Looking back: How was cross-coupling invented?

I was looking back at some reviews of the major cross-coupling reactions, and found that Ei-ichi Negishi's Nobel prize lecture is available both as a recording (https://www.nobelprize.org/prizes/chemistry/2010/negishi/lecture/) and also as a paper in ACIE (https://onlinelibrary.wiley.com/doi/abs/10.1002/anie.201101380). Similarly, Tamao (in the middle of the name of the Kumada-Tamao-Corriu cross-coupling reaction) published a short retrospective on their discovery of cross-coupling with organomagnesium reagents.

 Background:

The idea behind cross-coupling reactions did not spontaneously arise in anyone's head. The general idea of a "cross-coupling reaction" had existed since Victor Grignard's time, with the reaction of the nucleophilic organomagnesium reaction and an electrophile. However, these reactions generally didn't work. If you mix an organometallic reagent and a C(sp2) halide, you generally get no reaction. If you mix an organometallic reagent and a C(sp3) halide, you might get the reaction (likely through an SN2), but most of the time you get byproducts, such as elimination. 

 So what reactions did exist? Well, years prior Henry Gilman had been developing alkyl cuprate reagents, which were soon shown to be capable of doing SN2 reactions with alkyl halides (the Corey-House reaction). In 1959, Smidt reported a palladium-catalyzed oxidation of alkenes to carbonyls, which would soon be known as the Wacker oxidation . In 1965, Tsuji figured out that the same general idea could be used with allyl palladium complexes to form C-C bonds, leading to the initial discovery of what would become the Tsuji-Trost reaction. Corey of course joined in too, and he and Martin Semmelhack developed some reactions using allyl nickel complexes. Meanwhile, the Kumada group had developed a hydrosilylation of olefins using nickel catalysts and the fledgling Negishi group was developing reactions with boranes. 

Original references:

Corey-House reaction:  https://pubs.acs.org/doi/abs/10.1021/ja00991a049 and https://pubs.acs.org/doi/abs/10.1021/jo01348a012

Wacker oxidation: https://onlinelibrary.wiley.com/doi/10.1002/ange.19590710503 

Tsuji-Trost reaction: https://www.sciencedirect.com/science/article/abs/pii/S0040403900716741?via%3Dihub and https://pubs.acs.org/doi/abs/10.1021/ja00782a080

Corey and Semmelhack allylnickel: https://pubs.acs.org/doi/10.1021/ja01011a035

Kumada hydrosilylation: https://pubs.rsc.org/en/content/articlelanding/1970/c2/c29700000611 

Based on the account by Tamao, the plan for the Kumada group was to do some mechanistic studies on the nickel-catalyzed hydrosilylation. They originally expected to get the hydrosilylation product, but the major product actually had swapped out a chloride for a hydrogen, as if the nickel catalyst had done chlorosilylation and then reduced the alkyl chloride. To this end, Tamao and others began making a ton of relevant nickel-phosphine complexes: Ni(dppe)Cl2, Ni(dppp)Cl2, Ni(dppb)Cl2, Ni(dppf)Cl2, etc. Then, the goal was to make σ-organometallic complexes: complexes where carbon was directly bonded to the nickel. At the time, Tamao writes, there were approximately 20 known species like this, grouped into Ni(Ar)X(PR3)2, CpNi(alkyl)(PPh3), and Ni(bpy)(alkyl)2. That's not a lot of complexes. There were 3 key insights into the reactivity of nickel complexes gained from synthesizing a bunch of relevant complexes. First, reacting nickel halides with Grignard reactions is a general method to form Ni-C bonds (via transmetalation, as is know known). Second, alkylnickel complexes can be stabilized by bidentate ligands like bipy. Third, if you have 2 organic groups on the nickel, they can couple together to make a C-C bond. This last one was actually figured out by Uchino, Yamamoto, and Ikeda, who coupled two ethyl fragments together in a funky reaction with Ni(Et)2bpy and an aryl chloride (https://www.sciencedirect.com/science/article/pii/S0022328X00844757): 

That led to the idea that you could make the nickel σ-organic complex using a Grignard reagent, then add an an aryl halide and get selective bond formation between the aryl halide and the Grignard reagent. As Tamao, Sumitani, and Kumada write in the seminal 1972 paper (https://pubs.acs.org/doi/10.1021/ja00767a075):

"Thus, a dihalodiphosphine nickel reacts with a Grignard reagent to form the intermediate diorganonickel complex 1 which is subsequently converted to the (halo)-(organo)nickel complex 2 by an organic halide. Successive reaction of 2 with the Grignard reagent forms a new diorganocomplex 3 from which the cross-coupling product is released by the attack of the organic halide and thereby the original complex 2 is regenerated to complete the catalytic cycle."

This is what that looks like in the original paper, before chemdraw:

And this is what it looks like in chemdraw: 

 

I have intentionally not mapped it onto a catalytic cycle drawing, because there are some missing pieces. First, Ni(0) intermediates are not directly drawn, because at this point there is not sufficient evidence to propose such as cycle. With hindsight, we can easily propose that reductive elimination to form R-R provides a Ni(0) species that can then undergo oxidative addition with Ar-X. At the time, they could identify R-R as a byproduct, but proposed it formed after Ar-X engaged with the substrate. 

I feel confident saying that this is the first true cross-coupling paper, because it involves a transition metal catalyst forming a carbon-carbon bond selectively. 

After this, Negishi is one of many labs to become interested in the possibilities of other similar reactions, and starts conducting research into transition metal reactions. First, they try using alkenylboranes and borates. Negishi, after all, had originally been doing boron chemistry, so he's trying this out before the Suzuki reaction is established! Unfortunately, none of their efforts are successful, something that he attributes in hindsight to conducting rreactions at room temperature in THF. So, in 1976, what metals does the inventor of the Negishi reaction, the palladium catalyzed reaction with organozinc reagents turn to?

Nickel-catalyzed cross-coupling of aluminum reagents!

https://pubs.rsc.org/en/content/articlelanding/1976/c3/c3976000596b

Aluminum reagents can be nicely synthesized from alkynes using aluminum hydrides, which is the general category of reaction that Negishi was good at- he was making all sorts of alkenyl metal species previously for the boron chemistry. One other major development from this paper is the attempted use of a palladium catalyst. I'd like to draw attention to this conclusion:

"However, the palladium-catalysed reaction is much slower than the corresponding nickel-catalysed reaction and does not appear to offer any advantage over the latter."

Eventually, however, in attempts to develop a more stereocontrolled reaction (to synthesize E and Z alkenes), Negishi found that the palladium version, while slower, retained most of the original alkene geometry, leading to cleaner products. 

The final step in the journey was to screen other metals beyond aluminum (https://pubs.acs.org/doi/10.1021/ja00475a059). Using the same general idea- take alkynes, react them with aluminum hydrides, and then transmetalate to make different alkenyl metal species- the Negishi lab looked at Li, Mg, Zn, Hg, B, Al, Si, Sn, and Zr reagents. The idea was to couple different metals with an aryl iodide and see what works:

At this point, the year is 1978, six years after the Kumada/Tamao and Corriu papers came out, and the field is exploding in popularity. Unfortunately for Negishi, the use of allyltin reagents had been reported a year prior, but hits for organozinc and organoboron reagents were very useful. Eventually, Suzuki and Stille would engage in further development of cross-coupling reactions using organoboron and organotin reagents, respectively. At this point, the year is 1978, six years after the Kumada/Tamao and Corriu papers came out, and the field begins exploding. 

Caveats:

This is not intended to be a comprehensive history of the development of cross-coupling reactions, and I'm leaning heavily on the descriptions by Negishi and Tamao, which are also not comprehensive. I'm completely neglecting the development of the Heck reaction and many other groups that were also heavily involved the development of both cross-coupling reactions and organometallic chemistry as a whole. However, I think it is a helpful collation of the original publications related to cross-coupling reactions and a guide to how the field was invented. No discovery is made from nothing, and while the field of cross-coupling reactions did not exist in the same way prior to 1970, no one involved in its development had some immeasurably deep insight that instantly revolutionized chemistry. Progress was made one step at a time, incremental move by move.