Understanding silica – why matching TLC and flash silica is important for good flash column chromatography

Recently, I posted an article explaining why high performance TLC plates are not needed for method development for high-performance flash chromatography.  Based on some excellent feedback, I see a need to discuss silica chemistry and its impact on chromatography.

In a nutshell, no two silica brands are the same.  Sure, many have the same physical properties, such as surface area and porosity, but the surface chemistry between silica brands is almost always different.  While the surface of silica particles is comprised primarily of hydroxylated silicon (Si-OH), not all of that chemistry is the same. According to Ewe Neue in his text “HPLC Columns – Theory, Technology, and Practice (Neue, 1997)”, there are several different silanol types including single silanols (Si-OH), silanediols or geminal silanols (HO-Si-OH), and bridged silanols or silyl ethers (Si-O-Si), Figure 1.

Figure 1. Structures of the the three silanols types types. Top – geminal. Middle – single. Bottom – silyl ether.

Single and geminal silanols are quite polar and can be acidic. They are good at attracting polar compounds, organic bases, and other nucleophiles.  Silyl ethers, on the other hand, are relatively hydrophobic and not very reactive.

So, how do the differences noted above impact what you do with flash chromatography?  Well, consider the interactions taking place between your compound and the silica chemistry.  Let’s say you have a compound in your sample that is well retained on a silica TLC plate that happens to contain a high silyl ether percentage and you transfer the method to a flash column packed with a high single or geminal silanol content. With that scenario, it is very likely that your compound will elute later than you would expect from the flash cartridge.  In fact, you may need a very different, more polar solvent system to perform the purification.

The ratio of these three silanols types impacts selectivity as well.  Using unmatched TLC and flash silica has even caused compounds to change their elution pattern and/or selectivity.

I did not believe this early in my career but just one sample convinced me it is true.  The sample was a mixture of organic dyes and the eluent was 100% toluene for both the TLC and the flash purification.  Though I do not have a picture of the resulting flash chromatography, I do have the TLC data and the actual flash elution volumes to share, Table 1.

 Biotage
TLC
  Competitor
TLC
 Flash elution
volume
RfCVDCVRfCVDCVCV
0.841.190.821.221.25
0.342.941.750.333.031.813.00
0.185.562.610.254.000.975.50
0.1010.004.440.0911.117.119.25

What this data shows is that the more polar the compound (the lower the Rf) the more likely that a difference in retention and selectivity will be encountered when silica chemistry differs.  And remember, low Rf values tend to provide better separations so small Rf differences can cause big column volume (CV) and ΔCV shifts.

In the example above, the ΔCV for the last two compounds varies significantly between the Biotage plate and another supplier’s plate. Purifying the sample in a flash cartridge packed with Biotage silica (Biotage® KP-Sil) provided elution column volumes with better correlation to the Biotage TLC plate than with the competition’s plate.

Yes, I know you probably have a common TLC plate supplier and perhaps you have not experienced this problem so why change brands?  Well that is certainly your choice, but for me, I would rather not chance a problem occurring because it is just a matter of time before you have issues which may lead to confusion during purification when transferring your TLC-based method to flash.

Have you ever experienced an issue like this?  Share your experiences.

Published by

Bob Bickler

Technical Specialist, Biotage

6 thoughts on “Understanding silica – why matching TLC and flash silica is important for good flash column chromatography”

  1. We use cheap TLC plates to estimate where to insert an isocratic hold on our default gradient. It is just an estimate and the peaks can elute faster or slower than predicted but is still worth doing and will be better than a simple gradient. I always worry that it’s not just different silica but also comparing a dry TLC plate with an equilibrated column. For important work we have steel analytical columns packed with the same silica as the prep columns. This way it’s not just the same silica but we compare equilibrated analytical with equilibrated prep columns. The analytical column is 15um and the prep is 15um or 30um. Particulary usefull for CH2Cl2 / MeOH mixtures.

    1. Hi Derek,

      Equilibrated vs. non-equilibrated TLC can have an impact though I have not noticed much difference. Your approach using silica packed HPLC columns using the same media as the flash column is indeed a good method development approach. I have found, though, that most labs use HPLC and Prep HPLC only in reversed-phase and that for flash either TLC is used or the chemist will guess the gradient based on intuition or history using similar compounds.

      Bob

  2. I am happy you decided to tackle this issue. First, there is an error in the table – the competitor DCV for the last compound should be 7.11, not 11.11. It doesn’t really change much and your view is valid but to a point. My main issue is how Rf are measured. Rarely you get nice, round tiny spots. More often you end up with blobs that can induce ambiguity when measuring their Rf. So the rather poor resolution you get from TLC translates into a large spread of CV. Having said that, you’re of course right pointing to differences in silica chemistry between vendors. But it seems to me it has a negligible impact on developing column separation. From my perspective, TLC is supposed to give an idea what sort of selectivity is to be expected and if you want to fine tune your method, you’re better off running a small column and then scaling up. With the same flash silica chemistry of course 🙂

    1. Hi Tomasz,

      Excellent observation on the DCV – many thanks for seeing the error. I have corrected it.
      While it is true TLC is a scouting tool for flash I have seen situations, especially in a customer’s lab with their sample, where they get a separation of three spots on their TLC plate but using a Biotage flash column we get four separated peaks. I do agree with you and Derek that if separation characterization is critical the use of a packed column using identical silica will provide the best information for transfer to flash. It really is the best way to develop reversed-phase flash methods as well.

      Bob

      1. 10 g SNAP columns can be used for method development. But they are slightly too big for that purpose. On the good side though, if you set your method right from the start you already have some pure compound. Consider making smaller cheap columns just for method development 🙂

        1. There is a 5g Zip column available containing the same KP-Sil media. Use of the optimise function on the Isolera works well for then scaling up to purify the bulk material.

Leave a Reply

Your email address will not be published. Required fields are marked *