Modulation of Morphology and Electronic Structure ... - ACS Publications

Sep 25, 2014 - The signature of intercalation is cessation of molecule rotation, which is recognized by imaging of molecular orbitals. The electronic ...
0 downloads 0 Views 8MB Size
Article pubs.acs.org/JPCC

Interaction of C60 with Tungsten: Modulation of Morphology and Electronic Structure on the Molecular Length Scale J. Brandon McClimon,† Ehsan Monazami, and Petra Reinke* University of Virginia, 395 McCormick Road, P.O. Box 400745, Charlottesville, Virginia 22904-4745, United States S Supporting Information *

ABSTRACT: The evolution of morphology and electronic structure in sequential depositions of W and C60 on graphite has been studied by scanning tunneling microscopy/spectroscopy. The deposition sequence decisively controls morphology expression. W deposited on a graphite surface forms small clusters whose morphology is consistent with the predictions of a liquid droplet model in the size regime below 5 nm in diameter; these small clusters then agglomerate without sintering. These agglomerates are immobilized by the subsequent C60 deposition. C60 shows very little interaction with the W-cluster agglomerates, and the formation of typical close packed fullerene islands is observed. The inverse deposition sequence, W deposition on the surface of C60 multilayer islands, leads simultaneously to the formation of ultrasmall W clusters (d < 2 nm) due to limited mobility on the highly corrugated surface, and the intercalation of W in the C60 matrix. The signature of intercalation is cessation of molecule rotation, which is recognized by imaging of molecular orbitals. The electronic structure of C60 is not significantly modified by the presence of W agglomerates, clusters, and intercalation of W. However, if W is deposited on a single layer of C60 its impact on the electronic structure is considerable and expressed in a compression of the band gap, which might be attributable to charge screening due to image charges, or the onset of molecule breakdown. The morphology as well as the electronic structure of this layer is highly inhomogeneous and can be described as a composite of W and C60 due to accumulation of W at the graphite substrate−C60 interface.



INTRODUCTION The study of interfaces between metals and organic molecules has been a frequent topic of study over the last several decades due to the attendant technologies,1−3 such as organic light emitting diodes and organic solar cells where these interfaces are critical to functionality.4 In particular, C60 and its derivatives are of interest since C60 has a direct band gap and electron acceptor properties that make it and its derivatives particularly suitable for photovoltaic applications.5−7 The interaction of C60 with metal and semiconductor surfaces has been studied extensively, and in contrast, the deposition of metals on the C60 layers, where the metal is the highly mobile reactant, has only been observed in a few systems despite the frequent use of this process in the assembly of device structures.8−10 The metal−fullerene interaction can have a strong impact on the electronic structure of the molecular layer, and as such is critical to device design. Several studies of submonolayer (ML) films of C60 on single crystal metal surfaces have found strongly reduced band gaps compared to those present in the bulk phase.11−15 In all cases, the effect has been attributed to image charge effects arising due to electron donation from the metal to the lowest unoccupied molecular orbital (LUMO) of the C60 molecule. This explanation has been corroborated by ab initio calculations.13,14 If the metal is deposited on a C60 surface, a much larger variety of behavior is observed. For example, Au atoms have © 2014 American Chemical Society

very high mobility on top of the C60 layer and will travel to the edge of the layer before nucleating clusters.8 Cr nucleates into clusters on the C60 surface, while Ti and La form a conformal layer due to a chemisorption interaction.16 Numerous attempts have been undertaken to form intercalation compounds where metal atoms diffuse into a C60 matrix thereby forming exohedrally doped fullerene complexes. Such studies have been generally successful for alkali17 and alkaline earth metals,18 but not for transition metals.16 This difference has been attributed to the higher cohesive energies of the transition metals which make metal cluster formation thermodynamically preferred over intercalation.11 Nevertheless, this problem can be sidestepped by the use of nonequilibrium conditions which discourage metal clustering, as has been demonstrated for both Ti19 and Ag20 with C60. Two different reaction sequences are distinguished in the present study of C60-W composite material: the first type of experiment began by depositing the W on HOPG followed by the deposition of C60. In the second type of experiment the sequence was inverted and C60 was deposited first followed by W deposition. The inversion of the deposition sequence leads to significant differences in the final morphology of the thin Received: July 3, 2014 Revised: September 18, 2014 Published: September 25, 2014 24479

dx.doi.org/10.1021/jp506618b | J. Phys. Chem. C 2014, 118, 24479−24489

The Journal of Physical Chemistry C

Article

films, which is driven by modulation of the surface environment the reactants encounter; for example, W adatoms experience a flat graphite surface, where diffusion is nearly unhindered, in one scenario, and a highly corrugated C60 surface in the inverted scenario. The study of the impact of reaction sequence on morphology allows us to draw conclusions on metal− graphite−fullerene interactions. The surface morphology and electronic structure was studied using scanning tunneling microscopy (STM) and scanning tunneling spectroscopy (STS).

the absence of significant tip changes in the course of an experiment. To confirm quality and reproducibility of the STS maps, we compared single I/V curves in adjacent pixels, and similar topographic areas. Averaging over several adjacent pixels was sometimes used to improve the signal-to-noise ratio, but only after confirming that the variation of relevant spectral features was minimal. Topographic images were analyzed with Gwyddion 2.2923 while STS data and maps were analyzed using Igor Pro 6.22 (Wavemetrics). Cluster size distributions were measured with STM using an image mask to measure the projected volumes above the adjacent surface height (z-value), and those results were checked manually for select clusters using linescans. These measurements always carry an error as the tunneling current from the side of the tip to a nearby cluster invariably exaggerates the diameter of a cluster by several angstroms depending on tip and cluster shapes and the size of the cluster. Corrections can be made to such measurements if the shapes of the tip and cluster are well understood and can be modeled for spherical clusters or other simple geometric shapes.8,24 Such corrections were not undertaken here as many of the clusters are embedded in cluster agglomerates (see Figure 1) of slightly



EXPERIMENTAL SECTION All experiments were conducted in an Omicron Variable Temperature STM with a background pressure of