Quantum Confinement of Surface Electrons by Molecular Nanohoop

In these conditions, different molecules could be expected to produce nearly identical confining potentials, leading to a 2D-network of nearly identic...
1 downloads 6 Views 3MB Size
Letter pubs.acs.org/JPCL

Quantum Confinement of Surface Electrons by Molecular Nanohoop Corrals Benjamen N. Taber, Christian F. Gervasi, Jon M. Mills, Dmitry A. Kislitsyn, Evan R. Darzi, William G. Crowley, Ramesh Jasti, and George V. Nazin* Department of Chemistry and Biochemistry, Materials Science Institute, University of Oregon, 1253 University of Oregon, Eugene, Oregon 97403, United States S Supporting Information *

ABSTRACT: Quantum confinement of two-dimensional surface electronic states has been explored as a way for controllably modifying the electronic structures of a variety of coinage metal surfaces. In this Letter, we use scanning tunneling microscopy and spectroscopy (STM/STS) to study the electron confinement within individual ring-shaped cycloparaphenylene (CPP) molecules forming self-assembled films on Ag(111) and Au(111) surfaces. STM imaging and STS mapping show the presence of electronic states localized in the interiors of CPP rings, inconsistent with the expected localization of molecular electronic orbitals. Electronic energies of these states show considerable variations correlated with the molecular shape. These observations are explained by the presence of localized states formed due to confinement of surface electrons by the CPP skeletal framework, which thus acts as a molecular electronic “corral”. Our experiments suggest an approach to robust large-area modification of the surface electronic structure via quantum confinement within molecules forming selfassembled layers.

T

assembled layers caused by the presence of surface irregularities. An alternative approach, which would be relatively robust against imperfections in self-assembly, would have to rely on electron confinement within the molecules forming the self-assembled layer, with the individual molecules thus serving as electronic corrals. In these conditions, different molecules could be expected to produce nearly identical confining potentials, leading to a 2D-network of nearly identical quantum confined surface states (QCSS). Here, we report, for the first time, electron confinement within individual ring-shaped cycloparaphenylene (CPP) molecules [Figure 1a] forming self-assembled films on Ag(111) and Au(111) surfaces. [n]CPPs, made up of n paralinked phenylenes, are the shortest-possible fragments of armchair carbon nanotubesthus coined “carbon nanohoops” when first reported in 2008.14 The nanohoops can be synthesized with angstrom level control over diameter,15−17 on the gram scale,18 and with structural modifications that can lead to tuning of the electronic properties.19,20 Thus, these structures represent fascinating candidates for exploration by STM; however, to date, no studies have been reported. In this work, we investigated the local electronic structures with STS by using an ultrahigh vacuum (UHV) cryogenic STM system21 (see Experimental Methods for further details). [8]CPP molecules were thermally sublimated in situ onto Ag(111) and Au(111) substrates held at room temperature to achieve a submonolayer coverage, and the substrates were subsequently

he capability to controllably modify the electronic properties of surfaces of materials is essential for the development and optimization of applications involving interfacial electron transfer, with examples including (opto)electronic and photovoltaic devices, catalysis, and sensing. Among the vast diversity of surfaces and surface-modification techniques pursued in this field of research, the surfaces of coinage metals (Au, Ag, and Cu) corresponding to the (111) crystallographic orientation have received special attention due to the unique susceptibility of the electronic structures of such surfaces to the presence of adsorbates. This sensitivity stems from the existence of free-electron-like two-dimensional (Shockley) electronic states, tightly localized in the vicinity of such surfaces.1 Modification of the electronic structures of such surfaces relies on controlling the propagation of surface electrons by using adsorbate-induced electronic scattering, as has been demonstrated in the pioneering experiments on surface state confinement in electronic “corrals” constructed from individual atoms and visualized with scanning tunneling microscopy (STM) and scanning tunneling spectroscopy (STS).2−4 Formation of standing electronic waves has been since demonstrated in a wide variety of systems including artificial atomic5 and molecular6 structures, and molecular-sized adsorbate islands.7,8 Of particular interest are self-assembled atomic9 and molecular structures,10−12 where scalable large-area modification of surface electronic properties could be achieved.13 All previous investigations, however, focused on confinement of electrons in two-dimensional voids formed by porous self-assembled molecular layers, which presents a potential problem because the properties of such electronic states are inherently sensitive to imperfections in the self© XXXX American Chemical Society

Received: June 10, 2016 Accepted: July 26, 2016

3073

DOI: 10.1021/acs.jpclett.6b01279 J. Phys. Chem. Lett. 2016, 7, 3073−3077

Letter

The Journal of Physical Chemistry Letters

To compare the LDOS spectra of the skeletal part of CCP rings to those of the MIS, we measured LDOS as a function of spatial location along a line cross-secting adsorbed molecules [one representative data set collected across five molecules on Ag(111) is shown in Figure 2]. The progression of recorded

Figure 2. LDOS spectra measured across several [8]CPP molecules on Ag(111) revealing molecular interior states (MIS). (a) STM image of the molecular self-assembled monolayer [set point 5 pA, 1.0 V] (b) LDOS as a function of the bias voltage and position x along the path recorded along the dashed white line in (a). Labels A−E in (b) correspond to the respective molecules labeled in (a). (c) Individual LDOS spectra recorded in the locations indicated by the dashed white lines in (b) showing the relative intensities of MIS vs the molecular skeletal states.

Figure 1. STM imaging of [8]CPP molecules adsorbed on Au(111). (a) Model of a [8]CPP molecule. (b) High-resolution STM image (obtained using a functionalized STM tip) of a well-ordered, twodimensional crystal of [8]CPPs adsorbed on Au(111). Both individual molecules and the component benzene subunits of the [8]CPPs are discernible [set point 5 pA, 1.0 V.] (c) Image from (b) with overlaid molecular structures. (d−k) Bias-dependent images of a twodimensional crystal of [8]CPPs adsorbed on Au(111) showing “eye”-like spatial features in molecular interiors at higher voltages, starting from 2.7 V.

spectra (Figure 2b) shows clearly identifiable MIS peaks A−E at ∼2.6 V (this voltage polarity corresponds to unoccupied states) significantly enhanced in the interiors of the corresponding molecules. In contrast, LDOS spectra of the [8]CPP skeletal rings are dominated by strongly enhanced LDOS at ∼2.9−3 V [referred to as molecular skeletal states (MSS) in the following], even though they still contain shoulder-like features consistent with MIS. Athough both MIS and MSS are present in each spectrum of Figure 2c, these states clearly show disparate spatial localizations suggesting their different origins. The different nature of the MIS and MSS orbitals is further illustrated by STS measurements on a less tightly packed [8]CPP layer, where individual molecules are adsorbed in slightly different environments (Figure 3a). For such molecules, MIS appear at voltages ranging from 2 to 2.6 V (Figure 3b), whereas the variations in the MSS onset and peak positions are more subdued (Figure 3c). This is also seen in 2D LDOS maps of this area, which show the appearance of MIS over a range of different voltages (Figure 3d−f), with the MSS appearing in the LDOS maps at nearly the same voltage (∼2.8 V), resulting in more uniform LDOS distributions (Figure 3g−h). Maps shown in Figure 3d−f also corroborate the expectation that the MIS are predominantly localized in the interiors of individual [8]CPP molecules, whereas the MSS are localized on the [8]CPP skeletal rings, consistent with Figure 2b. This is particularly clear in Figure 3d,g, which are dominated by the states of molecule I, energetically downshifted with respect to the rest of the molecules in the scan area. (Note that the MSS of molecule I, as compared to the MIS state, is shifted relatively insignificantly, by less than ∼100 mV [see curve i in Figure 3c.] Possible reasons of these energetic shifts are discussed below.) The pronounced variations observed for the energies of the MIS (EMIS in the following) in Figures 3 and S2, as well as their localization in the interior of molecular rings (as contrasted to

cooled to ∼20 K for STS measurements. STM imaging shows that during the [8]CPP deposition the [8]CPP molecules had sufficient thermal energy to migrate on the metal substrate and form two-dimensional self-assembled molecular crystals (Figure 1b). By using functionalized STM tips, we visualize the structures of individual [8]CPPs, including the constituent benzene rings, which allows us to conclude that [8]CPP rings are adsorbed flat [with the individual benzenes approximately perpendicular to the surface] on the metal substrates (Figure 1b−c). In the following, however, we avoid using functionalized STM tips due to the potential impact of functionalization on the STS spectra, which, at low bias voltages (less than ∼2 V), results in topographic images showing regular 2D arrays of topographic depressions corresponding to the molecular interiors (Figure 1d−f). Intriguingly, though the molecular appearance in STM images at low bias voltages was consistent with the expected [8]CPP geometry, at positive higher voltages additional localized topographic protrusions appeared in the interiors of [8]CPP rings (Figures 1g−i and S1c−f). Because these topographic protrusions cannot be associated with structural features of [8]CPP molecules, they must be caused by electronic states localized in the molecular interiors, and we will thus refer to them as molecular interior states (MIS) in the following. Interestingly, MIS appeared at different voltages for different molecules (Figure 1g−i and S1c−f). To understand this behavior, we carried out STS measurements of the energydependent local density of states (LDOS) spectra by recording the differential tunneling conductance (dI/dV) as a function of the applied bias voltage (see Experimental Methods for details of the measurements). 3074

DOI: 10.1021/acs.jpclett.6b01279 J. Phys. Chem. Lett. 2016, 7, 3073−3077

Letter

The Journal of Physical Chemistry Letters

states but also the variation of EMIS for different molecules, and substrates [Au(111) versus Ag(111)], as described below. A potential source of molecule-specific variations in EMIS observed in Figure 3b may be the varied conformations of molecular adsorption. Indeed, inspection of the STM topographies in Figures 1b, 2a, and 3a shows varied deviations from circular shape for most molecules. To quantitatively characterize the shape of each studied molecule, we approximated the molecular shape (observed in STM topography) by an ellipse, and calculated the corresponding eccentricity as e =

1−

B2 A2

, where A and B are, respectively,

the major and minor axes of the ellipse [see Figure S3 for details]. Thus, obtained values of e and EMIS for 84 molecules on Au(111), and 81 molecules on Ag(111), show that the two quantities are correlated (Figure 4), with EMIS increasing with

Figure 3. Two-dimensional spatial mapping of LDOS for a [8]CPP submonolayer on Ag(111). (a) STM image of the [8]CPP submonolayer [set point 5 pA, 2.0 V]. (b) LDOS spectra recorded in molecular interiors, with labels A−Q corresponding to respective locations in (a−f). (c) LDOS spectra recorded at the molecular backbones, with labels a to q corresponding to respective locations in (a). (d−h) 2D spatial LDOS maps of the area shown in (a), measured at bias voltages from 2.05 to 2.80 V, as indicated in individual maps.

Figure 4. MIS peak voltages versus eccentricities of [8]CPP molecules on Au(111) [blue dots] and Ag(111) [red dots]. Also shown are corresponding theoretical curves calculated using the particle-in-anelliptical-box model described in the text. Simulated MIS peak voltages for Au(111) are higher than those for Ag(111) due to the lower effective mass of surface electrons on Au(111). The data correspond to 84 molecules on Au(111) and 81 molecules on Ag(111), with molecule “I” from Figure 3 designated by a cross.

higher e corresponding to less circular molecular shapes. In addition, the EMIS values for molecules on Au(111) are clearly higher than those for molecules on Ag(111). These observations are qualitatively consistent with the expected behavior of quantum-confined surface electronic states of the Au(111) and Ag(111) substrates, as shown by theoretical curves in Figure 4 and explained in the following. To qualitatively investigate the dependence of the electronic confinement on the molecular shape, we solved a 2D Schrödinger equation for surface electrons in a confining potential produced by a [8]CPP molecule modeled as an elliptical ring. Analogously to recent theoretical studies of confined surface electronic states, we leave out the atomic-level details of the molecular potential, and employ the isotropic effective mass approximation such that the corresponding 2D Schrödinger equation is mathematically equivalent to that of the free electron in a confining 2D potential.6,24,25 Indeed, in the absence of adsorbates, for sufficiently low energies, surface electronic bands show near-parabolic energy−momentum dispersion relations,26 with effective masses of 0.28 me27 and 0.4 me28 for Au(111) and Ag(111), respectively. For higher energies, the energy−momentum dispersions significantly

the relatively iso-energetic MSS localized on the molecular rings), suggest that the MIS are not of molecular origin but may be instead associated with the states of the substrate altered by the presence of the molecules. This scenario is suggested by the fact that 2D surface electronic bands susceptible to the presence of adsorbates are known to exist on both Au(111) and Ag(111) surfaces.22 These electronic states can be described as propagating surface waves, and are easily scattered by adsorbates, which results in the formation of standing electronic waves.23 In particular, when the adsorbates form a confining geometry on the metal surface, the formation of quasilocalized resonance states is observed, as demonstrated in experiments on electronic corrals assembled from individual atoms,2−4 missing-molecule defects in self-assembled molecular monolayers on such metallic surfaces,6 and self-assembled molecular structures.10 According to the physical picture developed in these studies, in the present case the formation of MIS in an individual molecule may be a result of quasiconfinement of surface electrons by the [8]CPP skeletal framework, which thus acts as a molecular electronic “corral”, with the MIS corresponding to the lowest-energy confined state. This interpretation explains not only the localization of the eye 3075

DOI: 10.1021/acs.jpclett.6b01279 J. Phys. Chem. Lett. 2016, 7, 3073−3077

Letter

The Journal of Physical Chemistry Letters

In addition to explaining the general trends observed for the MIS, the QCSS model offers an explanation for the existence of MIS outliers (at low eccentricities) in Figure 4. A representative example of this behavior is the MIS localized on molecule I in Figure 3a (marked by a cross in Figure 4). LDOS maps measured at the onset of the corresponding MIS peak offer an explanation to for the relatively low energy of this MIS (Figure S5). Indeed, the LDOS of this state is spread out significantly outside of the corresponding molecular ring, especially in the top part of the map (Figure S5), in the vicinities of molecules H, F, and J in Figure 3a. The lack of confinement may be a result of the specific adsorption geometry of molecule I: as can be seen in Figure 3a, the profile of this molecule is uneven, suggesting a reduced degree of alignment of the benzene rings comprising this molecule, which may be enabled, in part, by their reduced steric hindrance caused by the more circular shape of the molecule. The reduced torsional alignment of benzene rings may result in a nonuniform, and less confining, potential barrier for the surface electrons, thus leading to delocalized LDOS (such as that observed in Figure S5) and a lower (due to reduced confinement) MIS energy. In conclusion, our work demonstrates electron confinement within individual ring-shaped [8]CPP molecules leading to a dramatic modification in the electronic LDOS, transforming the relatively featureless LDOS of Au(111) and Ag(111) surfaces into a LDOS distribution dominated by a peak associated with the lowest-energy QCSS. Our experiments thus suggest an alternative approach for controlling the surface electronic structure, in contrast to the previously demonstrated approaches where electron confinement was realized within 2D voids formed in self-assembled adsorbate superstructures. The advantage of the new approach is in its inherent insensitivity toward structural imperfections associated with relative placement of the adsorbates. Though [8]CPP molecules used in the present study were found in varied conformations on the Au(111) and Ag(111) surfaces, which resulted in varied energies of the QCSS, molecules with more rigid macrocycles can be expected to produce more uniform energy distributions, especially because small deviations from circular shape should result in minimal QCSS energy changes according to the PIAEB model (Figure 4). The present approach thus provides a pathway for controllable and scalable modification of surface electronic structure through judicial choice of molecular geometry.

deviate from parabolic and are better described by formulas obtained for a 2D tight-binding model.28,29 To account for this, the corresponding Schrödinger equation would have to either include terms of higher-order spatial derivatives, or alternatively, the Schrödinger equation can be written in its conventional form, where the corresponding eigenenergy can subsequently be used to calculate the nonparabolic correction.30 For purposes of the qualitative treatment presented here, we further assume that the confining potential produced by a [8]CPP molecule is infinitely high. For this potential, the problem is reduced to that of the particle-in-an-elliptical-box (PIAEB), which is readily solved by using elliptical coordinates.31 We note that even though a potential with walls of finite height would likely be a better approximation, wave functions inside such a potential can be qualitatively modeled by wave functions obtained for an infinite potential well with appropriate (larger) dimensions. We further note that potential barriers associated with [8]CPPs can be expected to be permeable, similarly to barriers obtained for other molecular systems.6,24,25 However, quantitative modeling of tunneling across the [8]CPP ring is not possible without taking into account the other decay channel associated with coupling to bulk states,27 which is not well quantified in the energy range under consideration. Thus, we leave more quantitative modeling of MIS states incorporating more complex confining potentials to future work. Ground state energies calculated for MIS states on Au(111) and Ag(111) using the PIAEB approach and corrected for nonparabolicity of the energy−momentum dispersions are shown in Figure 4 (solid curves) as functions of the molecular eccentricities e. Despite its simplicity, the qualitative PIAEB model reproduces the main trends observed in the experimental data of Figure 4. For example, the PIAEB ground state energies for Au(111) are higher than those for Ag(111), which can be qualitatively explained by the lower effective mass of the Au(111) surface electrons as compared to that of Ag(111). Significantly, the PIAEB model reproduces the upward trend of EMIS with increasing eccentricity e observed in the experimental data. Indeed, EMIS can be expected to increase with e due to the strong electron confinement along the minor molecular axis at high e. In contrast, as mentioned earlier, pronounced variations in EMIS are not accompanied by similar variations in MSS energies (compare Figure 3b to 3c), suggesting the molecular origin of MSS, with the onset of the MSS likely dominated by the lowest unoccupied molecular orbital. Indeed, molecular orbital energies may be expected to be relatively insensitive to minor bending variations in conformation, as was observed recently for oligothiophene molecules.32 Additional support for the interpretation of MIS as QCSS is offered by the fact that QCSS were clearly observed in the sufficiently large voids between individual [8]CPPs molecules. Figure S4 shows LDOS mapping for one such void where, in addition to MIS and MSS states observed in LDOS maps of this area (Figure S4b), at lower voltages a pair of QCSS states is observed (Figure S4c) with spatial behaviors consistent with those expected for the ground and first excited states confined inside an elongated area, as further illustrated in Figure S4d,e. Here, the fact that QCSS in CCP layer voids were observed at considerably lower voltages is consistent with the expected scaling of the energies of QCSS with the lateral size of the confining area.



EXPERIMENTAL METHODS STM and STS measurements were conducted at ∼20 K, with a home-built cryogenic (closed-cycle cryostat-based) UHV STM system featuring a Pan-style scanner from RHK Technology.21 STS measurements were carried out using the lock-in technique (modulation frequency 570 Hz), which gives differential conductance (dI/dV) proportional to the local density of states (LDOS).33 To prepare atomically flat Ag(111) and Au(111) substrates in situ, gold films on mica were repeatedly subjected to a cycle consisting of Ne-ion bombardment and subsequent ∼300 °C anneals. The [8]CPP molecules were prepared by Jasti and co-workers.18



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.6b01279. 3076

DOI: 10.1021/acs.jpclett.6b01279 J. Phys. Chem. Lett. 2016, 7, 3073−3077

Letter

The Journal of Physical Chemistry Letters



(14) Jasti, R.; Bhattacharjee, J.; Neaton, J. B.; Bertozzi, C. R. Synthesis, Characterization, and Theory of [9]-, [12]-, and [18]Cycloparaphenylene: Carbon Nanohoop Structures. J. Am. Chem. Soc. 2008, 130, 17646−7. (15) Darzi, E. R.; Jasti, R. The Dynamic, Size-Dependent Properties of [5 - 12]Cycloparaphenylenes. Chem. Soc. Rev. 2015, 44, 6401−6410. (16) Iwamoto, T.; Watanabe, Y.; Sakamoto, Y.; Suzuki, T.; Yamago, S. Selective and Random Syntheses of [n]Cycloparaphenylenes (n = 8−13) and Size Dependence of their Electronic Properties. J. Am. Chem. Soc. 2011, 133, 8354−8361. (17) Takaba, H.; Omachi, H.; Yamamoto, Y.; Bouffard, J.; Itami, K. Selective Synthesis of [12]Cycloparaphenylene. Angew. Chem., Int. Ed. 2009, 48, 6112−6116. (18) Xia, J.; Bacon, J. W.; Jasti, R. Gram-Scale Synthesis and Crystal Structures of [8]- and [10]CPP, and the Solid-State Structure of C60@[10]CPP. Chem. Sci. 2012, 3, 3018−3018. (19) Darzi, E. R.; Hirst, E. S.; Weber, C. D.; Zakharov, L. N.; Lonergan, M. C.; Jasti, R. Synthesis, Properties, and Design Principles of Donor-Acceptor Nanohoops. ACS Cent. Sci. 2015, 1, 335−342. (20) Kuwabara, T.; Orii, J.; Segawa, Y.; Itami, K. Curved Oligophenylenes as Donors in Shape-Persistent Donor-Acceptor Macrocycles with Solvatofluorochromic Properties. Angew. Chem., Int. Ed. 2015, 54, 9646−9649. (21) Hackley, J. D.; Kislitsyn, D. A.; Beaman, D. K.; Ulrich, S.; Nazin, G. V. High-Stability Cryogenic Scanning Tunneling Microscope Based on a Closed-Cycle Cryostat. Rev. Sci. Instrum. 2014, 85, 103704. (22) Han, P.; Weiss, P. S. Electronic Substrate-Mediated Interactions. Surf. Sci. Rep. 2012, 67, 19−81. (23) Crommie, M. F.; Lutz, C. P.; Eigler, D. M. Imaging Standing Waves in a 2-Dimensional Electron-Gas. Nature 1993, 363, 524−527. (24) Klappenberger, F.; Kühne, D.; Krenner, W.; Silanes, I.; Arnau, A.; Garcı ́a de Abajo, F. J.; Klyatskaya, S.; Ruben, M.; Barth, J. V. Tunable Quantum Dot Arrays Formed from Self-Assembled MetalOrganic Networks. Phys. Rev. Lett. 2011, 106, 106. (25) Kepcija, N.; Huang, T. J.; Klappenberger, F.; Barth, J. V. Quantum Confinement in Self-Assembled Two-Dimensional Nanoporous Honeycomb Networks at Close-Packed Metal Surfaces. J. Chem. Phys. 2015, 142, 101931. (26) Jeandupeux, O.; Burgi, L.; Hirstein, A.; Brune, H.; Kern, K. Thermal Damping of Quantum Interference Patterns of Surface-State Electrons. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 15926− 15934. (27) Echenique, P. M.; Osma, J.; Machado, M.; Silkin, V. M.; Chulkov, E. V.; Pitarke, J. M. Surface-State Electron Dynamics in Noble Metals. Prog. Surf. Sci. 2001, 67, 271−283. (28) Bürgi, L.; Petersen, L.; Brune, H.; Kern, K. Noble Metal Surface States: Deviations from Parabolic Dispersion. Surf. Sci. 2000, 447, L157−L161. (29) Schouteden, K.; Lievens, P.; Van Haesendonck, C. FourierTransform Scanning Tunneling Microscopy Investigation of the Energy Versus Wave Vector Dispersion of Electrons at the Au(111) Surface. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 79, 195409. (30) López-Villanueva, J. A.; Melchor, I.; Cartujo, P.; Carceller, J. E. Modified Schrodinger-Equation Including NonParabolicity for the Study of a 2-Dimensional Electron-Gas. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 48, 1626−1631. (31) van den Broek, M.; Peeters, F. M. Confined States in TwoDimensional Flat Elliptic Quantum Dots and Elliptic Quantum Wires. Phys. E (Amsterdam, Neth.) 2001, 11, 345−355. (32) Taber, B. N.; Kislitsyn, D. A.; Gervasi, C. F.; Mills, J. M.; Rosenfield, A. E.; Zhang, L.; Mannsfeld, S. C. B.; Prell, J. S.; Briseno, A. L.; Nazin, G. V. Real-Space Visualization of ConformationIndependent Oligothiophene Electronic Structure. J. Chem. Phys. 2016, 144, 194703. (33) Chen, C. J. Introduction to Scanning Tunneling Microscopy, 2nd ed.; Oxford University Press: New York, 2008.

Expanded experimental details, additional STM images of [8]CPP on Ag(111) and STS of [8]CPP on Au(111) and Ag(111), and a description of the PIAEB model. (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The STM instrument used in this work was constructed with support from the National Science Foundation under Grant DMR-0960211. B.N.T., C.F.G, J.M.M., D.A.K., W.G.C., and G.V.N. gratefully acknowledge the National Science Foundation CAREER grant CHE-1454036. E.R.D. and R.J. gratefully acknowledge the National Science Foundation (CHE1255219), the Sloan Foundation, and the Camille and Henry Dreyfus Foundation for financial support. The authors also thank Jeffery Cina for his comments on the manuscript.



REFERENCES

(1) Shockley, W. On the Surface States Associated with a Periodic Potential. Phys. Rev. 1939, 56, 317−323. (2) Crommie, M. F.; Lutz, C. P.; Eigler, D. M. Confinement of Electrons to Quantum Corrals on a Metal Surface. Science 1993, 262, 218−220. (3) Heller, E. J.; Crommie, M. F.; Lutz, C. P.; Eigler, D. M. Scattering and Absorbtion of Surface Electron Waves in Quantum Corrals. Nature 1994, 369, 464−466. (4) Manoharan, H. C.; Lutz, C. P.; Eigler, D. M. Quantum Mirages Formed by Coherent Projection of Electronic Structure. Nature 2000, 403, 512−515. (5) Kliewer, J.; Berndt, R.; Crampin, S. Scanning Tunnelling Spectroscopy of Electron Resonators. New J. Phys. 2001, 3, 22. (6) Seufert, K.; Auwärter, W.; de Abajo, F. J. G.; Ecija, D.; Vijayaraghavan, S.; Joshi, S.; Barth, J. V. Controlled Interaction of Surface Quantum-Well Electronic States. Nano Lett. 2013, 13, 6130− 6135. (7) Manai, G.; Radican, K.; Delogu, F.; Shvets, I. V. RoomTemperature Self-Assembly of Equilateral Triangular Clusters via Friedel Oscillations. Phys. Rev. Lett. 2008, 101, 4. (8) Kim, H. W.; Takemoto, S.; Minamitani, E.; Okada, T.; Takami, T.; Motobayashi, K.; Trenary, M.; Kawai, M.; Kobayashi, N.; Kim, Y. Confinement of the Pt(111) Surface State in Graphene Nanoislands. J. Phys. Chem. C 2016, 120, 345−349. (9) Didiot, C.; Pons, S.; Kierren, B.; Fagot-Revurat, Y.; Malterre, D. Nanopatterning the Electronic Properties of Gold Surfaces with SelfOrganized Superlattices of Metallic Nanostructures. Nat. Nanotechnol. 2007, 2, 617−621. (10) Pennec, Y.; Auwärter, W.; Schiffrin, A.; Weber-Bargioni, A.; Riemann, A.; Barth, J. V. Supramolecular Gratings for Tuneable Confinement of Electrons on Metal Surfaces. Nat. Nanotechnol. 2007, 2, 99−103. (11) Lobo-Checa, J.; Matena, M.; Müller, K.; Dil, J. H.; Meier, F.; Gade, L. H.; Jung, T. A.; Stöhr, M. Band Formation from Coupled Quantum Dots Formed by a Nanoporous Network on a Copper Surface. Science 2009, 325, 300−303. (12) Klappenberger, F.; Kühne, D.; Krenner, W.; Silanes, I.; Arnau, A.; Garcı ́a de Abajo, F. J.; Klyatskaya, S.; Ruben, M.; Barth, J. V. Dichotomous Array of Chiral Quantum Corrals by a Self-Assembled Nanoporous Kagome Network. Nano Lett. 2009, 9, 3509−3514. (13) Müller, K; Enache, M.; Stöhr, M. Confinement Properties of 2D Porous Molecular Networks on Metal Surfaces. J. Phys.: Condens. Matter 2016, 28, 153003. 3077

DOI: 10.1021/acs.jpclett.6b01279 J. Phys. Chem. Lett. 2016, 7, 3073−3077