[Enthalpy]

Hello dear friends!

Some molecules are chiral: they exist in different left and right form. Drugs, optics... may want only one enantiomer. While biology often produces one enantiomer, this is more difficult to chemistry because most properties are identical.
http://en.wikipedia.org/wiki/Chiral_resolution
http://iupac.org/publications/pac/69/7/1469/pdf/

One route uses substances whose crystallization separates spontaneously pure left and right forms (most substance instead mix tightly the enantiomers one-to-one in the crystal). If such a substance is for instance a carboxylic acid, it can serve to purify a produced amine by giving salts whose properties differ with the amine's form. Or such a substance can serve in a chromatography column to separate a produced racemat.

For substances that separate spontaneously, a historical process puts two enantiomorph distant crystal seeds in a solution to better control the separation in two forms. Beware I ignore how much the process has improved meanwhile ; though, I dare to suggest here to adapt Czochralski's crystal growth process to chiral separation.
http://en.wikipedia.org/wiki/Czochralski_process
Sketch is below if you're logged in.

By pulling the crystal out of the melt as it forms, Czochralski's process has advantages  to produce semiconductor crystals, and I hope for chiral crystals as well:

• As the crystals are pulled away, they don't melt or dissolve again.
• The diameter is measured as it forms and adjusted by the pull speed.
• The melt or solution is slightly less than saturated, so no other crystal forms.
• Over a melt, the crystal is kept a bit cooler, by an amount that sets the growth speed.
• Over a solution, a different crystal temperature can over-saturate the liquid locally to control the growth.
• This enables to harvest the same amount of both forms and keep the liquid racemic.
• Silicon purity jumps from 1ppm to 0.1ppb during this (very careful) process: interesting for chiral molecules.

-----

Could chiral solvents be useful? I haven't read about them. I imagine they could dissolve racemats and, as they evaporate, let one enantiomer crystallize first and the other later. It looks too simple... must be already abandoned.

Marc Schaefer, aka Enthalpy

[discodermolide]

Two things.
1) how are you going to pull a crystal out of the mix and guarantee it's chirality and optical purity?
2) using chiral solvents would be too expensive for an industrial process and you are assuming that the solubility of the racemate would be sufficient, further as the solvent evaporates the racemate will crystallise back out.

[curiouscat]

Two things.
1) how are you going to pull a crystal out of the mix and guarantee it's chirality and optical purity?

If you start with a d-enantiomer seed the crystal that grows from it is d- and vice versa?

I don't know. But I thought that was the idea? Sort of like:

"In 1882 Pasteur went on to demonstrate that by seeding a supersaturated solution of sodium ammonium tartrate with a d-crystal on one side of the reactor and a l-crystal on the opposite side, crystals of opposite handedness will form on the opposite sides of the reactor."

Just speculating. Won't always work but might sometimes?

[Enthalpy]

Yes, that's the idea. Some substances separate spontaneously, and I hope Czochralski helps here as well, because it controls precisely where crystallization occurs and at what speed. The resulting big single-crystal is much purer, as the strong selectivity of crystallization produces it full effect, without crystal joints where impurities find a harbour.

For semiconductors, the growth tuned to months for 2m produces crystals with a packing perfection unseen elsewhere; supposedly not needed here.

It is my (limited!) understanding that most substances pack regularly one S molecule alternating with one R molecule in their crystal, as this is the combination that fits best, but for some other substances, S molecules pack better with S only, and R with R only, so in a plate they crystallize to a mix of separated S and R crystals.

Pasteur first sorted out the individual crystals , then triggered the separation using two enantiomeric seeds . This can be the method still in use now, or the industry has a better one, I don't know.

Czochralski - using two enantiomeric seeds, yes - would bring to these special substances the advantage of crystallizing only where desired and under controlled conditions. Bridgeman is one other process.

----------

I believe to understand (do I?) that these special substances are the only entry gate for technology into chiral synthesis. Special acids that separate spontaneously can be sorted out and then serve to purify amines that wouldn't spontaneously; special amines purify acids; and so on. Hence separating first the special substaces is useful, and I hope Czochralski helps here.

Next, when separating esters, a controlled crystallization by Czochralski can help as well - though there are other methods.

----------

Is a chiral solvent useful? As a drawback, it needs bigger amounts than an esterification; recycling looks possible in both cases.

It depends on the solvent's chiral selectivity. If it dissolves the S product far better, then R will crystallize first and can be harvested, and S later and can be harvested, before the solvent is reused.

Though, as I figure that a solvent molecule has few contact points with the solute, chiral selectivity is less probable than in a crystal where contact points between the molecules are numerous.

[curiouscat]

Forget chirality; for what large single organic crystals have you seen Czochralski  been used?

All the examples I know are inorganics.

[discodermolide]

Pasteur separated quartz crystals based on their right or left-handed twist. This does not mean that the SiO₂ is chiral, which it is not.
I know lots of methods for resolution of enantiomers in great detail, but unfortunately I can't divulge them here. If I can find the literature I will post them here, but I don't promise i will find them in my chaos.

[Enthalpy]

Why should organic compounds not fit Czochralski? They crystallize as well.

If their liquid-solid transition is convenient (cooling is acceptable), fine! Do it as for silicon, optical crystals or detectors.

If not, then use a solvent.

[discodermolide]

Why should organic compounds not fit Czochralski? They crystallize as well.

If their liquid-solid transition is convenient (cooling is acceptable), fine! Do it as for silicon, optical crystals or detectors.

If not, then use a solvent.

Sure they crystallise. But where and how are you going to obtain your chiral seed crystals? You would have to obtain them by normal methods (which are cheapish). Thus having worked out a process to do this one might as well scale it up instead of then going and trying out a new method.
I think this would be way too expensive for Pharma and certainly Agro.
Furthermore if you are thinking about applying this, say at the drug substance purification step or anywhere between the API and the drug substance, how would you control the polymorph which may form and be the wrong one or even a new one and the particle size? If you end up with different polymorphs then you have trouble.

[curiouscat]

Why should organic compounds not fit Czochralski? They crystallize as well.

If their liquid-solid transition is convenient (cooling is acceptable), fine! Do it as for silicon, optical crystals or detectors.

Mechanical properties? Can they be drawn out into one large boule that'll support its own weight and not shatter / crack  etc.?

[Enthalpy]

Seed crystals are cut from a previous boule for Czochralski silicon, I'd do the same for organic substances.

Strength depends fully on the substace. Sucrose for instance makes strong solids. Single-crystals tend to be stronger than amorphous or polycrystalline equivalents but more brittle. Jaws can hold the boule once the diameter is attained.

Did a message disappear? I thought I had written that already but can't find it.

At a single crystal grown regularly, any defect is very visible. After growth, reflection of polarized light on a polished area of the solid must suffice. If not, dissolve a sample.

Just for comparison: when growing silicon, we have only silicon in the melt, plus impurities soluble in the crystal. The kind of selectivity obtained is in the perfection of the crystal lattice and in purity (1ppm becoming 0.1ppb). Here we're talking about enantiomeric molecules that separate spontaneously from a solution - what impurities don't do in silicon. I'm convinced that the very controlled Czochralski process will produce much purer enantiomers.

[discodermolide]

You posted in the Boctane thread instead of here

[Enthalpy]

I had put it in the wrong thread indeed,  getting old... 
Or maybe a bug at ChemicalForums: when I log in, it reopens a different thread.

-----

I like your idea of sorting individual small crystals with a quick machine.

Light can be concentrated on little more than the crystal's size so the rotation effect is very perceivable. Better: analyze only the deflected light, which then comes through the crystal. Optionally, sort the crystals by size first, for instance by a fluidized bed.

Some crystal axis may rotate less, but just use several rays to avoid the bad directions.

Sorting machines have often an air jet to spew away the desired items. An air valve reacts in ~20ms, the rest of the circuit must be optimized.

-----

Polymorph crystals are in fact an argument for Czochralski. As I understand the process, it's meant to avoid multiple nucleations, by having temperature and concentration in the melt or solution that prevent the formation of a solid. Only the more favourable temperature near the already grown crystal (or at the beginning, the seed) allows the solid to form, and only at the surface of the already existing crystal.

So the seed's nature determines what the boule will be. Here we can choose freely the most stable crystal form as a seed, to maximize the chances.

[discodermolide]

How do you know it is the most stable crystal form? There may be others that you have not seen yet.

I think you may have confused *The Chemical Fora" with the number of threads that it just guesses when you log in and dumps you in one of them

[Enthalpy]

I always open many threads of the "Chemical Forums" at the same time, then log in at one of them, and after that the Forum drops me in the latest thread I opened, instead of the one from which I logged in. A possible cause why I put in the Boctane thread messages belonging here.

-----

Most stable crystal form? I don't know how to guess it for a new compound.

But for well-known compounds like chiral selectors, I suppose all crystal forms are perfectly known.

[discodermolide]

Most stable crystal form? I don't know how to guess it for a new compound.

But for well-known compounds like chiral selectors, I suppose all crystal forms are perfectly known.
[/quote]


Well it was not too long ago that a new crystal form of aspirin was claimed, after years of having stable polymorphs.