[AlphaScent]

Hey All,

I have been working on the diepoxidation of a simple, aliphatic terminal diene.  I am having issues with either reactivity or selectivity.  That is the low reactivity of the starting diene vs the high reactivity of the diepoxide product.

1,7-octadiene to 1,2,7,8-diepoxyoctane.  This seems like a straight forward reaction but my problem is I need to do this in a 1500L reactor and scalability of the process is paramount.

The literature shows mCPBA to work well but that cannot be used here on scale.  Various reasons, but cost being the most prohibitive.

I have tried various conditions from 32% peracetic acid in a Na2CO3/DCM and in situ trifluoroacetic acid from sodium percarbonate and trifluoroacetic anhydride.  There are some interesting Heme like catalysts that I have tried that give excellent conversion but poor selectivity due to the acidic conditions (MnCl2/2-picolinic acid catalyst/AcOOH and Fe(NO3)3/1,10-phenanthroline/AcOOH).

Also tried some racemic Shi type reactions with hexachloroacetone and acetone.  My next step with the Shi type is to try trifluoroacetone.

Also tried some Nyori H2O2/sodium tungstate chemistry that also proved futile.

I am just at a road block and if anyone has an idea or has experience with this type of chemistry, any insight is greatly appreciated.

Be well!

[rolnor]

Have you searched online; Scifinder, Reaxys? This seems like something thats been done.

[rolnor]

You can buy this so there is probably a good method already:

https://www.sigmaaldrich.com/SE/en/product/aldrich/139564

[AlphaScent]

Oh yes, I have searched both in detail.  The method found is mCPBA and is not something that is scalable for us due to the price point.  Even in the patent literature I cannot find anything. 

The only other method is a Mn(ClO4)2 catalyzed epoxidation, and we cant use perchlorates on scale. Chem. Eur. J. 2008, 14 7788.

I too thought this should be fairly straight forward, but the compound is much more reactive under acidic conditions than its monoepoxide cousin 1,2-epoxyoctane.  I would have thought the reactivities would have been similar.  I was wrong, haha.

I have some other methods I am trying (cumene hydroperoxide/Mo(CO)6, another in situ trifluoroperacetic acid method, methyltrioxorhenium/3-cyanopyridine/30% H2O2).

Also will be trying to repeat a lit prep using mCPBA just to see of the 95% yield is possible. 

Thanks Rolnor, Ill get back to the thread when I get some results.

[rolnor]

OK, that sounds good. Have you thought about mono-perphtalic acid?

[AlphaScent]

Not the mono, but magnesium  monoperoxyphthalate (MMPP) was something I thought about.  It has its advantages over mCPBA in that it seems to be cheaper, but commercially it is sold as a hexahydrate.  The MW is 494.64.  The atom economy is pretty terrible.  Wikipedia has a lit ref (Tetrahedron, 2009, 65(14) 2773) and they use 1.1 eq of the MMPP in 15 mL of moist (100 uL) DCM on a 1 mmol scale.  The solvent amount can be reduced on scale, but the amount of MMPP may be gargantuan....lets see what it is:

So on 100 KG scale:

diene: 100 KG, 907.4 mol.
MMPP: 2.2 eq. (1.1 eq. to olefin) 1996.4 mol, 987.5 KG
DCM (1 M): 907 L
Water: 82.5 L

Sigma Aldrich price is ~$400/KG for MMPP, could get that down to maybe $50 KG buying by the MT.  The atom economy is still quite poor.

Maybe if an in situ method could be developed from phthalic anydrude and H2O2, that could be cheaper.

Think it is possibly a thought to look at and price out.  Running mCPBA today/overnight.  If it works well, it may be something to look into.

I am trying Mo(CO)6/cumeme hydroperoxide afterall.  Cumene hydroperoxide is not very atom economical either.

Thanks for the brain stimulation rolnor!

[AlphaScent]

Of course....mCPBA works beautifully in 3 hours at room temp.  Will report back on yield tomorrow.

[rolnor]

Great!
Large-scale synthesis has other challenges than lab-scale. Its not very easy to find a solution sometimes. I understand that H2O2 would be better considering its low mw and low price.

[AlphaScent]

Ran a 5.5 g scale reaction using 2.7 eq. of mCPBA in DCM (0.5M) and obtained 97% crude yield by NMR.  DCM and MTBE present after rotovap.  Did not distill.  Looking at the cost of mCPBA, it just wont work for what the client wants to pay.  It should also be noted that they originally use peracetic acid.  Adding it to a suspension of sodium carbonate and DCM/diene (4M).  What do you think happens when you add peracetic acid (32% in acetic acid) to that? hahahaha...yeah the engineers lovveeee all that frothing ; )

[rolnor]

Hi, hi, not good…
Can you use calcium hydroxide as base maybe to avoid foaming?

[AlphaScent]

I have thought about that and it may come to that.  The problem is the conversion.  With peracetic acid you get a mixture of diene, monoepoxide, and diepoxide.  Then have to distill and recycle the diene and monoepoxide.  So we are looking into ways to get complete conversion (w/ good selectivity, epoxide ring opening happens readily under some of the more acidic conditions...duh! haha).

But that is a good idea.  I also buffered with sodium acetate and got similar yields and conversion.  The other issue under both conditions is you get 10% side product.  Addition of acetate ion to the epoxide (various regioisomers).

[rolnor]

Yes, acetate is a nuchleophile, it opens the epoxide. You can make a peroxide from acetone with Oxone that is very powerfull epoxidizer but that is probably dangerous on this large scale.
https://www.organic-chemistry.org/chemicals/oxidations/oxone-potassiumperoxomonosulfate.shtm

[MOTOBALL]

Can you use N-bromo/chloro-succinimide in aqueous conditions to form the di-halohydrin?
Then action of base to form the di-epoxide.

Regards,
Motoball

[wildfyr]

Motoball that is similar concept to how glycidyl ethers are made.

Epichlorohydrin reacts at the epoxide with an alcohol, then you do elimination of the chloride with strong base to give closed epoxide.

Alpha can you elaborate more on your failure modes? Good conversion by poor selectivity? But you are trying to epoxidize both double bonds, so what is the selectivity issue?

[AlphaScent]

Motoball,

I have thought about doing that too.  I used to make butadiene monoxide this way from butadiene.  Works very well.  The issue here is that it is a two step process to first from the bromohydrin and then close the ring in a second step.  Too much time and material.

Wildfyr,

I mean selectivity between diol and epoxide.  I have no proof of the tetraol being made other than mass balance.

By conversion I mean that all of the diene reacts and is gone by GC-FID.  Then the reaction is complete I work up and isolate to give only 35-55% yield.  The product is clean (sans minor MTBE solvent).  I know I am not losing it in the rotovap because I check the distillate and I have done an mCPBA reaction that worked great.  Problem is mCPBA is too expensive.

So I have been playing with cat MnCl2*4H2O (1 mol %, added dissolved in 1 mL water) this week.  Doing the reaction in 4:1 DCM:MeCN (4M) with various equivalents of 25% NaOH and 4.0 equivalents 32% peracetic acid.  When using 2.0 eq. of 25% NaOH is used I was afforded 48% yield of clean product.  The problem is too basic and it adds and too acidic and it adds. 

I am doing a cat. Mn(OAc)2/4:1 DCM:MeCN (2 M)/ solid Na2HPO4/PAA now at room temp.  We shall see....

I also found Methylrhenium trioxide works great, but its just crazy expensive even at 0.5 mol% loading!

Cheers Yall