Manufacturer "botanicals" recovery numbers

CannaBusiness
2

It depends mostly on the size and shape of your cannabis in whatever machine you’re using and it’s relation to solvent path.

I use a spray fuge with various alcohols. I think on a milled cannabis input (1/4" pieces or smaller) we average 92-95% efficiency in cannabinoid removal from biomass.

98% is doable, but I’d probably powder my cannabis to get that. But then I’m picking up way more junk, so more post processing required. Then in turn, we’ve got increases losses in post processing.

I’d say 90% start to finish from biomass to polished crude is a good goal?

5 Likes
3

If you’re doing solvent based extraction and not getting at least 95% cannabinoid recovery you’re doing it wrong.

Spray wash + virgin flush should get you well over that if your material is reasonably intelligently prepped.

3 Likes
4

You’re kidding right?

Why would you NOT test your spent biomass?!?

You’re using it as a washing machine?

Flooding and pretending that reversing every now and then equates to agitation?

I’ve certainly seen folks leaving 30% or more of their cannabinoids on the table playing it that way (that and not testing their spent biomass)

You running your solvent cold?

“Delta” probably wasn’t when they came up with that efficiency…

Have you read It washes the cannabis and then it washes the cannabis some more yet?

4 Likes
5

One of the minions got creative and made some changes to our extraction process without checking with me.

Then ran a bunch of our partner’s material that we took on a split.

Some of these changes were clever and we kept them.

But when we tested the resulting waste biomass, he had extracted only 70% of the cannabinoids.

So I changed a few things back and we got back to 95% after re-running that material.

Thankfully it was a small batch and only wasted a half day.

3 Likes
6

We are looking into a new novel solvent. Preliminary testing just at desktop scale showed the new solvent as far more efficient at cannabinoid removal than ethanol. We’ve moved to a larger scale test just using a reactor and stirring/agitation. We haven’t gotten those results back yet.
However, the PHD group is basing a lot of “potential” on these initial tests that are not overly reflective of a real world application. My argument is we should use at least (for now) for discussion, published claims on ethanol, and not these early initial experiment results which were done over a much longer timeframe than the 20-30 min times in industry.
Their argument is, that even though the experiment was done over a long timeframe (wash time), the efficiency difference should be valid. My argument is, that difference may not have been true at 20 minutes, 30 minutes, 25%/50% or 75% of the time the experiment took, but only peaking at 100% of the time. So they are discounting ethanols published efficiencies, yet taking theirs and applying it down to 30 minute washes. To settle this, ya we could run both in a full setup but I don’t want to spend $200k+ without some idea even using some real world numbers. We will have the novel solvent numbers from our test reactor setup fairly soon, and will move to ethanol to compare in the same environment, but it still isn’t completely indicitive of real-world. I’m just trying to get as much of an idea leading into deciding on expenditures to pursue this.

One of their other claims is, because the solvent is so much more efficient, you only need 60% of the actual volume of solvent compared to ethanol. Is it even possible to feed something like a cup30 that specs 100-110L of solvent, 60L and have it function properly or does it have to have that level of solvent? This would be a question more in general for all centrifuges.

7

Did you not read It washes the cannabis and then it washes the cannabis some more ?!?

~1l per lb of biomass.

Yes, there are better solvents.

1 Like
8

I’ll check it out.

1L/1lb for eth, i’m assuming because of ethanols efficiency? My q is is there any physical reason a centrifuge could not use less actual solvent.? IE I can have a motor oil that is twice as efficient at lubing motor parts, but I could not put 50% of the volume of what the motor spec’d and expect it to work.

9

Absolutely. But only if

Don’t do that…

10

We’re giving away some pretty solid tips here lol.

Build around your desired end product.

You can also extract everything warmer/room temp or higher, and get even more cannabinoids into your alcohols, but you just gotta do more cleanup afterwards.

There’s probably a million ways to skin this cat, and dozens of variables to fine tune.

Finding a “better” solvent is the least of your worries unless it’s producing an 80+% oleoresin without any thermal adjustments. Optimizing an extra 10-20% solvent capacity isn’t likely a goal worth chasing on scale. But that’s also highly dependent on your solvent removal process.

4 Likes
11

Our total solvent contact time is well under 5 minutes, and more than once we’ve seen >98% extraction yield of cannabinoids. We’re not using ethanol as part of our solvent matrix. And we’re definitely not using our first centrifugal stage as a washing machine - it’s a dryer, not a washer.

Our research suggests that in a beaker, with our virgin base solvent matrix, with dried biomass and CBD or THC inputs, 60 seconds with light agitation is sufficient for functionally complete extraction from the biomass into the solvent.

The important next trick is getting all of the cannabinoids off of the biomass, because there are usually cannabinoids in your solvent, and solvent on your biomass. You can leave 0% in the tricomes/biomass itself, but you will never* get your biomass 100% dry.

*for practical purposes in the real world

Food for thought, from someone who has to consider both science and production realities in the real world:

  • How long does it take your solvent to grab the cannabinoids when you’re using pure solvent?
  • What’s the maximum practical cannabinoid loading of your solvent?
  • How does the cannabinoid loading affect the extraction time?
  • Do you leave solvent on your biomass?
  • If so, how much?
  • What’s your solvent worth, in base solvent cost + cannabinoid content?
  • How much more solvent do you get back if you spin it at X speed for Y seconds?
  • What do those Y seconds cost you in lost production over the course of a day?
  • Should you spin for Z fewer seconds and allow a % increase in solvent losses to allow for N more production runs per day?

I’m just a dumb mechanical engineer, and I sure as shit don’t have a PhD, but my cup30 sized fuge is approaching 100kg/hr of dried biomass processing, and I’ll never let our gross extraction biomass-to-solvent metric slip below 95%.

@cyclopath has provided you a very helpful :spoon: or two. He’s one of the smarter cookies in this particular cookie jar, you might want to carefully consider what he’s said.

More than a few have gone down the route you have. It’s nice that you can afford PhD’s, but keeping them focused on production realities can be a chore.

My $0.02: hire a chemical process engineer or firm that’s built solvent extraction plants, ask them to tell you if and why you are barking up an incorrect tree, and then listen to them.

3 Likes
12

image
image480×640 122 KB

There are Cookies?!?

I’m just here cause someone said there was beer…

5 Likes
13

Just something to add.

Pending on where you live… Cannabinoids are going to become EVEN cheaper every year, and I highly doubt we’re even close to a commoditized price.

If your fancy new PHD solvent is quite expensive, leaving 2-5% of it behind in your biomass might eat up your margins pretty readily.

Using cheap, readily available solvents, that are already exposed to scaled processes; might be your only way to make $$ in the long run.

If your process can’t handle cannabinoids being worth under $100/kg and probably quite lower, it’s not likely an efficient process in the long run.

2 Likes
14

Other important factors:

Is it GRAS?

How hard is post-processing, including solvent recovery? Is a second solvent required for winterization?

15

This is part of the back and forth I have with them. It is more expensive, and I am assuming the loss per run is the same as ethanol ~3%. Where the cost is made up is the fact based on prelim testing that it is so much more efficient at removal than eth. So only 48% of equivalent volume is necessary to be used. Add to the fact that it can be kept pure, at full proof/strength with only 12L to be recovered per 8 hour shift of 3 runs per hour @ equivalent 100L solvent (based on cup30 spec ethanol) per 3 runs (re-using same solvent for 3 washes)… seems to work out on paper. But i’m not sure of the reality of a bunch of the assumptions. At 5% water content in biomass retention for eth, you hit 180proof at the end of day 6 having gained 13.7L of water per day, and still have 609L of eth/water mix (from a start of 1000L tote). I’ve heard rule of thumb was 185 proof before diminishing returns, and i’ve also heard extractors just running until it’s gone and maybe not caring how much their returns are dropping. Its is GRAS. I can’t say much on recovery parameters but uses same processes lower cost. Water retention not an issue. I’d have to model assuming the 3% loss is negated by drying for additional numbers.

1 Like
16

I’ve generally found proofing isn’t a big issue when needing to add solvent to make up for losses. I’ll generally end up with 91-93% ethanol averaged.

I use alkanes for live material almost always

Solvent removal is unlikely to be a bottleneck in process with current membrane tech.

But there’s some fine details to be worked out for sure.

The most beneficial solvent (in my opinion) would be the one that gets you closest to 80% cannabinoids in your oleoresin while allowing for efficient terpene preservation. But that’s assuming you’re looking out for Terps.

1 Like
17

sounds like you are asking all the right questions.

best of luck, and hope to hear about a breakthrough!

2 Likes
18

If you can make money at $50/kg, you’re probably ok in the medium to long term.

Say 10 kg per run. That’s 240kg per shift.

You’re planning to extract 240kg of biomass into 12L of solvent?

If you’re using more than that, how else are you getting the cannabinoids out of the solvent?

Depends on what you’re targeting. There can be damn good reasons to run substantially lower proof than that.

If your solvent recovery is required to be evaporative I would suggest stopping your research on it. Thermal separation processes will never be competitive in the long term with physical (membrane) separation processes. Don’t take my word for it, there’s an article in Nature that agrees. If there are membranes available, give’er.

19

No… based on delta numbers, you’d use 800L of ethanol per 8 hour shift using 100L for 3 batches of biomass. You’d gain 13.7L of water at 5% moisture.

This I can theoretically use 57L of this solvent for the same percent recovered cannabinoids per run. Over the day I will use far less volume. At the end of the day I have also gained 13.7L of water and used 456L of this solvent vs 800L of eth. As part of the recovery process I only need to distill 15.1L in this example to get the water out of it to bring the entire batch used back to full strength (with another costless step). I can’t go much more into those steps.

room temp, assume membranes.

20

Sounds like a single stage pervaporation process that’s used to remove the water, and a small fraction of the solvent comes along for the ride.

If you’re able to do room temp and membranes, it should work out to a first approximation. Solvent selection at that point is up to what you feel like using. Something that’s GRAS and doesn’t have any hazardous classifications would be ideal. If it’s hazardous and expensive, I would do a detailed cost analysis vs the standard off the shelf options that fit in the same buckets.

21

It is GRAS.