A previous implementation of optimal dynamic experiments for the identification of the Cardinal Temperature Model with Inflection, which describes the temperature effect on the microbial growth rate, revealed that the maximum growth temperature T(max) can only be estimated accurately when temperatures close to the true T(max), are included. Here, alternative approaches are considered. First, it is evaluated if the informative character of the optimal experiments can be improved by changing the temperature input structure. Simulations show that a simple profile, i.e., a constant phase followed by a linear temperature change and a second constant phase, is the most informative taking into account computational requirements and practical implementations. Second, experiments in which the temperature crosses the growth/inactivation interface are evaluated. The simulation study illustrates that the accuracy of T(max) largely depends on the selected temperature profile. A reliable T(max) is obtained when the initial temperature is slightly lower than the true T(max) Starting above T(max) result in larger uncertainties. Generally, information at temperatures close to T(max) is required for an accurate T(max) estimate. (C) 2011 Elsevier Ltd. All rights reserved.