NaH: Magical Oxidizing Pixie Dust
July 22nd, 2009 by excimerEDIT (12/23/2009): Retracted!
Not to channel Kyle or anything, but don’t you hate it when you do real science and other people manage to get giant turds published in good journals? Here’s one of said turds.
Okay. Maybe I’m being too harsh. Apparently it’s catalytic in NaH. I guess. I don’t know. They don’t really talk about it much. They don’t try other bases except sodium metal and sodium methoxide which don’t work as well, apparently. Other hydrides, though? Nope. Other solvents? Nope. They propose some mechanisms without any evidence. Ten bucks says their NaH is magically infused with something else. Like chromium trioxide or something. Or that they re-found a result that they didn’t bother to cite (seriously, has noone noticed that NaH makes benzophenone when treated with NaH?) They do, however suck a lot of potential reviewer cock by citing approximately 920539502 papers, of which maybe a third are actually relevant to this particular transformation. When you can’t dazzle them with science, stroke some science egos.
Now, I read every Tot. Syn. post. I also read the comments, which inevitably devolve into a discussion on whether a particular synthesis is “JACS-worthy.” I think it’s a tired argument. If the science is good, then who cares where it’s published? But now I need to say it too. This isn’t JACS-worthy. The science here is not that good. There was no curiosity, no limits being tested, just a bunch of random examples from a serendipitous result. Good science would require answering the question “How is my result not bullshit?” Maybe they’ll get around to it in the full paper. But I doubt it.
WAIT! Tot. Syn. is trying it. Looks like it’s working… but some of the commenters can’t get it to go. Of course, that doesn’t mean that it’s actually NaH doing the reactin’. Go blagosphoere, go!
EDIT: Here’s a paper from 1946 on how sodium hydride can reduce benzophenone in refluxing xylene, and also demonstrates that benzaldehyde can self-condense to form benzyl benzoate with catalytic NaH. Another, more relevant paper from a commenter on Tot. Syn.
maybe the real oxidant is the Mg2SO4 they dry over…
MgSO4. [Mg]2+[SO4]2-
hi, welcome to CBC. the first rule of teh blog is that
you cannot talk about fight clubyou can’t be an asshole here.srsly, we all know what JACrapS meant. don’t be that guy.
A transformative moment – first reaction that will be named to honor the journal editor
Why do you think that the science in the paper is not that good? Does JCAS require curiosity and limits being tested, or is just a really good result also okay?
Unless the result is a lie or has been done before as you say, I don’t see why it’s not JCAS worthy. It might have been published before as you say, but until we find the reference in SciFinder, it’s not a certainty. It’s a lot better than some Grubbs papers on olefin metathesis mechanisms from the last few years. I did a lot of oxidations back in the day, and I don’t remember seeing these conditions. I would have been happy to avoid the Swern oxidation. I never did oxidation on benzylic alcohols though.
A sodium hydride catalyst would really be something ‘curious’. Also, calling out some people from Shanghai is easy. Saying that Grubbs publishes crap that is not JCAS worthy would give you a lot more street cred.
I guess you would find the following paper not ‘Andjewandte worthy’ as well?
http://www3.interscience.wiley.com/journal/117926860/abstract?CRETRY=1&SRETRY=0
sorry but you only made the same assumption error as the editor who let this through. (You probably have an organometalic chemistry textbook from this editor) Read the paper carefully and ask yourself “is April 1st today?” And “if these dudes are for real, is there any more natural explanation – peroxides in THF, air oxygen etc – that could explain the puzzling results?” And if yes, then “what exactly did the authors do to exclude such a obvious possibility ?”
I agree, the first time I looked at the post, I was like, “w.t.f. this is bullshit”. I still think it’s highly unlikely, but I’m willing to give them the benefit of the doubt. Surely they must have tried to disprove this very strongly if it was published in JCAS?
Well, I’ll just wait until totally synthetic repeats the reaction (or fails to do so). He saved me some time trying to do it myself actually. The peroxides in THF is a very good explanation actually… plus benzylic alcohols to a pyridine like to oxidize over time for me anyways when left under simple air, so it could be that catalyzed by base. But they do have non-pyridine examples.
I just don’t want to condemn it outright before someone tries to repeat it.
Also, it’s very damning that there is no control experiment from what I can see in the paper and the supporting info. That is very sloppy. There should always be several controls. How do we know it’s not the silica gel after all?
When I get home I’ll do some back of the envelope calculations on the thermodynamic plausibility of all this. I can’t think well right now. But I’m not willing to dismiss this as crap just yet…
I’m not looking for street cred- if you think I give a crap about what “people on the street” such as yourself think of me, you’re sorely mistaken. That said, I no longer have carte blanche to say whatever I want about whoever I want. If that’s what you’re looking for me to do, go somewhere else.
This paper is ridiculous, though. Regardless of where it was submitted, the reviewers should have thrown up a red flag for the egregious lack of controls.
Er… I wasn’t insulting you or challenging you to say anything. It might have seemed that way though.
Anyways, as I said earlier, there are no controls and that’s very bad. So the reviewers didn’t do their job there. But your main point was that the science was not ‘JCAS worthy’. However, I think that if it all turns out to be true, then it is ‘JCAS worthy’ science. The result, if it is reproducible, and it can be actually proven that NaH is a catalyst, is JCAS worthy.
What about the paper that I linked to, did you look at it? Would you say that it’s Andjewandte worthy or not? Like I said earlier, you don’t need to show curiosity or test limits. For me, the science has to be good and reproducible as well as interesting. The paper accomplishes those goals if it is true. However, the controls should be done.
Sorry to be so disagreeable on this issue. For some reason it matters a little to me. Just a little however.
There are/should be lots of places to publish novel results quickly (OL? TL?) without having them entirely fleshed out. The main thing that (should) distinguish stuff published in levels of journals is the quality of the science – that was why ACIEE got pounded so hard for LaClair and then for the dimerizationXXXXXXXXXXXXZincke reaction papers (because there were significant flaws in the work that should have disqualified those works from being published there, and perhaps at all). If the distinction between qualities of science isn’t meaningful, then what exactly distinguishes a JACS paper (or Science, or ACIEE) from TL, or Arkivoc, or something someone published on their website?
I wouldn’t have submitted a rough draft of a paper read by my teachers in high school or college, much less to a magazine. Not having done control experiments is like submitting a rough draft of the work – but people reading it are expecting a finished product (or at least, a polished one). If the author isn’t willing to do it, it’s the journal’s job to either ask him to do so or not to publish it. If they aren’t going to do that, what’s their point?
Yeah, you’re right actually. But at least with the LaClair affair even I could immediately call bullshit on it after looking at the NMR spectra in the SI.
Also, I now decided that this paper isn’t good enough for JCAS as per my comment below. I was being stupid. I will now go home and get some sleep. Then wake up and go to the bar to fix the lack of alcohol issue.
Isn’t that why one has a liquor cabinet, though? Cheaper, too (unless your friends hang out there, in which case cheap isn’t the only consideration).
My friends (and I but I didn’t drink) hung out a bar for awhile, and they calculated that when the proprietor raised pitchers from $5 to $5.25, it cost them about $500 a year. I bet you won’t find that problem in an Algebra I book though.
In the place I’m in right now, people drink very rarely, and if they do drink, it’s very little. By this country’s standards, I just had enough liquor for a month in the last three hours. Which, is like, a normal Friday (or Wednesday) night back in the States. With this kind of attitude towards alcohol, you can bet that no one has a liquor cabinet and saving money is not an issue when you only drink 3-4 beers a month. This is actually the first time I had alcohol in the last two months. I’m becoming local…
Algebra I for Adults
I smell royalties.
The whole “on water” issue is an interesting one, I think, but I don’t know enough about it to actually comment on it. I’ve heard yeas and nays on both sides. At least the paper you linked to does some labelling experiments.
On thinking about this for a while and talking to some of my other lab members (one of whom told me nor to read all the crap I scan in journals and to get back to doing my own chemistry), I decided that the paper is not good enough even if the results turn out to be true. Please ignore my previous comments. I just remembered an oxidation reaction I did once with a metal oxide and some sodium hydroxide under oxygen atmosphere where the benzylic position was just a bridging CH2…
Yeah, you’re right. This paper wouldn’t be too novel or JCAS worthy.
A strong reducing agent performed a substantial oxidation. Hydride abstracts the hydroxylic proton to give H2. The alkoxide anion kicks out hydride. Mas Alllalalalah! The H*Y*D*R*O*G*E*N car has a recyclable catalytic fuel tank! (Beware autocatalytic higgledy-piggledyness.)
Now do it with diphenylmethylthiol. Benzophenonethione has such a lovely color.
Free H- is kind of ugly. Could Na+ activate THF as a hydride acceptor? (If peroxide is around it could work as an acceptor, though it might be better oxidizing the alcohol directly). If O2 were the hydride acceptor, that would be interesting as well.
Do they say anything about substituent effects? As a first guess, EWGs on the aryl rings should accelerate oxidation if deprotonation is important but might slow it down if a peroxyalcoholate is the intermediate.
We postulated in the lab that they were using old THF that had formed peroxides, which would give you the desired result. That was the best we could come up with.
They *claim* to use THF that was distilled from Na/benzophenone, which should take care of peroxides, but yeah, who knows how old the THF they used was.
good post.
There’s also a paper in JOC in 1968 on reduction of a camphor with NaH. The benzaldehyde paper wouldn’t seem to help, because the Cannizzaro reaction should happen with benzaldehydes using a reasonably strong base – there’s a good hydride donor once you can get alkoxides running around. Here, there’s no obvious place to put the hydride, unless it returns to solution (Uncle Al) – but NaH was supposed to be a miserable H- donor under most conditions, so how if the reductant getting generated so easily here? Is the fact that the H- ends up as H2 paying for the generation of H- from the alcohol (and if so, why here and not in other cases)?
I like the line that popped up in Google about NaH being an overlooked oxidation catalyst. It sounds like the equivalent of “gasoline has been often overlooked as a firefighting liquid”. Really?
If THF is the H- sink, then there should be an equivalent of 1-butanol floating around in the product mixture. Also, LiH (if you can push it into solution) should perform better (because if the M(THF)n complex is the hydride acceptor for the alkoxide, then Li should be better at activating the THF for ring cleavage than Na). Of course, in that case, the base (NaH, LiH) shouldn’t matter other than 1) generating alkoxide and 2) generating the M(THF)n complex to activate THF as a hydride acceptor; the latter would be pretty interesting if it were true (more than the use of NaH as an oxidation catalyst), but one would have expected it to be seen before.
This seems awfully weird. I’ll have to read this.
If THF is getting reduced, then the BuO- formed can act as the base for subsequent rounds (BuO- deprotonates the alcohol – alcohol donates H- to ring open a THF to butanolate) and the reaction would be catalytic in whatever base you choose.
I have never heard of NaH attacking THF. BuLi, yes, but not NaH. Or KH, for that matter.
It probably can’t be THF – the authors say the reaction runs poorly in other solvents, which I take to mean it runs in them, so THF can’t be the sole H- acceptor, at best. In addition, you would figure that people would have seen butyl ether byproducts in every NaH/THF etherification reaction were that the case. I just don’t see what the oxidant/hydride acceptor is. (Theoretically, the OH proton is like H+ so the hydride is transferring an electron to H+ to give H2, but I don’t know where the second is).
I don’t have anything reliable, but the dehydrogenation using numbers from The Chemist’s Companion might be about 10 kcal/mol exothemic, which is better than I thought.
I would bet money on oxygen. In whatever form – abstracted from ketyl on distillation, peroxides in THF, impurity in nitrogen, adsorbed on NaH (seriously, read the 1968 JOC paper – it will answer many question raised here). NaH IS able to reduce a carbonyl compound to a little but significant extent, even though it is completely insoluble in THF – they cite a paper saying that LiH in THF is not detectable down to 50 ppb. Probably LiH has a more robust crystal structure, but I would expect more than a few ppm of free hydride from NaH in solution. Which, apparently, is enough to start the process if the oxygen is present (and I believe it is).
We had a somewhat heated argument with a labmate who accused me of being a hater for immediate dismissal of this eyebrow-lifting report as bullshit. This view changed later (thanks TotSyn) to “very poor experiment design and academic presentation”. He is siding with Uncle Sam now.
Correction:
I would not expect for than a few ppm of free hydride.
I’m going with either physiosorbed oxygen on NaH or sodium oxide. An 18O scrambling/labelling experiment would be a nice one to try.
If you think this work is out there, check out a couple of papers by this guy back in ‘05 in Tet:
http://dx.doi.org/10.1016/j.tet.2005.05.059 and http://dx.doi.org/10.1016/j.tet.2005.05.058
Why is a 6-membered TS including Na-O-C-H…Na-O… completely out of discussion as it FORMALLY explains the re-generation of NaH and the formation of ketone from two alkoxide molecules? I am just electron pushing and have no clue of energetics.
Because NaH might not even be the oxidant- I think that’s the main point of contention. Also, I would think NaH would immediately deprotonate any alcohol in solution.
I agree that superstoichiometric amounts of NaH lead to complete deprotonation of the alcohol (with loss of H2). The question is: Is it possible to THEN generate NaH (+ketone) from a self-reaction of alkoxide in the absence of free alcohol (via the before mentioned TS)?
The thermodynamics of that transformation, whatever the cyclic transition state, are ludicrous IMO…
Wow, it sure is nice to come home and see that everyone else was as intrigued by this paper as I was.
Random thoughts that others have sparked or alluded to but not directly stated: (and no I haven’t taken the time to read the paper extremely carefully-trying to pay the bills today and make some drugs):
1) So why doesn’t this work with catalytic NaH, the mechanism I saw drawn by the authors wouldn’t need a stoichometric amount.
2) Two, if you deprotonate the C-H and sink the electrons on the heteroaryl nitrogen, you have a glorified enol/metalloenamine. Think tautomerization of enol -> ketone and then (cringing, but completing the circuit), rearomatization ejects a hydride–(or more likely), deprotonates another C-H. I think the need for excess NaH, is that most of the base is consumed by deprotonating the -OH, not the C-H. Therefore, 1.1 equiv of base is required. (Not buying all of this, but IF these results are real, offering a reasonable expectation. I’m always in the camp to believe the results until you have TANGIBLE reason not to).
3) Thermodynamics of Generation of NaH vs generation of 2-ketopyridine (aromaticity)? I’m too old for this (or just too tired), any help would let me stop worrying about it.
4) The use of LiH is screwed thermodynamically because of its ridiculous crystal lattice energy.
5) Excimer, I have to agree with your original comments to a point. In my opinion, this is completely the fault of the editor and referees. The bottom line is that IF this is truly related to NaH, it is pretty cool. The science that I did read, didn’t completely support/rule out less insane possibilities. Yes its the scientists (authors) job to do the science and rule out the other possibilities, it ISN’T their job to determine how high/low a journals standards are held. That’s up to the referees and editors. They don’t have to make the scientists do the work, but they can limit what DOES make it into the journal. (It’s like raising a kid, you want the dessert, you say please-run the controls). This happens all the time, its just sad to see such a uncritiqued article make it through the publishing process, especially after running stupid experiments that referees have demanded I run to get a paper in.
As a mentor of mine likes to say, the more impactful/outrageous a claim, the more support you need. If you think NaH is going to deprotonate methanol, no one will demand crystallographic proof of NaOMe, but in this case… I can’t imagine what the mechanistic group meeting blackboards look tonight across the academic landscape
uncle sam- JACS does claim to be the venue for the highest impact/highest quality science. Part of that (basic scientific method) is running controls. That is demanded to get in, in my mind- forget the fact that this is much ado about a quirky result at best. No one is going to ever think “Hey I can’t find DMSO/COCl2, lets chuck in NaH!”
Keep the debate raging!
i always enjoy it when you comment, dabut.
1) they claim it’s due to low solubility of NaH, but don’t actually do the experiment catalytic in NaH to back it up.
2) I like making shit up too. I agree, it’s fun to think about, but any mechanism proposal is completely dependent on experiments they haven’t done.
3) If they’re proposing going from sodium alkoxide to sodium hydride, the pKa difference alone there is like 30. There would have to be a massive energetic preference for aromatization, which I doubt.
4) yeah, oops. KH, then.
5) I agree, the authors aren’t the only ones to blame. Then again, they’re the ones who tried to submit it in earnest, so they deserve most of the blame. Inattentive reviewers/editors are also partly to blame.
My boss is fond of saying “it’s not science without the controls.” I think more than a few methodologists have forgotten that methodology is more than just about making crap in new, generally stupid ways- it’s about understanding what’s going on in those stupid reactions so that people will, at the very least, believe that, oh, I dunno, your sodium carbonate doesn’t have any palladium in it.
I’ ve written this down on Tot Syn but am coming back here… good job on covering this, and nice blog by the way.
The 1965 JOC paper
http://pubs.acs.org/doi/abs/10.1021/jo01018a080
describes not only reactions with benzaldehyde (as the title implies), but also the reaction with p-nitrobenzylalcohol. Read it, people, honestly: it replies all your questions here. The authors manage to oxidise this alcohol with sodium hydride in THF all the way to the carboxylic acid (81% yield). They manage to repeat this oxidation with NaOH and NaOOH. They conclude by speculating about mechanisms involving radical anions (after trying to get evidence for those) and are positive that oxygen/peroxides, either adsorbed on NaH or simply in the solvent, must be the culprit.
I would complete what excimer wrote: our brilliant advance of the state of Organic Chemistry as a science led us to a situation where a 1965 paper does a much (waaaay much!) more serious, careful, and scientific work than a 2009 JACS. Furthermore, the 2009 JACS was only concerned with filling up tables and tables of examples, citing piles of irrelevant literature (but not the crucial precedent of this 1965 JOC!) and has a Supporting Info with dozens of pages of spectra, but 5 or 6 paragraphs of real experimental data. And this piece of “science” was accepted by THREE independent referees in the 2nd highest-ranked journal for communications in Chemistry. Is this what we have come to?
EC I think you have it nailed. We tried an experiment with fresh NaH stored in a glove box and got no oxidation.
I agree Fink,
This paper does Stink
That’s not how my name is pronounced. It would be “stinky.” Thank you.
My bad, thank YOU
As a colleague mentioned a few days back, JACS is a journal in which everyone wants to publish but which nobody really wants to refer to for actual practical knowledge.
[...] Comments excimer on NaH: Magical Oxidizing Pixie DustCurious Wavefunction on NaH: Magical Oxidizing Pixie DustHap on NaH: Magical Oxidizing Pixie [...]
It’s not catalytic if you use a stoichiometric amount or an excess.
I am reminded of an oxidation reaction we run in our teaching labs which uses base and oxygen. I suspect that a) NaH is acting as a base rather than reductant, and b) at best there is water/oxygen contamination in their “dry, anoxic chamber”. It is also entirely possible that their THF includes peroxides. I know when I worked primarily with small quantities of THF in a glove box, we’d store it in there for months because bringing solvent in was such a pain. If they work similarly, it’s entirely plausible that peroxides could have developed too.
Really I’m inclined to disbelieve anything claiming “oxidation by a reductant!” in the TOC scheme. Maybe it’s a gimmick to get people to read their “novel” discovery. It doesn’t make me want to read the paper though — it makes me wonder what the editors and reviewers were smoking.
Verdict: So much to disbelieve and so much apathy on my part!
All that’s required for a catalyst to be a catalyst is that it not be consumed in the reaction. You can use superstoichiometric amounts of catalyst as long as you can recover it – it’s the expense and hassle of doing so (not to mention purifying the catalyst from components of the reaction mixture if the catalyst is expensive) that makes (super)stoichiometric catalyst not a good thing. Since NaH is so cheap, they can use it in excess and don’t have to worry about recovering it (though they say they can – how accurately the massed it afterwards I don’t know).
I don’t think anyone was smoking – there are lots of legal alternatives that could explain this behavior.
Not everyone agrees with your definition. Some add the specific requirement that a considerably substoichiometric amount of the proposed catalyst be used.
there’s a method in modern catalysis called the “just put it all in” method, whereupon things get a lot hotter and more reactive when you dump more stuff in. I mean, only putting a little in will get the desired result, but things go a lot faster if you put more in, and sometimes, for maximum rate-enhancement you need to put it all in.
…penis
Powdered milk also contains relatively abundant 5-hydroxymethylfurfural (C6H6O3) which produces an MH+ peak at m/z 127.0395. On collisional activation, 5-HMF happens to have produce exactly the same nominal-mass fragments as melamine (C3H6N6, MH+ at m/z 127.0732). Several people have been claiming to detect melamine by low-resolution MS/MS alone. Without high resolution and/or chromatography, you can’t distinguish melamine from 5-HMF. I tried to explain this when interviewed for an article recently, but the comments got cut for “lack of space”.
Oops — this comment was supposed to go with Part II: The Revenge…
Their MS-MS don’t have much of a 127 peak for the untainted raw milk (there is a peak, but it’s shifted and small compared to the 2 mM melamine peak), while the untainted baby formula has no 127 peak at all (the melamine-tainted formula has two peaks in that vicinity, with the 2 mM melamine peak being much bigger) . Is that reasonable? (Could the drying process generate 5HMF by dehydration of sugars, thus explaining its absence in the raw milk samples?)
This problem is not the first time that people find it. In a book which is wrote by Wiley, strategies for ogranic drug synthesis and design, Page 32, you can see an oxidation which is reacted by DIBALH. It is a reaction found by the industry. Dr wang just put this reaction on the desk for us to discuss, just like the organocatalyze, which was found in early years, was promoted by Barbas, List.
They had found a long time age, but without these people, like Dr wang, List, these reaction will not be noticed.
no, this is a truly atrocious paper that should not be acceptable even in some Indian journal – sloppy work, no controls, delusional interpretation of results, lousy reviewing and non-existing editorial oversight. Wang et al write that NaH is a known strong reducing agent – a fail answer on a freshman exam – and then he writes that “NaH is a catalyst that can be recovered in 100% yield by simple filtration” – this dude is smoking some strong stuff!
Sir, I can not find the words “NaH is a catalyst that can be recovered in 100% yield by simple filtration” from the paper, could you tell me where these words are? Thank you.
I think these words are a bit exaggerate.
“both NaH and THF could be quantitatively recovered by filtration of the reaction mixture under a nitrogen atmosphere and subsequent evaporation of the solvent. There is no such issue as the metal residue or decomposition waste that was incurred by the oxidant itself. ”
Sooo yeah.
[...] challenged by an online cabal of chemists over at TotallySynthetic. It was further questioned at CBC and has been the scorn of folk in the [...]
But I think it is not the main problem. Because it never happen when do filtration.
I focus on this reaction whether really happen. If the expression is exaggerate, It is the author and editor’s responsibility.
These things happen so often. Rarely do they get corrected. I am surprised that no one has brought up the shit that Carston Bolm has been pulling. He has been saying for all of these years that Iron has been catalyzing his cross-coupling, then he sends people over to Buchwald’s lab, because of course, Buchwald ain’t having/believing any of that shit. Turns out, it was the <.1% CuO present in the Iron that he was buying that was catalyzing the reactions. Check out his retraction last month in Angewandte, and his new paper, that makes no mention of Iron, but instead "Copper impurities." Lame.
People have been taking a dump on Fe cross-coupling ever since it started coming out. It was bound to happen. At least they owned up to what was really going on. I have a little more respect for people who are willing to do that. At least shows they’re humble.
At least they were not violating stoechiometry or thermodynamics.
By the way, I am amazed that it Bolm system works – I mean Fe(III) is a stronger oxidant than Cu(II) – it ought to have oxidized all Cu(I), and Cu(II) shouldn’t be catalytically active in arylation reactions….
I meant to type Cu(2)O. The Point is, you are right, Bolm was humble about it. But the fact that it took such intervention to do so is a bit fucked. If Buchwald and other sources are telling you for years that when they used Iron alone it didn’t work, wouldn’t you think about impurities? My boss bought the purest iron source he could get, new spatulas, the whole nine yards, and basically saw that no reaction occurred. It has been noted that the reaction works with extremely low catalytic amounts of Cu(I). Bolm knew this. Humbled, yes. Seems more like his humbleness was overlooked by 3 or 4 Angewandte papers…
[...] is one of them. But where are all the hardcore materials folk? Hiding out while everyone else aruges about NaH?[1] Perhaps I can entice some delurking with a picture of some solar [...]
Nope! we could not replicate the oxidation with NaH.
http://usefulchem.wikispaces.com/Exp243
http://usefulchem.blogspot.com/
Refer S. Ohta et al, Synthesis 1983, 291.
They said various benzylic alchols and aldehydes were oxidized in good yield using NaH with a catalytic amount of pyrazole in THF under air at room temperature. The paper suggests that Na2O2 is generated from O2 and NaH and which is the real oxdant.
At least, Chinese chemists had to refer the paper.
[...] IT is ON again: Na Promoted Aerobic Oxidation of Alcohols to Ketones, Tet. Lett., 2009, in press [tx mt]. [1] [...]
This may have been mentioned (or implied) by other comments which I haven’t read (I got through 50%… but then gave up), but for NaH to be the oxidant, you would either need to have to reduce Na+ to Na, or at least some sort of Na complex in which the Na has an oxidation state less than +1. I’m pretty sure that if they say stoichiometric formation of sodium metal in the reaction they would probably see that. either that or H- gets reduced to H-2, or at least an oxidation state less than -1, which is ridiculous to say the least.
The first thing I thought of when I read it was the massive acceleration you get with the anionic oxy cope compared to the oxy cope. the cope is obviously a different reaction to an benzylic oxidation, but I’m presuming the NaH is simply accelerating the reaction.
My money is on oxygen being the oxidant rather than peroxides. I don’t think you can get an equivalent or more amounts of ether peroxides to form even if you leave the solvent exposed to air for days, speaking from previous experience. I’m not saying that the small amounts of peroxides aren’t dangerous, but just that you can’t possible get that much. I’m assuming that these guys stored their THF under Ar, so even if it’s old, you can’t get an equivalent of peroxides in there.