Magnetic stirrers linked to issues with reproducing chemistry results

The magnetic stirrer plate is a stalwart of the chemistry lab, tirelessly turning stir bars in reactions around the world. But researchers in Russia now suggest that these trusty devices could be a significant source of variation between reactions, making procedures harder to reproduce (JACS Au 2025, DOI: 10.1021/jacsau.5c00412).

Valentine P. Ananikov of the Zelinsky Institute of Organic Chemistry first noticed the problem when a student in his lab was synthesizing palladium nanoparticles in chloroform. Each reaction produced nanoparticles of a different size, which affected their catalytic performance in subsequent procedures.

Determined to identify the source of this inconsistency, the team spent months checking conditions and purifying solvents and reagents, with no luck. Even simultaneously running half a dozen identical reactions in vials on the same stirrer plate produced inconsistent results. “We finally came to the conclusion that the position of the vials on the stirrer was affecting the outcome: the yield, the reaction rate, and the nanoparticle sizes,” Ananikov says.

His team systematically investigated this effect in other reactions, mixing multiple reactions at different positions and heights over a stirrer plate. These variations noticeably changed how the stir bar moved—sometimes it would spin freely, while other times it would grind against the side of the vessel or stop entirely.

In a reaction to prepare a palladium-on-carbon catalyst, for example, some positions on the plate resulted in the formation of particles in a narrow size range of 1–2 nm; others led to a wide distribution of 2–9 nm. Meanwhile, Suzuki-Miyaura cross-coupling reactions—using a commercial palladium catalyst—showed 15–20% variations in yield depending on the position. There was also variation between reaction vessels depending on how their shapes affected stir-bar movement.

When optimizing cross-coupling reactions, chemists often spend a lot of time trying a range of ligands, solvents, additives, and temperatures to identify the best conditions. But if the results of those trials are being skewed by stirring variations, the procedure could be far from optimal, Ananikov says.

The team used nickel granules to map the magnetic fields on half a dozen different models of stirrers and unsurprisingly found that the best stirring position is usually right in the middle of the plate.

Isn’t this just chemical common sense though? After all, every undergraduate is taught that stir bars should rotate freely in the middle of a vessel. Perhaps so, Ananikov says—but chemists can get sloppy, and it’s all too common in university labs to find multiple vials stacked inside a beaker atop a stirrer plate, their stir bars twitching pathetically.

“They make a valid point. It’s worth highlighting it, definitely,” says chemical engineer Nikolay Cherkasov, founder of Stoli Chem. The company makes flow reactors and other equipment, and Cherkasov previously developed a smart stir bar loaded with sensors to monitor reactions.

Cherkasov agrees with Ananikov that it’s useful to quantify the importance of stirring rates in reaction outcomes, not least because chemists often take uniform mixing for granted. But he points out that many reactions are less sensitive to stirring variations than those in Ananikov’s study, and he is not convinced that such variations have a big impact on reproducibility across the field.

The study wasn’t Ananikov’s first foray into rooting out irreproducibility. In 2019, for instance, he demonstrated that stir-bar contamination could smuggle unwanted catalytic metal nanoparticles into reactions. Following his stirrer plate investigation, he suggests that researchers should routinely include details such as stirring rates and vessel positions in their papers and even publish videos of their procedures.

Ananikov carried out an informal online survey of roughly 200 chemists from around the world, and many said that these kinds of variations were entirely normal. Indeed, about half of respondents said they are not surprised when a reaction cannot be reproduced at all. “But why should we observe a deviation in the yield of 20% and consider it normal? That’s too much!” he exclaims, adding that 5% should be the acceptable limit. “Chemistry,” he says, “should be an exact science.”