A new technique demonstrated by a Harvard group may provide environmental researchers, chemists and forensic investigators with a new tool to detect and identify trace molecules. The research combined the two separate concepts of surface-enhanced Raman scattering and multiphoton lithography, said Eric Mazur, physics professor and team leader. “A lot of innovation consists of putting two existing ideas together to enable something new. This is a prime example.” The result was a 27-fold improvement in Raman scattering and a molecule-identifying signal. This initial result can be improved through further optimization of the technique, Mazur said. Because it is sensitive to molecular bonds, a Raman signal can serve as a molecular fingerprint and allow the identification of unknown substances. The problem is that Raman scattering is small and the signal, therefore, weak. That makes detection of trace molecules virtually impossible. As the name implies, surface-enhanced Raman scattering boosts the Raman signal because of an interaction between molecules adsorbed to a surface and the surface itself. The enhancement can sometimes be quite large, increasing the signal by more than a billion times. It also can be much less than that. This close-up shows what could be the basis for a new trace chemical detection tool. The helium-ion microscope image shows a silver-coated surface-enhanced Raman spectroscopy substrate. Multiphoton lithography, performed with femtosecond laser pulses, selectively exposes photoresist. Molecules can then adsorb only at electromagnetic spots, leading to stronger and more repeatable total signal enhancement. Courtesy of Eric Mazur, Harvard University. This great variation means that the few sites with large enhancement contribute most of the signal. When there’s less than a single layer of molecules, which would be the case during trace analysis, such variation makes detection unreliable. One time, more of the molecule of interest will produce most of the signal – the next time, it may not. The Harvard group solved this problem through multiphoton lithography, a technique where the simultaneous absorption of two photons leads to a feature in a photoresist. The researchers first fabricated the surface-enhanced Raman substrate, using a femtosecond laser to create an array of microscopic cones spaced 500 nm center to center. They covered the cones with silver nanoparticles, leading to a substrate with an average millionfold Raman scattering enhancement. To isolate the hot spots, they coated the substrate with a resist layer thick enough to stop any molecules from reaching the surface. They then scanned the surface with laser pulses at 795 nm, almost twice the optimum resist exposure wavelength of 436 nm. Because of the silver, these pulses caused multiphoton exposure of the resist, but only at electromagnetic hot spots. The researchers then removed the resist from the hot spots and left it everywhere else, guaranteeing that only the spots with the highest surface enhancement were exposed for adsorption. When they tested the substrate, they found that it had 27 times the signal of an unprocessed one, as reported in the Nov. 18, 2009, Journal of the American Chemical Society. The group is now optimizing the process, and it hopes to eventually produce substrates with sufficient enhancement at every available molecular adsorption site for single-molecule surface- enhanced Raman-spectroscopy. “The ultimate goal is to enable new ways to perform trace-level detection of molecules with great specificity,” Mazur said.