fact, the formation of a significant amount of hydroxylamine lowered the yield of CMTMP in its preparation by the reaction of MTEMPO and AIBN. The molecular weight distributions of the resulting polymers were somewhat broadened as compared to those of the polymers obtained by the photopolymerization by AIBN and 4-methoxy-TEMPO in the presence of BAI [36,67]. This broad molecular weight distribution was probably due to the slow initiation by CMTMP having a low photo activity.

The polymerization was carried out for the different initial concentrations of CMTMP ([CMTMP]₀) at a constant BAI/CMTMP ratio, resulting in the molecular weight of the polymer decreasing with an increase in [CMTMP]₀. The plots of the molecular weight vs. the reciprocal of [CMTMP]₀ provided a linear correlation, implying the living nature of the polymerization (Figure 8). Figure 9 shows the GPC profiles of the polymers produced for each conversion. The curves shifted to the higher molecular weight side with an increase in the conversion. The plots of the molecular weight of the polymer vs. the conversion are shown in Figure 10. The molecular weight linearly increased with an increase in the conversion; however, the line did not pass through the origin and was almost parallel to the theoretical line. This phenomenon can be accounted for by the fact that a small amount of a polymer with a lower molecular weight was produced during the very early stage before the system reached the steady state of the living polymerization.

The ¹H NMR analysis revealed that the formation of the polymer, with a low molecular weight during the very early stage before the steady state, was caused by the phenyl radicals generated from CMTMP. Figure 11 shows the ¹H NMR spectra of the polymers obtained by the MTEMPO-mediated polymerization initiated by CMTMP and AIBN/MTEMPO in the presence of BAI. Whereas no BAI fragments were inserted into the polymer structure prepared from AIBN/MTEMPO, the polymer obtained from CMTMP contained the BAI fragment in its structure, because the aromatic proton signals originating from BAI were discerned at 7.21–7.40 and 7.85–7.92 ppm. The formation of the phenyl radical is supported by the decomposition mechanism of the photo-acid generator [68,69,70]. The initiation by the phenyl radical was also confirmed by the experiment in which the uncontrolled photopolymerization of MMA performed in the presence of BAI without CMTMP produced a polymer with Mn = 62,800 and Mw/Mn = 3.77 at 84% conversion within only 2 h. During the MTEMPO-mediated polymerization, the propagating chains formed by the phenyl radical probably participated in the controlled polymerization by MTEMPO by obtaining the counter radical from the hydroxylamine (Scheme 4) because the GPC curves were shifted to the higher molecular weight side without deactivation with the increasing conversion. In the MTEMPO/AIBN system, the propagation proceeded by repeating the dissociation and recombination of the C-ON bond, as well as the thermal polymerization, since no BAI fragments were contained in the polymer structure. However, electron transfer is expected to occur between MTEMPO and BAI in the excited state during the dissociation and recombination when it is taken into consideration that continued polymerization is difficult in the absence of BAI and that the excited diphenyliodonium salt easily receives an electron [68,69,70]. Consequently, the mechanism of the polymerization was thus proposed (Scheme 5).

The thermal NMP is often deactivated through the hydrogen elimination from the growing polymer radical by the nitroxide at the end stage of the polymerization [64,76,77]. The stability of the growing polymer chain ends for the MTEMPO-mediated photopolymerization of MMA was investigated by the block copolymerization with isopropyl methacrylate (ⁱPMA) [44]. The block copolymerization of ⁱPMA was performed using the PMMA prepolymer prepared through the photopolymerization of MMA by the r-AMDV initiator, the MTEMPO mediator, and the ^tBuS photo-acid generator. The block copolymerization continued until the reaction mixture was completely solidified. For the MMA polymerization by this system, the monomer conversion linearly increased up to 5 h, followed by a slight increase over 5 h to 10 h [41]. The GPC profiles of the prepolymer prepared by the MMA polymerization for 6.5 h and the block copolymer are shown in Figure 15. The resulting block copolymer contained the prepolymer due to the deactivation of the propagating chain end. It was found that ca. 75% of the prepolymer was deactivated on the basis of the area ratio of the respective molecular weights in the GPC curve of the block copolymer. The MMA polymerization was shortened to 5 h in order to prevent the deactivation of the growing chain end. As can be seen in Figure 16, the GPC curve of the block copolymer was shifted to the higher molecular weight side and contained no prepolymer. The molecular weight and its distribution of the block copolymer were estimated to be Mn = 43,200 and Mw/Mn = 2.14, while those of the prepolymer were Mn = 10,100 and Mw/Mn = 1.63. It was found that the deactivation of the propagating chain ends was caused by a decrease in the monomer concentration and occurred during the end stage of the polymerization. The absolute molecular weight of the PMMA-b-PⁱPMA diblock copolymer obtained by the MMA polymerization for 5 h was determined by ¹H NMR to be Mn = 36,000 based on the signal intensity of the methyl ester protons at 3.63 ppm and the methine protons at 4.87 ppm.

The deactivation of the growing polymer chain ends occurred during the MTEMPO-mediated photopolymerization, as well as the thermal NMP [44]. The growing polymer chain ends should be more stabilized in a solution than in the bulk, because their deactivation occurs at the end stage of the polymerization where most monomers are consumed [64,76,77], although the solution polymerization retards the propagation rate. In order to obtain more stable growing polymer chain ends, the solution polymerization was explored [45]. The polymerization was performed in acetonitrile at room temperature using the r-AMDV initiator and the MTEMPO mediator, and the ^tBuS photo-acid generator. The results are shown in Table 8. The solution polymerization allowed the conversion to reach 97%, whereas it was difficult for the bulk polymerization to increase the conversion over 85% [41].

The livingness of the polymerization was investigated at the MTEMPO/r-AMDV of 2.13 and ^tBuS/MTEMPO of 1.02. The time-conversion and its first-order plots for the polymerization are shown in Figure 17. The ln[M]₀/[M] plots showed different lines before and after 4 h, indicating that the radical concentrations were different for these two lines. The polymerization should be under the non-steady-state below 4 h and reached the steady-state over this time. The non-steady-state was not observed in the bulk polymerization [41]. It is suggested that MTEMPO could not effectively trap the propagating radical due to its low concentration in the solution polymerization, causing the occurrence of a normal termination between the propagating radicals under the non-steady-state. However, the proportion of the polymers produced by the normal termination is low based on its conversion (<18%). Figure 18 shows the plots of the molecular weight vs. the conversion for the polymerization. It was observed that oligomers with several thousand molecular weights were formed under the non-steady-state below the 18% conversion. However, the molecular weight linearly increased with the conversion under the steady-state, indicating that the polymerization proceeded by the living mechanism. The MWD was around 1.8, somewhat higher than that for the bulk polymerization. This broader MWD should be due to the presence of the oligomers produced under the non-steady-state.

The stability of the growing polymer chain ends for the solution polymerization was also investigated through the block copolymerization. The PMMA prepolymer was prepared by a 14 h polymerization. The block copolymerization was performed for 19 h using ⁱPMA as the second monomer. Figure 19 shows the GPC profiles of the prepolymer and the block copolymer. The curve of the block copolymer was shifted to the higher molecular weight side without deactivation, indicating that the prepolymer efficiently initiated the polymerization of ⁱPMA. The growing polymer chain ends were stabilized even at a high conversion over 85% for the solution polymerization.

The MTEMPO-mediated photo controlled/living radical polymerization has the potential to create a variety of architectures through designing TEMPO derivatives. The synthesis of a diblock copolymer using a macromediator of TEMPO is described in this section.

TEMPO-terminated poly(tetrahydrofuran) (PTHF-TEMPO) with Mn = 1,720 and Mw/Mn = 1.91 was used as the macromediator for the polymerization [78]. The photopolymerization of MMA was performed by AMDV, BAI, and PTHF-TEMPO [38]. Figure 20 shows the ¹H NMR spectra of the resulting block copolymer with Mn = 11,300 and Mw/Mn = 1.83. The observation of a signal of the methylene attached to the TEMPO at 3.36 ppm confirmed that the PMMA and PTHF blocks were connected through the TEMPO. The plots of the molar ratio of the MMA unit to the THF unit versus the conversion using their signal intensity are shown in Figure 21. The MMA/THF ratio linearly increased with the increase in the conversion, suggesting the living nature of the polymerization.

The living mechanism of the polymerization was also confirmed on the basis of plots of linear correlation of the molecular weight of the copolymer vs. the initial concentration of AMDV ([AMDV]₀). The MWD of the copolymers was the same as, or less than, that of PTHF-TEMPO. Linear increases in the first order time-conversion plots and the conversion-molecular weight plots endorsed the controlled/living mechanism.

The heterogeneous polymerization using the controlled/living radical polymerization has the potential to simultaneously control the molecular weight and particle size of a polymer. A number of the thermal controlled/living radical heterogeneous polymerization systems have been studied, and there is an outstanding review article on them [79]. The MTEMPO-mediated photo dispersion polymerization of MMA was performed at room temperature using r-AMDV, ^tBuS, and polyvinylpyrrolidone (PVP) as the surfactant in a mixed solvent of MeOH/water = 3/1 (v/v) [49]. Figure 22 shows SEM images of the resulting polymer particles. The photo dispersion polymerization in the absence of MTEMPO provided nonspecific particles. High molecular weight polymers should be produced by very fast uncontrolled polymerization, precipitated before being stabilized by PVP. On the other hand, the MTEMPO-mediated photo dispersion polymerization produced spherical particles of PMMA. The particle size distribution decreased as the PVP concentration increased. It was found that the size distributions were below 1.1 for the PVP concentrations of 60 and 75 wt% (Table 9). The spherical particles showed a comparatively narrow molecular weight distribution, indicating the simultaneous control of the molecular weight and particle size.

Figure 23 shows the time-conversion plots, the first order time-conversion plots, and the variation in turbidity of the polymerization system. The turbidity was determined by UV based on the absorbance at 230 nm. The turbidity of the system rapidly increased as the polymerization was initiated and became almost constant over 12 h. The constant turbidity suggests the limitation of the light passing though the system. The continued increase in the conversion in spite of the constant turbidity implies that over 12 h, the polymerization occurred outside of the particles. This inference was supported by a decrease in the molecular weight of the particles and the formation of a large amount of small particles outside of the secondary particles by the further extension of the polymerization time. The small particles had ca. a 1-μm diameter. It is considered that polymerization outside of the secondary particles did not occur by the transfer of the active species in the secondary particles, but by the remaining initiator in the solution or the autopolymerization. However, polymerization outside of the secondary particles should be slow because of the low monomer concentration at the later stage of polymerization. Therefore, the simultaneous control of the molecular weight and particle size was possible as long as light penetrated into the secondary particles.