Fabian Wittmers

Podcast: Massenhaft Fake-Studien durch KI (2) - So arbeiten Wissenschaftsdetektive KI macht es Paper Mills leichter denn je, wissenschaftliche Arbeiten zu fälschen. Immer wieder werden solche Fake-Studien veröffentlicht - auch weil die Prüfmechanismen an ihre Grenzen stoßen. Doch es...

I was recently interviewed by German journalist Florian Sturm about my sleuthing work for a podcast episode: www.ardaudiothek.de/episode/urn:...
(it's unfortunately only available in German language)

Even the authors requested this article to be retracted in 2020, but didn't get a response from the journal. Yikes, Elsevier....

(5/5) There is still more papers from this papermill out there, but Wiley and Elsevier have recently retracted some of the leftovers that they somehow forgot to address in 2020.

(4/x) I find this worrying, because to me it suggests that the publisher did as little as possible to get to the root of this papermill. They retracted some papers, just enough not to get themselves into trouble, but showed no interest in actually cleaning up the scientific record.

(3/x) Elsevier had addressed a portion of the problematic papers, but 'missed' many that remained on the record for an additional 5 years. Had the publisher tried, they would have easily been able to identify them all, given Smut did almost all of their work for them already.

(2/x) This paper was one of dozens that @smutclyde.bsky.social flagged in 2020 when he uncovered a prolific papermill used primarily (but not only) by Chinese authors. I flagged additional issues in 2025, but this, and many papers of the same template were known to be problematic.

(1/x) This paper was retracted recently: Liu et al. 2018 (DOI: 10.1016/j.biopha.2018.02.006). I flagged issues in June '25, and it took ~8 months to get this clearly made up article retracted. But there is more to this paper. pubpeer.com/publications...

This dumpster fire of a paper straight from Elsevier's papermill hell 'Environmental Research' just got retracted. Took just 8 months to conclude that completely nonsensical citations and insane COI between Editor and Authors warrants some consequenes... pubpeer.com/publications...

There is also extensive evidence for severe conflict of evidence between the authors, especially Jörk Rinklebe, and the editor of this paper. This is one of dozens of cases involving Rinklebe and his collaborators...

Yang et al. 2023 (DOI: 10.1016/j.jhazmat.2022.130308), co-authored by none other than Jörg Rinklebe. pubpeer.com/publications...
I am not a statistician but to me the maximum and minimum should not be labeled by the same letter/group, if none of the intermediate values are...

Arguing with authors about how similar two Kiwi's can look: Iqbal et al. 2025 (DOI: 10.1016/j.ijbiomac.2025.148178). These must be images of the same fruit, in my opinion. (the rotation is also not a great sign...).
I am a matter expert due to years-long Kiwi consumption! pubpeer.com/publications...

(10/10) The International Journal of Biological Macromolecules, if it keeps going at this pace is on track to publish more than 700 articles with data/image integrity issues in 2026 alone. That is 1 journal in 1 year. The scientific community is facing a wave of BS we are not prepared for.

(9/X) The journal assured us last year they had made a lot of improvements to their manuscript assessment system to increase their detection of these image/data integrity issues. If this is what the improved system looks like then I don't want to know what it looked like before - did it even exist?

(8/X): We can't forget about manipulated XRD data. A staple in every Elsevier journal that publishes materials-science related data:
Zehra et al. 2026 (DOI: 10.1016/j.ijbiomac.2025.149441) pubpeer.com/publications...

(7/X): There is also a huge amount of papers with less extensive but highly problematic issues. For example: Mohandesnezhad et al. 2026 (DOI: 10.1016/j.ijbiomac.2025.149584)
Figure 8B has a very obvious overlap, representing different treatments.

(6/X): Lan et al. 2026 (DOI: 10.1016/j.ijbiomac.2025.149785) pubpeer.com/publications...
I am sure the authors will claim the duplication was an accidental error during figure assembly, but how does one accidentally rotate a panel??

(5/X): Wang et al. 2026 (10.1016/j.ijbiomac.2025.149529) - a different one than in (4/X) - is an absolute mess. Basically all fluorescence images have overlapping regions for different conditions. Must have been intentional in this case, given the extend of issues. pubpeer.com/publications...

(4/X): Wang et al. 2026 (DOI: 10.1016/j.ijbiomac.2025.149607) pubpeer.com/publications...
The duplication in Figure 5 is so blatantly obvious that I am 100% sure that no reviewer or editor has actually looked at this figure for more than a second.

(3/X): Meng et al. 2026 (DOI: 10.1016/j.ijbiomac.2025.149484) pubpeer.com/publications...
The noise signature in Figure 6 should be random, but it is largely repetitive between multiple samples. This indicates potential manipulation of the underlying trace...

(2/X): Osman et al. 2026 (10.1016/j.ijbiomac.2025.149633) pubpeer.com/publications...
Even one overlap like this would have me question this experiment. But in Figure 6, most panels appear to overlap, supposedly representing different conditions (clearly not...).
#ImageForensics

(1/X) I have found ~2 problematic articles / day in 2026 so far Elsevier's 'International Journal of Biological Macromolecules'. The journal is an embarrassment. This comes after hundreds of problematic papers reported to them in 2025. Here is a 'worst-of' 2026 (so far):

(3/3) I have also contacted the EiC of the 'Journal of Hazardous Materials', but have not received a reply either. I find it ridiculous that the deposition of raw data is not required.
"Data will be made available upon reasonable request" is a ridiculous phrase, it is frankly pathetic.

(2/3) I have contacted the corresponding author 3 weeks ago and again last week to request the underlying raw data for this study to reproduce this table. I have unfortunately been ignored in this effort. (The emails didn't bounce & the study is also not old, so the data should still be available).

(1/3) Ahmad et al. 2021 (DOI: 10.1016/j.jhazmat.2020.124852).
This is one of many Rinklebe co-authorship papers with various issues, most notably COI between editor and author. But there is also some rather odd repetitions in the summary tables. No raw data is deposited.
pubpeer.com/publications...

Does someone want to guess the senior authors name in this article? Majeed et al. 2024 (DOI: 10.1016/j.csite.2025.106802) pubpeer.com/publications... Check out the reference section for a very subtle hint:

Yes, according to the Editor-in-Chief:

"The authors provided some of the original data to the Publisher, but this was insufficient to address the concerns raised. The Editor-in-Chief therefore no longer has confidence in the results and conclusions reported."

A case from the 'strange' category: This XRD pattern looks like the noise signature was trimmed with a cookie cutter? Abdin et al. 2025 (DOI: 10.1016/j.foodchem.2025.143841) in 'Food Chemistry'. pubpeer.com/publications...
Not sure why someone would use this method. Maybe get rid of peaks?

@mumumouse2.bsky.social & I have, at this point, reported almost 500 papers w/ issues to Elsevier's International Journal of Biological Macromolecules. Barely any have been addressed. Nothing has changed since: Shanmugam et al. 2026 (10.1016/j.ijbiomac.2025.149319) pubpeer.com/publications...

Awual reviewed for the journal 'Environmental Research' extensive for many years and I reported a lot more papers with exactly the same issues ~8 months ago. So hopefully this is only the first to be addressed.

Paper retracted this week because the reviewer (presumably Rabiul Awual, who's shenanigans I have been looking into for a while now) requested at least 8 of his (unrelated) papers to be cited. Authors didn't question and complied. Also unauthorized authorship changes... pubpeer.com/publications...