A dielectric barrier discharge (DBD) plasma, operating in air at atmospheric pressure, has been used to induce changes in the surface properties of polyethylene terephthalate (PET) films. The effects that the key DBD operating parameters: discharge power, processing speed, processing duration, and electrode configurations, have on producing wettability changes in the PET surface region have been investigated. The approach taken involves the application of an Taguchi experimental design and robust analysis methodology. The various data sets obtained from these analyses have been used to studies the effect of the operating parameters on the surface uniformity and efficiency of the said treatment. In general, the results obtained indicate that DBD plasma processing is an effective method for the controlled surface modification of PET. Relatively short exposures to the atmospheric pressure discharge produces significant wettability changes at the polymer film surface, as indicted by pronounced reductions in the water contact angle measured. It was observed that the wettability of the resultant surface shows no significant differences in respect to orientation parallel (L-direction) or perpendicular (T-direction) to the electrode long axis. However, there was significant differences between the data obtained from these two orientations. Analysis of the role of each of the operating parameters concerned shows that they have a selective effectiveness with respect to resultant surface modification in terms of uniformity of modification and wettability. The number of treatment cycles and the electrode configuration used were found to have the most significant effects on the homogeneity of the resultant PET surface changes in L- and T-orientation, respectively. On the other hand, the applied power showed no significant role in this regard. The number of treatment cycles was found to be the dominant factor (at significance level of 0.05) in respect of water contact angle changes at the processed PET surface in both orientations. The driven metal electrodes (stainless steel or aluminium) were apparently superior to the driven dielectric electrode (ceramic or quartz) configurations. The grounded electrode in each case was a silicon rubber-covered aluminium plate (see later). The nature and scale of the surface changes that originate from the various processing conditions employed have been considered so as to determine the optimum treatment conditions in respect of processing outcomes, properties and any orientation dependence. Thus, it was revealed that higher processing speeds and longer processing durations are key for uniformity along the electrode axial orientation, while lower processing speeds and short exposure durations are key considerations, in the corresponding perpendicular orientation. In general, longer processing durations (low processing speeds and a high number of treatment cycles) and higher plasma powers induced greater changes in the surface wettability of the PET, as demonstrated by the observed water contact angles. This behaviour is taken to indicate that different combinations of DBD operating parameters and electrodes produce discharge conditions that can result in different plasma chemical processes in respect of uniformity, treatment efficiency and orientation dependence. (c) 2005 Elsevier B.V. All rights reserved.