[Enthalpy]

Hello dear friends!

The nicely strained and stable bicyclo[1.1.1]pentane and oligomers are synthesized from [1.1.1]propellane usually, by breaking the central bond.
https://en.wikipedia.org/wiki/1.1.1-Propellane

[1.1.1]propellane is obtained from gem-di(chloromethyl)ethene (aka 3-chloro-2-chloromethyl-1-propene), on which dibromocarbene cyclopropanates the double bond, and then methyllithium closes efficiently two more cyclopropane cycles between the carbons that bear bromine and chlorine.
http://www.orgsyn.org/demo.aspx?prep=V75P0098
(Drawing appended here)

The cyclopropanation takes 4 days, a definite drawback for a product to be torched in hundreds of tons.
http://www.chemicalforums.com/index.php?topic=79637.msg290422#msg290422

• Do the chlorines slow down the cyclopropanation? From paragraph 29.2.1.1.6.1 in Houben-Weyl's Science of Synthesis (Google book): "Electron-deficient alkenes are less reactive towards dichlorocarbene"
• Do you believe the chlorines and bromines can be swapped in the synthesis as in the drawing's lower part? This has other potential advantages.
• Would the dibromoalkene react more quickly with the dichlorocarbene?

Thank you!

[discodermolide]

I think the alternative route may be too expensive.

[Enthalpy]

Being no cheap industrial compound, the intermediate di(chloromethyl)ethene must be produced.

The reference synthesis takes pentarythritol, converts three hydroxyls to chlorides, the fourth to an acid, then eliminates by heat 1*C, 2*O, 1*H, 1*Cl
http://www.orgsyn.org/demo.aspx?prep=V75P0089
http://www.dtic.mil/dtic/tr/fulltext/u2/a267508.pdf
Drawing appended.

----------

Orgsyn calls direct halogenation the "best alternative": 34% yield (starting from the monochloro), mixed products difficult to separate. Though, I feel it interesting for mass production:

• Isobutene and chlorine are really cheap, even for a bad yield. Bromine less so (4k$/t), but it can be recovered from the waste.
• The reaction must be swift and low-tech. Excellent GaN 405nm Leds dissociate Br₂, excellent AlGaN 385nm Leds and mp-Hg lamps dissociate Cl₂.
• Accurate distillation is a small worry at big scale, and freezing instead must separate the dihalo isomers better.

Random substitutions at 6 allylic sites would give 20% desired distant dihalo, 13% vicinal dihalo, 36% less halogenated and 31% more halogenated products at optimum 33% per site, so the 34% yield reported with elemental chlorine is already better than random.

• Addition to the double bond is hence limited. Little halogen and much light are said to help. This is easily done in a continuous process. Cl₂ molecules absorb 365nm light with 4*10^-20cm², or over 10mm at 1kPa and 300K, while Br₂ absorbs 15 times better at 405nm and other species nothing.
• Allylic substitution is very selective over vinylic, and other reactions are visibly under control.
• Precipitation of the dihalo may already limit the number of halogenations in the reported yield. This improves if the reactor distills the species (drawing appended), so that only dihalo exits.
• Steric hindrance may already hamper the gem-dichloro, offsetting the bond dissociation energy that favours it. Bromine would improve both the hindrance and the bond dissociation energy.

Taking ethane as a model available in "Bond dissociation energies" from Yu-ran Luo
http://staff.ustc.edu.cn/~luo971/2010-91-CRC-BDEs-Tables.pdf
we see that a first chlorine makes a second hydrogen abstraction easier locally, while a first bromine brings no clear preference.

BDE     +-
-------------------------
420,5   1,3   CH3CH2-H
-------------------------
423,1   2,4   CH2ClCH2-H
406,6   1,5   CH3CHCl-H
397,9   5,0   CH3CHCl2-H
-------------------------
415,1   8,4   CH2BrCH2-H
415,0   2,7   CH3CHBr-H
397,1   5,0   CH3CBr2-H
-------------------------

Would you see better tricks? NBS and NCS are known for allylic halogenation, but as they allegedly create X₂ first, steric hindrance should stay the same. Hypohalite acids? X-O-tBut? HX elimination bringing the double bond?

Comments, remarks, suggestions, even objections maybe? Thank you!
Marc Schaefer, aka Enthalpy

[Enthalpy]

I think the alternative route may be too expensive.

Thanks Discodermolide!

What makes dichlorocarbene-on-dibromobutene more costly than dibromocarbene-on-dichlorobutene: the reactants and their amounts, the lower yields, the slower reactions, the risks...?

If a launcher decides some day to consume every month 200t of the exotic stuff, I imagine it would be produced at a refinery, in an especially built annex, where byproducts like LiCl, LiBr, HBr... would be recycled into the LiCH₃, Br₂, HOBr etc used by the process. Would that make sense for these quantities already?

[discodermolide]

My gut feeling is that bromo compounds are more expensive than chloro.
Thus dichlorocarbene plus dichlorobutene may be cheaper. At industrial scale it may also be safer. The risks of carbene chemistry are large, very concentrated NaOH, dihalocarbenes, easily polymerised olefins, etc.
My option would be to consider an industrial flow system for this chemistry rather than batch.

[Enthalpy]

Could that route (drawing appended) be faster? It chlorinates later, threatening the cyclopropane instead of the double bond, but now the dibromocarbene acts on plain isobutene, taking hopefully less than four days. The chlorination step is magic for me, but literature exists for methyl- and dimethylcyclopropane.

On pages 2-5 (Pdf 8-11) of his PhD thesis
http://thesis.library.caltech.edu/3853/1/Rosen_aj_1964.pdf
Allan Joseph Rosen cites several groups who got 55-67% of chloromethyl instead of ring opening, and up to 16 times more chloromethylcyclopropane than chloroalkene through vapour-phase photochlorination by Renk et al.

Apparently, bromine uses to open the ring, hence better chlorine. Byproduced HCl is said to harm and should be removed. What do prior bromines do and does the second chlorination work? Opinions welcome!

Marc Schaefer, aka Enthalpy

[Enthalpy]

My gut feeling is that bromo compounds are more expensive than chloro.
Thus dichlorocarbene plus dichlorobutene may be cheaper. At industrial scale it may also be safer. The risks of carbene chemistry are large, very concentrated NaOH, dihalocarbenes, easily polymerised olefins, etc.
My option would be to consider an industrial flow system for this chemistry rather than batch.

If chlorine everywhere is possible, fantastic! But does methyllithium couple two carbons that wear chlorines? I've read it between bromine and chlorine up to now, and once with 1,3-dibromopropane. The last step of the reference synthesis, with 80% yield from dibromodichloro to propellane, suggests that methyllithium never couples chlorine-bearing carbons.

Flow rather than batch permits to draw the intermediates (especially from halogenations) before they react too much, to keep halogen proportions low... many advantages. But if the cyclopropanation takes 4 days, flow won't help: to produce 200t in a month, it would need >27t in the reactor, ouch. Hence a fast reaction is desired.

[discodermolide]

Think about a turbo Grignard reagent, e.g. described by Knochel.
Also a transition metal catalysed reaction may be possible for the propellane forming step.

[Enthalpy]

Close the cycles by other means, I dunno... Experimenters report many attempts, the one efficient way is LiCH₃ up to now, and only with the central bond first. But what if all three cyclopropane rings result from LiCH₃? This would avoid the carbene step.

The last intermediate to propellane would be 1,1,1-tribromo-2,2,2-tri(chloromethyl)ethane, tagged 23 on the appended drawing.

The suggested steps to 23, from isobutene (even more correctly represented now) and acetic acid, are most probably nonsense. Please consider them as an attempt to trigger off better ideas.

Marc Schaefer, aka Enthalpy

[wildfyr]

I'm fascinated by that thermal acetic acid elimination step. How would that come about?

[discodermolide]

I was also wondering about that step. Perhaps FVP and hope for radical formation and combination in the desired manner?

[wildfyr]

Radical formation is might polymerize this mess, so many alkyl halides that can homolytically cleave, and they are on nice highly substituted carbons

[discodermolide]

I did say "hope for radical formation and combination in the desired manner?", perhaps better with pray?

[Enthalpy]

Intermezzo at gem-di(chloromethyl)ethene.

Meanwhile I have doubts that the Cl and Br can be swapped before the cyclization to propellane as I had hoped in the first message. CH₃Li can also close cycles between two Br-bearing carbons, so a condition may be that all Br are vicinal and only the Cl remote. Opinions?

Which chlorinating agent and condition would you suggest to get di(chloromethyl)ethene from butene, instead of addition products or geminal dichloro? For instance, photochlorination of 1,1-dichloroethane gives 80-90% 1,1,1-trichloroethane industrially, probably the wrong way here. HClO, t-ButOCl, NCS, other?

Alternately, could bromochloromethane add to allyl chloride to give gem-di(chloromethyl)ethene after HBr abstraction, as on the appended sketch? The bromine is hopefully more reactive than the allylic chlorine, avoiding a dimer.

It would take reactants available at a vinylchloride or epichlorihydrin factory and use similar processes, for instance at an annex there.
If gem-di(chloromethyl)ethene is easy enough to produce, the polymer is worth a try too. And compound 24 might lead to 23 as well.

Marc Schaefer, aka Enthalpy

[kriggy]

Im wondering if something totaly different could work here to get the dichloroethene
maybe ?


Knoveagel followed by Hunsdiecker ? I mean those reactions work well so it might be better solution that looking for fancy halogenations