Curtain falls on controversy over creation of metal-oxo complexes as authors publish new work refuting their previous claims

A long-running saga in the rarefied world of metal–oxo chemistry appears to be drawing to a close following the voluntary retraction of three controversial papers – that some critics insist should never have been published in the first place. This was closely followed by the publication of a new paper that sets the record straight.

In 2004 a group led by Craig Hill, a respected and distinguished inorganic chemist at Emory University in Atlanta, US, published a paper in Science describing the preparation and characterisation of a compound containing a single oxygen atom bonded to a platinum.1 The paper caused immediate controversy within the transition metal community. While such terminal oxo bonds are commonplace for transition metals occupying the early and middle groups in the periodic table, groups three to eight, there has been a long-standing consensus that such bonds are not possible – except in a small minority of well-defined cases – for transition metals in the higher groups, beyond group nine. It is accepted that there is an ‘oxo wall’ at some point between groups eight and nine beyond which such atomic collusion is impossible because of excessive electronic repulsion.

Forbidden configurations

Hill’s team constructed huge polytungstenate complexes around the platinum metal centre, which they believed had stabilised the previously forbidden configuration. Two further papers were published in the Journal of the American Chemical Society, in 20052and 20073, describing similar structures containing terminal palladium-oxo and gold-oxo moieties. For each structure, exhaustive supporting information was provided, including x-ray crystallography, NMR and neutron diffraction data.

‘They provided an enormous amount of data using many different techniques,’ says organometallic specialist Clark Landis at the University of Wisconsin-Madison, US. ‘If one were to look at any one of the techniques, it might have been possible to pick holes in the analysis, but when taken together all the data could seem compelling and the referees decided that the results should be published. You need an extraordinary amount of data to be convincing, and they did provide a lot of data.’

Eyebrows  remained raised, however. ‘I am sure people were not entirely convinced, and some people were more vocal about it than others,’ Landis says.

Indeed, Hill himself admits that he and his colleagues felt the same. ‘The entire six-institution team – most being unquestionably top investigators in their respective disciplines – from the beginning of this story nine years ago were quite sceptical,’ Hill tells Chemistry World. ‘We were highly so at Emory University and that’s why I brought in all these expert collaborators. However, the initial evidence was unusually broad – 17 techniques used on the initially assigned gold-oxo species – and internally consistent, thus we all agreed to proceed to publish.’

Oxo wall still stands

Now, however, in the light of new experiments Hill accepts that the interpretation of the data was incorrect – oxo complexes with metals beyond group 9 are still not possible. He has published a new paper in Inorganic Chemistry4 titled: ‘Revisiting the Polyoxometalate-Based Late-Transition-Metal-Oxo Complexes: The “Oxo Wall” Stands’. The team describes new x-ray and NMR studies on the structures and concludes: ‘These and considerable other data collectively indicate that previously assigned terminal Pt-oxo and Au-oxo complexes are in fact co-crystals of the all-tungsten structural analogues with noble metal cations, while the Pd-oxo complex is a disordered Pd(II)-substituted polyoxometalate. The neutron diffraction data have been re-analysed, and new refinements are fully consistent with the all-tungsten formulations of the Pt-oxo and Au-oxo polyoxometalate species.’

Hill says that retraction of the earlier papers ‘was the only course’ thanks to the new data. ‘All 100% of the original data in all three papers are correct but the interpretation was wrong,’ he says. ‘Both we, the lead authors of these papers and the editor of the Journal of the American Chemical Society felt it was best to retract the initial papers for two reasons: one, we just published the new paper which not only gives considerable new data on these materials (originally formulated as noble metal terminal oxo compounds) but also refers to the data in the original papers and two, if we didn’t retract them, all the authors, but particularly me, would doubtless be getting emails on the initial papers for years to come. This would be a nightmare.’ The two JACS papers have been formally retracted, the Science paper’s retraction is being finalised.

Science works

Hill says that for him the episode is a demonstration of the scientific process working well. ‘We did make the wrong assignments of these compounds originally, but it is also we who got it right after another major, diverse and multi-year experimental effort.’ He adds: ‘It’s always a bit of a bruising experience when a hypothesis isn’t correct. I can’t speak for all my co-authors, but for me the experience has been much more part of the healthy – but sometimes harsh nature of the – scientific research process.’

Landis agrees that the saga demonstrates that science works. ‘I am rather encouraged by these developments and the way it has played out in many ways. First of all there is an ethical issue of what is the proper course of action: should one retract papers like this? An argument is made with good justification that the data are correct, no plagiarism, no misrepresentation, no lack of honesty. Under these circumstances retraction was the logical way to go as the only way of purging conclusions, which are incorrect, from the literature and do not continue to be cited.’

Ulrich Kortz, an expert in polyoxometalate (POM) chemistry at Jacobs University in Bremen, Germany, is scathing about the publication of these papers. ‘To all POM synthesis experts it was clear from the beginning that there was a problem,’ he says. It is a miracle to me how this work could be published in the first place. Clearly, the refereeing process was a complete failure. Not only once, but three times. This is unbelievable, especially for top – really? I start to wonder – journals such as JACS and Science.’

Do the retractions show that the system works? ‘For me it is very frustrating that it took eight years to reach this point,’ Kortz says. ‘Clearly, the system works, but not well.’