[discodermolide]

Yes I realised what you were getting at with the dissolving metals. I am not aware of this being done with cyclic amines as you describe.
Deprotonation is better achieved with NaH or similar. Dissolving metals mean solvated electrons and they may well cause havoc in this system.

[Enthalpy]

OK then, just gas phase reactions between the amine +chloride. The same real-time separation of the product must work for the diamine, being it distilled at one step (zeroamine and monoamine in the single reaction zone, diamine rains down, use the reactor alternately for both products) or in two steps (zeroamine in the upper reaction zone, then packing material to separate the monoamine, lower reaction zone, then packing material to separate the diamine), or the same reactor used for two successive aminations.

Atmospheric boiling points in °C as estimated (!) by Mpbpvp, route to dimethylamino-quadricyclane:

+191 Dichloronorbornene
+207 Chloro-(dimethylamino)-norbornene (harvest it)
+224 Bis(dimethylamino)-norbornene (avoid it)

Again Mpbpvp estimates, route to bis(dimethylamino)-quadricyclane:

+201 Trichloronorbornene
+228 Dichloro-dimethylamino-norbornene
+243 Chloro-bis(dimethylamino)-norbornene (harvest it)
+267 tri(dimethylamino)-norbornene (avoid it)

The differences are smaller than with azetidine and need some efficient distillation method. A lower pressure linders the temperature.

In addition to cyclopentadiene and di or trichloroethylene, dimethylamine is a mass product too. This fuel only needs to work.

[curiouscat]

Can you estimate, say, cyclohexane & styrene & trichlorophenol with Mpbpvp?

How accurate are these Boiling Points you are mentioning?

[Enthalpy]

Mpbpvp was integrated into the bulkier EpiSuite
http://www.epa.gov/oppt/exposure/pubs/episuitedl.htm
from which, after installation, Mpbpnt.exe can be started directly. Though, installing just the small Mpbp from the package needs light touch - something like a dll or a knowledge file not in the Mpbp subdirectory.

It gets Smiles as an input, has a data base of known compounds and also estimates properties. Bp use fairly accurate except for quadricyclane, mp are completely unusable. It's just Stein&Brown and Joback - but additive methods consistently fail on mp, like 50K wrong.

Meas  Pred  Compound
 +81   +90  Cyclohexane
+145  +147 Styrene
+246  +262 2,4,6-trichlorophenol

Not too bad. The vapour pressure must be about as accurate.

[curiouscat]

Pretty good I'd say.

[Enthalpy]

The amination must be controlled to leave one chlorine on norbornene for the next steps. A reactor example is appended.

Its fractional distillation, which could use a packed column, condenses the desired compound soon when it forms to harvest it. With a different pressure and possibly temperature, in the same reactor at a different time if desired, as well trichloronorbornene can become chloro-bis-dimethylamino-norbornene, or the amine can be azetidine - this isn't even limited to the present set of compounds.

Chloro-bis(dimethylamino)-norbornene can also result from adequate dibromochloronorbornene, which is expected to suppress the unwanted gem-bis-dimethylamino.

Not represented: the reactor needs cooling and a means to remove HCl. Or does HCl form dimethylammonium chloride here? That would be a reason to separate the distillation column from the reaction zone and collect the salt at the bottom or at a whirl.

Marc Schaefer, aka Enthalpy

[Enthalpy]

Light is cheap to produce quadricyclane from norbornadiene as it makes efficient use of easy wavelengths.

It needs a sensitizer, and many have been studied. Substituted norbornadienes tend to accept longer waves.

Acetophenone and plain norbornadiene to quadricyclane show 90% quantum efficiency at 366nm and 388nm:
Crc Handbook of Photochemistry and Photobiology, table 17.1 (Google book)
so for instance azetidylquadricyclane (C₁₀H₁₃N 147.22 g/mol) would need 7.5 mol/kg photons at 385nm (3.22eV = 311kJ/mol).

AlGaN diodes are 36% power efficient and produce 5W light at 385nm, while 405nm would be even better:
http://www.nichia.co.jp/en/product/uvled.html
so 1kg azetidylquadricyclane needs 6.5MJ electricity. At 100€/MWh = 28€/GJ, electricity costs 0.2€/kg or 40k€/200t only.

A production capacity of 200t in 20d*13h or 0.2kg/s needs 500kW light. If the diodes cost 1€/W the investment is reasonable. The claimed life expectancy of 50,000h lets 1W light produce 77kg azetidylquadricyclane, adding 0.01€/kg.

According to
http://www.nist.gov/srd/upload/jpcrd322.pdf (broad compilation, thanks Nist)
https://www3.nd.edu/~ndrlrcdc/Compilations/Tta/Tta0008.htm (acetophenone)
acetophenone extinguishes 2100 L/mol/cm at 406nm, so 1% in quadricyclane attenuates by exp(1) in 70µm.

With diodes that emit 5W and occupy 1cm², a sketch for a 500kW UV reactor is appended :

• h500mm * D38mm plungers can contain 8*50 diodes to emit 2kW and dissipate 4kW.
• Water would flow at 0.4m/s, but prefer phytane and similar as they insulate
  http://www.chemicalforums.com/index.php?topic=56069
• 48Vdc suffice in the plungers but are uncomfortable elsewhere.
• 250 plungers on a 40mm raster make the reactor 0.5m*1.1m*0.6m.

Marc Schaefer, aka Enthalpy