In science, there are results that are ho-hum and results that make everyone go ooooh. (There are also a few ho-hum results that still make everyone go ooooh.) In physics, that last category is dominated by single-molecule detection experiments.
Single-molecule detection is, basically, the limit for diagnosis. Imagine being able to pick up and read a single DNA strand or figure out if someone has an infection from the presence of a single protein. That is the noble objective. A lot of single-molecule detection results, however, are impressive but ultimately useless because they involve completely impractical conditions. This may be changing with a recent publication of single-molecule detection without any cheating.
Chris: Stop being mean
First off, single molecules are really hard to find. They are small and generally dont have much effect on the world. Yet we know that in biology, single molecules can have a huge influence—neurons are sensitive to single molecules. Some types of sperm get directional information from sensing single molecules. Nature, when needed, seems to have no trouble with single-molecule detection.
This is the kind of single-molecule detection that scientists aspire to.
To detect a single molecule, you need to get it to attach itself to something that is detectable and use the changes in that detectable object to infer that the molecule has arrived. That means the molecule has to cause a significant change in the properties of the object. That in turn suggests that the object also has to be quite tiny.
But something tiny is unlikely to stumble across something equally tiny like a single molecule. So this simply shifts the problem from finding a single molecule to waiting for the molecule to find you.
Put it like this: imagine that you need to find the one-and-only Ars Science editor, John Timmer, in New York. Following the approach described above, you would stand on a street corner with a sign saying, “Free beer for everyone named John Timmer.” Eventually you would capture our science editor. While waiting, though, you would have to chase away many, many false (beer loving) positives.
Grad students have died of old age waiting for the molecule to find their detector.
Upping the odds
A more effective solution to the science editor problem is to put a Timmer-trap on every street corner in New York. You then connect all the traps up, so that when one goes off, they all behave like theyve trapped a Timmer. Now, as soon as our editor leaves his home, we are likely to get a Timmer-signal lighting up the sky.
This is exactly what our intrepid scientists have done, though in their case they were searching for a protein.
The trick is all in the sensing surface. Start with a clean gold surface and grow a layer of polymer on it. Next attach your sensing molecule to the polymer layer—this is the molecule that is going to trap the protein we want to sense. Finally, fill in the gaps between sensing molecules. This prevents random non-target molecules from falling into gaps and spoiling the results.
The result is a surface that is a millimeter in size and has about a trillion sensing molecules poking out of it. All of this is electrically connected to a transistor. The current through the transistor changes when a molecule attaches to the sensing surface.
This approach really surprised me. To give you an idea of how unlikely an idea this is, let me put it in perspective: if one molecule should attach itself to the sensing plate, that should change the current in the transistor by one part in a trillion. For the currents the researchers used, we are talking 1 attoamp (10-18 amps). This is not easy to measure.
But this tiny change is not what the researchers observed happening. Instead, the current changes by about 0.2 microamps when a single molecule attaches itself to the sensing plate. Essentially, the influence of a single molecule is amplified by 100 billion.
Overly sensitive transistor
Why does the sensing layer respond so strongly? The researchers suspect, and use modeling to show, that the sensitivity is due to changes in the entire layer.
Before the target molecule binds, the polymer layer keeps itself together by a mechanism called hydrogen bonding. Basically, hydrogen atoms from neighboring polymers are weakly attracted to each other. These bonds form a network, which contributes to the electrical properties of the sensing plate.
When a single protein attaches to a single sensing molecule, it disrupts the hydrogen bonds locally. That has a knock-on effect, causing the hydrogen bonding network to reconfigure across a large fraction of the sensing plate. The researchers arent sensing the arrival of the molecule; they are sensing the disruption caused by the molecule arriving.
I suspect that this explanation will be the subject of debate for quite some time to come—certainly one of the reviewers was skeptical.
Regardless of the exact mechanism, this is quite a step in technology. The detection process is much closer to how natural single-molecule sensing works in living systems. That means there is a strong possibility that an overly sensitive transistor sensor may turn up in medical labs within a few years.
Nature Communications, DOI: 10.1038/s41467-018-05235-z (About DOIs)
[contf] [contfnew] 
Ars Technica
[contfnewc] [contfnewc]
