Optimizing Analysis Conditions, Part 2: Column Set
Last month I wrote about optimizing the mobile phase when analyzing a new sample with your OMNISEC system. In this post, I’m going to shift the focus to finding the right column set for a sample. Like last time, I’ll use example data to lead you through my process of trying to improve the resolution of the sample.
It should be noted that not every sample or sample type has a corresponding column set that will resolve every single component. As with the example below, sometimes the best solution doesn’t look perfect. Hopefully the strategy I employ here provides you with ideas and encourages you to try different things – you won’t know how a column set works until you try it.
Initial analysis – 2 x T6000M
Before any analysis of a new sample, I work through the series of questions below.
- Are there columns compatible with the chosen mobile phase?
- Is the sample compatible with the column type (e.g. cationic samples might require cationic columns)?
- Is there a targeted molecular weight / molecular size or is the sample unknown?
- Will the sample require resolution in a particular area (e.g. between sample & solvent)?
I didn’t know much about the sample going into the analysis. I knew it was soluble in THF and contained a mixture of components with a wide range of molecular weights. That answered two of the questions.
Since I knew THF would work as a mobile phase, and the samples required a wide molecular weight range, I decided to try the column set I use the most: two mixed bed organic columns (2 x T6000M). This is a great general-purpose column set because if offers good resolution over a wide molecular weight range, roughly 1,000 – 4,500,000 Da (for polystyrene).
As you can see in the data above, the chromatography clearly shows something, but it’s difficult to tell if the sample is completely resolved in the low molecular weight region (eluting right before the negative solvent peak around 25-26 mL in the RI signal).
The good news is that the sample is eluting well; there is no evidence of the sample interacting with the column. Therefore, all I needed to do was adjust the molecular weight range of the column set. Since all of the peaks present in the chromatograms were close to the solvent peak, I decided to try a set of low molecular weight columns next.
Second analysis – T2000 + T1000
In my experience, the T1000 low molecular weight column is the best for maximizing resolution between the sample and solvent peak. Since I didn’t want to have the entire sample exclude, I added a T2000 column to increase the molecular size and weight range of the column set. From the initial analysis, it was difficult to tell how big or small the size of the sample was.
Therefore, I replaced the 2 x T6000M column set with T2000 + T1000 (in that order) and re-analyzed the sample.
At first glance, this probably looks like a step backwards. However, the column set did do what I was hoping, which was increase the resolution in the low molecular weight range. Everything eluting from 14-20 mL had previously been smushed together in the peak eluting at 23 mL in the initial analysis with 2 x T6000M. Observation of the low molecular weight materials in the sample was one of the goals of the analysis, so I knew keeping the two low molecular weight columns was necessary.
It was also clear that a decent portion of the sample was excluding from the low molecular weight column set. The sharp rise of the sample peak in all detectors at 10 mL suggests that there are polymer molecules too large to diffuse into the pores of the low molecular weight columns. Since the resolution of that portion of the sample looked good with the 2 x T6000M columns, and I knew I also needed the T2000 + T1000 columns to resolve the low molecular weight materials, I next decided to try…
Final analysis – 2 x T6000M + T2000 + T1000
I tried not to overthink it. By simply combining the two column sets I knew I would create a column set that possessed the resolution the sample needed across its entire distribution.
Just like with the mobile phase optimization, I was able to select this example after seeing the results. But a difference between these examples is that the results of this process don’t look like a textbook example of chromatography. And that’s ok!
This sample contains multiple components. Even though some of the peaks still overlap, I was able to work with the person who submitted the samples and identify the key peaks. Even though the light scattering signal is only strong enough to provide reliable molecular weight data for peaks 1 and 2 (the materials in peaks 3-5 are too low in molecular weight and concentration), the ability of the RI to provide relative concentration for peaks 3-5 was enough to provide valuable information.
Final thoughts
I hope the example described in this post helps you understand the thought process behind my progression from one column set to the next, in the pursuit of maximizing the resolution of this sample. If you have any questions, please don’t hesitate to contact us or email me directly at kyle.williams@malvernpanalytical.com.
Related content