Faked figure sinks paper on potential new MRI contrast agent

langd5_v030i025.inddSurface chemistry journal Langmuir has retracted an article on a new MRI contrast agent — but only one of the authors agreed.

According to the notice:

The Editor-in-Chief of Langmuir, in consultation and agreement with one of the authors of the article, retracts the article “Development of Fe/Fe3O4 Core-Shell Nanocubes as a Promising Magnetic Resonance Imaging Contrast Agent” on the basis of reproducibility of the TEM image in Figure 1b of the article. The decision to retract the article as made after failure to obtain convincing raw data and images associated with Figure 1b from the authors of the article and following consultation with independent experts in electron microscopy. The questions about Figure 1 also raise questions about the conclusions within the paper.

Lyudmila Bronstein was the only one of the paper’s five authors to respond to our emails. According to her:

I agree with the Editor’s decision, this is the only thing I can say. Better talk to the journal editor.

Journal editor David Whitten cited confidentiality:

Unfortunately I am not at liberty to provide more information concerning the circumstances around the retraction due to my responsibility for editorial confidentiality.

Here’s the abstract:

Here, we report the synthesis, characterization, and properties of Fe/Fe3O4 core–shell nanocubes prepared via a simple route. It includes NaBH4 reduction of FeCl3 in an ethylene glycol solution in the presence of 2-mercaptopropionic acid (surfactant) and trisodium citrate (cosurfactant) followed by surface oxidation with trimethylamine N-oxide. The morphology and structure of Fe/Fe3O4 core–shell nanocubes were characterized using transmission electron microscopy (TEM), high-resolution TEM, selected area electron diffraction, X-ray powder diffraction, and X-ray photoelectron spectroscopy. All of the methods confirm a Fe/Fe3O4core–shell structure of nanocubes. Magnetic measurements revealed that the Fe/Fe3O4core/shell nanocubes are superparamagnetic at 300 K with a saturation magnetization of 129 emu/g. The T2 weighted imaging and the T2 relaxation time showed high MRI contrast and sensitivity, making these nanocubes viable candidates as enhanced MRI contrast agents.

The paper has been cited twice, according to Google Scholar.

Hat tip: Rolf Degen

21 thoughts on “Faked figure sinks paper on potential new MRI contrast agent”

  1. “editorial confidentiality”. That’s a new unwelcome euphemism. More transparency, and less excuses, please. Maybe another mess-up by scientists, but an equally bad lackof transparency by the editor. Science is sinking lower and lower by the day.

      1. @CR: very difficult to comment on pubpeer – heavily moderated. few of my comments filtered through and never appeared..too disappointed.

        1. try inputting comments using your account — unregistered comments are harder to be accepted — and avoid implying direct accusations, bad language, sarcasm, and focus on content issues and questions, then all comments should be fine. I never had any problems commenting in PubPeer, except when unregistered.

          1. alright then, can one register with anonymous identity? why don’t they disclose their identity if others are required to register with real identities? not fair, right?

          2. I understand your being extracautious, but this is the only way this system can work. Guess you would have to trust them on this, and you then understand very well why they did not yet reveal their own IDs. If you get in their shoes you will see that this is the only way it can keep on working in the long term, without being shut like SF and AS blogs in the past. Paul Brookes is a declared supporter of PubPeer.

          3. I have to agree with KK on this one, ot some extent. Who is “they” exactly? The “About” page doesn’t actually explain who owns, runs or moderates PubPeer (https://pubpeer.com/about). I believe that this is problematic. We can’t have anonymous ghosts moderating, although I do agree that civilized ghosts can rovide anonymous comments on papers.

          4. I agree however I am pretty sure they would have been taken offline already if everyone knew their IDs, like happened with Paul Brookes, who was also attempting to preserve his true ID. Obviously at some point soon their IDs will be disclosed (many probably know nowadays for you cannot hide such big things for too long) when PubPeer will be already settled by the support of many as to resist attacks.

          5. i thought the opposite. The blogs run anonymously would face problems rather than the ones with owners!Look at retraction watch! May be this is not an ideal comparison, though.

          6. Ideal comparison in these parameters is Abnormal Science Blog — lasted a few months and I am not sure what happened to the owner. Up to now PubPeer is the longest living and most popular, even if some will not trust the system. Retraction Watch is not in the same niche.

    1. If you get rid of confidentiality you prohibit whistle blowing. Or you limit it to people willing to throw away their career.

      Is that your intention?

      1. The term “confidentiality” is being totally abused by this editor, and you are also putting out of context. This is most likely publically funded research, so why the secrets? Surely the public and the scientific public have the right to know all details? Why, for example, does the retraction appear to have been more of a “negotiation” between one author and one editor? Why did all other authors refuse to sign and why was a retraction issued against their will, or without their approval (what happened to democracy in publishing?)? Isn’t this just editorial or publisher aggressiveness? Sorry, to say, but this story stinks of the lack of transparency and details. How is science expected to improve when the gatekeepers keep facts that should be in the public domain secret and don’t respond to queries by journalists and scientists? Confidentiality should be limited exclusively to whistle-blowing, but not to providing all other details of the case at hand.

    1. Western blot or other imaging techniques are not plagued. Since it is a visual, you can spot anomalies, duplications etc but they are far less susceptible to manipulation of data presented as bar and line graphs, % values etc where you will have to either have access to raw data or repeat the experiments to realize that they were manipulated.

      1. I am curious, how does one detect anomalies in SEM or TEM papers. Need some advice for the plant sciences. If anyne has any useful web-links that can give some training or guidance, this would be welcome.

        1. I guess some good sense, a good eye for pattern recognition and some knowledge about marks left by Photoshop and such is the way to spot manipulated images. I heard some alterations in the image contrast and other adjustments can reveal cropped parts and blurred areas which are indicative of manipulation. Yet I agree that images are just the easiest — fraud in numbers and statistics are probably more prevalent and much less detected. Often not even raw data can be any proof of reliability for random numbers can apparently be generated to fit patterns (like range and mean errors of mean, or linear regressions) by skilful mathematicians.

        2. Well, in this case it appears to me that figure 1b has been artificially created in photoshop. The figure 1b has four nano cubes, the top left nano cube looks like a prefect copy of the bottom right nano cube. This can be easily shown by using Photoshop to rotate the top right over the bottom left using the layers feature in photoshop with difference blending turn on. If the area you are interested in turns totally black then the two layers are identical. It does this for these two cubes. The text in the paper says “Figure 1b shows a representative HRTEM image of four nano cubes” it does not say the figure shows multiple copies of the same nano cube.

          1. Oh and the bottom left looks like an exact clone of the top left nano cube, this time rotated by 90 degrees clockwise and the top right looks like a flipped vertical with a 90 degree clockwise rotation of the top left. So in all to me I would say this is one nano cube cloned three more times.

          2. In fact even the core-shell nano cube may be a construction it self but that’s much more difficult to prove beyond doubt. From figure 1a (which is supposed to be a less magnified image of the same “sample”) we cannot anticipate the existence of such well defined core-shell structure.

            Also the XRD pattern looks somehow changed (look at the white background near the 311 label compared to the light grey background of the whole figure). In the XRD pattern, the 440 plane of Fe3O4 is missing, but it appears in the ED picture, wrongly labeled as the 400 plane (notice that XRD and ED are both reciprocal space techniques and the order of planes, from left to right in the XRD pattern and from center to border in ED must be the same!!!!) !!! Totally messed up!!!
            Figure 2 is also full of loose green and red spots between 3000 and 2500 cm-1… somebody was cleaning the whiteboard???

            In my opinion, the misconduct on this paper is not just about an assisted replication of inorganic material but it goes far beyond that.

Leave a Reply

Your email address will not be published. Required fields are marked *