B.R.S.M. Yield isn't everything


‘No Added Metals’ Is The New ‘Metal Free’

Base image from http://mrsec.wisc.edu/Edetc/nanolab/photonic/index.html

Did anyone else see that paper on Thursday in Chem. Comm. titled "Use of Dimethyl Carbonate as a Solvent Greatly Enhances the Biaryl Coupling of Aryl Iodides and Organoboron Reagents without Adding Any Transition Metal Catalysts", and think "here we go again"?[1] I immediately though it kind of appropriate that Chem. Soc. Rev. very recently published a history of transition metal contaminants in catalysis (DOI: 10.1039/C2CS15249E). However, on reading the Chem. Comm. paper, it seems the authors were very careful both to check all their reagents, and not make any grand claims. Not surprising, really, given the numerous examples of misunderstanding of such results in the literature. Even the title is cautious, saying 'without adding any transition metal catalysts', quite a step down from the bold claims of 'transition metal free' reactions seen in the literature of a decade or so ago.

A classic (and highly teachable) example is that of the Leadbeater group's 'Transition-Metal-Free Suzuki-Type Coupling' which began in 2003 with a triumphant Angewandte paper (DOI:10.1002/anie.200390362), and ended in 2004 with a slightly more subdued J. Org. Chem. paper (DOI:10.1021/jo048531j) when it turned out palladium in the sodium carbonate was present in large enough quantities to be catalysing the reaction.[2] The problem is that palladium is fantastically good at effecting coupling reactions, even in what people irritatingly call 'homeopathic quantities'.[3] Obviously, when doing a transition metal catalysed reaction, or, more importantly, publishing one, it's pretty important to make sure that the metal written on the bottle of reagent you're using is actually the one that's doing the magic. It's good to see that people are being a lot more careful these days.

It's not all bad, though. There are several example of reactions and methodologies being developed from investigation of trace impurities. One of the oldest and best known examples of this lead to what we now know as the Nozaki Hiyama Kishi reaction.[4] This chromium(II) catalysed coupling of vinyl and allyl halides with aldehydes is a useful and very mild coupling widely employed in total synthesis, especially for medium sized rings.[5] I've even done one. It was originally reported by Nozaki and Hiyama as a synthesis of homoallylic alcohols, but was found by Kishi and others to be a touch capricious with vinyl iodides. I'll let Kishi tell it (J. Am. Chem. Soc, 1986, 108, 5644):

"Unlike the Cr(II)-mediated coupling of allyl halides with aldehydes, the success of this coupling mysteriously depended on the source and batch of CrCl2... These facts naturally suggested an intriguing possibility that the success of this reaction might depend on some unknown contaminant in CrCI2. For this reason, we have examined the effect of transition  metals for the Cr(II)-mediated coupling reaction and found that NiCl2 and Pd(OAc)2 have a  dramatic effect... It is now possible to achieve the coupling using CrCl2 from any source with excellent reproducibility"

The secret? Just a touch of Ni (or Pd) to help with the formation of the organochromium species as these metals undergo oxidative insertion into the carbon halogen bond more readily, and then transmetallate with chromium:

Thanks, wikipedia.org

Important not to add too much, though, as more than 0.1 - 1.0 wt% in the CrCl2 causes problems with dimerisation of the vinyl iodides.

More importantly, but less interestingly for organic chemists, trace nickel impurities also played a key role in the history of Ziegla-Natter polymerisation, a Nobel prize winning reaction of incredible industrial importance. But that's another story...


Addenda and Random Musings

1. Actually, a couple of things prompted me to write something about this. For example, a friend of mine involved in a medchem project has been forced to spend much of the last couple of weeks getting the products of his rather palladium and tin-heavy route tested for trace metal impurities which might affect the upcoming biological tests. Also, a comment in Rawal's recent welwitindolinone paper (covered by me, with Garg's, here) concerning the requirement to use brand new virgin glassware in order to obtain a high yield in the hydrazone formation step.

2. Ultimately the key ingredient in gold-meditated Sonogashira reactions, and several different 'iron catalysed' processes, palladium is by far the most common culprit for such things.

3. I do so hate this term, which, while often incorrectly credited by Leadbeater to De Vries (in, for example the JOC paper above), actually first appeared in a Chem. Rev. by Beletskaya on the Heck Reaction (DOI:10.1021/cr9903048). The interesting thing is that as the concentration drops the reaction does get better (to a a point):

Fig 1. From Org. Lett., 2003, 5, 3285.

4. A previous supervisor of mine used to joke that when he was a PhD student the only green chemistry anyone did was chromium oxidations.

5. One of the things I'm hoping we'll see one day is the industrial route being used by Eisai for the production of eribulin. Modified from Kishi's work on the halichondrins but scaled up for hundred gram batches, I've heard that the route features a large (~1kg) NHK. I do hope they get all that lovely chromium out of the final product.



Comments (13) Trackbacks (0)
  1. Every time I think, “That’s an awesome post”, you come up with an even better one. Great job!

  2. A while ago, my grad school group had one of our old, beat-up stir bars analyzed by SEM. It was quite enlightening.


  3. A quick test for “metal free” reaction is to add a small amount of TMS-CN. If the reaction dies it is likely there was some metal involved after all.

    For cleaning Pd residues from glassware and stirbars aqua regia does a pretty good job. For other metals soaking in phosphate base bath (1 medium-sized jug of cheap anionic laundry detergent (about 36 loads), preferably unscented/dye free kind + 2 pounds of KOH + half a pound of K3PO4 dissolved in 10 gallons of water) does a pretty good job over a weekend.

  4. Nice post and thanks for the Leadbeater group shout out (even if it did highlight a somewhat embarrassing series of events) You wouldn’t believe how many times this topic has come up in the past few weeks in our group. This topic is clearly pretty big with my boss so when we discuss literature at group meetings we often encounter articles that we think (at least hypothesize) are more likely due to contaminants. The other day we came me and my lab-mate mike came across this one (don;t know if you saw it) http://pubs.acs.org/doi/abs/10.1021/ol2029178 which we think is copper contamination. K2CO3 is simply not powerful enough to deprotonate an alkyne. They do say the the deuterium exchanged does occur with amine bases but again that could be due to contamination on the stir bar as azmanam pointed out. I think its still a broad problem in organic today but certainly not as bad as it used to be.

    • “K2CO3 is simply not powerful enough”. I am not so sure. There is Mannich reaction of terminal acetylenes with imines derived from aldehydes or reactive ketones, which works fine. Also addition of acetylenes to ketones sometimes does not require a base more powerful than KOH

      • Fair enough, I was rationalizing it out in my head the other day (if there was no transition metal assistance) and it is possible that there is an equilibrium between the deprotonated form and protonated form. If the appropriate electrophile is used to trap the deprotonated form and the process is highly favorable (irreversible), than its possible that K2CO3 could be used.

      • One additional example (of exchanging deuterium on a not-so-acidic substrate in water): Few years ago our NMR dude was running some special solvent effect studies and he needed poorly deuterated DMSO, that is to say deuterated to something like 90% level. He called Cambridge Isotopes and learned that they would make it for him as a special order, for 5 thousand USD. So he cursed them, mixed their 99.8+ % d6-DMSO with water and KOH and de-labeled it to the level he needed by heating it to 50 C for few hours. (The exchange was reasonably slow so he could monitor it by NMR. His biggest problem was how to make the stuff anhydrous afterwards, I was distilling it for him.)

    • I thought you might weigh in on this one, in fact I think I might have even mentioned you in an earlier draft, but that seemed unfair as it all happened some time ago. Sounds like it’s still affecting the group’s view of the literature, although a little more skepticism can’t be a bad thing. I didn’t really think anything of that paper when I skimmed it, but I think I probably agree with milkshake that it’s not necessarily contamination (although it couldn’t hurt for the authors to make sure). A friend of mine is finding that nucleophilic aromatic substitution of chloropyridines and pyrimidines works a great deal better with inorganic bases, and in such cases you have to wonder…

      At any rate, it doesn’t seem the Leadbeater group at the time were too incautious, and at least they kept investigating and tracked down the true culprit themselves; it’s always worse when someone else tells you you’re wrong! Crazy that some cross couplings can occur with 2.5 – 50 ppb of Pd… kinda makes me feel somewhat decadent using 10mol%! Guess with the world’s supply of Pd dwindling this may not be a bad discovery.

  5. Has anyone tried systematically adding trace amounts of many different elements to various reactions to see if any of them increase the yield, alter the product, etc.?

    • Not as far as I’m aware, although there are some programs, for example Dave MacMillian’s unpublished ‘accelerated serendipity’, aimed at using robots to systematically mix different catalysts, substrates and conditions to try and discover new reaction types.

  6. I, on a whim, did a similar deuteration of an alkyne for a not-all-that-important reason. I stirred the hydrophobic alkyne rapidly with a 1:1 Et3N:D2O mixture for 48 h and got about 80% deuteration by 1H NMR which was good enough for the purpose I was interested in. I didn’t think it was very special and never bothered to try and optimize it. My point is that although the pKa of Et3NH+ is less than half that acetylene the equilibrium is on your side with an excess of D2O, similar in the case of the pot-carb example above (although pot-carb is about 1000 times less basic than Et3N in water).

    Another point about that deuteration paper. They didn’t actually run any 2H NMR analysis. In the mechanistic studies I’ve done with 2H labelled compounds my supervisor made the point quite strongly (also made by Carl Sagan) “Absence of evidence isn’t evidence of absence” and so I had to run 2H NMR on all the compounds to show there actually was a deuterium atom in them. They should of at least done it for one example in my opinion.

Leave a comment

five × = 45

No trackbacks yet.