Elsevier

Polymer

Volume 109, 27 January 2017, Pages 160-169
Polymer

In situ polymerization of styrene into a PMMA matrix by using an extensional flow mixing device: A new experimental approach to elaborate polymer blends

https://doi.org/10.1016/j.polymer.2016.12.045Get rights and content

Highlights

  • A new approach of reactive blending based on in-situ polymerization.

  • Use of a mixer based on extensional flow named RMX®.

  • Obtaining of nanometer particle size without the aid of compatibilizers.

Abstract

The present work consisted in the in situ polymerization of a liquid monomer (styrene) into a melt polymer matrix (PMMA) as an innovative method to prepare polymer blends at fine dispersions. That was possible by using an original mixing/reactor device called RMX®, which significantly develops extensional flow, greatly enhancing the mixing efficiency. Resultant morphology was characterized by TEM and average particle sizes were obtained by means of image analysis. A high distributive efficiency was observed at very fine dispersed phase morphology, tens of nanometers in diameter, in fact, in the absence of compatibilizing agents. It was found, by comparing self (thermal) and peroxide-initiated systems, that morphology strongly depends on the balance between polymerization rate and monomer diffusion into the polymer matrix (as enhanced by PMMA solubility into styrene) during mixing. Finally, a comparison of morphology between compounds prepared by reactive blending and those elaborated by conventional melt mixing was done.

Introduction

Different approaches have been used nowadays to elaborate new and functional plastic compounds. The synthesis of copolymers is one of them, either by combining monomers or by synthetizing new types of them intended for specific applications. Nevertheless, in a practical and economic way, a wide range of bulk properties for polymeric materials is still obtained by direct blending of polymers. Particularly, the mechanical melt blending of polymers was described for the first time in a patent registered in 1855 [1] to obtain a blend of cellulose nitrate with natural rubber and gutta percha. With the advent of synthetic rubbers, elastomer blends of natural rubber and styrene-butadiene rubber become popular in the 1940s [2]. Over time, the golden age for polymer blends arrives in the 60's. Ever since, academic and industrial works in this domain have exponentially grown [2].

Among different routes found in literature, it is clear that melt blending continues to be the most common method used to produce polymer blends at large scale because of the ease of implementation, high rates and volumes of production. On the context of this technique factors governing polymer mixing like the natural immiscibility of most polymer pairs, the viscosity ratio, p, of the blend's components and the flow geometry have been largely analyzed. However, despite the technical success of melt blending, important limitations linked to fundamental parameters, prevent it in general to go further than the elaboration of rather commercial compounds. For example, it is well-known that most of the conventional melt mixers work under shear flow regimes, which limits the efficiency of mixing to low and medium p values [3], [4], [5], [6], [7] (minimum size of dispersed domains hardly reaches 1 μm without the aid of compatibilizers). Moreover, the inherent downstream direction of the melt blending process by starting out from polymers as raw materials makes quite difficult the targeting or tuning of desired morphologies, thus, final properties of blends.

Different strategies other than direct melt mixing have been continuously evolving to prepare innovative polymer blends. Among the most interesting ones are those based primarily on the in situ polymerization approach applied to polymer processing, which is the frame of the present work to be detailed below. In this, unlike the conventional melt blending, one or more monomers in addition to polymeric components may be present in the starting mixture. In this way, a monomer generally soluble in a polymer component undergoes in-situ polymerization along with the mixing process. As long as the polymerization reaction proceeds phase separation and morphology development take place. Although the approach itself is not recent, the technique has had to be adapted over time to melt mixing procedures in order to foster both, the mixing efficiency and the potential for the continuous elaboration of polymer blends. From the existing literature, Parent and Thompson [8] had already essayed the in situ polymerization of methyl methacrylate into PS and styrene into PMMA. Also, Uchida et al. [9] polymerized styrene into PEG and PDMS. In both references, the reactive blending was carried out in small flasks under relatively controlled conditions. On the other hand, Cassagneau et al. (1998) [10] and Alam et al. (1995) [11] have attempted in situ polymerization to elaborate polymer blends via melt mixing, the first one by polymerizing PU into a PE matrix using a twin-screw extruder. The second one polymerized PC into chemically modified PP, using an internal mixer. Furthermore, some additional examples of materials produced by this technique include HIPS, PP/poly(epsilon-caprolactone) [12] and PS/polyepoxide [13], [14]. It is important to distinguish, though, the approach just described from that processes where only functionalization or cross-linking occur [15] i.e. reactive processing.

The potential of the in situ polymerization technique adapted to melt has to be unavoidably linked to quality of mixing. On this respect our strategy introduces a new brand mixing device, RMX®, in order to highly enhance the dispersion capacity. In literature, the influence of flow geometry in mixing efficiency has been extensively treated i.e. breakup mechanisms of dispersed droplets in a continuous matrix in shear flow [16], [17] or in extensional flow [18], [19], [20]. In particular, Grace [7] has studied the dispersive efficiency in shear flow fields using a Couette device and in extensional using a four roll device flow for different pairs of Newtonian polymers in a wide range of viscosity ratio. The Grace's study showed that the average size of droplets has a high dependence with the viscosity ratio in a mixer based on predominantly shear flow while under extensional flow the average size is smaller and almost independent of p. Other works confirm the general trends observed in Grace's work and the higher efficiency of extensional flow even for the case of polymers as viscoelastic materials [3], [4], [5], [6], [21]. Based on this, attempts to develop extensional mixers are widely describe by the literature [22], [23], [24], [25], [26], [27], [28] since the current processing and mixing technology are mainly based on shear flow. Among extensional flow mixers, an interesting type is that based on contraction flows where an extensional flow field is generated by flow acceleration in the convergence and/or divergence zones. Mackley's multipass rheometer [25], Son's microcompounder [27] and, more recently, the RMX® [29], [30], [31], [32], [33] are representative examples.

In our present work, we developed an original and powerful technique which, in one step, combines the in situ polymerization approach along with extensional mixing with the aim of preparing innovative polymer blends. By using this method, a fine and controlled balance between formulation design, desired morphologies and mixing efficiency is intended to be attained. Specifically, our strategy consists in polymerizing a liquid monomer into a melt polymer matrix within an extensional mixer in order to elaborate new type of material such as thermoplastic/thermoset/thermoplastic blends without the use of solvents. The reactor/mixing device to carry out the experimental work was recently developed and patented by our laboratory [33] with the name of RMX® (reactor and mixer device), a contraction flow type mixer capable of generate strong extensional flow fields. In this paper, we present our first experiment, PMMA/styrene system leading to PMMA/PS blends, to valid the feasibility of our technic.

Given the multivariable nature of the present approach, in this paper we focus on testing the conception of the idea, analysing the fundamental phenomena implied. In a second paper to be published shortly, we will discuss on the effect of changing different formulation and processing parameters.

As reactive raw materials we have selected Styrene (stabilized with 0.005% 4-tert-butylcatechol) from Sigma-Alrich as a monomer along with two initiators, Luperox® DI (L) provided from Sigma Aldrich and Trigonox® 311 (T) from Akzo Nobel. Table 1 and Table 2 present some important physicochemical properties for the monomer and initiators respectively. On the other hand, the polymer used in this study was Poly(methyl methacrylate) (PMMA: Altuglas® VM 100) from ARKEMA. A set of properties is shown in Table 3.

Fig. 1 shows a scheme of the overall mixer assembly from which the RMX®, where reactive mixing takes place, constitutes the central processing dispositive. As can be observed, the system also consists of two important outlying accessories: a single screw extruder aimed for feeding the melt polymer into the RMX® and a reservoir/pump device in charge of supplying the monomer. On the other hand, already discussed in detail elsewhere [29], [30], [31], [32], [33], [36], the key operational principle of the RMX® is based in the generation of strong extensional stresses by means of convergent/divergent flow. To this end, a pair of pistons running along two opposite chambers forces the material to pass through a central round die or mixing element fixed between the two chambers, thus creating accelerating (extensional) flow. A back and forth motion (cycle) allows for the multiple passage of the compound through high stress regions, which, in addition to the increased extensional flow contribution, represents a basic requirement for an efficient dispersive mixing [7], [37]. In practice, once the dimensions of the mixing element have been defined, a mixing sequence can therefore be commanded by controlling three main parameters: temperature, piston's speed and number of cycles. Moreover, Table 4 presents the RMX® specifications.

The central scope of this method consisted on the immediate start up of the mixing step once the blend's components have been finished to be fed into the RMX®. In this way, polymerization is intended to evolve along with the mixing process. For each blends, the total volume correspond to 50 cm3. Next, we present a detailed description of the approach:

  • 1)

    Introduction of melt polymers into the left chamber (Fig. 1) of the RMX® is carried out by the single screw extruder (Scamex) at 9 cm3/min. Both, the RMX® and extruder were kept at 195 °C.

  • 2)

    Introduction of liquid monomer as received with or without initiator into the right chamber was done through the use of a high pressure pump (MiltonRoy MILROYAL MD140G6M390). A counter-pressure built up to the piston of the introduction chamber in order to keep the styrene in the liquid form.

  • 3)

    Reactive mixing takes place, i.e. polymerization reaction with or without initiators along with the mixing mechanism. RMX® pistons motion start at the end of the monomer introduction, therefore, throughout the mixing sequence both, monomer conversion and blend morphology are expected to evolve simultaneously. In this step mixing conditions were fixed to 90 cycles at a speed of pistons of 100 cm/min which corresponds to 16 min of mixing.

  • 4)

    After the mixing/polymerization step (16 min) is finished, the compound was allowed to remain at rest within the RMX® a period of time, when needed, enough to ensure maximum monomer conversion according to the kinetics of each reactive system in Fig. 3. Thus, in the case of the thermal-initiated polymerization a resting time of 24 min in addition to previous 16 min of polymerization during mixing (summing up to 40 min) was considered sufficiently ahead the time for maximum monomer conversion to PS according to Fig. 3 (30 min). In the same direction, among the peroxide-initiated systems, the reactive blending of the less reactive of them was completed with 4 min of resting time whereas in the case of the most reactive one no resting time was added on considering that 16 min during the mixing/polymerization step was more than enough to ensure completion of the reaction. In summary, the total residence time for this method was considered as the addition of the mixing time plus the resting time when needed: 40min, 20 and 16min for the thermal-initiated system, Trigonox 311 and Luperox DI, respectively.

  • 5)

    The releasing of the final blend is carried out by opening up the exit valve and moving forward one of the pistons. Finally, the material is let cool down at room temperature.

Method B is indeed an inversion of method A. Here, the polymerization of the monomer is firstly allowed to go to completion within the RMX® before the start up of the mixing back and forth motion. This method is described as follows.

  • 1)

    Feeding of melt polymer. As in method A

  • 2)

    Introduction of monomer/initiator. As in method A.

  • 3)

    Once the components are fed into the RMX® the polymerization reaction takes place in static conditions until complete conversion of the monomer. The corresponding polymerization time varies according to Fig. 3 which shows the conversion kinetics of the different reactive systems. Accordingly, 40, 20 and 16 min for styrene without initiator, styrene/Trigonox® 311 and styrene/Luperox® Di, respectively (again, these times are well beyond those founded in Fig. 3 in order to ensure complete polymerization).

  • 4)

    Start up of the mixing step (no further or remaining polymerization is expected); dispersion and distribution of the in situ formed polystyrene throughout the PMMA matrix. The mixing step, as in A, was carried out at 90 cycles and 100 cm/min so as to represent 16 min of mixing as well. So this, total time in method B corresponds to the addition of the polymerization plus the mixing times.

  • 5)

    Releasing of final blends. As in method A.

Fig. 2 shows a schematic representation of both methods A and B for the sake of clarity.

Finally, in order to facilitate the identification of samples in the Figures caption the following nomenclature is followed: A-82v100N90, where A means method A, 82 refers to commanded polymer concentration (80/20 wt/wt) at a speed v of 100 mm/s and 90 cycles. A letter T or L preceding the concentration digits designs the use of peroxides in the formulation.

For the sake of morphology comparison, a conventional melt blending starting out from a pair of polymers was also done. Solid PMMA pellets and PS obtained from a previous polymerization reaction of styrene within the RMX® were introduced to an internal mixer chamber (Haake®). It is important to remark that the aim of using freshly polymerized PS was having the same molar mass and polydispersity as those present in the final blend coming from method A. The blending step was carried out at a temperature of 195 °C and 50 rpm. The mixing time was calculated to match the same specific mixing energy [36] than that of the RMX® in method A. The specific mixing energy was taken as a reliable criterion for morphology comparison since operation principles between devices greatly differ. The specific mixing energy is calculated as follows [32]:EHaake=iΩiTiΔtimwith: Ωi the rotation speed (rad/s), Ti is the torque (N.m), Δti the time range and m is the total mass of materials in the mixer.ERMX=ΔP·Nρwith, ΔP the pressures drop between the upstream and downstream chambers; ρ the mass density and N is the number of cycles.

Monomer conversion was characterized by a DSC Q200 (TA instruments) under nitrogen. A sealed stainless steel cup previously weighted was filled with around 10 mg of raw styrene or mixture of styrene and initiator, then proceeding as follows: equilibrium at 30 °C; isothermal for 1 min; jump to 195 °C at almost 90 °C/min; isothermal of 5, 10, 20, 30 and 50 min; cooling to 30 °C at 50 °C/min; equilibrium at 30 °C; isothermal for 2 min. After the experimental procedure cups were weighted to verify a possible mass loss. Finally, cups were thoroughly open and dry under vacuum at 50 °C, and weighted again. Thus, the rate of conversion is calculated as follow: 100·(massafterdrying massintroducedintheDSC).

Fig. 3 shows the kinetics of the polymerization reaction of styrene. As expected, according to the half-life times of each initiator, a conversion plateau is reached faster by using Luperox® DI, the most reactive agent. At 10 min, polymerization goes to completion under the influence of any of initiators. Differently, the pure thermal initiated polymerization reaches a conversion plateau until about 30 min but at the same percentage as in the case of initiators. As described in Section 2.2, this information has been critical to design and explore different reactive blending routes looking for a better mixing efficiency.

Fig. 4 presents the molar mass attained in the polymerization of styrene in the DSC with and without initiators. The evolution of the molar mass with time is not usual [38], [39] but consistent with the literature at temperature over than 180 °C [38], [40]. Values of Mn and Mw are quite similar than those found by Hui et al. [38] or Taherzadeh et al. [41] at 200 °C. In order to account for these findings a first approach is that at the early stage of the polymerization a strong competition between propagation and termination rates is present. Nevertheless, as recombination is said to be a significant mechanism of chain termination at this stage, the overall result is a rather relatively high molar mass polymer [42]. As long as the reaction proceeds the increase of viscosity normally hinders long radicals for recombination to occur and propagation is supposed to be favoured. However, in the case of the high temperature employed in the present case joined to very low PS molar mass the overall viscosity of the system is lower compared to conventional low temperature polymerization. As a result, a higher diffusion of active species is quite probable and termination modes other than and including recombination overcome propagation, specifically chain transfer to intermediates [43], [44], giving place to chains of low molar mass. Furthermore, chain transfer modes may keep a significant concentration of actives species as to initiate numerous short chains in small remaining portion of monomer, thus contributing as well to lowering molar mass.

On the other hand, Mw and Mn are significantly higher for the pure thermal activated polymerization of styrene. This may be explained since the presence of initiator generates, along the reaction, shorter and numerous propagating radicals, hence, able to lead to faster termination reactions due to a higher coefficient of diffusion.

The molar mass of PS synthetized in the presence of the PMMA matrix was obtained by dissolution of the PMMA phase in acetic acid followed by centrifugation. This operation was repeated four times in order to recover solely the PS phase. IR analysed confirmed the fact that only PS was recovered. The molar mass of the PS synthetized in the RMX without PMMA or in the PMMA matrix was analysed by GPC. The results are reported in Table 5.

The blend morphology was characterized using a Hitachi H 7500 Transmission Electron Microscope. Prior to the analysis, samples were cut into thin slices (80–100 nm thickness) by means of an ultra-microtome at a room temperature. For each blend, micrographs were taken on one sample at three different areas (size of the areas: 24 × 24 μm or 11 × 11 μm) using a voltage of 80kv. The obtained micrographs were then analysed using Image J, software for image analysis. At least 300 equivalent radii (ri) by micrograph were determined by measuring 2D surface area (Ai). The surface areas are not perfectly spherical, so we consider that ri correspond to the radius of an equivalent sphere with the same surface than the real surface area.ri=Aiπ

The number-averaged radius Rn, volume-average radius Rv, and molar mass distribution PĐI were calculated according to:Rn=inirniiniRv=inirni4inirni3PĐI=RvRn

Rn and Rv values to be shown below are based on the average of at least 3 micrographs. On the graphics, the bars are the average of the deviations from the average.

Dynamic rheological characterization of the raw materials and the blends was carried out at 195 °C with a strain-controlled rheometer (ARES, Rheometrics Scientific) with parallel-plate geometry. Time and strain sweep measurements were first carried out to determine the thermal stability and the domain of linear viscoelasticity. A frequency sweep from 100 to 0.5 rad/s with strain amplitude at 1% was performed to determine the rheological behaviour of the blends. It was completed by capillary rheometer (Rosand RH2000, Malvern Instruments) with both the Bagley and the Rabinovitch corrections.

In Fig. 5, we can observe that the viscosity behaviour of the PS synthetized in the RMX without (Mw = 40000 g/mol) and with initiator (Mw = 9000 g/mol) are pretty close to that of a Newtonian fluid. This behaviour may be explained in general by the vicinity and lower value, respectively, of the actual molar masses in comparison to the critical molar mass reported for the PS (Mcr(PS) = 35000–40000 g/mol [45]). According to literature [46], [47], in this conditions, where entanglements are absent or their effects negligible (M ≤ Mcr) the viscosity dependence of Mw approaches to: η = KMw1. In fact, an approximate viscosity ratio between the two PS (3.8) is not that far from the molar mass ratio (4.4) which is in good agreement with the discussed above.

Section snippets

Morphology of polymer blends obtained by in-situ polymerization, method A

Fig. 6 shows the comparison between the morphology obtained by running a just thermal-initiated monomer polymerization during the reactive blending, (a), and those employing initiators of low and high reactivity, (b) and (c), respectively. As observed in the pictures and confirmed by the image analysis, Fig. 7, the lowest dispersed domain average size corresponds to the blend obtained without the use of any initiator whereas among the peroxide-initiated systems a finer dispersion is attained by

Conclusions

The results of this article have demonstrated the success of a new approach of reactive blending based on in-situ polymerization along with predominantly extensional mixing exerted by the RMX® mixer. By means of this new approach is possible to obtain polymer blends morphologies at average particle size in the range of 100–200 nm (Rv) at concentrations 20% wt./wt. of in situ polymerized PS without the aid of compatibilizers. This average size order is lower in comparison to those obtained in

Acknowledgments

The authors wish to acknowledge gratefully Cathy Royer for the TEM micrographs, Christophe Sutter for software development and Thierry Djekrif for technical supports. We also wish to acknowledge the reviewers for their helpful comments.

References (53)

  • W.J. Milliken et al.

    J. Newt. Fluid Mech.

    (1991)
  • P. Cassagnau et al.

    Polymer

    (1999)
  • T.M. Alam et al.

    Polymer

    (2005)
  • R.W. Venderbosch et al.

    Polymer

    (1995)
  • M.R. Mackley et al.

    J. Newt. Fluid Mech.

    (2011)
  • J.K. Lee et al.

    Polymer

    (1999)
  • H.K. Jeon et al.

    Polymer

    (2005)
  • L.A. Utracki

    Polym. Eng. Sci.

    (1995)
  • L. Robeson

    Polymers

    (2014)
  • L.A. Utracki et al.

    Polym. Eng. Sci.

    (1992)
  • U. Sundararaj et al.

    Macromolecules

    (1995)
  • I. Delaby et al.

    J. Macromol. Sci.-Phy.

    (1996)
  • H.P. Grace

    Chem. Eng. Commun.

    (1982)
  • R.R. Parent et al.

    J. Polym. Sci. Part B-Polym. Phys.

    (1978)
  • T. Uchida et al.

    Des. Monomers Polym.

    (2002)
  • G.H. Hu et al.

    Macromolecules

    (1999)
  • L. Meynie et al.

    Polymer

    (2004)
  • R. Bella et al.

    Macromol. Symp.

    (2006)
  • G.I. Taylor

    Proc. R. Soc. Lond.

    (1932)
  • K. Min et al.

    Polym. Eng. Sci.

    (1984)
  • M. Meller et al.

    Polym. Eng. Sci.

    (2002)
  • M. Meller et al.

    Polym. Eng. Sci.

    (2002)
  • M. Meller et al.

    Int. Polym. Process.

    (1999)
  • B.D. Favis

    J. Appl. Polym. Sci.

    (1990)
  • M. Maric et al.

    Polym. Eng. Sci.

    (2001)
  • A.P. Harvey et al.

    Ind. Eng. Chem. Res.

    (2001)
  • Cited by (3)

    • Enhancing a multi-field-synergy process for polymer composite plasticization: A novel design concept for screw to facilitate phase-to-phase thermal and molecular mobility

      2020, Applied Thermal Engineering
      Citation Excerpt :

      When synergy angle α reaches zero, elongational flow occurs in the polymer melt; when it becomes 90°, the melt takes on a shear flow. It is generally known that the effect of elongational flow on physical mixing in composites is more effective than that of shear flow [26–31], which provides evidence for the feasibility of the field-synergy analysis method. This provides a new perspective to understand the essence of mixing enhancement.

    View full text