Separation and quantification of lanthanides in synthetic standards by capillary electrophoresis: a new experimental evidence of the systematic "odd-even" pattern observed in sensitivities and detection limits.

The systematic zigzag pattern of sensitivities and detection limits (LODs) of lanthanides, previously observed in mass spectrometric and chromatographic measurements, was once more investigated through the indirect photometric detection with capillary electrophoresis. Well-designed chemometric experiments were performed for the electrophoretic separation and detection of lanthanides using standards with similar concentrations. A fused silica capillary 355 mm x 75 microm was used. Complete separation for all 14 lanthanides was achieved in approximately 3 min at a capillary temperature of 15 degrees C. Indirect photometric detection at 214 nm using a voltage of +25 kV and a hydrostatic injection (100 mm for 20 s) were used. The background electrolyte used consisted of an optimum mixture of 0.004 M HIBA (as complexing agent) and 0.010 M UVCat-1 (as a UV-absorbing co-ion) with a pH 4.4. A good reproducibility in migration times (<2.7% RSD), peak areas (<3.8% RSD) and peak heights (2.7% RSD) were systematically obtained. Calibration curves based on both peak areas and peak heights (from seven replicates) were prepared using weighted least-squares regressions, which were employed for the correct estimation of individual sensitivities and LODs. For a better estimation of LODs, the lowest concentration standard was injected 30 times. A new experimental evidence of the systematic "odd-even" pattern was again observed in the lanthanide sensitivities (and therefore in LODs). The calculated sensitivities were greater for lanthanides with an odd-atomic number than for their corresponding neighboring element with an even-atomic number (i.e., (57)La-(58)Ce, (59)Pr-(60)Nd; (63)Eu-(64)Gd, etc.). Concerning the LODs, a systematic zigzag pattern was observed where the odd atomic number elements have lower LODs than the even atomic number neighbor elements (i.e., (57)La-(58)Ce; (59)Pr-(60)Nd; (63)Eu-(64)Gd, etc.). The possible origin of this "odd-even" effect is briefly discussed. Accuracy errors were less than 5% for lanthanide concentrations of three synthetic standard solutions, which were considered as "unknown" samples.