Sciencemadness Discussion Board

Scaling the Production of Phosgene to 100 grams

Brotato - 21-11-2025 at 21:11

A while back I had reported the liquefaction of anhydrous phosgene as a test concept for the reaction of sulfuric acid + diatomaceous earth with carbon tetrachloride to produce phosgene. Unfortunately due to the moderately low yields achieved with the procedure (~50%), this method was not viable for the future scale up operations. Instead a new method would need to be utilized to achieve both high purity and scalable yields. Initially, I saw the potential use of oleum in this process, however the larger amounts of oleum needed and still low literature yields obtained with this process. The next process I went to was the direct heating of carbon monoxide with elemental chlorine gas over a suitable catalyst. However this method was discounted quickly as it has the propensity to leave chlorine gas in the phosgene which would later have to be distilled out. Then I came upon a soviet paper using carbon tet, paraformaldehyde and aluminium chloride to slowly generate the phosgene upon heating. This was my intended route up until I was checking suppliers for phosgene and saw the decent price of triphosgene that was being offered (50 USD for 500 grams). Upon this discovery, I pursued the route described below.

The catalyst, triphosgene, and toluene were ACS grade and used as received. All glassware was flame dried, the tubing was viton fluoropolymer, and glass joints were sealed using Krytox 206 fluoroether grease. All procedures were carried out in a scrubbed and vented controlled atmosphere glovebox. A 50/50 v/v MeOH : 30% NH4OH solution was used in the scrubber.

Setup: A 500 mL one neck RBF placed in an oil bath was attached to a vigreux column, this was connected with a gas adapter to a viton tube connected to a fritted bubbler. This line was led through a suitable adapter and submerged into the toluene solution present in the stirring two neck 1000 mL flask. This flask was also equipped with a dimroth condenser cooled to -20C and was placed in an ice bath. Finally the output of the dimroth condenser was scrubbed into a wash bottle with a fritted inlet adapter and a suckback trap in between. Refer to the image for more detail.

To a 500 mL one-necked RBF was added 110 grams of triphosgene, 4.27 grams of Beta-CuPc (cas 147-14-8) and a PTFE stir bar. The second flask contained 462 mls (400 grams) of anhydrous toluene previously stored over molecular sieves. The oil bath was slowly heated to 80 C causing the triphosgene to melt and then to 100C causing evolution of phosgene which was bubbled into the toluene solution through a fritted inlet tube. After cooling the joint at the top of the vigreux column was opened and a dilute aqueous ammonia solution was added slowly to minimize any violent reaction with trace unreacted triphosgene. Overall 96 grams (87%) of phosgene was collected into the toluene, the solution is now stored at -20C.

This rxn seems to be a nice dependable way to generate phosgene for use in various processes. I will later detail the synthesis of a xanthine analog purported to have 100 times the affinity of caffeine for A1 and A2 using dipropylurea made from the phosgene in this process.

The yield could possibly be improved by allowing the reagents to mix more properly before heating giving a better interface between the compounds instead of allowing it to decompose to CCl4, HCl and one mol of phosgene per one mole of triphosgene. This is likely why the procedure was not quantitative. However it should be noted that simply blending these reagents together would start forming phosgene immediately and therefore not allowing it to be captured effectively.

Future projects: Soon I will need anhydrous HF gas in a cylinder aswell as SF4 for the trifluoromethylation of cubanes. I have also scaled the cubane synthesis to 22 L glassware and will do a mega post about that once Ive isolated more cubane (working on the photorxn optimization currently). Maybe eventually I will revisit phosgene in order to make a cylinder of the undiluted gas but that will be a long ways away after the above mentioned projects are completed entirely.

Pictures of the experimental setup is noted below:

Attachment: phpR1Ndnk (226kB)
This file has been downloaded 136 times Attachment: phpwsNFeV (493kB)
This file has been downloaded 126 times Attachment: phpSuwEgw (554kB)
This file has been downloaded 120 times

Attachment: phpvN6UOy (439kB)
This file has been downloaded 114 times

[Edited on 22-11-2025 by Brotato]

jackchem2001 - 28-11-2025 at 02:35

Scary procedure, nice thread and pictures
Quote: Originally posted by Brotato  
The next process I went to was the direct heating of carbon monoxide with elemental chlorine gas over a suitable catalyst. However this method was discounted quickly as it has the propensity to leave chlorine gas in the phosgene which would later have to be distilled out.


Maybe a mild reducing agent could scrub the chlorine but leave the phosgene unaffected?

chempyre235 - 28-11-2025 at 19:52

Honestly, scary as it sounds to work with, this is very impressive work, Brotato. Though, I'd probably try to edit this post to include a bold disclaimer, so some unexperienced kids don't get any bright ideas...

Metacelsus - 29-11-2025 at 04:41

Do you even need to prepare phosgene? Many reactions can be run with triphosgene or diphosgene as a direct substitute (with phosgene being generated in situ). As an undergrad I used diphosgene for this.

Niklas - 29-11-2025 at 12:18

For the mentioned application of making dipropylurea triphosgene directly would most definitely work (http://orgsyn.org/demo.aspx?prep=V86P0315, for the di-tert-butyl).

6dthjd1 - 29-11-2025 at 14:36

did you consider the reaction between Cl2 and CO in sunlight over time?

Brotato - 29-11-2025 at 18:32

Yes phosgene generally gives quantitative yields in the preparation of ureas whereas methods such as the one Niklas posted give around 80%. Propylamine is expensive due to the cold chain shipping thats needed because of the low boiling point so id rather use a reagent that gives that extra 20% yield especially at this larger scale I need since its a starting material for a 7 step synthesis.

The reaction of Cl2 and CO in sunlight is too slow to be practical but CHCl3 + O2 has been implemented on a lab scale with a high power mercury lamp as the UV source !

Radiums Lab - 29-11-2025 at 18:49

As I mentioned in other thread, steam+CCl4 gives phosgene.

Brotato - 29-11-2025 at 21:13

Where are you seeing this method for steam + CCl4 giving phosgene. First of all carbon tetrachloride is stable to decomposition at those temps and if you were to pressurize it, phosgene would react with the water anyways. Additionally the reaction of carbon tetrachloride at 700 C with oxygen is infeasible since it competes with the decomposition reaction: CCl4 --> Cl2 + C. So no, steam + carbon tet wouldn't give you phosgene.

foxofax474 - 29-11-2025 at 21:27

Quote: Originally posted by Radiums Lab  
As I mentioned in other thread, steam+CCl4 gives phosgene.


This is blatantly false? If you're thinking of CCl4 thermal decomposition, this doesn't occur until temperatures are well in the 600-700 degCs.

Niklas - 29-11-2025 at 23:16

Quote: Originally posted by Brotato  
Yes phosgene generally gives quantitative yields in the preparation of ureas whereas methods such as the one Niklas posted give around 80%. Propylamine is expensive due to the cold chain shipping thats needed because of the low boiling point so id rather use a reagent that gives that extra 20% yield especially at this larger scale I need since its a starting material for a 7 step synthesis.


Fair enough I guess, although honestly with a boiling point of 49 degrees it really shouldn’t require cold chain transportation. Even my acetaldehyde (even when purchased from Sigma) came without cooling, the bottle just has to be decent.
And even from TCI half a liter is just 33 bucks, so really propylamine shouldn’t be particularly expensive at all, definitely not that expensive to consider as a „limiting reagent“ even early on in a multistep synthesis.

[Edited on 30-11-2025 by Niklas]

Radiums Lab - 30-11-2025 at 02:40

Quote: Originally posted by Brotato  
Where are you seeing this method for steam + CCl4 giving phosgene. First of all carbon tetrachloride is stable to decomposition at those temps and if you were to pressurize it, phosgene would react with the water anyways. Additionally the reaction of carbon tetrachloride at 700 C with oxygen is infeasible since it competes with the decomposition reaction: CCl4 --> Cl2 + C. So no, steam + carbon tet wouldn't give you phosgene.


There is an old SM thread showing the reaction :
https://www.sciencemadness.org/talk/viewthread.php?tid=252
It is even started in chemical reaction part of this website:
https://www.chembk.com/en/chem/Carbon%20tetrachloride

This reaction is given in our syllabus so I just started it here there are sources too
But this should be performed because this method is not documented properly.
Some places suggest steam temperature to be 500C.

Brotato - 30-11-2025 at 07:33

Ah you omitted the catalyst which is ferric chloride. With this catalyst in the form of a bed in a tube furnace you could probably do this reaction, but from what I see the competing decomposition reaction is pretty unavoidable. So regardless you will get a stream of chlorine contaminated phosgene and maybe soot clogging the flow path within the furnace. At that point you might aswell just use Carbon Monoxide and Chlorine gas which dont require heating to get them into the vapor phase and are far cheaper to acquire than carbon tetrachloride. Also the CO + Cl2 reaction happens at a lower temperature over an even easier to acquire catalyst at atmospheric pressure, so why bother ?

There was a recent paper that Niklas showed me about using an appropriate catalyst (some trichloride of tetrabutylammonia iirc) to perform the reaction of CO and Cl2 all in the solution itself and get phosgene out. That would be easily viable and way more convenient but alas decomposing triphosgene provides high purity phosgene ready for synthetic use at a good price compared to purchasing phosgene solution from even the cheapest suppliers.

Niklas - 30-11-2025 at 07:38

Quote: Originally posted by Brotato  
There was a recent paper that Niklas showed me about using an appropriate catalyst (some trichloride of tetrabutylammonia iirc) to perform the reaction of CO and Cl2 all in the solution itself and get phosgene out. That would be easily viable and way more convenient but alas decomposing triphosgene provides high purity phosgene ready for synthetic use at a good price compared to purchasing phosgene solution from even the cheapest suppliers.


Yeah tetramethyl (or trimethylethyl or probably much of any tetraalkyl) ammonium trichlorides are apparently able to react with CO to make phosgene at room temperature, and while I believe that the paper was about the quantitative rather than catalytic reaction of CO with it as the preparation of these trichlorides is simply reacting the chloride with chlorine I don’t see why it shouldn’t be possible for it to act catalytically.

Niklas - 30-11-2025 at 07:39

Ah here I found it:
https://www.science.org/doi/10.1126/sciadv.abj5186

MaeBorowski - 25-12-2025 at 08:08

I think the best and safest route to phosgene is via diphosgene, which in turn can be made by the chlorination of methyl formate. Dimethyl carbonate is harder to source, requires tet for the chlorination, and can't be conveniently purified by distillation. I'm surprised that no one on the forum has tried it