Quantification in capillary electrophoresis-mass spectrometry: long- and short-term variance components and their compensation using internal standards.

Different approaches were chosen to examine ionization reproducibility of analytes after separation by capillary electrophoresis-mass spectrometry (CE-MS) in a commercially available sheath-flow electrospray interface. For this task three different standard samples were examined. Sample 1 contained neostigmine bromide (cationic), paracetamol (PCM) (neutral) and nicotinic acid (anionic component). Results were evaluated using internal standard (IS) calculations. Sample 2 represented an isotopically labelled IS of the quantified substance (PCM/D4-PCM), while sample 3 (neostigmine bromide/scopolamine hydrobromide) provided an IS closely migrating to the tested substance. Furthermore, short-time variations inside the interface were examined by multiple injections of the same substance. For sample 1, the relative standard deviations (RSD%s) were between 8 and 25% (n at least 58) for the peak area ratios. Multiple injected samples gave 5.5-19.4% (n = 25) for peak area RSD%. Using a closely migrating IS, sample 3, RSD%s between 6.5 and 10% (n at least 63) were achieved. With isotopically labelled IS, sample 2, an RSD% of 3-4% was achieved for peak area ratios over long periods (n = 25), for shorter periods (n = 9) even 1-2% RSD% was obtained. Keeping the instrument settings constant, the influence on the ionization efficiency and reproducibility was tested, varying the buffer pH, the organic buffer modifier and the sample concentration. Repeatabilities of migration time and peak area were measured and compared. Two 10 mM ammonium acetate buffers with pH 4.0 and 8.5 were investigated. No influence of buffer pH on peak area reproducibility was found. Isopropanol as organic buffer modifier significantly improved the ionisation leading to larger peak areas, but reduced reproducibility. The basic buffer produced slightly better RSD%s for migration times (2.5-4.0%) (n = 180) and faster analysis for the different test analytes of sample 1, while with the acetic buffer, RSD%s from 3.9 to 6.0% were obtained (n at least 163). The positioning of the capillary turned out to be the crucial parameter to ensure reproducible results. Thus, a procedure was established to ensure a defined ion-intensity level after capillary changes. The investigation of the different sample concentrations gave negligible differences in RSD%, showing that the signal-to-noise ratio was not the crucial parameter for reproducibility here, in contrast to CE-UV detection.