Elsevier

Polymer

Volume 45, Issue 4, February 2004, Pages 1223-1234
Polymer

Flow-induced chain scission as a physical route to narrowly distributed, high molar mass polymers

https://doi.org/10.1016/j.polymer.2003.11.051Get rights and content

Abstract

We present data showing a substantial narrowing of the polydispersity index (PDI) of high polymers occurring as a consequence of random chain scission events in a transient elongational flow field. In our experiments, semi-dilute aqueous solutions of high-molar mass, polydisperse polymers (PDI>1.4) were injected under pressure through an elongational flow field at the entrance of a capillary tube (i.d. 250 μm). Chain scission events occurring during multiple passes through the capillary entrance cause a marked decrease in PDI, to values as low as 1.12, along with the expected decrease of the average molar mass. The phenomenon appears to be entirely physical and independent of the chemical nature of the polymer, since similar results are obtained with polyacrylamide, polydimethylacrylamide, and poly(ethylene oxide). Statistical modeling of the evolution of the polymer molar mass distribution shows the results to be consistent with the random scission, near the mid-point, of those polymer chains that exceed a certain flow field-dependent critical chain length.

Introduction

Monodisperse high molar mass polymers are of interest for a number of industrial and research applications requiring stringent levels of material characterization and specific polymer phase behaviors [1], [2], [3], [4]. A few chemical methods currently exist that allow the generation of low-polydispersity index (low-PDI) polymers with high molar mass. For example, in living polymerization, chain propagation proceeds, ideally, without termination or chain transfer [5]. If short chains are terminated early, they may be ‘revived’ in some approaches by the use of highly active metallic catalysts and carried out to a target molar mass, producing a relatively narrow molar mass distribution even for large polymers. However, because of side reactions that can occur in some cases, living polymerization gives access to low-PDI high polymers (i.e. polymers with PDI<1.20 and Mw>200,000g/mol) of many, but not all chemical classes [5], [6]. One important living polymerization method for the preparation of monodisperse polymers is ring-opening metathesis polymerization (ROMP), which involves the catalytic polymerization of strained cyclic olefins in a variant of the olefin metathesis reaction [3], [5], [7], [8], [9]. The high activity of catalysts used in ROMP can cause intermolecular reactions in some classes of monomers, which can broaden the final molar mass distribution [5]. Generally, living polymerization methods require carefully controlled reaction conditions and special metallic catalysts [10], and are most convenient to use for those polymeric species without reactive side-chains and subgroups, unless protecting groups are employed. In the following, we present an entirely different route to the generation of low-polydispersity polymeric materials of high molar mass, by a physical method that exploits an iterative process of controlled polymer degradation in a transient extensional flow field. Before describing the method and our results, we give background on prior, related observations.

It has long been known that high shear from mechanical action can cause chemical bonds in polymer chains to break [11]. Some of the earliest work in this area used statistical methods to treat a process of random chain scission [12], [13], [14]. Most of these early theoretical studies modeled initial polymer populations that were monodisperse (PDI=1.0). Since then, the mechanical degradation of polymers in elongational flow fields, which leads to a reduction in average polymer molar mass, has been widely observed and studied. Passage through an elongational flow field exerts strong hydrodynamic forces upon a coiled polymer molecule in solution, causing it to stretch, orient and extend in the direction of flow. If elongational forces on the molecule are sufficiently strong, and the rate of chain stretching far exceeds the rate of chain relaxation, the polymer backbone can be severed. The precise mechanism of chain scission is not fully understood, and remains a matter of discussion in literature [15], [16], [17], [18], [19], [20]. It was theorized in 1944 by Frenkel that as the molecule becomes elongated in the direction of flow, the forces acting upon the molecule are greatest at its center [15], which can lead in turn to chain scission near the middle of the chain.

Several different flow geometries can lead to the creation of an elongational flow field. Generally, sudden contractions in the direction of flow generate a transient elongational effect, as velocity streamlines converge, which is localized in the region of the contraction. Full elongation of polymer chains most likely only occurs along the centerline of the flow field, where the molecules should not experience any shearing deformations [21]. Merrill and Leopairat designed an apparatus with a contraction ratio of 37.5 and demonstrated that chain scission occurred near the center of the molecule. They also were able to show that with an increase in the number of passes through the orifice, the net number of molecules that had experienced scission increased [22]. Nguyen and Kausch have performed extensive studies of polymer degradation in convergent flow using contraction ratios similar to those of Merrill and Leopairat [19], [23], [24], [25], [26]. Along with the expected reduction in molar mass, these authors observed a slight narrowing of the PDI of a high molar mass polystyrene sample as the total number of midpoint scission events increased [23]. At an applied strain rate of 275,000 s−1, a significant fraction of the polymers were found to undergo two degradation events during a single pass through the contraction [26]. In these transient elongational flow fields, polymer degradation was found to occur even when the residence time was insufficient to allow full elongation of the polymer molecules, and thus midpoint scission occurred in partially uncoiled polymer molecules. Other researchers have also studied the flow of polymer solutions through transient extensional flow fields created by an abrupt contraction, using contraction ratios on the order of 4:1 or 8:1 [27], [28], [29], [30].

So-called constant or homogeneous elongational flow fields, which can also lead to flow-induced chain degradation, are more difficult to achieve physically. The use of a cross-slot device, which generates an opposing-jets geometry, can create an elongational effect in which a given fluid element or polymer molecule will have a residence time long enough to achieve a constant strain rate. The geometry generates a stagnation point (zero velocity point) in the center of the flow field, where the molecules can experience full elongation. Hence, in a cross-slot device, chain extension is limited to a small area of the flow field. These quasi-steady-state flow fields have been studied by Odell and Keller and colleagues [20], [31], [32], [33], [34], [35], [36], [37], [38]. Owing to the small number of molecules that pass close to the stagnation point and undergo elongation and chain scission, a large number of passes through the system, at least 250, typically are needed to observe central chain scission [35]. Birefringence measurements were used to determine the extent of chain elongation [35]. In both transient and constant elongational flow fields, the critical fracture strain rates for chain scission are observed to decrease with increasing molar mass, with the specific dependence on molar mass (M−1 or M−2) differing as a function of the instrument and conditions used [24], [37].

In at least two other degradation studies in addition to that of Nguyen and Kausch mentioned above [23], a decrease in the polydispersity of a polymer solution was observed under some conditions [39], [40]. For example Ballauff and Wolf [39] examined the shearing of polystyrene in trans-decalin using a Couette device, while Tanigawa et al. [40] studied the degradation by ultrasonication of three nucleic acid polymers in water. For one of the DNA solutions, a reduction in PDI from 1.7 to 1.2 was reported (for DNA polymers with initial Mw of 3.1×105 g/mol and final Mw of 6.4×104 g/mol). However, such significant PDI reductions have not been shown for any other polymer classes, nor achieved by any other degradation method besides sonication.

In this work, we examine the evolution of the molecular mass distribution of three different, initially very polydisperse and high-Mw (>2.5×106 g/mol) polymer solutions as they are forced through a sudden constriction (contraction ratio∼100:1) comprising a 250 μm-ID capillary. To our knowledge, this is the first polymer degradation study involving iterative capillary entrance flow; we have examined the effects of this treatment on three industrially important water-soluble polymers, including linear polyacrylamide (LPA), polydimethylacrylamide (pDMA), which should have backbone–backbone bonds of similar intrinsic strength, and poly(ethylene oxide) (PEO). Our experimental results show that with multiple passes through the constriction, the expected progressive reduction in weight-average molar mass (Mw) is accompanied by a remarkable drop of the PDI for all three polymer samples, to values in the range of 1.12–1.15.

In order to better characterize this process, we also developed and applied to the data a simple statistical model with three physically motivated free parameters. One goal of this modeling study was to show that, based on a few reasonable assumptions having a firm basis in the literature, we could explain the dramatic PDI reductions that we observe. Analysis of the experimental data with this model indicates that the evolution of the molar mass distribution can be explained by a sequence of chain scission events, where the probability of breakage increases sharply beyond a flow field-dependent critical molecular size. We also conclude based on this analysis that scission occurs predominantly in the central ≅20% of the chain.

Section snippets

Experimental

High molar mass, polydisperse polymers (Mw>2.5×106g/mol, PDI>1.4) were injected in aqueous solution into fused-silica capillaries (Polymicro Technologies, Phoenix, AZ) by means of a custom-built pressure-loading device (Fig. 1). LPA (Mw 4.10×106 g/mol) and pDMA (Mw 4.30×106 g/mol) were prepared by free-radical polymerization. Acrylamide (Amresco, Solon, OH) and N,N-dimethylacrylamide (Monomer–Polymer and Dajac Labs, Inc., Feasterville, PA) were polymerized in an aqueous solution (7.0% (w/v) total

Experimental results

Results of the experiments are shown in Fig. 3(a)–(c), in plots of differential weight fraction vs. polymer molar mass. Each experiment for which results are shown in these figures was performed with a polymer solution concentration of c=0.1% (w/v) using a 15 cm long capillary with an inner diameter of 250 μm, and an applied injection pressure of 1000 psi. In these figures, the integrated area under a given peak represents the normalized sum of the total polymer mass in the sample. As expected

Conclusions

This study has demonstrated for the first time the potential of high-velocity capillary entrance flow to provide a general route to the creation of low-polydispersity high polymers, through a systematic reduction in PDI and Mw obtained by multiple passes through a well-defined contraction. It has never been shown before that such low PDI values (1.12–1.15) could be obtained via chain degradation, particularly for such high molar mass polymers (>200,000 g/mol). Experiments with the three

Acknowledgements

A.E.B. and B.A.B. acknowledge the Arnold and Mabel Beckman Foundation (Beckman Young Investigator Award) for support, and would like to thank Dr Wesley Burghardt and Victor Beck for helpful discussions on the subject and Donald H. Heckenberg for initial work on the project. G.W.S. and M.K. acknowledge a Research Grant and a Scholarship, respectively, from the Natural Sciences and Engineering Research Council of Canada (NSERC). M.K. would also like to thank the University of Ottawa for a

References (52)

  • T.Q. Nguyen et al.

    Polymer

    (1992)
  • T.Q. Nguyen et al.

    J Non-Newtonian Fluid Mech

    (1988)
  • T.Q. Nguyen et al.

    Polymer

    (1997)
  • L.M. Quinzani et al.

    J Non-Newtonian Fluid Mech

    (1994)
  • D.V. Boger et al.

    J Non-Newtonian Fluid Mech

    (1992)
  • P.J. Coates et al.

    J Non-Newtonian Fluid Mech

    (1992)
  • A.J. Muller et al.

    Polymer

    (1992)
  • S.P. Carrington et al.

    Polymer

    (1997)
  • W. Lee et al.

    Macromolecules

    (2000)
  • S.T. Trzaska et al.

    Macromolecules

    (2000)
  • B.R. Maughon et al.

    Macromolecules

    (2000)
  • A. Hirao et al.

    Acta Polym

    (1998)
  • O. Redakcji et al.

    Polimery

    (2001)
  • K. Matyjaszewski

    Macromol Symp

    (2001)
  • J. Qiu et al.

    Polimery

    (2001)
  • H.D. Maynard et al.

    Macromolecules

    (2000)
  • C.W. Bielawski et al.

    Macromolecules

    (2001)
  • M. Kobayashi et al.

    Macromolecules

    (1999)
  • H. Staudinger

    Die Hochmolekularen Organischen Verbindugen

    (1932)
  • W. Kuhn

    Ber Chem Dtsch Ges

    (1930)
  • W.Z. Kuhn

    Physik Chem (A)

    (1932)
  • E.W. Montroll et al.

    J Chem Phys

    (1940)
  • J. Frenkel

    Acta Physicochim URSS

    (1944)
  • M. Ballauff et al.

    Macromolecules

    (1981)
  • H.R. Reese et al.

    J Chem Phys

    (1990)
  • K.D. Knudsen et al.

    Macromolecules

    (1996)
  • Cited by (0)

    1

    Present Address: Department of Chemical Engineering, Stanford University, Palo Alto, CA 94305, USA.

    View full text