Enhanced melt free radical grafting efficiency of polyethylene using a novel redox initiation method

Abstract

In this work, for the first time, redox initiation was employed in the melt free radical grafting of glycidyl methacrylate (GMA) and maleic anhydride (MA) onto polyethylene (PE) by reactive extrusion. Since it is very challenging to obtain high grafting degrees of GMA and MA onto polymer backbones through conventional free radical initiation, especially in the extrusion process, the strategy of using a peroxide/reducing agent initiator was introduced. The redox initiation system was composed of dicumyl peroxide (DCP) and Tin(II) 2-ethylhexanoate (Sn(Oct)₂). The grafting reaction was monitored by torque rheometry and graft products were analyzed by Fourier Transform Infrared Spectroscopy. Effects of monomer concentration and DCP/Sn(Oct)₂ ratio on the degree of grafting were studied. The redox initiation proved to be very effective in improving the grafting degree of GMA and MA onto PE. It is reasoned that, in the presence of Sn(Oct)₂, the free radicals produced were not subject to a cage effect, yielding high initiator efficiency.

---

1. Introduction

Reactive polymers containing anhydride groups and epoxy groups are widely used as interfacial modifiers in many immiscible polymer blends and composites. Reactive groups can be incorporated into a polymer through polymerization or copolymerization of a monomer containing the reactive group, or through chemical modification of a preformed polymer. Free radical grafting of vinyl monomers is an important post-polymerization process used to functionalize polymers and expand their applications.^1–3 This process can be performed via solution reaction in batch reactors or by melt reaction in mixers and extruders.⁴ Reactive extrusion is a desirable method for industrial mass production due to being both fast and solvent-free.⁵ Although the free radical grafting reaction of polymers in extrusion has been the focus for mass production of functionalized polymers since 1970–80's,^3,6,7 improving grafting efficiency and achieving high grafting degree is still a challenge. This is illustrated by the inherent mechanism involved in the polymer melt graft reaction. Free radical grafting on polymers begins with the attack of the macromolecules by the primary radicals produced from homolytic decomposition of the initiator, which results in hydrogen abstraction of the macromolecules and formation of macroradicals. Compared with the primary radicals, the macroradicals exhibit significantly reduced accessibility to monomer molecules due to the steric effect of bulky substitutes. On the other hand, macroradicals can undergo side reactions such as crosslinking and chain scissions. For instance, during free radical grafting, polyethylene (PE) macroradicals are susceptible to crosslinking while polypropylene (PP) macroradicals tend to undergo degradation caused by β-scissions.^1,8,9 Further, the high viscosity of polymer melt generally leads to insufficient mixing of the reaction ingredients, especially within the limited residence time in extrusion, which further reduces the grafting efficiency.

Graft reactions of maleic anhydride (MA) and glycidyl methacrylate (GMA) onto polyolefins have received extensive investigations in the literature, but these reactions generally yield low conversions of monomers and equally poor grafting degrees. The steric effect is critical in the case of vinyl monomers with bulky substituent(s). For the grafting of MA, its electropositive effect which favors the attack of nucleophilic macroradicals is overshadowed by its steric hindrance.^10,11 For the grafting of GMA, both steric and electrostatic effects are unfavorable. A “co-monomer” approach was introduced to improve the grafting degree and the monomer conversion rate for MA grafted polyolefins.^4,12–14 Hu et al. evaluated the use of different co-monomers in the grafting of MA onto polypropylene.¹² They noted that the use of styrene (St) as a co-monomer effectively reduced the PP chain degradation and greatly improved the grafting yield of MA onto PP. Later, they also investigated the use of styrene as a co-monomer in the grafting of GMA onto PP and observed that the presence of styrene as a co-monomer also increased the GMA grafting degree with significant reduction in PP chain degradation.⁴ From these works to date, styrene has been used to improve melt grafting reactions of GMA and MA onto polyolefins. The concentration of styrene is often set to be equal to the concentration of the functional monomer, i.e., in a 1 : 1 molar ratio.^15–17 It was reasoned that the co-monomer could either activate the primary monomer (i.e. MA) by forming a charge transfer complex or directly react with the macroradicals first and then add to the primary monomer.^10,15 Styrene was found to be a good co-monomer which not only promoted the melt free radical grafting of both GMA and MA, but also reduced polypropylene chain scission and crosslinking of polyethylene macroradicals.^1,11

Another limiting factor in the free radical grafting reactions, which is often ignored, is the low initiator efficiency. In the case of thermal decomposition, each free radical initiator molecule produces two primary free radicals that are trapped in a solvent cage for some short period of time before diffusing apart.¹⁸ The primary free radicals in cages may initiate monomers, undergo recombination or are wasted by reacting with each other. The latter process is referred to as cage effect and is mainly responsible for the decrease in initiator efficiency. Unfortunately, cage effect has not been investigated in any melt free radical grafting reactions in the literature, which is probably because it is experimentally difficult to monitor such reactions in the melt reaction systems. Nonetheless, cage effect is known to increase with the viscosity of the reaction medium.¹⁹ Therefore, it is reasonable to believe that in melt free radical grafting reactions the cage effect is phenomenal and the initiator efficiency is much lower when compared to solution or bulk reaction. Cage effect is commonly noted in initiations involving thermal and homolytic dissociations of initiators. However, the peroxide/reducing agent redox initiation is an exception to the cage effect, because in the presence of a reducing agent one peroxide initiator molecule decomposes to give one primary free radical. In the literature, the redox initiation is generally used for free radical grafting of polymers or polymerization in solution.^20–24 To the best of our knowledge, redox initiation has not been used in the melt grafting reactions in the literature.

In this work, for the first time, a peroxide/reducing agent redox initiation method was used for the melt free radical grafting of GMA and MA onto PE. The reaction was performed in a co-rotating twin-screw extruder. Dicumyl peroxide was used as a free radical generator and Tin(II) 2-ethylhexanoate (Sn(Oct)₂) as a reducing agent. Sn(Oct)₂ was chosen as the reducing agent because it exhibits good miscibility with the molten polymer and does not impart color change to the polymer. The objective of this study is to demonstrate that in melt free radical grafting reaction redox initiation avoids the cage effect and is more effective in achieving a high grafting degree and grafting efficiency than by conventional thermal initiation.

2. Experimental section

2.1. Materials

Sugar-based linear low density polyethylene (LLDPE), I'm Green SLH218, was supplied by Braskem, Brazil. The ethylene/glycidyl methacrylate copolymer (E-GMA), Lotader AX8840, with a glycidyl methacrylate content of 8 wt%, was supplied by Arkema, USA. The maleic anhydride grafted polyethylene (PE-g-MA), Polybond 3009, with a maleic anhydride content of 1 wt%, was supplied by Chemtura, Brazil. These two commercial polymer products were used as standards to construct calibration curves. Glycidyl methacrylate (GMA, 97% purity) and maleic anhydride (MA, 95% purity) were the monomers used to functionalize the polyethylene. Styrene (St, 99% purity) was used as a co-monomer. Dicumyl peroxide (DCP, 98% purity) was used as a free radical initiator and Tin(II) 2-ethylhexanoate (Sn(Oct)₂, 95% purity) was used as a reducing agent. All these chemicals were purchased from Sigma-Aldrich Chemical Co. and were used as received.

2.2. Melt grafting in extrusion

The functionalization of polyethylene was carried out in a co-rotating twin-screw extruder, Leistriz ZSE-18HP, with a screw diameter of 18 mm and a length-to-diameter ratio (L/D) of 40. The screw speed was maintained at 100 rpm for all runs. The feed method and the temperature profile were chosen depending on the monomer to be grafted. For the preparation of PE-g-GMA, the mixture of all the reagents, GMA, St, DCP and Sn(Oct)₂ (if used), was injected into the extruder through a syringe pump connected to the third heating zone while PE was fed through a volumetric feeder. The temperature profile was set at 155, 135, 130, 150, 175, 180, 180, 180 °C from the first heating zone to the die. The temperature was reduced at the reagent feed zone to avoid early polymerization of the monomers before being fed into the extruder. Alternately, for the preparation of PE-g-MA, MA, St, DCP and Sn(Oct)₂ (if used) were first premixed with PE in a beaker and then the mixture was fed into the extruder. In this case the temperature was set at 180 °C for all zones. The average residence time of the materials in the extruder was ∼2 minutes. Unreacted monomers and others volatiles were removed through a vacuum venting port located at the sixth heating zone. The extrudate strands were cooled in a water bath and subsequently pelletized.

2.3. Sample purification procedures

Prior to characterization, the modified PE was purified to remove residual monomer and other possible secondary products (homopolymers of St or GMA and copolymers of St with GMA or MA). The crude grafted PE (PE-g-GMA or PE-g-MA) (1.5 g) was dissolved in 30 mL of xylene at 120 °C by stirring and refluxing for one hour and then was precipitated in acetone (100 mL) at room temperature. The precipitate was filtered, washed three times with acetone and then dried at 80 °C for 24 h.¹⁷

2.4. Determination of grafting degree

The grafting degrees (Gd) of GMA and MA onto PE were determined by FT-IR. A calibration curve was created using samples of known grafting degrees. A similar approach has been used by several other authors^9,25–27 and consists of constructing a calibration curve based on the known amount of the grafted monomer versus the ratio of the area of two FT-IR characteristic peaks related to the monomer and the polymer. Application of the FT-IR method over the titration method was selected because PE exhibits a poor solubility at room temperature which makes titration difficult to perform. To construct the calibration curves, two commercial functionalized polyolefins containing 8 wt% GMA (Lotader AX8840) and 1 wt% MA (Polybond 3009) were respectively diluted into PE using a torque rheometer to produce samples of different concentrations. Thin films of these samples were prepared using a hot press and then FT-IR spectra were recorded to construct the calibration curves. To determine the grafting degree of GMA the peak area ratios of A₁₇₃₀/A₁₄₆₂ and A₁₇₃₀/A₁₃₈₀ were used, where A₁₇₃₀ was the area of the peak at 1730 cm⁻¹ corresponding to the carbonyl group of GMA and A₁₄₆₂ and A₁₃₈₀ the areas of the peaks at 1462 and 1380 cm⁻¹ corresponding to the bending and symmetrical deformation of methyl groups in PE, respectively.^16,17 To determine the grafting degree of MA the peak area ratio A₁₇₈₀/A₁₃₈₀ was used, where A₁₇₈₀ was the area of the peak at 1780 cm⁻¹ corresponding to the symmetric carbonyl stretching of MA.²⁸ For determination of the grafting degree, the FT-IR analysis was repeated 3–5 times on average for each sample. Fig. 1 shows the calibration curves for GMA and MA. The equations for the grafting degree (wt%) were obtained from the calibration curves, as follows:

Fig. 1 Calibration curves for the quantitative measurement of (a) GMA and (b) MA grafting degrees onto PE by means of FT-IR.

For GMA:

For MA:

Once the grafting degree is determined, monomer and initiator grafting efficiency can be evaluated. In this study, monomer grafting efficiency (G_eff) is defined as the percentage of the monomer grafted on the basis of total monomer charged in the melt graft reaction. Besides abstracting hydrogen atoms from macromolecules to form macroradicals, the primary radicals also initiate the homopolymerization and copolymerization of the monomer and co-monomer. Therefore, the initiator efficiency of grafting is an important concern in this process. In this study, this efficiency is gauged by the molar ratio of the monomer grafted to the primary radicals (MPR). Because one peroxide molecule produces two primary radicals via thermal decomposition but gives only one primary radical in the presence of a reducing agent, MPR can be determined with the following equations:

Thermal decomposition:

Redox initiation:

where M_m is molecular weight of the monomer, M_I is the molecular weight of the initiator DCP, and [M_I] is the wt% of DCP added.

2.5. Fourier Transform Infrared Spectroscopy (FT-IR)

Samples of grafted polymers, after purification, were pressed into thin films with a hydraulic press at a temperature of 190 °C and a pressure of 2500 psi. Fourier transform infrared spectroscopy (FT-IR) spectra of the thin films were recorded using a Thermo Nicolet Nexus 670 spectrometer. The sample was scanned between 500 and 4000 cm⁻¹ with a resolution of 2 cm⁻¹ and 32 scans.

2.6. Torque rheometry

Change of the torque value during grafting reaction was monitored using a Haake Rheomix 600p torque rheometer, operating at 180 °C with a rotor speed of 50 rpm for 15 min. The torque value was recorded as function of mixing time.

2.7. Viscosity and viscoelasticity measurements

Rheological properties were measured using a HR-2 Discovery Hybrid Rheometer (TA Instruments) with a parallel plate geometry of 25 mm in diameter. Frequency sweep ranging from 0.05 to 100 rad s⁻¹ was performed with a strain within the linear viscoelastic region (computed by a strain sweep experiment, not shown here). Shear rate sweep tests were performed with the shear rate varying from 0.01 to 10 1/s. For both tests a gap size of 1 mm was used and the temperature was maintained at 200 °C.

3. Results and discussion

3.1. Torque rheometry

Fig. 2 shows how torque evolves as a function of time during the grafting reactions of GMA and MA onto PE. The total mass of each formulation in the torque rheometry test remained the same, e.g., 45 g, so the torque value could be compared. The initial sharp increase in torque for each sample was attributed to the melting of the polymer. Fig. 2a shows the effects of the additions of GMA, St, DCP and Sn(Oct)₂ on torque behavior of PE. For the neat PE as a control, its torque gradually reached a constant value after melting. For all other compositions the torque started to increase right after melting (at ∼1.5 min) and peaked at ∼3–4 min of reaction. The PE/GMA/DCP composition only exhibited a small peak at ∼3 min of reaction and then reached a similar torque value like that of the neat PE, indicating that the grafting only proceeded to a very limited extent. This result also indicates that the branching and crosslinking via recombination of PE macroradicals were not significant. When styrene was added to that formulation, the torque exhibited a much greater increase, suggesting an increased extent of the grafting reaction. According to the literature, styrene can increase the GMA grafting degree up to 4-fold.¹⁶ Furthermore, with additions of both styrene and Sn(Oct)₂ to that formulation, the torque reached its peak height, indicating that the greatest extent of grafting was attained when the redox initiation method was used. Fig. 2b shows the effects on the torque with the additions of St and Sn(Oct)₂ during the grafting of MA onto PE. With the exception of neat PE, after melting, all compositions exhibited a significant increase in torque, reaching a maximum, and then decreased gradually with reaction time. Similar tendency in the torque behavior was also observed by other authors.²⁶ Nonetheless, the additions of styrene and Sn(Oct)₂ in the MA grafting reaction did not cause clear change in torque, instead, all formulations exhibited quite similar torque vs. time curves. The overall level of torque value for MA grafting was higher than that for GMA grafting, which was probably because MA was solid but GMA was liquid at room temperature and GMA greatly contributed to the reduction of viscosity of the melt during reaction. Finally, it was noted that the use of Sn(Oct)₂ did not clearly interfere with the reaction, i.e., the reaction initiated by a redox initiator seemed to follow the same trend as the one initiated by DCP alone.

Fig. 2 Evolution of torque as a function of time during grafting reactions of: (a) GMA and (b) MA. The initial concentrations of GMA and MA were 5 and 2 wt%, respectively. Initial concentration of DCP and Sn(Oct)₂ was 0.5 phr. If used, styrene was added at the equal concentration (mol mol⁻¹) to that of GMA or MA.

3.2. Effects of the DCP/Sn(Oct)₂ redox initiation on grafting of GMA and MA onto PE

Fig. 3 shows the FT-IR absorption spectra of PE, PE-g-GMA and PE-g-MA prepared with and without Sn(Oct)₂. In Fig. 3a, the peak at 1730 cm⁻¹ was attributed to the carbonyl group of GMA and the three peaks of weak intensity at 990, 908 and 850 cm⁻¹ were attributed to symmetrical and asymmetrical deformation vibrations of the epoxy ring of GMA. A peak at 702 cm⁻¹ is also noted which is characteristic of polystyrene and corresponds to the vibration of the carbon chain of polystyrene with aromatic hanging nuclei.¹⁶ The presence of these peaks confirms that the grafting occurred successfully. It can be seen that with the addition of Sn(Oct)₂, redox initiation, the intensity of the GMA characteristic peaks was greatly increased, meaning the GMA grafting degree was increased. Also, the peak at 702 cm⁻¹ indicates that the styrene used as a co-monomer was also grafted onto PE (PE-g-(St-co-GMA)). Further, this result qualitatively indicates that the content of styrene grafted onto PE increased in the presence of Sn(Oct)₂. In Fig. 3b, the peaks at 1780 and 1850 cm⁻¹ are assigned to the symmetric and asymmetric stretching of carbonyl groups of cyclic anhydride.^15,26,29 These spectra showed that, as in the case of GMA, the use of a redox initiation system could significantly increase the grafting degree of MA. A peak at 702 cm⁻¹ was also noted in the grafted PE due to the grafting of styrene (PE-g-(St-co-MA)), and its intensity increased similarly when a redox initiation was employed in the grafting reaction.

Fig. 3 FTIR Spectra of (a) GMA grafted PE and (b) MA grafted PE prepared under the catalysis of DCP and DCP/Sn(Oct)₂, respectively. In the melt grafting reactions, the initial concentrations of GMA and MA were 10 and 2 wt%, respectively; the concentrations of DCP and Sn(Oct)₂ were both 0.5 phr; styrene was added at the equal concentration (mol mol⁻¹) to that of GMA or MA.

Table 1 summarizes the results of grafting GMA and MA onto PE with and without the use of Sn(Oct)₂ in the melt grafting reaction. It is noted that in both grafting reactions the presence of Sn(Oct)₂ significantly increased the content of the functional groups grafted onto PE. For the grafting of GMA, the grafting degree increased with the initial concentration of GMA, irrespective of redox initiation or thermal initiation of the grafting reaction. However, with the use of Sn(Oct)₂ the grafting degree of the resulting GMA grafted PE was 1.7–2.7 times that of the grafted PE obtained without using Sn(Oct)₂, depending on the initial GMA concentration. Alternatively, redox initiation also resulted in great increases in the monomer grafting efficiency (G_eff) and the molar ratio of monomers grafted to the primary radicals (MPR) for the grafting of GMA. While G_eff basically remained relatively unchanged, MPR increased continuously with the initial concentration of GMA. Because equal moles of styrene were present to assist the grafting during melt reaction, these results suggest that the GMA was grafted onto PE by adding to the propagating macroradicals proportional to the GMA concentration in melt reaction. For the grafting of MA, compared to thermal initiation, redox initiation also resulted in great increases in grafting degree, monomer grafting efficiency and initiation efficiency. However, compared to the grafting of GMA, the grafting of MA in general yielded grafted PE products with significantly lower grafting degrees, irrespective of redox initiation and thermal initiation in the reaction. Increasing MA concentration in the grafting reaction only resulted in very limited increase in the grafting degree. Consequently, G_eff decreased with increasing MA concentration, displaying a quite different trend from the grafting of GMA. A similar tendency in the grafting of MA onto PE was also observed by other authors.^30,31 The grafting of MA was further different from the grafting of GMA in that MPR remained relatively unchanged with increasing MA concentration. The low degree of grafting of MA was likely due to both the steric effect on the internal carbon–carbon double bond and the unfavorable inductive effects of the two carbonyl groups of MA. Though the electron-withdrawing carbonyls of MA can stabilize the propagating radical species, they also greatly decrease electron-cloud density of the MA monomer and hence reduce its reactivity in reacting with the propagating species. As a consequence, even with the assistance of styrene as a co-monomer, the addition of MA onto propagating macroradicals was still much more inactive than that of the grafting of GMA.

Table 1 Effect of Sn(Oct)₂ on the grafting of GMA and MA onto PE

	[M]i^c (wt%)	Without Sn(Oct)2^a			With Sn(Oct)2^b
		Gd^d (wt%)	Geff^d (%)	MPR^d	Gd (wt%)	Geff (%)	MPR
a Initiator concentration: 0.5 phr DCP.b Initiator concentration: 0.5 phr DCP and 0.5 phr Sn(Oct)2.c [M]i was the initial concentration of monomer and the styrene co-monomer was added at the equal concentration (mol mol^−1) to that of GMA or MA.d Gd is the grafting degree, Geff is the monomer grafting efficiency and MPR is the molar ratio of the monomers grafted to the primary radicals.
GMA	4.0	1.00 ± 0.09	25.0	1.9	1.70 ± 0.06	42.5	6.5
	7.0	1.13 ± 0.11	16.0	2.2	3.10 ± 0.25	44.3	11.8
	10.0	2.15 ± 0.04	21.5	4.1	4.70 ± 0.45	47.0	17.9
MA	2.0	0.34 ± 0.01	17.0	0.9	0.72 ± 0.02	36.0	4.0
	3.5	0.50 ± 0.02	14.3	1.4	0.75 ± 0.03	21.4	4.1
	5.0	0.46 ± 0.03	9.2	1.3	0.76 ± 0.03	15.2	4.2

---

To better understand the effect of Sn(Oct₂) on DCP initiated grafting of GMA and MA onto PE, a range of grafting reactions which were formulated with DCP and Sn(Oct)₂ in a fixed total weight (i.e., 1 phr) but with different weight ratios were performed. Table 2 summarizes the effects of the DCP/Sn(Oct)₂ ratio on grafting efficiency of GMA and MA. In general, the grafting degrees and monomer grafting efficiencies for both GMA and MA started to increase with the addition of Sn(Oct)₂ and achieved their maximum values when the DCP/Sn(Oct)₂ ratio was around 1. As the relative Sn(Oct)₂ content further increased, Gd and G_eff tended to decrease. This latter trend was likely due to the large reduction of the primary radicals because of low DCP content in the reaction. From these results it can be concluded that the highest grafting degree was achieved at the DCP/Sn(Oct)₂ weight ratio around 1 : 1 (w/w).

Table 2 Influence of the DCP/Sn(Oct)₂ weight ratio on grafting of GMA and MA onto PE

Entry	DCP (phr)	Sn(Oct)2 (phr)	GMA^a			MA^a
			Gd^b (wt%)	Geff^b (%)	MPR^b	Gd (wt%)	Geff (%)	MPR
a Initial concentration of GMA and MA was 10 and 2 wt%, respectively, and the styrene comonomer was added at the equal concentration (mol mol^−1) to that of GMA or MA.b Gd is the grafting degree; Geff is the monomer grafting efficiency and MPR is the molar ratio of the monomers grafted to the primary radicals.
1	1	0	2.05 ± 0.17	20.5	3.9	0.55 ± 0.05	27.5	1.5
2	0.75	0.25	1.81 ± 0.21	18.1	4.6	0.63 ± 0.02	31.5	2.3
3	0.5	0.5	4.70 ± 0.45	47	17.9	0.72 ± 0.02	36.0	4.0
4	0.25	0.75	1.77 ± 0.12	17.7	13.5	0.65 ± 0.01	32.5	7.2
5	0	1	0.50 ± 0.06	5	—	0.17 ± 0.01	8.5	—

---

It is interesting to note that grafting of GMA and MA still occurred when only Sn(Oct)₂ was present in the reaction though the grafting degree was significantly low (entry 5, Table 2). There are two likely explanations to account for the low grafting degree even without use of DCP in the reaction. The first is that the grafting reaction was initiated by free radicals generated by the thermo-mechanical processing. Several studies^32,33 demonstrate that the thermo-mechanical degradation of polyethylene involves formation of alkyl radicals through cleavage of covalent bonds, followed by reaction with oxygen to form hydroperoxides which initiates the grafting reactions. The second explanation is the presence of oxygen during the melt reaction which led to the formation of hydroperoxide groups on the macromolecules.³² In either case, the presence of Sn(Oct)₂ could promote more effective initiation of the grafting reactions. Such side reactions are usually inevitable in polymer melt processing though they only occur to a minimum extent and generally do not cause significant changes to mechanical properties and can be ignorable. To confirm this reasoning, the grafting reaction was performed without addition of both DCP and Sn(Oct)₂. It is noted that the grafting for both GMA and MA could take place even without the presence of both DCP and Sn(Oct)₂ in the reaction though it yielded much lower grafting degrees (Table 3). Furthermore, to ensure that the detected functionalization was not due to the possible residual monomers in the samples, the grafted PE samples were purified twice by solution precipitation. The results in Table 3 show that the samples from both one precipitation and two precipitations exhibited very close grafting degrees, suggesting a single purification was enough to remove the unreacted monomers. Other studies also indicated that a single purification was sufficient to remove residual monomers and homopolymers.^9,10,16,26

Table 3 Grafting degrees of GMA and MA onto PE without using DCP

Purification	[M]i^a (wt%)		Grafting degree (wt%)
			Without Sn(Oct)2		With Sn(Oct)2^b
	GMA	MA	GMA	MA	GMA	MA
a [M]i was the initial concentration of the monomer and the styrene comonomer was added at equal concentration (mol mol^−1) to that of GMA or MA.b The melt grafting reaction was performed in the presence of 1 phr Sn(Oct)2.
1st time	10.0	—	0.32 ± 0.05	—	0.50 ± 0.08	—
2nd time	10.0	—	0.28 ± 0.02	—	0.50 ± 0.06	—
1st time	—	2.0	—	0.14 ± 0.01	—	0.17 ± 0.00
2nd time	—	2.0	—	0.12 ± 0.02	—	0.16 ± 0.02

---

Unlike ferrous (Fe²⁺) or cuprous (Cu⁺) salts which are common reducing agents often used in redox initiation systems, the use of stannous (Sn²⁺) as the reducing agent is scarcely reported in the literature. In this work, Sn(Oct)₂ was used in the initiation system because it was miscible with PE and would not cause a color change to the compound. To the best of our knowledge, use of Sn²⁺ salts in initiation systems has only been mentioned in two patent applications; one used SnCl₂ in the suspension polymerization of vinyl chloride²⁰ and the other used the stannous salts of carboxylic acids for the polymerization of acrylic monomers.³⁴ By taking into account the greatly enhanced efficiency of grafting of MA and GMA onto PE by addition of Sn(Oct)₂ and the oxidation state of +4 of tin being more stable than +2, it could be deduced that the combination of DCP and Sn(Oct)₂ initiated the radical grafting via a oxidation–reduction reaction. Based on the known knowledge of redox initiation and styrene assistance for the grafting reaction, a redox initiation of the DCP/Sn(Oct)₂ mechanism in this work is proposed and shown in Scheme 1. In the absence of a reducing agent and upon heating, a DCP molecule undergoes a normal homolytic decomposition to yield a pair of primary free radicals which first exist in a cage. The free radicals can diffuse out of the cage and initiate the grafting reaction but may also undergo side reactions to form neutral molecules. The side reactions waste the initiator and reduce the initiator efficiency. In the presence of the reducing agent, i.e., Sn(Oct)₂ in this work, one DCP molecule gives one free radical and the cage effect is eliminated. The free radicals formed attack the PE molecules and abstract the hydrogen atoms, resulting in the formation of macroradicals. It is known that the direct attack of the PE macroradical to MA or GMA is significantly hindered by the steric effects of the polymer chains, resulting in low grafting efficiency.^4,10,11 When styrene is used as a co-monomer in the grafting reaction, the PE macroradicals prefer to attack styrene because styrene presents less steric effect and stabilizes the resulting styryl macroradicals by resonance.

Scheme 1 Comparison of free radical initiations by thermal decomposition of DCP and by a DCP/Sn(Oct)₂ redox initiation system.

3.3. Rheological properties of the grafted PEs

The rheological properties of a polymer are determined by its molecular structure. In general, both grafting and crosslinking leads to increased resistance to deformation of a polymer melt. As seen in Fig. 4a & b, both MA and GMA grafted PEs exhibited significantly higher complex viscosity than the neat PE and complex viscosity increased with grafting degree. The grafted PEs displayed weak shear-thinning behaviors over the entire range of frequencies scanned while the neat PE shows a typical pseudo-plastic behavior, with a Newtonian plateau at low frequencies and shear-thinning behavior at high frequencies.

Fig. 4 Complex viscosity as a function of the angular frequency (a and b) and apparent viscosity as function of the shear rate (c and d) for GMA and MA grafted PE. Initial concentration of GMA and MA was 10 and 5 wt% respectively. Styrene was added at the equal concentration (mol mol⁻¹) to that of GMA or MA. Initial concentration of DCP and Sn(Oct)₂ was 0.5 phr.

Similarly, the results of the static viscosity measurement also revealed the enhanced elastic behaviors for the grafted PEs (Fig. 4c & d). The static shear viscosity of the neat PE displayed a clear Newtonian plateau in the low shear rate region (<1 s⁻¹) and demonstrated a shear-thinning behavior with an increasing shear rate above 1 s⁻¹. In contrast, the static shear viscosity of the grafted PEs displayed a Newtonian plateau at much lower shear rates. Furthermore, the grafted PEs also exhibited much higher static viscosity than that of the neat PE, especially at the low shear rates. In other words, the grafted PEs had much higher zero shear rate viscosity than the neat PE. This result was probably attributed to the changes in molecular weight and branch structure incurred by grafting. However, as the shear rate increased, viscosity started to decrease and tended to converge.

4. Conclusions

Redox initiation of melt free radical grafting of GMA and MA onto PE was successfully performed by reactive extrusion. More specifically, the DCP/Sn(Oct)₂ redox system was used to initiate the grafting reactions. This approach proved to be more effective on melt grafting of PE compared to the conventional free radical grafting reactions initiated by thermal decomposition of peroxides. Significant increases in grafting degree and grafting efficiency were achieved when the redox initiation system was used. It is believed that use of redox initiation effectively eliminated the cage effect associated with the conventional thermal initiation, leading to high grafting efficiency. The optimum weight ratio of DCP/Sn(Oct)₂ was found to be around 1 : 1 to receive the highest grafting degree and monomer grafting efficiency for both GMA and MA onto PE. While the grafted GMA was proportional to the GMA concentration in melt reaction, the grafted MA exhibited little dependence on the MA concentration in reaction. This result suggests grafting of GMA onto PE could be effectively implemented with the assistance of the styrene co-monomer, but the grafting of MA was still largely dictated by its steric hindrance. In addition, rheological measurement showed that viscosity of the grafted PE increased with grafting degree.

Acknowledgements

The authors gratefully acknowledge CAPES, Brazil, for their financial support, proc. no. 17569-12-5, for the visiting study of Gustavo F. Brito at Washington State University. The authors are also grateful for the technical support and facilities provided by the Composite Materials and Engineering Center at Washington State University.

References

1. G. Moad, Prog. Polym. Sci., 1999, 24, 81–142.
2. A. Bhattacharya and B. N. Misra, Prog. Polym. Sci., 2004, 29, 767–814.
3. E. Passaglia, S. Coiai, F. Cicogna and F. Ciardelli, Polym. Int., 2014, 63, 12–21.
4. Y.-J. Sun, G.-H. Hu and M. Lambla, J. Appl. Polym. Sci., 1995, 57, 1043–1054.
5. C.-M. Chen, T.-E. Hsieh and M. O. Liu, React. Funct. Polym., 2008, 68, 1307–1313.
6. R. R. Gallucci and R. C. Going, J. Appl. Polym. Sci., 1982, 27, 425–437.
7. T. C. M. Chung, Macromolecules, 2013, 46, 6671–6698.
8. P. Chandranupap and S. N. Bhattacharya, J. Appl. Polym. Sci., 2000, 78, 2405–2415.
9. I. Pesneau, M. F. Champagne and M. A. Huneault, J. Appl. Polym. Sci., 2004, 91, 3180–3191.
10. H. Cartier and G.-H. Hu, J. Polym. Sci., Part A: Polym. Chem., 1998, 36, 2763–2774.
11. H. Cartier and G.-H. Hu, J. Polym. Sci., Part A: Polym. Chem., 1998, 36, 1053–1063.
12. G.-H. Hu, J.-J. Flat and M. Lambla, Makromol. Chem., Macromol. Symp., 1993, 75, 137–157.
13. Y.-J. Sun, G.-H. Hu and M. Lambla, Die Angewandte Makromolekulare Chemie, 1995, 229, 1–13.
14. G.-H. Hu and H. Cartier, J. Appl. Polym. Sci., 1999, 71, 125–133.
15. Y. Li, X.-M. Xie and B.-H. Guo, Polymer, 2001, 42, 3419–3425.
16. N. Torres, J. J. Robin and B. Boutevin, J. Appl. Polym. Sci., 2001, 81, 581–590.
17. Q. Wei, D. Chionna, E. Galoppini and M. Pracella, Macromol. Chem. Phys., 2003, 204, 1123–1133.
18. G. Odian, in Principles of Polymerization, John Wiley & Sons, Inc., 2004, pp. 198–349.
19. G. T. Russell, D. H. Napper and R. G. Gilbert, Macromolecules, 1988, 21, 2133–2140.
20. N. G. Gaylord, US Pat., 4 269 958, 1981.
21. B. N. Misra, I. K. Mehta and R. C. Khetarpal, J. Polym. Sci., Polym. Chem. Ed., 1984, 22, 2767–2775.
22. U. D. N. Bajpai, A. Jain and S. Rai, J. Appl. Polym. Sci., 1990, 39, 2187–2204.
23. J. Zhou, L. Bai, H. Qin and Y. Liu, Chin. J. Chem., 2011, 29, 351–355.
24. A.-C. Mortamet and R. A. Pethrick, J. Appl. Polym. Sci., 2012, 123, 1539–1547.
25. H. Cartier and G.-H. Hu, Polym. Eng. Sci., 1998, 38, 177–185.
26. D. C. Clark, W. E. Baker and R. A. Whitney, J. Appl. Polym. Sci., 2001, 79, 96–107.
27. J.-L. Li and X.-M. Xie, Polymer, 2012, 53, 2197–2204.
28. J. John, J. Tang, Z. Yang and M. Bhattacharya, J. Polym. Sci., Part A: Polym. Chem., 1997, 35, 1139–1148.
29. M. Sclavons, P. Franquinet, V. Carlier, G. Verfaillie, I. Fallais, R. Legras, M. Laurent and F. C. Thyrion, Polymer, 2000, 41, 1989–1999.
30. Y. Güldoğan, S. Eğri, Z. M. O. Rzaev and E. Pişkin, J. Appl. Polym. Sci., 2004, 92, 3675–3684.
31. W. Qiu, T. Endo and T. Hirotsu, Eur. Polym. J., 2005, 41, 1979–1984.
32. F. Gugumus, Polym. Degrad. Stab., 2000, 68, 337–352.
33. T. Andersson, B. Stålbom and B. Wesslén, J. Appl. Polym. Sci., 2004, 91, 1525–1537.
34. C. J. Schmidle and S. N. Varadhachary, US Pat., 4 389 514, 1983.

---

Footnote

† G.F. Brito is a visiting scholar at Washington State University.

---