Void Volume Question

[Gizmo]

Undertaking some WS vitamin analysis and we have a method that shows B12 as eluting in the void volume just a bit before calculated t0 (elutes at ~3.1, calcuated t0 is ~3.5 at the flow rate used) - other B vitamins (6, 3, 1, 2) are retained and elute at later times post t0 and appeared to have k values that were reasonable.

I am a bit stumped by this. Assuming an undamaged column is this even possible?

This is a C18 column, with a pretty non extraordinary HPLC RP protocol using isocratic MeOH/H2O at 55:45 UV monitoring at 230 and 280nm.

I'd throw up more details but really I'm just looking for the basic answer as to whether or not this should even be happening before I raise a ruckus. I'm inclined to question the method and perhaps the fact that we have a peak in the calculated void volume is enough, but test samples keep coming back close to their known concentrations providing an argument that the method and equipment are just fine to proceed with.

The method makes use of different injection volumes for producing the standard curve which I was sure was involved but with a uniform injection volume for the samples we are seeing data that matches (enough) to now rule that out.

Thanks in advance for any thoughts on this.

[Multidimensional]

You did not provide the column dimensions or the flow rate used so there is no way to confirm your column void volume estimate.

If your only question is 'can the estimated tzero time and measured tzero time vary?', then the answer is of course, YES, they can. The estimated Tzero time (as calculated from the calculated Column Volume and flow rate) are just that, estimates only. Relative to the column volume, sample volume (yes, injection volume and sample concentration can throw it off a bit) may effect it a bit, but the actual measured time of the unretained std is the one to use. We usually confirm this time by injecting a blank (mobile phase) and watching for the tiny blip on the detector which will correspond very closely to the true Tzero measured.

Note: Std curves should always be prepared from std solutions. You should use identical injection volumes for all stds and samples. Do not vary injection volume to 'create' different std concentration values. Using different volumes is incorrect technique and unreliable (because injection volume affects the exact peak shape and retention time).

For expert information on how to estimate AND measure Column Void Times and also K primes (very important), please refer to these free articles.

https://hplctips.blogspot.com/2011/05/d ... -time.html

https://hplctips.blogspot.com/2015/06/k ... actor.html

[Gizmo]

Thank you - and those articles were quite helpful. Regardless of exactly where t0 is (and we should determine this as you described) it seems that the "analyte" here is eluting at a point that is very far from ideal. Column dimensions are 4.6 x 150 mm and with flow at 0.5 ml/min I believe that the estimated t0 of ~3.5 minutes was reasonable based on the calculations.

[tom jupille]

Somewhere around three minutes is about right for t0 so one way or another you are looking at an unretained peak. The problem with an unretained peak is that you don't know for sure what you are looking at. Since the calibration plot was done by varying the injection volume, it is quite possible that you are simply looking at a system peak ("t0 noise") which is proportional to the volume injected -- I learned that the hard way many years ago! You could sort that out by re-running the calibration with constant injection volume.

That said, trying to quantitate a weakly- or un-retained peak is dangerously error prone. If that's really the way the method is written, it's a *bad* method.

[Multidimensional]

If your peaks are eluting at or near the Tzero time, then the "method" you have is invalid and must not be used. Failure to show good retention of any/all samples is a failure of the method. *This is why we calculate and measure Tzero during method development. It allows us to calculate our sample's K prime value to make sure we are proceeding in a logical manner. "K prime" is the measure of retention on the column.

When peaks elute at or near the Tzero time, no chromatography is taking place so no purpose exists for running an HPLC analysis as the results will match simply placing the sample in a spectrophotometer for analysis (this of course assumes you are using UV/VIS detection). The main purpose of chromatography is to develop a method which shows that it is specific and selective for a specific compound. This is accomplished by retaining it on the column in such a manner that is can be resolved away from any other compounds which may or may not be present at that time. It provides us with one dimension of separation, retention time, plus in many cases a second dimension of a spectral plot (again, for UV/VIS analysis by DAD, commonly used in HPLC method development). When proper method development technique is ignored and the peak(s) elute at or near the Tzero time, no chromatography and no proof of selectivity or specificity can be shown. The result in an unscientific method which can not be used.

[HPLCaddict]

Coming back to the original question:
1) You can not calculate the "true" void volume. All you can do is an estimation(!) of the true void volume. I don't know which formula you used for your estimation, but using the good old V0 = L * d^2 / 2 gives me a void volume of 1587 µL for your 150x4.6mm column. At the flow rate of 0.5 mL/min this gives a t0 of 3.174 minutes. This estimation (!) corresponds pretty good with the retention time of your peak in question, so I wouldn't say it elutes significantly before t0 but actually pretty much at t0.
2) All that said, yes, it is possible for analytes to elute before t0. t0 represents the time an unretained analyte spends in the column, meaning it doesn't interact with the stationary phase, but it goes all the way the mobile phase does. So when speaking of the usual silica-based columns, this means it penetrates into the pores. There can be situations when analytes do not enter the pores, e.g. for large molecules which are simply to big. Or very polar molecules might be excluded from the rather hydrophobic pores of reversed-phase materials. These analytes, which do not "travel" through the pores might elute before t0.

[Rndirk]

Theory aside, there is something very wrong with the method if vitamin B12 elutes before B1 and B3 in reversed phase. On the other hand, starting at 55% MeOH, B1 will have little to no retention.

From experience a rough order of elution in RPLC for your compounds, working with a generic gradient 5% MeOH to 95% MeOH, should be B1 > B3/B6 >> B12 / B2

It is possible that B12 doesn't elute at all in your isocratic method.

[Gizmo]

Thanks for all the replies - you all have confirmed my concerns. We have some work to do here.

[Rndirk]

We have some work to do here.

I would advice not to start from scratch. There are plenty of HPLC application notes for water soluble vitamins giving all the necessary details to set up a method.

In my experience the challenges for your compounds are, or should be:

1) Giving thiamine (B1) enough retention
2) Dealing with the low solubility of riboflavin (B2). It's just annoying to make stock solutions of this compound, you'll have to go <50ppm.