[curiouscat]

I am debugging a reaction at plant scale where there's an acid catalysed (H2SO4 95%) aldol condensation of  and .

Typically the major product is (Product a) . Greater than 99.5% typically.

It so happens that something has changed that has now caused increasing amounts of  ( Product c) to start appearing. The amounts are still small, about 1 to 1.5% but that's still a cause of worry.

Any tips in general what are the sort of parameter deviations that one might expect to boost the selectivity of ( c )over  ( a )? If Reaction temperature has wandered higher than the norm would you expect that to increase (Product c) levels?

Another point I'm suspecting is the catalyst acid, H2SO4. A supplier was recently changed and our sourcing analysis is a bit lacking. Could there be some impurity in the H2SO4 that might cause this?

Just brainstorming for ideas at the moment! Any tips?

[orgopete]

Can you return the product distribution to the 99.5%? I don't think the temperature should cause the formation of the by-product, however I can also think how it might. I am assuming the product is the thermodynamic product. In this case a higher temperature might trap an intermediate kinetic product. This is what I would do, look at the time course of the reaction. Inevitably, the products must form at different rates as they require different intermediates. If so, then a higher temperature may trap out one of the intermediates.

By the way, are the structures correct?

[discodermolide]

I would go for the sulfuric acid. Check the SO₃ content? Check for metals, Fe, Ni etc.
Make sure it meets the specs of the previous material.

[curiouscat]

I would go for the sulfuric acid. Check the SO₃ content? Check for metals, Fe, Ni etc.
Make sure it meets the specs of the previous material.

Thanks @disco.  That's a good idea. Analysing the sulfuric now.

Unfortunately, I don't think the previous material had specs. laid on these contaminants. QC was basically only checking for acid strength, color & some pretty gross indicators so far. Now that we've hit a problem I expect protocols to improve.

Since there's virtually endless impurities one could quantitate for, any idea what a catalyst acid should typically be assayed for?

[curiouscat]

Can you return the product distribution to the 99.5%?

That's the goal. Haven't succeeded yet. Success at doing that means identifying this problem.

By the way, are the structures correct?

I think so. Which one are you suspecting wrong? The undesired impurity is 4-hexen-3-one.

[discodermolide]

Any material left from the previous supplier, even a lab sample? Then you can compare.
I would check for metals, SO₃, even SO₂ and of course acid titration.

[orgopete]

Since the impurity is small, I didn't know if it's source could have been due to an impurity in one of the starting materials. However, an impurity may give a different product. So I didn't know how the impurity was identified, or whether it could be something else.

Re: Disco's suggestion, I guess I wondered whether you could just run the reaction in the lab with the new vs reagent sulfuric and see what you got. If they give different ratios, then I agree with Disco, if not, the I suspect the starting materials.

[curiouscat]

Since the impurity is small, I didn't know if it's source could have been due to an impurity in one of the starting materials. However, an impurity may give a different product. So I didn't know how the impurity was identified, or whether it could be something else.

The impurity is nothing exotic, merely the other crossed aldol. Instead of abstracting an alpha-H from a sec. C it is formed from the pri. C.

So basically its only a selectivity modification.

It was identified by GC-MS & then by injecting a standard & matching GC peaks.

[kriggy]

Also was the compound identified multiple times or was it jsut from single batch? It also might be something like short blackout that would for example stopped the heating of the reactions for short time or something like that

[curiouscat]

Also was the compound identified multiple times or was it jsut from single batch? It also might be something like short blackout that would for example stopped the heating of the reactions for short time or something like that

No, it wasn't a one off detection. Multiple batches. The trend seems clear.