Chemical Forums

Chemistry Forums for Students => Organic Chemistry Forum => Organic Chemistry Forum for Graduate Students and Professionals => Topic started by: Babcock_Hall on March 06, 2013, 08:30:04 PM

Title: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 06, 2013, 08:30:04 PM
Starting with dibenzyl phosphonate, I added LiN(TMS)2 at -75 °C, then TMSCl at 0 °C.  At this point I have an impure mixture by P-31 NMR that includes (PhCH2O)2P(OSiMe3), which is a tervalent phosphorus compound that is my nucleophile.  I usually observe a mixture at this point with related compounds, but sometimes it is a little cleaner.  So far this is all very close to published protocols.  Then I add para-formaldehyde that was dried using a vacuum pump, and not much seems to happen, or it is happening very slowly, again monitoring by P-31 NMR.  I do see an increase in area of a peak near 0 ppm that could be my desired product, but after 5 hours the reaction is less than half complete.  The desired compound at this point is (PhCH2O)2P(O)CH2OSiMe3).  Next would come removal of the TMS group to provide the alcohol.

When we do this with tert-butylphosphonate, we run the last reaction at 45 °C for a few hours.  We have tried making the dibenzyl version only twice before, and it was a long time ago.  The first trial was around the same temperature, and the second trial, the temp was not recorded.  I just checked the student's notebook from 13 years ago to refresh my memory.   My recollection is that the first trial did not go well, but then I thought that we dropped the temperature of the PFA reaction to room temp.  That is why I tried room temperature today.  I am going to let it stir overnight, but then raise the temperature.  I was thinking that I might add more PFA.  However, PFA does not look completely soluble in THF, my solvent.  It occurs to me that raising the temperature may make it more soluble.  Any thoughts?
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Dan on March 07, 2013, 06:32:04 AM
So the ultimate desired product is dibenzyl (hydroxymethyl)phosphonate:

P(=O)(CO)(OCc1ccccc1)(OCc1ccccc1)

You can just heat dibenzylphosphonate, paraformaldehyde and triethylamine solvent free. No need for LiHMDS. You have to give it a good kick in the balls but it works pretty well.

Example: http://www.sciencedirect.com/science/article/pii/S0960894X99005442
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 07, 2013, 09:48:32 AM
Thank you very much.  If this reaction goes awry, I will try the conditions you suggest.  Right now the reaction mixture is slowly changing by P-31 NMR, but there is still some of the phosphite present.  I may heat it up and see what happens.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: discodermolide on March 07, 2013, 10:16:23 AM
As Dan said you don't need LiHMDS. I used to use just hexamethyldisilazane at reflux, this is enough to convert the P(V) to the P(III). then you can follow it nicely by 31P-NMR, then add your paraformaldehyde and heat it  follow by NMR.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 07, 2013, 11:09:46 AM
I might have enough of the H-phosphonate for a small test reaction under the conditions of the 1999 paper, and I plan to order more of this reagent.  I have to admit that the other protocol looks much simpler and faster.  However, the reaction with about 24 grams of dibenzyl phosphonate and the other components (through para-formaldehyde) is continuing to react, and I need to do something with it.  It is a mixture by P-31, but one of the peaks is close in chemical shift to what my student observed.  Either it should be worked up, or it should be tossed.  BTW, I found in further reading of his lab notebook that he synthesized it a third time, and the temperature of the PFA addition was 25 °C, so apparently my memory was not faulty.

discodermolide,

The compound you made with hexamethyldisilazane was of the form (RO)2P(OTMS), is that correct?
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: discodermolide on March 07, 2013, 11:53:23 AM
Mine were phosphinites of the type
CCOP(O[Si](C)(C)C)C(OCC)OCC
Where we would alkylate the P(III) then hydrolyse the ketal to expose the other P-H for further alkylation.
I also did this chemistry with.
C[Si](C)(C)OP(CCCN([Si](C)(C)C)[Si](C)(C)C)O[Si](C)(C)C
generated it from
NCCCP(O)([H])=O
With refluxing HMDS then reacted the P(III) with paraformaldehyde at reflux. We published this in J. Med. Chem. back in 1995 or so. Phosphinic acid analogues of GABA parts 1 & 2

Just remembered the, exact reference is in the list of publications I posted. Unfortunately I mis-laid my copies.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: discodermolide on March 07, 2013, 01:31:22 PM
As an addendum; you should bbe able to take dibenzylphosphonate and triethylamine and heat it with paraformaldehyde.
By the way I found the papers as pdf's
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 07, 2013, 05:17:19 PM
discodermolide,

I will try the triethylamine reaction before too long, but I used up essentially all of my starting material on this last reaction.  It looks promising, though.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 08, 2013, 09:50:01 AM
Can para-formaldehyde go bad over time?  I just realized that my bottle is quite old, at least 10 years.  This compound and the starting H-phosphonate are common to both synthetic routes.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: discodermolide on March 08, 2013, 11:11:18 AM
I think paraformaldehyde is reasonably stable. You could always sublime it.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 21, 2013, 06:10:38 PM
I tried heating dibenzylphosphonate with paraformaldehyde and TEA (10 mole percent) for about 5 minutes at 130 °C.  There was a product at roughly 25 ppm that may be the desired product, but the reaction was less than 50% complete, and there were also side products.  I am thinking about running a time course on a slightly larger scale.  I don't presently have any hexamethyldisilazane, but I can order some.  Any thoughts?
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: discodermolide on March 22, 2013, 02:29:06 AM
Perhaps use triethylamine as the solvent? The BPt. is about 90°C and heat it overnight.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on March 22, 2013, 01:01:30 PM
Reflux it, in other words.  That sounds good.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on June 12, 2013, 07:58:31 PM
As per Dan's suggestion, I tried heating about 0.2 grams dibenzyl phosphite with 1 equivalent of PFA and 0.1 equivalent of triethylamine for 5 minutes at 123-127 °C.  The only change I made was to use a different batch of para-formaldehyde than I used in my previous attempt.  It is more powdery, whereas the other batch was in the form of small granules.  I saw a major peak at 25.5 ppm in the P-31 NMR, plus less than 10% of what might be the starting material.  This is looking much more promising than my previous reaction.

I also tried the same conditions except I used an impure preparation of di-tert-butyl phosphite.  After a couple of minutes I noticed some gas evolution that might have been isobutene.  I think that a pure preparation might have done better, but it could be that the heat just too much for the tert-butyl groups.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on June 14, 2013, 05:50:30 PM
I tried one more small test reaction with dibenzylphosphite, PFA, and TEA, using pelleted para-formaldehyde that I ground in a mortar and pestle and dried in vacuo.  By P-31 NMR it looked a little better than the sample without grinding, but not great (maybe a longer reaction time would have helped; there seemed to be unreacted starting material).  However, I scaled up using a different batch of PFA (one that was already powdered and that I had also dried), and it worked pretty well.  I will probably run a silica column on it on Monday.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on October 08, 2014, 10:12:14 AM
My student would like to make di-tert-butylhydroxymethylphosphonate, (t-butylO)2P(O)-CH2OH.  He found a procedure in the patent literature that suggested using microwaving for 20 minutes (90 °C) in the presence of 0.1 equivalents of triethylamine and an excess of para-formaldehyde.  Based on previous experience (mentioned upthread) I am a little bit worried about loss of the tert-butyl protecting groups, but I told him to go ahead on a small scale.  If it fails, we have previously used a route involving Michaelis-Becker chemistry.  It is unclear to me how the temperature is known in this procedure, inasmuch as it is done in a sealed vial.  Although others in our department do, we don't have any experience with microwave reactions, so any suggestions would be appreciated.

Reference
Synthesis of high relaxivity chelates for potential use as MRI contrast agents
Caravan, Peter D.; Jacques, Vincent; Dumas, Stephane; Sun, Wei-Chuan; Zhou, Kevin L.; Schmitt-Willich, Heribert
Assignee Epix Pharmaceuticals, Inc., USA
2008
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Dan on October 08, 2014, 05:19:19 PM
It is unclear to me how the temperature is known in this procedure, inasmuch as it is done in a sealed vial.

Our microwave detects the temperature and pressure in the vessel. You can set it to maintain either a certain pressure, temperature or wattage (these parameters are displayed graphically in real time as the reaction runs). Presumably there is a probe that goes through the septum - the mechanism is not accessible so I can't actually see what it does. I've only used it a couple of times, usually attempting to repeat literature reactions from India that used domestic microwave ovens (without success).

You can also just heat it normally, it will still work but may take longer.

Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on October 08, 2014, 07:52:31 PM
We obtained a mixture of two closely related compounds plus quite a bit of starting material.  I think a small fraction of the starting H-phosphonate lost a tert-butyl group.  Some of the para-formaldehyde ended up on the side of vessel, by the look of it.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on October 12, 2014, 12:39:06 PM
You can also just heat it normally, it will still work but may take longer.
We tried this once with an impure sample of HP(O)(OtBu)2, but the starting material appeared to decompose.  Yesterday, we tried our standard method of making the phosphonate, using LiN(TMS)2, followed by TMS-Cl, followed by para-formaldehyde and finally removal of the TMS protecting group.  The reaction with PFA is very sluggish, as monitored by P-31 NMR.  We usually see a complete reaction in 2.5 hours at 45 °C, but in this case even after the reaction has run overnight, the starting material, (OtBu)2P(OTMS) is still the dominant peak.  It is hard to come up with a mechanism for why a reaction is proceeding more slowly than it usually does.
Title: Re: para-formaldehyde reaction is unexpectedly sluggish
Post by: Babcock_Hall on November 05, 2014, 09:01:59 AM
We bought some new paraformaldehyde, and the formation of the P-C bond worked much more smoothly:  (tBuO)2POTMS + PFA  :rarrow: (TBuO)2P(O)CH2OTMS.  It is still unclear to me what is the matter with our older PFA samples, and what we can do to ensure a smooth reaction in the future.  In theory we could always buy a new bottle, but that gets expensive.