Investigation of the effect of some variables on terpolymerization process of vinyl monomers in CSTR by design of experimental method

1 Introduction

The polymerization process of acrylic monomers that have been successfully applied for production of acrylic homo- and copolymers is basically, dependent on the polymerization method, the chemical nature of comonomers and their contents. Solution and slurry processes are the most popular methods for polymerization based on acrylonitrile (AN) (1). Solution polymerization usually results in lower molecular weight because the organic solvent often acts as chain transfer agent. In slurry polymerization using deionized water (H₂O) with zero chain transfer constant, higher molecular weight polymer would be obtained. Ease of heat removal with temperature control, low medium viscosity and low level of impurities in the polymer product is considered as the advantages of this method (2). In addition, polymerization of acrylic monomers in the aqueous medium initiated by redox initiators has superiority in initiating efficiency at low temperature and high molecular weight of resultant polymers (3), (4), (5).

In the past decades, effects of various comonomers such as methyl acrylate, methyl methacrylate, acrylic acid, itaconic acid on synthesis and properties of AN-based copolymers produced via free radical polymerization in aqueous medium have been investigated by some researchers (6), (7), (8), (9). Despite these researchers, few authors have paid their attention to slurry copolymerization of AN and vinyl acetate (VAc). Cetiner et al. (10) studied the free radical copolymerization of AN and VAc with different feed ratios of VAc (wt%) by using ammonium persulfate as an initiator in the aqueous medium. The results showed that the conversion degree of the copolymer is lower than that of homopolymerization of AN, and the intrinsic viscosity of copolymer decreased as the VAc content increased.

The effect of pH value on the copolymerization of AN and VAc in a pilot-plant reactor was studied by Yao et al. (11). The results showed that the conversion degree increased with increasing pH value. The molecular weight of polymers produced at higher pH values was higher than those produced at lower pH value.

Marlet and coworkers (12) studied the experimental approach and statistic modeling in the copolymerization and terpolymerization of AN with VAc and methyl-2 propene-1 sodium sulfonate using the design of experimental method.

The aim of the present work is to study the simultaneous influence of three variables, including VAc concentration (wt%) in monomer mixture VAc/(VAc+AN), bisulfite- to-persulfate ratio in redox initiator system ([HSO3−][S2O8−2]) and bisulfite-to-monomer mixture ratio ([HSO3−]AN+VAc) on the conversion percentage of the monomers to polymer, intrinsic viscosity [(η)] and sulfur end groups (SEG) index of the synthesized polymers.

The effects of these variables are examined using response surface method (RSM) based on central composite design (CCD). RSM is applicable to optimization and modeling of many processes in which quadratic polynomial equations are suitable in explaining the effects of factors on responses (13).

2 Experimental

2.3 Polymer preparation procedure

Free radical slurry terpolymerization of AN with VAc and a constant amount of AMPS was carried out in a 15-l continuous stirred tank reactor (CSTR) shown in Figure 1. At first, two-thirds of reactor volume (10 l) was filled with DM water, and the temperature was raised to 60°C under atmospheric pressure. DM water as the polymerization medium and aqueous solutions of potassium persulfate and sodium metabisulfite as redox initiator system with given concentrations and recipe were continuously fed into the reactor for 30 min to create radicals needed to start the reaction (see the initial flow in Table 3). Agitation speed was 300–350 rpm, and during this time, sulfuric acid was also separately fed into the reactor to adjusting a constant pH value in the range of 2.7–3. In the second stage, the given monomer mixture with various weight ratios of AN and VAc along with an aqueous solution of AMPS comonomer was continuously fed into the reactor to start the polymerization reaction (see the secondary flow in Table 3). The feed rates of these reactants and water as shown in Table 3 were adjusted so that mean residence time in the reactor was 60 min and the water-to-monomer ratio was 1.9 approximately.

Figure 1:

Pilot plant reactor for acrylic polymer synthesis.

Table 3:

Feed rates of different flows.

Substance	Initial flow (g/h)	Secondary flow (g/h)
DM water	2400	8200
Sodium metabisulfite (20 wt% aqueous solution)	150–300	450–1100
Potassium persulfate (3 wt% aqueous solution)	200–250	350–600
Sulfuric acid (2 wt% aqueous solution)	300–400	200–300
(AN+VAc) mixture	–	4300
AMPS (3 wt% aqueous solution)	–	500

Reactant slurry flows out from top of the reactor and enters a termination tank. The resulting polymer was then mixed with ethylenediaminetetraacetic acid (EDTA) in order to stop the redox reaction and subsequently polymerization reaction. After 240 min of reaction time (four mean residence time), samples were taken, and the prepared polymers were purified by washing with deionized water and being filtered. The samples then dried in a vacuum oven at 70°C to a constant weight to remove the residual monomer and water. Eighteen polymer samples are synthesized at the same condition according to Table 2 in this study.

3 Characterization

4 Results and discussion

The experimental results of 18 runs with four replications at the center point has been shown in Table 4. A standard analysis of variance (ANOVA) was carried out to analyze the response surface models. ANOVA results of the quadratic models in Table 5 indicated that the model equations derived by RSM could adequately be used to describe the conversion percentage of the monomers to polymer, intrinsic viscosity [(η)] and SEG index.

The quality of fitting the experimental data with the polynomial model was expressed by the coefficient of determination (R²). A high R² value, close to 1, is desirable, and a reasonable agreement with adjusted R² is necessary. For the models, there was a good coefficient of determination (R²), which means that the quadratic polynomial models were appropriate to represent the actual relationship between responses and significant variables (15).

The importance of each coefficient was determined by the p-values; thus, smaller p-values corresponded to the more important coefficients. Data given in Table 5 show that the models were significant at the 5% confidence level since p-values were less than 0.05. Comparatively, high F-value indicates a significant effect on the response variables.

Figures 2–4 show the three-dimensional response surface plots as a function of the VAc concentration (wt%), bisulfite-to-persulfate ratio ([HSO3−][S2O8−2]) and bisulfite-to- monomer mixture ratio ([HSO3−]AN+VAc) and their interaction on the conversion percentage of monomers to polymer, intrinsic viscosity [(η)] and SEG index.

Figure 2:

Conversion as a function of VAc concentration, bisulfite-to-persulfate ratio and bisulfite-to-monomer ratio.

Figure 3:

Intrinsic viscosity as a function of VAc concentration, bisulfite-to-persulfate ratio and bisulfite-to-monomer ratio.

Figure 4:

Sulfur end groups as a function of VAc concentration, bisulfite-to-persulfate ratio and bisulfite-to-monomer ratio.

As it is seen in Figure 2A and B, the conversion percentage decreases with increase in the VAc concentration (wt%). It could be explained regarding the reactivity ratios of the AN-VAc monomer pair (r₁=4.05 and r₂=0.061), where the AN radical chain ends tendency to react with an AN monomer unit is much more than reacting with a VAc monomer unit. At the same time, a VAc radical chain end reacts much faster with an AN monomer unit than with another VAc monomer unit. Therefore, the incorporation probability of VAc units into the molecular chains is very lower than the AN units. Thus, in the various ratios of bisulfite-to-persulfate and bisulfite-to-monomer mixture (whether in low ratios or high), increasing the VAc concentration results in reduction conversion percentage. Furthermore, in a constant amount of VAc concentration, decreasing the amount of bisulfite-to-persulfate ratio (Figure 2A) and increasing the amount of bisulfite-to-monomer mixture ratio (Figure 2B) cause enhancement of conversion percentage, which can be attributed to the increasing the number of anionic radicals produced by the redox initiator system. The higher ratio of bisulfite-to-monomer mixture and the lower ratio of bisulfite-to-persulfate mean that the number of both sulfonate and sulfate radicals produced in the redox initiator system becomes more and the initiation step of reaction occurs faster; hence, chain growth and the conversion percentage of monomers to polymer increase, as shown in Figure 2C.

As seen in Figure 3A and B, the intrinsic viscosity of the samples increases with increase in the VAc concentration. This could be due to the transfer reactions of growing chains to the polymer containing VAc primarily at high conversion, where the concentration of polymer is high. Moreover, with increasing bisulfite-to-monomer ratio and decreasing the bisulfite-to-persulfate ratio (equivalent an increase in bisulfite ion), the intrinsic viscosity reduced as seen in Figure 3B and C. This is because the bisulfite ion ( HSO3−1 ), in addition to the reducing agent in redox initiator system, acts as a chain transfer agent that terminates the chain with hydrogen transfer while starting another with a sulfonate radical according to the following reactions that cause to reduce the molecular weight of the polymer.

The dye sites are sulfate and sulfonate end groups arising from the persulfate-bisulfite redox initiating system and also from sulfonate groups of AMPS monomer. Bisulfite ion is considered as a chain transfer agent. Chain transfer to bisulfite, however, terminates one chain with a hydrogen atom while starting another with a sulfonate radical. This increases the total dye site content of the polymer by reducing the polymer molecular weight. But simultaneously, this reaction produces chains with just one dye site. At the constant amount of AMPS monomer, by increasing the amount of bisulfite-to-monomer ratio, chain transfer reactions of bisulfite ion and subsequently total dye site of the polymer increase, as shown in Figure 4B and C. Also, as shown in Figure 4A and B, in the low values of the bisulfite-to-persulfate ratio and bisulfite-to-monomer ratio, increasing the concentration of VAc leads to an increase in the SEG.

Table 6 shows the solutions of optimum conditions with their responses obtained by optimization function of the design-expert software. Since VAc is less expensive than acrylic monomers, the optimum conditions were chosen in order to use the maximum amount of VAc in the polymerization process. It is worth mentioning that the optimum amounts of variables are determined by the requirements to produce polymer for dry spinning and fiber production, which are 75–85% monomer conversion, 1.35–1.4 dl/g intrinsic viscosity and 180–190 ppm SEG (see Table 6).

4.1 FTIR spectra analysis

The FTIR spectra of the terpolymer are shown in Figure 5, which displayed typical characteristics of the absorption band for PAN at 2243 cm⁻¹, corresponding to the stretching vibration mode of the nitrile groups for AN. The absorption bands at 2938 are assigned to the stretching vibrations of C-H in the main chain, whereas the absorption band of another mode of C-H vibration appeared at around 1451 cm⁻¹. The C=O stretching, C-O-C stretching and C-O stretching vibration peaks can be observed at 1736, 1232 and 1022 cm⁻¹, respectively, for VAc. The bands at 3437 cm⁻¹ (N-H stretching), 1670 (amide I band), 1232 cm⁻¹ (S=O stretching), 1042 cm⁻¹ (symmetrical O=S=O), 627 cm⁻¹ (C-S stretching) and 455 cm⁻¹ (C-N-C in amide group) are assigned to the structural peaks of AMPS core component (10).

Figure 5:

FTIR spectra of optimum sample.

4.2 Nuclear magnetic resonance (NMR) studies

The copolymer composition was determined by H-NMR spectra. The H-NMR spectrum of poly(AN-co-VAc) is shown in Figure 6. The resonance at around 2 ppm is assigned to the methylene protons of monomers in the main chain (proton a, in Figure 6). The signals of -CH proton of AN and VAc monomers occur at 3.0–3.2 and 5.13 ppm, respectively (protons b and c in Figure 6). The calculated copolymer compositions for almost 93 (mol%) AN and 7 (mol%) vinylacetate in feed were about 95.86 (mol%) AN and 4.14 (mol%) vinylacetate in copolymer chain. Because of fewer amount of AMPS comonomer against AN and VAc monomers in feed, there is not any clear peak attribute to this monomer in H-NMR spectra (16).

Figure 6:

H-NMR spectra of optimum sample.

4.4 XRD analysis

The XRD pattern of the terpolymer is shown in Figure 7. The strongest diffraction peak at about 2θ=17°, corresponding to the interplanar spacing of 0.53 nm that could be attributed to the (100) crystalline planes of the pseudohexagonal cell or (200) reflections of the orthorhombic structure of terpolymer. At the same time, a weak diffraction peak, present at about 2θ=29° (d=0.30 nm), is attributed to the (101) crystalline plane of the pseudohexagonal cell or (201) reflections of the orthorhombic structure. Both peaks are broad and weak, indicating the PAN homopolymer and terpolymers poorly crystallize.

Figure 7:

XRD curve of the optimum sample measured under room temperature.

A few amounts of VAc comonomer introduced into the PAN copolymers and terpolymers could block interactions between intermolecular C≡N groups and enhance the activity of PAN chain segments, which reduce the crystallinity of copolymer (17).

4.5 DSC study

Figure 8 shows the DSC curve of polymer heated at 10°C min⁻¹ from ambient temperature to 330°C. The exothermic peak at around 300°C was attributed to cyclization, dehydrogenation and other elimination reactions occurring in the temperature range. Among these reactions, the cyclization reaction was the most important one, which took place in nitrile groups and led to the development of ladder structures. Before the exothermic peak, an endothermic peak appears at around 290°C, which is probably due to the interactions between cyanide groups of AN and sulfonate groups of AMPS (6), (18).

Figure 8:

DSC curve of optimum sample.

4.6 SEM study

Particle morphology is largely determined by the way they are formed. The firstly formed primary particles (0.05–0.2 μm) can either absorb radicals or coagulate with other unstable particles in the aqueous polymerization. As they grow larger, their larger surface areas enable them to absorb more radicals and grow faster. Therefore, little-limited coagulation occurred between the secondary (0.5–3 μm) and the larger particles (4–50 μm), which mainly grow by surface polymerization (Figure 9).

Figure 9:

SEM images of the resulting particles.

5 Conclusions

To investigate the effects of VAc concentration (wt%) in monomer mixture, bisulfite-to-persulfate ratio in redox initiator system ([HSO3−][S2O8−2]) and bisulfite-to-monomer mixture ratio ([HSO3−]AN+VAc), slurry copolymerization of AN with vinylacetate and AMPS are carried out continuously in a 15-l CSTR pilot plant reactor by RSM method. Experimental data show that the conversion percentage of monomers to the polymer and the intrinsic viscosity of the prepared polymers were influenced considerably by VAc concentration and bisulfite-to-monomers ratio. Moreover, it has been found that the SEG index of the samples largely depended on the bisulfite-to-monomers and bisulfite-to-persulfate ratio. The optimum conditions derived via RSM for this polymerization were VAc=9 wt% (in AN+VAc mixture), bisulfite-to-persulfate ratio=9.6 and bisulfite-to-monomer mixture ratio=0.027. The amounts of conversion percentage of monomers to polymer, intrinsic viscosity and SEG under optimum conditions were 75%, 1.38 dl/g and 190, respectively. According to the obtained results, the very good adaptation between the experimental data and the predicted values was observed. FTIR, NMR and EA results confirmed the presence of vinylacetate comonomer in terpolymer chain. The XRD result indicated that terpolymer is semicrystalline, and the DSC analysis indicates that cyclization reaction occurs at high temperature. SEM images from sample powders showed particles have smooth surface and size distribution in the range of 10–50 μm.