The report of a metal-free Suzuki-Miyaura reaction by the Hai-Zhu Yu and Hua-Jian Xu labs in January must be the organic chemistry sensation of 2021 (blog here). In it, palladium is replaced by an aromatic amine as an organocatalyst. The report has been met with disbelief though (example here) because in this reaction type the general consensus is that it is nearly impossible to exclude the presence of at least some metal.
An 11 person strong team from Hungary and Germany took it on themselves to check the published procedure and they report about it in an article on ChemRxiv. Co-authors Gonda, Tolnai and Novak are not newbies when it comes to catalysis, for example, back in 2010 they unmasked a reported copper-catalyzed Sonogashira as well, a pbb-level palladium driven reaction. And the news is again not good, palladium was again found to be present in "homeopathic" quantities.
In the Yu / Xu procedure the organocatalyst was prepared in a classic Buchwald-Hartwig reaction with a relatively high palladium loading. The Hungarian detectives argued that If palladium had sneaked its way into the coupling reaction it must have been via this attack vector. They decided to make the catalyst in three different ways.
First they repeated the catalyst synthesis as reported by Yu and Xu (2-methylbenzene-1,3-diamine ,PdCl2(PCy3)2 ,1-bromo-2-methylbenzene ). By ICP-MS they found 512 mg/Kg of residual palladium in this sample, a figure much higher than that reported by Yu and Xu. More damaging was the find of quantities of phosphorous as well.
Another batch was synthesized using 0.01 equivalent of palladium instead of 0.03 equivalent (bis(cinnamyl palladium(II)) chloride , tBuXPhos and a polyethylene oxide surface surfactant). The new residual palladium level was 10 mg/Kg without any phosphorous present.
The catalyst was also prepared without the use of palladium at all using 2-methylbenzene-1,3-diamine and di-o-tolyliodonium tetrafluoroborate. The three batches were then tested in a typical Suzuki reaction with batch A resulting in a 98% yield, batch B in a 93% yield and batch C reporting no yield. The last result would certainly put the organocatalytic Suzuki hypothesis to bed, there may be other research groups on the procedure replication trail, we will have to wait and see.