An Alternative Approach to Non-Log-Linear Thermal Microbial Inactivation: Modelling the Number of Log Cycles Reduction with Respect to Temperature
Vasilis Panagiotis Valdramidis, Kristel Bernaerts, Jan Frans Van Impe* and Annemie Helena Geeraerd
BioTeC – Bioprocess Technology and Control Department of Chemical Engineering, Katholieke Universiteit Leuven, W. de Croylaan 46, B-3001 Leuven, Belgium
Article history:
Received September 21, 2004
Accepted June 23, 2005
Key words:
predictive microbiology, thermal processing, inactivation kinetics, modelling log cycles of microbial reduction, Escherichia coli
Summary:
A mathematical approach incorporating the shoulder effect during the quantification of microbial heat inactivation is being developed based on »the number of log cycles of reduction« concept. Hereto, the heat resistance of Escherichia coli K12 in BHI broth has been quantitatively determined in a generic and accurate way by defining the time t for x log reductions in the microbial population, i.e. txD, as a function of the treatment temperature T. Survival data of the examined microorganism are collected in a range of temperatures between 52–60.6 °C. Shoulder length Sl and specific inactivation rate kmax are derived from a mathematical expression that describes a non-log-linear behaviour. The temperature dependencies of Sl and kmax are used for structuring the txD(T) function. Estimation of the txD(T) parameters through a global identification procedure permits reliable predictions of the time to achieve a pre-decided microbial reduction. One of the parameters of the txD(T) function is proposed as »the reference minimum temperature for inactivation«. For the case study considered, a value of 51.80 °C (with a standard error, SE, of 3.47) was identified. Finally, the time to achieve commercial sterilization and pasteurization for the product at hand, i.e. BHI broth, was found to be 11.70 s (SE=5.22), and 5.10 min (SE=1.22), respectively. Accounting for the uncertainty (based on the 90 % confidence intervals, CI) a fail-safe treatment of these two processes takes 20.36 s and 7.12 min, respectively.
*Corresponding author:
++32 16 321 466
++32 16 322 991