I want to share one mistake I made in the past when attempting harmala separation to possibly help others avoid it...

To selectively precipitate harmine from harmaline according to the VDS protocol, one should dissolve the free base harmalas in vinegar and then slowly add ammonia solution until pH about 7.2-7.5 where harmine precipitates , can be filtered out and one can proceed precipitating harmaline.

Even though my syrian rue seeds are tested to be almost 1:1 harmine:harmaline, I usually ended up with LOTS of harmine and only tiny fraction of harmaline:



Now I could have finally solved the mistake. Instead of using vinegar, I simply dissolved harmala hydrochloride salt in warm water and started adding ammonia.

This neutralization process produces ammonium chloride salt, which is in itself acidic (at least in aqueous solutions) and skews the pH readings. This made me overshoot the splitting point and most of the harmaline precipitated in the first fraction.

According to the VDS protocol, one should use acetic acid and ammonia instead, which produces ammonium acetate, a weak, yet neutral salt. Although pH of ammonium acetate solution is not super stable, it should be stable around the range where harmine precipitates.

Now is it possible to reliably separate harmine and harmaline from harmala hydrochloride solution? Using sodium hydroxide solution, this should form sodium chloride, a strong neutral salt that certainly won't skew the pH readings. What do you think?

Of course, the sodium hydroxide used must be pure (lab grade, not common lye for cleaning) and diluted before dropwise addition.

I was also thinking about using potassium hydroxide, as it more soluble in water and thus allows higher concentration of harmalas (less water needed).

Being able to separate harmine/harmaline directly from HCl salt will save one step since the separation can be done right after the last Manske step and there would be no need to make free base only to redissolve it again in vinegar.

IIUC for the separation you need ammonia, NaOH won't work.

I've found the hydrochlorides to be a bit of a pain, as they need to be dissolved in hot water and hot water doesn't stay hot for long. When basing dissolved hydrochlorides with ammonia I've ended up with precipitates that were a mix of harmala HCl and harmala freebase. So what I do these days is dissolve them in hot water, base quickly with NaOH (the benefit of NaOH is that it bases quickly, before the solution cools enough for the hydrochlorides to precipitate again), decant / wash to reduce the amount of NaOH in the freebase, then dissolve in vinegar and do the VDS ammonia separation.

Okay thanks.

But do you know for what reason NaOH won't work? For both harmala acetates and HCl?

I would like to understand chemistry behind it better. In water, harmala.HCl dissociates to harmala base (+) and Cl(-) ions. Sodium hydroxide dissociates to Na(+) and OH(-).

Harmala base(+) + OH(-) -> harmala base (precipitates) + H2O

Both ammonium hydroxide and sodium hydroxide are considered "strong" bases. So why one works and the other doesn't?

NaOH will precipitate your harmalas, but it may not be so good for separation if it doesn't create the same clear pH depressions as ammonia. But maybe it does? VDS used ammonia for a reason, but that was after zinc reduction, where there were other reasons favoring ammonia.

By the way, don't rely on absolute pH values like 7.2 or 7.5, but rather look for relative changes. For me the cut-off point for harmine is often closer to 8.5. Variables like the concentration of ions (including spectator ones), impurities etc. all affect it.

Today I repeated the separation but this time with 3% acetic acid and ammonia...

...aaand it failed again. I got a little more harmaline this time so there is an improvement, but the two fractions are still something like 20:1 by weight, not even close to 1:1 like it should.

What I get is HUGE precipitation, I filter, and then almost nothing precipitates even after adding excess of strong base. The solution stays bright yellow.



What do you mean? The pH rises monotonically until precipitation of harmine happens, in which case pH raises and then quickly drops after adding base. The pH drops are more and more subtle until it is impossible for me to keep track of the changes - maybe with a graphing tool it would work but just from watching the numbers I cannot tell where should be the stopping point, so I simply target the pH 7.5.

Even if I stop and filter earlier, at pH 7.2 or so, where most of the precipitation happens, there is little alkaloids left in the solution. If I filter the first fraction at pH 8.5, I am sure there would be next to nothing left to precipitate.

I am trying to get as controlled conditions as possible to make this repeatable. Harmalas are 5x or 6x recrystallized and I use deionised water.

What concentration of alkaloids do you use? Are there any more parameters that can affect the separations? Maybe water temperature? Acid/salt concentration?

I really don't know why this still does not work for me reliably (even after 2 years of trying and getting periodically frustrated about this ... )

Maybe harmala concentration is too high? I use something like 10-15 g of freebase per liter of acidic water.

I observed the precipitation happens way before the pH 7.5 point (somewhere in the range of 7.0-7.2). It is very hard to get stable 7.5 - I add few drops of 10% ammonia, the pH jumps up to 7.56 and then gradually drops do 7.48 over the course of a minute or so.I repeat adding a whole lot more ammonia (15 mL or so) to get the stable pH 7.5 at which point I have a feeling there is too much base added, but it is not according to the pH reading...

So in conclusion the VDS protocol does not work for me or I have a bad batch of syrian rue or doing something wrong.

Maybe it would be wiser to filter the suspension as soon as harmine precipitates, because filtering out the alkaloids shifts pH by itself (the water pH won't then be affected by the presence of alkaloids).

I think I will try lower concentration of harmalas (maybe 7 g/L or so), using only slight molar excess of acetic acid and filtering early.

I've used NaOH with success (due to inaccessibility of suitable ammonia) but it was time consuming and required constant agitation of at least 20 seconds between drops to effectively read the PH changes with certainty.

Also, for me, the PH drop rarely correlated with the VDS PH points (usually lower). In fact, it seemed to vary from case to case - I think you're right about the harmala concentration being the operative element at play. I added dilute NaOH until the PH drop stablised and only collected the precipitate once I saw the PH climb at least 0.1 back up, irrespective of whatever the PH was. Overshooting wasn't unusual, and I'd re-treat both fractions to the same process to confirm - most of the time I'd get 80-90% to precipitate in the first instance.

I usually ended up with a rough 50/50 harmala split, confirmed differentiation by bioessay.

The pH drops and then climbs back up to where it started dropping. The consensus on the Nexus is to continue adding ammonia until that point + 0.5.

Example: With ammonia continuously being added, the pH climbs until it starts dropping at 7.5, drops to 6.8 and then starts climbing again. Continue adding more ammonia until the pH value of 7.5 + 0.5 = 8.0.

I don't understand why it is not working for you blue.magic. Your harmals look super clean and you surely are doing something right. Are you sure your pH meter is working as it should?

One suggestions if not tried already is to go old school and wash the harmine (with excess harmaline) with (warm) baking soda water. If I recall correctly, that should pickup any harmaline I think, you can confirm why doing this on a sample of the harmaline you already have.

To the baking soda wash, add a stronger base and see if you can recover any harmaline that got picked up.

I've never done the old school baking soda thing, I just read the old posts from back in the day from infundibulum and others. Good luck.

Yes it works correctly. This is the first thing I thought and went as far as to purchase a brand new pH probe. I have calibrated it with two buffers (4.01 and 7.01) right before the separation.

Maybe I will make a video of the whole process since I am really desperate now and maybe you or someone else watching the video will spot some mistake I am doing - it could be some small detail in the procedure that is actually crucial for the success.

It could be also possible the harmala extraction process removes harmaline, but this is unlikely as I use strong base for precipitation of harmalas (NaOH) and always add little more to the clear supernatant liquid after the alkaloids settle to see if there is any more precipitation. For final precipitation I use 20% sodium carbonate and also test for any residual precipitation.

Next time I will also use deionised water (chloride ions from tap water could spoil the separation perhaps ??) and also halve the harmala and ammonia concentration to e.g. 5 g/L and 5% NH4OH, respectively. Lower concentration might avoid problems like too fast precipitation causing crumbling and trapping harmaline or too thick solution causing spots of high pH (it equalizes quite quickly but anyway...)

Maybe I will make my own ammonia to be sure there are no salts that could skew the pH readings.



This is a very good advice. But why baking soda? The free base harmalas are not soluble in water. It should dissolve harmaline acetate regardless of baking soda. Or is just a pinch of baking soda to make sure the water is not acidic?

I've seen something similar mentioned in VDS paper - they washed the filtered harmalas with 1% ammonia (or 3% or sth like that). I don't know why the washing water is basic - it can as well be distilled water, right?

I believe harmaline should dissolve in a warm baking soda water solution while harmine will not. If your first dish does indeed contain harmaline among the harmine, this may be a way to recover and separate it.

I have not tried it, hope I'm not wrong.

Since you have harmaline in the second dish, you can test this out. Try to dissolve it in baking soda water. It should go into solution.

Interesting. I understand it slightly differently. The moment the PH starts to climb I take that as the harmine fraction. I don't return the PH to the original drop point. So using your example where the PH bottoms out at 6.8, I'd stop when the Ph bounces back up to 6.9 or thereabouts and take my harmine fraction then.

I found that the PH depression for the second (harmaline) fraction might often be only marginally higher than the first (harmine) fraction, but of course there's no reason to worry about overshooting.

Baking soda approach never worked for me, despite numerous attempts.

Yeah I second what Padawan said. The first pH drop is the harmine. I then keep adding ammonia till the pH again starts to rise and filter out harmine. Concentration is probably important. Cant remember exactly but I had about 10g harmalas in about 0.5 litres of water/vinegar. Also the ammonia was diluted as per VDS. Again, I dont remember the exact dilution, but a lot less than 10 percent. 0.5 maybe? More dilute ammonia makes it easier to catch the first pH depression. My first cheap pH meter was too inaccurate, so I got the voltcraft model, which works well when properly calibrated. Measured values were lower than those stated by VDS, and I filtered before pH went up as far as 7, same as Padawan.

* I've never had success using a strong base to get the pH depressions.

* Carbonates as a weak base has led to extra trouble: brown goo forming, CO2 forming that skew the pH and thus needs extra care to be eliminated.

* Where to stop adding base to harvest harmine: it is best you stop at the point where the pH depression started.
For example: titrating first climb up to say 7.2 where first precips fall out and then followed by a pH depression down to 6.3 then keep adding base till it reaches back to 7.2, then harvest the harmine.
If you harvest at the bottom point (6.3) then your harmaline will be likely contaminated with harmine. This is not a disaster really for our purposes but best separation happens at hitting 7.2 for the second time.

(7.2 is as an example, can be different for you.)

Okay thanks for all the good advice.

If harmaline is soluble in slightly basic water, then I probably lost it during water washes, since I wash mixed free base harmalas with water many times to remove soluble impurities. I thought both harmine and harmaline base are insoluble even in pH-neutral water.

On the pic on the bottom I roughly illustrated the pH evolution during the addition of base over time. Of course I always add small amount of base, observe pH raise (the curve raises), add more, observe raise and continue until pH drops. I usually add base until the dropping is very mild (less than pH 0.1 drop over 10 seconds or so) which happens around the pH 7.5 point.

Since the absolute pH values cannot be used for separation, how to tell when harmaline precipitates instead of harmine? By observing a bigger drop?

It is definitely a good idea to filter out harmine as soon as the first major precipitation happens (pH drop = precipitation).

One more thing could be important and that is the concentration of acetic acid.

The ammonium acetate formed is a pH buffer and maybe if too concentrated in the solution, it will buffer the base (preventing clean separation) until too much is added and then everything suddenly crashes out.

I think plain water washes on harmaline could be part of your issue. Below is a theoretical calculated plot from chemicalize for logS (natural logarightm of the solubility in mol/L) of harmaline. They have logS=-1.8 at pH 7, which means water may be able to hold up to some amount of harmaline (but it may dissolve very slowly).

Another possible issue is high concentrations as you mention. When stuff begins to crash depends on a lot of stuff. Ionic strength and high concentrations could make things crash earlier and bunch stuff together. Dilution may be your friend here, since you really want the pH swing to drive the precipitation, and you don't want precipitation to turn on early because you are too concentrated. For example if your harmaline is at a higher concentration than logS = - 1.8 (3.5g per 100ml I believe), the plot predicts it will begin to crash out at pH7. This should not be an issue at higher dilution. For example, if you start at a harmaline dilution of 1.7g/100ml (logS ~ - 2.5), then harmaline crashing won't begin until pH~8, and by then harmine should be done crashing.

Notice that this is not exact as we are ignoring ionic strength and this is just a theoretically calculated solubility plot we are looking at.

In your plot, your second crash begins early which means you may be at very high concentrations. How many grams of seed are you using? What volume of water? What acetic acid % did you start at? At what pH does harmine start to crash for you?

Again, I have not tested any of this. Its all purely theoretical and I could be wrong.

One can certainly wash away harmaline by water washes, if that happens then the wash water get a yellow-green tint in it, like fanta a bit.

About pH depressions:
there is only 1 pH depression per element to separate, not multiple!
The harmine will make the first depression, the harmaline the second.
(It is normal for the second to be less deep a depression and could be a plateau too.)

If you add a drop of base then one should not read out pH at once. Not yet!
Wait until some stirring is done and a new pH has established steadily, only then read the pH.

If one reads pH during the establishing of a new pH, then this is just a pH wobble before a new final pH value sets in. These wobbles are not the pH depressions VDS was aiming at, this is very important. Not any pH dance is a pH depression. The pH readouts for pH-depressions are done when the solution is back in a steady equilibrium.

For example when using NaOH then you get pH wobbles during adding base but I was never able to find pH depressions a la VDS. Those only occurred with me using a weak base.

And even when using a weak base, if the concentration is too low then the pH depressions stop happening too. There are conditions for the VDS pH depressions to occur. One could call it a niche in separation country.

Thanks for confirming Jess. Just to add to this, harmine should be much more difficult to dissolve in plain water. Therefore, it may be possible that plain water washes explain the result in the first post.

^^^^ adding to that and to be fair, I've no idea how much harmaline gets washed away actually.

Yeah and it think it would be difficult to quantify. Small pH changes in the incoming water could make big differences because of the strong logS slope.

Overall, I think we can say with some confidence that water washes will make you lose some product and skew the harmine/harmaline ratio higher than 1:1.

blue.magic, what yields did you get? If you got 2.5-3% in the harmine dish and <2% in the harmaline dish I think it makes sense that water washes caused this. If you got >3% on the "harmine" dish and 5-6% with both dishes combined then maybe you did not lose much product and it is a separation issue.