This paper assesses modeling parameters that affect the environmental performance of two state-of-the-art photovoltaic (PV) electricity generation technologies: the PVL136 thin film laminates and the KC120 multi-crystalline modules. We selected three metrics to assess the modules' environmental performance, which are part of an actual 33 kW installation in Ann Arbor, MI. The net energy ratio (NER), the energy pay back time (E-PBT), and the CO2 emissions are calculated using process based LCA methods. The results reveal some of the parameters, such as the level of solar radiation, the position of the modules, the modules' manufacturing energy intensity and its corresponding fuel mix, and the solar radiation conversion efficiency of the modules, which affect the final analytical results. A sensitivity analysis shows the effect of selected parameters on the final results. For the baseline scenario, the E-PBT for the PVL136 and KC120 are 3.2 and 7.5 years, respectively. When expected future conversion efficiencies are tested, the E-PBT is 1.6 and 5.7 years for the PVL136 and the KC120, respectively. Based on the US fuel mix, the CO, emissions for the PVL136 and the KC120 are 34.3 and 72.4 g of CO2/kW h, respectively. The most effective way to improve the modules' environmental performance is to reduce the energy input in the manufacturing phase of the modules, provided that other parameters remain constant. Consequently, the use of PV as an electricity source during PV manufacturing is also assessed. The NER of the supplier PV is key for the performance of this scheme. The results show that the NER based on a PV system can be 3.7 times higher than the NER based on electricity supplied by the traditional grid mix, and the CO2 emissions can be reduced by 80%. (c) 2006 Elsevier Ltd. All rights reserved.