My most cited paper of the last two years is wrong

Academics are assessed in a number of ways, and one proposal that keeps coming up to assess us is via citations. The idea being is that if nobody cites your work (i.e. refers to your paper in their paper) you must be doing boring work, but if your paper is being cited by many others then it is clearly popular so must be good. And almost whenever someone suggests this, it is pointed out that papers that are obviously wrong can be highly cited, particularly if they are published in a prestigious journal. This is a fair point. For example Wolfe-Simon et al.‘s Science paper of 2011 has been cited over 400 times, and is clearly wrong, as a number of the citing articles state.

My most highly cited research paper of the last two years is on the stratification — i.e. layering— of mixtures of small (shown in yellow) and big (blue) particles. The layers form during drying of an initially liquid thin film containing the mixture of particles; when the water has Captureevaporated we are left with a layer of small particles on top of a layer of big particles.

This work resulted from a EU-funded project. Right at the start of the work the postdoc paid under this grant and I, sat down and thought he should do computer simulations of mixtures of small and big particles in drying films. But we had a problem: simulating both the liquid (water) and the particles is very hard, you can easily end up simulating a million water molecules for each particle and simulating a thousand particles plus a billion water molecules is completely impracticable.

In science as in life, you do what you can do. So he simulated a model in which we just replaced the water molecules with a uniform background, and almost straightaway found stratification.

We thought this was cool, and experiments by the experimental postdoc also found stratification — in a real system with real water molecules. So we published the computer simulations and experiments in the prestigious physics journal Physical Review Letters. This paper has had quite an impact, which is both pleasing and good for my ego.

At the time we knew that throwing away the water was a risky approximation, and in the paper we confessed: “In our simulations we neglected any effect due to hydrodynamic flow of the solvent [the water].”

Early this year, I returned to worrying about this neglect of the water, and a friend of mine and I have now worked out the effect of the flow of water during evaporation. And it is a big effect. We have just published this and that paper adds one more cite for the original paper in Physical Review Letters. Clear evidence that you can boost your citation count by publishing papers with parts that need correction.

Now, almost all theoretical or computational physics papers published are wrong in the sense that they work with a model that leaves out important details of the real world, so our Physical Review Letters paper is not exceptional in being wrong in this sense. But I think this the first time I have published a high-impact paper that made this bad an approximation, and then corrected the approximation two years later. It is kind of fun, you first get to show people an exciting discovery, then later get to show them how you were a bit wrong first time round.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s