Introduction

Radical polymerization is advantageous for producing a large quantity of polymers from inexpensive vinyl monomers under mild conditions. Therefore, this approach is the most popular for the industrial production of polymers. The control of the molecular weight, molecular weight distribution and chain-end structures of polymers has been achieved by the development of living radical polymerizations during the past two decades [1,2,3,4,5,6]. In addition, polymer architectures can be designed with a precisely controlled structure using living radical polymerization [7,8,9,10,11]. Stereocontrol (i.e., tacticity of the polymers) via radical polymerization was also achieved via the addition of a Lewis acid and a selection of solvents [12, 13].

For a long time, many efforts focused on the synthesis of polymers with a complex structure through radical polymerization. Recently, much interest has turned to the sequence control of polymers using living radical polymerization [14,15,16,17,18,19], template polymerization [20,21,22], and predesigned reactive oligomers [23,24,25]. Radical alternating copolymerization is the oldest and simplest example of sequence-controlled polymerization [26, 27]. When an electron-accepting monomer and an electron-donating monomer are coupled, an alternating copolymer with a high molecular weight is readily produced in a high yield via a radical polymerization process in the presence or absence of a radical initiator. Alternating copolymerization is enhanced by a lack of homopolymerization ability of either or both monomers used for copolymerization. N-substituted maleimides (RMIs), which are the typical electron-accepting monomers, provide alternating copolymers by combination with olefins, styrenes, and vinyl ethers as the electron-donating counterparts [28,29,30,31,32,33]. The obtained copolymers exhibit a high onset temperature for thermal decomposition (Td5) and a high glass transition temperature (Tg) due to the stable imide ring and rigid poly(substituted methylene) structures in the main chain [32]. For example, the alternating copolymer consisting of N-methylmaleimide (MMI) and isobutene is highly heat resistant (Td5 more than 350 °C and Tg more than 150 °C) (Scheme 1) [33]. In addition, this copolymer possesses excellent optical properties (visible light transmittance more than 95%) and well-balanced mechanical properties (flexural strength more than 130 MPa and flexural elasticity more than 4.5 GPa) [33].

Scheme 1
scheme 1

Alternating copolymerization of MMI and isobutene

Full size image

Another type of olefin can provide different types of polymers with a controlled sequence structure. When the RMIs are copolymerized with an electron-donating monomer with considerable steric bulkiness, such as limonene (Lim) [34], β-pinene [35], diisobutene (DIB) [36], and 1-methylenebenzocycloheptane [37], AAB sequence-controlled copolymers (…M1M1M2M1M1M2…) are produced rather than alternating copolymers, as shown in Scheme 2. A repeating unit before the terminal repeating unit (i.e., penultimate unit (PU) effect) plays an important role in the production of AAB sequence-controlled copolymers. Many studies have been carried out to study the PU effect observed during radical copolymerization [34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54]. For example, Fukuda and coworkers reported their experimental results for copolymerization kinetics using a rotating sector method and kinetic analyses based on their robust theoretical background [40,41,42,43]. Davis and Coote et al. also investigated the PU effect using theoretical chemistry [44,45,46,47,48].

Scheme 2
scheme 2

AAB sequence-controlled copolymerization of RMI and DIB

Full size image

The unusual copolymerization behavior of the RMIs has been noted in the literature [55,56,57,58]. Recently, the PU effect was investigated in detail for the copolymerization of N-phenylmaleimide (PhMI) with various monomers. Satoh et al. reported the synthesis of AAB-type sequence-controlled copolymers with a controlled molecular weight, molecular weight distribution and chain-end structure using reversible addition-fragmentation chain transfer (RAFT) polymerization of PhMI with Lim [34] and other various olefins [38, 39]. Yamamoto et al. revealed the solvent effect on the PU-controlled copolymerization system with PhMI and β-pinene [35]. In that study, solvents with a high Lewis acidity significantly interacted with PhMI. Hisano et al. revealed that the PU control of propagation was significantly dependent on the ring number of 1-methylenebenzocycloalkanes during copolymerization with PhMI [37]. In contrast to the studies on the PU effect during copolymerization using a wide variety of olefins and styrene derivatives, no studies on the role of the N-substituents of the RMIs with regards to the PU effect have been reported. The PU effect may be due to the polar and steric effects of the substituents on the polymers and reacting monomers. However, these effects have not been clarified. In this study, we carried out the radical copolymerization of various kinds of RMIs with DIB and Lim (Fig. 1) and determined the monomer reactivity ratios based on the terminal and PU models to gain insight into the polar, resonance, and steric effects of the substituents.

Fig. 1
figure 1

Structures of monomers used in this study

Full size image

Experimental methods

Materials

Commercially available PhMI (Wako Pure Chemical Industries, Ltd., Osaka), DIB (Nacalai Tesque, Kyoto), and Lim (Nacalai Tesque, Kyoto) was used after recrystallization or distillation. The other RMIs (i.e., MMI, N-cyclohexylmaleimide (CHMI), N-tert-butylmaleimide (tBMI), N-tert-octylmaleimide (tOMI), N-acetylmaleimide (AcMI), N-(4-methylphenyl)maleimide (4MPhMI), N-(2-methylphenyl)maleimide (2MPhMI), N-(2-ethylphenyl)maleimide (2EPhMI), N-(2,6-diethylphenyl)maleimide (2,6DEPhMI), N-(4-methoxyphenyl)maleimide (4MOPhMI), and N-(4-ethoxycarbonylphenyl)maleimide (4ECPhMI)) were synthesized from maleic anhydride and the corresponding amines according to previously reported methods [28]. 2,2′-Azobis-(isobutyronitrile) (AIBN, Wako Pure Chemical Industries, Ltd., Osaka) was recrystallized from methanol. All solvents were distilled prior to use.

General procedures

Size exclusion chromatography (SEC) was carried out using Chromatoscience CS-300C, JASCO PU-2080PLUS, JASCO DG-2080-53, JASCO RI-2031-PLUS, TOSOH TSK-gel columns, GMHHR-N and GMHHR-H, and tetrahydrofuran (THF) as the eluent. Number- and weight-average molecular weights (Mn and Mw, respectively) as well as polydispersity (Mw/Mn) values were determined by calibration with standard polystyrenes. The NMR spectrum was recorded in CDCl3 using a JEOL ECS-400 spectrometer. The IR and UV-Vis spectra were recorded using JASCO FT-IR410 and Shimadzu UV-2400PC spectrometers, respectively. Thermogravimetric (TG) analysis was carried out using a Shimadzu TGA-50 with a nitrogen stream at a flow rate of 10 mL/min and heating rate of 10 °C/min. DFT calculations were carried out using Spartan’10 (Wave Function, Inc.) at the B3LYP/6-311 G* level.

Copolymerization procedures

Monomers, AIBN, and chloroform were placed in a glass tube. After the freeze–thaw cycles, the solution was heated at a determined temperature for a given time. Then, the polymerization mixture was poured into a large amount of methanol. The precipitated copolymers were filtered, washed, and dried in vacuo. The copolymer yield was gravimetrically determined. The copolymers were purified by precipitation using chloroform and methanol. The composition of the copolymers was determined by 1H NMR spectroscopy. The copolymer compositions of tBMI or tOMI with DIB were calculated based on the monomer consumption during copolymerization by 1H NMR spectroscopy using the peak intensity due to 2,4,4-trimethyl-2-pentene, which is included in DIB as the internal standard. We confirmed the good agreement between the results obtained from the two different methods. The copolymer composition was actually determined by analyses of the 1H NMR spectrum of the resulting copolymer and the monomer consumption during the copolymerization of CHMI and DIB with a 1/1 molar ratio in the feed. The CHMI content in the copolymer was determined to be 56.1 and 55.4 mol% by gravimetric and NMR analyses, respectively.

Determination of monomer reactivity ratios

All the copolymers were recovered at a low conversion to allow for analysis of the copolymerization parameters using the Mayo-Lewis equation (Eq. 1).

$$\frac{{{\mathrm{d}}[{\mathrm{M}}_1]}}{{{\mathrm{d}}[{\mathrm{M}}_2]}} = \frac{{[{\mathrm{M}}_1](r_1\left[ {{\mathrm{M}}_1} \right] + \left[ {{\mathrm{M}}_2} \right])}}{{[{\mathrm{M}}_2](r_2\left[ {{\mathrm{M}}_2} \right] + \left[ {{\mathrm{M}}_1} \right])}}$$
(1)

The monomer reactivity ratios (i.e., r1 and r2) were determined by the terminal and PU models.

For the terminal model, the r1 and r2 values are defined by Eqs. 2 and 3, respectively.

$$r_1 = k_{11}/k_{12}$$
(2)
$$r_2 = k_{22}/k_{21}$$
(3)

The Fineman-Ross [59] and Kelen-Tüdõs [60] methods yielded scattered plots, and the calculated comonomer–copolymer composition curves did not fit the experimental data in several cases (See Supporting Information). In contrast, the curve-fitting method using the nonlinear least-squares procedure resulted in good results over the nonlinear least-squares method entire range of compositions [37].

In general, the penultimate unit model involves the use of eight propagation reactions with four monomer reactivity ratios (Eqs. 4–7)

$$r_{11} = k_{111}/k_{112}$$
(4)
$$r_{12} = k_{122}/k_{121}$$
(5)
$$r_{21} = k_{211}/k_{212}$$
(6)
$$r_{22} = k_{222}/k_{221}$$
(7)

With these four monomer reactivity ratios, the copolymer composition can be represented as shown in Eq. 8.

$$f = \frac{{1 + r_{21}F\left( {\frac{{r_{11}F\, +\, 1}}{{r_{21}F\, +\, 1}}} \right)}}{{1 + \frac{{r_{12}}}{F}\left( {\frac{{r_{22}\, +\, F}}{{r_{12}\, +\, F}}} \right)}}$$
(8)

where F = [M1]/[M2] and f = d[M1]/d[M2]. In this case, the r11 and r21 values are zero due to the lack of homopolymerization ability. Therefore, Eq. 8 can be reformulated as Eq. 9.

$$\frac{{1 - 2f}}{{F(1 - f)}} = \frac{f}{{F^2(1 - f)}}r_{22} - \frac{1}{{r_{12}}}$$
(9)

Results and discussion

First, we investigated the PU effect for the copolymerization of MMI, CHMI, tBMI, and tOMI (M2) with DIB (M1) to discuss the steric effect of the N-alkyl groups. The comonomer–copolymer composition curves are shown in Fig. 2. The monomer reactivity ratios were calculated based on the terminal and PU models using the nonlinear least-squares method. The largest PU effect was observed in the copolymerization of MMI. The curve based on the terminal model did not provide a good representation of the experiment data. However, the curve considering the PU effect provide a good representation of the experimental data over the entire range of comonomer compositions in the feed, as shown in Fig. 2a. In contrast, the PU effect was negligible for the copolymerization of tOMI containing the most sterically hindered N-substituent. In this case, the r1 value that was determined using the terminal model as well as the r12 and r22 values that were determined using the PU model were in agreement (Fig. 2d). The copolymerizations of CHMI and tBMI exhibited an intermediate reaction behavior (i.e., a weak PU effect was observed) (Fig. 2b, c).

Fig. 2
figure 2

Comonomer–copolymer composition curves for the radical copolymerization of a MMI, b CHMI, c tBMI, d tOMI, and e AcMI (M2) with DIB (M1). Red and blue indicate the results analyzed using the terminal and PU models, respectively

Full size image

In Table 1, we summarize the reactivity and kinetic parameters for these copolymerization systems. The Es parameter is the Taft’s steric factor, which is represented as a negative value according to its steric bulkiness [61]. The larger absolute value indicates the larger steric hindrance. In this study, the steric bulkiness of the N-substituent was in the order of methyl < cyclohexyl < tert-butyl < tert-octyl groups. The r12/r22 value indicates the magnitude of the PU effect. A larger PU effect was observed for larger r12/r22 values. When no PU effect was observed during copolymerization, the r12/r22 value is equal to unity. The largest r12/r22 value was observed for the copolymerization of MMI, and this value decreased as the steric bulkiness of the N-substituents increased. Here, we can evaluate the absolute value of k221 using the r22 value determined in this study along with the kp (= k222) value reported in the literature [62]. As the bulkiness of the N-substituent increased, the k221 values decreased. This result suggests that significant steric hindrance is present between the N-substituents and the bulky olefin monomer when the propagating radical includes successive RMI repeating units (Fig. 3). In addition, the contribution of the polar effect on the PU control was investigated. The r12/r22 value for AcMI with an electron-accepting N-substituent was one tenth that for MMI with an electron-donating substituent. The PU effect was not observed for copolymerization of AcMI (Fig. 2e), which is in contrast to the significant PU effect observed for copolymerization of MMI. This result indicates that the introduction of an electron-withdrawing group is disadvantageous for induction of the PU effect.

Table 1 Monomer reactivity ratios and kinetic parameters for the radical copolymerization of RMIs with N-alkyl substituents (M2) with DIB (M1)
Full size table
Fig. 3
figure 3

Image for steric repulsion between the ~RMI-RMI radical and an olefin monomer

Full size image

Previously, significant steric hindrance has been observed for the radical polymerization of N-(substituted phenyl)maleimides and their copolymerization with styrene and methyl methacrylate [30]. Therefore, the steric effect of the alkyl substituents that are included in an N-phenyl moiety was investigated. In this study, we carried out the copolymerization of 4MPhMI, 2MPhMI, 2EPhMI, and 2,6DEPhMI (M2) with DIB (M1). The comonomer‒copolymer composition curves calculated by the terminal and PU models are shown in Fig. 4.

Fig. 4
figure 4

Comonomer–copolymer composition curves for the radical copolymerization of a 4MPhMI, b 2MPhMI, c 2EPhMI, and d 2,6DEPhMI (M2) with DIB (M1). Red and blue indicate the results analyzed using the terminal and PU models, respectively

Full size image

For these copolymerizations, the curves using the terminal and PU models were similar, and both methods provided a good representation of the experimental data. The r12 and r22 values computed from the results of the curve fitting process using the PU model were close to the r1 value determined from the terminal model. Therefore, the PU effect was much smaller during the copolymerization of these N-(alkylphenyl)maleimides. The r12/r22 value was 2.8 for 4MPhMI and 2MPhMI but close to unity for 2EPhMI and 2,6DEPhMI. To achieve 2:1 sequence control, the ~DIB-RMI radical should attack RMI rather than DIB, and the produced ~RMI-RMI radical must selectively react with DIB. Based on analysis of propagation rate constants [63, 64], the bulky N-substituent decreases the rate of both reactions. Due to an increase in the steric bulkiness of the alkyl substituents at the ortho position of the N-phenyl group, the kp value (and the k221 value) rapidly decreased. In addition, the overall polymerization rate decreased [64] (Table 2). The alternating tendency increased with an increase in the steric bulkiness due to a decrease in the homopropagation rate. The decrease in the kp (= k222) value was greater than that of the k221 value due to increased steric repulsion during homopropagation.

Table 2 Monomer reactivity ratios and kinetic parameters for the radical copolymerization of RMIs with N-alkylphenyl substituents (M2) with DIB and Lim (M1)
Full size table

A change in the reactivity is correlated to direct steric hindrance and the molecular conformation that accompanies the twisted structure between the maleimide and N-phenyl rings [65,66,67]. Figure 5 shows the most stable molecular conformation for 4MPhMI and 2EPhMI. These conformations were optimized using DFT calculations. The distorted conformation arises from the steric repulsion between the ortho-alkyl groups of the N-phenyl ring and the carbonyl groups of the maleimide ring. The observed spectral data and DFT results support the twisted conformation of the ortho-substituted RMIs. The torsion angles between the maleimide and phenyl rings were 47.0 and 76.9°, and the HOMO‒LUMO gaps were 4.04 and 3.64 eV for 4MPhMI and 2EPhMI, respectively. These calculated results were in good agreement with their UV spectral data. The observed λmax values were 312 and 293 nm for 4MPhMI and 2EPhMI, respectively. These twisted structures suppress the monomer reactivity due to the reduced resonance effect. Therefore, the high alternating tendency masked the PU effect.

Fig. 5
figure 5

Molecular models for 4MPhMI and 2EPhMI based on DFT calculations

Full size image

It is important to note that the appearance of the PU effect is dependent on the steric bulkiness of the olefins. No PU effect was observed during copolymerization of PhMI and other RMIs with isobutene. However, a significant PU effect was observed during copolymerization of PhMI with Lim [34], which has a bulky rigid structure. In addition, copolymerizations with DIB exhibited intermediate copolymerization features [36]. Olefins with a bulky and rigid molecular structure exhibited a larger PU effect during the copolymerization of PhMI. For example, using the r12 and r22 values reported in the literature, the r12/r22 values were calculated to be 9.0 and 69 for the copolymerization systems with 1-vinylcyclohexane and PhMI in dichloromethane and C6H5C(CF3)2OH, respectively [39]. Similarly, the r12/r22 values increased to 119 and 120 for the copolymerizations of 3-vinylcyclohexene and 3-isopropenylcyclohexene, respectively. The enhanced PU effect reflects the significant steric hindrance due to the rigid olefin structures, and the steric hindrance in the RMIs was not favored as described above.

In addition to the steric effect, the polar effect was investigated for the RMIs with substituents at the para position of PhMI (i.e., 4ECPhMI, 4MPhMI, and 4MOPhMI). The r12/r22 values were as small as 1.2‒2.8 during copolymerization with DIB. Therefore, the PU effect was studied for RMIs including the polar groups using copolymerization with Lim as the comonomer. Satoh et al. reported that Lim induced large PU effects for copolymerization with PhMI [34]. However, no data are available on the polar effect of RMIs. The bulky rigid structure of the olefin monomers is expected to enhance the PU effect. As shown in Fig. 6 and Table 2, the r12/r22 values were determined to be 45, 23, and 15 for the copolymerizations of 4MOPhMI, 4MPhMI, and PhMI, respectively. The order of the magnitude of the effect was in agreement with the order of the electron-donating properties of the substituents. Yamamoto et al. previously reported a significant solvent effect during radical copolymerization of PhMI with β-pinene [35]. When the Lewis acidity of the solvent increased, the PU effect increased. For example, the r12/r22 values increased in the following order: 29 in tetrahydrofuran, 44 in 1,2-dichloroethane, 78 in 2,2,2-trifluoroethanol, and 560 in perfluoro-tert-butanol [35]. An increase in the Lewis acidity can enhance the interaction between the carbonyl group of the RMIs and the solvent molecules, leading to an increase in steric hindrance around the RMIs. The electron-donating N-substituent may assist the RMI-solvent interaction due to an increase in the electron density on the oxygen atom of the carbonyl groups, which leads to enhancement of the PU effect.

Fig. 6
figure 6

Comonomer–copolymer composition curves for the radical copolymerization of a PhMI, b 4MOMI, and c 4MPhMI (M2) with Lim (M1). Red and blue indicate the results analyzed using the terminal and PU models, respectively

Full size image

Conclusion

In this study, we investigated the influence of steric, resonance, and polar effects of N-substituents on sequence control during radical copolymerization of RMIs with DIB and Lim in chloroform at 60 °C. The monomer reactivity ratios (i.e., r2 (= k22/k21) based on the terminal model as well as r12 (= k122/k121) and r22 (= k222/k221) based on the PU model) were determined. When the steric effect was studied using RMIs with bulky N-alkyl substituents, the PU effect was suppressed due to steric hindrance, leading to the formation of alternating copolymers with DIB. The introduction of bulky substituents at the ortho position of N-phenylmaleimides also predominantly caused alternating copolymer formation. However, the polar effect was observed during copolymerization of N-phenylmaleimides including various para-substituents with DIB and Lim. Therefore, less bulky and more electron-donating substituents that were introduced in the RMIs enhanced the PU effect for radical copolymerization of the RMIs with olefins. We have demonstrated that 2:1 sequence-controlled maleimide copolymers can be efficiently produced during radical copolymerization of Lim and N-phenylmaleimides with an electron-donating substituent at the 4-position of the N-phenyl group.