Enzyme Inactivation Kinetics in Carrot Extract The first-order inactivation kinetics model was tested for its applicability to thermal inactivation data for carrot extract. Linear regression of Eq. (2) was carried out using the least squares technique and the coefficients were deter- mined (Tables 1 and 2). Though it yielded good R2 (>0.88) and low SE values at all temperatures, the residual plot showed a patterned behavior (Eq. 2, Table 1). Consequently, the distinct isozyme model (Weemaes et al. 1998), based on the presence of two isoenzyme populations inactivated by the first- order mechanism, was tested. Although this model showed a high R2, low SE
values and random residual pattern at all the temperatures, the model gave equal values for parameters k1 and k2 (Eq. 3, Table 3), which nullifies the presence of isoenzymes. Therefore, both first-order and distinct isozyme models were rejected for fitting of the enzyme activity with the time of thermal processing.
The series model (Eq. 5, Table 4) and the fractional conversion model (Eq. 7, Table 5) were not found to satisfy the physical criteria because of the negative parameter estimates. The general nth order (Eq. 6) was consequently considered. A poor fit of the data with the model equation and the highly
patterned behavior of the residual plots led to the rejection of this model (Table 6). It may therefore be inferred that none of the models based on the preconceived mechanism could adequately explain the thermal inactivation pattern. The Weibull function (Eq. 8) was then considered for the estimation of POD inactivation parameters. Statistical (low SE, high R2), physical (nonnegative parameter estimates) and random residual pattern indicated that the Weibull distribution model is the best model to describe the kinetics of POD inactivation of carrot extract (Table 7). The values of the scale factor (b) ranged from 0.641 to 2.669, while the shape factor (n) varied from 0.568 to 1.117 for the entire temperature range.The POD activity curves fitted with the Weibull distribution function at the temperature range of 80–100C are shown in Fig. 7.Arepresentative residual pattern of POD activity at 80C as depicted by theWeibull distribution function is shown in Fig. 8.The parity between the experimental data and the activity predicted by the Weibull distribution is shown in Fig. 9. The broken lines in Fig. 9 represent the 5% confidence
interval of the experimental and predicted values. It can be seen that all the data lie within the 5% confidence interval.