[curiouscat]

I would think it's the solid. But at the liquid/gas interface you may get some deposition of the solid which could cause you some trouble, especially when stirring. This could cause some of the solid to contact each other and cause a minor friction induced decomposition.

Very good point. Very practical. This is exactly the sort of local phenomenon that causes process disasters. On average a process may seem safe but there might be local fluctuations that create dangerous conditions.

The overall summary I have in mind for now is somewhat like this:

There's no doubt there are severe hazards in the process

OTOH the commercial attractiveness for the route is also very high. The yields n selectivity are high. RM and solvents are relatively cheap. Single processing step and not too hard separations.

Competing routes seem more than marginally more difficult & expensive. More effluent problems.

The hazards do not seem insurmountable if tackled carefully. i.e. it does not seem a stupid, unwarrented or unmitigable risk

Other processes seem to have generated and lived with similar peroxide risks. So lessons may be learnt from those.

[discodermolide]

Carrying out an in depth risk analysis to identify hazards and methods for dealing with them will be required. This is not an impossible task. I'm sure that it can be done safely.

[curiouscat]

Carrying out an in depth risk analysis to identify hazards and methods for dealing with them will be required. This is not an impossible task. I'm sure that it can be done safely.

Indeed. I agree.

In any case, for now it is only lab experiments with H2O2 doses not exceeding 10 ml or so.

In case the results are promising I'll start the formal hazard analysis process before scaling up.

[Borek]

London bombings were carried out with acetone peroxide, it was discussed at CF in the past. I believe it is interesting from a thermodynamical POV as the reaction is mostly entropy driven.

[curiouscat]

London bombings were carried out with acetone peroxide, it was discussed at CF in the past. I believe it is interesting from a thermodynamical POV as the reaction is mostly entropy driven.

Here's a great paper on the entropic explosion part:

http://tx.technion.ac.il/~keinanj/pub/122.pdf

For oxidation to happen each molecule of TATP would need 10.5 molecules of O2. That's saying 1 gm of the compound needing ~8 Litres of air at STP. (if I did my math right!) That's like saying a volume of TATP would need 8000 equivalent volumes of air at STP to burn completely. Maybe this access is the hard part. I'm speculating that the combustion literally chokes itself. 

Decomposition to acetone is enough to cause the entropic  explosion OTOH. 3 gas molecules from one TATP.

Caution: Above calculations are my antics. Not in the article. I could be spewing nonsense.

[discodermolide]

Thanks for the link.

[Enthalpy]

TATP fully detonates without air. It's a high explosive (5300m/s). Only the complete combustion would need additional oxygen, what does not happen in the detonation (but the initial TNT-like detonation then transforms the reactor's contents in an air-breathing bomb, and may detonate the concentrated hydrogen peroxide as well).

The Wiki article seems voluntarily botched. The London bombings were carried using TATP.

In your risk analysis, will you detonate some TATP too see what happens then? To get an intuitive idea of how sensitive and powerful it is. Better try a tenth of a gram before having tons.

From the linked Pdf, second paragraph: "TATP is one of the most sensitive explosives known" and "with power close to that of TNT". This does of course not rely on additional air.

Beware the linked Pdf has only estimated the heat of formation by software. More serious sources measured +91kJ/mol and +151kJ/mol - and this does not include the partial combustion, for instance 6* 110kJ/mol for the CO, unless you do believe that a detonation produces ozone besides hydrocarbons.

Primary Explosives, By Robert Matyáš, Jiří Pachman (available from Google Book)
and
http://onlinelibrary.wiley.com/doi/10.1002/prep.201100100/abstract

Maybe you could drag Anders Hoveland in this conversation, if he's still reachable? He has practical experience with explosives, at least in lab amounts.