SUPPLEMENTARY INFORMATION: Harnessing Chemical Raman Enhancement for. Understanding Organic Adsorbate Binding on Metal. Surfaces

Transcription

1 SUPPLEMENTARY INFORMATION: Harnessing Chemical Raman Enhancement for Understanding Organic Adsorbate Binding on Metal Surfaces Alexey T. Zayak, Hyuck Choo, Ying S. Hu, Daniel J. Gargas, Stefano Cabrini, Jeffrey Bokor, P. James Schuck,, and Jeffrey B. Neaton, Molecular Foundry, Lawrence Berkeley National Laboratory, Berkeley CA 97, USA, The Moore Laboratory, Electrical Engineering, California Institute of Technology, Pasadena, CA 95, USA, Waitt Advanced Biophotonics Center, Salk Institute, La Jolla, CA 97, USA, and Department of Electrical Engineering and Computer Sciences, UC Berkeley, Berkeley CA 97-77, USA List of Figures S Calculated Raman active vibrational modes of gas phase BPE S Two-state model analysis of DFT results S Frequencies shifts To whom correspondence should be addressed Molecular Foundry, LBNL California Institute of Technology Waitt Advanced Biophotonics Center University of California at Berkeley

2 S Adsorbtion mode splitting S5 SERS measurements: different sets of data S Gold island binding energy S7 Gas phase orientation effect Here, we present geometries for several most Raman active modes of BPE. Vibrational mode assignments can be found elsewhere in the literature. 7 (555 cm-) ( cm-) Photon counts (experiment), Arb. units (theory) (97 cm-) (5 cm-) (9 cm-) VASP (PBE) 5 (59 cm-) 7 5 Experiment Raman shift (cm ) Figure S: Calculated Raman active vibrational modes of BPE. ( cm-) ( cm-) 9 (5 cm-)

3 References () Kim, A.; Ou, F. S.; Ohlberg, A. A.; Hu, M.; Williams, R. S.; Li, Z. Study of Molecular Trapping Inside Gold Nanofinger Arrays on Surface-Enhanced Raman Substrates. JACS,,. () Zayak, A. T.; Hu, Y. S.; Choo, H.; Bokor, J.; Cabrini, S.; Schuck, P. J.; Neaton, J. B. Chemical Raman Enhancement of Organic Adsorbates on Metal Surfaces. Phys. Rev. Lett.,,.

4 (a) E Vacuum (b) PDOS (States/eV) Adatom EFermi BPE LUMO PDOS (States/eV) Trimer (c) (Å /amu) R ZZ Bound - RZZ Isolated Adatom 5 Metal Deformation potential (ev/å) Trimer BPE HOMO Deformation potential (ev/å) 5 Pentamer 9 PDOS (States/eV) PDOS (States/eV) Deformation potential (ev/å) E-E Fermi (ev).5.5 Flat surface Pentamer Flat surface Deformation potential (ev/å) Figure S: (a) Schematic electronic structure of the metal-adsorbate interface. The position of molecule LUMO level relative to the metal Fermi level is the key parameter for the relative value (from one binding site to another) of the chemical Raman enhancement. (b) Actual electronic levels alignments calculated within DFT for four different binding sites. Electronic states projected on the molecule only are plotted with respect to the metal Fermi level, shown as the vertical dashed line. (c) Calculated enhancements of Raman tensor component, R Bound zz R Isolated zz, plotted against the deformation potentials, (E F E LUMO )/ Q n, for vibrational eigenvectors Q n. Overall, calculated Raman spectra for BPE on Au surface fit remarkably well to the two-state model, R Bound zz R Isolated (E zz F E LUMO ) (E F E LUMO ) Q n, similar to our prior work.

5 ω gas-phase - ω adsorbed cm - 97 cm - cm - 9 cm - 59 cm - 5 cm - Vibrational modes Experiment Adatom Trimer Pentamer Flat surface Figure S: Calculated and measured differences in vibrational frequencies between unbound and adsorbed molecules. Large frequency shifts are observed for the 97 cm modes. For the rest of modes, comparison between the theoretical and experimental shifts is not good. Unlike this case, in our previous study, DFT reproduced frequency shifts for benzene thiol molecule very well. However, we note that BPE exhibits relatively small shifts. Our calculations might not be sufficiently accurate to reproduce such small effects. Two degenerate modes at 97 cm - Two modes change and split symmetric anti-symmetric 97 cm- cm- Figure S: Here, we analyse the large shifts observed for the 97 cm mode. (Left) For the isolated molecule there are two degenerate modes: symmetric and anti-symmetric ring breathing. 5 (Right) Binding to gold breaks the degeneracy.

6 Takes: - h h h-h Takes: - h h h - h Photon counts Photon counts Takes: - Raman shift (cm - ) h h h-h Takes : - Raman shift (cm - ) h h h-h Photon counts 5 Photon counts Raman shift (cm - ) Raman shift (cm - ) Figure S5: Comparing of the extracted chemical contribution to SERS measurements from different exposures. Two samples, incubated for about min. and hours, respectively were used. Exposures were taken at different locations on the samples.

7 Figure S: Adatoms are frequently considered in the literature as likely sites for molecules to bind. The reason for this is that molecules bind to adatoms stronger than to other geometries. Thus, if we assume that all kinds of binding sites are available for the molecule on the surface, the molecule might prefer to bind to an adatom. Here, we look at this problem from a different perspective. Even though molecules would prefer to bind to adatoms, the latter might not be available on the surface. Our calculations show that adatoms are metastable on the surface and would always prefer to meet other atoms to form islands or other group arrangements with higher nearest neighbor coordination. 7

8 (a) (b) R(zz) cross section ( - cm /sr) Vertical BPE R(zz) cross section ( - cm /sr) Tilted BPE (c) R(zz) cross section ( - cm / sr) Vertical Tilted Raman shift (cm - ) Vibrational mode (d) Experimental wce Raman shift (cm - ) Vertical - Tilted relative enhancement Figure S7: (a) Raman cross section of gas phase PBE computed from only one component of the Raman tensor, R zz ; with BPE in this case vertically oriented. (b) Similar R zz spectrum, but this time with BPE tilted with respect to the Z axis. (c) Selected R zz intensities for the vertical and tilted orientation of BPE. One can observe that orientation alters relatve intensities of the peaks. (d) If we assume that experimental variations that we observed as a function of incubation time originate from changes of preferred orientation, then the experimental data could be compared to the differences between the vertical and tilted orientations. We find that orientation variations do not yield the main feature of our measurements: reversal of 59 and 5 cm peaks. None of the gas phase orientations could reproduce the peak reversal. Thus, we may conclude that from considered in this work effects only binding explains the experimental variations.