Using Separation Techniques to Get a Short Synthetic Procedure

Potential One Step Synthesis of 6,6-dimethylcyclohex-2-enone

My B.Sc. project at Carleton University in 1967 was the synthesis of 6,6-dimethylcyclohex-2-enone. One of the literature procedures was a one step synthesis from methyl isopropyl ketone and acrolein published by J. Colonge, J. Dreux and M.Thiers, Compt. Rendu. 243, 1425 (1956). The reported literature yield was just 12%.
When I attempted to repeat the procedure the product mixture in my inexperienced hands was either entirely or largely intractable with the volatile portion comprising 11 separate GC peaks, four of which were relatively large. Forty four years later I would like to reexamine how  this synthesis might be reasonably pursued.
The alternative route was six steps long. If the average isolated yield is 80% then the overall anticipated yield would be (0.8)6 = 26.2% for this longer route. The routes are comparable in yield. The time and cost of chemicals would be substantially less for the one step route. The question is can the terrible mixture expected from the acrolein condensation be separated reasonably on scale. Let us set devising a protocol to get clean product out of such a mass as the challenge.
The reaction was performed between methyl isopropyl ketone and acrolein. The acrolein is a neat liquid stabilized against polymerization with hydroquinone. The main by-products will likely arise from the polymerization of acrolein. The trick will be to get the ketone to react with the acrolein before the acrolein is destroyed by polymerization. When the reaction time is over the mixture must be cooled and neutralized. The presence of a free radical inhibitor such as galvinoxyl that is not interfered with by the methoxide base would seem wise.
The working out of the most preferred ratio of reactants and the most preferred mode of bringing them into reaction is a separate question and not the key one I would say. Critical is how are we going to separate the terrible mixture? The good news for this synthesis is that the poor yielding step comes at the beginning of the sequence- in fact the reaction constitutes the entire process. One will not be carrying along material that eventually becomes part of waste after numerous process stages. The waste will be formed quickly and not have extra work expended upon it!
The mixture will contain the degradation products of acrolein and sodium methoxide and the free radical catalyzed polymerization of acrolein to give CH2=CH-CO-[(CH2)2CO]n-(CH2)2CHO.The dimer of acrolein would still remain. The products of Cannizaro reactions involving the methoxide acting as a reducing agent are also possible. The main characteristic of the by-products would be that they are fairly high molecular weight. It might be desirable to quench with a mild reducing agent that will only hit aldehydes. The desired product is in fact distinguished in that it does not have aldehyde. Once the kill solution has been added steam distillation is likely to separate volatile materials from polymers. The desired alpha-beta unsaturated ketone if formed woulld be in this distillate. The main constituent of the steam distillation will be starting methyl isopropyl ketone. The recovered starting material can be separated from desire product by reaction with diethylamine. Only the alpha beta unsaturated carbonyl will react with the base by Michael addition and the product will become soluble in weak aqueous acid. The unreacted methyl isopropyl ketone will remain as the organic phase.
Upon back extracting the neutralized base into an organic layer it can be directly methylated. With base the quartenarized amine may be lost to return the unsaturated ketone. Unreacted beta amino compound is separable again by acid-base extraction.