Liquid Chromatography & HPLC – Troubleshooting & DIY

Liquid Chromatography & HPLC

Liquid chromatography is one of the most common processes in science. You may you know it as HPLC, and use it for analysis. Or perhaps you know it by a trade name, such as CombiFlash, and use it for separation of products. Whatever the situation, it’s rare to find a lab that doesn’t employ some form of liquid chromatography. In a sentence, LC involves passing a liquid through a tube of granular material that will interact with the samples dissolved within the liquid. The tube is called a column, and its solid contents are the stationary phase. The liquid that passes through the column is the mobile phase, called an eluent or solvent. The interactions between the stationary and the mobile phase happen at the molecular level – but (ideally) there is no reaction or filtration taking place.

LC Column

LC columns are classified as either normal or reversed phase. A normal stationary phase is neutralized silica gel (essentially high purity sand with its pH adjusted to around 7). Other varieties include silicate polymer gels bound to other polar groups, such as nitros or diols. Reversed phase consists of a hydrocarbon gel categorized by its chain length or monomeric sidegroup (C18, C8, phenyl, and so on). Column type is your first consideration in LC, and it will mainly be influenced by your sample’s molecular properties.

  • Reversed Phase (RP – Nonpolar): Hydrocarbon or aromatic polymer.
    • Interacts with non-polar compounds, slowing them down (longer retention time).
    • The most polar compound elutes first; most non-polar compound elutes last.
  • Normal Phase (NP – Polar): Neutralized silica or silicate polymer with polar side-group.
    • Interacts with polar compounds, slowing them down (longer retention time).
    • Most non-polar compound elutes first; most polar compound elutes last.
  • Special Types
    • Ion Exchange: Polymer gel with ionized side-groups. Reacts with specific ions in sample to quantify or purify.
    • Size Exclusion: Dextran polymer gel network. Hinders sample passage to separate compounds by molecular weight.

Important!  Polymers are less physically stable than silica-bonded columns. Always check your method against the maximum pressure specifications of the column.

Eluent (Solvent)

Your mobile phase selection will depend on your sample composition and available columns. Solvent quality is extremely important for LC. Impurities accumulate on the column and impede flow. Dissolved gases can lead to measurement error if not removed.

Protic solvents normally involve the use of a buffer solution to maintain the column at optimal pH, but these pose a risk to the instrument itself. Since salts and ionic solutions are corrosive, these must be purged from the system when not in use.

Important!  Never transition directly from a buffered solvent to a storage solution such as acetonitrile! Always flush buffers with water or another protic solvent first! Otherwise the salts will precipitate into the tubing or column fittings blocking flow and causing localized corrosion.

Verdant is an industry leader in solvent recovery technology. Recycling used solvent is a key focus of industrial sustainability efforts – and it’s also a great way to reduce operating costs. If you’re interested in learning about your options, ask us how we can help.

Flow Rate

Your instrument’s flow rate affects the pressure within the system. Your flow rates will typically fall within a range varying on the order of milliliters per minute (mL/min). Your flow rate selection will be influenced by your column dimensions, pore size, stationary phase composition, and solvent viscosity.

Data Sampling Rate

If your LC has a variable data sampling rate, faster is better – up to a point. If your sampling rate is too low, narrow peaks will tend to clip – meaning you’ll miss the tip of the peak and get an inaccurate reading. Set your data sampling rate to 40 Hz as a starting point, and then raise it as high as you can without increasing baseline noise.

Derivatization / Internal Standardization

Derivatization is an excellent option when you have trouble with solubility, detection ranges, or peak shape. By subjecting the sample to a chemical reaction, you can change the chemical structure into a derivative that’s better suited for analysis. It can also be used to work with a more limited set of  supplies by increasing sample compatibility with your available columns and solvents. There are more reactions than we could ever list here – entire books have been filled with the possibilities. But don’t worry, because we have those books! Contact us for assistance if you need help with this technique.

Internal standardization is a technique that’s better suited to UV detectors than LC-MS. This can help you to determine something elusive that can’t normally be detected with UV – such as a sample and solvent pair that absorb in the same range. By spiking your sample with a compound which absorbs outside of that range, you can quantify concentrations much more accurately.

Troubleshooting Liquid Chromatography & HPLC

I have a noisy baseline! (General Issues)

A noisy baseline is a common symptom with several different potential sources, which can make troubleshooting very time consuming. You might have a leak. The column could be contaminated or degrading. You could have a low signal to noise ratio brought on by solvent impurities, or a data acquisition rate that’s too high for your method. If the noise is periodic, or rhythmic like a wave, then it could be a failing pump or pulse dampener. If the noise is spiky and appears irregularly, it there could be bubbles within the system, or the detector lamp could be flickering/failing. If your instrument isn’t isolated on a power filter with ferrite chokes on the wires, it could also be caused by electrical noise from other devices within the lab.

My baseline won’t stay still! (Baseline Drift)

If your baseline repeatedly drifts in the same direction over the course of every gradient run, then your column is contaminated (if you normally run isocratic, you can run a gradient to see whether this is the case). Ensure that you’re taking adequate column protection measures, then clean or replace your column.

If you’ve ruled out column contamination, or if your baseline is drifting randomly or steadily in a particular direction over the course of multiple runs, it may be time to replace your detector lamp.

My peaks keep showing up earlier/later every time I run a sample! (Retention Creep)

As your column ages, it will be less effective at interacting with the sample, and the retention time will slowly get shorter. If this effect becomes significant enough to be noticeable between runs, check that you’re adequately maintaining the column’s pH between runs. If so, then it’s just time to change the column.

If your retention time is getting longer with each run, remember that LC leaks cause late peaks. If you can see the problem worsening, then you likely have a part wearing out and physically falling apart, which usually be accompanied by other erratic findings. You’ll most likely find the offending part, but LC is a sensitive technique, so don’t be surprised if the flaws aren’t visible to the naked eye. When in doubt, replace any parts which have been in service long enough to be at risk.

If your retention time is changing erratically, there is probably a bubble entering the system at injection time, or solvent cavitation occurring at the pumps. Ensure that your sample solvent and mobile phase are being adequately degassed, and check the inlet.

My peaks have peaks! (Split Peaks)

This means your sample is getting hung up on something. First, confirm that your sample and its solvent are compatible with your entire mobile phase (both solvents if you’re running a gradient). If compatibility checks out, then there is something in the way. This can either be a solid contaminant, or a bubble trapped in a component. Careful inspection will be necessary to determine which.

My peaks aren’t sharp enough to quantify! (Peak Tailing / Fronting)

This usually indicates a problem with the column caused by improper use. First, rule out whether your sample volume is overloading your column. Then check the column for voids, degradation, or channeling (all ways of saying: gaps within the column packing). For troubleshooting this issue, the simplest first step will be to replace the column and guard. Before running another sample, check for flawed processes. Confirm that the method is not exceeding the temperature limit of the column packing. Ensure that the column’s pH is being maintained correctly. Verify that the column is being properly stored between uses. Consider implementing a log recording the dates the guard columns are replaced, and increase the frequency if necessary.

If the problem is unique to particular samples, or certain portions of the chromatogram, you’re most likely seeing multiple components elute together, and should adjust the method to secure better separation.

My peaks are too wide! (Column Capacity)

Issues with peak width are mainly symptoms of sample volume. If your peaks are too broad, your column is most likely too small for your sample size, and is being overloaded. This is especially true for complex samples – your column interacts with everything being eluted, not just your sample. Ensure that you’re factoring the entire sample into column size determinations. If you’re certain that isn’t the case, first make sure your fittings are secure and that your target pressure is being reached. Then try slowing your gradient transition rate down to allow for a slower elution time.

My peaks are running together! (Poor Resolution)

This is usually a sign that it’s time to replace your guard column or filter, particularly if you’re noticing that pressures are beginning to rise. This is because poorly soluble contaminants tend to build up at the front of the column and get stuck.

My analyte isn’t showing up! (Missing Peak)

One possibility is that your analyte isn’t compatible with your solvent or column choice. This can include insolubility, UV absorbance masking, or samples that interact heavily with the column.

Another possibility that the sample is too dilute. A UV detector is linear across 4 orders of magnitude (104, or 10,000x concentration difference), so try increasing the concentration if the dilution factor is within your control.

Nothing is showing up at all! (No Peaks)

This could be an issue with your detectors or a wiring fault preventing communication. If you see noise in the baseline, the issue is probably not a loose connection or faulty wiring. Verify that your detector wavelength is set to a value that’s compatible with your solvent & analyte, and confirm that the connected solvent is what you intended to use. Check the signal gain and attenuation to ensure that your signal isn’t being suppressed into the baseline. Next check and re-zero or replace your lamp. If the problem still persists, then your detectors may be at fault.

I see peaks I don’t expect / the same peaks every run! (Ghost Peaks)

First, identify your ghost peaks by running your gradient method without any analyte. You can minimize ghost peaks by using HPLC grade solvents, and degassing, filtering, or using purification/water removal technologies.

Next, identify any injection issues by injecting a blank sample. Problems linked to injection indicate a mechanical part failure. You may have a poor seal that’s leaking, or you may have pump seals or injection rotors breaking down.

If your ghost peak is inconsistent, you might have a late bloomer in your sample. You can check for this by running your last sample and doubling the method time. The best thing you can do is to flush your injector between samples, and incorporate a “Solvent B” purge into the end of your method.

There’s something wrong with this LC Injection Syringe!

Even carefully maintained and properly used syringes wear out quickly with normal use, so it’s best to consider your syringes as consumables. Injection syringes are extremely fragile, and prone to failure at the slightest misuse. LC can involve corrosive liquids which can jam up a syringe pump within a few exposures.

Proper cleaning includes flushing the syringe with an appropriate solvent after every sample. Follow this with water and several acetone rinses, and then place the syringe back into its case.

We recommend that you always have spares on-hand. Replacement parts can be helpful, but undetectable flaws can make it difficult to change them without accidental damage to the new parts. It doesn’t take many bent replacements before it becomes cheaper and less frustrating to simply have an extra.

I heard you can regenerate a used column. How does that work?

As you run samples and solvents through your column, it will become less effective. While it is technically possible to regenerate a column, it isn’t always helpful to try. So before we delve into the process, let’s take a look some situations where you probably shouldn’t bother:

:(  If you’ve run samples with no filter or guard column
:(  If the column has already been regenerated several times
:(  If the column’s connections or fittings appear visibly worn
:(  If the column’s packing has begun to ooze out or channel (gaps within the material)

If any of those conditions are true, then regenerating the column won’t be your best option. Column regeneration involves the use of several different solvents which your lab may not carry – this can be a large initial expense, take up valuable lab space, and add to the regulatory overhead. The potential savings may not be worth the time investment for busier labs.

Important!  If your lab typically runs samples of unknown identity, your components probably have very short lifespans. But this doesn’t have to be the case! You can safeguard your column by pre-filtering your analyte. A standard vacuum filtration through a Nylon 66 membrane filter will remove anything that would get trapped on the column or in your instrument fittings. There’s no need to worry about sample loss, because anything trapped by this filter wouldn’t have been suitable for LC analysis in the first place. You can also pre-filter your solvents if you have concerns over their purity.

It’s also important to have a realistic expectation of what column regeneration can offer. When done properly, it will extend your column’s useful life and remove ghost peaks, but it won’t bring your column back to 100%. A regenerated column is like a tuned-up used car: it may run perfectly fine, but there will always be an increased of developing problems.

The table below provides the steps for the most commonly used columns. Different processes are required for special columns such as zirconia, or columns used for proteins and other biological samples. You also can design a customized process suited for the solvents and columns used within your lab.

Regeneration Processes for Silicate Columns

Column VolumesNormal PhaseReverse PhaseIon Exchange
20Hexane60 °C DI Water60 °C DI Water
20Methylene ChlorideAcetonitrileAcetonitrile
5IsopropanolIsopropanolIsopropanol
2060 °C DI WaterHeptaneMethylene Chloride
5IsopropanolIsopropanolIsopropanol
20Mobile Phase/Solvent AMobile Phase/Solvent AMobile Phase

Important! Hexane can be substituted for any HPLC grade hydrocarbon solvent. Isopropanol can be substituted with any anhydrous, HPLC grade alcohol, but it is the best transition solvent, but it is very viscous and should be run at a moderate flow rate.

Important! If your contaminants are metallic, these can be removed by including ETDA in the DI water pass.

Important! We don’t recommend “all in one” column regeneration mixtures, because no mixture of solvents can properly treat the column packing in one step.

Column volumes are the number of times the void volume for your column must completely transit the column to achieve the intended result. In plain english: For every milliliter per minute of flow rate (r), a full equilibration will take twenty times your column’s void volume (v). In an equation: 20 × (v mL) × (r mL/min) = # of mins.

The void volumes for most standard column sizes are given in the table below.

HPLC Column Void Volume (0.70 Average Pore Volume)

Column Size (mm x mm)Void Volume (mL)
300 x 50412.33
300 x 2172.74
300 x 1016.50
300 x 7.810.03
300 x 4.63.49
300 x 2.10.73
300 x 1.00.17
250 x 50343.61
250 x 2160.61
250 x 1013.75
250 x 7.88.36
250 x 4.62.91
250 x 2.10.61
250 x 1.00.14
150 x 50206.17
150 x 2136.37
150 x 108.25
150 x 7.85.02
150 x 4.61.75
150 x 2.10.36
150 x 1.00.08
125 x 50171.81
125 x 2130.31
125 x 106.87
125 x 7.84.18
125 x 4.61.45
125 x 2.10.30
125 x 1.00.07
100 x 50137.44
100 x 2124.25
100 x 105.50
100 x 7.83.35
100 x 4.61.16
100 x 2.10.24
100 x 1.00.06
50 x 5068.72
50 x 2112.12
50 x 102.75
50 x 7.81.67
50 x 4.60.58
50 x 2.10.12
50 x 1.00.03
30 x 5041.23
30 x 217.27
30 x 101.65
30 x 7.81.00
30 x 4.60.35
30 x 2.10.07
30 x 1.00.02
15 x 5020.62
15 x 213.64
15 x 100.82
15 x 7.80.50
15 x 4.60.17
15 x 2.10.04
15 x 1.00.01

If yours isn’t listed, you can determine its void volume for your column by calculating its volume and multiplying it by the average pore volume (usually 0.70, often 0.50). Convert all column dimensions to centimeters to get an answer directly in mL (1 cubic centimeter ≈ 1 milliliter), and divide the column’s inner diameter by 2 to get the radius!  v0 mL = 0.7 × π × l cm × (r cm)².

Important! Several manufacturers have copied mistakes from each others’ materials, leading to widespread confusion about void volumes. The volume of a cylinder is calculated from its radius squared – don’t square the diameter!

As always, we encourage you to contact us if you need help.

 

Is there something else that you think should be included on this page?

Let us know with a comment below!

 

Comments