Good point about the possible non 100% reduction.

I think to get me a TLC kit and see if that matches the LC/MS findings somehow.
Then also check if non pH-depression circumstances (like using a strong base as lye) would answer more to the good ole pH 8.75 separation rule (pKa inspired separation). If so, that would be a good thing to discuss with VDS.

All in all I'm so glad he did his works and got to know it !!!

Must study the results further, time management

OK guys, I was wrong: microscopy CERTAINLY CAN give a good idea about the composition and purity of our substances. I decided to check again on some of the samples and was shocked . The mostly uniform microscopic observations of a few months ago gave way to atypical combinations and crystal forms. That is some bad news because it might indicate that our samples deteriorate on standing (I already observed this in THH-solutions that quite quickly changed their UV-color from soft-blue to green. Addition of a little zinc changed the color back again). Whatever the cause of my observation (perhaps I only 'wanted' to see uniform crystals? although posts #106, 135, 160, 228, 260,... and my memory tell me otherwise), the GOOD NEWS is that there is a very strong CORRELATION BETWEEN MICROSCOPICAL APPEARANCE AND COMPOSITION/PURITY of the samples. In short: pure harmine is expected to give isolated or loosely connected long needles, DHH is expected to give isolated rhomboïdal/fern-like crystals and THH is expected to give smaller, isolated spherical crystals with a 'target'-appearance. These crystal-forms should be prepared in the following standardised manner: dissolve a speck of sample in a drop of vinegar on a microscopic slide, add a drop of strong ammonia, wait for 5 minutes and observe. No cover slip is necessary and might inhibit larger 3D-crystal formations from forming. With this information, I invite you to have a look at the pictures I made from the samples as I took them out of the freezer today. I must add that all samples except sample 7 were stored at room temperature, not airtight. Only when sample 7 was made, all other samples were put together with it in the freezer at -20°C, stored airtight. Possibly, these weeks of room-temperature storage may have degraded the samples or changed their compositions. I do not know this, but plan to leave sample 7 at room temperature and check on its microscopic appearance in a few weeks. It would also mean the samples might have been purer right after their preparation.
I did not shrink the images before uploading, as they might come in handy for future reference.

All ideas are welcome!

Another wonderful blast in this thread

I first used to drop the vinegar+harmala on the glass on the desk surface. When crystals developed then carefully carrying it under the lens. I've found the least of movements were enough to disturb the crystals in the drop, not overly but noticeable.
Then I also did not want to drop ammonia right under the lens.
Now I do it with the glass already on the microscope plateau on the edge, so not under the lens. The glass is now already on the right level. Then when crystals are developed (5 mins) then just slide the glass an inch to just under the lens (there is a lot less ammonia to notice by now). This slow sliding on the plateau itself creates almost no movement in the drop, giving the nicest results. Just saying for anyone in the mood to dig up a microscope.

The degradation of thh is another mystery but with the tools available now it should become determinable. Al it takes is time. Ewww time
For preserving, salt form does wink wink nudge nudge.

An1cca, you rule. that was frigging awesome. you single handedly advanced harmala research by miles. please write paper(s) or even a book on the topic.

Yes An1cca rules and the pics are also getting better
I'm glad there's a consent on ammonia for easiest workflow, and the nice numerical results are the cherry on the cake.

For a potential tek I suggest an ammonia base only tek, and then later thereafter maybe a separate tek for working with carbonates. So not a tek that combines both possible routes in one document.

A tek could start off with X gr seeds (i.e. 100), or start off with Y gr mixed harmalas (i.e 5). But there are many ways to extract the seeds and maybe we should let people choose their own go on it: boils, cold soaks, freez/thaws, whole seeds or ground, ... as long as they incorporated 5 manskes before their final FB + washing, this as a condition, I suppose they are set. So starting from Y gr of mixed harmala extract looks appealing to me. Like a separate module that emphasizes on separation + reduction.

Ideas/thoughts/remarks about this?

I think our colloborative research project rules, including the work of all people here .


I totally agree. Ammonia is pretty foolproof and the less complicated a tek, the better.



I totally agree. That's what this thread is talking about anyway.

What I would be interested in, looking at the current GC+LC/MS results, is how a freshly prepared (separated using ammonia and the pH-metric method) sample of harmine and DHH looks like under the microscope. In my samples, I see that in some regions, crystal form is chracteristic of the pure alkaloid, while in other regions on the same slide, contaminants and aberrant structures are observed. So, when using microscopy for purity estimation, I think one should honestly scan the whole slide on low magnification to ensure that the global appearance is homogenous and typical. Jees, do you have a 'pure' and/or fresh sample lying around that you might want to examine again to see if it looks any better than for example sample #1 from above? That would also help us with the degradation-theory. I do not have any mixed harmala's left, only a few more grams of sample #1. I do not have any DHH from a pH-metric separation left either. If we can confirm the uniformity of crystal shape to be correlated with high purity, this might be a low-tech way of getting important information without the need for specialized techniques. For identification purposes, a magnification of 200x seems enough. They sell USB-microscopes with this magnification for 40 euro's. If one adds a pH-meter with a resolution of 0.1 pH (50 euro's), one could prepare any amount of 'alkahuasca' of conformed purity for around 100 euro's.

My pH meter is not good enough for the harmine/harmaline separation but I still got good results.

With my new THH, I started off with an ayahuasca like McKenna’s; 280 mg of harmine, 25 harmaline, and 94 of THH. The THH is a great addition. Only one of my several purchases of caapi had much THH.

Rild, was that a pH separation done according to the VDS papers, or good ole pH 8.75 separation?

An1cca I'll come back, just need some time....

I have a good pH meter but the calibration was bad so I got about two parts harmine for each one part harmaline. After the zinc tek, there was almost no harmine or harmaline with the THH.

I love the THH but one person had lots of vertigo with it.

Bravo, all involved thus far!
I give you all a standing ovation!

I've just finished reading this entire thread, as well as about 8hrs yesterday reading all the others I could find. I've got a couple questions that I'll place in orange interspersed in the text. Let me see if I've got all this correct:

Per this thread and VDS, 5 manske followed by 5 pH-dependent precipitations yields a harmine/DHH separation efficiency approaching 95%. This can be done without a pH meter using NaHCO3/Na2CO3 respectively, or with a meter using any available base. Seems like the easiest choice is aqueous NH3+ due to it's ease and ability to fully evaporate from the precipitate.(VDS protocol 2.1)

The reduction reaction takes about 9hrs, continuously stirred, to proceed to a 95% conversion.(VDS protocol 3)
Has anyone done this under reflux to speed up the reaction?

Using ammonia after the reduction during the clean up takes advantage of the ligand exchange properties of ammonia and ZnXX to form a soluble at high pH amine complex. Allowing further basifying with NaOH if desired without the precipitation of ZnXX I see that NH4Cl was initially experimented with to do the same thing. Will this complex also form with Mg? What about Al? As I see both of these metals to be superior reducing agents.

Manske can be preformed on THH as well as Harmine/DHH, as long as the ZnXX has been removed from the equation. Although it doesn't really seem necessary given that a precipitation by ammonia followed by an ammonia wash produces very nice looking product.

Conversion of the freebase THH to the fumarate is diprotic requiring only one mol of acid to two mol of THH. Has this been tried with Harmine/DHH? It would be great if they follow the same rules.

The reduction reaction can be carried out with the same conversion efficiency using harmine/DHH mixtures if one isn't concerned with separation, but rather a more favorable alkaloid profile.

I look forward to running this reduction and hopefully running on TLC to verify using harmine/DHH mixture to convert to THH.

This deserves a standing ovation as well




No, only 1 pH-specific separation step is necessary when using a pH-meter.



Correct, ammonia is preferred for many reasons.



I can see you're on a busy schedule . Installing and cleaning up a reflux-setup takes time and effort as well...



I tried aluminum foil without effect, even after boiling for many hours. Galavanised (Zn) nails however work well (just for experimental purposes of course). I didn't test magnesium.



As you might have seen, my results after a 'combined' reduction were very poor. This might however be caused by the lack of stirring which seems to be crucial for effective hydrogenation.



Feel free to try!



This is what the article states. However, as I mentioned above, experimental conditions can greatly influence outcome. Constant stirring and slow basification seem beneficial.



Keep us posted

I'll be trying out all of the above questions when I can.

Thanks, I was a little confused about the separation conditions to achieve that efficient of a separation. I'll have to go back and reread some of that. I was only asking about the reflux because that is how most reductions are performed, but agree that it isn't necessary. I'll use any good excuse I can to play with my glassware.

I ran into someone in another thread talking about using Mg and HCl yesterday and kindly asked them to post up their results with regard to reaction conditions and the addressing of soluble salts of Mg in this thread. I hope they will. Personally, I'd rather use HCl and Mg under reflux, but agree that the Acetic/zinc would be safer for most.

Thanks for your interest BioTron.


5 x manske makes for a good purification. This seems to be a good bases whatever one does next.
The other 5 x bicarbonate/carbonate was a style that VDS proposed to separate the harmine from DHH, which seems to work with some loss as a price to pay, as shown by An1cca. It is a style that permits separation without pH monitoring.
But you can do that separation with lesser loss with ammonia in 1 go only, by monitoring pH, working trough the pH depression that VDS learned us about.

* * *

About not separating components and reducing the mix with Zn, then filtering the excess unreacted Zn away (to be reused), and then basing with ammonia or another base of choice that takes pH high enough: it does work by my experiment.
Even if this lead to a lesser complete reduction of the DHH, I had no noticing subjectively of DHH that was likely unreduced.

I've had notice of harmine solo, harmaline sole, THH solo, pure(?) as they come kitchen style. There's not much use of THH solo (referred weak MAOI) but more so as a duration extender, so harmine + THH works nice since the harmine alone has a fairly abrupt cut-out.
But on this combo I feel some condiment of DHH would not be a misfit at all, it's just that 50%harmine + 50%DHH (as rue seeds tend to offer) is quite a bit too much of DHH. So I was planning for something like 175mg harmine + 25DHH + 100THH in the next trials (to support the light). Numbers in freebase.
I mention this to say that reducing a mix of harmine+DHH (nothing separated) and if there would be a neat part of the harmaline not reduced, then this is nothing like a fail because I think some DHH is actually welcome in the end result, but not as much like rue's 50% natural composition.

Jees, that is exactly what I am after. Just about the same ratio you just mentioned.

Qualitative results are in for the question of harmine/DHH fumaric acid solubility.

A small amount (about 5mm deep in a 12mm ID test tube) of mixed harmine/DHH from a twice manske precipitated syrian rue extraction was treated with 3.5ml room temperature FASW. Most(about 75%, I'd guess) of the mixed alks went into solution displaying the characteristic fluorescent brown/yellow solution that is typical of acetate and hydrochloride salts. A small scoopula of dry fumaric acid was placed into the solution. The tube was vortexed(by finger flicking) and the remaining mixed harmalas went into solution leaving behind about a 1mm pile of undissolved/unreacted fumaric acid.

This mixture was treated with 3ml and 6ml dry IPA to see if the fumaric salts would precipitate. At room temperature, they did not. The remainder of the fumaric acid did go into solution. The solution was placed into an evaporating dish and dried over a steam bath, then further microwave dried to yield obviously unreacted fumaric acid, some yellowish crystals of mixed harmalan fumarates and some of the characteristic brown residue that is almost always present when drying harmalan salts, even if from fairly white colored freebase was observed.

The solids were treated with dry boiling IPA and the solids went back into solution, again displaying the characteristic fluorescence, the color was very pale yellow. An excess of IPA was used(50ml probably, as the harmalan fumarates were a little stubborn, but eventually did go into solution. This solution was placed in the freezer. No precipitation was observed. The solution was reduced to 10-15ml(estimate) over a steam bath,dried over MgSO4, filtered, and placed back into the freezer. It was left for approx 1 hour, no signs of precipitation.

The above solution was dried over a steam bath. FASW was used to wash the crystals in an attempt to wash the excess fumaric acid away. At room temperature, this seems to work well. So recrystallization was the obvious next test. Three ml FASW was added to the mixed harmalan fumarates, none seemed to disollve, but the FASW did acquire a yellow tint, indicative of some harmalan fumarates going into solution. The solution was heated to boiling and all harmalan fumarates went into solution. The solution was placed into an ice bath. Nearly white crystals precipitated. Supernatant was decanted, 3ml H20 was added to dissolve precipitate. Upon heating all of the precipitate dissolved. The solution was treated with NaOH, no clouding occured, indicating that harmalan fumarates are very soluble in water, even when already "saturated" with fumaric acid. The supernatant was treated with NaOH, clouding and precipitate did occur.

About the same quantity of salt contaminated mixed harmalan hydrochlorides were placed into a different dry test tube. About 6ml of 50% isopropanol was placed into the tube, the tube was heated gently. Some small translucent crystals remained in the bottom of the tube(salt, it looks like this may be a descent way to remove a bit of the salt contamination from manske). The alkaloid solution was filtered and an equal amount of saturated sodium carbonate was added, shaken, then allowed to separate. The isopropyl layer(containing the now freebased alkaloids) was removed and evaporated.(Interesting...) The resulting light tan freebase powder was washed with water, dried then treated with hot FASI dropwise with time allowed to ensure all material went into solution before more addition of FASI. The resulting solution was dried to produce yellow crystals with minimal residual fumaric acid.

Of course none of this was quantitative, that will be for another time. More work needs to be done to find losses on some of these steps, but they definitely have the potential for purification. A centrifuge would sure speed this process up quite a bit!

I went back and reread the results portion of the VDS paper. That is great, easier than I expected it to be. He pretty much wrote the tek. for us already, so long as his procedure is followed, the results should be(and thus far have been) replicable. Isn't science great! So, if harmine/DHH is being separated with ammonia and a pH meter. As the addition of ammonia proceeds, the pH increases pretty rapidly until the first depression, then harmine precipitation occurs. Harmine precipitation continues to occur progressively slower up until the point just before the second depression. Looks like a safe bet would be to stop precipitating the harmine at about pH 7-7.5, then dropwise add more ammonia until the next drop produced no increase in pH(this is where I need to get me a microscope!), filter off the rest of the harmine, then go ahead and drop out the harmaline.

Time to buy a new pH meter, my old one broke last year, a descent microscope wouldn't hurt either, it kinda defeats using TLC for these purposes.

BTW, those crystals grown from slow gassing with ammonia are gorgeous!

More literacure research has shown Mg to form complexes with ammonia, but Al to be precipitated as aluminum hydroxide.

Just a short tip about how to Manske multiple times and BTW to clean the alks from any NaCl.
Harmine, DHH and (not verified) THH in their HCl form are not very soluble into cold NaCl solutions.
What is less known is that they are neither very soluble in cold water.
The first Manske converts the alkaloids salts into their hydrochloric salts because they precipitate from the cold solution. From that point, one can just re-x them in clean water, no more NaCl is needed: dissolve the crystals in a small amount of hot water and precipitate by cooling.
An efficient way to achive that is to use a funnel as small as it can hold all the crystals into its "tail", which holds a loose cotton-ball at its bottom.
Once the very cold solution is filtered, the liquid is saved a part and the beaker is cleaned. Then boiling water is poured into the funnel by very small amount until the holded crystals are dissolved again. Then this solution is cooled in the fridge and cycle continues.
To works well, the beaker + funnel have to be kept inside the fridge during the cold filtration, and inside a high pan, with lid, containing some boiling water during the hot filtration, so the steam keeps the funnel hot.
A good speed of filtering is 1-2 drops/second max: more than that it is better to pour back the filtrate on the filter.
When the Crystals are clean enough, all the saved liquids are reunified and basified to precipitate as free-base a very small quantity of harmalas-HCl remained in solution at each cycle.
I hope that this tip can help someone, and sorry for my bad English...

Hello Michel,
thank you for your investigative mind but I do not agree that harmalaHCl's are not soluble in cold water. If I make a manske and due factors I fail to add enough salt, nothing will precipitate, really nothing in a very fridge cold situation. I'm feeling baffled by your suggestion to be honest. It's impossible to precipitate harmalaHCl's without adding salt, how can this ever be denied? This is really weird.

i observed exactly the same behavior jees talks about.
in fact, right now i have a proof of that in the fridge.

Thank you both for your interest !

Please, try this experiment: put some Harmine/Harmaline HCL (no freebase, no acetate) in a small funnel (with cotton ball) holded by a beaker. Then let drop, drop by drop, some boiling water in the funnel until the xtals are dissolved. At this point the filtrat has already start to crash because the bottom of the beaker is colder.
The key is to use as less very hot water as needed to dissolve the xtals.

You can do the test another way even more simple: put some clean HClxtals in a test-tube which is hold in a boiling, or near boiling, water bath. Add hot water slowly, drop by drop. At each drop, give the time for the dissolution to progress, moving a little bit the tube in the bath. When the xtals are dissolved, do not add more water and take the tube out. At room temperature it will start to precipitate almost instantly.

I am using a reverse-osmosed water, pH ~ 6.5, but I do not think that matters so much.

I you do not succeed, I will upload 1 or 2 pics.

I think what happens in your little experiment is the following: you are using salt but are not aware of it.
The harmalaHCl's are always contaminated with salt, which is inevitably. If you add a tiny bit of hot water drops just barely to dissolve the crystals and nothing more, and cool that, I bet the remains of salt contamination are enough to reach a 15gr/100ml level. It is then just another manske.

Use some more hot water please, at one point you will not perceive crystallization anymore, as you just dropped below (10 to) 15 gr salt/100ml concentration. The point where manske stops working for you. So I believe you examined very right, but concluded that there was no salt because you did not add salt, but were actually using residual salt of the first manske. My best bet on this.