HPLC Column Performance Check

[Milosh]

Dear collegues, I have one question. When I recieve the HPLC column which I order, in the certificate is mentioned which test solution or mixture is used by manufacturer to check the column preformances.
Our idea is to check the column with the same test solution, when we recieve it, and later, periodicaly. But, I have a problem to find and order such a mixtures.
If somebody have some experience with it, please help me.

Best regards!

[zokitano]

Dear Milosh,

Firstly, you should ask your column vendor about it. The one who sold you the column should be able to provide the specific column performance mixtures for you.

Regards

[danko]

Hi Milosh,

The only reason, I can think of, for testing the column with the manufacturer’s test solution is, if I don’t believe in the provided documentation. Now you don’t mention the brand so I can not be 100.000% certain of the manufacturer’s honesty. But even not knowing the brand I would be 99.9999% comfortable with the provided test data. The competition in this field is very tough so the column makers can not afford to lose customers.
However, if you have lifecycle test in mind, then you don’t need the manufacturer’s test compounds or procedure at all. On the contrary, the most relevant test for you would be your own system suitability test for the intended analysis. Just establish a SST parameters/values chart and log them (for every particular column) every time you run the relevant method. These parameters/values could be Resolution, Plate count, tailing or whatever is critical to the obtained results’ validity.
And there you have it: A perfect column lifecycle – unique and relevant to your needs.

Best Regards

[Bruce Hamilton]

One reason for testing the column using the manufacturer's test mixture is to ascertain how your system compares to the test system.

Many of the column test methods are isocratic, relatively easy ( mobile phase is often just CH3CN:H2O ) and it's always useful to know how the column behaves on your system before putting your gunk onto it.

I use the manufacturer's test composition to confirm how a column is performing during use, and if I have a dodgy column, I can return it, or get discount on a replacement.

I agree it's not that common, but I recall at least two situations where column performance did not match the certificate, and the failed column was exchanged. The replacement columns worked fine.

I've also had two other situations where a column degraded faster than expected, and providing the initial and subsequent test chromatograms to the manufacturer resulted in a signficant discount on the next column.

If you are a bulk buyer of columns, and a QC/QA user, your system suitability may detect an inadequate column, but it may not...

If you have put samples through the column, the supplier could claim the failure was due to your samples, especially for more fragile columns - such as C8, CN, etc.

I'd certainly recommend testing expensive columns (eg chiral, prep. ) before using on real world samples, especially if the samples may be full of gunk...

One word of warning. Various vendors use different calculations, so ensure you know how your CDS, and the vendor's CDS, calculate the data before comparing your actual data with certificate specification limits.

Please keep having fun,

Bruce Hamilton

[danko]

Hi Bruce,

While I recognize the validity of some of your ideas/thoughts, I will allow my self to comment a couple of them. Hope it’s OK.

One reason for testing the column using the manufacturer's test mixture is to ascertain how your system compares to the test system.

It smells of instrument qualification – more than a column performance test. I would much rather handle system and column issues separately.

If you are a bulk buyer of columns, and a QC/QA user, your system suitability may detect an inadequate column, but it may not...

If my SST does not detect any insufficiency, then I’m fine! Why would I worry as whether or not the column would perform better or worse with regard to some compounds, I’m not planning to analyse?
The role of the SST is precisely to document an adequate performance of the system (including the particular column) with regard to the tested samples/compounds.

I agree it's not that common, but I recall at least two situations where column performance did not match the certificate, and the failed column was exchanged. The replacement columns worked fine.

There you go: As you mention you recall a couple of situations where you found an inconsistency (I’m not sure whether or not it would have been an issue for your analytical method but let’s assume it would have)
Maybe the vendor would have exchanged these 2 columns anyway – assuming they didn’t perform as expected according to your SST.
But most importantly you should compare the time invested (hence the cost) in testing every column with different eluents and compounds to the cost of a couple of columns you potentially would dispose of, every 5 years, or something like that.
What is then most cost effective?

Best Regards

[Bruce Hamilton]

One reason for testing the column using the manufacturer's test mixture is to ascertain how your system compares to the test system.

It smells of instrument qualification – more than a column performance test. I would much rather handle system and column issues separately.

Maybe. But how many of your column suppliers detail their test systems on their certificates?. Few of mine do. For me, it's nothing to do with instrument qualification, it's about ensuring I understand how a brand new column performs on my system, and how it degrades during use on my system. IQ merely measures how my HPLC matches the manufactuers' specification, and should be designed to be column independent.

If you are a bulk buyer of columns, and a QC/QA user, your system suitability may detect an inadequate column, but it may not...

If my SST does not detect any insufficiency, then I’m fine! Why would I worry as whether or not the column would perform better or worse with regard to some compounds, I’m not planning to analyse?
The role of the SST is precisely to document an adequate performance of the system (including the particular column) with regard to the tested samples/compounds.

I noted that system suitability may detect an OOS column, but for users like me, where I use a column for a diverse range of analyses, the vast majority of which don't require a system suitability test, the manufacturers' tests are simple, effective, and relevant. YMMV.

I agree it's not that common, but I recall at least two situations where column performance did not match the certificate, and the failed column was exchanged. The replacement columns worked fine.

There you go: As you mention you recall a couple of situations where you found an inconsistency (I’m not sure whether or not it would have been an issue for your analytical method but let’s assume it would have)
Maybe the vendor would have exchanged these 2 columns anyway – assuming they didn’t perform as expected according to your SST.

The reason I instituted the testing of columns is because of an earlier situation, where a manufactuer refused to accept that a column was faulty because we had put samples through it before testing it. I had to swallow hard buy another and complete the project. I still try avoid columns from that manufacturer.

But most importantly you should compare the time invested (hence the cost) in testing every column with different eluents and compounds to the cost of a couple of columns you potentially would dispose of, every 5 years, or something like that. What is then most cost effective?

As I noted earlier, the vast majority of performance tests are simple isocratic systems, and easy to perform. The test solutions are stable, and the analysis typically only take about 15-20 mins. Some manufacturers, such as Phenomenex, may use their recommended storage solvent composition ( 65/35 CH3CN:H2O ) as the test mobile phase, so it's easy to do on most of their reverse phase columns. Other ratios of CH3CN:H2O are easy on many systems.

I only perform the test when I get the column, and when I suspect something is wrong. I also use it when a column forms a void and I repack it, often plates plummet to 20-30% of new, and come back to 90+%, which is fine for most of my work. Takes 5-10 minutes to fill a void in a column and compress it.

As I noted earlier, a column typically costs several hundred to several thousand dollars, and whether column testing time is justified depends on your environment. It is for me, but YMMV.

Please keep having fun,

Bruce Hamilton

[goxy43]

Dragi Miloše,
(let us keep the answer in English)
Each and every column manufacturer uses different test mixture, and they do mix this test sample themselves. You can order each component separately from Sigma or VWR and then prepare the test mixture yourself. It is quite simple to do so.
Testing the column is quite easy, as already written: isocratic separation ( we do it with 60% acetonitrile/40% water). You can contact me off-line for more details

[Rob Burgess]

I'm not totaly sure the test chromatogram report that accompanies new columns is completely relevant. The user needs to be sure that it can run pre-defined methods that the column type was initially developed on.

For instance, we have a method for basic pharmaceuticals that showed severe tailing peaks for one particular packing lot of columns. The "column test" is studied for neutral compounds. So the question is where is the relevance to test for suitability for my particular method - and this was meant to come from our company approved batch - which appears to be meaningless for our actual applications!

[goxy43]

Hi Rob,
I do agree that column testing is not relevant when comparing to actual application. However, column testing with neutral substances is the best way to say: OK, this is a good column! I mean: good packed column, no void volume etc. This is the only purpose of testing the column prior to ship it. I cannot imagine that the producer will test every and each column according to specific needs of any customer. Method development is up to the user. And I know from some of the manufacturer that they simply test the batch of the stationary phase and not every packed column. Some other do test every column (LCPackings (now Dionex)does this for every column they pack). Therefore, performing a column performance test with neutral compounds (usually polycyclic aromates) would be a great tool to check it and eventually claim a new one.
best regards

[danko]

I cannot imagine that the producer will test every and each column according to specific needs of any customer.

That’s right and no one would – and that is why the customer should test the column with a SST sample relevant to his/her particular method. As I mentioned earlier, if the column performs according to my expectations with regard to my application, then it is good and I don’t care about its performance with some compounds I’m never going to run on it.

And I know from some of the manufacturer that they simply test the batch of the stationary phase and not every packed column.

The brands I buy – and I buy many different brads - come with certificates for every single column. You can convince you self by comparing the test results for several different columns (otherwise identical) and you’ll see that “plate countâ€

[rick1112]

Hi

Regarding column performance could there be a universal method which can be applied to all RP-columns?? ( this test can at least used for internal usage)

[Sunjay]

I think it may not be a good idea to keep testing each and every column as per the supplier method, as this will a costly affair, as all the manufacturer uses the different chemicals.

We are testing the columns for SST for the product they will be used. We have returned number of column to the manufacturers based on our failing SST results. We were using the USP method, and column could not pass the SS crititeria mentioned in USP, we took up the matter and it was replaced.

Regards

[rick1112]

hi

well regarding my problem...we use a mixture or benzene and toulene for analysis of column performance , this method used to give satisfiable result for my RP-columns
recently we have started using ACE RP-columns, but my method does'nt seems to work in these columns , i just check four columns with above method none of them are giving base to base seperation of the compunds..? is this a problem with my method?? (actually i have never used this method personally to check column performance, my seniors says it works well for other column and gives base-to-base seperation), or is it some problem with the column??? (the column performance check that comes with the test chromatogram for ACE-column uses, benzene, bipenyl and other compunds (65%ACN, 0.2 ml flow rate, 25 degree column temp.)

[danko]

Rick,

The only conclusion you are able to reach to in this context is: The ACE column does not separate benzene and toluene as well as your previous utilized columns (unless incompatible separation parameters is the cause of the inadequate performance).
And that is exactly my point (see previous posts). Firstly, an “RP columnâ€

[Uwe Neue]

I remained outside of this discussion until now, but I thought it being now worth my while to comment:

At Waters, there are two different test methods. One test looks at the batch-to-batch reproducibility with probes that are sensitive to the surface chemistry, and the column test uses a simple hydrophobic probe, acenaphthene, to check plate count and peak asymmetry, which are a function of the packing conditions.

Most methods, especially those with smaller particles, are set up to work near the minimum of the van-Deemter curve. This is the spot where the test method is most sensitive to the packing method, and where one gets solid information about the true column performance. This is also the spot where extra-column bandspreading plays the largest role. You can get the exact same information only if you have a properly designed HPLC instrument with minimum bandspreading. Many HPLC instruments in the real world have more bandspreading, and are therefore less suitable for accomplishing the same test and for getting the same results.

The influence of the extra-column bandspreading increases with lower retention factors. You can check the quality of a manufacturers test instrument, if the manufacturer uses multiple probes to measure the plate count. If the plate count increases with retention, the chances are high that the instrument has a fair amount of extra-column bandspreading. Conversely, one will get higher plate numbers at low retention factor, if the test instrument is good enough for measuring this.

From the standpoint of the user, I think that all these points are things to be aware off, but they point to the fact that column testing under some standard conditions is not easy nor is it necessary. I agree with the attitude that the user should go to his method and verify the column performance under his use conditions.