ACS PRF | ACS
All e-Annual Reports
44259-AC5
Charge Transfer between Semiconducting Substrates and Organic Molecules
Molecular electronics is leading to the development of novel electronic and photonic devices. These devices will require them to be attached to solid substrates. This work was undertaken to understand the chemical interface between the molecule and silicon substrates and its effects on charge transfer between them. Specifically, how chemical attachment of the conjugated system affects band alignment and hence the formation of charge injection barriers, is studied. Silicon substrates offer the ability to tailor the substrate's valence and conduction levels with respect to the Fermi level by adjusting the type and amount of dopant.
Progress:
Phenylacetylene was chosen as the first molecule. Phenylacetylene is conjugated over the entire molecule. The well-characterized silicon (111) surface was used as the substrate.
Chemical attachment of phenylacetylene on hydrogen terminated Si (111) (H:Si) was probed using X-Ray photoelectron spectroscopy (XPS) and Fourier Transform Infrared Spectroscopy (FTIR). XPS performed on silicon substrates show a large increase in the C 1s intensity after exposure. Also, a &pi -&pi * shakeup, evidence of a conjugated molecule, is evident after exposure. These data clearly demonstrate the attachment of phenylacetylene to the silicon surface. H:Si exposed to phenylacetylene has an FTIR spectrum very similar to the gas-phase spectrum, including an obvious acetylene stretching mode and the phenyl breathing modes; suggesting attachment of the phenylacetylene at the acetylene termination.
Attachment of phenylacetylene on Si (111) 7x7 reconstructed surfaces was monitored using scanning tunneling microscopy (STM). Si (111) 7x7 surfaces are prepared by several flash anneals to above 1450 K. Bare surfaces show the expected surface reconstruction, as well as a few features, which are undoubtedly defects. After 0.05 L dosing (1 L = 10E-6 Torr-s), an STM scan of the same region shows new features. Another 0.5 L increases the coverage to 0.014 monolayer. At this exposure the 7x7 reconstruction is still evident. After an additional 30 L, the 7x7 reconstruction is broken, making molecular and substrate features indistinguishable.
Measurements of the valence band density of states were made using ultraviolet photoelectron spectroscopy (UPS). Bare Si (111) 7x7 surfaces were prepared as above, and exposed 1, 10, and 100 L phenylacetylene. The UPS spectrum of unexposed silicon shows the surface states expected for the 7x7 reconstruction. With increasing dosing, changes in the valence band spectra are evident. After 10 L exposure, the silicon surface states are no longer present, consistent with the STM data. Features on the 1 L exposed surface show agreement with the theoretical density of states calculated by DFT. Comparison of the experimental and calculated spectra show that the highest occupied state with contributions from the phenylacetylene molecule lie approximately 1 eV below the silicon Fermi level; which aligns within 0.1 eV of the bulk silicon valence band edge.
As unoccupied states are inaccessible to photoelectron spectroscopy, other methods must be used to measure the conduction band/LUMO alignment. An initial attempt has been made to make such measurements using bias dependent STM imaging. The apparent height of a feature in an STM image is dependent on many parameters including, the density of states in the sample below the applied bias. Thus, as the applied bias is raised, more states are available to tunneling, resulting in an increase in the measured height. By tracking the apparent height of molecular features with changing applied bias, the molecular energy levels can be found.
STM images were taken of the 0.1 L exposed surface at biases between 0.7 and 2.4 V for unoccupied states. Scans of the same region at 0.7 and 0.9 V show markedly different features; at 0.7 V, only silicon features are evident, but at 0.9 V, molecular features are obvious. This suggests that the molecular LUMO lies between these two voltages. Between 0.9 and 1.2 V, the apparent height of the phenylacetylene features increases as expected, but after 1.2 V, the apparent height unexpectedly drops until after 1.8 V, where the silicon features dominate.
Considering the UPS and STM results, the band alignment for Si (111)/phenylacetylene systems can be drawn. The molecular HOMO aligns fairly well with the bulk silicon valence band edge, with a charge injection barrier below 0.1 eV. The barrier between LUMO and conduction band edge, though, with the LUMO lying 0.7 eV above the band edge.
Future Work:
We are extending this work to other molecules with different chemical structures, to probe how changes in chemical structure effect the band alignment. Styrene will elucidate the precise role of the binding mechanism on the charge injection barrier. The structure of styrene is essentially that of phenylacetylene, with the acetylene replaced by vinyl. This small change, though, has been shown to result in very different binding on silicon surfaces. We have also recently constructed an inverse photoelectron spectrometer, which will give us direct access to the density of states of the unoccupied electronic levels.
