At 7 kelvin — a temperature so frigid that even air would solidify — a solitary molecule rests on a copper surface. By most standards, it is among the tiniest entities ever intentionally examined: a ring composed of four carbon atoms and one nitrogen, with hydrogen atoms extending from its periphery. A tungsten tip hovers mere nanometres above it, gauging the quantum flow of electrons. Then, the laser fires.
The molecule rotates.
Not in a random manner, nor thermally, but in direct response to an infrared light beam fine-tuned to resonate with the exact frequency at which its nitrogen-hydrogen bond expands. The tip records the transition. The transition represents the spectrum. And for the inaugural time, a single molecule has delivered its own vibrational signature, one note at a time, to a patient listener able to perceive it.
The technique, devised by Shaowei Li and associates at UC San Diego and published this week in Science, combines two robust yet previously mismatched tools. Infrared spectroscopy has served as chemistry’s essential method for more than a century: shine light of the appropriate frequency on a molecule, observe its absorption, and uncover which bonds exist. The complication lies in resolution. Infrared light has wavelengths measured in micrometres, and the diffraction limit implies you can never focus it narrowly enough to single out a solitary molecule from the multitude. Scanning tunnelling microscopy, on the other hand, can visualize individual atoms with ångström precision. However, its sensitivity to vibrations has always been hindered by the limited energy windows that electron tunnelling can access.
Li’s team merged infrared excitation with tunnelling current detection, naming the resultant platform IRiSTM. The concept is refreshingly straightforward in theory: if you adjust your laser to a frequency that stimulates a specific molecular vibration, and that vibration induces even a slight movement of the molecule, the tunnelling current will vary. However, implementing this required eliminating numerous confounding signals, particularly the thermal expansion of the tip itself, which the researchers managed to suppress with an active piezoelectric feedback mechanism.
To confirm the approach, they began with a small target. The ethynyl radical, consisting of just two carbon atoms and one hydrogen, was deposited onto copper and illuminated across a broad infrared spectrum. At 3,169 wavenumbers, correlating with the carbon-hydrogen stretch, the molecule began rotating rapidly among its four equivalent orientations on the surface. Detune from that frequency and the rotation decelerates. Swap the hydrogen for a heavier deuterium atom and every peak shifts predictably to lower energy, as basic physics dictates. The isotope substitution performed its usual duty in spectroscopy: it verified the assignment without a doubt.
Having validated the method, the team proceeded to examine something biologically significant. Pyrrolidine — the five-membered ring central to the amino acid proline — is one of the most influential small molecules in structural biology. The “proline kink” it creates disrupts the regular configuration of proteins, causing bends and twists that shape enzyme active sites, membrane receptors, and the triple helix of collagen. Pyrrolidine’s ring is not fixed: it moves between two conformations, axial and equatorial, a behavior chemists refer to as ring puckering. At temperatures pertinent to biology, that puckering occurs spontaneously and continuously. At 6.3 kelvin on a copper surface, it requires a push.
The laser provides the impetus. Scanning nearly the entire mid-infrared range, the team assessed how different vibrational excitations influenced the switching rate between the two conformers. The resulting spectrum was remarkably richer than anything traditional methods had unveiled. In addition to the fundamental nitrogen-hydrogen and carbon-hydrogen stretches, the spectrum revealed overtones — the molecular counterpart of harmonics — at double and triple the fundamental frequencies, and combination bands where two distinct vibrational modes shared their energy. Some of these characteristics, notably a cluster between 4,500 and 5,500 wavenumbers attributed to modes coupling into the ring-puckering motion, were completely undetectable by standard infrared techniques.
That obscurity is significant. IRiSTM adheres to different selection principles than conventional spectroscopy, as the signal relies not only on whether a vibration absorbs infrared light, but also on whether it translates into perceivable nuclear motion. The carbon-hydrogen stretch, clearly discernible in bulk infrared measurements, was missing from the IRiSTM spectrum of pyrrolidine: it absorbs light sufficiently, but the resulting energy dissipates quickly into other pathways without driving the ring toward its conformational switch. The ring-deformation overtones behave oppositely: their substantial nuclear displacements couple directly into the puckering coordinate, allowing them to stand out distinctly despite being weak absorbers in ensemble measurements.
The team validated this observation by substituting hydrogens with deuteriums in controlled manners — replacing just the nitrogen-hydrogen, or substituting all eight carbon-hydrogens, or both. Each isotopologue