Sciencemadness Discussion Board

Crystallisation - truly a dark art

Eddygp - 20-11-2020 at 02:42

Often described as unpredictable, mysterious, based on luck, esoteric practices and with many people reporting their own ritualistic approaches to guarantee good crystals... with that background, there is no doubt there has to be some truth in crystal growing being more magic than science.

However, some of the basic premises we all assume hold true regardless - namely, slow crystallisation yields better crystals than fast removal of solvent.

Imagine my surprise when a complex that had failed to crystallise in over 5 crystallisation attempts, yielding a powder instead, formed huge needles (0.8 mm x 1mm x 5 mm) upon removal of residual solvent on the Schlenk line vacuum. These comparatively huge crystals only grew on the points where the solution was BUMPING. Elsewhere, only a fine white precipitate formed.
I have since got an XRD structure from these.

I was thinking that these apparently unexpected situations may be quite an interesting topic to discuss.
Have you encountered any similar anomalous crystallisation conditions for your products?

karlos³ - 20-11-2020 at 06:57

I have seen something similar when I precipitated 1-benzylpiperazine as dihydrochloride from an aqueous solution with acetone.
They suddenly formed beautiful needles 3-4mm long, unlike a slow evaporation would yield.
Imagine my surprise!

njl - 20-11-2020 at 07:17

I suppose it isn't very special relative to the crystallization you mentioned, but I love seeing nitrostyrene spontaneously come out of solution with a seed crystal

karlos³ - 20-11-2020 at 08:52

Quote: Originally posted by njl  
I suppose it isn't very special relative to the crystallization you mentioned, but I love seeing nitrostyrene spontaneously come out of solution with a seed crystal

Thats true, the first time I saw it happen accidentally, put a little piece of ice into the postreaction mixture and suddenly several cm long needles grew in a matter of ~30sec.
Truly impressive!

Heptylene - 20-11-2020 at 09:13

I've had non-reproducible crystallisations with 4-nitrobromobenzene. First time, the crude product produced 5-mm sized beautiful, transparent yellow crystals. When these were recrystallized again, the crystals were only 1-mm although I used the same procedure (hot ethanol cooled down slowly).

Other problems with crystallisations:
- nucleation and growth on the walls of the container, which makes removal of the product tricky, especially with small amounts and small crystals.
- oiling out for no apparent reason.
- Crystallization that takes days even though thermal equilibrium has been reached and no solvent is evaporating (closed vial).

As you said, it is a dark art...

monolithic - 20-11-2020 at 12:37

https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.70...

Corrosive Joeseph - 21-11-2020 at 04:17

/CJ

Attachment: J. W. Mullin - Crystallization (Butterworth-Heinemann) (2001).pdf (6.1MB)
This file has been downloaded 661 times

Heptylene - 21-11-2020 at 04:28

That was a great read, thanks monolithic!

Eddygp - 28-2-2021 at 14:18

Bumping this topic in order to avoid posting similar threads.
First, just wanted to point out that the crystal formed upon bumping seemed to be extremely good (as seen from the resulting CIF, R1 value and the like). This is remarkable considerig it crystallised within a few milliseconds of exposure to high vacuum. Strange.

Also, I would like to point out that I have never read about one of the methods I like for crystallising... say your compound is soluble in dichloromethane and practically insoluble in alkanes: if you dissolve it in a DCM:heptane mixture with the minimum quantity of the former and then allow the dichloromethane to evaporate from the mixture, you will slowly reach saturation in a progressively more alkane-rich solvent.
Does anyone use this technique too? I find it gives better results than, for this hypothetical case, just evaporating dichloromethane.

DraconicAcid - 28-2-2021 at 14:42

Quote: Originally posted by Eddygp  
Bumping this topic in order to avoid posting similar threads.
First, just wanted to point out that the crystal formed upon bumping seemed to be extremely good (as seen from the resulting CIF, R1 value and the like). This is remarkable considerig it crystallised within a few milliseconds of exposure to high vacuum. Strange.

Also, I would like to point out that I have never read about one of the methods I like for crystallising... say your compound is soluble in dichloromethane and practically insoluble in alkanes: if you dissolve it in a DCM:heptane mixture with the minimum quantity of the former and then allow the dichloromethane to evaporate from the mixture, you will slowly reach saturation in a progressively more alkane-rich solvent.
Does anyone use this technique too? I find it gives better results than, for this hypothetical case, just evaporating dichloromethane.


You can also take your mixture of DCM/heptane and stick it in a jar of pentane, so that pentane is diffusing in while the DCM is diffusing out.

wakatutu - 1-3-2021 at 21:50

From the link provided by Monolithic (https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.70...)

"Choose solvents carefully. Many solvents will react with your compound (e.g.CH2Cl2 should not be used with Cu+ because of Cl- abstraction by copper)."

I looked through the references listed but all of them were to crystallization resources and none pointed directly to a source for this quote about Cl- abstraction from DCM by copper ions. I tried doing a basic google search and a search through my educational institution's library about this but I could not find anything resembling a research paper specifically dedicated to this process or explaining a mechanism or, obviously, a thermodynamic examination of this reaction. Is this just so common of knowledge that it doesn't merit a citation?

Having been poisoned by what seems most likely to have been mono-chloromethane by way of extracting a dried reaction mixture of copper salts, bromide salts, and vanillin derivatives with dichloromethane, I spent quite some time with the old pen and paper trying to work out just what had happened which led me to my abysmal state. I had a less than sufficient understanding of chemistry at the time, and though I have taken quite a few more chem classes in the intervening time (and thankfully experienced a mostly complete recovery, with only the occasional twinge of off-balance and facial numbness), I have yet to satisfactorily settle on the exact reaction that took place.

Anyone knowledgeable about this specific phenomenon, please do pipe up!

And thanks to all the posters here for their contributions to the sciencemadness community's literature on the "dark art"

DraconicAcid - 1-3-2021 at 22:04

I have never heard of chloride extraction from DCM by any transition metal, and DCM was our research group's go-to solvent for late transition metal chemistry. I've done numerous reactions involving silver triflate and silver tetrafluroborate in DCM.

unionised - 2-3-2021 at 14:00

I have seen DCM which has actually turned green when left in contact with a porous stainless steel frit in sunlight.
Not copper- probably chromium, maybe nickel.
So the idea that it reacts with copper compounds wouldn't surprise me

Panache - 6-3-2021 at 16:56

i have no basis in theory or literature to forward this obsefrvation with. Ill proceed anyway. Its a discussion. On the slow is better rule. I don't agree, well not as a blanket rule. Well it depends upon why is it going slowly.
If you take a super saturated solution (i generally finds its possible to make these reliably from the melt of your 'to be crystallised compound' and some of the intended solvent at boiling) and you quench cool this in dryice/solvent bath. Holding it there at baths temperature you will eventually get many small crystals driven by the fact that the soln is super saturated and slowed by the cold (although this increases the solubility driven aspects as the solution is more supersaturated). However if instead you make the same solution and cover it and hold it just below the boiling point of the solvent in an oven (has to be an oven to minimise thermal gradients and prevent crashing, has to be covered to stop evaporation and even worse thermal currents) you get excellent large crystals and quickly.
There is a trade off, the slower you go the more opportunity the solution has to find other nucleation points and hence more numerous smaller crystals. I find this especially true as you drop below 0. The kinetic slowing of crystallisation seems more pronounced. Likely there's sound phys chem to support this observation.

wxyz - 7-3-2021 at 19:31

Quote: Originally posted by Eddygp  

Imagine my surprise when a complex that had failed to crystallise in over 5 crystallisation attempts, yielding a powder instead, formed huge needles (0.8 mm x 1mm x 5 mm) upon removal of residual solvent on the Schlenk line vacuum. These comparatively huge crystals only grew on the points where the solution was BUMPING. Elsewhere, only a fine white precipitate formed.


When the solvent is bumping it is evaporating explosively, that will cause the temperature to drop rapidly, at least locally. I wonder if that is related to the crystal formation.

I certainly have seed. a similar effect when using a vacuum suction when filtering crystals from a saturated solution: more crystals invariable come out in the filtrate. (As the pump cases the solvent to evaporate/boil).


[Edited on 8-3-2021 by wxyz]