In March 1989, two chemists, Stanley Pons and Martin Fleishmann, stunned the world by claiming that an electrical current from a palladium electrode immersed in a test tube of water had resulted in nuclear ‘cold fusion’ at room temperature.¹ For many years, attempts to understand and reproduce this seemingly impossible result- with sketchy information from the inventors- continued feverishly, until finally being debunked and condemned to the archives of pathological science. Smoke and mirrors at the palladium? – perhaps.

However, it’s not the only occasion that palladium has caused controversy.  In 2005 Arvela et al re-evaluated their palladium-free Suzuki chemistry, only to find that palladium contamination at 50 ppb in the commercial sodium carbonate they used as base in their reactions was responsible for generation of the biaryl not, as previously published, a transition-metal free pathway.² Other metals have also pulled the wool over the eyes of unsuspecting researchers. A review paper by Kazuhiko Takai some years ago describes several examples of reproducibility issues in organometallic chemistry that were eventually attributed to the presence of traces of a second metal. Bolm has also published a review specifically on the effect of trace metal impurities in catalysis.³ In addition to reagents, metal contamination in intermediates and solvents can also be problematic and reaction vessel history can be a source of active rouge-metals.

A few other notable cases involving other metals include the chromium(III) catalyzed Nozaki-Hiyama-Kishi coupling of an aldehyde with an allyl- or vinyl halide that only worked effectively with traces of nickel impurities, iron catalyzed coupling reactions in which, according to Bolm and Buchwald, are most likely driven by traces of copper, and catalytic amounts of lead present in metallic zinc reducing the reactivity of the latter towards iodoalkanes in the Simmons- Smith reaction.⁴

A number of seemingly ‘palladium free’ or ‘metal free’ catalytic reactions have been described in the literature, however in a kind of parody of homeopathy, PTFE magnetic stirrer bars can become impregnated with metals- almost like retaining a memory that remains ready and waiting for that next catalytic reaction. A recent paper by Ananikov et al with the unnerving title ‘Phantom Reactivity in Organic and Catalytic Reactions as a Consequence of Microscale Destruction and Contamination-Trapping Effects of Magnetic Stir Bars’ demonstrated that surface defects in a stirrer bar can trap nanoparticulate metals and, in the case of palladium, the residual activity is sufficient to catalyze a “metal free” Suzuki cross coupling.⁵ They also found that these metal particles could leach into reaction mixtures and were detectable by ICP-MS. In fact, it is now possible to use transition metal impregnated Teflon coated stirrer bars as a reusable catalyst system.⁶

Akira Suzuki shared the Nobel prize for chemistry in 2010 for his contributions to metal catalysis, specifically the development of the palladium-catalysed cross coupling reaction that bears his name. The Suzuki reaction (or more correctly Suzuki-Miyaura reaction), discovered in 1979, has been run millions of times in a high throughput capacity to make compounds for biological screening in drug discovery and is also used in many industrial manufacturing routes to important pharmaceuticals and agrochemicals. For example, the broad-spectrum crop fungicide Fluxapyroxad (Xemium, BASF) is manufactured via a late stage 1000 tonne scale Suzuki coupling reaction (Figure 1). An extremely efficient reaction- really the only fly in the ointment, so to speak, in the cost of palladium and the long-term sustainability of metal-catalysed reactions.

In light of all this history it would take a brave person to claim that you do not need palladium for a Suzuki-type cross coupling reaction. And you would think that the barrier to publication of such a claim would be extremely high. That said, in January of 2021 a group led by Hua-Juan Xu from the Hefei University of Technology reported in Nature-Catalysis an amine-catalyzed coupling of aryl boronic acids with aryl halides (the standard Suzuki reaction coupling partners) without the need for palladium (or metals of any kind- a truly organocatalytic process).⁷ Understandably this paper was received with a great deal of scepticism by the wider community- with much discussion on social media platforms. Several popular science news websites reported the storey almost in a parody of David and Goliath. As was the case with cold fusion, many academic and industrial groups began repeating the work and reporting back on their findings. A concern that began to propagate fairly rapidly was that the amine organocatalyst ‘star of the show’ (Figure 2) was prepared via a palladium-mediated coupling reaction. Quite rightly chemists began asking if low levels of residual palladium could still be present (and active) after purification of the organocatalyst. However, in science every hypothesis needs to be considered and tested with full due diligence, despite prejudices and a strong history of conflicting prior art. Xu’s group ran a number of experiments to rule out palladium contamination from synthesis of the amine- enough to convince those involved in the peer review process that the results were real- however these experiments now seem likely to be false negatives.

Several groups have gone public with their conclusions and the general consensus is that a during purification of the amine, prepared via a palladium-catalysed Buchwald-Hartwig C-N cross coupling reaction (as descried in the Nature catalysis paper), stable palladium complexes are formed with the tricyclohexyphosphine (PCy₃) ligand used as an additive in the Buchwald reaction. These complexes were readily entrained during chromatographic purification of the organocatalyst and it is postulated that it is these (very active) species that drive the catalysis. Pre-catalysts that contain PCy₃ ligands are known to be active at very low palladium loadings. In fact ligand design has enabled palladium loadings as low as parts per billion to be effective for cross coupling reactions- a major advantage for large scale reactions.

Kazunori Koide from the university of Pittsburgh estimates that a 5 mol% catalyst loading of Xu’s amine containing 65ppm of palladium equates to 9.3 x10-4 mol% of palladium present in the reaction mixture. This gives a catalytic turnover number of at least 108 x10³ suggesting the residual palladium species were indeed very active.  Koide has developed fluorometric methodology for quantifying trace palladium in complex matrices, including synthetic organic materials, a technology particularly useful for studying catalytic reactions.^8a

In hindsight the amine should have been synthesised using procedures that didn’t involve metal catalysis. An easy thing to say but perhaps not an easy thing to do. Currently (July 2021) an editor’s note on the Nature articles webpage alerts readers that the conclusions of the paper are ‘subject to criticisms that are being considered by the editors ‘. It should be noted that the papers disputing Xu’s claims have not, at the time of writing, undergone peer review or been accepted for journal publication- they are, however, available on the publicly accessible Chemrxiv pre-print server.

What this chain of events does highlight is the need for standardisation and guidelines for the development of new catalytic processes, most important of which is identification of any extrinsic adventitious metals that could catalyse the reaction. Using a catalytically active metal to prepare a putative organocatalyst that is known to effect the transformation being investigated (for example palladium in a C-C cross coupling reaction) is also a definite red flag. Other recommendations put forward by Novak et al in their analysis of Xu’s results include using alternative sources of metals, bases and any additives in ultra-high purities (and cost!).^8b Pre-treating mixtures with metal scavenger resins is also a pragmatic approach. Reaction components should also be analysed for a range of trace metals known to be catalytically active. Good batch history records for substrates and back-traceable handling are standard good laboratory practices (no sticking that nickel spatula in the bottle!). Solvents and volatiles reagents should be distilled before use.

Needless to say, new stirrer bars should always be used vide supra and the integrity of ancillary equipment (needles, vessels etc.) scrutinised. Elevated reaction temperature and/or corrosive reagents are a possible source of metal contamination. New or rigorously cleaned equipment would be preferrable- and for small-scale work fresh reaction vials used for each new reaction. Running blank reactions without catalyst to eliminate background ‘noise’ and duplicate or triplicate reactions in screening reactions should be done as part of the standard workflow.

All that said, Xu’s paper was an interesting read with very detailed in-silico modelling of a putative organocatalytic mechanism. The possibility always remains that there’s more going on below the surface. Real magic, smoke and mirrors or just another night at the palladium?

**This article was previously published in Chemical Knowledge Magazine under the title “Palladium: the controversial element that cannot be ignored” June/July 2021 ** Link Here

References:

1. Too hot to handle: the race for cold fusion, Frank Close, W. H. Allen publishing, 1990, ISBN 1 85227 206 6.
2. A reassessment of the transition-metal free Suzuki-type coupling methodology: R. Arvela et al, Org. Chem. 2005, 70, 161-168.
3. Trace amounts of second metal elements can play a key role in the generation of organometallic compounds: K. Takai, Chem. Soc. Jpn. 2015, 88, 1511-1529; Trace metal impurities in catalysis (tutorial review): C. Bolm et al, Chem. Soc. Rev. 2012, 41, 979-987.
4. Catalytic effect of nickel(II) chloride and palladium(II) acetate on chromium(II) mediated coupling reaction of iodo olefins with aldehydes: Y. Kishi et al, Am. Chem. Soc. 1986, 108, 5644-5646; On the role of metal contaminants in catalyses with FeCl₃: S. Buchwald et al, Angew. Chem. Int. Ed. 2009, 48, 5586-5587; A dramatic effect of a catalytic amount of lead on the Simmons-Smith reaction and formation of alkyl zinc compounds from iodoalkanes. Reactivity of zinc metal: activation and deactivation: K. Takai et al, J. Org. Chem. 1994, 59, 2671-2673.
5. Phantom reactivity in organic and catalytic reactions as a consequence of microscale destruction and contamination-trapping effects of magnetic stir bars: V. Ananikov et al, ACS Catal. 2019, 9, 3070-3081.
6. Turning Teflon-coated magnetic stirring bars to catalyst systems with metal nanoparticle trace deposits – A caveat and a chance: C. Janiak et al, Catal. A 2012, 178–183.
7. The amine-catalysed Suzuki-Miyaura-type coupling of aryl halides with arylboronic acids: Hua-Jian Xu et al, Chem. 2021, 4, 71-78.
8. a) K. Koide et al, https://doi.org/10.26434/chemrxiv.14423579.v2; b) Z. Novak et al, https://doi.org/10.26434/chemrxiv.14071247.v1; c) R. Bedford et al, https://doi.org/10.26434/chemrxiv.14237288.v1.