[discodermolide]

Silicon boule are hold at their seed only because of their outstanding strength. As soon as the final diameter is grown from the seed, a normal material can be grasped there.

Strength depends on the material... Sugar for instance can make big solid parts that don't break. Single-crystals also differ from polycrustalline or amorphous materials; they're often much stronger but brittle.

I think you placed this in the wrong place?
Anyway I was talking about polymorphs of any particular crystal. If you generate a seed from one process it will have morphology , let's say A. It may be meta-stable and when you place it in your solution to grow it may easily revert to a different morphology, say B. Now if the compound you are interested in has morphology A then you are royally screwed. This mainly because all the tests and evaluations you have done, even registration has been done with A, if you suddenly produce B then the whole lot must be repeated. You may even find different bio-availabilities,rates of dissolution and the like. So the whole process of crystal growth is a minefield. Which is why Pharma companies (I can only speak for them) have crystal engineering departments.

[curiouscat]

@disco

On a flight of fancy, I was wondering if one could make a machine with a high resolution camera and a nanofluidic channel that'd sort individual crystals as they went by?

Something kinda like  a flow cytometry cell sorter. Or are these crystals too small to be able to do that?

[discodermolide]

Depends what you require for size. You can control that by optimising the crystallisation parameters for particle size. So such a flow system may just work.
As you probably know this is how phase separations are done in the plant, using refractive index changes at the solvent(s) interface.

[curiouscat]

As you probably know this is how phase separations are done in the plant, using refractive index changes at the solvent(s) interface.

Yep. But that sounds a lot easier. RI is a macroscopic property needing only a sensor.

I was thinking of crystal detection with some real time image processing algorithm. Although now that you mention is just an optical rotation sensor might suffice?

[discodermolide]

Yes if it can measure fast enough, and the temperature would need to be carefully controlled as would the concentration.
Optical rotation is ok if it's large enough, what happens when the rotation is very small? Then you usually measure at different wavelengths. And the rotation depends upon the solvent you use for the measurement.
So maybe your first idea is better?

[Enthalpy]

I've put my commentary to the seducing selection of small crystals back in the original thread.
So sorry!

[discodermolide]

Here is a compound you might like to consider?


It appeared in the Journal of the American Chemical Society 2013, doi.org/10.1021/ja3104582 along with a host of other compounds of interest to you. The coupling method is especially nice.

[Enthalpy]

Spiroheptane is a good start, sure.
Boron is advantageous if the compound is affordable and not toxic. LiBH4 is one compound that would improve performance over hydrocarbons.
But... Would you have the same one without oxygen? I'd accept a different colour and smell if needed.

[discodermolide]

The point here is that these pinacolboronates can be used to couple this fragment to other molecules.

[Enthalpy]

A rather recent ladderane production method holds two identical molecules close together, each with a C=C-C chain, and irradiates them with UV. If one could make the same with two (substituted) benzene rings to produce hexaprismane, it would bring 2/3 of cubane's advantage as a rocket fuel...

Hexaprismane was nonsense  - and nobody tells me a clear word... 

- Cubane was the first prismane to be synthetized, just to picture the difficulty.
- Strain energy is slightly less in pentaprismane (which has been synthesized) than cubane but worse in hexaprismane - which has not been seen up to now.
- Aromatic rings don't stack willingly, quite the opposite.
- I had hoped to hold by carbon chains two benzene stacked. It's banal and called a cyclophane
http://www.org-chem.org/yuuki/cyclophane/cp_en.html
but if a [2.2]cyclophane gave a hexaprismane, this one would include [2.2.2]propellanes which are unstable, and a longer cyclophane is bad as a rocket fuel precursor.
- Guess what, many people tried precisely that and haven't succeeded up to now
http://sciencelinks.jp/j-east/article/200415/000020041504A0465587.php
including by irradiation, after which the compounds rearrange a lot.

I've also vaguely tried to run a software (hence unreliable) on metal-cyclopentadiene and metal-cyclooctatetraene sandwiches, and PM3 claims that it is not a way to bring two C5 or C8 rings (flat aromatic anions) closely stacked, neither within a single complex nor between two complexes.

[curiouscat]

A rather recent ladderane production method holds two identical molecules close together, each with a C=C-C chain, and irradiates them with UV. If one could make the same with two (substituted) benzene rings to produce hexaprismane, it would bring 2/3 of cubane's advantage as a rocket fuel...

Hexaprismane was nonsense  - and nobody tells me a clear word... 

....and who wrote that nonsense?

[Enthalpy]

  I've quoted the author together with his nonsense!

Just for fun:

In the solid, benzene rings arrange rim-to-face if I understand properly. A pressure like 15,000 bar (still accessible to steel cylinders and pistons) should bring them parallel, to the same 300pm distance as in a [2.2]cyclophane, ready for photoexcitation, but flat and without the [2.2] links that bring unstable propellanes in the desired prismane and introduce rearrangement paths.

Provided the benzene rings really get parallel, they wouldn't necessarily be aligned, but maybe some substitutions bulkier than in xylene would help - or take the [3.3]cyclophane and compress it to the distance of the [2.2].

Then, irradiate through the ceramic piston: if the volume shrinks you're on the right path, but if the cylindre detonates you better have stayed out of the path.

Well, the pressure relies on energy computations by AM1, hence suspect... All by a non-chemist, even more suspect.

Marc Schaefer, aka Enthalpy

[Enthalpy]

The easily obtained vinylcyclobutane could make the dicyclobutyl (Boctane) and tricyclobutyl, as suggested in #30 and #31.
http://www.chemicalforums.com/index.php?topic=50579.msg233818#msg233818
In the alkenes-to-cyclobutane photoaddition, an excited electron is said to spread to both alkenes when they meet. Exciting only the bigger alkene diluted in ethylene would be more selective, but meanwhile I doubt that the excited electron from the bigger alkene has enough energy to spread to the smaller one.

The photoaddition remains possible, with an Xe₂ lamp at 172nm exciting the ethylene. It will also excite the vinylcyclobutane (represented by vinylisopropane in the appended spectra), so the partial pressures can't differ too much, and the reaction is less selective, giving as well the useful tricyclobutyl and byproduct cyclobutane - see the appended sketch.

Let's imagine an equimolar mix, equal cross-sections and equal cycloaddition probabilities. Four absorbed photon moles would yield 164+110g of products and waste only 56g ethylene. If the lamp is 40% efficient and the reactions 50%, 1kg fuel needs 29 mol photons and affordable 14kWh.

Marc Schaefer, aka Enthalpy

[Enthalpy]

UV light costs a bit, and a path to cyclobutane without light is known from
"Science of Synthesis: Houben-Weyl Methods of Molecular Transformations" ch. 48.3.1.2
(from 1,4-dichlorobutane and gaseous K+Na at 220°C and 0.1 torr with 70% yield), so it's interesting again to "oligomerize" cyclobutane.

The C-H dissociation takes 404kJ/mol for cyclobutane as for tertiary unstrained alkane, according to
Neumann, "Organic Chemistry", chap 11.4
which is said to ease free-radical halogenations. Though, I haven't found direct reports of cyclobutane halogenation. Comments welcome!

Free-radical halogenations may exceed the intended degree, but

• I want anyway a mix of bis- and tris-cyclobutyl (appended sketch) to raise the flash point from +20°C to +55°C
• Plentiful isomers are welcome to lower the melting point
• If some isomers freeze too easily, they can be filtered out at cold
• I plan to separate the products as the reactor synthesizes them.

Since unselective halogenation is fine, maybe Cl₂, HBrO, Br₂+Cl₂ or even H₂O₂ can accelerate this step and save UV light further. Could the reaction rate be controlled by slowing the halogen inflow?

The oligomers shall result from Wurtz coupling (or a better idea).

• The competing cyclobutene production is energetically more defavoured than with straight alkanes.
  
• More than three rings aren't supposedly desired as they freeze more easily. They can be distilled or freezed away.
• Bis- and tris- simultaneously maximizes the proportion of monohalocyclobutane in the reactor hence reduces the too heavy products.
• What if the mono and dihalo bear different halogens? I imagine more reactive dibromocyclobutane kept in excess chlorocyclobutane by controlling the inflow in the continuous process. Opinions?
• Maybe the metal can be gaseous if advantageous.

How easy would it be to couple a halocyclobutane with plain cyclobutane? For one lorry a week of course. Avoiding the metal would be nice, and hopefully the degree of oligomerization is easier to control.

Details should come about the reactor that separates the mono-, di-... halocyclobutanes.

Marc Schaefer, aka Enthalpy

[Enthalpy]

For the Wurtz coupling among mono- and dihalo-cyclobutane, I've suggested a more reactive dihalo in excess monohalo. Though, the potential drawback is the production of bicyclobutane. Highly strained, but obtained very efficiently from 1,3-bromochlorocyclobutane - probably because all the atoms for the internal coupling are already there once the bromine has reacted.

The other way may be better then: have a more reactive (bromo) monohalo- and less reactive (chloro) dihalo- cyclobutane, and enrich the reaction zone with dihalo to get the desired proportion of tris- versus bis- cyclobutane.