【正文】 ther chains present in the same region, due to their lower concentration. The overall force acting on the test chain is thus zero, and hence the chain remains ideal. The current picture of melt morphology emerged from this evidence. It is considered that chains in polymer melts are unperturbed, behaving as ideal. Energetic interactions involving segments of the same, or di erent, chains are inexistent, and changes in chains’ free energy result only from entropic variations originated by changes in chain conformations. Chains of ideal chain models consist of immaterial links, each one joined to two nearest neighbors, having no interactions with solvent molecules or with other links of the same, or another chain. In concentrated solutions or polymer melts, it is considered that entanglements between chains act as topological restrictions to chain motion. Entanglement e ects on the ow of polymer melts are described by tube model, reptation movement and several additional mechanisms. 3,4 At the present, it is generally accepted that polymer melts are nearly ideal. Deviations from ideality have been ascribed to longrange intramolecular correlations, 5,6 and to the existence of shortrange order. Some recent results indicating “solidlike” behavior (or shortrange order) in melts 7,8 have generated controversy. 9,10 The argument is that the experiments in their basis were performed in the limit of the instruments’ working conditions, casting doubts on the results validity.Apparently uncontroversial experimental results demonstrating the existence of shortrange order can be classi ed in three types. The rst are those of phenomenological nature, such as the work of Zondervan et al. 11 that can only be understood if shortrange order is considered to exist in the liquid phase. In the second type we include the experiments by Fisher et al. 12 that quanti ed shortrange order with a correlation volume below (10 197。) 3 and also those by the group of Spiess that quanti ed this order with an order parameter (S = ), 13 and assigned to the heterogeneities present in the melt a dynamic length scale (≈ 30 197。).14 Finally, other works provided nerdetails on shortrange order by quantifying the separation distance for intermolecular interactions. In this class we include the wideangle Xray di raction experiments of Londono et that demonstrate for polyethylene melts the existence of intermolecular correlations at 4 197。. A similar result was also obtained by Mitchell et al. 16 that assigned the breadth of di use interchain peaks in scattering patterns from melts to the limited spatial correlation between chain segments, concluding further that correlated segments consisted of three to four skeletal bonds in the trans conformation .The first known work on the evaluation of shortrange order by molecular dynamics simulations in nalkane chains dates back to 1988. Rigby and Roe modeled sequences of 10 and 20 CH2 groups in ensembles of 500 or 2000 united Force elds used are similar to those of more recent They evaluated the intermolecular radial distribution function and the orientational correlation function at di erent temperatures for di erent sets of CH2 units, from 1 up to 10 (for the chain with 20 segments). It was found that the height of the peaks in the radial distribution function decreases by increasing the number of segments considered in the chains and temperature. On the other hand, by decreasing temperature the peak positions shift to slightly shorter distances and the peaks shapes sharpen. Results for the orientational correlation function showed a similar behavior, and demonstrated that in the simulated melts there exists a signi cant degree of shortrange order. The authors concluded that, on a very local scale, a shortsubchain is surrounded by a bundle of about seven parallel subchains. Furthermore, the nding that the degree of order increases by decreasing temperature led the authors to evaluate a temperature for the “possible” isotropic?nematic phase transition. The buildup of nematic order was also found by Weber et in Monte Carlo simulations of dense polymer melts. The Hamiltonian used in these simulations favors stretched bond angles and short bonds, naturally yielding therefore a Gaussian coil at high temperatures and a rigid rod at very low temperatures. Shortrange order in polymer liquids was accessed also by Abrams and Kremer 20 with molecular dynamics simulations of bead?spring models. They analyzed the ratio of bond length to excluded volume bead diameter, l0/d0 , on the equilibrium structure of bulk melts posed of freely joined bead?spring polymer chains. It was found that the characteristic disorder existing when l0/d0 1 disappears for l0/d0 1. In this last case, if bonds are long enough, signi cant intermolecular correlations exist, and the polymer uid can display high shortrange order. It was concluded then that the manner in which excluded volume is distributed along a chain of simple spheres determines the intermolecular structure in dense liquids of these chains. Here we identify and characterize shortrange order in polymer melts. We started by simulating a melt of density as similar as possible to available experimental results. Chains of di erent length were simulated with the united atom representation, and their characteristic ratio was evaluated at di erent temperatures. Since models used in the description of the ow behavior consider the Kuhn monomer as the basic repeating chain unit, the Kuhn statistical segment length, l k, and the Kuhn monomer diameter, dk were evaluated at each temperature. The only assumption made in the de nition of shortrange order is that aligned chain segments (ACS) may be de ned by con ning all chain segments to a tube of diameter lower than dk. After this de nition, a search is made for interactions of a tagged ACS with other ACS belonging to the same chain or to other chains, located at a reasonable separation