What is the optimal sample to sorbent ratio for dry loading in flash column chromatography?

For chemists preferring or needing to dry load their crude sample mixtures to get an acceptable flash purification result, using the right ratio of sample to sorbent can be quite important.  Too much sample and solubility issues can ensue, too little sample and significant band broadening occurs, reducing the separation quality.

In this post, I propose an acceptable ratio range based on my own experimental data.

What is dry loading? For those who may be unfamiliar with the term, dry loading is a common flash chromatography loading technique used to facilitate purification of poorly soluble compounds and samples dissolved in chromatographically strong solvents.

For most chemists, liquid loading is used because it is quick (I often use it).  However, liquid loading can also be problematic.  If the dissolution solvent is too strong, it can interrupt the sample components’ affinity for the column’s media and distort what should be a good purification.

Dry loading eliminates these issues, at least in theory.  However, even with dry loading there are some caveats mostly centered around the sample/sorbent ratio. Let’s explore the following caveats.

  • Ensure the sorbent is inert towards the target compound
  • Maintain a suitable ratio of sample to sorbent

Silica is most commonly used as a sorbent but may not be the best choice due to chemical interactions and irreversible adsorption.  Other media such as diatomaceous earth Celite® (or Biotage® HM-N), Florisil®, and alumina are also useful.  Even reversed-phase media and ion exchangers can be used.

When using silica, Florisil, or alumina (all relatively high surface area media), I use a ratio no more concentrated than 1 part sample to 3 parts media (1:3) to 1 part sample to 4 parts media (1:4), by weight not including the solvent.

The reason I recommend this range is two-fold. 1) In my experience,  making the mixture more concentrated than 1:3 may result in solubility, leading to broad elution bands and poor resolution. 2) Alternatively, if I choose to make the sample less concentrated than this ratio, I have observed that the sample is spread over too much sorbent and band-broadening (peaks spread) results and often deteriorates my resolution. In addition, when using diatomaceous earth, which has lower density and surface area, I have found that a greater amount of media can and should be used.  For ion exchange media, using a slight excess based on equivalents is best to ensure full ionic binding.

As an example to support my proposal, the data in Figure 1 shows the impact that different sample/silica ratios have on separation quality. A 1:2 ratio actually reduces the amount of separation and a 1:20 ratio, while providing a better separation, also broadens the peaks.  At a 1:4 ratio, separation quality is obtained and peaks are more narrow.  For this work I used an isocratic hexane/ethyl acetate solvent system.

Figure 1. Chromatograms depicting the results of different sample/silica dry load ratios. Top – 1 part sample to 2 parts silica provides insufficient separation. Middle – 1 part sample to 20 parts silica provides a good separation but with broad elution bands. Bottom – 1 part sample to 4 parts silica provides an excellent separation with narrow elution bands.

These are my results. I am interested in your experiences as well.

I will address diatomaceous earth and ion exchangers as dry load sorbents in future posts.

Thank you for making this blog a success and please take a minute or two to take the survey so I can learn more about your blog improvement ideas and other topics about which you want more information.


Published by

Bob Bickler

Technical Specialist, Biotage

6 thoughts on “What is the optimal sample to sorbent ratio for dry loading in flash column chromatography?”

  1. I also like 1:4 for normal phase and 1:5 for reverse phase. It works well for most crude reaction mixtures but It’s flexible. I like to see a free flowing powder after evaporation, something which is fully porous to the mobile phase and doesn’t contain sticky impermeable lumps. Liquids seem to require less than solids, maybe 1:3. I also make a choice based on the amount of crude I have, the difficulty of the separation and the size of columns available.
    Routinely I use dry loading silica to account for 10% of the total silica, that’s 4.4g for a 40g column and 1.1g of crude. If it’s an easy separation then I might use up to 17g of silica (30% of total!) and 4.25g of crude. For a harder separation I would use less crude but not necessarily less dry load, it doesn’t seem to make a big difference.
    I find that if under flash conditons with 10% dry load, the peak elutes over say 2 column volumes (isocratic 50um silica), then with 30% dry load I guess the peak will elute over 2.2 column volumes, not a big difference. If it was hplc and peaks were narrow then I would keep the injection volume to a minimum.
    I’ve just bought some HM-N for dry loading after having read a previous post but will have to guess how much to use. I notice it is quite soft and easy to compress or maybe I shouldn’t squash it!

    1. Hi Derek,

      Excellent information, thank you for sharing. The 10% rule (dry load sorbent to column silica mass) is what I believe is optimal as well. The HM-N, as you know, is a type of diatomaceous earth. It is not as dense as silica and is compressible to a degree. I have used it successfully in our Samplet cartridges. As long as you can make a good liquid-tight seal in your dry load vessel without compressing the HM-N too much you should be fine.


Leave a Reply

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