The yields were 19.5 g "harmine" and 1.95 g harmaline, so it is exactly 10:1 "harmine":harmaline.

I then reprocessed the supposed harmine and got roughly the same ratio (don't remember the weighs).

I will do a third run, separating at the first pH depression as suggested here. I guess this will finally separate all the harmaline.

Form rue, I usually get around 5% total alkaloids (free base, after all the washings). It's something like 65 g from 1 000 g seeds (6.5%) but after washing and Manske, it goes down to sth. like 40 g.

OK, so we on you got 19.5g and 1.95g, how many grams of seed did you start with?

When you separate the alkaloids, what concentration are you at?

It's hard to tell, I start with mixed alkaloids from several previous extractions.

The concentration is 20 g/L purified mixed harmalas (after Manskes and everything) the solvent is 3% acetic acid. I add 30 mL glacial acetic acid to 1 L of warm filtered water, then add harmalas, wait until everything dissolves and then start adding 10% ammonia solution (using a pipette). The pH probe is held inside the beaker for continuous measurement.

OK, one thing you can try is diluting that further to say 10g/L of harmalas. I don't know if it will help in separation, but if you look at the plot in this post, you are at a concentration that is near the large slope of solubility change with pH. Since you want the pH changes to dominate crashing (and not have a high concentration initiate early harmaline crashing while strong solubility changes to pH are still ongoing), maybe being more dilute could help.

Also, as others have said consider switching to a more dilute ammonia source as you finish neutralizing the acetic acid.

Finally, next time don't do any neutral water washes. Instead, do dilute ammonia water washes only.

Just a few suggestions, not sure if they will help and/or work.

Yes this is all helpful, at least I have something to work with.

I will re-process the "harmine" and see how much harmaline will separate this time. This would confirm the concentration is a significant factor.

Then I will try it with harmalas from fresh batch and report back.

Okay I finally did a new separation run with fresh harmala freebase. This time it was not washed to keep harmaline slipping away, just Manske'd 4 times and freebased.

I dissolved 20 g harmala free base in 1 300 mL DI water containing 45 mL glacial acetic acid:



While stirring, I added 10% ammiona solution until pH 5.5 was reached, then added 7% ammonia in small portions (about 2-4 mL at a time).

To my suprise, precipitation started pretty soon, at pH 6.0 or so!!



Here is the supposed harmine fraction:





After the removal of harmine, pH was 6.15:



I dumped in the rest of 7% ammonia, about 50 mL of it, crashing out the supposed harmaline. pH went up to almost 8.0:



After drying, I was left with 5.62 g 'harmine' and 11.92 g 'harmaline'. This adds up to almost the original 20 g (I regularly lose gram or more after each freebasing):





I am sure I overshoot the harmine precipitation (I expected it would occur at 7.0 or higher) so I repeated the procedure with my old "harmine", this time being more careful about hitting that pH drop and indeed got precipitation at pH 6.07.

There was HUGE amount of 'harmaline' in that second sample. I compared the first and second separation results:



Top dishes contain harmine (left) and harmaline (right), bottom is the second run. As you can see, the top 'harmine' is little lighter as it is probably contaminated by harmaline. The harmaline on right is similar in colour although the 2nd one is slightly lighter.

The second separation yielded 3.78 g of 'harmine' and 16.54 g 'harmaline'. This adds up to 20.32 g but I used 25 g so again almost 5 g was lost. I even added excess sodium carbonate but no more alkaloids crashed out of solution, perhaph reducing it would help reclaiming the lost amount.

I took notes on the pH development from adding base in portions, in the 2nd run (4 mL of 10% ammonia at a time):

5.02
5.05
5.13
5.21
5.31
5.42
5.56
5.73
5.95
6.07 (precipitates)

I noticed NO PH DROP at all so I am absolutely confused. The pH rose by slightly larger value each time, then it jumped to 6.29 when the "last drop" was added and precipitation followed. Then pH stabilised at 6.07, which is HIGHER than previous stable value of 5.95.

I used new and freshly calibrated pH probe, deionised water, lower and higher concentration of both harmalas and base and all with same weird result. No pH drop observed.

I will try to purify the fractions by doing another two runs and then make THH from the 'harmaline', then it might be clearer if there was harmine contam. I am too lazy to do TLC at this point but maybe it would be necessary to monitor separation with TLC as I don't trust the pH values alone, especially as this separation does not seem reproducible.

^^^In that second run I think you had to keep poring base and not stop at 6.07. You have a lot of alks to treat there. I suppose you took the pH hill slope partially instead of over the top into the drop. Coincidentally rhymes with the low harmine content, your harmaline might be 'contamed' with harmine?

PS, personally I prefer more or less continuous base adding over batch wise. It renders the pH drop better imho.

pics

The problem with continuous addition is that it takes time until the pH stabilises and during that time, more base is added so we can also skip the pH drop (which I never observed - the mysterious pH drop is already a mystical legend for me )

Next time I think I will add just enough acetic acid to dissolve alkaloids (pH 5 or so), then start adding base in very small controlled batches.

I will keep adding after pH 6 up to say pH 7.5 ?

Hopefully the pH drop will appear.

I had 8.5 g harmine (from a harmine/harmaline separation) that I thought was more contaminated with harmaline than I liked, so I decided to do another separation step.

I dissolved it in tap water with distilled vinegar and proceeded to add a dilute ammonia solution drop-wise, under stirring and with a pH pen in it.

The first pH depression appeared around pH 6. It went down to 5.80 or so, then I let it creep back up to ~6.10. I let it settle and collected precipitate #1 (6.8 g).

Then I continued adding the dilute ammonia solution drop-wise and after climbing to maybe 6.20 and clouding like crazy, it dropped to under 6 and went back to 6.5. It looked like all the harmaline had precipitated. But I continued adding base and cut it off at 9.5 just to be sure. I collected precipitate #2 (1.2 g).

Then I added NaOH to the supernatant to see if any more would crash out, but nothing did.

So it looks like both the harmine and harmaline precipitated around pH 6, very likely below 7 and definitely below 9.5.

I bioassayed precipitate #1 with Chacruna tea and it was great, the element I found annoying in the original product was absent, it felt very much like Caapi.

The result is somewhat unexpected; I wouldn't have thought that such a small amount of harmaline would have affected the experience so much; I thought the harmine:harmaline ratio was closer to 1:1. Also the low pH precipitation points are surprising.

Thanks for the confirmation!

I will do TLC on the harmine and harmaline fractions to see whether there is cross contamination and how much.

So far I judging by the colour and this confirms what you said. Maybe the pH separation points are different in solutions buffered by ammonium and acetate ions (when excess acetic acid has been used).

I repeated the separation on the two batches I had (one mostly harmine, the other mostly harmaline) but again, each separation went differently.

The pH separation method is still very unreliable even though I carefully controlled the conditions.

First of all, the precipitation and pH drop DO NOT happen simultaneously. Precipitation started as soon as at pH 5.2 ... The pH drop happened at around pH 6.4 (dropped to 5.9 or so).

In both batches, the pH drop point was around the same pH but each batch resulted in completely different separation judging by the colour.

In the most-harmine batch, the resulting fractions were quite mixed, the harmine was lighter than it should and harmaline somewhat darker than it should be.

In the most-harmaline batch, the harmine fraction was very dark (as it should be) and harmaline somewhat darker than it should be.

I have another batch of isolated harmaline, which is much brighter than the harmaline I isolated today.

I am pretty much lost, since following the VDS protocol does not work and following the pH drop point does not work either.

The pH separation method seems to be very sensitive on 1) concentrations of acid to dissolve the alkaloids (using excess acetic acid seems to provide cleaner separation), 2) on the alkaloid concentration in the water, 3) maybe even on water temperature.

I am now thinking about choosing some more reliable method, such as running a chromatographic column.

The pH drop is caused by precipitation ( = deprotonation, the freed up protons add to the protons already in solution), but it's a tug-of-war between pH rising (caused by the addition of base) and pH dropping caused by precipitation. Even a small amount of precipitation can cloud the solution, but it may not win the tug-of-war, so the first signs of clouding don't have to be accompanied by a pH drop.

In one of my separations I've had both the harmine and harmaline precipitate under pH 7. In my last one, everything precipitated under pH 8.

Yes, and in that VDS thread it was stated to follow the protocol as good as possible to have as clear as possible a separation. This is immediately our critics to Van Der Sypt that he claims generality of his used method, while it is rather specific in nature. He has proved the protocol's sensitivity himself by publishing a full 1pH drop using ammonia, whereas his curves using sodbicarb only gave him a shallow 0.25pH drop.

I also suspect its closeness of adherence to his precise conditions that are likely causing the difference in peoples results. Particularly the pH 8 versus pH 6 thing.
When I did a run of this protocol I tried to use his exact alkaloid concentration as well as his proportional acetic acid excess. I did use stronger base, due to An1ccas success in that regard, and I used 5% vinegar instead of 7% but I adjusted volume of water to compensate. The centrifugation was a bonus purification step, I've learned to centrifuge harmalas any time harmala base is dissolved in vinegar to boost purity. All water was distilled water, including the water used to make the 1M NH4OH.
Keep in mind that after raising pH it takes several seconds for precipitation to begin and pH to drop.
My results were right in line with the paper and, while I did not do TLC, my harmine•HCl gave no harmaline feel even when dosed at 650 mg. That was so nice
Here's the outline I made for that run and the notes from the actual run: