Skip to content

PIDA cyclisation #97

Closed
alintheopen opened this Issue Oct 24, 2013 · 41 comments

6 participants

@alintheopen
OpenSourceMalaria member

Poor yielding and very hard to visualise the product or SM for that matter.
Have tried using KMnO4, Vanillin, Ninhydrin, and Goofy. Will update with more dips once I've made/stolen/borrowed some. Any ideas?
Product and SM very insoluble in DCM, CHCl3, MeCN, EtOAc, even MeOH so this could be another problem.
PIDA cyclisations typically in DCM, fluorinated alcohol solvents, sometimes DCM.
Ideas to improve reaction:
Use of MeCN as solvent - currently being tried
Use of HFIP or TFE as solvent
Swap to more reactive PIFA
Heat reaction? CRO synthesis claims 38% at rt.

@JoieG @PatrickThomson @mattodd

@JoieG
JoieG commented Oct 24, 2013

Hi Alice, I wouldn't worry about the stain since the SM and Products should both be UV active. Upon completion are you working up the reaction or concentrating it to then purify (I was unsure from your notebook)? I would make sure to work up the reaction - dilute with DCM and wash with saturated bicarbonate and brine. This should make it easier to handle the crude product and cause less problems in the purification. If it isn't soluable in DCM or EtOAc you can try 2Me-THF as a solvent in the extraction. It solubalizes almost anything and seperates from AQ layers.

@PatrickThomson

My students put on an opportunistic overnight stir on 100mg of the nitrile SM with PIDA, but about 10x diluted (to 5ml from 0.6) and they've said the NMR/tlc has completely changed - I'm meeting with them now so will get data looked at and uploaded.

@alintheopen
OpenSourceMalaria member

Edited as I posted as @newgitontheblock by accident needed to clarify UV vis products.
Awesome, I think the dilution will do it. Every other PIDA reaction that I've done worked much better at greater dilution. Joie, I initially did no work up as the CRO prep had none but I worked up the reaction from today. Had some issues updating the lab book but will do tomorrow when the server is working again. I know prod and SM should be uv active but my tlc said otherwise....despite the fact I had product (albeit very little) by NMR....weird! Looking forward to seeing Devon and Eduvie's tlc - hope they have their target.
Sorry, just to clarify I couldn't see the pyrazine very well in the UV, could see hydrazine, could just about see imine, but couldn't see cyclised product...perhaps just because my first attempt didn't work so well. Will check the more dilute reaction in the am. See ELN for TLCs ;)

@alintheopen
OpenSourceMalaria member

Here is a PDF of the product NMR for reference @PatrickThomson http://malaria.ourexperiment.org/data/7703.html

@PatrickThomson

Extremely helpful - we've got that showing up in our crude NMR alongside what I think is iodobenzene and possibly some SM as well.

@alintheopen
OpenSourceMalaria member

Great, yes. You should see iodobenzene, which will fly off the column and I wouldn't be surprised if there was some SM (the CRO reported 38% yield). I was rather hoping to improve this yield a bit if possible.

@mattodd
OpenSourceMalaria member
mattodd commented Oct 24, 2013

(newgitontheblock = alintheopen in case that's not obvious)
Joie - the TLC behaviour is most odd. So are we all using crude NMR spectra here - i.e. that works? i.e. could be followed that way?

@JoieG
JoieG commented Oct 24, 2013

First suggestion, would be to change the HNMR solvent. It does not look like the material is that soluble in CDCl3. You may be able to pick out the peaks better and understand the by-products, if you take a small aliquot and take the HNMR in DMSO. You should be able to monitor dissapreance of the imine proton this way as well, but it looks like you are getting some product! Second question, are you analyzing the TLC in both long and short wave? Is there a difference in the spot intensity with one or the other - do you see anything? Do you have an LCMS available to use - this could help you understand product formation?

@PatrickThomson

We're going to have difficulty playing with the NMR in Edinburgh a lot due to a major facility upgrade that's happening from now until some time in November.

@sdebbert

If PIDA is giving you trouble, I found a ACIE paper that has better luck with Chloramine-T as the oxidant in a similar reaction:
DOI: 10.1002/anie.201001999
They also use 2-MeTHF, as mentioned earlier, to good effect in their workup. Almost all the NMR data in the supporting info is in DMSO-d6.
I also found a paper where they use PIFA in a solvent-free reaction: DOI: 10.1007/s10593-012-0899-0 . Needless to say, I'm a little more skeptical of that one (if only because I've had bad luck with a mortar and pestle outside the context of making pesto), but if it were legit that would be fantastic.

@alintheopen
OpenSourceMalaria member

Thanks @sdebbert! That's great...I also agree with you about the P&M. I think the PIDA is working OK now, it was a dilution issue but if the yield doesn't improve I'll try the prep you cite. Now onto the cyanides...just waiting for a new bottle of TMSCN. Cheers everyone.

@alintheopen
OpenSourceMalaria member
@mattodd
OpenSourceMalaria member
mattodd commented Oct 29, 2013

File has nothing in it? (zero bytes)...

@alintheopen
OpenSourceMalaria member

apologies...I am having a few problems with 'empty' uploads....this should work... My NMR: http://malaria.ourexperiment.org/data/7739.html, CRO NMR: http://malaria.ourexperiment.org/data/7703.html

@mattodd
OpenSourceMalaria member
@alintheopen
OpenSourceMalaria member

1g scale synthesis led to a whopping 12% yield of the product....although of course both previous reactions weren't purified (as per the CRO experimental) so its probably a tad higher. Still not amazing, so I'll try to optimise this and perhaps Eduvie and Devon can try this too @PatrickThomson whilst waiting for SM. It is important to note that the Rf is more like 0.6 rather than 0.1 as reported in the CRO experimental.

@JoieG
JoieG commented Oct 30, 2013

@alintheopen If there is left over AcOH from the reaction than this can potentially change the Rf of the product. What are your ideas for reaction improvement?

@PatrickThomson

@alintheopen Yes, Devon found that a 100% EtOAc column ran off product and major impurities in the same first few fractions - we're slowing it down to get a small reference sample.

The way to go forward with this that I'd suggest is to find a bunch of alternative conditions, and anything where I have material available in-house (thanks to Edinburgh's chemical inventory manager) we'll try, or anything with cheap reagents that we'll get delivered quickly. Devon has a gram of the hydrazide available for 100mg-scale test reactions, and we can get more.

@PatrickThomson

@JoieG @alintheopen we found that the Rf was more like 0.6 - after a basic workup. We ran the fractions off the top of the column.

Question is, is that low yield from side-products, incomplete reaction, or E/Z discrimination? If it's the latter, can we equilibrate the isomers in-situ?

@PatrickThomson

Just a quick update - we did a quick reaction with Dess-Martin Periodinane and got nothing back. Going to try both conditions from @sdebbert earlier post - Chloramine-T, PDF and mechanochemical with PIDA pdf

@mattodd
OpenSourceMalaria member
mattodd commented Nov 21, 2013

Sounds good, Patrick.

@PatrickThomson

Oh man, the Chloramine-T reaction has certainly done something appropriate... It looks like the elevated temperatures might even cause hydrazone "epimerisation" - I'm literally shadowing a TLC plate now.

@PatrickThomson

We'll get a mass yield tomorrow but it looks like chloramine-T has done the job - eyeballing the NMR, there's maybe >50% product, a few other misc. aromatic peaks possibly the same C-cyclisation product Alice found.

@PatrickThomson

Ok, the crude reaction mixture evaporates to this: 1:0.7 of product to mysterious other thing that has a peak at 8.6 and some other peaks, possibly C-cyclisation as opposed to N-cyclisation, or possibly unreacted hydrazone isomer. image

@PatrickThomson

The two large peaks at 7.6 are an unknown contaminant present in other NMRs - I suspect the undergraduates have gotten into our chloroform. The doublet at 7.3 appears to integrate 2:3 with the chloramine-T methyl peak at 2.46. So in here, then, we have product, and one of the hydrazone isomers and absolutely nothing else. We'll try increasing the reaction time, and possibly some acid or base to labilise the hydrazone.

@mattodd
OpenSourceMalaria member
@PatrickThomson

We're finding that Chloramine-T is still leaving us with a mixture of one of the hydrazone isomers (presumably Z) and product, but much cleaner than with PIDA. Solubility is higher, etc. Running it overnight didn't make any improvement, running it with acetic acid didn't make a difference, and we're running it with TFA now to see if we can coerce the hydrazone to "epimerise" and give us a form of Dynamic Kinetic Resolution. However, an old paper seems to imply that heavy metal oxidants such as lead tetra-acetate will cause the desired effect - every other literature hit about this oxidation uses the E-isomer and gets >85% yields, whereas ours are slightly favouring Z it seems.

@PatrickThomson

acid addition seems to hinder the oxidation, looks like we might have to try the distinctly un-friendly lead conditions.

@mattodd
OpenSourceMalaria member
@PatrickThomson

Unfortunately I'm visiting family then. I'll prepare a slide and upload it tomorrow, but it looks unlikely that I'll have enough technology with me on holiday to commit to attending.

The labbook seems to be down? Or at least, inaccessible from the UK. In any case, I can't see the agenda.

@PatrickThomson

Single slide about my oxidative cyclisation stuff - the Amide work is basically at the "we've synthesised a bunch of starting materials" so nothing interesting yet.

@mattodd
OpenSourceMalaria member
@sabinllm

@PatrickThomson Could you send me the best protocol you have for the cyclisation step?

@PatrickThomson

2-methyl THF (15 ml/mmol), chloramine-T (1.12eq), 60°C (lit, or 75 ours due to oil bath drifting), 1.5 hours. Lab book here but I've just noticed the millimole quantities for Chloramine is wrong consistently - I'm chasing it up now.

This will only cyclise the E isomer of the hydrazone, and do it cleanly. the Z is a continuing problem (will be tackled here on friday)

Edit: Original paper provided by @sdebbert in this comment

@PatrickThomson

Commenting here to get it on paper, but this is important: The main conclusion from Devon's project is that the Z-isomer of the hydrazone will not react under reasonable conditions, whereas the E-isomer will cyclise cleanly with PIDA, and even cleaner with Chloramine-T. In most general cases of hydrazone formation, they are E-selective so this isn't a problem. However, our pyrazines generate a mixture of about 0.8:1 E:Z, resulting in low yields. My suggestion would be to form the crucial pyrazine-N bond with cross-coupling on a terminal hydrazone, as the formation of the terminal hydrazones are pretty E-selective.

@mattodd
OpenSourceMalaria member
@PatrickThomson

Yes, Devon wrote this up in her project report - getting my hands on an electronic copy of this is actualy proving far from trivial, as they have both gone into full-on "exam mode" and stopped answering twitter. We have NMR evidence that there's product and unreacted Z isomer, and the different oxidative conditions or addition of acids to the reaction mixture is unhelpful - a failure to induce Dynamic Kinetic Resolution. We haven't looked at changing the selectivity of the hydrazone formation itself.

@mattodd
OpenSourceMalaria member
@PatrickThomson

After some (internal to my own head) wrangling, I can share the reports out. Because I'm not the author, I can't post them to our repo or figshare, but I can put copies here. I'm not sure how useful it'll still be 2 months on...

https://drive.google.com/file/d/0B6rAjqmPhHy5NkRLaWx3ZEtpUjg/edit?usp=sharing

https://drive.google.com/file/d/0B6rAjqmPhHy5VDltWWZPMk1tUFk/edit?usp=sharing

@mattodd
OpenSourceMalaria member
mattodd commented Jun 29, 2014

Hi Patrick - this is great. So these are PDFs that we need to make sure are available. Presumably there's no problem with posting them to the ELN as "shared data"? And linking to them out of the wiki? Ideally the contents would be incorporated into the wiki in advance of incorporation into a Series 4 synthesis paper. Question on raw data - what % of the raw data do you think is on the ELN? Is there a way of zipping up everything and putting it somewhere, so you think?

@mattodd
OpenSourceMalaria member
mattodd commented Dec 10, 2014

Hi @PatrickThomson I uploaded the reports to the ELN here:
http://malaria.ourexperiment.org/osdd_malaria_shared/11335
When you've time it'd be great to incorporate any conclusions into the "synthesis" section of the Series 4 wiki, linking back to data. There may be nothing, but it'd be useful to fold in anything concrete, in preparation for writing the Series 4 paper.
This Issue is already linked from the wiki, so now closing.

@mattodd mattodd closed this Dec 10, 2014
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.