The precision of instruments used in 3 collaborative studies conducted within the Food and Drug Administration over a 4-year period (1981, 1982, 1984) for water activity (aw) determinations according to the official AOAC method is evaluated. Calibration responses of the instruments were tested for linearity over the aw range from 0.75 to 0.97. Average absolute percent difference between predicted and assigned aw values for the linear model ranged from 0.3 to 0.7% for a fiber-dimensional hygrometer (Abbeon) and 3 electrical hygrometers (Beckman, Rotronics, and Weather Measure). The calibration responses for another electrical hygrometer (Hygrodynamics) were nonlinear. The fiber-dimensional hygrometer yielded mean aw values and precision estimates that did not differ significantly from those obtained with the electrical hygrometers for (NH4)2SO4slush, KNO3 slush, sweetened condensed milk, pancake syrup, and cheese spread. However, the mean aw value for a soy sauce was 0.838 for the electrical hygrometers compared with 0.911 for the fiber-dimensional hygrometer. The fiber-dimensional hygrometer was affected by a volatile component(s) in the soy sauce that caused an erroneously high aw value. Pooled estimates of reproducibility (Sx) in the 3 studies were 0.008 for the fiber-dimensional hygrometer and 0.010 for the electrical hygrometers; these values were not significantly different from those reported in the study that verified the current official AOAC method.
Read full abstract