When the snow begins to melt and the birds are singing, it reminds us that spring is upon us. It also reminds us that we need to get ready for the upcoming warmer months by perfecting our shape. Our HPLC peak shape, that is. While the perfect body shape is not realistic, the perfect HPLC peak shape is: long, sharp, and symmetrical. A Gaussian.

Before getting into the six most common shape problems, it is important to know why it matters. If a peak becomes too broad or its tail is too large, it might overlap with or hide another peak. In the case of preparative chromatography, this also means mixed fractions are collected.

The cause behind peak distortions can be either physical or chemical in nature. With physical causes such as contamination of the particle bed or column components, incompatibility of the injection solvent with the mobile phase, or void volume and extra-column components, all peaks on the chromatograms would show the same distortions. On the other hand, chemical issues such as unwanted interactions between the analytes and the mobile phase or stationary phase, kinetics, or overloading, means all peaks would behave differently.

The importance of diagnosing the right source of the peak distortion is that it allows to quickly resolve the issue and get back to the beautiful chromatograms you were hoping to see in the first place. To get ready for the summer, let's look at the six most common peak challenges in HPLC and how to overcome them, while still getting to eat your cake.

1. Peak Tailing

Peak tailing indicates that a fraction of the sample is more retained than the rest and is diffusing throughout the column. One reason this can happen is when your analyte can be both charged and neutral depending on the pH. This results in different retention times and peak tailing occurs. At the same time there could also be interactions with free silanols in the stationary phase.

If either of these are possible (analyte is ionic or can react with active sites on the stationary phase), use a pH modifier and buffer or an ion-pairing reagent to ensure that the whole sample stays at the same state for the entire run.

To find the best pH, think of your analyte's pKa. At a low pH, when a basic analyte is protonated (i.e., charged), it is extremely polar and retained on a normal phase. At a higher pH when the base is not protonated (i.e., neutral), the retention time of the analyte is different. At the compound's pKa, it protonates and deprotonates so there is a mixture of the two states, which causes multiple peaks or shark fin shaped peaks. Hence you should be at least 2 pH units higher or lower than the compound's pKa, but make sure to stay within the pH stability range of your grafted phase. Changing the pH can also reduce the interactions with residual silanols.

Unreacted silanols are Bronsted acids. They can thus donate a proton to the mobile phase, therefore producing an ionic group that can react with the analytes. It should be noted that the distorted peak shape due to interactions with silanols can improve when a greater quantity of sample is loaded: as larger sample loads are injected, the unreacted silanols become saturated and can no longer interact with the analyte resulting in a lower fraction of the sample being affected, thus creating less tailing in comparison.

Another possible reason (and easy to fix) for peak tailing is mass overload. As the term suggests, this happens when you inject too much sample into the column. Simply solve this by injecting a smaller quantity. When doing so, you might notice that your peaks will become more symmetrical. It is even possible that a single broad/tailing peak will become two sharp separated peaks.

To check if unreacted silanols are the cause for distortion, simply add more sample and watch whether it improves the peak shape (the analytes are reacting with silanols) or deteriorates the peak shape (there is mass overload). Please note that both can be true at the same time, in which case you should reduce the loading and adjust the pH.

Backflushing – why do we do it?

We want to get rid of impurities that might be retained in the first few millimetres of the column. Going the usual way might contaminate the rest of the column or consume enormous amounts of solvent, so backflushing allows to remove them by the way they entered. Using a guard column would also prevent these impurities to touch the column or clog the frit in the first place.

If injecting smaller quantities or adjusting the pH do not fix the problem, the frit may be partially or completely blocked. When changing frits, make sure not to disturb the bed surface. Cleaning the frit can be done by sonication in methanol. Backflush the column: disconnect it and reverse it on the support, connect the solvent line into what used to be the column exit, and then connect it to the pump but not the detector to protect your detector cell. You can direct the flow towards the waste. Pump solvent at half the usual flow rate for 5-10 column volumes. Do not forget to reverse back your column after your backflush.

Remember that if all peaks show the same problem, peak tailing in this case, then it is most likely a system issue. For example, extra dead volume (volume of the system, excluding the column) can cause peak tailing. While it is impossible to completely remove dead volume, try to reduce it as much as you can by using shorter tubes with smaller internal diameters before and after the column, reducing the injection volume, and the detection cell's volume. If dead volume is the cause, the early peaks should be more affected than the later peaks since they reside longer on the column's stationary phase compared to the time spent in dead volume.

When it comes to peak distortions, the most likely connections that could leak or be obstructed are the ones connecting items between the sample injector and the detector. Apart from leaks and obstructed tubes, keep an eye open for tubes with an end that is not square. The extra space on one side of the tube causes the analyte's band to change shape as it passes in the gap, and this can cause peak tailing. To avoid this kind of issue, make sure to use pre-cut capillaries, or use the proper tools when cutting them yourself to make sure the tube's end is perfectly square.

To summarize peak tailing, optimize the mobile phase's pH, avoid overloading, backflush the column, and reduce dead volume.

2. Peak Fronting

On the other hand, peak fronting happens when part of the sample is less retained than the rest of it and races quickly down in the column.

If all peaks show fronting, this is because of packing issues. This happens when the silica bed density is uneven, and analytes travel in a preferential (rather than straight) path. This can happen when packing a column in the lab, or with a pre-packed column that has been exposed to repeated pressure cycles causing the particle bed to rearrange over time. Either be careful to create a more homogenous packing when doing it by hand or it may be time to change the column for a new one.

Another possible cause is mass overload. The analyte may be too concentrated in the injection volume and cause overload. The analyte molecules saturate the silica at the inlet end of the column and the excess molecules move down the column with fewer interactions than they should. This decreases the analyte's retention time. Reducing the amount of analyte injected should resolve this.

Fronting or shark fin peak shape can also happen when the sample solvent is not compatible with the mobile phase. Whenever possible, the mobile phase should be the injection solvent. When this is not possible due to poor solubility, you can use other solvents such as acetonitrile, methanol, DMSO, DMF, or DMA. Use as little solvent as possible to not cause volume overload at the injection.

In brief for peak fronting, do not overload your sample and consider your injection solvent carefully.

Both peak tailing and fronting can also be an illusion caused by two peaks being too close together. By changing the selectivity or improving the resolution, the peaks may separate. To achieve better resolution or selectivity, changing the mobile or stationary phase might be in order.

3. Split Peaks or Shoulder Peaks

If only one peak is splitting, it is likely to be a chemistry issue with either the wrong mobile phase and/or stationary phase as the culprit. If all peaks are splitting, other causes are possible.

Selection of the injection solvent is consequential since it may cause split peaks if it is too strong. If it is not possible to inject in the mobile phase, the chosen injection solvent should be weaker than the mobile phase. Otherwise, the analytes race through the column instead of eluting in a symmetrical band.

If there is a disruption in the sample path and it follows multiple paths through the column instead of passing through in a straight line, split peaks can appear on the chromatogram. There can be several factors causing this: the column could be poorly packed, or the packing bed drifted under the system pressure, the mobile phase pH could be too high causing dissolution of the silica and thus creating a void at the column inlet, or the column frit could be partially blocked or obstructed. All these issues would cause the flow through the column to be uneven, leading to different mobile phase velocities which will result in peak deformations.

Always stay in the pH range recommended for your sorbent. To solve a clogged frit issue, backflush the column and run a regeneration procedure. Take a look at your guard column, consider changing it for a new one. Using inline filters and sample preparation are generally good practices to protect the inlet frit from debris accumulations from very dirty samples.

If all the peaks show the same splits or shoulders, consider changing the column. Remember to treat your column with love before it is too late and maintain it in good condition following the recommended cleaning and storage conditions to keep it strong and healthy. This means running a cleaning procedure regularly and consider using sample preparation for especially dirty samples before injecting them on your precious HPLC column.

To summarize peak splitting, consider the chemistry (mobile and stationary phases), the injection solvent, the mobile phase's pH, and filter the sample before injecting it.

4. Broad Peaks

It's all about sharp and symmetrical peaks. When they are broad, it reduces the gap between the peaks and overlap is more likely, leading to poor separations. When peaks are broad, the analytes diffuse too much during elution, and increasing the flow rate should resolve this.

5. Peaks Too Small or Absent

An issue with the detector could cause this. If no peaks show at all, your detector might be off, have a loose connection or there could be some broken wires. Check all connections. The flow could be interrupted or obstructed: check if your mobile phase is immiscible or your buffers precipitated, blocking the flow.

Have you forgotten to inject your sample, maybe? It will be our little secret, just act normal and start over with the right sample with the detector turned on. Or maybe you injected it alright, but it has deteriorated? What is your sample's stability in time? In solvents?

Check the lamp status and auto-zero. Listen for unusual noises and consider changing the seals if you hear anything worrisome.

6. Negative Peaks or Ghost Peaks

Check the recorder's polarity. It might be reversed if you notice your peaks are negative. Make sure your sample's refractive index is higher than your mobile phase's and that it is also more absorptive. You might want to change mobile phase for one that does not absorb at the chosen wavelength.

Ghost peaks come from your system instead of from your sample. They arise when there is contamination in the injector or the column. Flush your column between all your analyses and include a final wash step in gradient analyses to remove all the strongly retained compounds.

In Conclusion

At SiliCycle, we could not care less about thigh gaps; we care about gaps between peaks. Distorted peak shapes are signs that something with the method is off and should be looked at to ensure efficient separation. SiliCycle has HPLC experts on staff with up to 30-years' experience to help you solve your common and not so common HPLC issues. If we've peaked your interest, keep an eye open for the upcoming article on other common problems in HPLC.

Learn more


Eliane, M.Sc

Eliane is the Scientific Content Specialist here at SiliCycle, with years of experience both in the lab and client support. She studied at Laval University, for both her Bachelor’s and Master’s in Chemistry. In fact, her thesis was in organic electronics in which purity is of the utmost essence, making her in-tune to the purification needs of chemists.