Session N4 - Crystallization.
INVITED session, Wednesday morning, March 24
517C, Palais des Congres

[N4.001] Evidence for a multi-stage route followed in polymer crystallization

Conventional wisdom assumes for crystallizing polymers a direct growth of the layer-like crystallites into the melt, associated with activation steps encountered during the transfer of chain segments from the melt onto the growth face. Crystal thicknesses and growth rates then have to be controlled by the supercooling below the equilibrium melting point. This was taken for granted during long time by all workers in the field, however, a comprehensive SAXS-study on s-PP and derived copolymers then yielded a contradicting result: Thicknesses of crystals obtained by isothermal crystallization turned out to be inversely proportional to a temperature which is more than 20C above the equilibrium melting point and furthermore, to be independent of the co-unit content. We understood the behavior as indicating that the growth of polymer crystals follows in general a multi-stage route which includes a passage through transient states. Based on the observations, a model was developed which proposes a route via mesomorphic and granular crystalline layers to the final lamellar crystallites [1]. To check the model more experiments were carried out, in recent time a) An investigation on the effect of co-units and diluents on the lamellar structure of PE (with T.-Y. Cho and B. Heck [2]):While in the presence of the two n-alkanes thicker crystals form, no effect arises for the methyl-anthracene and octene co-units. b) A study of the crystallization of s-PP from the glassy state (with M.Grasruck [3]): Annealing of sPP after a rapid quench from the melt to a temperature below the glass transition leads for an annealing temperature 0C to a mesomorphic phase, for 25C to the helical crystalline phase. Crystal thicknesses and long spacings are equal. c) A study of the crystallization of PCL (with P.Kohn): The WAXS-pattern is composed of three contributions, associated with melt-like regions, crystals and a mesomorphic phase. [1] G.Strobl: Eur. Phys. J. E 3, 165, 2000. [2] B.Heck, G.Strobl, M.Grasruck: Eur.Phys.J.E 1,117, 2003. [3] M.Grasruck, G.Strobl: Macromolecules 36, 86, 2003.

[N4.002] Shifting Paradigms in Polymer Crystallization

We have investigated the molecular mechanisms of primordial stages of polymer crystallization from solutions using Langevin dynamics simulations and theoretical models. The experimental observation of finite lamellar thickness (much smaller than extended chain dimension) has been historically attributed to kinetic origins, and it is believed that equilibrium thickness is the extended chain dimension. Our simulations and theory show that the finite lamellar thickness is actually the equilibrium result. The key feature that distinguishes polymer crystallization is that a single chain can participate in several nuclei. The resulting entropic frustration leads to spontaneously selected finite structures during nucleation. In addition, the crystallization growth is found in our simulations to be dominated by chain adsorption and highly cooperative dynamics of all chains. Finally, we have derived criteria for the onset of mesomorphic states in polymer crystallization. Our results contradict the conventional assumptions and theories of polymer crystallization.

[N4.003] Discovery of Reversible Crystallization of Macromolecules

For 10 years “reversing melting” was observed with temperature-modulated differential scanning calorimetry, TMDSC. This reversing melting is the first harmonic response beyond that caused by the heat capacity of a metastable, semicrystalline macromolecular sample. Before one can identify “reversible melting,” the calorimeter response must be corrected for loss of linearity, stationarity, frequency, amplitude, and instrument lag, or proper experiment-design must avoid these problems. Using quasi-isothermal TMDSC, the following observations were made [Prog. Polymer Sci. 28 (2003) 383-450]: Equilibrium crystals of polymers may melt at the equilibrium melting-temperature, but crystallization needs supercooling, even in the presence of crystal nuclei, making the overall process irreversible. Metastable, folded-chain crystals of the same molecules also melt irreversibly, however, may have some specific reversibility. Flexible, linear molecules of up to 10 nm length may melt fully reversibly. Macromolecules of less flexibility may lose the ability to melt reversibly. Decoupling of molecular segments, molecular nucleation, segregation of molar masses, rigid amorphous fractions, effects of equilibrium point defects in crystals and glasses, and transition-less ordering and solidification will be discussed in some detail.

Supported by NSF, Polymers Program, DMR-0312233, and the Div. of Mat. Sci., BES, of DOE at ORNL, managed by UT-Battelle, LLC, for the U.S. Department of Energy, DOE-AC05-00OR22725.

[N4.004] Kinetics in melting of polymers

With the help of controlled chemistry it is now possible to synthesise polyethylenes having a "single chain forming single crystal". Following the build-up modulus, entanglement process of these "monomolecular crystals" at 180°C is followed. The molten state shows strong heating rate dependance on the rheological properties. For example, the melt obtained by slow heating of crystals maintains partially "dientangled state" compared to the crystals that are heated fast from room temperature to 180°C. Heating rate dependance on different melt structure invokes role of kinetics in melting of monomolecular crystals. The paper will also address physical properties of different melt structures obtained in the same polymer.

[N4.005] Structure Formation in Supercooled Polymer Melts - Some Ideas from Molecular Dynamics Simulations with Slightly Coarse-Grained Models

We developed a simplified polymer model which appeared to be very efficient for the study of polymer crystallization from the dense melt using molecular dynamics simulations. It was investigated with different temperature protocols: (i) continuous cooling to study the nucleation process of short chains; (ii) quench below the melting temperature and subsequent isothermal relaxation of long chains to study the formation of chain-folding and lamellar formation as a function of supercooling. The key results are the following:

(i) The chain length dependence of crystallization and melting temperatures is similar to that of alkanes. Visualization of the critical nucleus shows a thickness of about 10 monomers for all chain length. There is a cross-over from extended-chain to folded-chain nuclei in a domain where the final crystal is still formed of extended chains. For longer chains (N=100), crystallization may start from an intra-chain nucleus formed by tight hairpins on a chain which is more compact than the average chain.

(ii) For long chains (up to N=400 monomers) in an isothermal relaxation protocol, it was possible for the first time in MD simulations to determine crystallization and melting lines: the average thickness of crystal domains is smaller the larger the supercooling, and the melting temperature increases with average stem length, as expected from experiments. The formation of lamellar structures does not proceed at a sharp growth front but in a growth zone several chain diameters thick. Hairpins seem to be very active during the growth.

It is particularly interesting that these results were obtained with a model containing no explicit attraction between monomers. This means that the conformational statistics and the orientational ordering are sufficient driving forces for the formation of chain-folded lamellae.

\smallskip

[1] H. Meyer and F. Müller-Plathe, J. Chem. Phys. 115 (2001) 7807; Macromolecules 35 (2002) 1241; H. Meyer in Lecture Notes in Physics 606, J.U. Sommer, G. Reiter (eds.) Springer 2003.

Part N of program listing