Some LC troubleshooting topics are never out of date, as there are issues in LC practice, even as instrument technology improves over time.There are many ways in which problems can arise in an LC system and end up in poor peak shape.When issues related to peak shape arise, a short list of possible causes for these results helps simplify our troubleshooting experience.
It’s been fun writing this “LC Troubleshooting” column and thinking about topics each month, because some topics never go out of style.While in the field of chromatography research certain topics or ideas become obsolete as they are superseded by newer and better ideas, in the field of troubleshooting, since the first troubleshooting article appeared in this journal (the LC Journal at the time) since some topics are still relevant) in 1983(1).Over the past few years, I have focused several LC Troubleshooting sections on contemporary trends affecting liquid chromatography (LC) (for example, the relative comparison of our understanding of the effect of pressure on retention [2] New Advances) Our interpretation of LC results and how to troubleshoot with modern LC instruments.In this month’s installment, I’m continuing my series (3), which started in December 2021, which focused on some of the “life and death” topics of LC troubleshooting — elements that are great for any troubleshooter are essential, no matter the age of the system we are using.The core topic of this series is highly relevant to the LCGC’s famous “LC Troubleshooting Guide” wall chart (4) hanging in many laboratories.For the third part of this series, I chose to focus on issues related to peak shape or peak characteristics.Incredibly, the wall chart lists 44 different potential causes of poor peak shape!We can’t consider all of these issues in detail in one article, so in this first installment on the topic, I’ll focus on some of the ones I see most often.I hope young and old LC users will find some helpful tips and reminders on this important topic.
I find myself increasingly answering troubleshooting questions with “anything is possible”.This response may seem easy when considering observations that are difficult to interpret, but I find it often appropriate.With many possible causes of poor peak shape, it is important to keep an open mind when considering what the problem might be, and to be able to prioritize potential causes to begin our troubleshooting efforts, focusing on those most common possibilities ,this point is very important.possible.
A key step in any troubleshooting exercise — but one that I think is underrated — is recognizing that there is a problem that needs to be solved.Recognizing that there is a problem often means recognizing that what happens to the tool is different from our expectations, which are shaped by theory, empirical knowledge, and experience (5).The “peak shape” referred to here actually refers not only to the shape of the peak (symmetrical, asymmetrical, smooth, fluffy, leading edge, tailing, etc.), but also to the width.Our expectations for actual peak shape are simple.Theory (6) well supports the textbook expectation that, in most cases, the chromatographic peaks should be symmetrical and conform to the shape of a Gaussian distribution, as shown in Figure 1a.What we expect from peak widths is a more complex issue, and we will discuss this topic in a future article.The other peak shapes in Figure 1 show some of the other possibilities that could be observed—in other words, some of the ways things could go wrong.In the remainder of this installment, we’ll spend time discussing some specific examples of situations that can lead to these shape types.
Sometimes peaks are not observed at all in the chromatogram where they are expected to be eluted.The above wall chart indicates that the absence of a peak (assuming the sample actually contains the target analyte at a concentration that should make the detector response sufficient to see it above the noise) is usually related to some instrument issue or incorrect mobile phase conditions (if observed at all). peaks, usually too “weak”).A short list of potential problems and solutions in this category can be found in Table I.
As mentioned above, the question of how much peak broadening should be tolerated before paying attention and trying to fix it is a complex topic that I will discuss in a future article.My experience is that significant peak broadening is often accompanied by a significant change in peak shape, and peak tailing is more common than pre-peak or splitting.However, the nominally symmetric peaks are also broadened, which can be caused by a few different reasons:
Each of these issues has been discussed in detail in previous issues of Troubleshooting LC, and readers interested in these topics can refer to these previous articles for information on the root causes and potential solutions to these issues. More details.
Peak tailing, peak fronting, and splitting can all be caused by chemical or physical phenomena, and the list of potential solutions to these problems varies widely, depending on whether we are dealing with a chemical or physical problem.Often, by comparing the different peaks in a chromatogram, you can find important clues about which is the culprit.If all peaks in a chromatogram exhibit similar shapes, the cause is most likely not physical.If only one or a few peaks are affected, but the rest look fine, the cause is most likely chemical.
The chemical causes of peak tailing are too complex to discuss briefly here.The interested reader is referred to the recent issue of “LC Troubleshooting” for a more in-depth discussion (10).However, an easy thing to try is to reduce the mass of the injected analyte and see if the peak shape improves.If so, then this is a good clue that the problem is “mass overload”.In this case, the method must be limited to injecting small analyte masses, or the chromatographic conditions must be changed so that good peak shapes can be obtained even with larger masses injected.
There are also many potential physical reasons for peak tailing.Readers interested in a detailed discussion of the possibilities are referred to another recent issue of “LC Troubleshooting” (11).One of the more common physical causes of peak tailing is a poor connection at a point between the injector and detector (12).An extreme example is shown in Figure 1d, obtained in my lab a few weeks ago.In this case, we built a system with a new injection valve that we had not used before, and installed a small volume injection loop with a ferrule that had been molded onto a stainless steel capillary.After some initial troubleshooting experiments, we realized that the port depth in the injection valve stator was much deeper than we were used to, resulting in a large dead volume at the bottom of the port.This problem is easily solved by replacing the injection loop with another tube, we can adjust the ferrule to the proper position to eliminate the dead volume at the bottom of the port.
Peak fronts like those shown in Figure 1e can also be caused by physical or chemical problems.A common physical cause of the leading edge is that the particle bed of the column is not well packed, or that the particles have reorganized over time.As with peak tailing caused by this physical phenomenon, the best way to fix this is to replace the column and keep going.Fundamentally, leading edge peak shapes with chemical origin often arise from what we call “non-linear” retention conditions.Under ideal (linear) conditions, the amount of analyte retained by the stationary phase (hence, the retention factor) is linearly related to the concentration of the analyte in the column.Chromatographically, this means that as the mass of analyte injected into the column increases, the peak becomes taller, but not wider.This relationship is broken when the retention behavior is non-linear, and the peaks not only become taller but also wider as more mass is injected.In addition, nonlinear shapes determine the shape of chromatographic peaks, resulting in leading or trailing edges.As with mass overload that causes peak tailing (10), peak leading caused by nonlinear retention can also be diagnosed by reducing the injected analyte mass.If peak shape improves, the method must be modified to not exceed the injection quality that causes the leading edge, or the chromatographic conditions must be changed to minimize this behavior.
Sometimes we observe what appears to be a “split” peak, as shown in Figure 1f.The first step in solving this problem is to determine whether the peak shape is due to partial co-elution (ie, the presence of two distinct but closely eluting compounds).If there are actually two different analytes eluting close together, then it’s a matter of improving their resolution (for example, by increasing selectivity, retention, or plate count), and the apparent “split” peaks are related to physical Performance has nothing to do with the column itself.Often, the most important clue to this decision is whether all peaks in the chromatogram exhibit split shapes, or just one or two.If it’s just one or two, it’s probably a co-elution issue; if all peaks are split, it’s probably a physical issue, most likely related to the column itself.
Split peaks related to the physical properties of the column itself are usually due to partially blocked inlet or outlet frits, or reorganization of particles in the column, allowing the mobile phase to flow faster than the mobile phase in certain areas of the column channel formation .in other regions (11).Partially clogged frit can sometimes be cleared by reversing the flow through the column; however, in my experience, this is usually a short-term rather than a long-term solution.This is often fatal with modern columns if the particles recombine within the column.At this point, it is best to replace the column and continue.
The peak in Figure 1g, also from a recent instance in my own lab, usually indicates that the signal is so high that it has reached the high end of the response range.For optical absorbance detectors (UV-vis in this case), when the analyte concentration is very high, the analyte absorbs most of the light passing through the detector flow cell, leaving very little light to be detected.Under these conditions, the electrical signal from the photodetector is heavily influenced by various sources of noise, such as stray light and “dark current”, making the signal very “fuzzy” in appearance and independent of analyte concentration. When this happens, the problem can often be easily resolved by reducing the injection volume of the analyte—reducing the injection volume, diluting the sample, or both.
In chromatography school, we use the detector signal (ie, the y-axis in the chromatogram) as an indicator of the analyte concentration in the sample.So it seems odd to see a chromatogram with a signal below zero, as the simple interpretation is that this indicates a negative analyte concentration – which of course is not physically possible.In my experience, negative peaks are most often observed when using optical absorbance detectors (eg, UV-vis).
In this case, a negative peak simply means that the molecules eluting from the column absorb less light than the mobile phase itself immediately before and after the peak.This can occur, for example, when using relatively low detection wavelengths (<230 nm) and mobile phase additives that absorb most of the light at these wavelengths.Such additives can be mobile phase solvent components such as methanol or buffer components such as acetate or formate.One can actually use negative peaks to prepare a calibration curve and obtain accurate quantitative information, so there is no fundamental reason to avoid them per se (this method is sometimes referred to as “indirect UV detection”) (13).However, if we really want to avoid negative peaks altogether, in the case of absorbance detection, the best solution is to use a different detection wavelength so that the analyte absorbs more than the mobile phase, or change the composition of the mobile phase so that They absorb less light than analytes.
Negative peaks can also appear when using refractive index (RI) detection when the refractive index of components other than the analyte in the sample, such as the solvent matrix, is different from the refractive index of the mobile phase.This also happens with UV-vis detection, but this effect tends to be attenuated relative to RI detection.In both cases, negative peaks can be minimized by more closely matching the composition of the sample matrix to that of the mobile phase.
In part three on the basic topic of LC troubleshooting, I discussed situations in which the observed peak shape differs from the expected or normal peak shape.Effective troubleshooting of such problems begins with knowledge of expected peak shapes (based on theory or prior experience with existing methods), so deviations from these expectations are obvious.Peak shape problems have many different potential causes (too wide, tailing, leading edge, etc.).In this installment, I discuss in detail some of the reasons I see most often.Knowing these details provides a good place to start troubleshooting, but doesn’t capture all possibilities.Readers interested in a more in-depth list of causes and solutions can refer to the LCGC “LC Troubleshooting Guide” wall chart.
(4) LCGC “LC Troubleshooting Guide” wall chart.https://www.chromatographyonline.com/view/troubleshooting-wallchart (2021).
(6) A. Felinger, Data Analysis and Signal Processing in Chromatography (Elsevier, New York, NY, 1998), pp. 43-96.
(8) Wahab MF, Dasgupta PK, Kadjo AF and Armstrong DW, Anal.Chim.Journal.Rev. 907, 31–44 (2016).https://doi.org/10.1016/j.aca.2015.11.043.