Data: potent & efficient lipid infusions via percolation (21 g batches)

[batsona]

Hey Folks

Wanted to share data I generated by percolating MCT through a few batches of Sweet Rainbow Hemp.

Over the past week or so I have been working with @TheKid420 on home counter-current lipid infusions. He is using coconut oil so heat is absolutely necessary. In my experiments here I simply used MCT at ambient temperature.

What I’m mainly trying to show is that you can create potent infusions by either percolation or maceration if you thoughtfully stage your batches and washes. On any scale–from home growers to farmers looking to process their own biomass into an effective (full plant synergy) and potent product. So I see building awareness of infusions, with data, as a way to also build interest in the equipment my company (https://www.hempharvestinnovations.com) builds. Those interested should also see my topic on economics of lipid infusion: The Economics of Lipid Infusion

Maceration turns in to a lot of work on small scales so I put together the following percolator. It is composed of two six inch spools with a mesh gasket screen and a valve at the bottom. I like to fill it up about halfway with moderately-packed biomass and leave the upper spool empty to pour carrier oil in to. Basically the goal of the experiment was to produce baseline data on how much carrier oil is needed to sufficiently extract a batch of flower in addition to how the vial potency can be stepped up by running subsequent batches.

For potency tests I use an hplc that I have calibrated for cbda and cbd. Potency of starting and spent biomass were Soxhlet extracted with ethanol to determine potency. I only did one hplc sample per vial sample so the reported values may have a few percent uncertainty.

Experiments:

1. freeze biomass at -30degC

2. hand grind biomass and filter through a ~2mm sieve

3. place a scouring pad filter at bottom of percolator

4. pack 21 g sieved material into percolator (batch 1)

5. place ~50 micron felt filter on top of packed biomass

6. fill six 60 mL vials with fresh MCT oil

7. start pouring oil into top of percolator and start timer

8. Collect the bottom effluent back into the six 60mL vials

9. label the most potent vial “1,” the next vial “2”, the next “3,” and so on

10. stop timer once the 5th vial is filled

11. squeeze remaining oil into 6th vial

12. record masses of oil in each vial

13. mix and sample each of the 6 vials

14. empty percolator by removing bottom triclamp

15. reload percolator with filtration and biomass (batch 2)

16. sequentially pour in vials 1, then 2, and so on into the percolator

15.(a) wait for the liquid level to drop below the felt to pour the next vial

16. collect the effluent with stepped-infusion back into vials 1 through 6.

17. repeat steps 9 through 16 to collect data and run batch 3

Results -

• The flower tested at 16.5% total potential (t.p.) cbd. I don’t know the cbd/cbda breakdown for sure because I think the Soxhlet causes a bit of decarb. From this potency, each of three 21 gram batches has 3465 mg t.p. cbd.

• My first batch of biomass came back 57.84 grams and the second was 57.91 grams. The third batch of biomass came back 48.46 grams because I let it sit overnight while draining in to vial 6.

• I only tested the first spent biomass potency at 8.68 mg t.p. cbd per gram, which, with 57.84 grams of spent material it means there remained 502 mg t.p. cbd in this batch. Thus from the input and output biomass potencies I infused (3465 - 502)/3465 x 100% = 86% of the available cbd. Not amazing but pretty good in my opinion… could do better perhaps with a finer biomass grind or using a little heat.

• On the infusion side of things, vials 1 through 6 contained, respectively (61.4, 24.3, 9.4, 2.9, 1.5, 0.5) % of the infused cbd. Their potencies were (84.8, 34.1, 12.9, 4.0, 2.1, and 1.6) mg t.p. cbd/g. Notably you see that the vast majority-95%-of the infused cbd ends up in the first 180 mL of collected infusion.

• When I look at the total potential cbd in all of the vials, at the end of the three batches, I count only 7800 mg t.p. cbd which together with the 3465x3 = 10396 mg t.p. input suggests an extraction efficiency of 75%. This efficiency is lower than the 85% I cited above but I also removed about 0.5 grams from each vial after each batch.

• Looking at the potencies of the vials as a function of the cumulative mass taken from the system shows that, in my opinion, I could have done this infusion with 3 vials.

• Really interesting to me is that the potency change after each batch is nearly the same for each vial. Looking closer, it seems to me that the most potent vials gained just a bit more cbd from batch 2 and even more from batch 3. To me this is a result of an increase of the infusion viscosity — or adherence to the biomass — because the flow rates of the more potent infusions slowed considerably for batches 2 and 3. So the simple gravity flow through the column gives some slight nonlinearity but it is mostly linear… from this data I can predict how vial potencies will increase for future tests.

• The useful implication of this linearity is that one can design a process that runs well more than 3 batches to efficiently produce oil at a given potency. That is, if your spec is 30 mg/g then you start taking oil out of the process, i.e. vial 1, once it exceeds the spec. (In my experiment you could take out vial 1 after every other batch since you know it gains ~25 mg/g in potency per batch. Then, vials 2 through 6 become the new vials 1 through 5 and you fill the last vial with fresh oil.)

Happy New Year!

figs:

percolator:

start biomass:

ground biomass:

mid percolation:

vials after second batch:

2020-12-30 percolation tests.xlsx (13.9 KB)

2020-12-30 percolation tests.pdf (40.0 KB)

[TheKid420]

@batsona

So the simple gravity flow through the column gives some slight nonlinearity but it is mostly linear

What about slowing the flow down by partially closing the ball valve?

[batsona]

totally! I don’t see much issue with closing the valve completely to give a longer steep, either.

[batsona]

This figure presents the data I posted above by showing the total amount of (total potential) cbd that has been infused into to the MCT vials after 3 batches of biomass:

I think this representation shows more clearly that batches two and three infused more cbd into the oil than the first batch. With each batch of biomass initially having 3465 mg total potential cbd the extraction efficiencies for batches 1, 2 and 3 are 71, 69 and 86%. Somewhat counter intuitive that batch 3 extracted most efficiently but the vial washes for this batch also had the longest residence times.

The data above can be fit nicely with the hyperbolic tangent function by adjusting two parameters (see the values of m1 and m2 for each batch). Certainly only a leading order model but the trend similarly follows what I saw (dashed green line) for infusing with our production scale equipment:

[batsona]

update! I ran 7 more minibatches of 21 grams biomass each for a total of 10 minibatches, 210 grams of biomass. Here’s a pic after the 9th minibatch:

Vial potencies are substantial–216 is the highest I’m aware of anyone getting with a lipid infusion–expressed in mg total potential cbd per gram of oil:

v1: 216.3
v2: 138.6
v3: 66.4
v4: 30.9
v5: 16.2
v6: 7.1

vial masses each have 55 to 56 grams of oil, the total potential grams cbd in each vial are

v1: 12.2
v2: 7.7
v3: 3.8
v4: 1.7
v5: 0.9
v6: 0.4

for a total of 26.7 grams of cbd in the vial oils. At 16.5% biomass, 210 grams is 34.7 grams, i.e. and infusion efficiency of 77%. Not bad for an ambient temperature oil extraction! My biomass potency could also be off, who knows. For the first minibatch I added 336 grams of MCT, then after each minibatch I added 40 grams for a total of 696 grams of MCT used.

Next step is to decarb the first vial, filter, and get a COA

[TheKid420]

You inspired me. I bought the pieces to build a percolator.

I’m building a stand from scrap I have laying around so it’ll be a bit rough but it’ll be functional.

Using your method I’ll try it but smaller scale. Say 4 mini batches @ 7g per. It seems you were using about a 16:1 ratio oil to biomass. I’ll have to look through your data more when it’s not so late but I think I’ll run 112mg oil initial batch and assuming same makeup oil ratio I’ll need about 13g oil.

I will only use 4 vials as the time/benefit looking at your data for me isn’t there for the last 2. It’s only a 1% loss in the first batch. Even if I only get 70% t.p. Cannabinoids but save 33% time that’s worth it in my senario.

But this is fantastic information. Thank you so much for sharing.

[batsona]

Sure thing! You’re right about cutting down the number of vials, especially if your desired potency is still down around 30 mg/g. For higher potency specs, it does help your efficiency to run higher liquid:solid ratio per minibatch because you have to build up a more substantial gradient across the vials.

If you’re doing 4 minibatches (with vials holding 1/3rd of the mass of oil I ran, because you are using 1/3rd of the biomass) then I expect you will be up around 100 mg/g in your most potent vial when it’s all said and done. I seem to believe your biomass is in the same ballpark of potency as what I ran.

[TheKid420]

I was misinformed about the capsule size initially and ideally want to use #2. Which is dramatic decrease in volume than initially planned.

90mg/g is desired spec.

#2 holds 0.36 ml, coconut oil 0.92 ml/g, desired dose 30mg.

(0.36*0.92)*30 = 90.5 mg/g.

Finished painting stand. Maybe tomorrow I’ll run a batch.

I’ll post an update when I do in my topic as not to detract from yours. Fantastic data. I appreciate your time.

[stoopkid]

Sounds like an MCT tincture with extra steps

[Capttripppp]

But… percolate…

[batsona]

Indeed! Making MCT tinctures all day. Though I would hope the data & method I presented would convince you that percolation is in fact quite simple. I guess, in other words, what I’m trying to show is how to effect a counter-current extraction in your kitchen. And we all know that counter-current processing is generally the most efficient means to drive transport of heat, mass, etc…

From this link:

Percolation tinctures are the main way that pharmacies and drug companies made herbal extracts (back before the 1940s when pharmacies and drug companies used herbal medicine). That’s because not only is it faster, it’s also stronger. In a maceration tincture all the different chemicals in the plant are competing for space in the tincture, and some are more soluble and some less. In a percolation tincture you have fresh solvent coming through the cone and picking up the most soluble constituents and dripping them out, then you have more fresh “blank” solvent coming through and getting the next most soluble ingredient, and so on. Therefore, a percolation extract is a more complete extraction technique.

[batsona]

Here are some pics of a decarbed (mostly) and filtered Sweet Rainbow infusion. Made from vials 2 through 4 above:

The taste resembles the strain well but also has a good bite to it which isn’t uncommon for MCT hemp infusions. Notably I have an infusion made with another strain that has hardly any bite to it at all, and the taste is great. Soon I want to send off for terpene COAs to see if there’s some kind of correlation…

[ky_cbd]

Direct lipid infusions are actually mct tinctures with less steps I mean it’s just like making weed butter but u use mct instead

[batsona]

I ran new tests to show a twist on the data I reported above: instead of endlessly stepping up the vial potencies, to generate the data below I removed a small fraction (26 grams) of the first bits of product oil leaving the column after each minibatch of biomass. This process approaches a steady state where both the product and vial-to-vial potencies are relatively constant after each batch. This product removal procedure effectively mirrors how my company’s larger scale (20 lbs per shift, 50 lbs per shift) equipment works.

To promote equilibration and uniform residence times from vial to vial and minibatch to minibatch, I now use a peristaltic pump to inject the vials into the bottom of the column. Specifically for the data below I rapidly pumped in a new vial, shut off the pump, closed the drain valve, and waited for 20 minutes before doing the same for the next vial. Here’s a pic from a previous olive oil infusion:

As before, the vial fractions are each about 57 grams. After filling the 5th vial, the pump is stopped, tubing disconnected, and the (weak) infusion remaining in the column is drained into a beaker. The potency of this beaker oil was not measured since it is really low but nevertheless it is pumped back into the column for subsequent minibatches (after all the vials have been run).

Here are the product fraction & vial potencies from a test of 8 minibatches:

By minibatch 2, the product oil saturates to about 60 mg cbd/g oil. The vial potencies however keep increasing little by little pretty much up until the last minibatch.

The significant jump in product and vial 1 potencies after the 6th minibatch coincided with refilling my 1L beaker with biomass. I thought I took biomass from the same container I used for minibatches 1-5 but maybe I screwed up? Whatever the case it appears the input bio potency increased.

In total I measured 186 grams of product oil after filtration:

Averaging the potency data for the product oil, those 186 grams ought be around 62 mg cbd/g. Thus after the 8th minibatch the product oil contains 11600 mg cbd and vials 1 through 5 contain {2400, 730, 520, 280, 220} mg cbd. Likewise, about 73% of the infused cbd is in the product oil, 88% is in product oil + vial 1, and 93% is in product oil + vial 1 + vial 2. If I were to run more minibatches, the amount in the product oil relative to all of the vials will only keep increasing.

All told, I used ~790 grams of MCT to process 168 grams of biomass. Given how low potency vials 3, 4 and 5 are… one could see similar results by omitting vials 3 4 and 5. Only notion you have to keep in mind is that more vials are generally needed for larger target potencies. I.e., if I only took out 15 grams of product oil after each minibatch, the potencies associated with all the vials would increase accordingly, and efficiency will decrease if, say, vials 4 and 5 are omitted.