method development problem

[lube chemist]

I think I have a pretty basic problem but I'm stuck and hoping you guys can help me out. We have a Waters 2695 equipped with a 410 RI detector and a 996 PDA. It has always been used as a GPC system. Recently, I have been working on synthesizing an ester and would like to use an HPLC column to analyze the purity of the ester as well as to monitor the reaction.

I found a waters Nova-Pak C18 (3.9 x 150mm) column stashed away in the lab. It had never been used (package still sealed) but it is of unknown age. I got myself some HPLC acetonitrile and water and installed the column. I have been trying to develop a method to analyze my materials and am not having much luck at all. The PDA doesn't seem to really pick anything up. I don't expect this ester to have a big absorbance but I would think there should be something at least down by 220 or so. On the 410, I get a big peak at < 1 minute, no matter what solvent ratio I use and that's it. I've run the range between 100% ACN down to 40% ACN (all isocratic). The samples are roughly 20-30 mg/mL in ACN, I'm injecting 20 ul (also tried 40) and running at 2 mL/min for 10 minutes.

Any ideas on what my problem is?

[Uwe Neue]

The first question is how much analyte you are injecting and what your expected signal is going to be. At the high organic concentration, your ester is likely to be unretained. So you could calibrate yourself what kind of signal you should expect. You could even inject without the column, if necessary.

The RI is not that helpful, since it will see both the difference in solvent compostion and maybe your ester. From that standpoint the low UV with acetonitrile/water is the better choice to look for your analyte. Also, does the analyte have other functional groups, or just the ester function?

[lube chemist]

The vials are approx. 20-30 mg/mL and I'm injecting 20 or 40 uL. I think the big peak I am seeing at the beginning is the ACN solvent. The product ester (a diester) has only the ester functionality. I'm also analyzing the starting materials, an acid and an alcohol.

I figured that the ester would elute right away, but I thought that as I increased the water content that it would shift. I've tried up to 60% water. Maybe I just didn't go far enough yet?

The peak I'm seeing is around 500 mv on the RI detector.

[Uwe Neue]

If I understand you correctly you are getting a peak, and this is at the solvent front. If this is the case, this might very well be your analyte. Go to a higher water content, in steps that do not exceed 5%. Stop at 5% organic. If you still have an unretained peak in 5% acetonitrile, you need to get a different column, Atlantis T3.

[lmh]

(1) Any particular reason for not running a gradient from very low organic upwards? It gives a quicker idea of the rough hydrophobicity of your analyte than running huge numbers of isocratic methods.

(2) If your ester has virtually no hydrophobicity, look at what it does have. Does it have phenyl groups that would interact with a phenyl column? Would it do better by ion exchange? Is it a case for hilic or normal phase?

[lube chemist]

I tried several more things and still could not get anything to be retained on the column. I called Waters and based on the lot number, they estimated that the column is 13 years old! She said that they normally only recommend a 2 year shelf life once solvent is in the column, so there is probably no C18 left on this column.

Hiopefully I will be able to order a new one and get my analysis done!

[Uwe Neue]

It is not clear if the lack of retention is due to the lack of C18, or due to the polarity of the analyte. I strongly recommend to use the Atlantis T3 column, since it will give you the flexibility to go to 100% water, if needed.

Do you recall how the column was stored?

[danko]

Lube chemist,
13 years should not be a big problem, provided the column has never been used. And that’s what you told us in your initial post.

Also,

there is probably no C18 left on this column.

I wonder where it’s gone.
Finally: Did you confirm that the compound elutes un-retained? As Uwe suggested earlier, did you inject the sample with no column installed? In this way you should at least be able to obtain the UV spectrum of the compound and compare it to the spectrum of the peak you see when there is a column on.
Maybe you won’t need the RI detector, which will allow you to run gradients – until you find a viable mobile phase composition anyway.

Best Regards

[Uwe Neue]

Actually, Danko is right. I forgot that the column box had never been opened. 13 years ago is 1996. This is long after we started shipping all columns in acetonitrile. Acetonitrile does not affect the SiC bond, so you got yourself a brandnew column that is just 13 years old.

The C18 is there. But you still have not yet told us anything anout your analyte that would allow us to make a judgement under which solvent composition it should be retained.

The RI is not a good choice for method development, if you can't find your peak. The unretained peak could be anything - your analyte, your solvent, even minute differences in the solvent composition between your sample and the mobile phase, even if you used the mobile phase to make up your sample.

[lube chemist]

The column package said acetonitrile and water. THe tech support person at Waters told me that the C18 hydrolyzes off of the silica support over time, accelerated if it is in water or methanol.

My analyte is a diester, no phenyls or any other functional groups. Not particularly polar, molecular weight around 400. I'm also running the starting materials, a long carbon chain acid and a diol. Nothing was retained.

I am no chromatography expert, but I really don't think that this anaylsis should be all that complicated.

Thanks for all your advice!

[Uwe Neue]

If the product is indeed 13 years old, I don't give a damn what the care and use manual says. It was shipped in 100% acetonitrile.
The tech support people are right in the sense that if it would have been shipped with an aqueous mobile phase, you would need to put the column into the trashcan. But see comment above!
From the additional info that you are providing now I would guess that your analyte is reasonably hydrophobic. A long carbon chain and a MW of over 400 suggests to me that you should run a mobile phase with 100% acetonitrile and inject a sample that is dissolved in acetonitrile and expect some retention, monitored at both 210 nm and RI.

The complication with your analysis is exclusively the detection method.

[HW Mueller]

I don´t see the sense in making any further suggestions unless we know how the diester looks if run through the system without a column (as has been suggested several times).

[shaun78]

Oh how about one more suggestion just for fun.

Last post I saw it in, you were injection 20 to 40 uL of a solution that had a concentration of what appeared to be 20-30 mg/mL.

This is way too concentrated and you are likely overloading the column, which is why you can not get your analyte retained.

[lube chemist]

Thank you for the additional suggestions. I made up a new sample at 2mg/mL and injected 10 uL. I also put the column back in 100% acetonitrile. One thing that is encouraging to me is that when I did this, I monitored it during equilibration and saw a few peaks elute.

I did two runs - one with and one without the column, both 10 uL injections and run at 1mL/min. Here are the results. Obviously the blue line is the run w/o the column.

[DR]

So, you do have some retention - and possibly 2 peaks. Now, you need to figure out what will work in terms of isocratic conditions (if you want to use RI in addition to UV), or you need to optimize gradient conditions to get peaks separated in a short time (if you're OK using just UV, as I would do).

I'd do the typical DryLab prescribed runs (20', 60' gradients from - in your case - 5%-90% ACN), then you can start to dial in more optimal conditions, figure out if reasonable isocratic options exist etc. etc.