The new device consists of a glass microscope slide that has metal leads placed on it. The gaps between the leads are filled with a film of gold nanoparticles. The researchers, led by Francesco Stellacci and Bartosz Grzybowski, begin by measuring the conductance of the film. They then immerse the whole device in water, remove and dry it, and then measure the conductance again. If the water contains toxic ions, like those of heavy metals, for example, the conductance of the film increases.

The device works because the nanoparticles are able to trap the toxic metal ions thanks to the fact that they are functionalized (or coated) with n-hexanethiol and alkanethiols terminated with ethylene glycol units. These units have a striped ligand structure that generates supramolecular pockets able to selectively trap positively charged ions. By varying the length of the nanostripes, the researchers are able to target different types of pollutant.

The team, which includes scientists from the Swiss Federal Institute of Technology (EPFL), Northwestern University and the University of Michigan, found that the device is particularly sensitive to methyl mercury – the most common form of mercury pollution that accumulates in fish, such as tuna. Indeed, the sensor can detect concentrations of the toxin down to levels as low as 600 methyl mercury ions per cubic centimetre of water.

The device could be a quick, easy and cheap way to measure toxins in rivers and fish, and would beat sending samples to the laboratory and waiting for the results to come back, says Stellacci.

“Our technique is also one of the first – if not the first – that can selectively bind metal toxins and that does not rely on a lock-and-key mechanism,” he told “We thus have a selective binding with what, apparently at least, is a ‘floppy’ system. Chelation in the biological world is also sometimes based on such floppy architectures, but ours is the first fully synthetic one.”

The team tested its technique by measuring methyl mercury levels in Lake Michigan and comparing its results to those undertaken by the US FDA using conventional methods. The two analyses were in agreement. A second test on mosquito fish from the Everglades backed up measurements made by the US Geological Survey. “We made measurements on fish tissue that had been dissolved in acid to see if we could detect even very minuscule quantities of methyl mercury,” said Grzybowski, “since we know that this fish is too low in the food chain to accumulate much toxin.”

The team, which reports its device in Nature Materials, would now like to better understand the secret behind the nanoparticles’ selectivity in trapping toxins as the length of the ligand stripes is changed. “Another important point is the remarkable combination of selectivity and sensitivity that the nanostripes have,” said Stellacci. The next grand challenge for Bartosz and I is surely to build a nanostripe-based ‘nose’ that can analyse different metal ions at once.”