Baseline Problems Chromatogram

Baseline Problems (Chromatogram) Common baseline problems are illustrated with several diagnostic examples shown in Figures 10.6 and 10.7. The description of each symptom and its possible remedial action is as follows. 

Plot baseline with chromatograms in reports to allow detection of integration problems. 

Very often baseline problems are related to detector problems. Many detectors are available for HPLC systems. The most common are fixed and variable wavelength ultraviolet spectrophotometers, refractive index, and conductivity detectors. Electrochemical and fluorescence detectors are less frequently used, as they are more selective. Detector problems fall into two categories electrical and mechanical/optical. The instrument manufacturer should correct electrical problems. Mechanical or optical problems can usually be traced to the flow cell however, improvements in detector cell technology have made them more durable and easier to use. Detector-related problems include leaks, air bubbles, and cell contamination. These usually produce spikes or baseline noise on the chromatograms or decreased sensitivity. Some cells, especially those used in refractive index detectors, are sensitive to flow and pressure variations. Flow rates or backpressures that exceed the manufacturer s recommendation will break the cell window. Old or defective source lamps, as well as incorrect detector rise time, gain, or attenuation settings will reduce sensitivity and peak height. Faulty or reversed cable connections can also be the source of problems. 

Plot baseline with chromatograms in reports and watch out for integration problems. Use more advanced integration algorithms (e.g., ApexTrack) if possible for complex chromatograms with sloping baselines. 

There are several other problems that can occur with uv flow cells. Air bubbles in the cell can produce a series of very fast noise spikes on the chromatogram, or pronounced drift in the baseline, followed by sudden changes in the baseline position, or both effects at once. 

The major problem in data reduction is to select the relevant range of data from a chromatogram such as the one shown in Figure 5. We do not limit our data collection process to the data that will actually be used to calculate the distribution parameters because we find that values outside of the range used in data reduction help to characterize baseline drift and provide indications of the validity of the results. 

The baseline is chosen to be a straight line connecting these two points in the chromatogram. Additional schemes for dealing with the baseline are available for specimens containing large amounts of low molecular weight species. The results obtained for such specimens are, however, dubious because of problems associated with calibration and because of interference from injection artifacts... 

There are two modules in a strip chart the electronics and the mechanical. The trick is to look at one at a time. Disconnect the detector leads, turn off the bed, and watch the pen. Does it sit quietly or chatter up and down Noise at this point comes from the strip chart electronics. If it s quiet, turn on the bed and let the pen trace to see if the baseline is flat. If so, short across the leads and make sure the pen deflects without sticking (this would show up as a plateau in a chromatogram). Lubrication or drive wire replacement would fix the problem. Do not get much oil on the slide bar it just traps dust. Spray some WD40 on a Kimwipe. Wipe the bar, then wipe off the excess. Next, use a stopwatch and time the bed. Is it accurate at 0.5cm/min where you will be using it If it passes these tests, we re ready to hook up the detector leads and move on. 

Figure 6.1 Example chromatogram illustrating noise (short-term fluctuations, approximate peak-to-peak value shown by lines on expanded section) and drift (the slope in the baseline, indicated by the arrow). The signal-to-noise ratio for the first three peaks should be sufficient for reasonably adequate quantitation, but the broader fourth peak on the noisy, sloping baseline would pose problems for most integrators.

---