LC Ghost Peaks

The discussion of ghost peaks for this blog post refers to extraneous peaks that appear when injections of presumably clean solvent are made. For the purpose of this discussion, I am not including extraneous peaks that occur only in prepared samples. In addition, I am not including extraneous peaks encountered for PFAS analyses, due to the presence of PTFE in most LC systems.  Of most concern are ghost peaks that could be interpreted or easily mistaken as target analytes. These represent regularly occurring interferences that hamper proper integration of target analyte peak(s). Usually these appear as a regular pattern, not random. To begin the troubleshooting process, you should start with these steps.

1. First, if you have any guard cartridges or in-line filters attached to the front of the column, remove those and try running some blanks without those in line. If you can achieve a clean baseline with no filter or guard in line, then you will simply need to replace the filter or guard cartridge and rinse the guard holder thoroughly with solvent. If you are still seeing contaminant peaks without them in line (even if it is slightly better), then proceed to the remaining steps without them in line.
2. Make sure you are not recirculating/recycling the mobile phase. This is occasionally done with isocratic methods to conserve solvent. This practice should be paused until the current issue with ghost peaks is resolved. After stopping the recirculation, make sure to purge the instrument lines with fresh mobile phase and then inject a mobile phase blank.
3. When troubleshooting with blank injections, be sure to inject mobile phase (at starting conditions) with nothing added and no other solvents if possible. If injecting other solvents, additional factors are introduced and may interfere with the ability to diagnose the current issue
4. After establishing results with injection of mobile phase, cycle through the method without making any physical injection. This should be made without activation of the sample injection valve if possible. Most instruments have a way to generate a start signal without engaging the autosampler to do this. Sometimes it is as simple as pressing a “start” key. If this is not possible, sometimes the sample injection valve can be manually bypassed. If the data generated this way shows no ghost peaks, that confirms that the source of the ghost peaks is coming from somewhere in the autosampler or the injected sample. Try thoroughly cleaning the rinse vials and syringe, replacing with new if needed. If the issue remains, there is likely contamination in the tubing or valves of the autosampler. If using PEEK tubing, it is usually easier (and not very expensive) to replace it rather than try to clean it. If flushing multiple times does not help, valves may require replacement and possibly servicing by your instrument manufacturer rep. 
5. If cycling through the method with no injection still produces ghost peaks, then the source is not in the autosampler, and could be one of several things. The easiest thing to check first and most common issue is the mobile phase itself. Try making fresh mobile phase with water and solvents from a known source. If water is used from a purifier in your lab, try purchasing certified LC or LCMS grade water from an outside source that is reliable instead. Make sure you are using LC grade organic solvent from a reliable source and make sure to carefully clean all glassware that is used to mix the mobile phase.  Also, make sure to inspect the mobile phase filters and sonicate or replace as needed. Bacterial growth can be an issue and is not visible inside a stainless steel filter. Glass ones like our catalog #26431 can be an advantage in this regard.
6. Try removing the column and replace with a clean union to see if the peaks still show up or if the baseline is clean. If the contamination is coming only from the column, you should see a pretty flat signal with no column in line. (Please keep in mind that you will not see a true baseline because there is not much delay in separating the peaks without the column, i.e., there is no chromatography happening.) If the signal plot does not look clean, then the issue is either coming from the mobile phase, from instrument parts in the LC module or the detector. It could be coming from contaminated tubing or fittings, switching valves, or something inside the detector, like the detector flow cell.
7. If the test you performed in the last step pointed toward the column as the source of ghost peaks, you will need to try a few things to resolve this. If the column itself is the culprit, try flushing the column with the series of solvents described in the instructions located here: Restek Liquid Chromatography (LC) Columns — Cleaning Recommendations
8. If flushing the column does not remove the contamination, then it is time to order a new column.

Here are some resources you might also find helpful:

LC Troubleshooting- Baseline Problems (video)

Getting Your LC Up and Running Again

Live Long and Prosper-How to make your HPLC columns last longer

Routine LC Maintenance: Simple Steps to Preventing Unexpected Downtime

How do I Replace my LC Tubing?

Ghost Peak Investigation in a Reversed-Phase Gradient LC System, LCGC North America-05-01-2011, Vol 29, Issue 5

Causes and Measures for Addressing Ghost Peaks in Reversed Phase HPLC Analysis, Shimadzu Technial Report Vol. 45

Ghost Peaks from Nitrile Glove Contamination in Reversed-Phase LC Drug Analysis, LCGC North America, January 2016, Vol 34, Number 1, pages 50-55

Liquid Chromatography Issues: Ghost Peaks, Agilent Technologies Support Document A25637.doc

I hope you found this helpful and thanks for reading.