Presenting theoretical arguments and numerical results we demonstrate long-range intrachain correlations in concentrated solutions and melts of long flexible polymers which cause a systematic swelling of short chain segments. They can be traced back to the incompressibility of the melt leading to an effective repulsion u(s) ≈ s/ρR 3 (s) ≈ c e / √ s when connecting two segments together where s denotes the curvilinear length of a segment, R(s) its typical size, c e ≈ 1/ρb 3 e the "swelling coefficient", b e the effective bond length and ρ the monomer density. The relative deviation of the segmental size distribution from the ideal Gaussian chain behavior is found to be proportional to u(s). The analysis of different moments of this distribution allows for a precise determination of the effective bond length b e and the swelling coefficient c e of asymptotically long chains. At striking variance to the short-range decay suggested by Flory's ideality hypothesis the bond-bond correlation function of two bonds separated by s monomers along the chain is found to decay algebraically as 1/s 3/2 . Effects of finite chain length are considered briefly. PACS numbers: 61.25.Hq,64.60.Ak,05.40.Fb * Electronic address: jwittmer@ics.u-strasbg.fr † URL: http://www-ics.u-strasbg.fr/~etsp/welcome.php 1 I. FLORY'S IDEALITY HYPOTHESIS REVISITEDA cornerstone of polymer physics. Polymer melts are dense disordered systems consisting of macromolecular chains [1]. Theories that predict properties of chains in a melt or concentrated solutions generally start from the "Flory ideality hypothesis" formulated already in the 1940s by Flory [2,3,4]. This cornerstone of polymer physics states that chain conformations correspond to "ideal" random walks on length scales much larger than the monomer diameter [1,4,5,6]. The commonly accepted justification of this mean-field result is that intrachain and interchain excluded volume forces compensate each other if many chains strongly overlap which is the case for three-dimensional melts [5]. Since these systems are essentially incompressible, density fluctuations are known to be small. Hence, all correlations are supposed to be short-ranged as has been systematically discussed first by Edwards who developed the essential statistical mechanical tools [6,7,8,9, 10] also used in this paper.One immediate consequence of Flory's hypothesis is that the mean-squared size of chain segments of curvilinear length s = m − n (with 1 ≤ n < m < N) should scale as R Both equations are expected to hold as long as the moment is not too high for a given segment length and the finite-extensibility of the polymer strand remains irrelevant [6].Deviations caused by the segmental correlation hole effect. Recently, Flory's hypothesis has been challenged both theoretically [11,12,13,14,15] and numerically for threedimensional solutions [16,17,18,19,20] and ultrathin films [21,22]. These studies suggest 2 that intra-and interchain excluded volume forces do not fully compensate each other on intermediate length scales, l...

We present results of molecular dynamics simulations for free-standing and supported thin films of a nonentangled polymer melt using a coarse-grained (beadspring) model. Our discussion is mainly concerned with the equilibrium properties of the films above the critical temperature (T c ) of mode-coupling theory, although we also determine the glass-transition temperature (T g ) by measurements of the film thickness h upon cooling. We explore the influence of confinement on the structure and dynamics of the polymer films. We find that the dynamics in the films is accelerated compared to the bulk, that this enhanced mobility originates from the surfaces, and that the effect is larger at the free than at the supported surface. Thus, the films have lower T c values relative to the bulk. T c depends on film thickness h; this dependence can be well parametrized by T c (h) = T bulk c /(1 + h 0 /h), a function proposed by experiments on supported polystyrene films.

The scaling of the bond-bond correlation function P1(s) along linear polymer chains is investigated with respect to the curvilinear distance, s, along the flexible chain and the monomer density, ρ, via Monte Carlo and molecular dynamics simulations. Surprisingly, the correlations in dense three dimensional solutions are found to decay with a power law P1(s) ∼ s −ω with ω = 3/2 and the exponential behavior commonly assumed is clearly ruled out for long chains. In semidilute solutions, the density dependent scaling of P1(s) ≈ g −ω 0 (s/g) −ω with ω0 = 2 − 2ν = 0.824 (ν = 0.588 being Flory's exponent) is set by the number of monomers g(ρ) contained in an excluded volume blob of size ξ. Our computational findings compare well with simple scaling arguments and perturbation calculation. The power-law behavior is due to self-interactions of chains caused by the chain connectivity and the incompressibility of the melt. This study suggests a careful reexamination of the operational definitions used for the experimental determination of the persistence length.PACS numbers: 05.40. Fb, 05.10.Ln, 61.25.Hq In this Letter we study the correlations of the directions of bonds along polymer chains in semidilute solutions and melts [1,2,3]. We focus on flexible monodisperse chains of N monomers (cf. Fig. 1) under good solvent conditions in three dimensions (d = 3) where both the bond length l and the excluded volume screening length ξ [2, 4] are always much smaller than the chain end-to-end distance R e . In principle, once the bond-bond correlations are computed all other conformational single chain properties can be derived. Importantly, being the (second) derivative of the spatial distances along the chains, they allow us to probe directly -without trivial ideal contributions -the non-gaussian corrections proposed recently [5]. As we shall see, these corrections are crucial to make the description of dense polymer systems, first proposed by Flory [3] and later corroborated by Edwards [2, 4], fully self-consistent.The bond-bond correlation function P 1 (s) is generally believed to decrease exponentially [3]. This belief is based on the few simple single chain models which have been solved rigorously [3,6] and on the assumption that all long range interactions are negligible on distances larger than ξ due to the screening mechanism described by Edwards [2,4]. Hence, only correlations along the backbone of the chains are expected to matter and it is then straightforward to work out that an exponential cut-off is inevitable due to the multiplicative loss of any information transferred recursively along the chain [3].We demonstrate here that this assumption is in fact incorrect and that unexpected long range correlations remain. They are responsible for a scale free power law regime with P 1 (s) = c a (ρ)s −ω for g(ρ) ≪ s ≪ N (g(ρ) being the number of monomers per blob at monomer density ρ) characterized by an exponent ω > 1 and a density dependent amplitude. Our simulation results are pre- sented first and discussed together wi...

We develop coarse-grained force fields for poly (vinyl alcohol) and poly (acrylic acid) oligomers. In both cases, one monomer is mapped onto a coarsegrained bead. The new force fields are designed to match structural properties such as radial distribution functions of various kinds derived by atomistic simulations of these polymers. The mapping is therefore constructed in a way to take into account as much atomistic information as possible. On the technical side, our approach consists of a simplex algorithm which is used to optimize automatically non-bonded parameters as well as bonded parameters. Besides their similar conformation (only the functional side group differs), poly (acrylic acid) was chosen to be in aqueous solution in contrast to a poly (vinyl alcohol) melt. For poly (vinyl alcohol) a non-optimized bond angle potential turns out to be sufficient in connection with a special, optimized non-bonded potential. No torsional potential has to be applied here. For poly (acrylic acid), we show that each peak of the radial distribution function is usually dominated by some specific model parameter(s). Optimization of the bond angle parameters is essential. The coarse-grained forcefield reproduces the radius of gyration R G of the atomistic model. As a first application, we use the force field to simulate longer chains and compare the hydrodynamic radius R H with experimental data.