Do I need a Guard Column?

[pippin24]

Hello,

I wonder whether having a guard column is a good idea for my LC application.

The analytical column we use is Waters Sunfire C18 analytical column, 4.6 x 150 mm, 5um. The mobile phase is a mixture of 50:50 (v/v) MeCN and 0.1% dilute phosphoric acid. It's for simvastatin detection.

I wonder whether a guard column is necessary for the separation.

Thanks!

[tom jupille]

- If the method was validated with a guard column, then you should use it.

- If the method is unvalidated or was validated without a guard column, then the answer depends on how much crud is in your sample. If samples are clean and predictable, then a guard column provides little or no benefit.

[Uwe Neue]

My own experience with guard columns has been so positive that I always recommend them, no matter what you are doing. I also know that this is a rather extreme standpoint, and that in many applications you may be happier by buying a new column once in a blue moon.

[lpnian]

if your sample is dirty your should use a guide column i think.

[Ron]

Not very scientific, but my rule of thumb is if you have an analysis where you consider using a guard column you probably should.

[Supercritical]

Unless the method calls for it, I never use a guard column. I filter mobile phases, solvents and samples if needed. I never specify a guard column when I develop a method. I've actually seen guard columns cause more problems than they solve. Leaks, band broadening, and peak shift to name a few.

[HW Mueller]

We have been through this before, nevertheless:
since almost all plugging of columns in this lab turned out to be due to frits I always use a guard frit, rarely a guard column (the latter only with highly concentrated biological fluid extractions). Even with relative clean biological samples one should implement a rigid cleaning regimen (also discussed previously), which is best done by reversing flow on the column. Even if you use a precolumn. (The washing should be done at least once a day before shutting down for the night, don´t leave pure organic on the column over night, wash the precolumn separatly if used, or renew).
Another reason for reducing the use of guard columns in this lab is that stat phase repairing kits are usually no longer available (used to fill Upchurch, etc. empty 2x20mm, etc., precolumns with that). The expense of commercial guard columns is out of proportion with their usefulness iin my opinion (and experience).

[Uwe Neue]

Well, here are real life data:
1. Clean sample:
Without guard column, the column lasted for 2000 injections of a very clean sample. With guard columns, the column did not show any sign of degradation even after 10 000 analyses. Repeated twice with different column chemistries.
2. Cleaned plasma sample:
Without a guard column, the separation deteriorated after 250 injections. With the guard column, the analytical column lasted for over 1000 injections with the same dirty sample. Also repeated twice.

I personally do not mind if people are buying lots of columns

[MG]

I run mostly dirty samples (urine, plasma extracts) and am definitely in favor of guard columns. Usually when I see guard column failure, it is not due to plugging. Instead I see increased peak tailing and retention time variability. Now I have a triple-quad, so separation schmeparation, but the peak tailing cuts into my LOD, and QA really doesn't like the retention time variability. Reversing the flow and flushing the guard doesn't help, so I replace it. Then I think to myself, "man, I'm glad that was only my guard column, not my column".

[HW Mueller]

Uwe,
your first example was not that X-Terra at high pH?....sorry, can´t help it sometimes. But seriously, there must have been some disagreable chemistry there. On the 2nd example it, of course, depends on what you mean by "cleaned". Also, in my cases I cleaned the column after about 20 injections, so I can well imagine what you got after 250 injections. Did you change your guard column after every 250, or so inj. (in the experiment with guard)? Now a proper wash might be cheaper after all. Please don´t misunderstand me, I am also for guards in some cases.

MG,
you ran your stuff until peak form changed yet the analytical column was unscathed? Very interesting. What size guard was that? What did you use to flush? Organic? If so you probably cemented some proteins to the guard. No chaotropics or detergents, used? Did your guard develop a void? I can´t recall ever to get an indication to replace a guard other than a pressure increase or shoulder formation due to voids.

(I am just trying to reduce this "feeling" (art) aspect of chromatography, also trying to get away from my habit of not trying something again after a bad experience [my TFA aversion sits a little deeper, though]).

[MG]

you ran your stuff until peak form changed yet the analytical column was unscathed? Very interesting.

I won't say it was unscathed, but the immediate problem is nearly always solved by replacing the guard. I did have some food sample extracts where it helped to backflush the column with high organic, in addition to changing the guard.

What size guard was that?

2.1 x 10 mm, Betasil Javelin C18

What did you use to flush? Organic? If so you probably cemented some proteins to the guard.

In many of my methods, I'll have a mostly aqueous hold to wash off any polar crud or salts, a fast gradient (10 min or less) to elute my compounds, then a mostly organic hold to wash off any nonpolar crud. It is possible that I am cementing proteins to the guard. Any way to avoid this, or is it better that they cement to the guard rather than going through the column?

No chaotropics or detergents, used?

Not usually, but some plasma extracts may contain ion pair reagents.

Did your guard develop a void? I can´t recall ever to get an indication to replace a guard other than a pressure increase or shoulder formation due to voids.

I always kind of assumed that it was due to the formation of a void, but I don't know. It starts out looking like a tail, with increasing retention times accompanying. If I let it go long enough, it will start to look like a shoulder on the back end of the peak. Occasionally when I replace one, I will notice that a high backpressure accompanies the chromatographic problem, but the backpressure is usually still tolerable. In other words, the deterioration in peak shape is what motivates me to change the guard, not the backpressure.

[HW Mueller]

MG,
the trouble is that one doesn´t usually study such things in themselves so the "believe" aspect is strong here. If you do study it, like Uwe did, it´doesn´t necessarily resemble reality. I take such studies very seriously, though, as they provide a framework for action. So I agree, if you are in doubt, use a guard, you can always drop it later if it didn´t help.
More specifically: In a run it is better to adsorb the proteins in the guard, if you want to wash it you should get rid of them as close to completely as possible.
Ion pair reagents may prevent adsorption, but they also do the opposite. Detergents are similar, but the right one at the right concentration may be best for removing proteins (especially if they are caked out). For either chromatography or washing it is best to start with mild reagents to keep proteins from caking (cementing) out. Sometimes it is best to clean up more before doing RP-HPLC. In the analysis of cortisol in plasma/serum we first used a Na2SO4 precipitation, then a restricted access column (Pinkerton) or ultrafiltration, then a two-step RP.
On crud vs void, can you look into your guard? Adsorbed dirt can give rise to very wierd peak shapes, very equivocal. I have seen columns with huge amount of crud behave normally, others misbehaved with little.

[Uwe Neue]

I do not agree with the statement that a carefully designed experiment does not reflect reality. A carefully designed experiment will eliminate extraneous and confusing effects, but it is still very, very real.

Part of the reason why things do not appear as neat with protein samples as with small molecule samples is the fact that proteins denature in the presence of organic solvents, and that a recovery from such a process is practically impossible. In addition, the denaturing process is not instantaneous, thus not all of it can be caught by a guard column, which by its very nature has a finite residence time.

On the other hand, it can also be shown that under favorable conditions and with proper care, a HPLC column is nearly indestructable. While again the favorable conditions may not be the standard use conditions, this experiment demonstrates at least that column life is not like human life, limited to a bit over 100 at best.

[robin]

Hi! great to have the opportunity to join in this discussion.
I'm a bit of purist -never used guard columns in 32 years of practicing chromatography and I still have columns in regular use which are as efficient as when first used. Lets be honest, most guards are used to mask poor chromatography techniques though I do accept that there are exceptions. IF you have to use them then you accept operating separations under less efficient methodology, and don't forget that the guard column will also deteriorate affecting your column performance. I would look to sample preparation as the key factor and much of this is now automated and highly applicable to biosamples. If you have to go direct then use inert materials in guards to mimimise peak broadening . Protect columns from unavoidable particulates (eg pump generated component residues )with inline filters- even these can be avoided with highly controlled maintenance regimes.
Just a concluding point - To my mind guard columns negate the development of ultra high efficiency separations so lets develop our technology /techniques to do without them[/quote]

[Uwe Neue]

Robin,

A properly designed guard column does not deteriorate the separation. As a matter of fact, a well designed and well packed guard column using the same material as in the analytical column is a small column extension that ADDS to the column plate count. It thus IMPROVES the separation.

In some cases, a good sample preparation can eliminate the need for a guard column. However, in some cases it is virtually impossible to do so. This is the case when one is working with samples of biological origin: plasma, urine, meat, etc. Under these circumstances, a column contamination is practically unavoidable.