This is a guest post by Philip Moriarty, Professor of Physics at the University of Nottingham
Thank you for your statement which explains your position on the striped nanoparticles controversy and the reasons underpinning your decision not to disclose, from the start, your affiliation and your professional connection with the work of Stellacci et al. It is to your immense credit that you have apologised for this lapse in judgement and that you realise that it compromised the openness of the scientific debate. While I cannot agree with your position that there was no conflict of interest in posting comments without making clear your role as an editor of Nature Materials, I nonetheless welcome your admission that a disclaimer in your very first comment would have been appropriate.
I would also like to take this opportunity to apologise again for the rather ‘robust’ language in one of my responses to some of your comments at this blog. Despite standing firmly behind my criticisms of your support for the experimental and data analysis protocols of Stellacci et al., the tone of my language in that particular response was clearly not at all helpful. The following is a rather more ‘well-measured’ rebuttal.
Although your statement is an important contribution to the debate, you make a number of points with which I take particular issue as they do not provide an accurate assessment of our criticisms of the work of Stellacci et al. Moreover, you make strong accusations – such as claiming that we “misrepresent” the facts, that we “cherry-pick” results, and that our criticisms are “biased” – without providing any supporting evidence. We would have liked to debate these issues directly with you at your website but, unfortunately, you have stated that you “prefer to refrain from discussion on the matter”, a stance which, unfortunately, rather undermines the laudable claims earlier in your statement that you stand by your scientific comments and that science benefits from openness.
I simply cannot accept that we “misrepresent” the facts. It is unfortunate that you have decided not to engage in further debate because I would be very keen to see your evidence for this claim. Our criticisms of the work of Stellacci et al. do not focus on a single result in a single paper; as detailed at Raphael Levy’s blog, there are issues with the data in a considerable number of papers. The key direct experimental ‘evidence’ for the existence of stripes remains scanning probe microscopy (both STM and AFM) images. These images are awash with experimental artifacts, as again described in considerable detail at Levy’s blog. A number of scientists, myself included, have systematically critiqued the data. Perhaps most damningly, virtually identical ‘stripe patterns’ have been acquired on samples without adsorbed nanoparticles using the same STM as was used for the ‘striped nanoparticle’ imaging in the Stellacci group.
In what sense are we misrepresenting the facts?
It is interesting that you claim that we have only partially, rather than ‘holistically’ (to use your term), assessed the experimental evidence. A very similar argument has been put to me in e-mail correspondence with Francesco Stellacci (FS). (Note that Francesco has recently asked me to keep future e-mails between us confidential. This request will of course be honoured, but the discussion of ‘partial’ vs ‘holistic’ consideration of the data arose in earlier e-mails where confidentiality was not requested). In any case, not only are the SPM data riddled with artifacts, but the TEM images are far from compelling and apparently also suffer from a rudimentary artefact that is well known in the TEM community (see, in particular, this comment at Douglas Natelson’s blog). There is an interesting parallel here with the issues surrounding FS’s STM data acquisition and interpretation.
A core issue remains the lack of reproducibility; Stellacci et al. have not credibly reproduced the same stripe pattern in the same nanoparticle in two consecutive scans. Yu and Stellacci attempted to show reproducibility of features in their response to Levy et al. in Small but it can be conclusively shown that they cherry-picked the data so as to select a fortuitous ‘alignment’ of pixels of noise. (I have contacted the editor of Small to ask whether a Comment on this aspect of Yu and Stellacci’s response can be considered for publication, as it represents a highly misleading approach to the acquisition and analysis of STM data). I remain deeply surprised by your attitude, voiced in a comment at Levy’s blog, that you do not see why the lack of reproducibility of the features in Stellacci et al.’s STM data is an issue. This is a worrisome stance, to say the least, from an editor of Nature Materials and I cannot but mention again that it does not inspire confidence in editorial standards at the journal.
To reiterate — we have identified multiple examples of poor experimental technique and highly questionable data analysis spanning a series of papers. How does this misrepresent the facts or “cherry-pick” the data?
You might question why I did not mention the molecular dynamics simulations of Glotzer et al. which have been used to provide support for FS’ ‘striped nanoparticle’ interpretation of the experimental data. We stress that the simulations have, to the very best of our knowledge and experience, been carried out in a rigorous and systematic fashion, in stark contrast to the acquisition and analysis of the STM data. The key point is this: even if strong experimental evidence is found for the existence of stripes, this does not in any way vindicate the Stellacci group’s approach to the acquisition and analysis of STM data.Nor would it in any way support your argument that reproducibility is not a criterion for publication in journals of the calibre of Nature Materials.
Your counter-argument regarding Stellacci et al.’s re-use of data is similarly far from convincing. Dave Fernig has already described the deficiencies in your argument at his blog. It is also important to note, however, that the data re-use discussed in your statement is not an isolated case. There are multiple examples of the re-use of STM images in Stellaci et al.’s papers (without reference being made to the publication in which the data were originally published). Other examples, in addition to that which Fernig has highlighted, include the following:
– Both Fig. 1(a) and Fig. 8 of AM Jackson et al., J. Am. Chem. Soc. 128 11135 (2006) are taken from the same image as shown in Fig. 2 of Hu et al., J. Scanning Probe Microscopy 4 24 (2009)
– Fig. 3 of Y. Hu et al., J. Phys. Chem. C 112 6279 (2008) is a cropped version of Fig. 3 of Hu et al., J. Scanning Probe Microscopy 4 24 (2009)
– The central panel of Fig. 1 of A. Centrone et al. PNAS 105 9886 (2007) is a cropped and contrast-adjusted version of Fig. 1(a) of A. M. Jackson, J. B. Myerson, and F. Stellacci, Nat. Mat. 3 330 (2004). [The two nanoparticles in the bottom right hand corner of the latter have been used for the PNAS paper].
Finally, the most troublesome aspect of the striped nanoparticle story is the extent to which the peer review process and the associated editorial oversight have failed in this case. As you will know better than me, publications in NPG journals can make or break the career of a young researcher – or, indeed, more established scientists – because an inordinate amount of prestige and importance (and, via the UK REF, financial recompense to departments and schools) is given to papers in Nature and its ‘daughter’ journals. NPG journals unquestionably publish a great deal of exciting and ground-breaking science, and have traditionally been associated with exceptionally high reviewing standards. This case, however, has highlighted that even those journals at the peak of the ‘publications hierarchy’ are not immune to serious deficiencies in editorial and reviewing standards.
School of Physics and Astronomy,
University of Nottingham