Yours Chromatographically, Kaushik Zala

1. Introduction
2. The HPLC
3. The Column
4. Solvents
5. Gradient
6. Flow Rate
7. Sample Preparation
8. Optimizing the Separation
9. What is HPLC Equilibration?
10. Preparing for the First Run of the Day
11. After the Last Run of the Day
12. Should I Control Column Temperature?
13. How Much Protein Should I Load?
14. Links to Related IonSource Material
15. Links to External HPLC Information

1. column dimension (size)
2. particle size and pore size
3. stationary phase

1. Since columns are tubular, column dimensions usually take the following format, internal diameter X length (4.6mm X 250mm).  As a mass spectroscopist you will encounter columns ranging in internal diameter from 0.050 to 4.6 mm or even larger if you are performing large scale preparative chromatography.  For mass spectrometry a short reverse phase column will work nearly as well as a longer column and this is an important fact because shorter columns are generally cheaper and generate less back pressure.  Why is less back pressure important?  If a column runs at low pressure it allows the user more flexibility to adjust the flow rate.  Sometimes shorter columns are used to do fast chromatography at higher than normal flow rates.  In terms of length we routinely run 100 mm columns, however 50 mm or 30 mm columns may be adequate for many LC/MS separation needs.
2. The most common columns are packed with silica particles.  The beads or particles are generally characterized by particle and pore size.  Particle sizes generally range between 3 and 50 microns, with 5 µm particles being the most popular for peptides.  Larger particles will generate less system pressure and smaller particles will generate more pressure. The smaller particles generally give higher separation efficiencies.  The particle pore size is measured in angstroms and generally range between 100-1000 angstroms.  300 angstroms is the most popular pore size for proteins and peptides and  100 angstroms is the most common for small molecules.  Silica is the most common particle material.  Since silica dissolves at high pH it is not recommended to use solvents that exceed pH 7.  However, recently some manufactures have introduced silica based technology that is more resistant to high pH, it is important to take note of the manufactures suggested use recommendations.   In addition the combination of high temperature and extremes of pH can be especially damaging to silica.
   
   
3.  The stationary phase is generally made up of  hydrophobic alkyl chains ( -CH₂-CH₂-CH₂-CH₃ ) that interact with the analyte.  There are three common chain lengths, C4, C8, and C18.  C4 is generally used for proteins and C18 is generally used to capture peptides or small molecules.  The idea here is that the larger protein molecule will likely have more hydrophobic moieties to interact with the column and thus a shorter chain length is more appropriate.  Peptides are smaller and need the more hydrophobic longer chain lengths to be captured, so C8 and C18 are used for peptides or small molecules.  Here is an interesting note: Observations have been made that C8 columns are actually better for capturing smaller hydrophilic peptides, the theory here is that the longer C18 chains lay down during the early aqueous period of the gradient and the more hydrophilic peptides are not captured.  We use C8 routinely for all peptide work and this particular alkyl chain length works equally well if not better than C18 for all peptides.  

Our Preferred Solvent System for ESI LC/MS
A = HPLC grade Water, 0.1 % formic acid
B = HPLC grade Acetonitrile, 0.1% formic acid

The sample is normally reconstituted in the A solvent to maximize binding to the column.  The sample should not be dissolved in an organic solvent or it may not stick to the stationary phase.  The sample should not be dissolved in detergent containing solutions.  Some detergents may bind to reverse phase columns and modify them irreversibly.  In addition detergents preferentially ionize in electrospray mass spectrometry and can obscure the detection or suppress the ionization of the analyte.

Once you have a separation you may want to optimize it.  You may wish to shallow out the gradient to improve the separation, or you may wish to shorten the run time.  Taking the illustration above one can see that all of the peptides are out by 40 minutes.  This does not mean that we can change this 80 min run into a 40 min run, but there is room for improvement.  The first step in the optimization is to determine the %B at which the last peak elutes.  If you look at the blue gradient line you might guess that the last peak elutes near 40%B but this would be incorrect.  All HPLC systems have a gradient delay.  The gradient delay is the time between when the software tells the pumps to start pumping at a certain mobile phase composition and the time it takes for that solvent composition to reach the column and have an effect.  A good guess for a gradient delay is 10 minutes.  This would mean that our guess for the final mobile phase composition for the 40 min peak would be approximately 30%B.  To observe the gradient delay time look at the illustration above and observe that the baseline returns to the starting conditions at 70 minutes and not at 60.1 minutes when our pumps have gone back to 2% B.  One must take care to avoid having the last peak elute on the "equilibration cliff", (at 70 min. in this example).  This can be avoided by ending the gradient at a %B that is slightly higher than that required by the last component.

This would make the gradient shallower and possibly give a better peak separation.  To shorten the run time one could rewrite the gradient to look like this:

This last change would cut 30 min. from the analysis time.  Shorter analysis times are always better for work efficiency.  With every minute you can cut from the HPLC method without sacrificing your chromatographic goals you will be rewarded with better work efficiency. With this change the last peak would most likely still elute at 40 minutes and the peptide separation would essentially remain the same as in the initial 60/60 analysis.