CBD to THC Conversion - Patent Linked

Hey All,

I spend a decent amount of time lurking around the forum and I contribute where I can.

I commented on a post and got my ass handed to me from a chemistry perspective, so I figured I’d better dive deeper into the subject.

I see all kinds of people attempting to convert CBD to THC with very, VERY strong acids—often with little regard for relative weights of cannabinoid/solvent/reagent and often disregarding inert atmospheres or residual moisture in solvent. For these reasons, and because I happened to be down a rabbit hole on the topic, I figured I’d link the patent for the synthesis of Dronabinol, which is a D9/D8 compound that is quite pure. Typically 97%+ D9 with residual D8.

The really interesting thing about Dronabionol, is that the people who have created this drug found a way to convert CBD to THC with extremely high selectivity and great yields—and it seems extremely simple, relative to how it’s been presented in the forum.

I’ll summarize the steps below, but it is REALLY worth the time to read if you’re interested in pursuing this process. I haven’t done it myself, but now I’ve got to try.

  1. Mix anhydrous solvent with CBD at a 1:1 ratio. There are options for solvent choice and desiccant that can be used to make the solvent anhydrous. Typically something like pentane, hexane, or heptane dried with sodium sulfate.

  2. Optional—In a ratio of 3:1 Acid:Solvent or Acid:CBD, add Lewis Acid. This isn’t necessary but serves to improve yield and selectivity.

  3. Inert the reflux atmosphere. This is critical. The author of the patent argues that residual H2O will reduce the selectivity to D9 and result in higher percentages of D8 in the final product.

  4. Reflux with 4A zeolite molecular sieve. This blew my mind. The author argues that the molecular sieve is what is really allowing the cyclization to occur in the first place. The Lewis acid just potentiates that effect.

That’s the basics. Additionally, the manufacturers found that mixing the formulation with sesame oil was enough to prevent oxidation at room temperature for at least 3 months.

Anyways, I figured it was worth a share. I thought it was cool as shit and I don’t remember seeing this in other conversion threads. If this is just a repeat, someone toss it in the echo chamber :grimacing:


Is it possible that you misread? I don’t see a statement in the patent that mol sieves serve as the primary cyclization catalyst, and conversion is very poor without a Lewis acid. They even only list reaction examples where quite a lot of Lewis acid is used. It has long been known that the exclusion of water and oxygen is necessary for high yields of D9.

It looks like they just needed some creative wording that made it seem like they discovered something new, which there isn’t really.

Overall, if you look at the numbers, the bump in yield is only achieved with huge excesses of mol sieves and the yields aren’t that much higher than with BF3 for example, like Mechoulam did it many decades ago getting roughly 60% D9.


Literally in the introduction.

Edit: they spend basically the entire patent explaining what the purpose and method of using molecular sieves is.

Basically, a 1:1 ratio, by weight, AS MENTIONED previously is what is ideal. So NO there is not an excess of mol sieves needed.

I’d ask that if you’re just going to put me on blast you actually read what is linked instead of guessing at the content of a patent, you’ve essentially told me that I’m wrong about things that are clearly listed on the patent.

So…what’s that about “60%”???

Also…”something new”??? What the fuck are you talking about? It was filed in 2005. Additionally, it’s been sold in Germany since 1998.

Are you trolling me today or something? @y_tho

If you’re gonna act like you’ve got some answers…at least READ the publication you’re trying to refute. I’m not gonna debase myself calling you names, this is just frustratingly bad science. When people make wild conjecture it hurts the conversation in this community—the very same type of conjecture that lead me to spend hours researching this last night.


Had a client hit low 70% range d9 conversion on their first run ever from my sop on cbd-d9 :call_me_hand:t3::facepunch:t3: Great link & the use of 4a beads is one of the many parts used in higher yeilds :wink::smirk:


This is legitimate criticism, so i’d like to ask you to adjust your tone when addressing me as i find it utterly disrespectful.

There is not a single example in the patent giving high yield, selectivity or even just conversion using mol sieves alone, only in combination with ZnBr2, so addition of the latter surely is not “optional” for achieving barely acceptable yields.

In the second screenshot you posted, “95%” refers to the specification for the assay of Dronabinol in the respective monograph and has nothing to do with the yield of the isomerization whatsoever.

I’m not refuting the publication at all, just correcting your statements, which are misleading or incorrect:

There is no such statement anywhere in the patent. And of course not, because it is not true. With Lewis acids you’ll get roughly 60% D9 while with mol sieves it goes up to around 75%. Still not stellar and mostly attributable to the effective scavenging of water.

PS: Dronabinol is the INN of D9


Cause no one speaks on the 4a beads

Maybe you misread? Just as you DIDNT read the links or anything I sent you when you dm’ed me.

So if you’re gonna come then please do so correctly.


Why 4a zeolite? Doesn’t 3a remove water?

1 Like

It does but doesn’t allow the seperation of the compound it’ll hold onto the surface of the 3a

What do you mean by “separation of the compound”?

1 Like

Once done with your rxn, using the beads in the bf will control acidic water from distilling off with the compounds you’re wanting to pull off being d9. I also advise a packed head


There seems to be a few posts here, and a few people that are very adamant that zeolite reacts with our catalyst; potentially producing side reactants.

What’s your chromatograph look like? Any unknown peaks?

1 Like

The last one I did consulting a client I’ll ask if I can show, but no mystery peaks that I’m aware of they hit 89 tac with 17%cbd left to be consumed still. 30-45 more minutes he’s got a full conversion.

I just looked at the chroma, not mystery’s at all(not even the tiniest peak in between anything), only 17%cbd 72%d9

How do you get a 100% cbd, to 89% tac, without mystery peaks…

What’s the 11%? Unknown but not mysterious? :stuck_out_tongue:

Or is the 11% d8?


It doesn’t account the 11% on the chroma, but the standard is on there to show baseline for it. I believe it was a disty run they did not iso

1 Like

“The inventors found out that, surprisingly, the Δ9-THC yield considerably increases in relation to the formed Δ8-THC, if the Lewis acid is replaced by a molecular sieve.

Cyclization is considerably speeded up by the use of a molecular sieve, as described in the present invention. It is possible that the molecular sieve serves as a catalyst in this process.”

Pretty sure you didn’t read it because this is specifically stating that it appears that the sieve beads are possibly the catalyst in the process.


Read again:

“the Δ9-THC yield considerably increases in relation to the formed Δ8-THC, if the Lewis acid is replaced by a molecular sieve.”

If you replace something it means you have removed it and used something else in its stead, does it not? So I can safely assume that they directly mean that the Lewis acid is not necessary with the appropriate amount of sieve beads.


Literally the first two examples demonstrate that 1) The overall yield is 3x higher when using molecular sieves alone vs zinc (II) bromide alone. 2) The selectivity between isomers is 6.6:1 and 3.2:1 in the same respective situation as 1.

Example 1
If CBD is reacted to Δ9-THC without zinc(II) bromide as Lewis acid and with a molecular sieve,

  • The percentages Δ9-THC (23.0%) and Δ8-THC/iso-THC (3.5%)** are determined by means of HPLC analysis.

  • The ratio of Δ9-THC to Δ8-THC/iso-THC is 6.6:1.

Example 2
If CBD is reacted to Δ9-THC with zinc(II) bromide as Lewis acid and without molecular sieve,

  • The percentages of Δ9-THC (approx. 6.5%) and Δ8-THC/iso-THC (approx. 2.0%) are determined by means of HPLC analysis.

  • The ratio of Δ9-THC to Δ8-THC/iso-THC is 3.2:1.

From the patent: “Surprisingly, the inventors of the present invention found out that the use of a molecular sieve in a method for the preparation of dronabinol (Δ9-THC) by cyclization of cannabidiol (CBD) in an organic solvent had a strong impact on the selectivity regarding the Δ9-THC/Δ8-THC-ratio as well as on the reaction rate…The inventors found out that, surprisingly, the Δ9-THC yield considerably increases in relation to the formed Δ8-THC, if the Lewis acid is replaced by a molecular sieve.

From the patent: “The use of a molecular sieve results in an increase in the Δ9-THC yield, in a considerable shift of the relation of Δ9-THC to Δ8-THC/iso-THC in favour of the Δ9-THC, in a considerable decrease of the sum of by-products, and in a reduction of by-products resulting from the degradation process of Δ9-THC because of thermal load.”

From the patent: “The weight ratio of…(CBD)/molecular sieve is preferably 5:1 to 1:5, more preferably 1:1.”

I didn’t argue anything about 95% yield–I said:

I also said:

Which is TRUE!!! That’s the reason they filed for a patent!

Just because conversion selectivity and yield are high, does not mean that the product won’t require downstream purification…This process in and of itself is not going to produce 97% Δ9-THC. But that’s not the point, and I wasn’t arguing it.

In fact, the patent goes on to clarify: “Upon the successful completion of the cyclization of cannabidiol (CBD) to dronabinol (Δ9-tetrahydrocannabinol (Δ9-THC)) and the side-product Δ8-THC, the desired Δ9-THC may be purified from the mixture of Δ9-THC/Δ8-THC, for example, by means of a preparative HPLC column with modified silica gel, for example type C8 or type C18.



I’m not quite sure why its 4A specifically, the patent just says that zeolite molecular sieves are ideal…

Here are some direct quotes:

  • “It is possible that the molecular sieve serves as a catalyst in this process.”

  • “For the use in the framework of the present invention, the molecular sieve preferably exhibits a pore size of 0.2-1 nm, preferably of 0.4 nm.”


Does it state what acid was used to impregnate the zeolite?

I’m assuming its the acid sites on the media thats allowing cyclization to happen.

1 Like