Skip to main content

Chemical News

The Good Ol’ Grignard

Once in a while I’ll see someone studying undergraduate organic chemistry, and I’ll mention to them that those reactions that they’re learning – well, a reasonable number of them – actually get used out in the real world. (The students are generally surprised by this news). I think that a prototype of this sort of thing is the Grignard reaction. It’s a staple of sophomore organic chemistry questions in the “Synthesize this compound from these starting materials” format, as it should be: it’s a great way to bring in a “methyl minus” carbanion nucleophile or to functionalize an aryl bromide. And a lot of such courses do one in the lab, or at least they used to. When I was a TA, there was a phenyl Grignard reaction done in ethyl ether, and the poor ventilation in the lab chased me out into the hallway by the time I taught my third lab section of the week. But most of the students got it to work, and I think the “Try crunching up the magnesium, try stirring hard, try adding a crystal of iodine” part of getting the reaction initiated gave them an honest look into what organic synthesis can be like sometimes.

And you can, of course, buy a big long list of commercial Grignard reagents to use at the bench. I’ve also been involved in a couple of scale-ups of more customized Grignard reactions to make material for a first two-week tox run; it really does get used, although I’m sure that no one is enthusiastic about doing it with a hundred-kilo charge of magnesium turnings (although I’m equally sure that it’s done on that scale from time to time!)

The funny thing about the Grignard, though, is that the mechanism is still a bit unclear, even after all these years. There is evidence for a polar mechanism (that carbanion), and there’s evidence in other systems for a single-electron radical mechanism. It’s also obvious that the Schlenk equilibrium is operating, shuttling you back and forth between (at the very least) methylmagnesium bromide (as a prototype), magnesium bromide, and dimethylmagnesium, with the ratio of those things very much dependent on the conditions. That “methylmagnesium bromide” isn’t quite as drawn, either – there are certainly more complex alkylmagnesium halide species in there, in equilibrium with each other as well. The harder you look, the more you see.

This paper is a nice recent overview. It’s an intensive look at methylmagnesium chloride in THF with either acetaldehyde or fluorenone, about as plain-vanilla a set of Grignard conditions as you can imagine (well, short of maybe using benzaldehyde, with which every nucleophilic organometallic species known to science has been reacted at one point or another). High-level calculations indicate that the reaction is indeed balancing on several energetic knife edges, and there are a *lot* of species involved. That diagram at right, for the acetaldehyde case, will give you the idea. You’ll note that the THF solvent is very much a participant, which Grignard aficianados already know well from the sometimes-different reactivity of the diethyl ether solutions versus the THF ones. In the end, as we know, the Grignard is not just one reaction at all:

The coexistence of multiple species in rapid exchange due to the Schlenk equilibrium and the evidence that the activation energy range of all possible reactions is relatively modest indicate that the Grignard reaction should not be described by an individual process. Instead, it should rather be thought of as an ensemble of transformations that can occur simultaneously in solution. It is likely that improvements of the Grignard reaction, for example, by alkali or copper salt additives occur by interference with one or more of the possible pathways. . .

It sure is. The concentration, solvent, counterions, starting halide, and electrophile partner all have roles to play, and that’s even before we start getting fancy with the additives. So when I next see some student working on that section of their first organic chemistry course, I’ll still tell them that the Grignard is real and that the Grignard lives – but that it’s not quite as straightforward as their textbook has it! I’ll also still tell them another famous bit of trivia: that it’s surely the most famous reaction that was actually named after the graduate student (Victor Grignard, who eventually won the Nobel for the work), rather than the professor (Philippe Barbier, of the often-closely-related Barbier reaction). Grignard himself said he would rather have shared the prize with Barbier, actually!

64 comments on “The Good Ol’ Grignard”

  1. neo says:

    Someone needs to put “The harder you look, the more you see.” into Latin and print it on a t-shirt.

    1. loupgarous says:

      Qui vult videre magis, vide fortius would be the elegant translation.

      1. Dan Berger says:

        “He who would see more, looks harder.” Not quite the same, but in Latin perhaps that’s the more common order for conditionals.

  2. Joseph says:

    Another piece of impressive graduate work: Van Der Waals derived his eponymous equation of state in his PhD thesis.

    1. Mad Chemist says:

      Don’t forget Arrhenius.
      His dissertation on ionic dissociation formed a significant part of his Nobel work. Ironically, his committee originally wanted to fail him because they didn’t believe in ions. They passed him with the lowest score possible but later were persuaded to bump it up a notch.

      1. Martin (still not Shkreli) says:

        mmmh…. Mossbauer, anyone?

        1. Anonymous says:

          Re: Mossbauer. For those not knowing the reference, Mossbauer discovered the Mossbauer effect and Mossbauer spectroscopy as a grad student and he won the Nobel Prize a few years later.

          Cerenkov discovered the Cerenkov Effect as a student / staff researcher in 1934. He didn’t get his PhD (on the Cerenkov Effect) until 1940. He also won a Nobel Prize.

          Fujishima discovered the Fujishima Effect as a grad student with Honda. No Nobel Prize yet, but I think it’s deserving.

          Not all PIs are as accommodating of junior co-workers making new discoveries in their labs or they will try to take credit for the discovery.

  3. ScientistSailor says:

    “The harder you look, the more you see.”

    True for just about everything in life, no?

    1. Blind man says:

      Looking inside an empty glove box under high vacuum may be an exception?

      1. tim Rowledge says:

        Oh, just look harder; a seething mass of virtual particle creation and annihilation.

  4. MTK says:

    In thinking about it the Grignard is probably the #1 reaction in terms of moles of material produced in my career.

    After years of experience and an early, and scary, near miss in terms of initiating one, I settled on the dry stir a big excess of turnings under nitrogen for a few days then the slow dropwise addition of a solution of the halide and dibromoethane. Works every time and gives you a chance to stop or slow things down should things get hot.

    Actually if your halide is an aryl iodide or highly electron deficient, transmetallation with iPrMgCl is even better.

    1. ScientistSailor says:

      YES! I love the pre-stirr procedure. Too few people know about it.

    2. Derek Lowe says:

      I agree on the dry-stir! My impression, though, was that it worked better under argon than nitrogen? Formation of a magnesium nitride layer was the explanation I heard, but I haven’t done a head-to-head test.

      1. milkshake says:

        I can confirm that dry stirring overnight works even better under Ar with few drops of neat Br2 added, so that the turnings get crushed under Ar in vapor of Br2. Looks scary, but gives reproducible results with dry THF or ether, with minimum induction period the next day.

      2. Martin (still not Shkreli) says:

        Well I do remember my main-group-organometallic teacher telling us about the dry-stir activation, and that you should not do it overnight under N2 because of the Mg nitride.
        Maybe if you keep it short, you have only little nitride formation, so it works, but if you leave it longer (as you would under Ar), the nitride has time to form and cover the Mg with nitride…

      3. MTK says:

        From all the other comments I guess you’re right, it was Ar. It’s been at least 8 or 9 years since I’ve done one, so my memory may be hazy.

        Anyway, the dry stir really works well.

        Hap, you do end up getting a spin bar with a nice metal sheen to it.

        1. MTK says:

          Maybe it was nitrogen that I used.

          The paper claims that either N2 or Ar can be used.

    3. Michal says:

      What is a dry stir?

      1. Dionysius Rex says:

        Dry stir = (Mg turnings in a sealed RBF) + inert atmos + magnetic stir bar, without solvent. Stirred. Gets lots of nice fresh magnesium exposed without chance to oxidise.

        1. Hap says:

          Will that scrape up the stir bar?

          In undergrad, they did the PhMgBr and benzophenone? or Et benzoate? to give triphenylmethanol in lab. Even for me, who was not good, it generally worked if you swept magnesium turnings off the benchtop, put them in a test tube, added bromobenzene, and ground it with a glass rod; you could use the reaction to initiate other people’s Grignards, I never did it with anything that might have been more sensitive, though.

        2. A Nonny Mouse says:

          I have always used a quick scrape in a pestle and mortar and into the reaction flask coating with a little solvent and iodine. When the colour goes start adding the bromide.

    4. Anonymous says:

      Mg activation methods: some give a quick wash with dilute aq acid (HCl or NH4Cl), wash with water, solvent, and dry under vacuum or Ar flush.

      1. Another Anonymous Chemist says:

        I second this. Stir even the worst, oldest, inherited-from-the-lab-across-the-hall Mg turnings in 1N HCl (note: evolves H2 and presumably heat, do not do this on the kilo scale!), then pour in Buchner funnel, rinse with ethanol, followed by diethyl ether, then immediately throw in rbf under hi vacuum and hit it with the heat gun. This has been infallible in my hands; it’s also super quick and, if the turnings are REALLY old, viscerally satisfying when they emerge shiny and new. May be considered “cowboy chemistry” by some, not sure.

        1. VikingChemist says:

          I remember trying the dry stirring method under argon in grad school and it worked quite well but I preferred washing Mg turnings with HCl, water, ethanol and drying under vacuum overnight. Add Br2 and the halide in THF and sonicating until reaction would initiate. Worked every time.

  5. Lane Simonian says:

    I was just glossing over this one, because it is well beyond my capacity to understand. But largely off topic, can anyone tell me about the antioxidant capacities of tetrahydrofurans. To date this all that I have been able to find:

    1. Nick K says:

      Sorry, Lane. I’m not going to drink THF under any circumstances, even if it is an anti-oxidant.

    2. Hap says:

      They tend to react with oxygen to yield peroxides, so they seem like they would be bad antioxidants – they would convert oxygen to things that are likely worse for you (peroxides and aldehyde hemiacetals, etc.), but I don’t know for certain, because biology is weird.

      1. Lane Simonian says:

        Thank you, for this cautionary, helpful answer, Hap.

  6. Anonymous says:

    Sometimes, it’s a question of time resolution. If two separate events (1e, then another 1e) occur in a time span smaller than your time resolution, it looks like one event (simultaneous 2e transfer). Henry Linschitz was a pioneer in flash photolysis and, in the 1950s-1960s, his cutting edge methods were referred to as “ultra-fast spectroscopy.” At the end of his career (2010s), as picosecond and femtosecond resolution spectroscopy became routine, his old methods were still being used but were referred to as “ultra-slow spectroscopy.”

    JF Normant got a big boost to his career when he discovered that vinyl Grignards can be preferentially prepared in THF rather then Et2O (Compt. Rend., 1954, 239, 1510). (It really took 50+ years to figure that out? Whatever …)

    For DECADES, undergrads had been preparing Grignards in UG lab in open test tubes, using non-distilled (WET) solvents. CJ Li (McGill) went further and did Grignard reactions in water w/o co-cosolvent. It’s supposed to be “green” but some people argue the aqueous waste is harder to dispose of than conventional organic waste. (Any process folks care to comment?) … Come to think of it, the original Barbier Reaction (1899) was carried out in mixed THF/H2O. I can’t access the original Grignard papers (1901) – did Grignard also use H2O as co-solvent?

    1. Nick K says:

      I’m astonished to read that one can do Grignards in wet solvents. Could you give a reference to that?

      1. milkshake says:

        you can do Barbier-style Grignard reactions (with carbonyl partner present in the react mixture) with non-dried ether if you use sonicator bath. It is inferior and harder to scale up but easier to set up on small scale. The biggest problem with moisture (apart from solvolysing Grignard) is formation of Mg hydroxyhalides on Mg surface, and sonication apparently takes care of this problem

        1. Nick K says:

          Most intriguing. Do you need to use an excess of halide and magnesium to offset the losses due to reaction with water?

          Perhaps the reaction is occurring through the radical anion bound to the metal surface rather than through a discrete Grignard.

          1. milkshake says:

            yes you lose some material due to moisture. this is much less of a problem with Barbier-style couplings done with Zn powder (organozincs are far more sensitive to air but hydrolyze relatively slowly)

      2. Anonymous says:

        Grignards in water: I think this was Li’s first paper on the subject: J. Am. Chem. Soc. 1998, 120, 9102-9103.

        And to the follow-up Q, Yes, you usually need excess halide.

        There are several reviews on the topic.

        1. Nick K says:

          Many thanks for the reference. I’ve learned something useful today.

  7. Anon says:

    I thought this is well known. The same is true for alkyllithium and alkylaluminum compounds. Nothing is simple.

    1. Anonymous says:

      Anon, re: “this is well known”: Which “this” are you referring to?

  8. Another Guy says:

    After trying iodine, dibromoethane etc I finally broke down and resorted to making finely-divided magnesium powder using MgCl2 and potassium metal. It worked, but it was a lot of effort. Wish I knew about the dry-stir.

    1. Anonymous says:

      Another Guy: MgCl2 + K —> finely divided Mg powder. That would be Rieke Mg, n’est-ce pas? That is HOT stuff! It glows red and burns upon exposure to air.

      1. Another Guy says:

        Indeed it is. Should use a blanket of argon, anhydrous and O2-free solvents, transfer via cannula .. etc etc. I never found out if it was pyrophoric and that is probably a good thing. Potassium is a kick-butt reagent. Even the tiniest speck that managed to stick to the spatula would smoke long after clean-up was supposed to have neutralized it.

    2. Anonymous says:

      I had a very difficult vinyl magnesiation problem. Yield, stoichiometry, stereochemistry, … I had read about “triply sublimed magnesium” that had been prepared by a group at Dow (also for a Grignard, I think). I wrote to them and asked for a piece and they sent it. It was a big chunk of Mg. One surface was fairly smooth and curved to the shape of what was probably a round bottom flask or trap from which it was probably chipped away. It was about 1 to 1.5 cm thick and the “inner” (I suppose) surface was finely striated, like the soft aluminum heat exchangers on a car or air conditioner radiator … but hard as a rock. … It is known that magnesiation of vinyl halides is not stereospecific and triply sublimed Mg doesn’t change that.

    3. Another method of activation that I have come to appreciate is DIBAL-H. I use overhead agitation a lot of the time (magnesium salts precipitate out near the end of the reaction at process concentrations) so the dry stir method is not always an option. Link in handle is to an OPR&D paper investigating activation of an aryl bromide.

      Aside from the paper, I can offer the anecdotal evidence that I have been using the same 250 g bottle of Mg turnings for the last 3 years, and they have been stored under my fume hood (not in an oxygen-free environment). Suspend the turnings in minimal THF under N2, charge 2 – 4 mol% relative to Mg of DIBAl-H (available as a 1 M – 25% solution in hydrocarbon solvent) and observe the mild exotherm as it takes out the MgO. Go grab a cup of coffee/quick bite, then start adding your alkyl/organic halide dissolved in THF at whatever temperature it needs. Enjoy the reproducibly activated Mg.

      A bonus of this is that it makes it easier for operators to follow when done at larger scales.

      1. Nick K says:

        Interesting method! It sounds very straightforward. I’ll give it a try.

  9. milkshake says:

    I was making BrMg(CH2)3MgBr from dibromomethane (there is a JACS paper from young G. Whitesides that says it cannot be made directly in this way). So I made it in 45% yield at RT in ether (the rest of material goes to cyclopropane), but it does not work in THF. The trick was to use large excess of Br2 and mechanically preactivated Mg turnings in ether, and syringe pumping 1,3-dibromopropane over 2 hour period, to limit the formation of dimeric BrMg(CH2)6MgBr. So in this one case it looks like single electron transfer with radical formation might be involved, and the solvent role is critical.

    1. milkshake says:

      sorry for the typo, 1,3-dibromopropane was the actual starting material

  10. Jack Straw from Wichita says:

    i teach the Grignard reaction to ~170 students that all think they are going to be MDs. please do not add any layer of complexity to this reaction. students already have enough trouble with it.

    1. Tourettes of Chemistry says:

      Recalling the prof’s first statement on opening day for the same type of organic class for ~400:

      ‘Physicians are to medicine what paperboys were once to the news – they merely deliver the goods.’

  11. Karl Miller says:

    Interesting new looks at the Grignard and the Schlenk

    Raphael M. Peltzer et al. How Solvent Dynamics Controls the Schlenk Equilibrium of Grignard Reagents: A Computational Study of CH3MgCl in Tetrahydrofuran, The Journal of Physical Chemistry B (2017). DOI: 10.1021/acs.jpcb.7b02716

    Raphael Mathias Peltzer et al. The Grignard Reaction – Unraveling a Chemical Puzzle, Journal of the American Chemical Society (2020). DOI: 10.1021/jacs.9b11829

  12. Thomas McEntee says:

    Decades ago, at the Arapahoe Chemicals plant in Boulder, CO, we made all kinds of Grignard reagents for sale. Depending on customer needs, these were produced in either THF or diethyl ether…in 750- to 2000-gallon vessels. For production campaigns, we pumped out the finished Grignard reagent and used the gunk at the bottom of the reactor, the “heel”, to kick off the next batch. Boulder was a great place to produce Grignards since the humidity levels never got much above 30 percent and in the cold winters, could drop down to the single digits. One plant at our original site was outfitted with 750-gallon vessels and most of the ancillary equipment–pumps, mechanical stirring drives, were air-driven to reduce potential for electrical sparks. The process developed in Boulder for production of the dl-acid that ultimately led to naproxen was a Grignard reaction run in 4000-gallon vessels.

  13. David Edwards says:

    Although I’ve yet to take a peek at the paper, this already looks fascinating. And leads me to wonder what other venerable reactions regarded in the past as “simple”, will be revealed by a similar analysis to possess this interesting ensemble nature? The organic textbooks are littered with potential candidates for this – or, at least were, back in the days when “organic textbooks” meant the 2-volume I. L. Finar tome that was already showing its age when I first found it in my local library at the age of 10. Whether the most recent textbooks still cover 19th century vintage reactions or not, I’ll have to rely on others here to tell me, but Derek has already pointed to at least one reaction that’s been superseded by a modern alternative in his TIWWW piece on dimethylcadmium. (Yes, I pay attention!)

    Incidentally, while musing on the matter of organometallics such as Grignards … most of the carbon-metal bonds I’ve seen discussed here on Derek’s blog, have been tremblingly weak, to the point of making the resulting compounds terrifyingly pyrophoric. tBuLi, trimethylaluminium and the dialkyl zincs merely being the most infamous instances covered. Are there any carbon-metal bonds that are strong enough to resist attack by atmospheric oxygen? The C-Se bond in selenocysteine strikes me as a candidate, if one stretches the definition of ‘metal’ to an element such as selenium (a pyrophoric amino acid isn’t going to see much use in the biosphere, I suspect), but are there any reasonably stable C-M bonds, where M is an element with only positive oxidation states? Other than carbonyls, of course, of which nickel carbonyl has received at least some attention from Derek in the past.

    I suspect the full time chemists will have fun introducing me to various exotic compounds upon reading the above, and may even surprise me with something I should consider relatively familiar even though my last lab occupancy was a long time ago.

    1. Ken says:

      The nitrogenase cofactor has a FeS lattice with a carbon inside, bonded to six of the iron atoms. That always seemed a little weird to me.
      I have no idea how it would react with oxygen. The nitrogen-fixing bacteria very definitely do not allow oxygen anywhere near it.

    2. Cob says:

      Co-C bond of the cobamides (vitamin B12 et al) come to mind as being pretty stable, especially in cyanocobalamin

      1. David Edwards says:

        I knew I’d miss at least one obvious one! Vitamin B12!

        I take it I can also add haemoglobin to the list, and those copper based haemocyanins found in crustaceans?

    3. Baltic says:

      Some metallocenes are reasonably air-stable, as well as some transition metal complexes.

      If you’re looking for a more or less discrete sigma bond between carbon and metal atoms, well, trimethylplatinum iodide probably qualifies, as do several organotin and organolead compounds – tetraethyllead was stable enough to be used as fuel additive back in the day, for example.

      1. Hap says:

        I forgot ferrocene. I think organoleads are reasonably stable (tetraethyllead as fuel additive). I thought methylmercury complexes were stable to air, but I don’t know (they last in sediments, but I don’t now if that’s because of lack of oxygen or not).

        1. Baltic says:

          Yeah, I believe organomercury compounds are air-stable. I once planned a synthesis route that involved an organomercury intermediate (not even close to my first choice, but all other routes had failed, and I was getting desperate enough to consider working with such stuff), so I reviewed the relevant literature, and the compounds (internal arylmercury carboxylates, one could call them) were described without mention of anything that would suggest any sort of instability to air.

          I cancelled my plan after checking out the articles citing the original one, and finding that nobody who had tried could reproduce the preparation of those compounds.

      2. David Edwards says:

        And this is why I come here … I learn things.

    4. Tourettes of Chemistry says:

      RuBisCO – – Mg as the metal and O2 is a processed substrate that competes non-destructively with CO2 fixation – aka wasteful photo-oxidation – the enzyme handles the atmosphere as it is and is not inactivated

      CaC2 – rather stable non-biological acetylene precursor

      1. cookingwsolvents says:

        some Pd, Pt, Ir, and especially Au alkyls are air-stable under ambient conditions. Et4Pb and related main-group alkyls are, too.

  14. drocto says:

    on the 24th Jan you mentioned phosphorylation, and now the grignard. I would hazard a guess that you like many other medicinal chemists around the world are looking at the remdesivir synthesis route quite hard at the moment.

    1. Tourettes of Chemistry says:

      Thanks for the interesting observation on remdesivir. Since the Red Sox equipment truck just left on the 3rd for Florida, it sounds like guessing pitches is back in season:

      Swing and a miss.

      Keep in mind that Burgess reagent is very handy for some of those hostile hydroxyls noted just the other day and the current interest in anti-virals is very much at the fore in many minds as was just pointed out.

      Frankly, the SMILES –> trials commentary has been the most intriguing thing appearing In the PL’s recent past IMHO.

  15. Ted says:


    In the mid 2000’s we made most of the Buchwald ligands on scale using a combination of Grignard additions to “in situ” generated benzynes. The elegant part is that after the initial regiospecific addition to the benzyze, you recovered a grignard, which was then quenched by the appropriate phosphine chloride.

    Nothing like trying to get kg of Mg to go after you’ve already dumped a liter of dibromoethane in… 2-Me THF seemed to be more finicky as the solvent, even though it was typically drier than drumstock THF. More often than not, the trick was to make a starter batch in the lab (where ‘grinding’ Mg is easy) and then charging it into the mains. For a few of our multi-batch campaigns, we’d just preserve a few liters of our previous batch like a sourdough starter, just as Thomas described. All of this stuff had to be in Hastelloy reactors, as the Mg turnings would scratch glass something fierce.

    I never bought the magnesium nitride argument, mostly because I thought the documented temps of that were far in excess of reactor bake-out temps? Fun stuff, I learned to love the chemistry, as well as Knochel’s ‘turbo’ variants…


    1. milkshake says:

      An old trick with starting Grignard in not-so-anhydrous solvent was to add some Red-Al solution to it. It eats the moisture and reduces the oxide coating on Mg surface. Of course you have some reducing agent in there afterwards, in your Grignard. I suppose your starter batch was just drying the solvent/Mg. One thing you could have done is to install couple of activated alumina columns from MBraun in series – you can dry ten liters of THF from the original 50ppm waterdown to about 2-3 ppm, close to the Karl Fisher detection limit, within a reasonable time

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.