DOI: 10.1002/adma.200502115

Nanostructured Organic Material: From Molecular Chains toOrganic Nanodots**

By Javier Méndez,* Renaud Caillard, Gonzalo Otero, Nicoleta Nicoara, and José A. Martín-Gago

We present a new strategy for nanostructuring organic mol-ecules in order to create either molecular chains or organicnanodots. The method is based on the use of patterned sur-faces as templates for nucleation of the organic molecules. Asthe patterned substrate we use iron-island arrays grown onAu(111). These metallic arrays are formed spontaneously onthe gold reconstructed surface via strain-relief nucleation. Weshow that deposited 3,4,9,10-perylene-tetracarboxylic-dianhy-dride (PTCDA) molecules anchor at the iron islands formingnew structures that are not observed for PTCDA adsorptionon clean gold. Namely, PTCDA forms either chains of mole-cules connecting contiguous iron clusters, or aggregates sur-rounding the metallic core. The resulting arrangement of or-ganic dots exhibits a different density of states than the two-dimensional self-assembled molecular monolayer. Scanningtunneling spectroscopy (STS) studies indicate that the mainmodifications are the absence of the gold surface state andthe appearance of an iron state induced in the PTCDA high-est occupied molecular orbital–lowest unoccupied molecularorbital (HOMO–LUMO) gap.

Organic molecules are one of the most promising candi-dates to substitute inorganic materials in nanoscale devices.Up to now, these organic materials have found several opto-electronic applications such as organic light-emitting de-vices,[1,2] organic transistors,[3,4] organic lasers,[5,6] organicrectifiers,[7] and organic memories.[8] Most of these devicestake advantage of the confinement properties due to thenanostructuring of the organic molecules in 2D thin films or1D nanowires. Naturally, organic 0D nanostructures are envi-sioned as systems with potential applications. Among thestrategies for nanostructuring organic materials, landing ofmolecules or tip-induced manipulation have been successfullyused for single-molecule set-ups.[9–12] More extended assem-blies can be achieved by self-organization of molecules, bychoosing the appropriate chemistry of the molecules,[13,14] or

by using metallic atoms as linkers to form molecular net-works.[15,16]

Small amounts of iron—as well as cobalt, nickel, and palla-dium—deposited on the strain relief (22 ×

���

3�

) gold recon-struction lead to the formation of islands at the edges of theso called herringbone reconstruction.[17–20] Figure 1 shows anarray of iron islands nucleated on Au(111) after deposition of0.08 ML (ML: monolayer) at room temperature. Iron atomshave nucleated forming clusters at the edges of the goldreconstruction. These iron atoms are arranged in a close-packed face-centered cubic phase,[21] leading to the truncatedtriangular shapes observed in Figure 1b. The average size ofthe iron islands, determined from the distribution shown inFigure 1c, is close to 4 nm × 4 nm for this particular coverage.Similar metallic arrays have been obtained on strain-reliefpatterns,[22] where dislocations repel diffusing adatoms. Thesepatterns are obtained spontaneously when one or two mono-layers of a material are deposited on a substrate with a differ-ent lattice constant.[23] In our case, it is the Au(111) recon-struction itself that provides the dislocation network, theelbows of the reconstruction acting as nucleation centers fordiffusing adatoms.[24]

On the other hand, it is also known that PTCDA moleculesdeposited on clean Au(111) form highly ordered 2D films.[25]

The weak molecule–substrate interaction[26] allows the organ-ic molecules to diffuse at a high rate at room temperaturewithout being altered by (or altering) the gold reconstruction.Molecules establish intermolecular interactions with theirneighbors, linking the O atoms of one molecule with the Hatoms of the next one, to form the characteristic herringboneand square phases.[27] In both cases, PTCDA molecules arearranged perpendicular to each other.

Deposition of PTCDA organic molecules on a Au(111) sur-face tailored by iron cluster 2D arrays results in a completelydifferent growth mode. Figure 2a shows several organic aggre-gates pinned to the iron islands. We observe several molecularstructural arrangements that are never formed for PTCDAdeposition on clean noble-metal substrates. Such small aggre-gates are not stable at room temperature without the presenceof nucleation centers (as step edges or contaminants). More-over, for this low PTCDA coverage (just 0.04 ML) thePTCDA characteristic structures are not observed. First, letus consider the distribution of these new aggregates. Most ofthe iron clusters in Figure 2a are unaltered (without nucleatedmolecules) and only a few of them have one or two moleculesattached. PTCDA nucleation is preferred at iron clusterswhere other molecules are already nucleated. This irregular

CO

MM

UN

ICATI

ON

S

2048 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2048–2052

–[*] Dr. J. Méndez, Dr. R. Caillard, G. Otero, N. Nicoara,

Dr. J. A. Martín-GagoInstituto de Ciencia de Materiales de Madrid, CSICCampus de Cantoblanco28049 Madrid (Spain)E-mail: jmendez@icmm.csic.es

[**] The authors thank Dr. J. Cerdá, Dr. F. Zamora, and Prof. A. Baró forfruitful discussion and critical reading. Financial support from theComunidad Autónoma de Madrid (CAM) under grant no. GR/MAT/0699/2004 and the Spanish Ministerio de Educación y Ciencia,grant no. MAT-2005-3866 is gratefully acknowledged.

distribution suggests that the iron-mediated PTCDA–PTCDAinteraction is energetically more favorable than the interac-tion between PTCDA and the iron cluster itself.

Second, let us observe the structure of the aggregates. Forinstance, Figure 2b corresponds to chains of PTCDA mole-cules pinned to iron clusters. The molecules are placed withthe longer axis parallel to the chain axis; therefore, the oxygenterminations of two consecutive molecules face each other.We suggest that this unusual structure is stabilized by ironplaced between the molecules in a similar way to the supra-molecular noncovalent interactions described by Lehn,[28] orthe reported metallo-organic networks steadied by ironatoms.[15,16] In our case, the anhydride groups of the PTCDAare already saturated, and therefore an oxidation reactionwith iron atoms is not possible.[29] Ab-initio calculations cur-rently in progress[30] point out that the linkage between con-secutive PTCDA molecules is constituted of two iron atomsinteracting with the carboxylic groups of the PTCDA. Themodel for this linkage is shown in Figure 2c, where twoPTCDA molecules are connected by two iron atoms.

We have also found other PTCDA structures induced byiron, such as the ladderlike structures observed in Figure 2d.For these kinds of aggregates, perpendicular PTCDA mole-cules connect two parallel chains, and three PTCDA mole-cules are attached by the same iron linkage.

Annealing the substrate up to 370 K after deposition favorsmolecular-chain formation, leading to longer PTCDA chains(molecular wires) connecting several iron islands, mainlyalong the [112] crystallographic direction. Further annealingto higher temperatures (420 K) breaks up the PTCDA–ironlink, allowing iron atoms to diffuse to the step edges and thePTCDA molecules to form islands with the “herringbone”phase characteristic of PTCDA on Au(111). The interactionbetween PTCDA and the iron is therefore reversible upon an-nealing. The lack of OH groups in the PTCDA molecules isalso in concordance with the observed reversibility for theiron–PTCDA linkage in the present metallo-organic aggre-gates.[29]

In analogy to standard crystal growth modes,[31] one may ex-pect the organic aggregates either to grow in concentric circlesaround a nucleation point for high substrate temperatures, orto form dendrite chains for low temperatures. However, weexperimentally observe the opposite behavior: formation ofchains at higher temperature and rounded structures at lowertemperature. Assuming that the diffusion of PTCDA mole-

CO

MM

UN

ICATIO

NS

Adv. Mater. 2006, 18, 2048–2052 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.advmat.de 2049

2 4 6 8 10 12 14 160

20

40

60

80

100

120

Nu

mb

er

of

Isla

nd

s

Area [nm2]

a b c

Figure 1. Iron island array on Au(111). a) Scanning tunneling microscopy (STM) image (145 nm × 145 nm) of Au(111) substrate after evaporation of0.08 ML of iron. The iron nucleates forming islands at the gold reconstruction edges [17]. b) Derivative image (24 nm × 24 nm) showing iron islandsat the edges of the reconstruction. c) Island distribution plot of the image in (a) with an average size of the islands around 4 nm × 4 nm.

a

b

d

c

Figure 2. Pinning of PTCDA molecules at the iron islands. a) STM image(120 nm × 120 nm) after evaporation of 0.04 ML of PTCDA molecules onan iron array (iron coverage, hFe ∼ 0.08 ML). The organic molecules formaggregates pinned by the iron. The observed arrangements, such as mo-lecular chains or ladderlike structures, are never observed for PTCDA onclean gold. Iron atoms in between the organic molecules stabilize thesearrangements. b) Chains of PTCDA molecules pinned at two iron clus-ters (brighter features) (18 nm × 16 nm). c) Model for the chain struc-ture with two iron atoms as linkage between two consecutive PTCDAmolecules. d) Ladderlike structure. (13 nm × 11 nm).

cules is high enough in the temperature range accessible inthese experiments, the promoted formation of molecularwires for higher temperature suggests that the crucial param-eter is the amount of free iron adatoms diffusing on the goldsubstrate. For higher temperature, the number of diffusingiron adatoms increases, being enough to couple to the car-boxylic groups and to form chains where iron atoms link twoconsecutive molecules. On the contrary, cooling the substrateafter iron deposition, or reducing the iron coverage, meansthat most of the iron atoms are nucleated at the reconstruc-tion corners and leads to the formation of organic aggregatesexhibiting a more circular shape (see Fig. 3), which we havenamed organic nanodots (ODs).

Therefore, by tuning the experimental growth conditions(temperature and iron coverage) we can switch from the for-mation of molecular chains connecting consecutive iron clus-ters to organic (unconnected) aggregates sited at the gold re-construction corners.

In Figure 3 we present an STM image with several organicnanodots. These are grown around small iron clusters of a fewtens of atoms (the iron coverage was lower than 0.01 ML). Insome organic aggregates a bright feature inside, related to theiron-cluster seed, is observed. In others, the presence of ironis supported by the different intermolecular assembling of thePTCDA molecules. As is shown below, a different electronicbehavior is also observed. In the image of Figure 3 the goldreconstruction lines are clearly resolved. The organic dots arepinned to the reconstruction elbows, as these are nucleationcenters for the iron. The average diameter of these dots isabout 4–6 nm and they follow the network of the iron clusters,presenting, therefore, a high degree of ordering. Nanostruc-tured dot formation occurs in a narrow range of experimentalparameters. Indeed, it is a compromise between the amountof free diffusing iron that favors molecular wires versus organ-ic nanodots, and a minimum amount of deposited iron thatforms an extended iron network at every reconstruction

elbow. Again, increasing the temperature leads to the separa-tion of the phases: ordered PTCDA phases and iron nuclea-tion at the steps.

In order to explore the electronic properties of the organicnanodots, we have performed STS measurements for differentsituations. Previously, we have recorded STS curves on cleanAu(111) and on the PTCDA monolayer grown with a herring-bone structure.[26] These plots are used here as a reference,checking that the tungsten tip is “in good condition”, that is,free of contaminants or absorbed molecules.

Spectroscopic conductance plots on the Fe–Au(111) systemshow the different electronic behavior between the iron clus-ters and the gold substrate. In Figure 4, the dI/dV (I: current;

V: voltage) plot obtained for the gold substrate (dashed line)shows the gold surface state feature around –0.4 eV. The spec-trum obtained with the same tip in the center of the iron clus-ters (continuous line) does not show any feature at this posi-tion, but a bump close to –0.6 eV. We consider this lastfeature as being indicative of the iron.[32]

Spectroscopy on the metallo-organic aggregates describedin this paper shows clear contrast with respect to the gold sub-strate. The spectra on the organic nanodots display featuresrelating to the iron peak together with the characteristicPTCDA peaks. To emphasize the difference between PTCDAmolecules in the organic nanodots and molecules in a com-plete layer (with and without the effect of the iron) we haveincreased the PTCDA coverage close to the monolayer. Thetopographic STM image of Figure 5a corresponds to the Fe–Au(111) substrate covered with 0.9 ML of PTCDA. Three dif-ferent regions are distinguished: A, PTCDA areas on the goldterrace; B, PTCDA molecules close to the iron clusters; andC, the iron cluster. Far from the nanodot, the organic mole-cules form the well-known herringbone arrangement of thePTCDA (A), covering the surface without disturbing thestrain-relief gold reconstruction. Bright features at the recon-struction elbows can be seen, corresponding to the iron core

CO

MM

UN

ICATI

ON

S

2050 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2048–2052

Figure 3. Organic nanodots. STM image (40 nm × 40 nm) where PTCDAmolecules have grown around small iron islands nucleated at the goldsubstrate reconstruction’s elbows. The reconstruction lines of the goldsubstrate are clearly observed in the background. The organic nanodotsare 4–6 nm in diameter. The arrows point to the iron seeds in two nano-dots. The coverages are hFe < 0.01 ML and hPTCDA ∼ 0.15 ML.

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8

0.0

0.5

1.0

1.5

2.0

2.5

3.0Fe

Au

dI/d

V [

nA

/V]

Voltage [V]

Au(111)

Fe on Au(111)

Figure 4. Tunneling spectroscopy on the iron clusters. STS spectroscopicplot (dI/dV) obtained over an iron island (continuous line) in compari-son with a plot obtained on the gold terrace (dashed line). On the gold,the characteristic surface state peak is obtained, while on the iron clustera bump close to –0.6 V is the only observed feature.

(C) and surrounded by PTCDA molecules (B) in a less-or-dered disposition than the molecules in (A).

Figure 6 shows STS curves obtained by averaging I–V plotsrecorded on different areas. These curves are presented as thederivative dI/dV for voltages between –2 and +2 V. ForPTCDA areas with the well-ordered herringbone structure, asthe region marked A, we have obtained curves similar tothose previously observed for the PTCDA monolayer ongold.[26] Curve A is characterized by the presence of a peak at–0.4 eV, corresponding to the Shockley surface state ofAu(111), visible through the PTCDA layer.[26] Curve C, re-corded on the iron core, exhibits an electronic feature at–0.7 eV, which resembles the one observed on iron clusters,and therefore can be assigned to an iron electronic state.[32]

The conductance curve for PTCDA areas surrounding theiron core, plot B, shows a peak at –0.7 eV corresponding tothe iron, while the Au(111) surface state at –0.4 eV is not ob-served. The PTCDA molecules of the organic nanodot, modi-fied by the iron, have a different electronic structure than themolecules in a single monolayer. The main difference con-cerns the transparency of the molecules to the gold surfacestate. In contrast to a charge-transfer process,[33] where a shiftin the STS plot would be expected, we assume here that there

is an iron-induced state in the surrounding PTCDA molecules.Furthermore, as a consequence of the induced state, theHOMO–LUMO gap value of the PTCDA opens for the mole-cules close to the iron core. The position of these peaks variesslightly with the tip and for different organic dots, but wereproducibly observe that the feature obtained over the ironpersists in the surrounding PTCDA molecules and that thesemolecules present a different spectroscopic behavior. We cannot exclude the possibility that the observed iron feature atthe organic nanodots originates from iron atoms situated be-tween the PTCDA molecules in the dot, but the lack of thegold feature is a clear evidence of the change in the electroniccharacter in the organic nanodot.

Conductivity (dI/dV) maps, extracted from current imagingtunneling spectroscopy (CITS),[34] are presented in Fig-ure 5b–d. These images are measured simultaneously with thetopography in Figure 5a and correspond to bias voltageslabeled as F, G, and H in Figure 6, respectively. At F, thenanodot-related electronic state is the main contribution tothe total current (Fig. 5b). At H, the main peak is the goldsurface state; therefore, the molecules exhibiting this state ap-pear brighter in the spectroscopic image, while the nanodotappears dark (Fig. 5d). In between, at G, molecules at theboundaries of the organic nanodot are the brightest as theyreceive contribution from both peaks (Fig. 5c). Notice thatthe perimeter of the dark area (Fig. 5d) is smoother than theorganic dot itself (Fig. 5a), that is, the electronic effect is alsoextended to nearby molecules with the herringbone phase.The extension of this dark area strongly depends on the volt-age. For low negative voltages the area corresponds to the sizeof the dot. When the voltage decreases towards negative val-ues the effect that the iron cluster in the PTCDA moleculehas decayed, which results in a shrinking of the dark area.

The iron cluster induces an electronic state inside the gap ofthe surrounding molecules (similar to a metal-induced gap

CO

MM

UN

ICATIO

NS

Adv. Mater. 2006, 18, 2048–2052 © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.advmat.de 2051

a b

c d

Figure 5. Spectroscopic images of the organic nanodots. a) TopographicSTM image (40 nm × 40 nm) for 0.9 ML PTCDA covering the Fe–Au(111) surface. The marked areas correspond to: A, PTCDA moleculeson the gold terrace; B, PTCDA molecules close to the iron cluster; and C,the iron clusters. b–d) Conductivity maps (dI/dV) extracted from CITScorresponding to bias voltages labeled F, G, and H, in Figure 6. b) At F,the organic dot appears brighter, corresponding to the iron state. c) AtG, the molecules at the boundaries appear brighter due to the contribu-tion from both peaks. d) At H, the PTCDA molecules away from the or-ganic dot appear brighter, corresponding to the gold surface state ob-served through the molecules [26]. Notice that the darker area issmoother than the structural shape of the dot.

-2 -1 0 1 2

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

FeAu

HGF

LU

MOHO

MO

dI/d

V [

nA

/V]

Voltage [V]

A: PTCDA on gold

B: PTCDA at OD

C: iron

Figure 6. Spectroscopy on the organic nanodots. STS spectroscopic plots(dI/dV) measured for the different areas shown in Figure 5. For A, char-acteristic PTCDA HOMO–LUMO peaks at –1.8 and +1.0 V, and theAu(111) Shockley type surface state at –0.4 eV are observed [26]. For B, apeak at –0.7 V is induced in the organic molecules. This peak is also ob-served for the iron core C.

state)[35] at –0.7 eV. As we move away from the iron core, thecontribution of this state to the PTCDA molecules decays, un-til we reach the 1 ML regime for molecules in A. The inclu-sion of this electronic state in the organic molecules slightlyopens the HOMO–LUMO gap, as observed in the STS curves.It is also remarkable that the gold reconstruction is not ob-served through the PTCDA molecules of the nanodot sur-rounding the iron clusters (see Fig. 5a), while it is clearly visi-ble through the PTCDA layer far from the dot. This fact againindicates that the iron modifies locally the electronic structureof the molecular nanodot, opening the gap and changing thetransparency of the organic layer to the gold states.

In summary, we conclude that by controlling the amount ofiron deposited on a gold surface, we are able to switch fromPTCDA organic chains to organic nanodots. These nanostruc-tures possess a local density of states different from that ob-tained in a molecular monolayer.

This method, to fabricate controlled shaped nanoscopic mo-lecular objects, could be used to guide molecular chainstowards contacts to experimentally determine conductionthrough a single molecule, or to form arrays of nanoscale or-ganic aggregates with potential application in optoelectronicdevices.

Experimental

Experiments were performed in an ultrahigh-vacuum chamber withcommercial Omicron STM1 using Nanotec’s electronics and software.The base pressure of the chamber was 1 × 10–10 mbar (1 bar =100 000 Pa). A gold single crystal was prepared by sputtering–anneal-ing cycles. Iron was evaporated by means of electron bombardmentof a clean iron bar. The evaporation rate was kept below 1 Å min–1.PTCDA molecules were sublimated from a cell at a rate of0.3 Å min–1. STM tips were made of electrochemically etched tung-sten. Tips were cleaned and resharpened in situ by annealing at highvoltage until stable field emission was obtained. Tunneling spectrosco-py was measured by local IV plots and using a CITS technique. Allthe results were checked by observing the characteristic Shockleystate for the clean gold prior to and after performing the spectroscopicmeasurements.

Received: October 5, 2005Final version: April 1, 2006

–[1] J. H. Burroughes, D. Bradley, A. Brown, R. Marks, K. Mackay,

R. Friend, P. Burns, A. Holmes, Nature 1990, 347, 539.[2] Z. Shen, P. E. Burrows, V. Bulovi, S. R. Forrest, M. E. Thompson,

Science 1997, 276, 2009.[3] F. Garnier, R. Hajlaoui, A. Yassar, P. Srivastava, Science 1994, 265,

1684.[4] W. A. Schoonveld, J. Wildeman, D. Fichou, P. A. Bobbert, B. J. van

Wees, T. M. Klapwijk, Nature 2000, 404, 977.[5] N. Tessler, G. J. Denton, R. H. Friend, Nature 1996, 382, 695.

[6] V. G. Kozlov, V. Bulovi, P. E. Burrows, S. R. Forrest, Nature 1997,389, 362.

[7] R. M. Metzger, Chem. Rec. 2004, 4, 291.[8] S. Möller, C. Perlov, W. Jackson, C. Taussig, S. R. Forrest, Nature

2003, 426, 166.[9] H. Park, J. Park, A. K. L. Lim, E. H. Anderson, A. P. Alivisatos,

P. L. McEuen, Nature 2000, 407, 57.[10] S. Kubatkin, A. Danilov, M. Hjort, J. Cornil, J.-L. Brédas, N. Stuhr-

Hansen, P. Hedegård, Th. Bjørnholm, Nature 2000, 425, 698.[11] S. J. Tans, A. R. M. Verschueren, C. Dekker, Nature 1998, 393, 49.[12] G. V. Nazin, X. H. Qiu, W. Ho, Science 2003, 302, 77.[13] T. Yokoyama, S. Yokoyama, T. Kamikado, Y. Okuno, S. Mashiko,

Nature 2001, 413, 619.[14] J. Weckesser, A. De Vita, J. V. Barth, C. Cai, K. Kern, Phys. Rev.

Lett. 2001, 87, 096 101.[15] S. Stepanow, M. Lingenfelder, A. Dmitriev, H. Spillmann, E. Del-

vigne, N. Lin, X. B. Deng, C. Z. Cai, J. V. Barth, K. Kern, Nat. Mater.2004, 3, 229.

[16] M. A. Lingenfelder, H. Spillmann, A. Dmitriev, S. Stepanow, N. Lin,J. V. Barth, K. Kern, Chem. Eur. J. 2004, 10, 1913.

[17] B. Voigtlander, G. Meyer, N. M. Amer, Surf. Sci. 1991, 255, L529.[18] S. Padovani, I. Chado, F. Scheurer, J. P. Bucher, Phys. Rev. B 1999,

59, 11 887.[19] D. D. Chambliss, R. J. Wilson, S. Chiang, Phys. Rev. Lett. 1991, 66,

1721.[20] A. W. Stephenson, C. J. Baddeley, M. S. Tikhov, R. M. Lambert,

Surf. Sci. 1998, 398, 172.[21] J. A. Stroscio, D. T. Pierce, R. A. Dragoset, P. N. First, J. Vac. Sci.

Technol. A 1992, 10, 1981.[22] H. Brune, Surf. Sci. Rep. 1998, 31, 121.[23] H. Brune, M. Giovannini, K. Bromann, K. Kern, Nature 1998, 394,

451.[24] J. Méndez, J. Gómez-Herrero, J. I. Pascual, J. J. Sáenz, J. M. Soler,

A. M. Baró, J. Vac. Sci. Technol. B 1996, 14, 1145.[25] T. Schmitz-Hübsch, T. Fritz, R. Sellam, R. Staub, K. Leo, Phys. Rev.

B 1997, 55, 7972.[26] N. Nicoara, E. Román, J. M. Gómez-Rodríguez, J. A. Martín-Gago,

J. Méndez, Org. Elec., DOI: 10.1016/j.orgel.2006.03.010.[27] I. Chizhov, A. Kahn, G. Scoles, J. Cryst. Growth 2000, 208, 449.[28] a) J. M. Lehn, Polym. Int. 2002, 51, 825. b) J. M. Lehn, Prog. Polym.

Sci. 2005, 30, 814.[29] During the preparation of this manuscript nanostructuring of TPA

molecules has been reported by S. Clair, S. Pons, H. Brune, K. Kern,J. V. Barth, Angew. Chem. Int. Ed. 2005, 44, 7294. The link betweenorganic molecules of TPA and iron atoms is reported to be irrevers-ible upon annealing. The authors suggest a strong bonding with theiron by breaking the OH group, which is not the case for PTCDA.

[30] R. Caillard, G. Otero, P. de Andrés, J. A. Martín-Gago, J. Méndez,unpublished.

[31] J. A. Venables, Introduction to Surface and Thin Film Processes,Cambridge University Press, Cambridge 2000, p. 144.

[32] A similar peak has been observed at –0.6 eV by photoemission for5 ML of iron growth on Cu(111). A. Brodde, K. Dreps, J. Binder,C. Lunau, H. Neddermeyer, Phys. Rev. B 1993, 47, 6609.

[33] P. G. Piva, G. A. DiLabio, J. L. Pitters, J. Zikovsky, M. Rezeq, S. Do-gel, W. A. Hofer, R. A. Wolkow, Nature 2005, 435, 658.

[34] R. Hamers, R. Tromp, J. Demuth, Phys. Rev. Lett. 1986, 56, 1972.[35] F. Flores, C. Tejedor, J. Phys. C 1987, 20, 145.

______________________

CO

MM

UN

ICATI

ON

S

2052 www.advmat.de © 2006 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Mater. 2006, 18, 2048–2052