Troubleshooting HPLC- Loss in Response for Some, but Not All Analytes

Just as in GC applications, patterns in your HPLC chromatography can exhibit telltale signs that point toward probable sources of error.  In this brief series of posts, we will look at possible scenarios that you may encounter in the laboratory and how you might approach resolving those difficulties.

We occasionally get calls or emails from folks that are seeing lower than expected response for peaks of interest. This may be a matter of method development if you have not done this analysis previously. Suggestions in the post are provided assuming you are working with an established method and have had successful results over an extended period of time (until recently). These suggestions also assume that the response is low for quantitation standards as well as in prepared QC samples. If the problem appears only in your QC samples, then please review the earlier post on method troubleshooting. Otherwise, please see below, where we begin with a case where all peaks are equally low across the board for multiple analytes.

50 cent

This is often caused by:

1. Calculation/dilution error- This is one of the most common and easiest problems to resolve, although it is often the least suspected. A dead giveaway is an exact ratio of error, such as results that are close to exactly 50% low (2-fold).  Try having someone who has not been involved in your project prepare your standard or do the calculations starting from scratch to see if they obtain a different result. This can be a classic case of tunnel vision. Make sure the established method is being used for this and has not changed.

2. Injection volume is being measured incorrectly. This depends on how your specific autosampler and system are configured. This could be caused by:

a. Incorrect programming of injection volume

b. Incorrect size of syringe or sample loop, or inadequate volume of sample loaded onto the autosampler.

c. Malfunction of sample injection valve or metering device.

Depending on the situation, issues of this type might give results that are pretty consistent, similar to a calculation or dilution error.  If there is any doubt, try performing a manual injection to see if your response increases. (Consult your instrument manufacturer if you need help determining how to do this with your system.)

3. Injection volume is partially lost due to leaks. This could occur at either of the following locations:

a. Leakage in the autosampler from fittings or tubing or could be due to a malfunction/leakage from the sample injection valve.  An issue of this type may or may not be fairly consistent, although it would be surprising to see a numerical correlation such as in a dilution error. You may be able to see such an issue visibly, but if there is any doubt, try performing a manual injection to see if your response increases.

b. Leaks downstream from the injection valve- If you have a large enough leak between your valve and the detector, either before or after the column, it can result in low response. If the leak is before the column your system pressure will be lower also, thus, another good reason to keep a record of your operating pressures. If your leak is after the column or in the detector, you might see disturbances in the baseline, such as might be caused by air bubbles being present.

4. Peak broadening- When peaks become broader, the peak height is always lower. Try comparing your peak areas instead of heights and see if this explains what you are seeing. Broader peaks are also more difficult to get integrated properly, which makes matters worse. If this occurs for all of your analyte peaks, it is likely due to one of the following:

a. Excessively large injection volume. Check your operating parameters and see if you are using a higher volume that you did initially for this method. Try injecting a smaller volume of a more concentrated standard instead. (See the FAQ section on our website for suggested volumes.)

b. A change in system dead volume, resulting in increased extra-column band broadening. This effect is more pronounced for smaller ID columns. Try shortening those lengths of tubing and/or use smaller ID tubing. If you have changed detectors, see if the flow cell is smaller than the one you used previously and replace if needed.  Click here for more background information on the topic.

5. Detector maintenance needed- Detector lamp may be near the end of its life and need replaced. If baseline or background signal is high, the flow cell may be contaminated and should be cleaned. These issues are usually accompanied also by baseline drifts and/or noise and a steady decline in response over time.

6. If you have exhausted all of the above possibilities, check out the next post in this series, where we will discuss sources of loss for only certain analytes, but not always all of them. Of course, if you have only one analyte, or a handful of very similar analytes, the second post may be more applicable to your situation.