in Wiley InterScience (www.interscience.wiley.com).Terminal and penultimate models for controlling copolymer composition distribution (CCD) in ethylene and norbornene (NB) copolymerization were developed by taking into account the variation of active site concentration with the initial comonomer ratio. The models were validated by batch polymerization experimental data. The terminal model gave better correlation with the composition data whereas the penultimate model had a better fit to the rate data. The terminal model was then used to design NB feeding policies in semicontinuous processes for targeted CCD profiles. Based on the model results, a series of ethylene-NB copolymers with various NB contents were prepared. With the same NB content, the semicontinuous process produced a uniform composition, whereas the batch process yielded broad CCD. The batch samples had lower T g values and broader transition ranges, even yielded crystalline materials. In contrast, the semicontinuous samples overcame the disadvantages. (a) Copolymer yield vs. reaction time; (b) NB conversion vs. reaction time. Conditions: catalyst concentration 5 4 3 10 25 mol/l, Al/Zr 5 1500, toluene 5 50 ml, P 5 1 atm, T 5 708C, initial NB/E 5 19.6, total reaction time 5 90 min. Conditions: catalyst concentration 5 4 3 10 25 mol/l, Al/Zr 5 1500, toluene 5 50 ml, P 5 1 atm, T 5 708C.Figure 5. Copolymer yield/NB conversion vs. reaction time in the batch process for initial NB/E 5 1.3 (For the model calculation, c * varies with the initial NB/E according to Figure 4). (a) Copolymer yield vs. reaction time; (b) NB conversion vs. reaction time. Conditions: catalyst concentration 5 4 3 10 25 mol/l, Al/Zr 5 1500, toluene 5 50 ml, P 5 1 atm, T 5 708C, initial NB/E 5 1.3, total reaction time 5 10 min.