Industrial catalysis is typically done by flowing gaseous or liquid reactants over a solid catalyst. Engineers strive to probe the thermodynamics of such reactions in situ to better model the reaction kinetics and optimize the reactor’s design, among other objectives. But in laboratory-scale systems such as microreactors or lab-on-a-chip devices, it’s often impossible to probe the temperature without disrupting the flow and thereby the reaction itself. Chemists led by UCLA’s Louis-Serge Bouchard have now developed a thermal mapping technique that extracts temperature information noninvasively from a nuclear magnetic resonance (NMR) spectrum. Here’s how: They place the reactor in a weak magnetic field gradient, apply a broadband RF pulse to tip the spins of the hydrogen nuclei, and record the NMR signal as the spins relax, precess, and decohere. Relatively cold molecules diffuse little in a given time, so in the magnetic gradient their spins precess with a range of resonance frequencies; the effect broadens the NMR line width. Warmer molecules cover more ground and experience more of the average magnetic field, so their line widths are narrower. With standard magnetic resonance imaging methods, the researchers could spatially map line-width-based temperatures to the different “voxels” of the reaction chamber and create an image. As proof of principle, they measured the temperature variations during the hydrogenation of propylene in a reactor NMR tube held at about 418 K. The two-dimensional map of variations about that temperature, shown here (with temperature averaged in the third dimension), was accurate to within 16 K, with a spatial resolution of about 100 µm. (N. N. Jarenwattananon et al., Nature 502, 537, 2013.)—R. Mark Wilson
