SAM Interfacial Structure
 |
| Figure 1 |
Self-assembled monolayers (SAMs) based on a sulfur head group have been widely studied using numerous surface analysis techniques, inspired by both fundamental studies of self-organization and potential technological applications. Despite this extensive scrutiny, the structure of the adsorbed species and the details of the interfacial region are still a topic of controversy. Many groups have investigated the nature of the interface between the standing c(4×2) phase of 1-decanethiol (C10) and the Au(111) surface. Conclusions based on these studies have often been extrapolated to describe the analogous interface for the lower density pinstripe phase. It is our contention that this extrapolation is inappropriate, leading in some instances to erroneous structural conclusions.
A characteristic diffraction scan of the chemisorbed, low-density striped phase of C10/Au(111) obtained via a ~20 L dose at the target is shown in Figure 1a. An average domain size of at least 400 Å is inferred from the width of the diffraction peaks because there is no broadening added to the substrate specular peak width. The peak-to-peak spacing in this spectrum is 0.198 Å-1, which agrees with previously published helium diffraction from this system. This spacing alone is not sufficient to distinguish between an (11´Ö3) and an (11.5´Ö3) unit cell because the atomic spacing in the topmost layer of the substrate, which serves as the reference for the overlayer nomenclature, is not probed. However, LEED and STM studies have clearly demonstrated that the C10 striped phase forms a p(11.5´Ö3) structure. Assuming this conclusion to be true, therefore, the 0.198 Å-1 peak spacing in the helium diffraction proves that the substrate nearest-neighbor spacing must be 2.76 Å. The nearest-neighbor spacing for Au(111) is 2.88 Å while the Au(23´Ö3) spacing is 2.76 Å. Thus, our helium diffraction data considered in the context of the available LEED and STM data prove that the gold is reconstructed underneath the striped phase SAM.
| Figure 2 |
 |
Looking closer at these data, moreover, reveals additional, smaller peaks centered between the principals (Figure 1b). These ½-order peaks can only arise from periodicity on a scale exactly twice that of the C10 dimer-namely, 63.4 Å. Interestingly, and we will argue not coincidentally, this distance is identical to the long dimension of the Au(23´Ö3) reconstruction unit cell. Figure 3 depicts a proposed schematic structure of the C10 pinstripe phase adsorbed on a still-reconstructed Au(111) surface. Displacement along the [112¯] direction has been exaggerated to clarify the non-linear structure composed of four decanethiol molecules. If it were not for the fortuitous match between the length of the (23´Ö3) reconstruction unit cell and the length of two (11.5´Ö3) C10 repeat units, the ½-order peaks would be imperceptible. A second possible explanation for the additional peaks is an as yet unobserved, long-range periodic structural differentiation between alternating thiol dimers adsorbed on deconstructed gold.
| Figure 3 |
 |
To test these competing theories we studied a chemically analogous striped phase with a different molecular length, effectively removing the unit cell size correspondence present between C10 and Au(23´Ö3). 1-octanethiol (C8), known to form a (10.4×´Ö3) unit cell, is chemically very similar to C10 and has a vapor pressure well suited to gas phase dosing, and yet the Au(111) reconstruction cannot accommodate an integer number of C8 dimers without an awkward gap between them. A diffraction scan of the chemisorbed striped phase of C8/Au(111) showing four orders of diffraction is shown in Figure 2. As with C10, an average domain size of at least 400 Å is inferred from the width of the diffraction peaks. The peak-to-peak spacing in this spectrum is 0.218 Å-1, implying either a (10´Ö3) or a (10.4´Ö3) unit cell depending on the reconstruction status of the substrate. This agrees with published STM data. In contrast to the C10 spectrum, there is no evidence for ½-order peaks between the principal diffraction peaks. Note, however, that this result does not imply that the surface is deconstructed. Even with a reconstructed substrate, the periodicity of this SAM does not coincide with the reconstruction and will therefore adopt a random relationship with the stacking fault transitions. This negative result does, nonetheless, illustrate that the ½-order peaks are not a simple byproduct of alkanethiol striped phase structure. C10, due to its convenient dimension, serves as a window to the (usually hidden) substrate below. The presence of ½-order peaks in the C10 diffraction spectrum is strongly redolent of substrate reconstruction.
| Figure 4 |
 |
To complement the reciprocal-space studies, STM experiments were also performed on the C10/Au(111)
pinstripe system. Particular attention was paid to the initial stages of monolayer growth to allow for
simultaneous imaging of the SAM and the substrate. Before exposure to C10, the gold crystal displayed expansive
domains exhibiting both the (23´Ö3) and herringbone reconstructions (Figure 4a).
Figure 4 depicts a series of images obtained following progressively larger exposures of C10, beginning
with a clean substrate. As observed by Poirier, the bright parallel lines indicative of the
(23´Ö3) reconstruction stacking fault dislocations are clearly visible on those areas
of the surface not directly covered with striped phase decanethiol. The herringbone periodicity in which the
parallel lines zigzag on a scale of 250 Å is affected by alkanethiol adsorption (and, interestingly, vice-versa),
but the (23´Ö3) structure endures throughout. Indeed, (23´Ö3)
bright parallel lines are discernible immediately adjacent to the growing C10 islands. Furthermore, no vacancy
islands are observed on the gold surface. Vacancy islands have been shown to accompany deconstruction of the
(23´Ö3) superlattice, so their absence rules out destruction of this reconstruction.
We note that pits have been observed to exist simultaneously with the pinstripe phase of mercaptohexanol adsorbed
on gold, but none were observed in the course of the experiments described here. Indeed vacancy islands are only
observed in association with the denser standing phase. These results agree well with the helium diffraction
data presented above.
95. "Coexistence of the (23´Ö3) Au(111) reconstruction and a striped phase self-assembled monolayer"
