kilomentor

Friedel Craft Work Up of Aluminum Chloride catalyzed Reaction At Scale

kilomentor | 13 November, 2012 17:53

Larry Fertel asks a question about his Friedel Craft reaction at the

Organic Process Research & Development Networking Group on LinkedIn.

Larry wrote, “I am running a classical F-C reaction based on a process given to us from our customer: benzene derivative, benzoyl chloride (1.1 eq.), AlCl3 (1.1 eq), nitrobenzene solvent, 85 deg C, . i.e standard conditions. When I cool and quench into water and HCl or add water/HCl to the reaction, I am supposed to see 2 layers, then separate the aqueous and go forward with the isolation of the product in the organic layer, etc..

Instead, after the exotherm of quenching, I get massive amounts of solids, the stirrer jams, etc... a real mess. The solids are presumably Al salts, I don't see the nitrobenzene sitting in the flask, it seems to be incorporated into the solids. Note that the reaction goes to completion, no s.m. is seen at all. Is there a "standard" recipe for the workup for this reaction to avoid formation of solids. Note that the customer received the process from their previous manufacturer who is loath to give more details. Also, no time or money to investigate other methods, catalysts, etc..”

From Larry’s description of the reaction methodology I assume the following:

1. The reaction mixture is homogeneous at the end of the heating period. {I assume this because aluminum chloride forms a soluble complex with nitrobenzene. This is the reason for its popularity in F-C reactions. Otherwise it is not a particularly practical solvent since it is high boiling and is usually removed in the end by steam distillation.}

2. When you cool the reaction mixture before quenching it is not a mess yet.

3. You are adding the aqueous HCl into the nitrobenzene solution or slurry. {I assume this because otherwise you would probably have described what happens when a small amount of the quench solution is added, and a little more and so on, with the mixture getting thicker and thicker.}

Larry does not mention details of how he was instructed to do this quench or to what temperature the reaction contents were initially cooled. I think it is very important to keep the reaction mixture very cold during the quench. In fact it is for this reason that a mixture of water/ice and HCl is so often used. Reaction mixtures often thicken so much that wall cooling is probably most often going to be inadequate. If the quenching mixture overheats some hydrolysis of the aluminum chloride to an aluminum hydroxide gel is likely to occur. This I am guessing is giving the mess you report.

Put another way that is to say, it is very important that the solution of aluminum chloride hexahydrate that forms not get warm because the chlorine atoms can be replaced by hydroxyls to give trihydroxyaluminum, which is a gel.

Wikipedia teaching seems to confirm this analysis when it states: “Aluminum chloride is hygroscopic, having a very pronounced affinity for water. It fumes in moist air and hisses when mixed with liquid water as the Cl- ions are displaced with H2O molecules in the lattice to form the hexahydrate AlCl3·6H2O (also white to yellowish in color). The anhydrous phase cannot be regained on heating as HCl is lost leaving aluminum hydroxide or alumina (aluminum oxide) (my italics):

Al(H2O)6Cl3 → Al(OH)3 + 3 HCl + 3 H2O”

Looking for a standard Friedel-Craft acylation reaction with nitrobenzene as solvent, I found the synthesis of methyl naphthyl ketone in Organic Vogel [ATextbook of Practical Organic Chemistry , Vogel, Third Edition, Longmans, pg. 731].
In their procedure HCl is driven off by reducing the internal pressure rather than heating to 85 C as Larry does. The quench is with “an excess of crushed ice”. This suggests to me that so long as the temperature is controlled no additional hydrogen chloride is required although it doesn’t hurt but and reaction mixture must be mixed together with a consistent excess of ice. This is not do-able at scale because adding solid ice cannot be done quickly enough if at all. The quench of the mixture into an excess of ice and enough water to make it stirrable seems a better bet.

I am assume that Larry’s product is soluble in nitrobenzene since the procedure you have been given separates the phases and isolates the product from the nitrobenzene. Probably increasing the amount of nitrobenzene a bit until a solution is worked out will make the experimentation easier. Then when one has something more workable reduce the nitrobenzene back.

General | Permalink | Comments (3) | Trackbacks (0)

 
Powered by LT - Design by BalearWeb