[Enthalpy]

Some of the suggested reactions are favoured only if the byproduced salt is immediately solid. Potassium looks possible even if the salt condenses after the reaction, and the reactions with metal vapours I've read up to now use potassium... Cadmium and zinc wouldn't need their salt to condense immediately. Other reactions need it, especially to make carbenes.

In this (hypothetic) situation, a fast reaction needs a big contact area between the gaseous reactants and the produced salt: fine salt can be introduced in advance, the grains can be ground if they grow too much. It needs also to move the powder against the gases. While sound, jolts or a fluidized bed achieve it, I suggest a rotating drum to create the movement more gently, as in the appended sketch.

The whole vessel can rotate, but gas joints are easier with a separate drum. The drum can have small side walls and a 3D shape that keeps the powder within.

Marc Schaefer, aka Enthalpy

----------

Toxicity: I've heard about metallic cadmium and will have a look at the salts. How critical do you feel it is? There is no good alternative, since zinc is little volatile. At least, the produced salts can be recycled in situ, for instance by electrolysis - it's an advantage I see to the process: no halogens, no metals on the road nor on the bill.

[Enthalpy]

Toxicity of cadmium salts! Surely this makes this method very unattractive?

Cadmium halides are indeed yuk, thanks Disco!

Zinc halides look less bad. Zinc vapour needs a higher temperature, for instance 397°C for 10Pa or 337°C for 1Pa. Fine chemicals would suffer from it, but the strained alkanes I need are stable enough. Some examples at 400°C:

• 0.004/h cyclopropane to propene
• 0.07/h cyclobutane to ethylene
• 0.4/h 1,4-dichlorobutane to butadiene and HCl

unless, of course, something like ZnCl₂ meddles in.

Plain zinc is solid at 400°C but some alloy must be liquid if that helps. It boils down to surface dirt preventing the evaporation. The already suggested laser shots solve it too.

[Enthalpy]

I've found data for the destruction rate of spiropentane at heat - Nist is just invaluable:
http://kinetics.nist.gov/kinetics/ReactionSearch?r0=157404&r1=0&r2=0&r3=0&r4=0&p0=-10&p1=0&p2=0&p3=0&p4=0&expandResults=true
at 370°C it's 0.1/h and at 430°C it's 4.3/h, so even spiropentane permits a significant vapour pressure of some metals. Provided, of course, that other species don't catalyse the destruction.

[Enthalpy]

[..] The reaction actually forms gaseous AX molecules
[(g)A + CH₃X -> (s)AX + .CH₃X considered by Enthalpy]
With poor reaction enthalpy because AX tends to be high-boiling. [..]

Eventually, I've found thermodynamic data about gaseous salts there, very nice, many thanks:
http://chemister.ru/
which tells that even gaseous salts produce enough heat to make most such reactions exothermic hence easy. So the reactions wouldn't need a big contact area with pre-existing salt to proceed.

I plan a clean update of my message of 26 January 2014 about making monoradicals.

[Enthalpy]

With data mostly from http://chemister.ru/ I've evaluated the heat of reaction (this time negative means exothermic) when gaseous alkaline metals and halomethane make a methyl (to represent other radicals) and a salt, produced gaseous before it condenses elsewhere - see the appended image.

This is at +220°C just to mimic a synthesis of cyclobutane from 1,4-dichlorobutane and potassium vapour (0.1 torr) found in chapter 48.3.1.2 of "Science of Synthesis: Houben-Weyl Methods of Molecular Transformations".

• The enthalpy explains such reactions. It doesn't preclude to provoque the reaction at the salt's surface.
• Methyl is among the most difficult radicals. t-Butyl and cyclobutyl are 34kJ easier.
• Alcohols would need exceptional conditions.
  
  • Converting them to halides looks easier
  • Or react at the surface of the salt if that works
  • Or react at the surface of the liquid metal
  • Or use Grignard-like intermediates. Zinc vapour?
• Chlorides look as good as bromides and iodides, but iodides can be easier to recycle at the site. Even fluorides seem possible.
• K is more volatile than Na. If necessary, eutectics are liquid at RT. Rb would help difficult cases.

For small hydrocarbons, the relative vapour pressures fit the reactor sketched here on May 03, 2015.

Marc Schaefer, aka Enthalpy

[Enthalpy]

A pulsed laser, for instance vanadate or Ti:sapphire, can also evaporate refractory elements as reactants, like boron, silicon or carbon. One setup is described here on Oct 12, 2014
http://www.chemicalforums.com/index.php?topic=72951.msg280320#msg280320
similar lasers already clean stones and cut metals, where ultrashort pulses leave the underlying material undamaged.

Other methods are possible, like electricity exploding a fed fibre when it touches a counter-electrode. They should be quick enough that propelled atoms meet the other reactant, rather than the other reactant meeting a hot surface. Whatever the method, a very inert carrier gas like helium or argon can usefully thermalize the ejected atoms and spare the other reactant.

The reaction of such atoms would be unselective, but maybe useful to obtain exotic compounds from small reactants, as products or as short-lived intermediates for other reactions. Examples with carbon:

• Cyclopropane (more probably propene and ethene) from ethane.
• Cyclopropene (more probably allene or cyclopropane and methylidenecyclopropane) from ethene.
• Bicyclo[1.1.1]pentane and bicyclo[2.1.0]pentane (more probably methylidenecyclobutane, maybe bicyclobutane or cyclobutadiene) from cyclobutane.

It depends on whether a carbon atom only abstracts two hydrogens from the other reactant or inserts immediately in the molecule, and how much the radicals rearrange.

Marc Schaefer, aka Enthalpy

[Enthalpy]

Atomic carbon may well be what produces acetylene in Berthelot's synthesis since 1862 - not so new hence. Thermalization by the carrier gas differs in my proposal, less current through the gas too.

Atomic carbon isn't as reactive as I thought. I binds a first hydrogen atom much more weakly than carbon in a molecule binds its last one, while carbene makes the strongest bond with one more hydrogen, according to Yu-Ran Luo and Nist's ancestor
www.nist.gov/data/nsrds/NSRDS-NBS31.pdf

Atomic carbon shouldn't subtract a hydrogen from the target molecule, but like the more reactive carbene, insert on double bonds instead. The resulting strained carbene must be short-lived, so while bicyclobutane from propene and spiropentane from ethene are worth a try, products of cyclopropene like propadiene and propyne are more reasonable. Substituted cyclopropenes live longer.

Marc Schaefer, aka Enthalpy

[Enthalpy]

I suggested here pulsed laser light to evaporate the solid reactant (metal, graphite...)
http://www.chemicalforums.com/index.php?topic=72951.msg280320#msg280320
but a xenon arc lamp does it faster and cheaper.

Consuming 150W to 20kW, they convert >50% to broadband light. Their colour and tiny emission volume match >6000K, so a concentrator lets sublimate or evaporate any metal or metalloid, even graphite. They cost around 10k€ for 10kW_e and last few continuous years. Serve in cinema theatres and sunlight simulators.

The joined sketch has no uniform scale, nor does it show the probable focal plane between both optics. Some carrier gas and reactants blows at the window shall keep it cool and clean. Wavelengths that heat the lens and the windows or damage the products or reactants can be filtered away. The reactor is coupled with an important cooling and separation unit to reinject the carrier gas and reactants.

10kW light would sublimate 50mol/h graphite, or 100mm³/s: feed a Do=10mm Di=8mm tube at 3mm/s, or maybe a bunch of fibres.

3289K give 1kPa graphite vapour pressure, and most other elements are easier. The usual method estimates then 0.27mm/s sublimation speed, or 2s to volatilize the tube, consistent with 3mm/s feed and ~10mm hot tip.

A L=10mm D=10mm tip at 3289K radiates 2kW. This increases as T⁴ and the sublimation rate at some T²⁸, so a smaller hotter tip saves power: prefer a narrower thicker tube or a rod if light can be concentrated enough.

The rod (100W/m/K for hot graphite in plane direction) conduces 1kW away across 10mm drop but the feed speed brings much back. Steady gas (100mW/m/K when hot) conduces 0.2kW away from R=5mm at 3000K; convection increases this.

Sufficient reactant pressure lets the vapour atoms make the product rather than rebuild a solid - or take a semi-vacuum where vapour atoms can spread far enough. The evaporated atoms should collide a few times more often with a carrier gas (like argon) than with the reactant so they have thermalized.

The destruction rate of the product is difficult to predict, but Berthelot's synthesis uses harsher conditions and the fragile acetylene survives. By the way, Berthelot's simpler apparatus is worth trying on other reactants like propylene, butene... with a carrier gas. The xenon lamp setup is more caring with the gaseous compounds here.

If producing 24/7 a C₄H₆, 3k€ electricity and 1k€ lamp wear make 2000kg a month at perfect yield for a single reactor. Typical products (...if any) would be small and energetic: cyclopropenes, maybe bicyclobutanes, and similar with graphite, while metals and metalloids could couple reactants or make organometallics.

Marc Schaefer, aka Enthalpy

[Enthalpy]

Atomic carbon for chemical synthesis is known.
https://en.wikipedia.org/wiki/Atomic_carbon
https://en.wikipedia.org/wiki/Phil_Shevlin

An electric arc has been used between graphite electrodes. Diffusion through tantalum and desorption at >2200K is an other source with little C₃ and C₂ in contrast to thermal evaporation
https://aip.scitation.org/doi/10.1063/1.4895806

Addition to olefins can make spiropentanes or allenes, Web search
"Atomarer Kohlenstoff reagiert mit Olefinen"
"Cyclopropylidene to allene rearrangement"

One synthesis of fullerenes evaporates graphite by a laser. A xenon arc lamp may be faster, unless a much higher transient temperature is wanted.

[Enthalpy]

Diffusion through a tantalum wall and desorption of atomic carbon is seducing. Reportedly, the hot wall dosn't destroy all the reactants and products.

Diffusion and desorption must be slow, but a (set of) narrow long tube(s) improves the area. Bending and coiling would pack it in a compact reactor. Tantalum being extremely ductile, drawing a tube through a matrix with a kernel makes the tube longer and thinner-walled, which accelerates the diffusion. Biassed rolling is known too to make thin tubes.

It should be easier to fill the long narrow tube(s) with a liquid or paste and pyrolyze it. Toluene and others spring to mind. The process could look like:

• Fill the tube with a liquid or paste.
• Seal the ends by fusion, possibly in a chosen atmosphere.
• Pyrolyze, possibly in a separate location. Hydrogen diffuses quickly through the walls.
• Use in the reactor at a higher temperature as a carbon source.
• Cut the end(s), reuse the tube(s).

Closely packed narrow tubing wastes less energy by radiation. The reactor's walls can reflect the light back to the tubes.

The flow of reactants and products can be optimized, for instance axial through a stack of spiralling tubes, or radial through helices.

Most elements are easier to atomize than carbon (717kJ per atom): nitrogen (473kJ), boron (565kJ), silicon (450kJ), lithium (159kJ), sodium (108kJ), magnesium (147kJ), calcium (178kJ), zinc (130kJ)... Some tend to make polyatomic gases, which the diffusion-desorption prevents. Interactions with the cooled tube are a limit. Other refractories include niobium (cheaper), tungsten and rhenium, zirconium and hafnium (reactive), molybdenum (brittle), nickel and cobalt (moderate temperature).

Atomic nitrogen is as exotic as atomic carbon and may provide new syntheses too. Few years ago, a group claimed N(NO₂)₃ might be stable and interesting, for instance as a rocket oxidiser. Could it result from atomic N in NO₂ gas? Just NO is a more probable product.

Marc Schaefer, aka Enthalpy