It is well known that poor quantitative reproducibility substantially limit
s the practical implementation of capillary electrophoresis (CE) separation
s in chemical analysis. The principal sources of variance in observed peak
areas are irreproducible flow rate, which influences on-column detector res
ponse, and inconsistent injection volume or amount. An overview of studies
by researchers to address the reproducibility issue will be presented. In a
ddition, current efforts in our laboratory to assess sources of quantitativ
e variance for separations of dansylated amino acids using an automated CE
system are presented and related when appropriate to the body of existing k
nowledge on this important topic. A comparison of different injection metho
ds (hydrostatic vs. electrokinetic) and approaches (e.g., high vs. low pres
sure), the effect of random changes in electroosmotic flow (EOF) due to air
bubbles in the CE capillary, and choice of certain peak integration parame
ters in terms of peak area reproducibility are presented. Under optimum con
ditions relative standard deviation (RSD) values in raw peak area are typic
ally 2.0%. With nonoptimum conditions (e.g., with air bubbles in capillary)
, RSD values can substantially degrade. However, normalizing with retention
times, internal standards, or observed electrophoretic current produces RS
D values in a range of 1.4-2.3%.