The preparation and characterization of randomly cross-linked polybutadiene (PB) networks containing unattached PB chains are presented. These were used as model systems for the experimental study of the dynamics of linear chains trapped in networks, which is the basic model of the reptationin-a-tube model of Doi and Edwards. The preparation consists of the conversion at random of a small fraction of the PB chains' double bonds into epoxide groups and in the subsequent selective use of these as cross-linking centers in the presence of unmodified PB chains; these last remain chemically unattached to the network. The selectivity of the cross-linking reaction was examined by swelling-extraction of networks containing known fractions of relatively short "monodisperse" unattached chains. The polymeric phase composition of the networks with unattached chains (nuc) was tested by glass transition temperature measurements. Dynamic-mechanical measurements were shown to be sensitive to the polymer-phase composition of nuc, as well. The dynamics of the linear PB chains in the melts and in the networks was studied by dynamic-mechanical measurements. The experimental results for the longest relaxation time τ L were compared with the predictions of the reptation-in-a-tube model with and without chain-end effects. Within experimental accuracy, the obtained scaling of τL relative to the chain mass M (up to 9 × 10 5 g/mol) is the same in both media: τL ∼ M x with x ) 3.35 ((0.10 as against x ) (3 from the pure reptationin-a-tube model; correcting for chain-end effects in the form of contour length fluctuation, results in a good description of the experimental data for the τLs of unattached chains. τL for the chains trapped in networks were found to be 1.9-3.2 times longer than the τL of the respective melts. Data on 14 PB prepared by anionic polymerization and covering the range of weight average molar masses 1.0 × 10 4 to 1.5 × 10 6 g/mol are presented. IntroductionRubber networks with unattached chains 1-7 are ideally one-phase, two-component systems; the first component is one macroscopic molecule (the network) and the second consists of linear chains "dissolved" into the network. The materials (with or without phase separation) are also classified in the literature as "semiinterpenetrating networks" (sIPN). 8 Because the short time dynamics (glass and transition regions 1 ) of a rubber network is essentially the same as that of a high polymer melt, 1 the rubber networks containing small concentrations of unattached chains permit one to study the dynamics of almost isolated chains in the rubber plateau and the terminal zones. 1 Therefore, these systems can be used to test the predictions of one-chain theoretical models (specifically the reptation-in-a-tube model 9-12 ) and as references for the study of polymer melts. On the other hand, the same systems are well-fitted for the investigation of the longtime (eventually equilibrium) properties of rubber networks as depending on the degree of cross-linking. The analysis of the con...