Every AFM Scan Is Worth Something Now – Even the Ones That Never Made a Paper

Author: Mansoor Nellikkal

Somewhere on an old hard drive, I still have several hundred Atomic Force Microscopy (AFM) images of perovskite and photovoltaic thin films. I know the exact folder structure. I could probably still tell you which scan came from which sample batch, on which day, with which tip.

Only one or two of those images ever made it into a published paper. Reviewer 2, inevitably, still had questions about that one.

That is not a confession of laziness. It is just how AFM research usually goes. You spend hours calibrating the instrument, dialling in the setpoint, chasing away drift and noise. You scan the same film a dozen times, at different spots, different scan sizes, different humidity levels, because thin films are inconsistent and grain structures don’t sit still and wait for you. Eventually, out of all that effort, one image looks clean enough, representative enough, “publication quality” enough, to go in a figure. The rest get filed away as “backup” and never looked at again.

At the time, that felt normal. A paper only needs one good picture to make its point. But looking back now, in an era where AI models are hungry for exactly the kind of structured, real-world image data that AFM produces, it stings a little. Every one of those unpublished scans came with metadata: scan parameters, tip type, environmental conditions, sample composition. Data that, on its own, could have helped someone else calibrate faster, train a model, or simply avoid repeating an experiment that didn’t need repeating. Instead, it’s sitting idle, format increasingly obscure, on a drive that will eventually fail or get lost in a house move.

I doubt my hard drive is unusual. Multiply it by every AFM lab that has ever existed, and the scale of what’s quietly going unused becomes enormous.

Forty years of images, and no shared home for them

2026 marks 40 years since Gerd Binnig, Calvin Quate, and Christoph Gerber introduced the first Atomic Force Microscope (AFM). The landmark paper[1] describing the instrument was published in 1986. Later that year, Binnig shared the 1986 Nobel Prize in Physics with Heinrich Rohrer for their design of the scanning tunnelling microscope (STM). Building on that breakthrough, Binnig, Quate, and Gerber developed the AFM to image insulating and other non-conductive materials beyond the reach of STM. Over the past four decades, AFM has evolved from a laboratory curiosity into a workhorse instrument across materials science, biology, and semiconductor research, enabling measurements ranging from nanoscale surface topography to mechanical, chemical, electrical, and magnetic properties.

And yet, as a community, we’ve never really solved the problem of what happens to all that data after the paper is published. It sits in proprietary formats, on personal drives, described (if at all) in whatever labelling convention made sense to one person on one day. Fragmented, inconsistent, and effectively invisible to anyone else who might benefit from it.

To mark AFM’s 40th anniversary, the Physical Sciences Data Infrastructure (PSDI) has partnered with Frontiers in Nanotechnology on a new Research Topic: AFM Data for the Community: Towards Reproducible, Interoperable, and Open Science. It’s now open for submissions, and it’s built on a simple idea: let researchers publish their scientific findings and deposit the AFM datasets behind them, properly documented, openly accessible, and reusable by others.The collection is co-edited by Prof. Lidija Siller (Newcastle University) and Dr. Rimal Isaac (Noorul Islam Centre for Higher Education) – and I am glad to be joining them as the co-editor. It’s a genuinely well-connected place to put your next AFM paper in front of the people already shaping this conversation.

What the editors are actually looking for

The scope is broad by design. Submissions are welcome across materials science, nanobiology, nanoelectronics, nanofabrication, and surface characterisation, as Original Research Articles, Data Reports, Methods papers, Reviews, or Perspectives. Reviews and Perspectives are encouraged to critically assess current data practices; Methods and Data-focused submissions should offer approaches the wider AFM community can actually adopt.

The one recurring theme editors want to see: AFM datasets treated as first-class scientific outputs, not an afterthought filed away after publication. That means manuscripts that show how the underlying data is structured, documented, and made reusable, ideally deposited alongside the paper.

Somewhere to put the images that never made the cut

That’s where PSDI’s Community Data Collections come in, specifically the Atomic Force Microscopy Database (AFMDB). It’s an open repository built for exactly this kind of data: any AFM instrument file format, plus the processed image outputs, described using a FAIR-compliant metadata schema. Anyone can deposit data. Anyone can search and reuse it. Deposition isn’t mandatory for submission, but it’s strongly encouraged, and it’s the kind of thing that makes a paper stand out in a collection explicitly about reproducibility.

For the processing side, PSDI also supports TopoStats, an open-source Python toolkit for processing AFM data at scale. It can batch process AFM datasets, identify and trace surface features, and generate image outputs and quantitative statistics, helping researchers apply more consistent analysis workflows and make results easier to compare across studies.

None of this requires giving up the “hero image” model of publishing. It just adds an option that didn’t really exist before: the ten, the hundred, the several hundred images that didn’t make the final figure can still have a life. They can sit in a repository with proper metadata, waiting for someone, a future PhD student, a data scientist training a model, a researcher trying to reproduce your calibration, to find them useful.

Why submit here rather than bury the data again?

Because the alternative is what this whole article started with: good data, quietly expiring on a hard drive, uncredited and unfound. This collection is a rare case where the “extra effort” of documenting your dataset properly is actually rewarded, with visibility in a dedicated 40th-anniversary collection, association with a strong editorial board, and alignment with exactly the FAIR and AI-ready data standards that funders and journals are increasingly asking for anyway.

I can’t get back the images already lost to a dead hard drive or a forgotten format. But the next batch, and the next manuscript, doesn’t have to end the same way.

Ready to submit? If you have an AFM paper, dataset, or even just an idea for one, send a short abstract to [email protected] by 16 September 2026, and we will get you an invite to submit to the collection.

Reference: [1] *Binnig, Gerd, Calvin F. Quate, and Ch Gerber. “Atomic force microscope.” Physical review letters 56.9 (1986): 930.