USING NUMERICAL METHODS IN CHEMOMETRICS: 
MOISTURE CONTENT DETERMINATION BY AUTOMATED KARL FISCHER TITRATION

Róbert Rónyai1, Róbert Rajkó2

1Environmental Protection Inspectorate of Lower-Tisza region,
6721 Szeged, Felső Tisza-part 17.

2 Department of Unit Operations and Environmental Engineering
College Faculty of Food Engineering University of Szeged
H-6701 Szeged, POB 433. E-mail: rajko@sol.cc.u-szeged.hu

We have developed an analytical procedure to moisture content determination by automated Karl Fischer titration based on chemometric way. The ISO 900x regulations were fully considered during developing the controlling program, so the determination can be inserted to a quality control and/or assurance process.

Conditions of the application of the Least sum of Squares Method (LSM) as the basis of the algorithm were exhaustively verified. The following conditions have to be fulfilled for the support of the advantageous properties of the LSM estimation (unbiased and efficient, simple computer code etc.):

  1. the fitted function is bordered and closed, i.e. compact, on the investigated domain,
  2. Taylor series expansion of the function exists on all points of the investigated domain,
  3. the measurement errors are additive,
  4. the measurement errors have zero expected value, and they are stochastically independent from each other,
  5. the measurement errors have finite variance and they are homoscedastic,
  6. the independent variable is not a random variable, but if it is, its error can be neglected according to the error in the dependent variable.
In our investigation the fulfillment of the above conditions was proved in detail, so the algorithm with LSM estimation developed to control the moisture content determination by automated Karl Fischer titration works efficient.