# Fileset

[PTCDA_InSi.pdf](https://mdr.nims.go.jp/filesets/bb4b8a65-2627-47cc-8e5f-3c9c6c8215e7/download)

## Creator

Shunsuke Inagaki, Narunori Ebara, Takahiro Kobayashi, Ryota Itaya, Kenta Yokota, Isamu Yamamoto, Jacek Osiecki, Khadiza Ali, Craig Polley, H. M. Zhang, L. S. O. Johansson, [Takashi Uchihashi](https://orcid.org/0000-0003-0811-5665), Kazuyuki Sakamoto

## Rights

©2024 American Physical Society[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Effects of adsorbed molecular ordering to the superconductivity of a two-dimensional atomic layer crystal](https://mdr.nims.go.jp/datasets/b188bc28-c68c-493e-86e8-3a1fb9d417ca)

## Fulltext

Effect of adsorbed molecular ordering to the superconductivityof a 2D atomic layer crystalShunsuke Inagaki,1 Narunori Ebara,1 Takahiro Kobayashi,2 Kenta Yokota,3, 4 Isamu Yamamoto,5 Jacek Osiecki,6Khadiza Ali,6 Craig Polley,6 H.M. Zhang,7 L. S. O. Johansson,7 Takashi Uchihashi,3, 4 and Kazuyuki Sakamoto1, 8, 91Department of Applied Physics, Osaka University, Osaka 565-0871, Japan2Department of Material and Life Science, Osaka University, Osaka 565-0871, Japan3Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan4International Center for Materials Nanoarchitectonics,National Institute for Materials Science, Ibaraki 305-0044, Japan5Synchrotron Light Application Center, Saga University, Saga 840-8502, Japan6MAX IV Laboratory, Lund University, Lund 221 00, Sweden7Department of Engineering and Physics, Karlstad University, SE-651 88 Karlstad, Sweden8Center for Spintronics Research Network, Osaka University, Osaka 560-8531, Japan9Spintronics Research Network Division, OTRI, Osaka University, Osaka 565-0871, Japan(Dated: November 28, 2022)The effect of 3,4,9,10-perylene-tetracarboxylic-dianhydride (PTCDA) adsorption to the physicalproperties of a two-dimensional (2D) atomic-layer superconductor (ALS) In/Si(111)-(√7×√3) hasbeen studied by angle-resolved photoelectron spectroscopy, transport measurements and scanningtunneling microscopy. Hole doping from the adsorbed molecules has been reported to increase theTc of this ALS, and the molecular spin to decrease it. Owing to its large electron affinity and itsnonexistent spin state, the adsorption of PTCDA was expected to increase the Tc. However, thePTCDA adsorption dopes only a small amount of holes and causes a suppression in Tc with a sharpincrease in the normal-state sheet resistance, followed by an insulating transition. Taking disorderingof the arrangement of PTCDA into account, we conclude that the increase in conductivity resultsfrom the localization effect resulting from the random potential that is induced by the disorderedPTCDA molecules. The present result also indicates the importance of the crystallinity of 2Dmolecular film adsorbed on ALSs.I. INTRODUCTIONIn the past decades, atomic-layer superconductors(ALSCs) epitaxially grown on semiconductor surfaceshave been intensively studied as a very fascinatingplatform for studying the physical properties of two-dimensional (2D) superconductors (SCs) [1–5]. Intrigu-ing physical phenomena, which require understanding be-yond a 3D conventional SC [6], have been reported suchfor 2D systems, e.g., enhanced in-plane critical magneticfield above the Pauli paramagnetic limit [7–10]. Fur-thermore, the combination of spin physics, which arisesfrom the presence of both the spin-orbit coupling (SOC)and broken inversion symmetry, with superconductivityhas possibility of creating novel superconducting states[11, 12]. In terms of applications as well as fundamen-tal science, further development of this field demandsimprovement and control of superconducting properties.By utilizing the high sensitivity of ALSs to surface ad-sorbates, a new method for tuning the superconduct-ing transition temperature (Tc) by constructing 2D het-erostructures consisting of well-ordered organic molecule,has been reported [13]. In this study, two different metalphthalocyanines, CuPc and MnPc, were adsorbed on anALS In/Si(111)-(√7×√3) (referred to as (√7×√3)-In,hereafter). The (√7×√3)-In consists of double In atomiclayer on Si(111), and form a (√7×√3) supercell with aquasi-square lattice [14–16]. It has spin-polarized metal-lic surface bands in normal state [17, 18], and become su-perconducting at approximately 3 K [1–3, 19, 20]. Eventhough the two molecules have similar ordered structureson (√7 ×√3)-In, Tc was found to be enhanced by 5%when using CuPc, while the Tc was rapidly suppressedfor MnPc adsorption. This Tc modification was explainedas the consequences of a competition between a posi-tive effect of hole doping [21] from the molecules into(√7×√3)-In and a negative effect of the exchange inter-action between the conduction-electron spin and the lo-cal spin in the molecules. However, no organic moleculesexcept CuPc and ZnPc have ever been reported to in-crease the Tc [13, 22–24]. In order to better conceptualizethe effect of molecular adsorption on the physical prop-erties of ALSCs, further investigations using differenttype of organic molecules are needed. 3,4,9,10-perylene-tetracarboxylic-dianhydride (PTCDA) is known to formhighly ordered layers on selected metals [25, 26], and thushas been used as a prototypical molecule to study prop-erties at the organic molecule-metal interface. Further-more, PTCDA is also used as an electron acceptor dueto its large electron affinity, whose value is much largerthan that of CuPc in the isolated state [27, 28], and hasno spin states. This means that PTCDA would be a po-tential candidate to achieve a higher Tc of (√7×√3)-In.In this paper, we report the electronic structure, con-ductivity, and atomic structure of PTCDA adsorbed(√7×√3)-In, which were established by high-resolutionangle-resolved photoelectron spectroscopy (ARPES),transport measurements, and scanning tunneling mi-2croscopy (STM), respectively. Although a large amountof hole doping and an increase in Tc were expected fromthe electronic properties of PTCDA, only a small amountof hole doping and a suppression in Tc were observed.Moreover, a sharp increase in the normal-state resistancethat eventually undergoes an insulating transition with-out the destruction of the metallic character of (√7×√3)-In was observed. Taking the disordering of adsorbedPTCDA into account, we conclude that the suppressionin superconductivity and the insulating transition are dueto the random potentials induced by disordered-arrangedPTCDA molecules, which possibly scatter the conduc-tion electrons and cause carrier localization in the 2DIn layer. Our findings suggest that molecular films withhigh crystalline quality are indispensable to improve thesuperconductivity in 2D systems consisting of an ALSCand organic molecules.II. EXPERIMENTAL DETAILSHigh-resolution ARPES measurements were conductedat the Bloch beamline of MAX IV, Sweden, which isequipped with a DA30-L analyzer (Scienta Omicron),and the beamline 13 of Saga Light Source, Japan,equipped with a A-1 analyzer (MB Scientific AB). Thetransport experiment was performed using the four-pointprobe method in a home-built UHV apparatus [2, 19] andthe STM observation was done using a low-temperatureSTM (UNISOKU Co. Ltd.). An n-type Si(111) sub-strate (1-5 Ωcm) was used for PES and LT-STM, anda non-doped one (>1000 Ωcm) for transport measure-ments. (√7 ×√3)-In was prepared by depositing ap-proximately three monolayers of In on a clean Si(111)surface at 300 K, followed by a 600 K annealing for acouple of minutes. The sample quality was confirmedby the observation of sharp spots in low energy electrondiffraction (LEED). PTCDA (purity > 98%) was well de-gassed in UHV before deposition. The PTCDA thicknesswas monitored with a quartz balance and calibrated byPES and STM. The sample was maintained below 20 Kduring all PES measurements and at 4.7 or 80 K in STM.All sample preparation and experiments were performedunder ultrahigh vacuum (UHV) conditions.III. RESULTS AND DISCUSSIONThe PTCDA coverage-dependent valence band spec-tra and change in work function obtained with a photonenergy (hν) of 40 eV are shown in Figs. 1(a) and (b),respectively. In the valence band spectra, the intensity ofthe prominent peak observed at a binding energy (EB) ofapproximately 1 eV on the pristine (√7 ×√3)-In (indi-cated by a red dashed line in Fig. 1(a)) decreases, and theintensities of the four peaks (indicated by black dashedlines) develop as the molecular coverage increases. Tak-ing into account the MO spectra reported in previousFIG. 1. Coverage-dependent (a) valence band spectra and (b)work function of PTCDA adsorbed (√7×√3)-In. The spec-trum indicated by a gray solid line in (a) shows the calculatedDOS of and isolated PTCDA molecule.studies [29–31], we attribute the peak located at approx-imately EB = 2.1 eV to the highest-occupied molecu-lar orbital (HOMO) of PTCDA. Furthermore, by con-sidering the good agreement between the relative EB ofthe observed four peaks and those of the theoreticallyobtained molecular orbitals (MOs) for an isolated PD-CDA (the gray spectrum), we conclude the origin of allthe four peaks to be the MOs of PTCDA. (The calcula-tion was performed using the GAMESS program [32, 33]with the B3LYP method and 6-311G** basis set, and theMO energies were shifted so that the HOMO peak coin-cides the experimental spectra.) The negligible shift inEB of the MOs and the invisible molecule-derived fea-tures within the energy gap region of PTCDA, unlikethe case of strongly interacting systems [34], suggest therather weak interaction between PTCDA and (√7×√3)-In. This weak interaction is supported by the satura-tion in work function at ca. 1 ML where the value isin accordance with that of the work function of a thickPTCDA film [35]. The negligible change in work functionat coverages higher than 1 ML indicates that the surfacedipole of PTCDA adsorbed on (√7×√3)-In and that ofa thick PTCDA film are equivalent, and therefore thatthe charge transfer between PTCDA and (√7 ×√3)-Inis negligible and also that PTCDA is hardly distorted on(√7×√3)-In.We have measured the Fermi surfaces (FSs) at differentPTCDA coverages with hν = 40 eV. Figure 2(a) showsthe FS of the pristine (√7×√3)-In obtained from a sum-mation of the photoelectron intensity within a 20 meVenergy window from the Fermi level. This FS, which con-sists of two types of FSs (a circular Fermi surface that ismainly formed by electrons located at the outermost Inlayer and a butterfly-shaped Fermi surface mainly formedby electrons located at the In layer connected to Si), isin good agreement with the experimental and calculatedones reported previously [10, 13, 17, 18, 21]. The circu-3FIG. 2. (a) FS of the pristine (√7×√3)-In. The green solidlines show the√7×√3 Brillouin zone, and the red and bluesolid lines enclose a part of the circular and butterfly-shapedFSs, respectively. (b) FSs of PTCDA adsorbed (√7×√3)-Inat different coverages. (c) PTCDA coverage-dependent MDCsobtained from a summation within a kx range of 0 ± 0.015Å−1. The open circles are the experimental data and the solidlines overlapping them show the Lorentzian fittings. The peakposition in each MDC is indicated by a triangle.lar and butterfly-shaped FSs are indicated as C-FS andB-FS, respectively, in the figure. The FSs obtained afterPTCDA deposition [Fig. 2(b)] show that the increase ofPTCDA coverage leads to an intensity drop of the FSbut not to a significant modulation in its shape. This re-sult indicates that the interaction between PTCDA and(√7×√3)-In does not involve chemical bonding and thushardly affects the electronic structure of (√7 ×√3)-In.In order to obtain more detailed information about thecharge transfer at the interface, we have analyzed the ra-dius of the circular Fermi surface. Figure 2(c) displaysthe momentum distribution curves (MDCs) at kx = 0and ky > 0. Hole doping from PTCDA into the In layerwould shrink the FSs, and thus increase the separationbetween the two circular FSs at kx = 0. As shown in Fig.2(c), the position of the peak in MDC, which is obtainedby fitting using a Lorentzian function, shows a small shifttoward larger ky as the PTCDA coverage increases. Thismeans that the adsorbed PTCDA molecules slightly dopeholes into the In layer. Table I summarizes the obtainedradius of the circular FS at different PTCDA coverage,and the number of holes transferred to an In atom fromPTCDA estimated from the experimental result. (Thenumber of transferred holes was obtained by compar-ing the area of the circular Fermi surface and the Bril-louin zone formed by the In atoms as in Ref. [21].) TheTABLE I. Radius of the circular FS at different PTCDA cov-erages and the number of holes transferred per In atom esti-mated from the radius.PTCDA coverage Radius ∆ holes(ML) (Å−1) (/In atom)0.0 1.4309 ± 0.0002 -0.32 1.4288 ± 0.0002 0.0052 ± 0.00050.55 1.4279 ± 0.0002 0.0073 ± 0.00051.1 1.4262 ± 0.0005 0.0115 ± 0.0012.0 1.4233 ± 0.002 0.0184 ± 0.005amount of hole doping continuously increases even abovea PTCDA coverage of 1.0 ML. This results from the factthat PTCDA does not grow layer-by-layer on (√7×√3)-In and regions uncovered by PTCDA remain at above 1.0ML as will be discussed below. Despite its large electronaffinity, which is larger than that of CuPc, the PTCDAadsorption causes less hole doping into (√7×√3)-In thanCuPc. This result indicates that discussing the chargetransfer at the interface between organic molecules andmetal substrates based on the energy level alignment onlyis not sufficient.In order to obtain further information about the in-teraction between PTCDA and (√7 ×√3)-In, we havemeasured the coverage-dependent C 1s and O 1s core-levels. In Fig. 3, we show the C 1s and O 1s spectraobtained with hν = 650 eV at different PTCDA cover-ages. Both the C 1s and the O 1s spectra show onlyincrease in intensity and no remarkable change in thespectral shape as the PTCDA coverage increases. TheEB of two O 1s components and their relative intensityagree well with those of thick PTCDA films [36–38]. Onthe other hand, although the relative intensity of the twoC 1s components shows agreement with those of thickfilms [36–38], there is a difference in the separation intheir EB . That is, the separation of the two componentsin Fig. 3(a) is narrower than those reported in the litera-ture for thick PTCDA films. By considering the origin ofthe small component at higher EB , we conclude that thedifference in EB separation results from the small chargetransfer from In to the C atoms in the carboxylic group,i.e., C atoms bonded to O atoms, like the case of PTCDAadsorbed on Sn covered Si(111) surface [38].Since the small hole doping and the observation ofcircular FS even after PTCDA adsorption indicate thepossibility of change in Tc of (√7 ×√3)-In, we haveperformed coverage-dependent transport measurements.The superconducting transition of the pristine (√7×√3)-In was confirmed by a steep decrease that reaches 0Ω/□ in sheet resistance at approximately 3.0 K, whichis consistent with the previous reports [1–3, 19, 20]. Asshown in Fig. 4(a), the Tc is suppressed when increas-ing the PTCDA coverage. This is in contrast to thechange in Tc expected from hole doping as in the caseof CuPc adsorption [13]. Furthermore, the normal-statesheet resistance at 3.5 K shows a sizable increase af-4FIG. 3. (a) C 1s and (b) O 1s core-level spectra from PTCDAon (√7×√3)-In at different coverages. The upper left inset ineach figure displays the molecular structure of PTCDA withchemically inequivalent C and O atoms.ter PTCDA adsorption, which was not observed in caseof several metal-phthalocyanine adsorptions [13, 22, 23].Moreover, the disappearance of superconductivity andthe increase in sheet resistance when decreasing temper-ature at 1.9 ML reveals the semiconducting or insulatingbehavior of (√7 ×√3)-In at this coverage. Since theARPES measurements confirm that the electronic struc-ture is only modified by a small amount of hole dopingand the metallic bands still remain up to 2.0 ML, thistransition in conductivity would not be a phase transi-tion accompanied with a gap opening at the Fermi level.It should be noted that the sheet resistance increases upto more than 50 kΩ/□, which drastically exceeds theuniversal critical sheet resistance h/4e2 (= 6.45 kΩ) ofthe superconductor-insulator transition in the 2D limit[39–41]. Since PTCDA has no spin states, the magneticeffects cannot be the origin of the suppression in super-conductivity. In superconducting thin films, the Tc hasbeen reported to decrease and the sheet resistance to in-creases by the presense of disorder [39, 41]. This meansone possible origin of the results shown in Fig. 4 wouldbe the disordered potential formed on (√7 ×√3)-In bythe adsorbed PTCDA.Figure 5(a) shows a 200 × 200 nm2 STM image of(√7 ×√3)-In covered with 1.0 ML of PTCDA. The co-existence of the first and second PTCDA layers and thepresence of uncovered (√7 ×√3)-In regions indicatesthat PTCDA deposition follows an island growth modeand not a layer-by-layer one as in case of CuPc. Thissuggestion is consistent with the PES results discussedabove, where the amount of hole doping does not sat-urate and/or show a maximum at a PTCDA coverageof 1 ML. [The PTCDA molecules of the first layers ap-pear as dark as the (√7 ×√3)-In region in the STMimage, but since no significant changes are observed inthe electronic structure of (√7 ×√3)-In upon PTCDAadsorption this is due to an effect of the electronic statesinstead of the height of atoms.] The STM image withmolecular resolution of the first PTCDA layer reveals aFIG. 4. Temperature-dependent sheet resistance of thePTCDA adsorbed (√7×√3)-In at PTCDA coverages below1.0 ML (a) and at 1.9 ML (b).mixture of ordered and disordered regions [Fig. 5(b)].This means that PTCDA molecules do not form a per-fectly ordered molecular layer on (√7×√3)-In like metal-phthalocyanine molecules [13, 22, 23]. In both regions,PTCDA molecules are flat lying like PTCDA adsorbedon various substrates [26]. As shown in the inset of Fig.5(b), PTCDA molecules form a herringbone structurein the ordered region, which is commonly seen in thefirst few layers on metal substrates [31, 42, 43] or metal-terminated semiconductor surfaces [44–46]. On the otherhand, PTCDA molecules are randomly arranged in thedisordered region. The disruption in ordering might bedue to the defects of the substrate surface that act astrapping sites and cause different interactions to the sur-rounding molecules. In the herringbone structure, theelectropositive H atoms of PTCDA always face the elec-tronegative O atoms of the neighboring molecule andcancel out the charge distribution within the molecule.In the disordered region, however, the charge distribu-tion cannot be cancelled and thus would induce a ran-dom electrostatic potential to the underlying In layer.This would lead the PTCDA molecules to play a roleas electron-scatterers that cause carrier localization inthe 2D In layer [47, 48]. We therefore conclude thatthe observed suppression in superconductivity, increasein normal-state sheet resistance, and insulating transi-tion in conductivity would result from the random po-tential induced by disordered PTCDA molecules. Thisresult provides a new insight toward a rational design forenhancing the superconducting properties of ALSCs.IV. CONCLUSIONSIn conclusion, we have measured the coverage-dependent electronic structure, resistivity, and molecu-lar arrangement of PTCDA adsorbed (√7 ×√3)-In, tounderstand the effect of organic molecule adsorption onan ALSC. The FSs at different PTCDA coverages indi-cate that the adsorption causes a slight hole doping into5FIG. 5. STM image of the PTCDA adsorbed (√7 ×√3)-Insurface at a PTCDA coverage of 1.0 ML. The sample biasand tunnel current are -2.0 V and 10 pA. (a) 200× 200 nm2STM image. The three white arrows indicate the regions ofthe bare In surface and the first and second layers of PTCDA.(b) Closer view of the first PTCDA layer. Ordered and dis-ordered regions are shown below and above the white dashedline. The inset shows enlarged view of the 5 × 5 nm2 areaoutlined by the white solid square, where PTCDA moleculesform a herringbone structure. The yellow rectangle in theinset indicates the unit cell.(√7×√3)-In without making appreciable alteration to itselectronic structure. Regarding the transport measure-ments, a suppression in Tc with an increase in the normal-state sheet resistance, followed by an insulating transitionhas been observed when increasing the PTCDA coverage,though the hole doping and the absence of spin states inmolecule were expected to increase the Tc. The unex-pected results in conductivity come from the presence ofdisordering of PTCDA molecule adsorbed on (√7×√3)-In, which induces random potentials and thus lead to thelocalization of conduction electrons in the In layer. Thisshows that the 2D superconductivity is destroyed dueto potential disordering introduced by molecular adsorp-tion, and therefore indicates the importance of the highcrystallinity of 2D molecular film in 2D systems consist-ing of an ALSC and organic molecules.ACKNOWLEDGMENTS[1] T. Zhang, P. Cheng, W.-J. Li, Y.-J. Sun, G. Wang, X.-G. Zhu, K. He, L. Wang, X. Ma, X. Chen, et al., NaturePhysics 6, 104 (2010).[2] T. Uchihashi, P. Mishra, M. Aono, and T. Nakayama,Physical review letters 107, 207001 (2011).[3] M. Yamada, T. Hirahara, and S. Hasegawa, Physical re-view letters 110, 237001 (2013).[4] A. V. Matetskiy, S. Ichinokura, L. V. Bondarenko, A. Y.Tupchaya, D. V. Gruznev, A. V. Zotov, A. A. Saranin,R. Hobara, A. Takayama, and S. Hasegawa, Phys. Rev.Lett. 115, 147003 (2015).[5] C. Brun, T. Cren, and D. Roditchev, SuperconductorScience and Technology 30, 013003 (2016).[6] J. Bardeen, L. N. Cooper, and J. R. Schrieffer, Phys. Rev.108, 1175 (1957).[7] T. Sekihara, R. Masutomi, and T. Okamoto, PhysicalReview Letters 111, 057005 (2013).[8] C. Brun, T. Cren, V. Cherkez, F. Debontridder, S. Pons,D. Fokin, M. Tringides, S. Bozhko, L. Ioffe, B. Altshuler,et al., Nature Physics 10, 444 (2014).[9] T. Nakamura, H. Kim, S. Ichinokura, A. Takayama,A. Zotov, A. Saranin, Y. Hasegawa, and S. Hasegawa,Physical Review B 98, 134505 (2018).[10] S. Yoshizawa, T. Kobayashi, Y. Nakata, K. Yaji,K. Yokota, F. Komori, S. Shin, K. Sakamoto, andT. Uchihashi, Nature communications 12, 1 (2021).[11] M. Smidman, M. Salamon, H. Yuan, and D. Agterberg,Reports on Progress in Physics 80, 036501 (2017).[12] G. C. Ménard, S. Guissart, C. Brun, R. T. Leriche,M. Trif, F. Debontridder, D. Demaille, D. Roditchev,P. Simon, and T. Cren, Nature communications 8, 1(2017).[13] S. Yoshizawa, E. Minamitani, S. Vijayaraghavan,P. Mishra, Y. Takagi, T. Yokoyama, H. Oba, J. Nitta,K. Sakamoto, S. Watanabe, T. Nakayama, and T. Uchi-hashi, Nano Lett. 17, 2287 (2017).[14] J. W. Park and M. H. Kang, Physical Review Letters109, 166102 (2012).[15] K. Uchida and A. Oshiyama, Phys. Rev. B 87, 165433(2013).[16] K. Uchida and A. Oshiyama, Physical Review B 87,165433 (2013).[17] E. Rotenberg, H. Koh, K. Rossnagel, H. Yeom, J. Schäfer,B. Krenzer, M. Rocha, and S. Kevan, Phys. Rev. Lett.91, 246404 (2003).[18] T. Kobayashi, Y. Nakata, K. Yaji, T. Shishidou,D. Agterberg, S. Yoshizawa, F. Komori, S. Shin,M. Weinert, T. Uchihashi, et al., Physical Review Let-ters 125, 176401 (2020).[19] T. Uchihashi, P. Mishra, and T. Nakayama, NanoscaleResearch Letters 8, 1 (2013).[20] S. Yoshizawa, H. Kim, T. Kawakami, Y. Nagai,T. Nakayama, X. Hu, Y. Hasegawa, and T. Uchihashi,Physical review letters 113, 247004 (2014).[21] R. Sagehashi, T. Kobayashi, T. Uchihashi, andK. Sakamoto, Surf. Sci. 705, 121777 (2021).[22] N. Sumi, Y. Yamada, M. Sasaki, R. Arafune, N. Takagi,S. Yoshizawa, and T. Uchihashi, J. Phys. Chem. C 123,8951 (2019).[23] T. Uchihashi, S. Yoshizawa, E. Minamitani, S. Watan-abe, Y. Takagi, and T. Yokoyama, Molecular SystemsDesign & Engineering 4, 511 (2019).[24] K. Yokota, S. Inagaki, W. Qian, R. Nemoto,S. Yoshizawa, E. Minamitani, K. Sakamoto, and T. Uchi-hashi, Journal of the Physical Society of Japan 91,123704 (2022).[25] E. Umbach, C. Seidel, J. Taborski, R. Li, andA. Soukopp, physica status solidi (b) 192, 389 (1995).[26] N. Nicoara, J. M. Gómez-Rodŕıguez, and J. Méndez,physica status solidi (b) 256, 1800333 (2019).6[27] P. C. Rusu, G. Giovannetti, C. Weijtens, R. Coehoorn,and G. Brocks, Physical Review B 81, 125403 (2010).[28] W. Wu, L. Rochford, S. Felton, Z. Wu, J. Yang, S. Heutz,G. Aeppli, T. Jones, N. Harrison, and A. Fisher, Journalof Applied Physics 113, 013914 (2013).[29] D. R. Zahn, G. N. Gavrila, and M. Gorgoi, ChemicalPhysics 325, 99 (2006).[30] J. Gustafsson, E. Moons, S. Widstrand, M. Gurnett, andL. Johansson, Surface science 572, 32 (2004).[31] Y. Zou, L. Kilian, A. Schöll, T. Schmidt, R. Fink, andE. Umbach, Surface Science 600, 1240 (2006).[32] K. Ishimura, L. McMurchie, E. Davidson, and J. Rys, J.Comput. Chem 14, 1347 (1993).[33] G. M. Barca, C. Bertoni, L. Carrington, D. Datta,N. De Silva, J. E. Deustua, D. G. Fedorov, J. R. Gour,A. O. Gunina, E. Guidez, et al., The Journal of chemicalphysics 152, 154102 (2020).[34] M.-C. Lu, R.-B. Wang, A. Yang, and S. Duhm, Journalof Physics: Condensed Matter 28, 094005 (2016).[35] M. S. Khoshkhoo, H. Peisert, T. Chassé, and M. Scheele,Organic Electronics 49, 249 (2017).[36] A. Schöll, Y. Zou, M. Jung, T. Schmidt, R. Fink, andE. Umbach, The Journal of chemical physics 121, 10260(2004).[37] J. Gustafsson, H. Zhang, E. Moons, and L. Johansson,Physical Review B 75, 155413 (2007).[38] H. Zhang and L. Johansson, Chemical Physics 439, 71(2014).[39] D. B. Haviland, Y. Liu, and A. M. Goldman, Phys. Rev.Lett. 62, 2180 (1989).[40] T. Pang, Physical review letters 62, 2176 (1989).[41] A. M. Goldman and N. Markovic, Physics Today 51, 39(1998).[42] T. Schmitz-Hübsch, T. Fritz, F. Sellam, R. Staub, andK. Leo, Physical Review B 55, 7972 (1997).[43] T. Wagner, A. Bannani, C. Bobisch, H. Karacuban, andR. Möller, Journal of Physics: Condensed Matter 19,056009 (2007).[44] J. Gustafsson, H. Zhang, and L. Johansson, Physical Re-view B 75, 155414 (2007).[45] D. Shin, Z. Wei, H. Shim, and G. Lee, Applied SurfaceScience 372, 87 (2016).[46] N. Nicoara, J. Méndez, and J. Gómez-Rodŕıguez, Nan-otechnology 27, 365706 (2016).[47] P. W. Anderson, Phys. Rev. 109, 1492 (1958).[48] E. Abrahams, P. W. Anderson, D. C. Licciardello, andT. V. Ramakrishnan, Phys. Rev. Lett. 42, 673 (1979).