Impact of Fluorination on Initial Growth and Stability of Pentacene on

Mar 8, 2012 - In-Depth Atomic Structure of the Pentacene/Cu(110) Interface in the Monolayer Coverage Regime: Theory and X-ray Diffraction Results...
2 downloads 0 Views 4MB Size
Article pubs.acs.org/JPCC

Impact of Fluorination on Initial Growth and Stability of Pentacene on Cu(111) H. Glowatzki,*,† G. Heimel,‡ A. Vollmer,§ S. L. Wong,† H. Huang,† W. Chen,† A. T. S. Wee,† J. P. Rabe,‡ and N. Koch‡ †

Department of Physics, National University of Singapore, 2 Science Drive 3, 117542 Singapore Institut für Physik, Humboldt−Universität zu Berlin, Newtonstr. 15, D-12489 Berlin, Germany § Helmholtz-Zentrum Berlin für Materialien und Energie GmbH Bessy II, Albert-Einstein-Str. 15, 12489 Berlin, Germany ‡

ABSTRACT: Structure, growth, thermal stability, and electronic properties of thin films of the fully fluorinated analogue to the archetypical organic semiconductor pentacene (PEN), perfluoropentacene (PFP), were investigated on Cu(111) at room temperature by scanning tunneling microscopy (STM), low energy electron diffraction (LEED), ultraviolet photoelectron spectroscopy (UPS), and X-ray photoelectron spectroscopy (XPS). In contrast to PEN, where molecules could only be imaged by STM at full monolayer coverage, PFP was seen to stabilize in disordered clusters already in the submonolayer regime. Furthermore, while long-range order was observed for closed PEN molecular monolayers, PFP only formed a disordered first wetting layer. Highly ordered domains were not observed until the formation of the second layer of PFP. In this layer, the molecular planes are inclined to the surface, as supported by additional STM measurements on graphite and theoretical modeling. Careful consideration of the structural details in the transitional growth regime from molecular mono- to multilayers thus emerges as the key ingredient to achieving a deeper understanding of metal/organic interfaces relevant for organic electronic devices.



electron mobility to more than 0.2 cm2 V−1 s−1.3 However, the exact height of charge injection barriers into organic semiconductors is governed not only by the molecular properties but also, and maybe predominantly, by complex interfacial mechanisms.4,7−10 Since electrical contacts in OFETs are mainly made of coinage metals, detailed studies of their interfaces with organic semiconductors are inevitable to reach a fundamental understanding and, eventually, full control. Pursuing this goal, the electronic and structural characteristics of PEN have been extensively investigated on various model surfaces including Au(111),11−17 Au(100),18 Au(110),19−21 Cu(119),21,22 Cu(110),23−25 Cu(111),26 Ag(110),27 and Ag(111).28−32 For its fluorinated analogue, PFP, the number of similar reports is comparatively limited.32−35 In a previous study, the vertical interaction strengths of PEN and PFP with Cu(111) were investigated focusing on the conformation of PFP, which becomes nonplanar in the monolayer.35 Here, the focus will be on the lateral structure of PFP on Cu(111), its electronic properties, its relation to the thermal stability of PFP, and notably, on growth beyond the first monolayer. In this regard, the present work relates to a series of publications, in which the mono- to multilayer transition in the growth of PEN on Ag(111) was investigated and controversially discussed: Eremtchenko et al. proposed that PEN forms a weakly bound,

INTRODUCTION In the field of organic electronics, the development and optimization of thin film field-effect transistors (OFET) has received considerable attention due to their potential use in low cost applications and flexible electronics.1 Because of its high hole mobility of up to 5.5 cm2 V−1 s−1 in OFETs,2 pentacene (PEN, C22H14, Figure 1a) is widely studied as an active layer

Figure 1. Molecular structures of (a) pentacene (PEN) and (b) perfluoropentacene (PFP).

material in this context. At the same time, PEN exhibits poor characteristics as an electron transporting material,3 mostly because its low electron affinity (ca. 2.8 eV4) introduces large barriers for electron injection from common OFET electrode materials. Full fluorination in the derivative perfluoropentacene (PFP, C22F14, Figure 1b) significantly lowers the energies of the frontier molecular states5,6 and was found to enhance the © 2012 American Chemical Society

Received: September 6, 2011 Revised: January 18, 2012 Published: March 8, 2012 7726

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

Figure 2. STM height images of pentacene on Cu(111) at room temperature. (a) Monolayer pentacene forming a highly ordered structure. Parameters of the unit cell marked by the blue box are (1.6 ± 0.1) nm × (0.9 ± 0.1) nm (89 ± 1)° (US = −1.0 V, IS = 0.6 nA). The inset shows the corresponding LEED image. (b) Zigzag structure of the pentacene layer. The progression of molecules within different domains is highlighted by the white lines. The inset shows a magnification in which also single molecules can be identified (US = −1.2 V, IS = 0.3 nA).

disordered first layer of flat lying molecules on Ag(111).30 Remarkably, they present evidence for the subsequent growth of ordered molecular layers on top of this first layer with a possible tilt of the molecules out of the surface plane. These findings were contested by Käfer et al., who claimed that, in contrast, PEN forms a single wetting layer with an apparent 13° tilt angle of the molecular planes with respect to the Ag(111) surface and adopts a bulk-like structure with an average tilt angle of 34.5° immediately beyond this first monolayer.31 The authors effectively suggest that the ordered second layer proposed in ref 30 actually represents the monolayer rather than a bilayer. Eventually, Dougherty et al. presented scanning tunneling microscopy (STM) data recorded at 50 and 300 K, which strongly suggest that (i), indeed, a disordered monolayer of PEN exists on Ag(111) at room temperature that turns into an ordered monolayer when cooled down to 50 K and that (ii) the ordered PEN layer observed in ref 30 at room temperature indeed represents the second molecular layer that grows on a first, disordered wetting layer.36 In the present work, a similar behavior is observed for the growth of PFP on Cu(111). To arrive at these conclusions, structural analysis of thin PFP films was performed by STM and low energy electron diffraction (LEED), while the thermal stability and electronic properties were explored by X-ray (XPS) and ultraviolet (UPS) photoelectron spectroscopy. Additional STM measurements were performed on highly ordered pyrolitic graphite (HOPG), in order to compare the structural behavior of PFP on Cu(111) to that on a nonreactive substrate, and for PEN on Cu(111), in order to explore the impact of fluorination on molecular ordering and packing.

system at the National University of Singapore housing an Omicron low temperature STM. Ex situ cut PtIr-tips and etched W tips were used for STM imaging. Voltages for STM images presented in this work refer to sample bias. UPS and XPS measurements were done at the synchrotron endstation SurICat, beamline PM4 at BESSY II, Berlin. For XPS and UPS experiments, excitation energies of 620 and 35 eV were used accordingly (F1s core levels were measured in second order). In order to clear the analyzer work function, the secondary electron cutoff (SECO) was measured at −10 V sample bias. In all cases, sample preparation proceeded in preparation chambers (base pressure < 2 × 10−9 mbar), which were interconnected to the respective analysis chambers (base pressure < 5 × 10−10 mbar) allowing sample transfer without breaking UHV conditions. Cleaning of Cu(111) was accomplished by repeated cycles of heating and Ar+-ion sputtering, and clean surfaces were verified by photoelectron spectroscopy. PEN (Sigma-Aldrich) and PFP37 were evaporated from resistively heated crucibles (substrates kept at room temperature during deposition) and the deposited mass thickness was monitored with quartz crystal microbalances. Molecular orbitals were calculated within the framework of density functional theory (DFT) for the relaxed geometry. The B3LYP functional38 was used in conjunction with the 6311G** basis set. Calculations were performed using Gaussian03.39 To simulate STM images, we then performed periodic DFT calculations (PW91 functional40) on freestanding monolayers of PFP (i.e., no metal substrate) using the Vienna ab initio simulation package (VASP).41,42 There, valence−core interactions were treated within the projector augmented-wave approach,43,44 allowing for the low kineticenergy cutoff of 20 Ryd for the plane-wave expansion of the valence pseudo Kohn−Sham wave functions. The lateral unit cell was taken from experiment, and the layer geometry was optimized under the constraint of planarity until residual forces were below 0.1 eV/Å. The 2D Brillouin zone was sampled with a 4 × 7 Monkhorst-Pack k-point grid.45 Subsequently, the charge density associated with the molecular HOMO-derived states was extracted from the calculations and STM images were computed in the Tersoff−Hamann approximation,46 using



EXPERIMENTAL DETAILS AND THEORETICAL MODELING STM experiments at room temperature were performed at the Humboldt−Universität zu Berlin with an Omicron variable temperature STM in a customized ultrahigh vacuum (UHV) system also equipped with a channelplate low energy electron diffraction (LEED) system. LEED patterns served to support structural analysis done by STM. Additional STM measurements were conducted at 77 K using a custom-built UHV 7727

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

Figure 3. STM height images of perfluoropentacene on Cu(111) at room temperature. (a) Submonolayer after deposition of PFP. Some of the molecules sticking to the step edges and point defects at terraces are marked by the arrows. In the inset, individual molecules can be identified (US = −1.2 V, IS = 0.5 nA). (b) Submonolayer PFP after annealing at 150 °C for 45 min (US = −1.4 V, IS = 0.3 nA). (c) Submonolayer PFP after annealing at 150 °C for 135 min in total (US = −1.2 V, IS = 0.3 nA). Some second layer structures are marked by arrows. The bottom inset shows a magnification of ordered molecules in the second layer. (d) Closed layer of PFP forming highly ordered domains. Domain boundaries are defined by Cu step edges (US = −0.8 V, IS = 1.0 nA). In the left inset, single molecules are clearly visible exhibiting a unit cell of (1.75 ± 0.05) nm × (1.0 ± 0.1) nm (92 ± 1)° marked by the white box (US = −0.8 V, IS = 0.5 nA). The right inset shows the according LEED image.

interaction may indeed be strong, but the lateral (intermolecular) interaction may be weak enough to allow for highly mobile molecules on the surface at room temperature. Indeed, high lateral mobility of monolayer PEN molecules on Cu(111) was reported in a dynamic force microscopy study by Kawai et al.50 In contrast, individual PEN molecules on Cu(111) at submonolayer coverage can be imaged at low temperatures.26 At full monolayer coverage (nominally about 4 Å), we observed a closed and well-ordered PEN layer by STM also at room temperature (Figure 2). Within this layer, two different packing motifs were found. The first is shown in Figure 2a. There, the molecules form a close packed structure in large-area regular arrays comprising a side by side arrangement of PEN with parallel long molecular axes. The unit cell (marked by a blue box in Figure 2a) was determined from STM and LEED analysis (inset of Figure 2a) to be a = (1.6 ± 0.1) nm, b = (0.9 ± 0.1) nm, and α = (89 ± 1)°. In the second structural motif, a similar molecular packing was found (Figure 2b). The

the averaging procedure outlined in ref 47 to account for finite tip size.



RESULTS AND DISCUSSION We start by discussing the experimental results of STM investigations. In order to explore the impact of fluorination, pentacene (PEN) on Cu(111) was first studied as a reference system for the subsequent experiments with perfluoropentacene (PFP) since no according STM data for this material combination was available. In a previous study, XPS results of monolayer PEN on Cu(111) have suggested that the organic/ metal interaction is rather strong, even chemisorptive.35,48 Nevertheless, after preparing a sample with submonolayer PEN coverage (i.e., a nominal thickness of 2 Å), we observed no signature of molecules by STM. This phenomenon may occur at room temperature and is commonly interpreted in terms of adsorbates forming a two-dimensional molecular gas.30,49 In such a situation, the vertical (molecule-metal binding) 7728

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

Figure 4. (a) Detailed structure of PFP derived by correlation averaging of the ordered structure shown in Figure 3d. (b) Line profiles according to the lines shown in panel a. Profile A is referring to the line along the long axes, whereas B follows the line perpendicular to it. (c) Calculated probability density of the PFP HOMO. (d) Simulated STM image for flat lying PFP molecules. (e) Proposed model for molecular short axes tilted by 15° with respect to the substrate plane. (f) Simulated STM image according to molecules arranged as proposed in panel e.

form small clusters. In the zoom shown in the inset of Figure 3a, individual molecules can be clearly identified as rod-like features with random orientation. Still, most of the terrace areas appear to be free of molecules, and we found a clear mismatch between the number of molecules imaged by STM and the number expected from the nominal coverage. As discussed for PEN above, the majority of PFP molecules might be adsorbed in a 2D gas-like phase (i.e., not bound to step edges or defects), thus diffusing across the surface during imaging.49 By that they contribute to the overall background of the tunneling current, but remain invisible in STM as defined entities due to their high lateral mobility. Upon annealing this sample at 150 °C for 45 min, more individual PFP molecules appear (Figure 3b). Apparently, the density of nucleation sites has increased. In fact, as will be substantiated and discussed below, PFP reacts with the Cu surface upon annealing. The magnification in the inset of Figure 3b indicates local ordering of the molecules, which sometimes exhibit a side by side arrangement. The same sample was then annealed for an additional 90 min at 150 °C, leading to a total annealing time of 135 min. A corresponding STM image is shown in Figure 3c. The majority of the surface appears now covered with PFP molecules and no individual, separated clusters can be identified as they appear all interconnected. However, no long-range intermolecular order

difference lies in a frequent change in the orientation of molecules, forming an overall zigzag structure with 120° kinks at typically every 2 to 5 nm. This is highlighted by the white lines in the STM image of Figure 2b. Individual molecules can be clearly identified in the magnification (inset of Figure 2b). To emphasize the molecular arrangement, the chemical structure of PEN is overlaid in the inset. While in Figure 2a, the resolution does not allow to clearly distinguish the boundaries of individual molecules along their long molecular axes, their defined shape in Figure 2b exhibits a rod like appearance with bright spots at their ends. The same appearance was described by Lagoute et al. for isolated PEN molecules on Cu(111) at low temperature and explained by an enhanced tunneling conductance.26 Next, PFP on Cu(111) was investigated following the same experimental protocols. In this case, individual molecules could be imaged at room temperature even at submonolayer coverage (nominally 3 Å). As can be seen in Figure 3a, single molecules and small clusters of molecules are visible at step edges and even within terraces (some marked by white arrows in Figure 3a). The step edges, which commonly are preferred adsorption sites,51,52 are almost completely covered by PFP. Adsorption sites within terraces are mainly defined by point defects.51 Once a molecule sticks to such a defect, others can attach to it and 7729

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

experiment, where the middle one appears higher than the neighboring two. This is in contrast to the two features expected from the calculated HOMO (Figure 4c). Since the HOMO of an isolated molecule does not necessarily reflect the appearance of the molecule in STM, we simulated STM images for a periodic monolayer of flat lying PFP (Figure 4d). Still, there are five features along the long molecular axis, but only two perpendicular to it are found, which is in contrast to experiment. However, because of inherent shortcomings of (semi)local DFT and limited computational resources to treat PFP on Cu(111) rigorously, the simulation was done for a freestanding monolayer of PFP, thus neglecting the (presumably weak) interaction with the underlying first layer of PFP and the Cu(111) substrate. To validate the quality of the simulation for the situation of PFP on a weakly interacting substrate, we conducted additional experiments for PFP adsorbed on HOPG. The corresponding STM image of a PFP monolayer is shown in Figure 5. While the packing of molecules is quite similar to

was observed, and the layer is not completely closed (inset of Figure 3c). Despite this, some molecules are found to already adsorb on top of the first layer, which can be seen in Figure 3c as white spots and the white line in the lower right part of the image (marked by arrows). In high magnification (bottom inset), these white lines appear to be composed of molecules in a side by side arrangement. Apparently, PFP in the first layer exhibits strong molecule−metal interaction after annealing, which prevents the formation of an ordered layer. In contrast, the interaction strength of the second molecular layer with the first (and the Cu substrate) can be expected to be smaller, thus possibly allowing for intralayer intermolecular interactions to trigger ordering. Motivated by these findings, a distinctly higher molecular dosage (nominally 7 Å) was deposited on a freshly prepared Cu(111) surface. Indeed, this led to the formation of a close packed and highly ordered molecular layer (Figure 3d). On the basis of the results discussed above, as well as similar observations for PEN on Ag(111),30,36 we presuppose that the ordered PFP layer in Figure 3d actually represents the second rather than the first molecular layer, an assumption that will be corroborated in the UPS section. Apparently, the PFP molecules lie in a side by side arrangement, this time extending over entire terraces with orientation changes only occurring between different terraces. Three domain orientations (two of them visible in Figure 3d) were observed, which differ by azimuthal 60° rotations of the long molecular axis. Notably, these three orientations are in accordance with the symmetry of the underlying substrate. Even though we expect weak interaction between Cu and the (supposedly) second molecular layer, the substrate might still trigger the preferential growth directions, e.g., step edges are still starting points for layer formation as discussed in refs 30 and 36. However, because of the mostly irregular shape of the terraces in these experiments, we cannot exclusively attribute the growth behavior of PFP on Cu(111) to this effect. Therefore, we have to take into account that molecule−metal interaction possibly extends into the second layer, either directly or indirectly. In particular, it could be speculated that the first molecular layer is locked into a similarly ordered structure by the presence of the second.30 In the magnification shown in the inset of Figure 3d, individual molecules can be clearly identified. Even an intramolecular structure is visible, whose origin will be discussed in detail in the next section. The unit cell (shown as the white box in Figure 3d) determined by STM and LEED analysis (inset) has the parameters a = (1.75 ± 0.05) nm, b = (1.0 ± 0.1) nm, and α = (92 ± 1)°, which is only slightly larger than that found for PEN. Domains with frequently alternating orientations, as found for PEN (Figure 2b), were not observed for PFP on Cu(111). The intramolecular contrast of PFP (inset of Figure 3d) needs particular attention. To obtain a statistical average of this contrast, STM images were processed by correlation averaging including many molecules over a large area, the result being shown in Figure 4a. To discuss the intramolecular features, line profiles of a molecule along the long molecular axis (labeled A) and perpendicular to it (labeled B) are shown in Figure 4b. For comparison, the calculated electron density distribution of the highest occupied molecular orbital (HOMO) of PFP is depicted in Figure 4c. The five distinct maxima along profile A clearly correspond to the five HOMO lobes along the long molecular axis. From the profile perpendicular to the long molecular axis (profile B) three features are identified in the

Figure 5. STM height image of the PFP monolayer on HOPG at 77 K. The inset shows a magnification in which the inner structure of PFP is clearly visible (US = −2.0 V, IS = 0.1 nA).

the (at this stage still supposedly) second layer PFP on Cu(111), the intramolecular contrast is markedly different. The molecules lie flat on the HOPG surface, and their appearance is in perfect agreement with that predicted by the simulated STM image in Figure 4d. On the basis of this finding, we rule out the possibility that the appearance of bilayer molecules on Cu(111) could be affected by weak interaction with a first molecular layer and/or indirect interaction with the copper substrate. Instead, we turn toward the possibility of noncoplanar orientation. Such an inclination of the second-layer molecular planes with respect to the surface is plausible because PFP exhibits a herringbone arrangement of molecules in the bulk crystal structure.37 Furthermore, we note that this scenario (as suggested for PEN on Ag(111)30) was recently invoked also for PFP on Ag(111).34 Encouraged by the agreement between theoretical and experimental results in the case of PFP on HOPG, we simulated STM images for different angles between the short molecular axes and the surface plane as shown in Figure 4e. Specifically, the tilt angle of the molecules was varied 7730

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

Figure 6. (a) C1s and (b) F1s core level spectra of monolayer PFP on Cu(111) before (top curve) and after annealing at 150 °C (bottom curve). Black points represent the measured data points, while solid lines are derived from fitting the data with appropriate pseudo-Voigt profiles.

changed (bottom curve in Figure 6a). Now, the peak at low BE has higher intensity than the others, and a clear shoulder at 1 eV lower BE becomes apparent, i.e., a fourth peak appears, which can be assigned to carbon after decomposition of PFP as discussed further below. This is fully consistent with the STM observations (Figure 3), which showed that the density of immobile molecules increases upon annealing. Together with the XPS results, we conclude that immobilization is brought about by a chemical reaction of PFP with Cu. A detailed analysis of the C1s core level spectra reveals that the two high BE peaks change upon annealing, i.e., BE shifts of ca. 0.1 eV and broadening (see Table 1). Most notably, the intensity ratio

in steps of 5° between 0° and 20°. Judging from visual inspection, the best agreement with experiment, in particular the emergence of three features perpendicular to the long molecular axes and their relative apparent height, is found for a tilt angle of 15°; the corresponding simulated STM image in Figure 4f is to be contrasted with the measured one in Figure 4a. On the basis of the step size of the manual variation, the error of the tilt angle can be estimated as ±3°. Note that a similar behavior was reported for PEN on Cu(111) by Kawai et al.50 In contrast to flat-lying molecules within the first layer, they report a tilting of the short molecular axis by 30° for molecules in the second layer, which differs from the bulk structure. Further support is given by a DFT study of PEN on Ag(111).53 From their simulations, the authors report an orientation of second-layer molecules (on top of a first wetting layer) that differs from the bulk structure as well. We now turn toward the chemical and electronic properties of PFP monolayer and multilayer films on Cu(111). First, the origin of annealing-induced strong PFP−Cu interaction will be revealed. Such a temperature induced reaction at the interface was suggested in ref 35 and will now be supported by additional photoemission data. XPS spectra of the C1s and F1s core levels taken at monolayer PFP coverage on Cu(111) are shown in Figure 6. The top curves are from the sample just after the (room temperature) deposition of PFP, and the bottom curves were recorded after annealing at 150 °C for 135 min. For the freshly prepared sample, three peaks in the C1s core level region can be identified (top curve in Figure 6a). The two high binding energy (BE) peaks (at 285.8 and 287.3 eV) are separated by 1.5 eV and can be attributed to chemically inequivalent carbon atoms within PFP, i.e., C atoms bound only to other C atoms and those that are bound to F.35 These peaks are virtually identical to the ones found for PFP on Au(111),35 where weak, physisorptive molecule−metal interaction prevails.33 A lower BE peak at ca. 284.5 eV binding energy is observed, however, with lower spectral weight than the other two. This low BE peak is assigned to residual carbon contamination occurring during sample handling, on the one hand, and also to PFP molecules that have reacted with the substrate surface, on the other hand (i.e., the molecules bound to step edges and point defects on terraces that could be imaged by STM in Figure 3a). Upon annealing the sample at 150 °C, the ratio of the peak spectral weights drastically

Table 1. C1s Peak Positions and Widths before and after Annealing a Monolayer of PFP on Cu(111) at 150 °C fresh sample

after annealing

peak

position (eV)

width (eV)

position (eV)

width (eV)

1 2 3 4

287.31 285.83 284.49

0.83 0.63 0.86

287.29 285.93 284.36 283.42

1.21 1.08 0.95 1.00

of the two high BE components is reversed. The finding that all components in the entire C1s region are altered supports a strong reaction of a large portion of PFP molecules with the Cu(111) surface. This, again, is in line with the STM data, where most molecules appear immobile after prolonged annealing (Figure 3c). Yet, further support for the proposed strong reaction is found in the F1s core level spectra (Figure 6b). For the fresh sample, a single peak at 687.5 eV BE is identified as belonging to F atoms covalently bound to C,54 indicating the presence of structurally intact PFP molecules. In contrast, annealing results in two F1s peaks, the first peak at 687.0 eV BE and an additional peak at 683.2 eV BE. The low BE component most likely originates from F atoms that have been abstracted from the molecule and are directly bound to the copper surface. This last proposition will be correlated with according changes of sample work function below. Here, we note as a last point that further heating of the sample up to 250 °C leads to pronounced desorption and the F1s core level signal to vanish, while some C1s intensity remains (not shown). 7731

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

Figure 7. UPS spectra of PFP on Cu(111). From bottom to top spectra of the clean Cu(111) substrate, the submonolayer PFP, monolayer, multilayer, and the sample after annealing at 150 °C are presented (according to nominal coverage of 0 Å, 1 Å, 3 Å, and 7 Å). (a) Valence band region. For better comparison, all spectra have been normalized to their intensity at the Fermi-edge. (b) Corresponding secondary electron cutoffs SECO following the same order.

the monolayer HOMO decreases, and a second peak at 0.35 eV higher binding energy is visible at a nominal coverage of 7 Å (the coverage where the STM images in Figures 3d and 4a were obtained). We identify this peak as the HOMO of the bilayer and assign its energetic shift to two effects: the dominant contribution arises from the fact that, due to the larger distance from the metal surface, photohole screening is weaker in the bilayer than in the monolayer, which commonly shifts the valence band features of organic semiconductors by 0.2−0.3 eV toward higher binding energy.33,55 The even larger shift in our experiments can be rationalized through molecular orientation.10,34,56 Intramolecular dipoles in PFP (predominantly the highly polar C−F bonds) and the π-electron distribution10,56 lead to orientation dependent ionization energies of ordered molecular assemblies.10,56−58 From the theoretical modelingbased interpretation of our STM data (Figure 4), we proposed a tilt angle of about 15° of the molecular planes with respect to the surface. This results in a nonvanishing projection of the C− F dipoles onto the surface normal and, consequently, an additional surface dipole layer that can account for the higher mono- to bilayer HOMO shift than expected from pure photohole screening. In the bulk herringbone arrangement that PFP adopts in thick films, the tilt angle of the molecular planes with respect to the surface is more than 40° and the orientation dependence leads to an increase of the PFP ionization energy in excess of 0.4 eV.34 Looking at the SECO taken for the coverages just discussed reveals that the Cu(111) work function is lowered by about 0.35 eV upon initial adsorption of a monolayer of PFP and remains constant throughout the multilayer regime. This shift is due to the push-back effect of metal surface electrons59 and, additionally, by the adsorptioninduced distortion of monolayer PFP molecules described in detail in ref 35. Overall, the UPS data strongly support our initial assumption that the ordered PFP layer imaged by STM in Figures 3d and 4a indeed represents a second molecular layer with slightly inclined molecular planes, which grows on top of a first, initially disordered layer.

In this last part, we discuss the UPS data for PFP on Cu(111). From bottom to top, the respective valence-band spectra and secondary electron cutoffs (SECO) are shown in Figure 7 for the pristine Cu(111) surface, submonolayer PFP coverages, bilayer, and the sample after annealing at 150 °C. Again, we start by discussing the thermal stability of PFP. While distinct features are visible in the (sub)mono- to multilayer UPS spectra (arrows in Figure 7a), the valence spectrum of the annealed PFP/Cu sample (top) appears to be featureless, similar to that of the pristine Cu(111) surface (bottom). The annealing-induced reaction of PFP molecules with the metal surface and the concomitant loss of π-conjugation thus likely results in a molecule−metal chemical species with a large energy gap, where emission from the highest occupied level is masked by the strong copper d-band emission at BE values higher than ca. 2 eV. Notably, the corresponding SECO shifts to higher kinetic energy upon annealing (Figure 7b), resulting in an even higher work function (by 0.2 eV) than that found for pristine Cu(111). As already suggested by the F1s core level spectra above (Figure 6b), some F atoms might have been abstracted from the molecules to directly react with Cu. In this case, the strongly electronegative fluorine would point outward from the surface, thus forming a surface dipole layer that could fully explain the observed increase in work function. In conjunction with the observations in STM and XPS, it can be concluded that annealing the sample leads to a strong reaction of PFP with the copper surface, which is accompanied by significant changes of the molecular chemical and electronic structure. Finally, we turn to the UPS spectra of (sub)mono- and bilayer. Starting at a coverage of nominally 1 Å, the emission from the HOMO of PFP appears with a low BE onset at ca. 1.40 eV and the maximum at ca. 1.5 eV. Its intensity increases with coverage up to nominally 3 Å, the film thickness where also the STM images in Figure 3a−c were taken. The line shape of the HOMO emission is strongly reminiscent of that obtained for PFP physisorbed on Au,33 thus corroborating a weak PFP− Cu interaction at room temperature and the 2D gas-like behavior with highly mobile molecules as seen in STM. Upon further deposition of PFP, the intensity of photoemission from 7732

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C



Article

(11) Schroeder, P. G.; France, C. B.; Parkinson, B. A.; Schlaf, R. J. Appl. Phys. 2002, 91, 9095−9107. (12) France, C. B.; Schroeder, P. G.; Forsythe, J. C.; Parkinson, B. A. Langmuir 2003, 19, 1274−1281. (13) France, C. B.; Schroeder, P. G.; Parkinson, B. A. Nano Lett. 2002, 2, 693−696. (14) Kang, J. H.; Zhu, X.-Y. Chem. Mater. 2006, 18, 1318−1323. (15) Beernink, G.; Strunskus, T.; Witte, G.; Wöll, C. Appl. Phys. Lett. 2004, 85, 398−400. (16) Käfer, D.; Ruppel, L.; Witte, G. Phys. Rev. B 2007, 75, 085309. (17) Kang, J. H.; Zhu, X. Y. Appl. Phys. Lett. 2003, 82, 3248−3250. (18) McDonald, O.; Cafolla, A. A.; Carty, D.; Sheerin, G.; Hughes, G. Surf. Sci. 2006, 600, 3217−3225. (19) Menozzi, C.; Corradini, V.; Cavallini, M.; Biscarini, F.; Betti, M. G.; Mariani, C. Thin Solid Films 2003, 428, 227−231. (20) Guaino, P.; Carty, D.; Hughes, G.; McDonald, O.; Cafolla, A. A. Appl. Phys. Lett. 2004, 85, 2777−2779. (21) Baldacchini, C.; Mariani, C.; Betti, M. G. J. Chem. Phys. 2006, 124, 154702. (22) Annese, E.; Viol, C. E.; Zhou, B.; Fujii, J.; Vobornik, I.; Baldacchini, C.; Betti, M. G.; Rossi, G. Surf. Sci. 2007, 601, 4242− 4245. (23) Söhnchen, S.; Lukas, S.; Witte, G. J. Chem. Phys. 2004, 121, 525−534. (24) Lukas, S.; Witte, G.; Wöll, C. Phys. Rev. Lett. 2002, 88, 028301. (25) Chen, Q.; McDowall, A. J.; Richardson, N. V. Langmuir 2003, 19, 10164−10171. (26) Lagoute, J.; Kanisawa, K.; Fölsch, S. Phys. Rev. B 2004, 70, 245415. (27) Wang, Y. L.; Ji, W.; Shi, D. X.; Du, S. X.; Seidel, C.; Ma, Y. G.; Gao, H. J.; Chi, L. F.; Fuchs, H. Phys. Rev. B 2004, 69, 075408. (28) Danisman, M. F.; Casalis, L.; Scoles, G. Phys. Rev. B 2005, 72, 085404. (29) Zhang, H. L.; Chen, W.; Huang, H.; Chen, L.; Wee, A. T. S. J. Am. Chem. Soc. 2008, 130, 2720−2721. (30) Eremtchenko, M.; Temirov, R.; Bauer, D.; Schaefer, J. A.; Tautz, F. S. Phys. Rev. B 2005, 72, 115430. (31) Käfer, D.; Witte, G. Chem. Phys. Lett. 2007, 442, 376−383. (32) Wong, S. L.; Huang, H.; Huang, Y. L.; Wang, Y. Z.; Gao, X. Y.; Suzuki, T.; Chen, W.; Wee, A. T. S. J. Phys. Chem. C 2010, 114, 9356− 9361. (33) Koch, N.; Vollmer, A.; Duhm, S.; Sakamoto, Y.; Suzuki, T. Adv. Mater. 2007, 19, 112−116. (34) Duhm, S.; Hosoumi, S.; Salzmann, I.; Gerlach, A.; Oehzelt, M.; Wedl, B.; Lee, T.-L.; Schreiber, F.; Koch, N.; Ueno, N.; Kera, S. Phys. Rev. B 2010, 81, 045418. (35) Koch, N.; Gerlach, A.; Duhm, S.; Glowatzki, H.; Heimel, G.; Vollmer, A.; Sakamoto, Y.; Suzuki, T.; Zegenhagen, J.; Rabe, J. P.; Schreiber, F. J. Am. Chem. Soc. 2008, 130, 7300−7304. (36) Dougherty, D. B.; Jin, W.; Cullen, W. G.; Reutt-Robey, J. E.; Robey, S. W. J. Phys. Chem. C 2008, 112, 20334−20339. (37) Sakamoto, Y.; Suzuki, T.; Kobayashi, M.; Gao, Y.; Fukai, Y.; Inoue, Y.; Sato, F.; Tokito, S. J. Am. Chem. Soc. 2004, 126, 8138−8140. (38) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (39) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;

CONCLUSIONS To summarize, in a combined multitechnique experimental and theoretical study, we investigated the initial growth, the thermal stability, and the electronic properties of perfluoropentacene (PFP) on Cu(111) surfaces and contrast its structure to that of pentacene (PEN). For the latter, only full monolayers could be imaged by scanning tunneling spectroscopy (STM), and they exhibited two different packing motifs. The first is composed of molecules maintaining the same orientation over large areas, whereas in the second, a zigzag arrangement of domains exists, where the azimuthal orientation of the long molecular axes alternates on the few nanometer length scale. In contrast, for PFP, single molecules and small clusters could be imaged already at submonolayer coverage due to pinning at step edges and to point defects within terraces. Upon thermal annealing, the density of immobile molecules on the surface increased. A strong reaction of PFP with the Cu surface was identified as the cause by photoelectron spectroscopy (PES). While PFP molecules in the first layer are laterally disordered, the formation of highly ordered molecular domains was observed in the second layer. From PES, complementary STM experiments on HOPG, and theoretically modeled STM images, we conclude that PFP molecules in the second layer are inclined by about 15° to the surface plane. DFT calculations that fully optimize the geometry of the entire bilayer/substrate system would potentially allow further insight, provided that van der Waals interaction between molecule and metal is adequately taken into account.60 Notably, we found the same unusual growth mode for PFP on Cu(111) that was previously discussed for PEN on Ag(111) in the literature.30,31,36 This highlights that ordered molecular overlayers can form directly on top of a disordered first wetting layer and that this phenomenon is not limited to a single particular material combination. Therefore, in future studies, it is advisable to carefully investigate the initial growth beyond the first monolayer under this aspect to reliably determine the structural characteristics of organic−metal interfaces and, consequently, interpret their electronic properties.



AUTHOR INFORMATION

Notes

The authors declare no competing financial interest.



REFERENCES

(1) Klauk, H. Chem. Soc. Rev. 2010, 39, 2643−2666. (2) Lee, S.; Koo, B.; Shin, J.; Lee, E.; Park, H.; Kim, H. Appl. Phys. Lett. 2006, 88, 162109. (3) Inoue, Y.; Sakamoto, Y.; Suzuki, T.; Kobayashi, M.; Gao, Y.; Tokito, S. Jpn. J. Appl. Phys. 2005, 44, 3663−3668. (4) Kahn, A.; Koch, N.; Gao, W. J. Polym. Sci., Part B: Polym. Phys. 2003, 41, 2529−2548. (5) Delgado, M. C. R.; Pigg, K. R.; Filho, D. A. S.; Gruhn, N. E.; Sakamoto, Y.; Suzuki, T.; Osuna, R. M.; Casado, J.; Hernndez, V.; Navarrete, J. T. L.; et al. J. Am. Chem. Soc. 2009, 131, 1502−1512. (6) Duhm, S.; Salzmann, I.; Heimel, G.; Oehzelt, M.; Haase, A.; Johnson, R. L.; Rabe, J. P.; Koch, N. Appl. Phys. Lett. 2009, 94, 033304. (7) Braun, S.; Salaneck, W. R.; Fahlman, M. Adv. Mater. 2009, 21, 1450−1472. (8) Hwang, J.; Wan, A.; Kahn, A. Mater. Sci. Eng., R 2009, 64, 1−31. (9) Bröker, B.; Hofmann, O. T.; Rangger, G. M.; Frank, P.; Blum, R.P.; Rieger, R.; Venema, L.; Vollmer, A.; Müllen, K.; Rabe, J. P.; et al. Phys. Rev. Lett. 2010, 104, 246805. (10) Duhm, S.; Heimel, G.; Salzmann, I.; Glowatzki, H.; Johnson, R. L.; Vollmer, A.; Rabe, J. P.; Koch, N. Nat. Mater. 2008, 7, 326−332. 7733

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734

The Journal of Physical Chemistry C

Article

Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision D.1; Gaussian, Inc.: Wallingford, CT, 2003. (40) Perdew, J. P.; Wang, Y. Phys. Rev. B 1992, 45, 13244. (41) Kresse, G.; Furthmüller, J. Comput. Mater. Sci. 1996, 6, 15−20. (42) Kresse, G.; Furthmüller, J. Phys. Rev. B 1996, 54, 11169−11186. (43) Blöchl, P. E. Phys. Rev. B 1994, 50, 17953−17979. (44) Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59, 1758−1775. (45) Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976, 13, 5188−5192. (46) Tersoff, J.; Hamann, D. R. Phys. Rev. B 1985, 31, 805−813. (47) Heimel, G.; Romaner, L.; Brédas, J. L.; Zojer, E. Surf. Sci. 2006, 600, 4548−4562. (48) Ferretti, A.; Baldacchini, C.; Calzolari, A.; Felice, R.; Di; Ruini, A.; Molinari, E.; Betti, M. G. Phys. Rev. Lett. 2007, 99, 046802. (49) Stranick, S. J.; Kamna, M. M.; Weiss, P. S. Surf. Sci. 1995, 338, 41−59. (50) Kawai, S.; Pawlak, R.; Glatzel, T.; Meyer, E. Phys. Rev. B 2011, 84, 085429. (51) Wandelt, K. Surf. Sci. 1991, 251/252, 387−395. (52) Pascual, J. I.; Jackiw, J. J.; Kelly, K. F.; Conrad, H.; Rust, H.-P.; Weiss, P. S. Phys. Rev. B 2000, 62, 12632−12635. (53) Mete, E.; Demiroglu, I.; Danisman, M. F.; Ellialtioglu, S. J. Phys. Chem. C 2010, 114, 2724−2729. (54) Bismarck, A.; Tahhan, R.; Springer, J.; Schulz, A.; Klapötke, T. M.; Zeil, H.; Michaeli, W. J. Fluorine Chem. 1997, 84, 127−134. (55) Koch, N.; Heimel, G.; Wu, J.; Zojer, E.; Johnson, R. L.; Brédas, J. L.; Müllen, K.; Rabe, J. P. Chem. Phys. Lett. 2005, 413, 390−395. (56) Salzmann, I.; Duhm, S.; Heimel, G.; Oehzelt, M.; Kniprath, R.; Johnson, R. L.; Rabe, J. P.; Koch, N. J. Am. Chem. Soc. 2008, 130, 12870−12871. (57) Chen, W.; Qi, D.-C.; Huang, H.; Gao, X.; Wee, A. T. S. Adv. Funct. Mater. 2011, 21, 410−424. (58) Heimel, G.; Salzmann, I.; Duhm, S.; Koch, N. Chem. Mater. 2011, 23, 359−377. (59) Witte, G.; Lukas, S.; Bagus, P. S.; Wöll, C. Appl. Phys. Lett. 2005, 87, 263502. (60) Sun, D.; Kim, D.-H.; Le, D.; Borck, O.; Berland, K.; Kim, K.; Lu, W.; Zhu, Y.; Luo, M.; Wyrick, J.; et al. Phys. Rev. B 2010, 82, 201410(R).

7734

dx.doi.org/10.1021/jp208582z | J. Phys. Chem. C 2012, 116, 7726−7734