Contents     Index   

11. Repeatability problems
Peak size repeatability problems can be caused by a number of chemical, physical, or computational issues. 

1. Excessive baseline noise. An excessively noisy baseline is probably the most common cause of integration problems. Before addressing repeatability problems, you should check to confirm that your baseline noise level is within specification. 

2. Poor resolution from an interference. If resolution between the analyte and an interfering peak is marginal, then even small amounts of noise can affect the allocation of area between the peaks. 

3.  Variation in injection volume. This will usually affect all of the peaks in the chromatogram by the same proportion. The "pseudo internal standard" test can be a useful diagnostic.

4. Variation in sample concentration. This will usually manifest as a systematic change in response (i.e., the peak size will steadily increase or decrease as a function of time). Steady increases in the size of all peaks in a chromatogram suggests evaporation of solvent from the sample vial. Steady decrease in the size of some peaks (and possibly increase in the size of others) suggest sample degradation.

4. Variation in detector response. Variations in lamp output or photodetector response usually manifest themselves as excessive baseline noise. The problem can be exacerbated if your detection wavelength is located on a steeply sloping portion of the absorbance spectrum, because small variations in either detector wavelength or sample spectrum will result in large changes in absorbance.

A certain amount of fluctuation in lamp output is normal; the magnitude increases with the age of the lamp. In most cases, the fluctuations are short-term, and result in excessive baseline noise. Most UV detectors are effectively "double beam", so that variations in lamp output should cancel. 

5. Variation in peak area integration. The final possibility is that the peaks are actually the correct size, but that the data system is integrating incorrectly. This is usually due to one of three things:

A. Changes in peak shape. These are usually systematic (i.e., steady increase or decrease in peak area) and can show up as excessive tailing, abnormally wide peaks, shoulders, flat top peaks or, in extreme cases, split peaks. They should be diagnosed in the same fashion as any other peak shape problems
   
B. Excessive baseline noise.  

C. Incorrect integrator settings. These usually result in larger errors for smaller peaks or for noisier baselines. 

Ref: LC-GC, 15(1) 24 (1997)
Ref: LC-GC, 15(6) 516 (1997)
Ref: LC-GC, 15(12) 1118 (1997)
Ref: LC-GC, 16(10) 910 (1998)



© 2006 LC Resources, Inc. all rights reserved