Sorting of Enhanced Reference Raman Spectra of a Single Amino Acid Molecule Thibault Brule ́ , † He ́ le ̀ ne Yockell-Lelie ̀ vre, †,‡ Alexandre Bouhe ́ lier, † Je ́ re ́ mie Margueritat, †,¶ Laurent Markey, † Aymeric Leray, † Alain Dereux, † and Eric Finot* ,† † Laboratoire Interdisciplinaire Carnot de Bourgogne, UMR 6303 CNRS, Universite ́ de Bourgogne, F-21078, Dijon, France * S Supporting Information ABSTRACT: In this contribution, we report the identification of the principal reference Raman spectra of a single cystein molecule. To that purpose, we design an active Surface Enhanced Raman Spectroscopy (SERS) template based on surfactant-less Chebyshev nanoparticles operating in a microfluidic platform. A principal component analysis is obtained from fluctuating spectra to sort the reference spectra of cystein. The assignment of Raman bands brings new insight into the conformation of an amino acid adsorbed onto gold nanoparticle. ■ INTRODUCTION As a very sensitive spectroscopic technique capable of providing precise chemical information on trace amount of molecules, surface enhanced Raman scattering (SERS) is a powerful analytical tool used in biochemistry and biomedical research. 1-3 In this context, SERS identification of proteins is a formidable task because of their complex structures and often dynamic nature. It was shown recently that the SERS spectrum of a protein 4 is essentially driven by all the amino acids in the vicinity of the protein-metal interface. 5 Experimentally, a spectrum is obtained from either an integrated response of a relatively large amount of amino acids adsorbed as a monolayer over a millimeter-scale area or from a time-integrated response of a few molecules adsorbed on nanoparticle aggregates. 6 The tabulated SERS responses are thus an average of a collection of possible adsorption conformations of the amino acids onto the surface weighted by the stochastic distribution of enhancement sites. Furthermore, vibration fingerprints strongly depend on the molecular environment such as the nominal pH value of the buffer solution, 7-9 the chemical properties of the surface, 10 or the presence of gauche and trans rotamers. 11 Orientation of amino acids as well as specific competitive interactions of their functional groups with the gold surface also provide specific SERS spectra. 12 A clear assignment of SERS lines is effectively mitigated when spectra result from the averaged contributions of various molecular conformations. Distinguishing known spectra in a pool of signatures is usually made possible by statistical approaches. For instance, the bianalyte SERS method 13 uses a mixture of two resonant identifiable molecules to clearly and unambiguously demon- strate counting of molecules at high concentration. Another powerful statistical method relies on principal component analysis (PCA) to extract all redundant information using a simple graphical representation. 14,15 PCA requires also distinguishable resonant molecules with one distinct Raman fingerprint per species. In low concentration assay, the detection of a few nonresonant proteins highlighted the high temporal variability of the SERS spectra. 16 Thus, single proteins cannot be identified by using a unique spectrum but rather by several vibrations of molecular groups near multiple possible anchoring points of the protein. Our work provides a comprehensive SERS methodology developed both to detect unequivocally single amino acid molecules and to identify its principal spectra. Our approach based on a unique experimental strategy and statistical analysis enables us to interpret the temporal fluctuations of spectra in terms of molecular identification. To illustrate the methodology described in this contribution, we detect the SERS response of a cystein molecule which is a sulfur-containing amino acid known for interacting with metallic surface. 17,18 Many types of adsorption configurations of L-cysteine were found on the surface of silver 19,20 and gold: 11 through the sulfur atom; in zwitterionic form; with the COO - groups; 20 and with electrostatic interaction between the positive charge of the NH 3 + group. Because of the adsorption specificities and the surface reactivity (oxidation), several SERS spectra are thus expected. For detecting SERS spectra of a single amino acid, we improved existing hybrid fabrication methods involving self-assembly of colloidal nanoparticles into prepatterned templates 21-23 by integrating complex surface shape Chebyshev gold nano- Received: May 5, 2014 Revised: July 9, 2014 Published: July 9, 2014 Article pubs.acs.org/JPCC © 2014 American Chemical Society 17975 dx.doi.org/10.1021/jp504395c | J. Phys. Chem. C 2014, 118, 17975-17982