A probability theory is presented to judge the utility of the internal
standard method in high-performance liquid chromatography (HPLC). The
relative standard deviation (RSD) of experimental results in the inte
rnal standard method is theoretically predicted based on the recently
proposed uncertainty theory for sample dilution and HPLC measurement.
The experiments comprise the taking of sample solution, addition of in
ternal standard, dilution, measurement on HPLC and signal processing.
Two kinds of correlation between concentrations of analyte and interna
l standard are formulated distinctively in the sample preparation proc
ess and in the HPLC measurement. The RSD values for the total experime
ntal process are calculated from the volume error of pipette, signal s
hape of analytes and some probabilistic properties of baseline drift i
n chromatograms and are shown to be in good agreement with the observe
d ones over a wide range of dilution degree. The external calibration
is also examined.