Study on MMA and BA Emulsion Copolymerization Using 2,4-Diphenyl-4-methyl-1-pentene as the Irreversible Addition–Fragmentation Chain Transfer Agent

As a chain transfer agent, 2,4-diphenyl-4-methyl-1-pentene (αMSD) was first introduced in the emulsion binary copolymerization of methyl methacrylate (MMA) and butyl acrylate (BA) based on an irreversible addition–fragmentation chain transfer (AFCT) mechanism. The effects of αMSD on molecular weight and its distribution, the degree of polymerization, polymerization rate, monomer conversion, particle size, and tensile properties of the formed latexes were systematically investigated. Its potential chain transfer mechanism was also explored according to the 1H NMR analysis. The results showed that the increase in the content of αMSD could lead to a decline in molecular weight, its distribution, and the degree of polymerization. The mass percentage of MMA in the synthesized polymers was also improved as the amounts of αMSD increased. The chain transfer coefficients of αMSD for MMA and BA were 0.62 and 0.47, respectively. The regulation mechanism of αMSD in the emulsion polymerization of acrylates was found to be consistent with Yasummasa’s theory. Additionally, monomer conversion decreased greatly to 47.3% when the concentration of αMSD was higher than 1 wt% due to the extremely low polymerization rate. Moreover, the polymerization rate was also decreased probably due to the desorption and lower reactivity of the regenerative radicals from αMSD. Finally, the tensile properties of the resulting polyacrylate films were significantly affected due to the presence of αMSD.


Introduction
Due to the growing environmental concerns, emulsion polymerization has become a vital method for polymer production with its environmentally friendly process, high polymerization rate, high monomer conversion, and excellent heat dissipation performance [1]. Many engineering efforts have also been put on the multiscale modeling of this process [2,3]. Chain transfer agent (CTA) as an important component, including reversible addition-fragmentation chain transfer (RAFT) and traditional CTAs, have been widely employed to control the molecular weights of synthesized polymers during the emulsion polymerization. Although many works about RAFT CTAs have recently been conducted, the complex processes and strict operation conditions of RAFT CTAs limit their wide applications. [4][5][6] For traditional CTAs, including mercaptans [7][8][9][10], alcohols [11], and halides [12], R X C Z Y Scheme 1. General chemical structure of the irreversible addition-fragmentation chain transfer (AFCT) agent.
2,4-Diphenyl-4-methyl-1-pentene (αMSD), shown in Scheme 2, is a typical irreversible AFCT which can be prepared by α-methylstyrene dimerization in the presence of an aqueous sulfuric acid catalyst in the liquid-liquid mode of operation [25,26]. Currently, αMSD has been introduced into solution and bulk radical polymerization in order to synthesize polystyrene with decreased molecular weights and narrowed molecular weight distribution [27,28]. For the specific regulation mechanism of αMSD in the polymerization (Scheme 3) Fisher et al. [29] proposed that the chain transfer would proceed in two cases. In the first case, alkyl hydrogen carrying a free electron abstracted from an αMSD was combined with a chain radical to form a dead chain (Scheme 3a1), whereas the residual structure of αMSD carrying a free electron could reinitiate the polymerization process (Scheme 3a1). In the second case, a chain radical was directly added to αMSD and then formed a transient state radical. An alkyl hydrogen radical divided from the αMSD structure of the latter radical was then transferred to a monomer to form a new radical. Afterwards, the new radical was able to reinitiate the polymerization. Finally, the residual part of the transient state radical became a dead chain, and the end group of the dead chain was an internal olefin (Scheme 3a2). However, according to the results of the thermal polymerization of styrene, Yasummasa et al. [30] concluded that the polymer chain radical would first be added to the terminal double bond of αMSD to produce an intermediate radical, and then this radical would decompose into a polymer chain with a terminal double bond and a cumyl radical. Finally, the cumyl radical would reinitiate the polymerization (Scheme 3b1) [31][32][33][34]. Furthermore, the regulation mechanism study of αMSD in the radical bulk and suspension polymerization of styrene has been observed to support Yasummasa's theory. 2,4-Diphenyl-4-methyl-1-pentene (αMSD), shown in Scheme 2, is a typical irreversible AFCT which can be prepared by α-methylstyrene dimerization in the presence of an aqueous sulfuric acid catalyst in the liquid-liquid mode of operation [25,26].
Polymers 2020, 12, x FOR PEER REVIEW 2 of 14 halides [12], several disadvantages have also been observed, such as an unpleasant smell for mercaptans and serious environmental pollution for halides. In the case of alcohols, the amount used as a molecular weight regulator in the polymerization is typically large (10%-20% relative to monomer mass), and it is also not considered to be an environmentally friendly agent. An irreversible addition-fragmentation chain transfer (AFCT) agent has been developed to solve the above problems. It displayed higher chain transfer efficiency, lower toxicity, and lower dosage. The first report about the AFCT agent was presented in 1988, when Meijs et al. [13] used α-benzyloxy-styrene as a chain transfer agent for the preparation of poly(methyl methacrylate) and polystyrene with low molecular weights. Then, various new types of AFCT agents were reported in research papers [14][15][16][17][18][19][20][21] and patents [22][23][24]. The common structure of the irreversible AFCT agent is shown in Scheme 1. C=Z is a double bond that is vulnerable to be attacked by the chain radical. R-X is a single bond that can be broken easily, and Y is a group to activate the reaction. The general mechanism for the irreversible AFCT agent mainly contains three steps. Firstly, the chain radical is added to the unsaturated double bond of C=Z. Then, the addition product offers a leaving group radical (R•) by the decomposition. Eventually, the novel radical reinitiates the polymerization process. 2,4-Diphenyl-4-methyl-1-pentene (αMSD), shown in Scheme 2, is a typical irreversible AFCT which can be prepared by α-methylstyrene dimerization in the presence of an aqueous sulfuric acid catalyst in the liquid-liquid mode of operation [25,26]. Scheme 2. Chemical structure of 2,4-diphenyl-4-methyl-1-pentene (αMSD).
Currently, αMSD has been introduced into solution and bulk radical polymerization in order to synthesize polystyrene with decreased molecular weights and narrowed molecular weight distribution [27,28]. For the specific regulation mechanism of αMSD in the polymerization (Scheme 3) Fisher et al. [29] proposed that the chain transfer would proceed in two cases. In the first case, alkyl hydrogen carrying a free electron abstracted from an αMSD was combined with a chain radical to form a dead chain (Scheme 3a1), whereas the residual structure of αMSD carrying a free electron could reinitiate the polymerization process (Scheme 3a1). In the second case, a chain radical was directly added to αMSD and then formed a transient state radical. An alkyl hydrogen radical divided from the αMSD structure of the latter radical was then transferred to a monomer to form a new radical. Afterwards, the new radical was able to reinitiate the polymerization. Finally, the residual part of the transient state radical became a dead chain, and the end group of the dead chain was an internal olefin (Scheme 3a2). However, according to the results of the thermal polymerization of styrene, Yasummasa et al. [30] concluded that the polymer chain radical would first be added to the terminal double bond of αMSD to produce an intermediate radical, and then this radical would decompose into a polymer chain with a terminal double bond and a cumyl radical. Finally, the cumyl radical would reinitiate the polymerization (Scheme 3b1) [31][32][33][34]. Furthermore, the regulation mechanism study of αMSD in the radical bulk and suspension polymerization of styrene has been observed to support Yasummasa's theory. Currently, αMSD has been introduced into solution and bulk radical polymerization in order to synthesize polystyrene with decreased molecular weights and narrowed molecular weight distribution [27,28]. For the specific regulation mechanism of αMSD in the polymerization (Scheme 3) Fisher et al. [29] proposed that the chain transfer would proceed in two cases. In the first case, alkyl hydrogen carrying a free electron abstracted from an αMSD was combined with a chain radical to form a dead chain (Scheme 3a1), whereas the residual structure of αMSD carrying a free electron could reinitiate the polymerization process (Scheme 3a1). In the second case, a chain radical was directly added to αMSD and then formed a transient state radical. An alkyl hydrogen radical divided from the αMSD structure of the latter radical was then transferred to a monomer to form a new radical. Afterwards, the new radical was able to reinitiate the polymerization. Finally, the residual part of the transient state radical became a dead chain, and the end group of the dead chain was an internal olefin (Scheme 3a2). However, according to the results of the thermal polymerization of styrene, Yasummasa et al. [30] concluded that the polymer chain radical would first be added to the terminal double bond of αMSD to produce an intermediate radical, and then this radical would decompose into a polymer chain with a terminal double bond and a cumyl radical. Finally, the cumyl radical would reinitiate the polymerization (Scheme 3b1) [31][32][33][34]. Furthermore, the regulation mechanism study of αMSD in the radical bulk and suspension polymerization of styrene has been observed to support Yasummasa's theory. Although there have been some developments for the radical bulk and suspension polymerization using αMSD, its utilization in the emulsion polymerization has never been reported. In the present study, different levels of αMSD were incorporated into poly-(meth) acrylate emulsion via the semi-continuous emulsion copolymerization of MMA and BA to investigate its effects on the properties of the resulting polymer emulsion, such as molecular mass, polymerization rate, and particle size. The αMSD regulation mechanism was also discussed.
Emulsion polymerization: The emulsion polymerization of (meth) acrylates was carried out in a 500 mL 4-neck glass reactor equipped with a mechanical stirrer, a reflux condenser, a nitrogen gas inlet, and a thermometer. These following substances and amounts were mixed into the DDW (158.40 g): SDS (2.60 g), NP-10 (0.65 g), OS (0.65 g), and Na2HPO4·12H2O (0.40 g), followed by mechanical stirring to form an aqueous solution. The (meth) acrylate monomers MMA (65 g) and BA (65 g) were homogeneously mixed with various amounts of αMSD under a mechanical stirring. The formed organic phase was then added dropwise into the aqueous solution to form an emulsion solution. A total of 10 g of the resulting solution and 0.55 g of initiator solution (0.05 g APS in 0.50 g DDW) were poured into the glass reactor and then heated to 75 °C. Finally, a residual pre-emulsion and 6.6 g of initiator solution containing 0.60 g APS were added dropwise simultaneously into the reactor to polymerize at 75 °C for 4 h under a nitrogen atmosphere.
Characterization: Monomer conversion was determined by a gravimetric analysis of samples withdrawn from the glass reactor during polymerization, and these samples were quenched and dried at 120 °C for 2 h. The conversion calculations were based on the total amount of solid of each emulsion sample taken at the beginning of the polymerization. The number and weight-average molecular weights of the polymers were obtained by the gel permeation chromatography (GPC) via a Malvern Viscotek3580 system, which was calibrated using standard narrow molecular weight polystyrene samples. The dried polymer films were dissolved in GR-grade tetrahydrofuran eluent and filtered through a polytetrafluoroethylene (PTFE) filter (0.22 μm) before the injection. Analysis Although there have been some developments for the radical bulk and suspension polymerization using αMSD, its utilization in the emulsion polymerization has never been reported. In the present study, different levels of αMSD were incorporated into poly-(meth) acrylate emulsion via the semi-continuous emulsion copolymerization of MMA and BA to investigate its effects on the properties of the resulting polymer emulsion, such as molecular mass, polymerization rate, and particle size. The αMSD regulation mechanism was also discussed.
Emulsion polymerization: The emulsion polymerization of (meth) acrylates was carried out in a 500 mL 4-neck glass reactor equipped with a mechanical stirrer, a reflux condenser, a nitrogen gas inlet, and a thermometer. These following substances and amounts were mixed into the DDW (158.40 g): SDS (2.60 g), NP-10 (0.65 g), OS (0.65 g), and Na 2 HPO 4 ·12H 2 O (0.40 g), followed by mechanical stirring to form an aqueous solution. The (meth) acrylate monomers MMA (65 g) and BA (65 g) were homogeneously mixed with various amounts of αMSD under a mechanical stirring. The formed organic phase was then added dropwise into the aqueous solution to form an emulsion solution. A total of 10 g of the resulting solution and 0.55 g of initiator solution (0.05 g APS in 0.50 g DDW) were poured into the glass reactor and then heated to 75 • C. Finally, a residual pre-emulsion and 6.6 g of initiator solution containing 0.60 g APS were added dropwise simultaneously into the reactor to polymerize at 75 • C for 4 h under a nitrogen atmosphere.
Characterization: Monomer conversion was determined by a gravimetric analysis of samples withdrawn from the glass reactor during polymerization, and these samples were quenched and dried at 120 • C for 2 h. The conversion calculations were based on the total amount of solid of each emulsion sample taken at the beginning of the polymerization. The number and weight-average molecular weights of the polymers were obtained by the gel permeation chromatography (GPC) via a Malvern Viscotek3580 system, which was calibrated using standard narrow molecular weight polystyrene samples. The dried polymer films were dissolved in GR-grade tetrahydrofuran eluent and filtered through a polytetrafluoroethylene (PTFE) filter (0.22 µm) before the injection. Analysis was conducted with a flow rate of 1 mL per minute at 40 • C. Average particle size (d p , light intensity-weighted) and particle-size distribution index (PDI) were measured by a laser light-scattering method using a Malvern ZETASIZER Nano at 25 • C. All latex samples for the measurement were diluted 50 times with DDW. The PDI was a dimensionless coefficient denoting the broadness of the distribution diameter, and it was produced by a second-order method of cumulant analysis [35].
The 1 H NMR spectra of the resultant polymers were measured with a Bruker DMX 300 NMR. The polymer samples were dissolved into CDCl 3 (10-15 wt/vol %) at 25 • C. The tensile properties of the dried polymer films were measured on electronic tensile machine CMT7504 with a crosshead speed of 50 mm/min and a load cell of 250 N at room temperature. The stress-strain curves of the polymer films were plotted to estimate the films' strength and toughness variations.

Effect of αMSD on the Physichemical Properties of Polymers
Different levels of αMSD were introduced into the emulsion polymerization of (meth) acrylate in order to effectively regulate the physical properties of the resulting polymers. Its effect on average molecular mass, polymerization degree, and monomer conversion is shown in Table 1. As the addition of αMSD increased from 0% to 0.1%, the number-average molecular weights were evidently decreased from 40.56 × 10 4 to 28.58 × 10 4 g/mol, whereas monomer conversion underwent a slight decrease to 97.8%. When αMSD concentration increased from 0.1% to 1%, both molecular weights and monomer conversion were significantly decreased, as the presence of αMSD was able to effectively interfere with the polymerization process. In terms of the dispersity (Ð) of the molecular weight, the addition of αMSD contributed to the gradual decrease of the dispersity to nearly 1.60 due to a mathematical artefact in the definition of dispersity [36]. However, despite its efficiency in adjusting molecular weights and Ð, only 44.6% monomer conversion was accomplished when the concentration of αMSD was 3%. These results indicate that the utilization of αMSD as a chain transfer agent is effective in the adjustment of the molecular weights and the dispersity Ð of synthesized polymers, but a further addition of αMSD might compromise the monomer conversion. In addition, n X in Table 1 was obtained from Equation (1) [37]. As MMA was expected to polymerize first, followed by the polymerization of BA [38], Equation (1) is shown as follows.
where x is the mass fraction of MMA in the synthesized copolymers; F1 as the mole fraction of MMA in the copolymer is calculated by Meyer's [37] copolymerization equations (Equations (2) and (3)); m1 and m2 are the relative molecular masses of MMA and BA, respectively; f1 0 and f1 are the mole fractions of MMA in the monomer phase at the beginning and after a certain conversion of the polymerization, respectively; f2 0 and f2 are the mole fractions of BA at different stages; C is a monomer conversion; α, β, γ, and δ are coefficients relating to the properties of monomers, which are shown in Table 2.
Note: r1 and r2 are reactivity ratios of MMA and BA (shown in Table 4), respectively.
The relative variation tendencies of x and 1-x could be seen in Figure 2. As the amounts of initiator and αMSD in the polymerization process were very small, their proportions in the polymer were ignored. It is evident that x increased gradually with increasing amounts of αMSD. This indicated that αMSD affected not only the molecular weights of polymers, but also its composition. In addition, X n in Table 1 was obtained from Equation (1) [37]. As MMA was expected to polymerize first, followed by the polymerization of BA [38], Equation (1) is shown as follows.
where x is the mass fraction of MMA in the synthesized copolymers; F 1 as the mole fraction of MMA in the copolymer is calculated by Meyer's [37] copolymerization equations (Equations (2) and (3)); m 1 and m 2 are the relative molecular masses of MMA and BA, respectively; f 1 0 and f 1 are the mole fractions of MMA in the monomer phase at the beginning and after a certain conversion of the polymerization, respectively; f 2 0 and f 2 are the mole fractions of BA at different stages; C is a monomer conversion; α, β, γ, and δ are coefficients relating to the properties of monomers, which are shown in Table 2. Table 2. Formulas and values of α, β, γ, and δ.
Note: r 1 and r 2 are reactivity ratios of MMA and BA (shown in Table 4), respectively.
The relative variation tendencies of x and 1-x could be seen in Figure 2. As the amounts of initiator and αMSD in the polymerization process were very small, their proportions in the polymer were ignored. It is evident that x increased gradually with increasing amounts of αMSD. This indicated that αMSD affected not only the molecular weights of polymers, but also its composition. To analyze the potential reason, the reactivity ratios were calculated according to a semiempirical formula, equation Q-e (Equations (4) and (5)).
where Q and e represent the conjugation degree and polarity of monomer substituents, respectively. ra and rb are the reactivity ratios of monomers a and b, respectively. In the case of ra, a higher value means a larger self-polymerization trend. Alternatively, a lower value indicates a larger copolymerization trend. During the calculation, the Q and e values for αMSD were estimated using Q and e of alpha methyl styrene (AMS), as αMSD was a dimer of AMS. The specific Q and e values of AMS, MMA, and BA were listed in Table 3, and the calculated reactivity ratios of the three monomers were given in Table 4.  where r1 and r3 are the reactivity ratios for MMA to BA and AMS; r2 and r6 are the reactivity ratios for BA to MMA and AMS; and r4 and r5 are the reactivity ratios for AMS to MMA and BA, respectively. As given in Table 4, r1 > r2 revealed that MMA had a greater polymerization activity than BA, which contributed to a higher x (>0.5) in the early polymerization period. For the sample without αMSD addition (Run 1), BA's mass fraction in the polymers gradually increased up to 50% in the later polymerization period. However, this process would be terminated earlier in the presence of αMSD. Because of the chain transfer reaction of αMSD, the propagation process was terminated significantly earlier with an increase in the amount of αMSD. As a result, the growth of BA's mass To analyze the potential reason, the reactivity ratios were calculated according to a semi-empirical formula, equation Q-e (Equations (4) and (5)).
where Q and e represent the conjugation degree and polarity of monomer substituents, respectively. r a and r b are the reactivity ratios of monomers a and b, respectively. In the case of r a , a higher value means a larger self-polymerization trend. Alternatively, a lower value indicates a larger copolymerization trend. During the calculation, the Q and e values for αMSD were estimated using Q and e of alpha methyl styrene (AMS), as αMSD was a dimer of AMS. The specific Q and e values of AMS, MMA, and BA were listed in Table 3, and the calculated reactivity ratios of the three monomers were given in Table 4. Where r 1 and r 3 are the reactivity ratios for MMA to BA and AMS; r 2 and r 6 are the reactivity ratios for BA to MMA and AMS; and r 4 and r 5 are the reactivity ratios for AMS to MMA and BA, respectively.
As given in Table 4, r 1 > r 2 revealed that MMA had a greater polymerization activity than BA, which contributed to a higher x (>0.5) in the early polymerization period. For the sample without αMSD addition (Run 1), BA's mass fraction in the polymers gradually increased up to 50% in the later polymerization period. However, this process would be terminated earlier in the presence of αMSD. Because of the chain transfer reaction of αMSD, the propagation process was terminated significantly earlier with an increase in the amount of αMSD. As a result, the growth of BA's mass fraction in the Polymers 2020, 12, 80 7 of 14 polymer was terminated earlier and earlier. Finally, MMA's mass fraction in the polymer increased, whereas the mass fraction of BA was lowered with an increase in the amount of αMSD.
In addition, r 1 > r 3 indicated that MMA had a greater tendency to copolymerize with AMS than BA. Similarly, r 2 > r 6 indicated that BA tended to copolymerize with AMS rather than MMA. In other words, both MMA and BA were more likely to copolymerize with AMS (αMSD). This could indirectly explain why αMSD effectively controlled the molecular weights of the polymer. For AMS, r 4 ≈ r 5 , there was an almost equal possibility of AMS to copolymerize with MMA or BA.
To compare the regulation reactivity differences of αMSD as a chain transfer agent to MMA and BA more accurately, its transfer coefficients (C tr ) to each single monomer were calculated with the following expression (Equation (6)). It was derived by Alfrey [39] for the case of binary copolymerization in the presence of small amounts of CTA, based on the assumption that the only effect of the chain transfer reaction was to terminate one growing molecular chain and start another, i.e., ignoring cases in which a chain transfer also leaded to the termination of the kinetic chain.
where A and B are mole fractions of monomer A and monomer B, respectively, in the polymer; [A] and [B] are the mole fractions of monomer A and monomer B in the whole monomers; r a and r b are the reactivity of the two monomers; and K is the slope of a Mayo-like plot (i.e., 1/X n vs. total monomer concentration, Equation (7)) [40,41]. 1 where [C] and [M] are concentrations of the CTA and whole monomers, respectively; X n is the average degree of polymerization for the polymer at a given [C]; and X n,0 is for the polymer in the absence of CTA. In this work, MMA and BA were regarded as monomers A and B, respectively, as αMSD was used as the CTA. On the basis of the above analysis, the values of the individual transfer coefficients (C tr,a and C tr,b ) for a particular monomer ratio were successfully obtained by the two simultaneous equations (Equations (6) and (7)). Figure 3 displayed the fitting result of the Mayo plot, and K was derived from the slope of the plot: 0.56 (R-Square = 0.986). K remained unchanged as long as the monomer ratio of MMA and BA was a constant. As a consequence, C tr,a and C tr,b could be calculated through Equation (6) by choosing any two runs (including αMSD) in Table 1, such as Run 2 and Run 4, and the result was shown in Table 5.
Polymers 2020, 12, x FOR PEER REVIEW 7 of 14 fraction in the polymer was terminated earlier and earlier. Finally, MMA's mass fraction in the polymer increased, whereas the mass fraction of BA was lowered with an increase in the amount of αMSD.
In addition, r1 > r3 indicated that MMA had a greater tendency to copolymerize with AMS than BA. Similarly, r2 > r6 indicated that BA tended to copolymerize with AMS rather than MMA. In other words, both MMA and BA were more likely to copolymerize with AMS (αMSD). This could indirectly explain why αMSD effectively controlled the molecular weights of the polymer. For AMS, r4 ≈ r5, there was an almost equal possibility of AMS to copolymerize with MMA or BA.
To compare the regulation reactivity differences of αMSD as a chain transfer agent to MMA and BA more accurately, its transfer coefficients (Ctr) to each single monomer were calculated with the following expression (Equation (6)). It was derived by Alfrey [39] for the case of binary copolymerization in the presence of small amounts of CTA, based on the assumption that the only effect of the chain transfer reaction was to terminate one growing molecular chain and start another, i.e., ignoring cases in which a chain transfer also leaded to the termination of the kinetic chain.
where A and B are mole fractions of monomer A and monomer B, respectively, in the polymer; [A] and [B] are the mole fractions of monomer A and monomer B in the whole monomers; ra and rb are the reactivity of the two monomers; and K is the slope of a Mayo-like plot (i.e., 1/ n X vs. total monomer concentration, Equation (7)) [40,41]. In this work, MMA and BA were regarded as monomers A and B, respectively, as αMSD was used as the CTA. On the basis of the above analysis, the values of the individual transfer coefficients (Ctr,a and Ctr,b) for a particular monomer ratio were successfully obtained by the two simultaneous equations (Equations (6) and (7)). Figure 3 displayed the fitting result of the Mayo plot, and K was derived from the slope of the plot: 0.56 (R-Square = 0.986). K remained unchanged as long as the monomer ratio of MMA and BA was a constant. As a consequence, Ctr,a and Ctr,b could be calculated through Equation (6) by choosing any two runs (including αMSD) in Table 1, such as Run 2 and Run 4, and the result was shown in Table 5. [ ] ,  If this polymerization system was treated as a terpolymerization of MMA, BA, and αMSD, the two C tr s of αMSD could be calculated by the reactivity ratios of the three monomers on the basis of Q-e values. According to the definition of C tr , the C tr s values of αMSD should be kinetically equivalent to the reciprocals of the reactivity ratios of any two monomers. The calculation results were shown in Table 5.
As shown above, the C tr s values of αMSD determined by different methods were obviously different. C tr,MMA and C tr,BA were 0.62 and 0.47 according to Alfrey's theory, respectively. However, when using Q-e method, the former was 2.12 and the latter was increased to 14.28. The difference could be explained from two aspects: First, the C tr calculations based on the Q-e method utilized AMS's reactivity ratios to represent αMSD's. Nevertheless, as a dimer of AMS, αMSD's structure was more complex than AMS, and the former had more steric effects when it reacted with other monomers. Thus, their reactivity ratios as well as C tr s should be different. Second, there were some inherent errors in calculating the reactivity ratios using the Q-e method.
In contrast to the Q-e method, the substitution of AMS for αMSD was avoided in Alfrey's theory, so the C tr s values of αMSD should be more accurate. In addition, the almost same values of C tr,MMA and C tr,BA indicated that αMSD has almost same chain transfer possibilities with both MMA and BA, which was consistent with the above analysis of reactivity ratios in Table 4.
As described above, few studies have been conducted to use αMSD as a chain transfer agent in the emulsion polymerization. Therefore, we have used NMR analysis to reveal its potential mechanism. In the polymerization process, chain propagation would be terminated by the irreversible AFCT agent. Therefore, a part of the polymer's terminal structures consists of the AFCT agent. On this basis, the mechanism for αMSD could be revealed by a polymer terminal structure analysis. The 1 H NMR analysis of the resulting polymer using 1% αMSD showed that two signals of similar intensities at approximately 5.0, 5.1, and 5.6 ppm were observed as the characteristic peaks of the two syntonic hydrogen atoms of the terminal double bonds (Figure 4). Moreover, multiple signals at~7.0 ppm were present, which were ascribed to the substituted benzene ring from the αMSD. Accordingly, these results confirmed that emulsion polymerization of acrylates using αMSD as a chain transfer agent followed Yasummasa's mechanism. The detailed regulation mechanism is illustrated in Scheme 4.
Polymers 2020, 12, x FOR PEER REVIEW 8 of 14 If this polymerization system was treated as a terpolymerization of MMA, BA, and αMSD, the two Ctrs of αMSD could be calculated by the reactivity ratios of the three monomers on the basis of Q-e values. According to the definition of Ctr, the Ctrs values of αMSD should be kinetically equivalent to the reciprocals of the reactivity ratios of any two monomers. The calculation results were shown in Table 5. As shown above, the Ctrs values of αMSD determined by different methods were obviously different. Ctr,MMA and Ctr,BA were 0.62 and 0.47 according to Alfrey's theory, respectively. However, when using Q-e method, the former was 2.12 and the latter was increased to 14.28. The difference could be explained from two aspects: First, the Ctr calculations based on the Q-e method utilized AMS's reactivity ratios to represent αMSD's. Nevertheless, as a dimer of AMS, αMSD's structure was more complex than AMS, and the former had more steric effects when it reacted with other monomers. Thus, their reactivity ratios as well as Ctrs should be different. Second, there were some inherent errors in calculating the reactivity ratios using the Q-e method.
In contrast to the Q-e method, the substitution of AMS for αMSD was avoided in Alfrey's theory, so the Ctrs values of αMSD should be more accurate. In addition, the almost same values of Ctr,MMA and Ctr,BA indicated that αMSD has almost same chain transfer possibilities with both MMA and BA, which was consistent with the above analysis of reactivity ratios in Table 4.
As described above, few studies have been conducted to use αMSD as a chain transfer agent in the emulsion polymerization. Therefore, we have used NMR analysis to reveal its potential mechanism. In the polymerization process, chain propagation would be terminated by the irreversible AFCT agent. Therefore, a part of the polymer's terminal structures consists of the AFCT agent. On this basis, the mechanism for αMSD could be revealed by a polymer terminal structure analysis. The 1 H NMR analysis of the resulting polymer using 1% αMSD showed that two signals of similar intensities at approximately 5.0, 5.1, and 5.6 ppm were observed as the characteristic peaks of the two syntonic hydrogen atoms of the terminal double bonds (Figure 4). Moreover, multiple signals at ~7.0 ppm were present, which were ascribed to the substituted benzene ring from the αMSD. Accordingly, these results confirmed that emulsion polymerization of acrylates using αMSD as a chain transfer agent followed Yasummasa's mechanism. The detailed regulation mechanism is illustrated in Scheme 4.

Effect of αMSD on the Polymerization Rate
As αMSD has an obvious influence on polymer molecular weights, it was necessary to study if its presence would also affect the polymerization rate (Rp). In the emulsion polymerization, the constant speed stage was typically used as the indicator of total polymerization rate, and it could be calculated from the conversion-time curve. Figure 5 showed the monomer conversion versus reaction time in the presence of αMSD at various concentrations. The slopes of the curves directly reflected Rp. It was clear that the presence of αMSD could decrease the Rp as compared to samples without αMSD. Moreover, as the increase in the quantities of αMSD, Rp was decreased more significantly. For example, the monomer conversion only reached 0.4 after 250 min when 3% of αMSD was added, whereas its conversion approached 100% without αMSD addition. Furthermore, the effect of αMSD concentration ([αMSD]) on Rp was further correlated via constructing a log-log plot of Rp and [αMSD], as shown in Figure 6. As expected, the result was quantitatively estimated to be Rp∝[αMSD] −0.945 (R-Square = 0.905).  Figure 5. Monomer conversion as a function of reaction time at various molar ratios between αMSD and acrylate monomers (see Table 1).

Scheme 4.
Chain transfer mechanism of αMSD in MMA and BA emulsion copolymerization.

Effect of αMSD on the Polymerization Rate
As αMSD has an obvious influence on polymer molecular weights, it was necessary to study if its presence would also affect the polymerization rate (R p ). In the emulsion polymerization, the constant speed stage was typically used as the indicator of total polymerization rate, and it could be calculated from the conversion-time curve. Figure 5 showed the monomer conversion versus reaction time in the presence of αMSD at various concentrations. The slopes of the curves directly reflected R p . It was clear that the presence of αMSD could decrease the R p as compared to samples without αMSD. Moreover, as the increase in the quantities of αMSD, R p was decreased more significantly. For example, the monomer conversion only reached 0.4 after 250 min when 3% of αMSD was added, whereas its conversion approached 100% without αMSD addition. Furthermore, the effect of αMSD concentration ([αMSD]) on R p was further correlated via constructing a log-log plot of R p and [αMSD], as shown in Figure 6. As expected, the result was quantitatively estimated to be R p ∝[αMSD] −0.945 (R-Square = 0.905).

Effect of αMSD on the Polymerization Rate
As αMSD has an obvious influence on polymer molecular weights, it was necessary to study if its presence would also affect the polymerization rate (Rp). In the emulsion polymerization, the constant speed stage was typically used as the indicator of total polymerization rate, and it could be calculated from the conversion-time curve. Figure 5 Figure 5. Monomer conversion as a function of reaction time at various molar ratios between αMSD and acrylate monomers (see Table 1). Figure 5. Monomer conversion as a function of reaction time at various molar ratios between αMSD and acrylate monomers (see Table 1). Polymers 2020, 12, x FOR PEER REVIEW 10 of 14 The potential reasons for the decrease in Rp in the presence of αMSD may be explained from two aspects. First of all, the overall rate constant of the scission process of the leaving group may be lower than propagation [20]. The efficient transfer reaction of an irreversible AFCT agent required that transitional radicals, created by the addition reaction of the irreversible AFCT agent and chains radicals, could easily fragment to generate new radicals to reinitiate the polymerization process. As a result, retardation would happen. In fact, the novel radical (isopropyl benzene radical, seen in Scheme 4) in this paper, had less access to reinitiate polymerization than the primary radical (sulfate radical anion) due to its steric hindrance. Thus, the macroscopic reaction rate decreased. Second, radicals derived from both the initiator and leaving groups might escape from micelles or latex particles (desorption) and then be terminated with radicals in either the water phase or polymer particle (reabsorption). This desorption phenomenon had been successfully verified in the emulsion polymerization, and the related theoretical model had been previously established [42][43][44]. The sulfate radical anion is more hydrophilic than the isopropyl benzene radical. As a result, the desorption number of the former was much greater than latter, and the latter tended to stay in the original particles. This led to serious loss of the high-active free radicals, and in addition, the remaining free radicals in the latex particles were not active enough, which further exacerbated the decline in Rp.

Effect of αMSD Particle Size and Relevant Parameters
The influence of αMSD concentration on latex particle size(dp), its distribution (PDI), and particle number (Nc) were also explored, and the results were shown in Table 6. The particle number of the resulting emulsion was calculated using an Equation (8).
where m is the solid content of the dispersed phase (containing monomer and polymer at a given conversion and expressed in g/mL −1 ), and ρ represents the particle density. As shown in Table 6, dp declined from 109.38 to 46.88 nm as αMSD dosage was varied from 0% to 3%, whereas Nc was enhanced from 1.67 × 10 17 to 20.79 × 10 17 mL −1 . Moreover, PDI started to dramatically increase when the αMSD dosage was higher than 1%.
According to recent reports [45,46], the exit and entry rates of radicals and monomers into particles have strong effects on miniemulsion polymerization kinetics. In this study, isopropyl benzene radicals derived from αMSD increased the whole radical concentration in the particles, which contributed to the greater desorption of sulfate radical anions. Thus, these sulfate radical anions were resorbed by other micelles, and more micelles could be transformed into polymer particles. This explained the increase in Nc and the decrease in dp. Generally speaking, the increase in The potential reasons for the decrease in R p in the presence of αMSD may be explained from two aspects. First of all, the overall rate constant of the scission process of the leaving group may be lower than propagation [20]. The efficient transfer reaction of an irreversible AFCT agent required that transitional radicals, created by the addition reaction of the irreversible AFCT agent and chains radicals, could easily fragment to generate new radicals to reinitiate the polymerization process. As a result, retardation would happen. In fact, the novel radical (isopropyl benzene radical, seen in Scheme 4) in this paper, had less access to reinitiate polymerization than the primary radical (sulfate radical anion) due to its steric hindrance. Thus, the macroscopic reaction rate decreased. Second, radicals derived from both the initiator and leaving groups might escape from micelles or latex particles (desorption) and then be terminated with radicals in either the water phase or polymer particle (reabsorption). This desorption phenomenon had been successfully verified in the emulsion polymerization, and the related theoretical model had been previously established [42][43][44]. The sulfate radical anion is more hydrophilic than the isopropyl benzene radical. As a result, the desorption number of the former was much greater than latter, and the latter tended to stay in the original particles. This led to serious loss of the high-active free radicals, and in addition, the remaining free radicals in the latex particles were not active enough, which further exacerbated the decline in R p .

Effect of αMSD Particle Size and Relevant Parameters
The influence of αMSD concentration on latex particle size(d p ), its distribution (PDI), and particle number (N c ) were also explored, and the results were shown in Table 6. The particle number of the resulting emulsion was calculated using an Equation (8).
where m is the solid content of the dispersed phase (containing monomer and polymer at a given conversion and expressed in g/mL −1 ), and ρ represents the particle density. As shown in Table 6, d p declined from 109.38 to 46.88 nm as αMSD dosage was varied from 0% to 3%, whereas N c was enhanced from 1.67 × 10 17 to 20.79 × 10 17 mL −1 . Moreover, PDI started to dramatically increase when the αMSD dosage was higher than 1%.
According to recent reports [45,46], the exit and entry rates of radicals and monomers into particles have strong effects on miniemulsion polymerization kinetics. In this study, isopropyl benzene radicals derived from αMSD increased the whole radical concentration in the particles, which contributed to the greater desorption of sulfate radical anions. Thus, these sulfate radical anions were resorbed by other micelles, and more micelles could be transformed into polymer particles. This explained the increase in N c and the decrease in d p . Generally speaking, the increase in N c should be accompanied by inhibition of the termination reaction, leading to a higher reaction rate. However, the actual R p was decreased. In fact, actual variations of N c and d p were not necessarily related to R p . A mass transfer resistance occurred during the transfer process of sulfate radical anions among the micelles, water, and particles. This would counteract the slight increase in R p caused by the lower degree of termination. As a result, the variations of N c and d p failed to improve R p . Table 6. Collection of latex particle size (d p ) and its distribution (PDI) and particle number (N c ) with αMSD at different concentrations.

The Effect of αMSD on Tensile Properties of Polymer Films
The tensile properties (E, σ p , and ε b ) of the latex films prepared from polyacrylate emulsion at various αMSD concentrations were determined, and the stress-strain curves were shown in Figure 7. The results were summarized in Table 7. As expected, αMSD addition had an impact on the mechanical properties of the latex film. With the increase in αMSD from 0% to 1%, E and σ p of the resultant films were reduced by nearly 75% and 50%, respectively, whereas ε b was only slightly affected. Furthermore, the obvious yield point of the resultant polymers gradually became less obvious when the amount of αMSD was above 0.3%, indicating that the polymer films changed from hard to soft [47]. When the amount of αMSD reached to 3%, E and ε b were only 2.00 and 0.23 MPa, respectively, and ε b had decreased to 467%. The toughness of the resultant films was significantly decreased, and the film was soft and weak. The decrease in the molecular weights of polymers was the main reason to lead to the poor mechanical properties.
Polymers 2020, 12, x FOR PEER REVIEW 11 of 14 Nc should be accompanied by inhibition of the termination reaction, leading to a higher reaction rate. However, the actual Rp was decreased. In fact, actual variations of Nc and dp were not necessarily related to Rp. A mass transfer resistance occurred during the transfer process of sulfate radical anions among the micelles, water, and particles. This would counteract the slight increase in Rp caused by the lower degree of termination. As a result, the variations of Nc and dp failed to improve Rp. Table 6. Collection of latex particle size (dp) and its distribution (PDI) and particle number (Nc) with αMSD at different concentrations.

The Effect of αMSD on Tensile Properties of Polymer Films
The tensile properties (E, σp, and εb) of the latex films prepared from polyacrylate emulsion at various αMSD concentrations were determined, and the stress-strain curves were shown in Figure  7. The results were summarized in Table 7. As expected, αMSD addition had an impact on the mechanical properties of the latex film. With the increase in αMSD from 0% to 1%, E and σp of the resultant films were reduced by nearly 75% and 50%, respectively, whereas εb was only slightly affected. Furthermore, the obvious yield point of the resultant polymers gradually became less obvious when the amount of αMSD was above 0.3%, indicating that the polymer films changed from hard to soft [47]. When the amount of αMSD reached to 3%, E and εb were only 2.00 and 0.23 MPa, respectively, and εb had decreased to 467%. The toughness of the resultant films was significantly decreased, and the film was soft and weak. The decrease in the molecular weights of polymers was the main reason to lead to the poor mechanical properties.

Conclusions
In this study, it was shown that αMSD was a highly effective CTA to control the molecular weights in MMA and BA emulsion copolymerization. Molecular masses and their distributions, as well as the mass proportion of MMA in the polymers, were gradually decreased when increasing the dosage of αMSD. Meanwhile, the final monomer conversion was also declined as the αMSD concentration increased above 1%, indicating its apparent inhibition ability of the polymerization. The chain transfer constant of αMSD to MMA was approximately the same with αMSD to BA. The αMSD regulation mechanism was verified by NMR analysis, which was consistent with Yasummasa's theory. In addition, the R p and d p of the acrylate latex were decreased as the increase in the dosage of αMSD, while the N c of the emulsion was increased. Furthermore, the PDI was also affected when αMSD dosage was more than 1%. Finally, the decrease of the tensile properties of the resulting polyacrylate films was mainly due to the reduction in the molecular weights of synthesized polymers after the addition of αMSD.