Letter pubs.acs.org/NanoLett
Consecutive Charging of a Molecule-on-Insulator Ensemble Using Single Electron Tunnelling Methods Philipp Rahe,*,† Ryan P. Steele,‡ and Clayton C. Williams† †
Department of Physics and Astronomy and ‡Department of Chemistry, The University of Utah, Salt Lake City, Utah 84112-0830, United States S Supporting Information *
ABSTRACT: We present the local charge state modification at room temperature of small insulator-supported molecular ensembles formed by 1,1′-ferrocenedicarboxylic acid on calcite. Single electron tunnelling between the conducting tip of a noncontact atomic force microscope (NC-AFM) and the molecular islands is observed. By joining NC-AFM with Kelvin probe force microscopy, successive charge build-up in the sample is observed from consecutive experiments. Charge transfer within the islands and structural relaxation of the adsorbate/surface system is suggested by the experimental data. KEYWORDS: Localized charging, molecular charge state modification, single electron tunnelling, Kelvin probe force microscopy, ferrocene, calcite
R
adatoms17,18,21,33 as well as molecules19 on ultrathin insulating films. A Cu complex molecule on NaCl thin films was found to undergo a geometric reconfiguration while changing its charge state16 and charge stability was further revealed for trap states within oxide thin films.34 The situation when relaxation is present in the system can be well described within the Marcus theory.18,35,36 Here, we follow the bottom-up strategy using on-surface selfassembly37,38 for fabrication of a ferrocene-on-insulator system that is stable at room temperature. We use the tip of an atomic force microscope (AFM) for addressing, modifying and reading the charge state of small ferrocene ensembles using singleelectron tunnelling (SET) methods,39−41 applied to this system type for the first time, to allow the transfer of quantized amounts of charge at the single-electron limit. Interestingly, small ensembles of ferrocene molecules stabilize charge, with multiple charge states evident in our data. The experimental findings will be explained by a model involving three processes, namely, tip-island tunnelling, intra-island electron transport, and molecular relaxation for charge stabilization. A relaxation mechanism is discussed using density functional theory (DFT) calculations. Sample preparation and AFM experiments in the frequencymodulated noncontact (NC) mode42 were performed under ultrahigh vacuum (UHV) conditions at room temperature. In NC-AFM, the measurement channel frequency shift Δf is related to tip−sample interaction forces.43 Tip and backside
ealizing information processing at the level of single molecules1 has significantly progressed with the successful implementation of diverse single-molecule functionality such as diodes, transistors, or rectifiers.2,3 Besides this intensively followed, current-based approach for realizing computational devices, alternative paradigms such as field-based information propagation in quantum cellular automata (QCA) have been pursued during the past decade.4 Within the context of memory devices5 as well as chargebased QCA, redox-active molecules are promising materials.6,7 Especially ferrocene with its well-known and robust redox states8 is recognized for its potential.4,9,10 However, combining and interconnecting the functionality of single molecules into devices remains a challenge, due to the competing needs of isolating them from the environment yet offering local access for information exchange.11 Local charge state control of individual surface-supported species and surface defects has been presented before, mainly with the species in the gap of a scanning tunnelling microscope.12−19 With energy alignment of the donor or acceptor state with the external electrode an either transiently stable charge state13,20 or charge stability after relaxation16,17,19,21 has been observed, and the coupling to the external electrodes has been identified22 as a critical property. Transient charge states could be induced in quantum dots,23−25 nanoparticles,26 dopants,27−29 and molecules12−15,20 and fingerprints in the tunnelling current of a molecule in a doublebarrier tunnelling junction has been described in detail.30 In contrast, fewer studies identified stabilized charge states. Charge trapping is a phenomenon observed for quantum dots,31,32 while relaxation of the underlying or surrounding substrate has been found to stabilize oxidized or reduced metal © 2015 American Chemical Society
Received: September 15, 2015 Revised: December 15, 2015 Published: December 29, 2015 911
DOI: 10.1021/acs.nanolett.5b03725 Nano Lett. 2016, 16, 911−916
Letter
Nano Letters
In order to induce a single electron tunnelling event from the tip of an AFM (see Figure 1c), the quantum mechanical tunnelling rate (which can, as with scanning tunnelling microscopy,52 be adjusted by changing the tip−sample distance) is required to be comparable to the inverse measurement time, and the energy condition for elastic tunnelling has to be satisfied. For the latter, the applied tip voltage shifts the Fermi energy relative to the sample states; elastic tunnelling requires sample states within this voltageaccessible window. The tunnelling event itself is identified by the NC-AFM technique as a small-but-rapid change in the oscillation frequency of the cantilever due to the change in the electrostatic interaction,23,33,40,53 and in case of a charge stabilization, the charge state can afterwards be identified using Kelvin probe force microscopy (KPFM).44,46 Charge states are typically stabilized by a geometric reorganization within the sample35,36 as we will discuss later. Our protocol to modify the charge state of FDCA-on-calcite uses frequency shift-versus-voltage (Δf(V)) curves to induce and measure electron tunnelling events. With the tip fixed above an island at vertical distance ztunnel and starting at either zero bias or at the KPFM voltage, we ramp the tip voltage in either the positive (electron depletion or oxidation of molecules) or negative (electron accumulation or reduction of molecules) direction while sampling the frequency shift Δf. We manually abort the voltage ramp in each curve shortly after a rapid step in the frequency shift signal and subsequently ramp the voltage back to zero under the same conditions. A single Δf(V) curve (polynomial background subtracted54) acquired with this strategy is reproduced in Figure 2a. The forward (fw) curve (voltage ramped from zero to negative tip potential) clearly presents two tunnelling events at voltages Vevent of about −5.52 V and −5.57 V, each having a step height of about (3.3 ± 0.5)Hz. Remarkably, the backward (bw) curve (presented with a vertical offset in Figure 2a) does not reveal any reverse tunnelling events. Apart from possibly a few stochastic events, this absence is a common finding in our bw data54 and indicates that the injected charge is trapped in the FDCA-on-calcite system. Repeated Δf(V) curves at the same position with negative (positive) tip voltages aiming for reduction (oxidation) are performed in sequence A (B and C), respectively; data are reproduced in Figure 2b. A quantitative analysis of the clear shift of Vevent is plotted in Figure 2c; we find a linear dependence for each set until a downward shift occurs after curve 19, followed again by a temporary increase until curve 24. If the curves are executed with alternating positive/negative sweep directions, this systematic progression is absent.54 The absence of systematic reverse tunnelling events in the bw curves54 already suggests trapping of injected electrons on the time scale of our experiment, and the approximately linear increase of Vevent observed during successive voltage ramps furthermore suggests a gradual charging of the FDCA-oncalcite system in which multiple charges are stable. This finding is further substantiated by measuring the KPFM signal before and after each tunnelling sequence A to C; see Figure 3. The KPFM data56 reveals a clear negative shift of the addressed island after the first sequence A (Figure 3b), while the island charge state within the experimental measurement uncertainty is recovered to the initial value during sequences B and C. Charge transfer is a ubiquitous process in biological and physical as well as chemical systems and was first successfully
PtIr-coated as well as highly doped Si cantilevers were used for the tunnelling and high-resolution imaging experiments, respectively. Molecular charge states were identified44 with Kelvin probe force microscopy (KPFM) in the frequency modulation (FM) mode.45,46 All voltages including the KPFM voltage ΔV refer to the tip potential, and the fast and slow scan directions are marked by arrows next to the physical channel in all image data; further experimental details including the driftcompensation strategy can be found in the Supporting Information. Density functional theory (DFT) calculations were performed on model complexes of this system with the ωB97X functional47 and the 6-31G(d,p) atom-centered basis set, using the Q-Chem quantum chemistry software package.48 The ionic lattice substrate was modeled as an isolated 2 × 1 × 1 slab, fixed in the experimentally known bulk structure. A single molecule was adsorbed onto this surface and allowed to relax in several possible configurations. While the interaction between adsorbate molecules is known to affect the morphology of the monolayer,49 the strong interaction found between the adsorbate and the surface suggests that this single-adsorbate model should capture the majority of the binding motif. Further details on the methods are given in the Supporting Information. Submonolayer coverages (about 0.3 ML) of 1,1′-ferrocenedicarboxylic acid (FDCA, see Figure 1b) are deposited on the
Figure 1. (a) Bulk-truncated calcite surface structure, (b) sketch of the FDCA molecule, (c) geometry of the tip−sample system, (d) NCAFM topographic image of FDCA on the calcite(10.4) surface, a mean island area of about 150 nm2 is measured, corresponding to about 185 molecules per island. (e) High-resolution image of the (2 × 1) structure within one island.
insulating calcite(10.4) surface (Figure 1a), forming small, single-layered, ordered but well-separated molecular islands (Figure 1d). The FDCA molecules assemble on a rectangular (2 × 1) superstructure within these islands (Figure 1e, further details will be published elsewhere). The assembly is stable at room temperature over at least several days. Our experiments to modify the charge state of the moleculeon-insulator system herein follow the SET methodology, which operates on nonconducting systems in contrast to atomic or molecular charge state modification based on tunnelling currents.16,17 Force detected tunnelling methods have successfully been employed earlier to investigate, for example, single defect states in thin oxide films39,40 as well as charging of quantum dots,23 single metal adatoms on thick NaCl films,33 and nanoclusters on an oxide.32 An experimental mode based on dynamic tunnelling has recently been developed to provide full images.50,51 We here apply this single-electron tunnelling method for the first time to this molecule-on-insulator system. 912
DOI: 10.1021/acs.nanolett.5b03725 Nano Lett. 2016, 16, 911−916
Letter
Nano Letters
Figure 2. (a) Single Δf(V) forward (fw) and backward (bw) curve (background subtracted using a polynomial of order 3rd (6th) of fw (bw) data, respectively). (b) Consecutive sets (labeled A, B, C) of Δf(V) curves. Before and after each set A, B, and C, KPFM images were acquired; see Figure 3. Δf(V) curves are shifted along the Δf axis for clarity. The upper left inset shows a representative total interaction force curve54 as a function of the tip height with a tip−sample distance of about 9 Å (relative to Fmin) chosen for the tunnelling experiments. The used tip height was identified to here constitute the threshold for observing tunnelling events. (c) Tip voltages of the first tunnelling event determined for each curve in panel b. We find from linear fits slopes of (−270 ± 50) mV/curve ((160 ± 20) mV/curve) for curves 1−8 (9−19), respectively. Note that the voltage axes in panels b and c are interrupted.
Figure 3. (a−d) KPFM data and (e) topography data for one set of tunnelling experiments. The tip with an estimated radius of 20 nm is paused on the lower left island as indicated in (e) and tunnelling with a negative (positive) tip voltage is performed before panel (b) (panels c and d), respectively. The charge neutrality of the islands was reestablished after preceding tunnelling experiments. DFT calculations reveal a small dipole within individual FDCA molecules in the gas phase, pointing in the on-calcite geometry with the positive end outward from the surface. This model is in agreement with measuring a positive KPFM ΔV voltage for the as-deposited, neutral FDCA ensembles.44 We furthermore see no evidence in control experiments (data not shown) or computed properties for an initial net charging of the species during the sublimation or during the first tip contact. We can also exclude that these observations are caused by charging the tip55 during this experiment as the Kelvin signal on a reference island nearby (marked by “ref”, see line profiles in f) remains identical within the noise limit. The blue colorscale applies to all KPFM images. The topography image is acquired at a large tip−sample distance with reduced short-range interactions to minimize the risk of tip contamination.
described by Marcus theory.35,36 Charge transfer between molecules is usually attended or driven by a change in geometry of the participant molecules and/or surrounding medium which ultimately stabilizes the transferred charge. For example, relaxation of a single molecule after ionization has been mapped.16 We performed ab initio calculations for the neutral, anionic, and cationic FDCA molecule in vacuum and found a large stabilization of the FDCA anion due to molecular relaxation (see Figure 4b), which includes movement of the carboxylic acid groups toward each other. This stabilization is actually sufficient to render the anion more stable than the neutral analogue, in this new geometry, as subsequent removal of the excess electron is calculated to be an uphill process by about 0.2 eV. In contrast, the FDCA cation (see Figure 4c) remains higher in energy than its neutral analogue at either geometry. While this inner-sphere stabilization of the FDCA anion in vacuum would already explain our experimental finding of negative charge stabilization, this analysis is expected to oversimplify the situation by neglecting the supporting environment. In solution, for example, the oxidized species (ferrocenium) is surrounded by a polar medium or, if present, by counterions.57
DFT calculations reveal an upright geometry of FDCA on calcite (10.4), characterized by a bidentate, hydrogen-bonded tethering (see also Figure 4a) between the molecular carboxyl groups and the surface carbonates. The Fe center of the ferrocene is located above an interstitial site of the underlying calcite lattice. Due to the use of this (uncapped) finite-slab model, contaminating slab-edge orbitals−which would artificially ionize prior to the adsorbate−prevent direct simulation of the ionized adsorbate in the presence of the calcite bulk and surface. Nonetheless, an environmental influence from the calcite environment is extremely plausible. When injecting a negative charge from the AFM tip, the carbonate CO32− and calcium Ca2+ ions of calcite likely undergo relaxations, and we expect a similar response as observed for surface ions in the case of Au− or Ag− on NaCl films18,21,33 where the surface cations (anions) relax toward (are repelled from) the molecular 913
DOI: 10.1021/acs.nanolett.5b03725 Nano Lett. 2016, 16, 911−916
Letter
Nano Letters
Figure 5. (a) Topography image, (b) KPFM data before and (c) KPFM data after a single Δf(V) curve. Line profiles in d reveal an increase of the KPFM signal at the tip position (marked in a with estimated tip radius of 20 nm), (e) KPFM signal on reference island as an estimate for the measurement accuracy. All line profiles are averaged over 20 pixels. (f) A large number of tunnelling events occur during a single voltage ramp in DTFM mode50 and positive tip voltage scan direction without aborting after the first event. Island outline after massive electron tunnelling between island and tip. In this experiment, the tip was periodically approached and retracted along z with an applied voltage. Due to instabilities occurring at these conditions, temporarily reduced tip-island distances allowing for contact charging cannot be excluded.
Figure 4. Numerical calculations at the DFT level. (a) Relaxed FDCAon-calcite structure. Relaxations within the unbound FDCA molecule for the anionic (b) and cationic (c) case (energies calculated relative to the free neutral molecule). Step 1: e− injection/removal. Step 2: anion/cation nuclear relaxation. Step 3: removal/restoration of e−. Step 4: neutral nuclear relaxation. (Energy levels not to scale).
anion, respectively, in all cases causing a stabilization by inducing an energy barrier for the reverse, electron-removal process. In agreement with our data shown in Figure 2, a larger positive tip potential is necessary to allow for electron extraction from the reduced FDCA-on-calcite. Between curves 8 and 9, we measure a difference of the applied voltages of about 9.7 V between charge injection and extraction, and we find a voltage of about −5.5 V to be necessary for the first charging step. Thus, in agreement with the earlier findings, the charge is stabilized, even at zero bias. Movement of the ions in the calcite (10.4) surface has, in fact, been suggested already in the presence of scanning probe tips terminated by ions.58 We furthermore note that only a fraction of the actual applied voltage is dropped between the probe tip and the surface due to the tip geometry and the finite thickness of the dielectric calcite sample. When assuming54 a probe formed by half-sphere, shank, and cantilever with tip radius of 20 nm, we find the true energy difference between the last injection (curve 8) and the first extraction (curve 9) to be about 2.0 eV. The multiple charging that we observe in repeated experiments is unlikely to occur within a single FDCA molecule because it is energetically unfavorable; we rather consider the presence of multiple sites, each carrying an injected charge. Consequently, our data suggests that sites adjacent to the tip apex tunnelling junction are accessible. Evidence for intraisland charge transport is given from the results presented in Figure 5. Although KPFM data is subject to spatial signal broadening due to the finite tip size,46 which can often render the determination of the exact size of the charged region inconclusive unless KPFM is operated under for this sample system so far unaccessible conditions,59 we can clearly resolve single FDCA islands with KPFM and especially find their imaged width to be smaller than the estimated tip radius of 20 nm. Clear differences between islands being charged with very few electrons (see Figure 5a−e) and islands that have been
charged repeatedly are apparent from the KPFM data. Furthermore, islands that were subject to massive electron injection undergo a structural transition (see Figure 5f). As shown in Figure 3 before, consecutive charging results in a homogeneous shift of the KPFM signal across the island, suggesting a homogeneous charge distribution as a result of mobility within the FDCA island, likely induced by Coulomb repulsion between the charges of same polarity. In contrast, we can reveal a gradient across an FDCA island (see Figure 5a−e) after injection of very few electrons as a result of only a single Δf(V) curve. Remarkably, however, the transfer of a large number of electrons leads to destruction of the island as is shown in Figure 5f. From the total step size of 800 Hz in the corresponding Δf(V) curve and based on the finding that transfer of a single electron typically causes a change in frequency shift of maximum 6 Hz, we estimate that at least 100 electrons are transferred between sample and tip in this experiment. Because areas at distances from the tip site larger than the tip radius are affected, these data strongly suggest that charge transport across whole islands is possible. Consequently, we propose the existence of three processes explaining our experimental findings: First, electrons tunnel from the tip into one or very few adjacent FDCA molecules at the tip apex position with rate rt−s(z,ΔV) depending on the tip−sample distance z and the tip−surface potential difference ΔV. The latter parameter is defined by the applied tip potential and the charge present within the island, in agreement with our experimental observation that further tip−sample tunnelling is only possible at increased magnitude of the applied voltages due to the changed island potential. Second, this step is followed by a discharge in favor of charging other FDCA sites nearby. This discharge can be modeled by a nearest-neighbor transport rate rn−n. Such electron transfer within the island 914
DOI: 10.1021/acs.nanolett.5b03725 Nano Lett. 2016, 16, 911−916
Letter
Nano Letters Present Address
would be reasonably expected between (nearly) identical redox centers which are in van der Waals contact. Hopping to closeby redox centers has been observed before, where rates in the order of 100−107 s−1 have been measured in proteins at room temperature.60 Interestingly, electron exchange between ferrocene moieties has also become evident for ferrocenes tethered to Au surfaces.61 Third, the finding of a large gap between reduction and reoxidation of the FDCA-on-calcite system requires the presence of a relaxation mechanism at rate rrelax. Superficially, the latter two effectsthe nonadiabatic charge transfer between adsorbate sites as well as the charge-trapping, adiabatic localization due to molecular relaxationmay appear at odds with one another. However, our data reveals that the two are compatible for the present system (rn−n > rrelax). Initial injection from the tip to the adsorbate produces an FDCA anion, temporarily in the original adsorbate geometry before molecular relaxation. The extent of electron transfer within the island would then simply be dictated by the relative magnitude of hopping and molecular relaxation rates. In the present interpretation, some initial transfer is observed, whereas eventual stabilization of the injected charge at a given trap site may lead to slowed transfer and the observed hysteretic charging effect. Future investigation into the magnitude and time scales of these effects, as well as their role in the final distribution of the injected charge, is encouraged. In conclusion, we realized localized charging and discharging of the FDCA-on-calcite system by addressing individual molecular islands with the tip of an AFM. Signatures of controllable single-electron tunnelling are observed during the charging experiments and reveal sample states accessible within the tip-defined voltage range. This system furthermore allows control over the magnitude of charges present on the molecular island from performing a predefined number of consecutive single-electron tunnelling experiments. Multiple charge states were stable and could be mapped with KPFM without modifying their values. The experiments suggest the presence of three processes within the FDCA-on-calcite system, namely tip−sample tunnelling, intra-island electron transfer and molecular relaxation. The success of charging small molecular structures individually addressed by a probe on a truly insulating surface opens the door for future studies of electronic interactions, relaxation and charge exchange between single molecules at the atomic scale.
■
P.R.: School of Physics and Astronomy, The University of Nottingham, University Park, NG7 2RD Nottingham, UK. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This study has been supported by the University of Utah Research foundation. P.R. gratefully acknowledges financial support from the Alexander von Humboldt-Foundation. The support and resources from the Center for High-Performance Computing at the University of Utah are gratefully acknowĺ edged. We furthermore thank Pavel Jelinek for most helpful discussions.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b03725. Sample preparation and methods. Determination of the absolute tip height during the tunnelling experiments. Estimation of the effective surface potential. Polynomial background subtraction of the Δf(V) data. Presentation of all backward Δf(V) curves. Alternating voltage sweeps (PDF)
■
REFERENCES
(1) Aviram, A.; Ratner, M. A. Chem. Phys. Lett. 1974, 29, 277−283. (2) Carroll, R. L.; Gorman, C. B. Angew. Chem., Int. Ed. 2002, 41, 4378−4400. (3) Song, H.; Reed, M. A.; Lee, T. Adv. Mater. 2011, 23, 1583−1608. (4) Lent, C. S.; Isaksen, B.; Lieberman, M. J. Am. Chem. Soc. 2003, 125, 1056−1063. (5) Lindsey, J. S.; Bocian, D. F. Acc. Chem. Res. 2011, 44, 638−650. (6) Liu, Z.; Yasseri, A. A.; Lindsey, J. S.; Bocian, D. F. Science 2003, 302, 1543−1545. (7) Quardokus, R. C.; Wasio, N. A.; Forrest, R. P.; Lent, C. S.; Corcelli, S. A.; Christie, J. A.; Henderson, K. W.; Alex Kandel, S. Phys. Chem. Chem. Phys. 2013, 15, 6973−6981. (8) Gritzner, G.; Kuta, J. Pure Appl. Chem. 1984, 56, 461. (9) Li, Q.; Mathur, G.; Gowda, S.; Surthi, S.; Zhao, Q.; Yu, L.; Lindsey, J. S.; Bocian, D. F.; Misra, V. Adv. Mater. 2004, 16, 133−137. (10) Zhu, H.; Hacker, C. A.; Pookpanratana, S. J.; Richter, C. A.; Yuan, H.; Li, H.; Kirillov, O.; Ioannou, D. E.; Li, Q. Appl. Phys. Lett. 2013, 103, 053102. (11) Heath, J. R. Annu. Rev. Mater. Res. 2009, 39, 1−23. (12) Fernández-Torrente, I.; Franke, K. J.; Pascual, J. I. Phys. Rev. Lett. 2008, 101, 217203. (13) Fernández-Torrente, I.; Kreikemeyer-Lorenzo, D.; Strózė cka, A.; Franke, K. J.; Pascual, J. I. Phys. Rev. Lett. 2012, 108, 036801. (14) Mikaelian, G.; Ogawa, N.; Tu, X. W.; Ho, W. J. Chem. Phys. 2006, 124, 131101. (15) Hauptmann, N.; Hamann, C.; Tang, H.; Berndt, R. Phys. Chem. Chem. Phys. 2013, 15, 10326−10330. (16) Leoni, T.; Guillermet, O.; Walch, H.; Langlais, V.; Scheuermann, A.; Bonvoisin, J.; Gauthier, S. Phys. Rev. Lett. 2011, 106, 216103. (17) Repp, J.; Meyer, G.; Olsson, F. E.; Persson, M. Science 2004, 305, 493−495. (18) Olsson, F. E.; Paavilainen, S.; Persson, M.; Repp, J.; Meyer, G. Phys. Rev. Lett. 2007, 98, 176803. (19) Swart, I.; Sonnleitner, T.; Repp, J. Nano Lett. 2011, 11, 1580− 1584. (20) Kocić, N.; Weiderer, P.; Keller, S.; Decurtins, S.; Liu, S. X.; Repp, J. Nano Lett. 2015, 15, 4406−4411. (21) Gross, L.; Mohn, F.; Liljeroth, P.; Repp, J.; Giessibl, F. J.; Meyer, G. Science 2009, 324, 1428−1431. (22) Gohda, Y.; Pantelides, S. T. Nano Lett. 2005, 5, 1217−1220. (23) Stomp, R.; Miyahara, Y.; Schaer, S.; Sun, Q.; Guo, H.; Grutter, P.; Studenikin, S.; Poole, P.; Sachrajda, A. Phys. Rev. Lett. 2005, 94, 056802. (24) Zaknoon, B.; Bahir, G.; Saguy, C.; Edrei, R.; Hoffman, A.; Rao, R. A.; Muralidhar, R.; Chang, K.-M. Nano Lett. 2008, 8, 1689−1694. (25) Tekiel, A.; Miyahara, Y.; Topple, J. M.; Grutter, P. ACS Nano 2013, 7, 4683−4690. (26) Azuma, Y.; Kanehara, M.; Teranishi, T.; Majima, Y. Phys. Rev. Lett. 2006, 96, 016108.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. 915
DOI: 10.1021/acs.nanolett.5b03725 Nano Lett. 2016, 16, 911−916
Letter
Nano Letters (27) Teichmann, K.; Wenderoth, M.; Loth, S.; Ulbrich, R. G.; Garleff, J. K.; Wijnheijmer, A. P.; Koenraad, P. M. Phys. Rev. Lett. 2008, 101, 076103. (28) Song, C.-L.; Jiang, Y.-P.; Wang, Y.-L.; Li, Z.; Wang, L.; He, K.; Chen, X.; Ma, X.-C.; Xue, Q.-K. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 045441. (29) Cui, Y.; Nilius, N.; Freund, H.-J.; Prada, S.; Giordano, L.; Pacchioni, G. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 88, 205421. (30) Nazin, G. V.; Wu, S. W.; Ho, W. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 8832−8837. (31) Hummon, M. R.; Stollenwerk, A. J.; Narayanamurti, V.; Anikeeva, P. O.; Panzer, M. J.; Wood, V.; Bulović, V. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 115439. (32) Zheng, N.; Johnson, J. P.; Williams, C. C.; Wang, G. Nanotechnology 2010, 21, 295708. (33) Steurer, W.; Repp, J.; Gross, L.; Scivetti, I.; Persson, M.; Meyer, G. Phys. Rev. Lett. 2015, 114, 036801. (34) Winslow, D.; Williams, C. C. J. Appl. Phys. 2011, 110, 114102. (35) Marcus, R. A. Rev. Mod. Phys. 1993, 65, 599−610. (36) Barbara, P. F.; Meyer, T. J.; Ratner, M. A. J. Phys. Chem. 1996, 100, 13148−13168. (37) Whitesides, G. M.; Mathias, J. P.; Seto, C. T. Science (Washington, DC, U. S.) 1991, 254, 1312−19. (38) Barth, J. V.; Costantini, G.; Kern, K. Nature (London, U. K.) 2005, 437, 671−679. (39) Bussmann, E.; Williams, C. C. Appl. Phys. Lett. 2006, 88, 263108. (40) Bussmann, E.; Kim, D. J.; Williams, C. C. Appl. Phys. Lett. 2004, 85, 2538−2540. (41) Bussmann, E.; Zheng, N.; Williams, C. C. Nano Lett. 2006, 6, 2577−2580. (42) Albrecht, T. R.; Grütter, P.; Horne, D.; Rugar, D. J. Appl. Phys. 1991, 69, 668−673. (43) Giessibl, F. J. Phys. Rev. B: Condens. Matter Mater. Phys. 1997, 56, 16010−16015. (44) Neff, J. L.; Rahe, P. Phys. Rev. B: Condens. Matter Mater. Phys. 2015, 91, 085424. (45) Nonnenmacher, M.; O’Boyle, M. P.; Wickramasinghe, H. K. Appl. Phys. Lett. 1991, 58, 2921−2923. (46) Melitz, W.; Shen, J.; Kummel, A. C.; Lee, S. Surf. Sci. Rep. 2011, 66, 1−27. (47) Chai, J.-D.; Head-Gordon, M. J. Chem. Phys. 2008, 128, 084106. (48) Shao, Y.; et al. Phys. Chem. Chem. Phys. 2006, 8, 3172. (49) Steele, R. P.; DiStasio, R. A., Jr; Head-Gordon, M.; Li, Y.; Galli, G. Phys. Chem. Chem. Phys. 2010, 12, 82−96. (50) Johnson, J.; Zheng, N.; Williams, C. C. Nanotechnology 2009, 20, 055701. (51) Wang, R.; King, S. W.; Williams, C. C. Appl. Phys. Lett. 2014, 105, 052903. (52) Binnig, G.; Rohrer, H.; Gerber, C.; Weibel, E. Appl. Phys. Lett. 1982, 40, 178−180. (53) Klein, L. J.; Williams, C. C. Appl. Phys. Lett. 2001, 79, 1828− 1830. (54) For details see Supporting Information. (55) Teobaldi, G.; Lämmle, K.; Trevethan, T.; Watkins, M.; Schwarz, A.; Wiesendanger, R.; Shluger, A. L. Phys. Rev. Lett. 2011, 106, 216102. (56) Shifts in the KPFM voltage are additionally revealed when fitting a polynomial function to the Δf(V) curves during the tunnelling experiments. (57) Rowe, G. K.; Creager, S. E. Langmuir 1991, 7, 2307−2312. (58) Foster, A. S.; Shluger, A. L.; Nieminen, R. M. Appl. Surf. Sci. 2002, 188, 306−318. (59) Mohn, F.; Gross, L.; Moll, N.; Meyer, G. Nat. Nanotechnol. 2012, 7, 227. (60) Gray, H. B.; Winkler, J. R. Annu. Rev. Biochem. 1996, 65, 537− 561. (61) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am. Chem. Soc. 1990, 112, 4301−4306. 916
DOI: 10.1021/acs.nanolett.5b03725 Nano Lett. 2016, 16, 911−916
