DCVC high column load run on crude


This started as dark brown concentrate testing 65% nominal cannabinoid. The crude was dissolved into hexane and then wet loaded on top of a used column. I load it right on top of the Celite 545 layer (Amazon) from a dry loaded run. The Celite layer works GREAT when wet loading a bunch. This run started at one ounce. The column handled it with zero problems. This run was not dewaxed or processed at all prior to this first run.

The solvent system is generally Ethyl Acetate and Hexane starting with hexane alone. Then the Ethyl Acetate is added in 5% increments and is collected 20ml fractions at a time. Once the gradient is over about 30% Ethyl Acetate to hexane the majority of wanted fractions have eluted. Then I add Ethyl Acetate alone along with methanol to move the waxes and chlorophyls and just general gunk off the column. The cannabinoid fractions can be identified visually. They always elute after the highly colored bright yellow (or bright orange according to starting strain) fractions and at about fraction six the cannabinoids are seen in the test tubes as the brown fractions between the bright yellow first fractions and the green and later fractions.

Here is the starting crude dissolving in hexane.

Here is the crude now wet loaded onto a previously used column. The compound is loaded on top of Celite used in a previous lab as a dry loading agent. Wet loading allows much larger quantities to run. Dry loading can resolve close fractions better imo. The Celite layer is on top of the standard chromatography silica gel 60 A pore, 35-75 micron from Analtech (Amazon).

Here is the column now just as the cannabinoids are getting ready to elute and are followed above them by the green chlorophyls. The green chlorophyls always elute well after the cannabinoid and take much more polar solvent to move off the column. The dark browns seen above the chlorophyls represent the tough to move waxes and such which only will flush out with methanol. Methanol is not used at all to move the cannabinoid off the column.

Now the chlorophyls are ready to come off.

Now finally the last of the waxes and gunk following chlorophyls and eluting with the use of methanol.

Here is the wrap up. Fractions collected top to bottom left to right. Larger catch jar was used to catch the waxes and methanol column flush.

The brown cannabinoid fractions will be combined and purged of solvent. Then I will proceed with molecular distillation. As a starting point removing the bulk of unwanted compounds is a breeze with a wet loaded column.

Cannabis Alchemy Extraction Method vs. Modern Etoh Extraction Methods
Reading test results

Hi Beaker, thank you for your posts, as always they are very informative.

I haven’t ran any chromatography myself, but that DCVC video you uploaded is really excellent and I will surely have to give it a go at some point.

Is the brown in the cannabinoid fractions some of the thermal decomposition products you mentioned in your other thread? Have you ever evaporated off and used this fraction without distillation? Also have you ever ran a DCVC on unactivated crude?

With all of the pigments etc removed it seems that many of the undesirable flavor / health components gone, and DCVC so much easier (RE: cheaper), I am wondering about the quality of your finish material here without further refinement.



This crude came to me without any heat treatment at all. This was raw crude and this was the first process. I used to often only do DCVC before I had a good set up for distillation. I have evaporated and vaped portions of every color over time. There is nothing of real interest outside the brown fraction. I do not know why the cannabinoid fraction looks brown.

The end product is excellent when vaped without further processing except purging solvent. The entire flavor profile is still present but toned down and much smoother. The colors seen here are not thermal breakdown because it never got heated to this point. The difference here between vaping the raw but very good concentrate and vaping it after a DCVC run is the difference between hamburger and T-bone steak. Done right the DCVC run can turn ho hum but good dabs into “wow” dabs. It really is worth the effort and is VERY fast compared to say short pathing. Plus with DCVC you can enjoy a product that has never been boiled or really heated beyond about boiling temp of water for solvent purge. THCa remains intact.

Multiple runs through a column will also improve quality however there is a caveat with DCVC if you plan multiple runs. I use a solvent gradient. This means the silica gel starts out soaked with hexane alone. Think of a packed column like a very tight sponge in a way. It gets soaked by the solvent passing by. I add ethyl acetate in steps making the gel soaked in ever more polar solvent. Finally to flush the column it really takes methanol along with ethyl acetate. This means that the gel is soaked with highly polar solvent at the end of each run. So if I want to run the product again then I have to replace the gel. The reason is that it is nearly impossible to completely restore that used silica gel back to its original state of being soaked in only non polar solvent (hexane) as a starting point. Even slight traces of polar solvent left over will throw it off although with DCVC this is an issue only for the first few fractions until the gradient caught up to the residual solvent (messy).

It takes more money in solvent flushes to restore the column to starting polarity than just packing a new column. So multiple runs while simple do eat up resources. The solvents cannot really be used again either except repurposed as cleaners owing to azeotropes that make solvent purification impractical.


Great work!
How necessary is the solvent gradient? Can the column just be flushed with a large quantity of heptane (or hexane) to recover cannabinoids or is a gradient necessary to elute them?

Also, dumb question here, what is the purpose of the celite? is there a specific ratio of celite:silica that should be used? (If i had a 2L column, how much of each should I use?)

Finally, if you know… does pH impact this process? I know sometimes silica is rinsed with a buffer solution prior to use.



What is shown here is a column that had already been used once. When loading a column like this with compound you can wet load it or dry load it. The first use of this column was as a dry loaded column. This means I dissolved the compound first in a solvent and then soaked up,the solvent with Celite 545 (Sigma Aldrich ordered from Amazon) which is also used in industry as shop absorbant. The Celite Soaks up the solvent with the compound which is then dried onto the Celite. The dried Celite powder then is loaded on top of the silica gel.

What dry loading does is present the absolute most uniform loading of the compound onto the column that is possible. The Celite has no real affinity for anything in these extracts so it releases the compound when the solvent begins to flow. Uniformity of loading and packing is everything for good results.

For those pictures I simply flushed the column and reused it. The Celite is this case serves to protect the silica gel from churning up. For wet loading a sand bed could also have been used and is typical however the column was already packed. The Celite layer made a good wet loading buffer and for a crude run wet loading is as a practical matter as good as dry loading. Wet loading allows far more compound to be loaded but is much trickier to get to run properly at first if you are trying for very fine seperations.

The difference between a solvent gradient and one that is a single solvent is huge with DCVC. Without a gradient the solvent system is called isocratic. A DCVC column of course can be run with an isolcratic solvent system. However this is not the power of this simple technique.

The idea put simply is this; the silica gel is packed very tight. The compound represents a bad stain on top of the gel. With standard gel like I use Hexane is a pretty weak solvent to wash the stain off the gel. It is so weak that many parts of the stain simply do not wash off or through and just stick. Some parts like terpenes are not so sticky so hexane will wash those through.

If you just kept dumping hexane in it is not strong enough to wash most of the stain through. Hexane is called a non polar solvent and is weak on standard gel as a washing agent. So then a bit at a time a much stronger chemical capable of washing things off the gel is introduced a little at a time into the weak hexane. Bit by bit as the solvent gets stronger (more polar) more and more parts of the compound begin to get washed through and out (elute). This keeps going with a gradient run until all compounds have eluted.

Some solvent combos during the run are almost strong enough to wash out the stain but not enough to do it all the way. So that part of the compound moves further down the column but not out. Then as the gradient becomes stronger with each fraction it moves faster and faster until it elutes.

Your question about using an isolcratic column is a good one. Here is the problem with it. You need to make the solvent just strong enough to barely move the slowest parts but easily move the faster parts. If you just dumped pure polar solvent in then all would elute at once and no seperation. So what happens is that you end up never moving the slow compound very fast at all. Worse in order to get the slow compounds to move then likely the fast compounds through the column are closely grouped together unlike the slow ones spread far apart.

The downside of course to an isocratic column here is that orders of magnitude more solvent are used to completely flush the slow compounds through. DCVC is particularly easy to establish a precise gradient because the column is vacuumed dryish with each solvent gradient used. Initial banding if any is simply erased as the gradient catches up. The internet “father” of DCVC out of Copenhagen University even claims that DCVC as he and I practice it provides more resolution than Thin Layer Chromatography, noting that at times a TLC run will show one compound but a DCVC run often reveals a second. This is an extremely powerful tool in this respect.

Ph modifiers are not used by me. I have no experience with them. I would first have to identify the theoretical reason and working theory as to their use. I have gotten proficient at DCVC but am a long way off from perfecting the system I use. Each run I learn a bit but it will take another year or so to get it nailed. My plan is to stick with my solvent system as is and make adjustments there. I need to slow the ramp up of polar solvent a bit for instance but it will take a few more labs to figure out just where.


Hi Beaker. Question about cleaning a used column. What do you think about a heavy acetone wash. I just tried it, and it looked like it washed the column pretty well, but I still do have some bright colored impurities at the top of the column that I just cant seem to move. I have never even tried to reuse columns when I was in industry because I didn’t have to worry about the cost of Silica, but now that I am paying, I want to save every penny! I would like to get several runs out of this now dedicated “first pass roughing column.”


Very interesting post. The results look great and it’s such a simple method.

Any thoughts of manipulating temps/pressures to make n-tane a viable non-polar solvent for use in this method? Instead of pulling solvent through with vac it could be pushed through with positive pressure to keep butane as a liquid. Would be fairly easy to put together a setup that would run like this using n-tane as an isocratic solvent but figuring out a way to pass gradients would be tricky. As well as figuring out a polar solvent that will be miscible with butane at low temps. Gets me thinking though…

Thanks for sharing this great info


The most thorough flush is the most polar flush that is not water - a methanol flush. Silica gel beads will form a colloidal after about 10% methanol or so and be able to pass the glass frit a bit but that is all junk anyway.

However it is harder to reset the column to a non polar phase after flushing it with methanol. Hexane and methanol are immiscible. Flushing methanol out in favor of hexane would present unique problems because of this mehtinks. So I doubt I would reuse a column I ran crude through because for sure there are gobs of gunk that needs a methanol flush really and normally I do not flush it then. I just toss it.

So the trick to reuse for me is to reuse a column only after a seperation like I do between first and second distillation runs. A DCVC run after distillation will seperate just trace thermal degredation products which elute well in front of cannabinoids typically. So the column needs no really polar flush like methanol in that case and is good to reuse. In order to restore a lightly used colum to its non polar state I currently do a soak through run of hexane. Gravity feed only through the column. At least two complete columns worth by gravity (about 200ml). For sure you will see colored traces elute from the previous run but after about 100ml of flush it should clear. The idea is not to flush impurities so much at all. The idea is to give the gel the maximum soak time to pull all Ethyle Acetate (the polar solvent) out of the gel by allowing a lot of time for it to soak out by swamping it with hexane.

Then I pull the column dry under vacuum for a minimum of 30 minutes or so. The idea behind leaving the vacuum running is to evaporate much of the solvent out of the gel. There will be at least a trace of EA polarity left in the gel even then but this is negated in DCVC after the first gradient increase so in a way is even a bit helpful. A trace of polar solvent gets the first fraction moving in a better and more even line down the column. DCVC is incredibly forgiving but the column must begin in a highly non polar state which for me is hexane.


I believe the patent that deals with this idea incorporated aluminum oxide as the static phase instead of gel and incorporated three columns inside the supercritical chambers which the solvent would flow through. With each successive chamber conditions made such that each successive column refined things further. In theory then once the oil gets puked out of the supercritical fluid extractor it is near perfect purity. I understand where the idea is going with it but… I do not think anyone has implemented this idea in a machine yet.

In theory the practice would work but in practice the theory might not. I can tailor my DCVC runs according to what I see going on plus with a vacuum arrangement once the column is packed it won’t channel (a failure mode). The packing is simple but even better once shown how to pack a DCVC column correctly it is visually verifiable that it is correct. I hope to do a short vid to show how fast and easy it is to pack a boss cool tight column in just a few minutes loading everything dry first.

If that was to be done inside a SFE device the complexity of course increase. No sin there of course but I was an engineer for a long time on machinery and my instincts tell me that more and more complexity is generally something to avoid. German engineers love to make things complex. I perceive an inline having enormous problems with fouling a column and even channeling without a way for the operator to detect or correct. I also suspect the patent author had not ever actually seperated crude himself in any fashion so is working strictly from published data on the compound which really leaves off just how messy this stuff can get.


It is interesting to ponder the possibility of an inline chromatography column being incorporated into an extractor design but I do agree that the engineering required would be rather in depth and there would be many many complications on the way.

Seeing as the method you have provided is so simple and effective on its own there seems to be no need to incorporate this seperation into the initial extraction process.

The main reason I think about ways to use butane over hexane or heptane as a solvent is that for small scale extractors/hobbyists the starting crude is usually a butane based liquor, and I like the idea of minimising the number of solvents involved in the refining process. Although anyone using this method to prepare crude for a distillation would have no worries of residual solvents due to the nature of the distillation process.


In HPLC (analytical) I used pH modifiers to be sure the analyte was far enough from the pKa to be either fully protonated or deprotonated, not 50/50 (like at the pKa). But that was reverse phase, and I’m not really a chemist, so who knows.

Also in HPLC we use all different kinds of bonded silicas. Maybe changing up the packing will allow to run isocratic and recover solvent?

From another post, I saw the pKa of carboxylic acids might be around 4, so a slight modifier might be useful. Just throwing you ideas. Good work beaker! I wish I could use alkanes sometimes.


Generally speaking as a preparatory stage the Chromatography as I use it is almost used as a disposable filter. I think the gel costs roughly about two bucks an inch in my standard size column so,most columns have about $7 of gel. The hexane after shipping is over $65 gallon. The other chemicals are less but do cost of course. Saving the gel by flushing for me to do one more run is really more a convenience because the column is packed already and if it can be reset close to non polar for a rough run then great.

However I am purposely loading these columns likely far beyond what a typical analytical application would do. In other words it is my intent to use them more as a filtration than pure seperation because I only seek a single component and have no desire to quantify presicely the make up of compound I just want them seperated. So changing the packing is very attractive from a learning point of view but anything not standard silica gel gets spendy. It would then of course be cost effective to flush the spendy because solvents are cheaper but cost effective means spendy use of more solvents lolz.

Even an isocratic column, meaning same solvent system (not single solvent necessarily) is extremely difficulty for solvent recovery except solvent is recovered for repurposed cleaning use. Nearly all of the solvent systems I know of will form azeotropes with one another. Even Hexane and methanol which are not miscible in each other form an azeotrope if you try to fractional distill them pure!

Certain azeotropes can be broken, like the methanol and hexane one can be done by simply adding water to the mix. The methanol and water are miscible and mix and then can be seperated this way but generally speaking just look at a chart of azeotrope with hexane and all the rest. Lolz it is hella tough to return these solvents to their pure form for use in Chromatography.


Great info, that makes a lot of sense. Damn azeotropes. Ethanol/water especially.


Yup, I love the idea. its totally possible on different fronts too. from in-process solid phase extraction, to in–line butane flash chrom. Lots of variales to play with here, but with proper equipment I think its worth exploring


It seems like it would be an interesting experiment to add an inline sight glass filled with silica to a cls and to force a butane rich runny extract through it with n2 to force through. Collecting the different fractions in their own containers would be a pain but not impossible if you have the parts to play with.

Alternatively dumping a similar butane rich extract into a chamber with the celite substrate in it and then removing the butane would give the “loaded” substrate for dry loading a DCVC column. From there it could be loaded straight onto a silica column and run with the solvent system suggested in the SOP above.


What kind of solvent gradient would you use? Could butane-propane work or do you need something more polar? Isocratic would be better for recovering solvent, but a gradient should make collecting fractions way easier.


My understanding is that both propane and butane are non polar hydrocarbons and hence any combination of the two should also be non polar and have a very weak ability to elute the target molecules from the silica substrate. There would need to be a polar solvent mixed with the prop/butane solvent system to increasce the polarity of the solvent system. Due to the volatility of prop/butane it would be tricky to make the several different gradients of solvent required to elute the targets as they would each need to be stored in their own pressure vessels. The complexity required to pull that off makes me think it may not be worth the effort.

However if you could find a chromatography substrate that is less polar than the standard silica gel then perhaps flash chromatography with an isocratic non-polar solvent system (prop/butane) could provide good resolution of our target molecules.


I was thinking 2 separate tanks, manually varying the ratio with valves to make a sort of of gradient.


I have been thinking about this a bit over the past few days and i think that you are onto something with the two tanks idea. And as far as the solvents for making the gradient I think that acetone or ethanol could provide the polarity required to elute the targets off the stationary phase. I am not sure about the miscibility or any potential reaction that may occur between propane/butane and acetone or ethanol. Assuming they are miscible and wont react upon mixing, having a tank of your hydrocarbon extraction solvent and another tank of the miscible more polar solvent should work. As long as you have a way to measure the volume of solvent you introduce from each of these tanks into a 3rd mixing tank, you can vary the ratios to create different polarity solvent systems. From this 3rd mixing tank the solvent could then be dumped through an inline sight glass packed with silica to elute the target molecules. The column would however need to be loaded with the initial hydrocarbon extraction solution prior to passing the solvent gradients through the column to ensure there is an even loading of the column.


Yeah me too. The different alkanes do have slightly different dielectric constants, which I think means different degrees of non-polar pretty much, so maybe there is some room for fractionation there. Tane and ethanol is probably faster and easier to separate back out at the end. I wish I could try this out. Float type flow meters to give you approx flow rate and fine needle valves. Pure tane in 1, EtOH pressurized with tane in the other. Increase the ethanol as you go. Maybe a detector or sight glass to read fractions coming off the column as you go.