Macromolecules, Vol.42, No.8, 3170-3178, 2009
Linear and Nonlinear Elastic Behaviors of Star Polymers
In this paper, we investigate the elastic behavior of M-arm star polymers in good solvents under an external force X exerted on each end of the arms in the x direction. In the region of small X, there is a slight difference in the linear elastic behaviors of a single chain and star polymers. In the case of a single chain, the mean x component of the end-to-end vector < x(N)> is rigorously equal to (R0X)-X-2/3k(B)T, where R-0(2) = < r(N)(2)>(0) is the mean square end-to-end distance at X = 0. In contrast, the mean x component of the center-end vectors < x(N)((i))> of star polymers is not simply given by R-0(2)(M)X/3k(B)T, where R-0(2)(M) = < r(N)((i)2))(0) is the mean square center-end distance at X = 0. Instead, we show that the rigorous relation < x(N)((i))>/X = [R-0(2)(M)/3 + (M-1)< x(N)((i))x(N)((i))>(1 not equal j)(0)]/k(B)T holds for small X, where < x(N)((i))x(N)((i))>(1 not equal j)(0) denotes the mean product of the x components of the center-end vectors of the ith and jth arms in the absence of X. Here we perform large-scale Monte Carlo simulations of star self-avoiding walks (SAWs) on a simple cubic lattice with a lattice constant a. By using the enrichment algorithm generating 1000 000 samples of star SAWs with M = 2, 3, 6 and 9 arms, each of which has the same chain length N up to 300, under the external force X, we find that the elastic behaviors of star polymers very much resemble those of a single chain obtained in our previous paper (Macromolecules 2008, 41, 4037.), and are characterized as < x(N)((i))> proportional to X (for very small X) and < x(N)((i))> proportional to X-2/3 (for 0 << X similar to k(B)T/a). The crossover between these two behaviors is smooth, and the crossover point is very much close to that of a single chain. Next, using the enrichment Monte Carlo simulation in the absence of the external force, we find that R-0(2)(M) is an increasing function of M and < x(N)((i))x(N)((i))>(i not equal j)(0) is a small negative quantity. This is consistent with our Monte Carlo result for the small X behavior of < x(N)((i))>/X that weakly decreases with increasing M. Also, by using a renormalization-group epsilon = 4 - d expansion up to the first order in epsilon, we confirm that < x(N)((i))x(N)((i))>(1 not equal j)(0) is a small negative quantity, whose small absolute value is independent of M and proportional to N. Finally, we briefly comment on the application of the present study to more general polymer networks.