Determination of iodine adlayer structures on Au(111) by … JChemPhys 107... · and the 223A3...

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: Jul 23, 2021

Determination of iodine adlayer structures on Au(111) by scanning tunnelingmicroscopy

Huang, Lin; Horch, Sebastian; Zeppenfeld, Peter; Comsa, George

Published in:Journal of Chemical Physics

Link to article, DOI:10.1063/1.474419

Publication date:1997

Link back to DTU Orbit

Citation (APA):Huang, L., Horch, S., Zeppenfeld, P., & Comsa, G. (1997). Determination of iodine adlayer structures onAu(111) by scanning tunneling microscopy. Journal of Chemical Physics, 107(2), 585-591.https://doi.org/10.1063/1.474419

https://doi.org/10.1063/1.474419

https://orbit.dtu.dk/en/publications/a971407c-5399-4ae7-bc1d-73096daad2bf

https://doi.org/10.1063/1.474419

Determination of iodine adlayer structures on Au(111) by scanningtunneling microscopy

Lin Huang, Peter Zeppenfeld,a) Sebastian Horch,b) and George ComsaInstitut fur Grenzflachenforschung und Vakuumphysik, Forschungszentrum Ju¨lich GmbH,D-52425 Ju¨lich, Germany

~Received 22 August 1996; accepted 31 March 1997!

Chemisorbed iodine adlayers on Au~111! films on quartz were studied using scanning tunnelingmicroscopy~STM! at room temperature in air. The iodine was adsorbed by successive deposition ofdroplets of a dilute solution of iodine in methanol. As a function of coverage, various adlayerstructures were obtained. By changing the tunneling parameters, either the iodine adlayer or theAu~111! substrate can be imaged with atomic resolution. In this way, the adlattice properties suchas periodicity, orientation and the local absolute coverage have been characterized with highaccuracy. In the low coverage range (u,0.33), due to the high mobility of iodine atoms, only theunreconstructed Au~111! substrate lattice could be imaged. Atu;0.33, a (A33A3)R30° structureis evident. With increasing coverage, a (p3A3) structure of higher iodine packing density isobserved, which can be described as a uniaxially compressed~striped! phase. Finally, nearmonolayer saturation coverage the iodine atoms form a hexagonal moire´-like pattern withlong-range height modulation. The results are compared with previous measurements on this systemunder UHV and electrolyte conditions. ©1997 American Institute of Physics.@S0021-9606~97!00126-8#

-eexb

o-yn-

iadysul

XS

Ssllgra

cr

onedls-ate-c-calumsellns.

eri-of

e-

or

cedingsot.onicope,er-e-pledisu-ith

re

I. INTRODUCTION

Iodine adsorption on the Au~111! surface has been studied intensively since the early 1980s. The investigations wperformed under various conditions and using differentperimental methods. The iodine structures were studiedlow energy electron diffraction~LEED! in situ in UHV,1 andex situ following emersion from iodide solutions under ptential control.2 Recently, scanning tunneling microscopwas employed to study this system in electrolyte solutiounder electrode potential control3,4 as well as in organic polar solvents and in air.5,6 The results obtained from LEEDand surface x-ray scattering~SXS!7,8 in reciprocal spacecombined with those in real space using STM should,principle, enable the exact determination of the iodinelayer structures. However, due to the complexity of this stem there are still some uncertainties in the published resIn the low coverage range, the (A33A3)R30° structure re-ported by LEED and STM was not observed in the Smeasurements. At higher coverages, a (53A3) structure wasfound by LEED and STM,2,3,5 while a (333) structure wasalso observed.4 In the corresponding coverage range, SXmeasurements revealed a (p3A3) phase which compressecontinuously with increasing potential in the electrolyte ceFurthermore, near iodine monolayer saturation coveraSXS experiments found a rotated-hexagonal incommensuphase, whereas a high-order commensurate (737)R21.8°structure was also suggested as a possible model to desthe moire-like pattern observed by STM.3,5

a!Author to whom correspondence should be addressed. Current addInstitut fur Experimentalphysik, Universita¨t Linz, A-4040 Linz, Austria;Fax:143-732-2468-9677; [email protected]

b!Current address: Institute of Physics, University of A˚ arhus, DK-8000Aarhus, Denmark.

J. Chem. Phys. 107 (2), 8 July 1997 0021-9606/97/107(2)/58

Downloaded¬29¬Nov¬2005¬to¬192.38.67.112.¬Redistribution¬subject¬

re-y

s

n--ts.

.e,te

ibe

The structure and phase diagram of iodine adsorbedAu~111! is to some extent comparable to that of physisorbrare gas adlayers on graphite9,10 and single-crystal metasurfaces.11 Indeed, I/Au~111! can be viewed as a model sytem to study 2D phenomena, such as commensurincommensurate~C-I! phase transitions of the surface strutures. The study of iodine adsorption also finds practiapplications, for example, as protective coating of platinsingle-crystal surfaces.12 One important advantage of thisystem is its stability: It can be investigated in UHV, as was in electrolytic solutions and under ambient conditioComparing the phase diagram of I/Au~111! as a function ofcoverage and temperature obtained under different expmental conditions may further improve the understanding2D systems in general.

II. EXPERIMENT

The STM experiments were performed using a hommade modified ‘‘beetle’’ type instrument.13 The microscopewas operated at room temperature either in rough-vacuumin a neutral gas atmosphere. A glass bell jar~Leybold!,which serves as the vacuum chamber for the STM, is plaon the open flange of a metal pot. There are several openwith flanges distributed around the side wall of the pThrough these flanges the pumping system and electrsignals are connected to the chamber and the microscrespectively. The STM head is mounted on a levmicrometer, by which it can be positioned on top of or rmoved from the sample. The sample is fixed on the samholder by a special ‘‘walk-ring’’ with a slope of 0.3 mm usefor the tip-sample coarse approach. The whole systemplaced on an air damped table for vibration isolation. Usally, the chamber is first evacuated and then either filled w

ss:

5855/7/$10.00 © 1997 American Institute of Physics

to¬AIP¬license¬or¬copyright,¬see¬http://jcp.aip.org/jcp/copyright.jsp

nbnchploiorta

neimvealethesao-

n-re, ttro.2ryteghmlinu

d-etbrcsusee

te

M

he

g-mtio/e

n

ab

on

c--omd-andssivewasa-tions-ha-talpo-ents,es.rre-ed.

ese-

ly-a

tly

of

-g ahasHV

in

ces

tednTheg.he

d-on-y of

re-

na-evi-

586 Huang et al.: Iodine on Au(111)

neutral gas~Ar, N2) or kept under vacuum. With this desigthe influence of thermal drift and acoustic vibration caneffectively reduced. Furthermore, the sample surface cakept free from water or other contamination in air for a mulonger period of time. Note, however, that during the sampreparation and transfer to the STM, it is impossible to avthe exposure of the sample surface to the air in the labtory. On the other hand, the design is still convenientchange the tip and sample. The pumping system is usudisconnected from the chamber during image acquisitionorder to minimize the vibrations.

All STM images were obtained using constant curremode. One important feature of the electronics is the usan adjustable electronic high-pass filter for recording theages. This acquisition mode allows suppression of the aage inclination of the sample and, thus, the full gray sccan be used to display the surface morphology. Anotherfect caused by this high-pass filter is the impression thatsurface is illuminated from the left. In some particular casthe contrast of the images was artificially enhanced after dacquisition. No other ‘‘filters’’ were used in the image prcessing.

One of the major problems in STM is due to the uknown structure of the tunneling tip during the measuments. There are several ways to prepare tunneling tipsmost popular methods are mechanical forming or elecchemical etching. In this study mechanically polished 0mm diameter iridium tips were used. Since iridium is vehard, the mechanical grinding method can be easily adopAlso, iridium is rather stable against oxidation. The hiquality atomic-resolution images obtained with Ir tips deonstrate that Ir is indeed a good choice for use as tunnetip in air, and that the preparation method used here is scessful.

The Au~111! surface is a well-suited substrate for asorption studies. Although gold single crystals are the ‘‘bter’’ substrates for STM studies, high cost is the main prolem that prevents them from being widely used. The seafor lower-cost, easy to handle substitutes has been theject of many investigations. The interest has mainly focuon gold thin films deposited on various kinds of substratsuch as mica,14–17 quartz glass18 and silicon.19 In this studywe have used thin gold films deposited on quartz. Afproper preparation~described below!, these Au~111! thinfilms are indeed well-suited substrates, particularly for STstudies in air.

The deposition of the gold films was carried out at tInsititut fur Schicht und Ion-entechnik~ISI! at the Fors-chungszentrum Ju¨lich. The background pressure durindeposition was 531025 mbar. In order to improve the adhesion of gold films to the quartz substrate, 20 Å of chromiuwere deposited on the clean quartz prior to the evaporaof gold. The gold was evaporated at a rate of about 10 Åduring evaporation the substrate was held at room tempture. The film thickness~about 2000 Å! was monitored by aquartz crystal thickness controller. After short annealing ibutane flame~30–40 seconds at;1000 K! and coolingdown in air, the average grain size increased consider

J. Chem. Phys., Vol. 10


ebe

eda-ollyin

tof-r-ef-e,ta

-he-5

d.

-gc-

--hb-ds,

r

ns;ra-

a

ly

and the 223A3 surface reconstruction could be observedthe large~a few hundred nanometers wide! Au~111! terraces.

From the results of a combined LEED and Auger eletron spectroscopy~AES! study,1 it is known that the desorption of iodine atoms in the monolayer coverage range frthe Au~111! surface starts at about 200 °C. Obviously, asorption at room temperature is essentially irreversible,the desired coverage can be reached by adding succeiodine doses onto the gold substrate. The iodine adlayerformed by depositing droplets of an iodine solution in methnol on the freshly annealed gold substrate. The concentraof the solution was in the range of 0.1–1 mM. After depoiting the droplets we waited for the evaporation of the metnol and the sample was transferred to the STM. The toiodine coverage was controlled either by consecutive desition on the same sample between the STM measuremor by depositing more or fewer droplets on different samplUsing this procedure various iodine adlayer structures cosponding to different monolayer coverages can be obtainThe exact local iodine coverage was determineda posterioriby counting the number of iodine atoms in the STM imagand making use of the ‘‘selective imaging’’ technique dscribed in Section III B.

III. RESULTS

A. Au(111) thin films on quartz

After deposition and prior to annealing, the STM anasis revealed that the Au films consist of small grains withtypical diameter of about 100–500 Å, as shown in Fig. 1~a!.After flame annealing, the average grain size is significanincreased and large facets with~111! orientation are formed@Fig. 1~b!#. The large atomically flat terrace in the centerthe image extends over;4000 Å.

It is well known that the clean Au~111! surface under-goes a characteristic 223A3 reconstruction at room temperature, which has been the subject of many studies. Usinvariety of surface science techniques, the reconstructionbeen observed and investigated not only under Uconditions,20,21 but also in air, organic solution18 and aque-ous electrolytes.22 The observation of this reconstructionair and various solutions indicates that the Au~111! surface israther inert under ambient conditions. On the large terraof our flame-annealed samples, the 223A3 surface recon-struction was routinely observed; the STM images presenin Figs. 1~c! and 1~d! show the details of this reconstructiopattern on a large and the atomic scale, respectively.223A3 unit cell is marked by the white rectangle in Fi1~d!. The STM results reveal that the morphology of treconstruction of Au~111! in air are indistinguishable fromthose in UHV.20,21 Due to the unavoidable presence of asorbates under ambient conditions, the mobility of the recstruction patterns in air is higher. In rare cases, the decathe reconstruction and a phase transition from the (223A3)into a ~131! phase were also observed on the freshly ppared samples.

Since this reconstruction is very sensitive to contamition, the presence of the reconstruction can be taken as

7, No. 2, 8 July 1997


tddarcr

eighionila

o

e

uc-isicce

a

Auges

cal

hen

redngs

ge,di-ngas

veTheM

:

tion

587Huang et al.: Iodine on Au(111)

dence for the clean Au~111! surface. Also, we found thaafter flame annealing, the samples which had been storeair for several months gave similar results as the freshlyposited samples. Therefore, these low-cost, easily prepsamples can be used as ideal substitutes for gold singletals in STM studies.

B. Iodine adlayer structures on Au(111)

1. (p3A3) structureAfter iodine adsorption the Au~111! surface was found

to be transformed into the unreconstructed (131) phase. Atsmall iodine coverage (u,0.33), iodine atoms could not bdirectly imaged with STM presumably because of their hmobility at room temperature. Apparently, iodine adsorptleads to an enhanced mobility of the Au step atoms, simto the case of Na adsorption on Au~111!.23 The steps are nolonger straight and parallel to the@110# high symmetry di-rections, instead, they meander on the surface and bec‘‘frizzy.’’ This situation is illustrated in Fig. 2~a!. At theatomic scale@Fig. 2~b!#, a well-ordered hexagonal latticwith nearest-neighbor spacing~NNS! of a5c52.9 Å andb52.7 Å is observed, which is ascribed to the atomic strture of the unreconstructed Au~111! surface. Due to the mechanical asymmetry of the STM-scanning head, thereslight distortion in the observed hexagonal structure whcould in principle be corrected for using the known lattiparameter of the Au~111! surface (aAu52.884 Å!. We havemeasured the average NNSs and used these values to c

FIG. 1. Surface morphology of gold thin films on quartz.~a! After deposi-tion without further preparation (600036000 Å!. Tunneling parametersVt 5400 mV, I t51.0 nA. ~b! After flame annealing for 40 s at; 1000 K.A large atomically flat terrace extends over almost the entire 450034500 Åscanned area. Tunneling parameters:Vt5400 mV, I t51.5 nA. ~c! Step andterrace morphology of the reconstructed Au~111! surface (150031500 Å!.Tunneling parameters:Vt520 mV, I t520 nA. ~d! The 223A3 reconstruc-tion at the atomic scale (52352 Å!. Tunneling parameters:Vt510 mV,I t510 nA.



ine-edys-

r

me

-

ah

lcu-

late the area of a unit cell and, hence, the number ofsurface atoms in each image. The number of atoms in imaof the same scan size as in Fig. 2~b! is calculated to 400610, which will be used as reference to determine the loiodine coverage.

The first stable iodine adlayer structure is observed wthe iodine coverage reachesu.0.33. At this coverage theiodine adatoms are forced into registry~most likely intothree-fold hollow sites! on the Au~111! surface to form a(A33A3)R30° structure. The image shown in Fig. 3~a!shows an approximate hexagonal lattice with the measuNNS along the three high symmetry directions beia55.1 Å,b54.8 Å andc55.0 Å, respectively. These NNSare close to the value ofA3aAu55.00 Å expected for the(A33A3)R30° structure. To determine the local coverathe number of the atoms in this image is calculated andvided by the number of substrate atoms yieldiu50.3260.01. Thus, the observed structure is identifiedthe commensurate (A33A3)R30° structure.

If the iodine coverage is only slightly increased abou50.33, the adlayer structures change correspondingly.(53A3) structure reported in previous LEED and ST

FIG. 2. Surface morphology of iodine covered Au~111! (u,0.33). ~a!‘‘Frizzy’’ monatomic steps meander on the surface after iodine adsorp(160031600 Å!. Tunneling parameters:Vt5200 mV, I t55.4 nA. ~b!Atomic resolution image of the (131) Au~111! surface (52352 Å!. Tun-neling parameters:Vt510 mV, I t510 nA.

FIG. 3. Iodine adlayer structures:~a! (A33A3)R30° phase (52352 Å!.Tunneling parameters:Vt510 mV, I t53.8 nA. ~b! (p3A3) (p 5 5) struc-ture (52352 Å!. Tunneling parameters:Vt510 mV, I t57 nA. ~c! ‘‘Selec-tive imaging’’ of the Au~111! substrate, same area as in~b! (52352 Å!.Tunneling parameters:Vt510 mV, I t529 nA.

7, No. 2, 8 July 1997


hetu

teesei-

igthnd’’-crocaumde

-d-onh

ec

etinea

iaue

hetplt-0thtso4thit

ts.meenreantheera-heagereteous

e

veainaslls

-ig.toth

near

theer-

oveed,

all


studies,2,3 and the (p3A3) structure found in SXSexperiments8 both suggest a uniaxial compression of t(A33A3)R30° structure to form a striped phase. This siation is similar to the behavior of the Xe/Pt~111! system,24

for which the symmetry breaking commensuraincommensurate~C-I! transition from a commensurat(A33A3)R30° structure into a striped domain wall phahas been observed.11 Theoretical studies predict this transtion to be of second order.25

At first glance, the STM image in Fig. 3~b! obtained forslightly higher coverage is similar to its counterpart in F3~a! for u50.32. Further analysis, however, reveals thathexagonal lattice is slightly distorted. To determine this kiof structure more accurately, the ‘‘selective imagingmethod was used.5,26 By simply changing the tunneling parameters, either the iodine adlayer or the substrate latticebe imaged at the same place. In this way, the adlattice perties such as periodicity, orientation and local coveragebe determined with high accuracy. Directly after the acqsition of Fig. 3~b!, the tunneling current was increased fro7 nA to 29 nA at constant voltage, corresponding to acrease of the tunneling gap impedance from 1.4 MV to 0.34MV. The image then obtained is shown in Fig. 3~c!. It isobvious that this image with the average NNS of 2.860.1 Åis very similar to the one in Fig. 2~b!: It represents the unreconstructed Au~111! lattice. With this reference of the gollattice, the iodine structure in Fig. 3~b! can be easily determined. It is found that the distance between the atoms althe rows running from the upper left toward the lower rigcorner maintain the commensurate NNS~5.0 Å!, whereas theatomic spacing along in the other two close packed dirtions are reduced to 4.460.1 Å. This value is close to thevalue of 4.39 Å expected for the (53A3) structure reportedin the previous STM studies.3,5 The corresponding unit cell isindicated by the white rectangle in Fig. 3~b!. Analysis of theimages yields a local iodine coverage ofu50.4060.01. Theangle between the two compressed iodine rows and the nest gold close packed direction is measured to 26°, deviaby about 4° from the value characteristic of th(A33A3)R30° structure. The selective imaging process wfound to be reproducible with different samples and tips. Wcould image the gold substrate by either lowering the bvoltage at constant current or increasing the tunneling crent at constant voltage, which indicates that the decreasthe gap impedance~and hence the tip-sample separation! isresponsible for this interesting phenomenon.

A large number of experiments were performed in tcoverage range between 0.33 and 0.4. The results showthe adlayer structure is more complex than the sim(53A3) structure. While one of the NNS of the iodine latice seems to always keep the commensurate value of 5.the other two can take various values associated withcontinuous change of the local coverage. We found thathe coverage range between 0.37 and 0.4, the compresincreases continuously with increasing coverage, with a cresponding change of the compressed NNS from 4.6 toÅ. These two values are close to those expected for(p3A3) structures withp58 and 5 iodine atoms per un



-

-

.e

anp-ni-

-

gt

-

ar-g

sesr-of

hate

Å,einionr-.4e

cell ~4.59 Å! andp55 with 4 atoms per unit cell~4.39 Å!,respectively. The SXS measurements on I/Au~111! in elec-trolyte under potential control yielded very similar resulFor the (p3A3) phase, SXS found coverages varying fro0.365 to 0.409 with corresponding compressed NNS betw4.62 and 4.32 Å. It is remarkable that similar results aobtained under such different experimental conditions. It cthus be concluded that for the anionic iodine adsorbate,adlattice structures are determined by coverage and tempture rather than the particular environment. Although tpreparation method prevents us from varying the covercontinuously, the data based on the large number of discmeasurements can to some extent simulate this continuchange of coverage.

The formation of domain walls during the C-I phastransition has been predicted by theoretical studies,27 andwas observed experimentally for various systems.11,28 TheLEED study of the I/Au~111! system1 reported a splitting ofthe originalA3 beams into triplets at iodine coverage abo0.33, which may be caused by the formation of domwalls.29 Although the uniaxially compressed structure widentified in the previous STM studies, such domain wawere not observed so far. The STM image shown in Fig. 4~a!together with its Fourier transform@Fig. 4~c!# and a proposedhard sphere model@Fig. 4~b!# provides evidence for the existence of such domain walls in this system. As seen in F4~a!, the NNS along the rows running from upper leftbottom right varies periodically, forming a striped array wistripes running along the perpendicular direction@white ar-rows in Fig. 4~a!#. In the Fourier transform@Fig. 4~c!# of thisimage, extra spots marked by white circles are observedthe commensurate positions~marked by white rectangles!.Further analysis shows that this pattern corresponds todiffraction pattern expected for a striped phase with supheavy domain walls.29 Note that only one out of threeequivalent domains is present in the STM image. The abresults clearly demonstrate that the iodine adlayer, inde

FIG. 4. ~a! STM topograph showing a uniaxially compressed domain wphase (92392 Å!. Tunneling parameters:Vt510 mV, I t53.8 nA. ~b! Pro-posed hard sphere model.~c! Fourier transform of the image in~a!.

7, No. 2, 8 July 1997


th

teina

cg

ahem

aaxuab

a-ae

eouc

th

eng

th

ictintasm

t-lleicthTt

cth

actofbyes-

atn-ts,tricnDy

3°

ace

e

erus

nsu-pe-rn.d ined inhe

ed

-


forms a uniaxially compressed domain wall phase atcoverage.

To determine the structure more accurately, a (131)grid is superimposed on the STM image in Fig. 4~a! in sucha way that the iodine atoms in the supposedly (A33A3)R30° commensurate regions fall right into the three-fold siof the Au~111! substrate. The position of the atoms withthe domain walls can then be determined. A possible hsphere model for this structure is shown in Fig. 4~b!. Notethat the superstructure is discerned by a (53A3) unit cell. Itcan be seen that due to the stress induced by the presenthe wall, the atoms within the walls may relax by shiftinaway from it ~probably toward the nearest bridge site!. Thissituation is indicated by the small arrows drawn in the bmodel. By the atomic relaxation within the domain walls tatoms take up positions close to those in a uniformly copressed phase: Comparing the (53A3) model presented inFig. 3~b! with the ball model discussed above, it is clear ththe first one is derived from the second one after full relation of the domain walls. The strong relaxation and flucttions may explain why such domain walls are rarely oserved in the STM images.

2. Moire -type structures at higher iodine coverage

Haisset al.5 found a structure with long-range modultion at iodine saturation coverageu50.44 and suggested(737)R21.8° structure as a possible model. Also, intermdiate structures consisting of patches of local (53A3) andmoire-like patterns at coverages aroundu50.42 were re-ported. In the same coverage range, SXS measuremfound that the adlayer undergoes a first-order transition frthe (p3A3) to a hexagonal incommensurate rotated strture at a critical coverage ofuc50.409. The NNS in thishexagonally compressed phase first increases to 4.43 Ådecreases to 4.3 Å at a coverage of 0.445.

Figure 5~a! shows an image of a well-ordered iodinlattice at monolayer saturation coverage with clear lorange periodicity. The zoom-in image in Fig. 5~b! reveals theadlayer structure at the atomic scale. The analysis ofFourier transform~FFT! of Fig. 5~a! @shown in Fig. 5~c!#provides additional information about both the superlattand the atomic structure. Two sets of spots can be disguished in the FFT image: The hexagonal set in the cecorresponds to the superstructure; the outer spots whichsplit into triplets reflect the atomic structure. Although thetriplets resemble the LEED patterns obtained in UHV frothis system1 and from Kr/graphite,9 we find that this patternslightly differs from those LEED patterns. In the LEED paterns, one of the spots in the triplets is displaced radiaoutward along theA3 direction in reciprocal space. Also, thoff-axis splitting of the other two spots is highly symmetrwith respect to this direction, indicating a compression ofadlayer along theA3 direction in real space. In the FFimage shown in Fig. 5~c!, it is obvious that if we assume thaone of the spots is displaced along theA3 direction, thetriangle would not be symmetric with respect to this diretion. This implies that either the compression is not along



is

s

rd

e of

ll

-

t---

-

ntsm-

en

-

e

en-erree

y

e

-e

A3 direction or that the adlayer is rotated from the exA3 orientation in real space. Such a ‘‘rotational epitaxy’’the adlayer above a critical misfit is, in fact, predictedtheory,30 and has been observed in many experimental invtigations of physisorbed adlayers.10,11 If we assume a uni-form compression of a rotated lattice, this would imply ththe first-order diffraction spots were split from the commesurateA3 position in such a way that the center of the tripleshould remain at theA3 position in reciprocal space. This isindeed, verified by the results presented below. A symmesplitting of theA3 spot into triplets centered at the positioof the originalA3 spot was likewise observed in a LEEstudy of S/Ru~0001!.28 Then, the orientation and periodicitof the superlattice is easily determined@see Fig. 5~c!#. As aresult, the superlattice is found to be rotated by aboutfrom the R30° direction and has a periodicity of;19.5 Å.The NNS calculated from the triplet splitting is 4.460.1 Å,in excellent agreement with the values obtained in real spfrom the STM topograph in Fig. 5~b!. The coverage is cal-culated to beu50.4560.01, which is believed to be thsaturation coverage of iodine on Au~111!.8

The observed structure is very similar to the high-ordcommensurate (737)R21.8° phase as suggested in previostudies performed in air and electrolyte solution.3,5 It ispointed out that these authors also realized the incommerate nature of this structure and have observed differentriodicities and rotation angles of this hexagonal patteBased on the FFT results and the STM images obtainereal space, we propose the hard sphere model presentFig. 5~d! for the structure observed here. The unit cell of t(737)R21.8° is also indicated~dashed lines!. It can be seenthat the proposed model deviates from the (737)R21.8° bya rotation of 5° whereas the length of the unit cell is reducfrom 20.2 Å to;19.5 Å.

FIG. 5. STM images of well-ordered moire´-like structures at saturationiodine coverage.~a! Large-scale overview (4003400 Å!. Tunneling param-eters:Vt520 mV, I t515 nA. ~b! Zoom-in image revealing the atomic structure (52352 Å!. Tunneling parameters:Vt520 mV, I t515 nA. ~c! Fouriertransform of the image in~a!. ~d! Proposed hard sphere model.

7, No. 2, 8 July 1997


aob

onener

.s-

tobe

iorirr

ldegein

ougt

antluisic

theofthenceede as

zy’’rlyture.pe-°-alb-o-ri-

ethodc-

e.of

n-heasntA

ichTheac-andyerisbe

ine


From our experiments, we cannot determine the exvariation of the rotation of the iodine adlayer as a functioncoverage. However, the x-ray scattering results obtainedOcko et al.8 indicate a linear dependence of the rotatiangle on the incommensurability, in qualitative agreemwith the Novaco–McTague theory. To our knowledge, thexists no theoretical study of the I/Au~111! system and therelevant binding and corrugation energies are not knownquantitativecomparison with theory is, therefore, not posible.

Besides the homogeneously compressed phase, ostructures which are not as well-ordered, were oftenserved near saturation coverage. Figure 6 shows twoamples of this kind of structures. As seen in Fig. 6~a!, theadlayer structure differs from the (p3A3) structure by itslocal height modulation. On the other hand, this modulatis not as regular as in the case shown in Fig. 5. The Foutransform of this image reveals three sets of spots cosponding to~i! the gold substrate lattice,~ii ! the iodine ad-lattice and~iii ! the iodine superlattice. Although the gosubstrate is not discernible in the STM image presented hit can be observed in the high magnification zoom-in imaWith the reference of the gold lattice, the average iodNNS is determined to be 4.460.1 Å. The coverage isu50.4360.01. The adlayer lattice makes an angle of ab26° with respect to the gold substrate. Using this FFT imawe were able to test the assumption made earlier thatcenter of the triangle should lie at theA3 position. Indeed,we find that the angle between connections from the~0, 0!spot in reciprocal space to the center of the trianglesthose to the nearest spots of the gold lattice is 30°, andlength ratio between them is 1.8, which is close to the vabeing expected for theA3 position. As mentioned above, thimplies that the compression of the adlayer is isotrop

FIG. 6. ~a! Intermediate structure composed of moire´-type and (p3A3)patches (1003100 Å!. Tunneling parameters:Vt5260 mV, I t510 nA. ~b!Moire-like structure across a ‘‘frizzy’’ step (52352 Å!. Tunneling param-eters:Vt510 mV, I t58.3 nA. ~c! Fourier transformed image of~a!. ~d!Fourier transformed image of~b!.



ctfy

te

A

her-x-

nere-

re,.e

te,he

dhee

,

whereas the adlayer is rotated by a few degrees fromcommensurate R30° direction. Furthermore, this kindhexagonal rotated structure was found to coexist with(p3A3) structure on the same sample area. This coexisteis suggestive of a first-order transition from the strip(p3A3) to the rotated hexagonal incommensurate phaspredicted by theory.25,27

Figure 6~b! depicts the moire´-like pattern near saturationcoverage on two adjacent terraces separated by a ‘‘frizmonatomic step. The local distortions in this image cleaevidence the incommensurate nature of the adlayer strucA quantitative analysis reveals that the superlattice has ariodicity of ;21 Å and is rotated about 7° from the R30direction. It is well known that the periodicity and orientation of moire patterns are extremely sensitive to the locmisfit, and the relative orientation of the adlayer and sustrate lattices.31 Therefore, small distortions and local hetergeneities will result in a significant change of the local oentation of the pattern.

To determine the registry of the iodine lattice with thsubstrate more accurately, Gao and Weaver used the meof changing potential in the midway of the scan in the eletrolyte to obtain ‘‘composite’’ images,3 which show the io-dine adlayer and the Au~111! substrate in the same imagWe have adopted this idea in our experiment. Insteadchanging the potential, the tunneling impedanceRt waschanged roughly midway during the STM scan, which eabled us to obtain similar ‘‘composite’’ images through tselective imaging effect. The image shown in Fig. 7 wrecorded with a bias voltage of 10 mV; the tunneling curreused for imaging the top half and bottom half was 25 n(Rt50.4 MV) and 2.1 nA (Rt54.8 MV), respectively. Twodifferent well-ordered hexagonal lattices are observed, whcorrespond to the gold and iodine structure, respectively.registry between the iodine adlayer and substrate can becurately determined. The angle between the adlatticegold substrate in this case is 24°, the NNS of the adlastructure is 4.360.1 Å. The selective imaging processfound to be reversible, i.e. the iodine adlayer seems to

FIG. 7. ‘‘Composite images’’ revealing the registry between the iodadlayer and the gold substrate (1003100 Å!. Tunneling parameters: Vt 510 mV, It 5 25 to 2.1 nA.

7, No. 2, 8 July 1997


e

rethaoeithpu

e

edhioeislin

ail

5tothg, ttethnay

.4.tha

ssramt

rara.gar

ntoem-g’’ocal

ut-

-

as,

L.

lid

Sci.

rf.

Sci.


only temporarily affected by the stronger interaction betwetip and sample at the lower gap impedance.

The complete understanding of the selective imagingquires a calculation of the tip-sample interactions andbonding strength between I–I and I–Au atoms, as wellfurther experiments. At present, we can only speculate abthe possible mechanisms:~i! An electronic state effect can bexcluded because the phenomenon is observed by echanging the bias voltage or the tunneling current, i.e. simcaused by the reduction of the tip-sample separation throthe lowering of the tunneling gap impedance.~ii ! From atheoretical study on the energy dissipation in STM,32 it wasinferred that the tunneling processes may induce a rissurface temperature by as much as 30 K atRt5105 V andVt51 V on metal surfaces. In our case, the tunneling impance is of the order of 105 V, but the voltage used is mucsmaller than 1 V. Thermal annealing experiments of thedine monolayer1 indicate that the desorption of the iodinfrom Au~111! only starts at about 200 °C. Therefore, itunlikely that the heat transfer associated with the tunneprocess could account for the observed phenomenon.~iii !The reduction of the tip-sample distance leads to an increof the tip-sample interaction, which in some cases wmodify the sample surface.33 In fact, the ‘‘critical’’ gap im-pedance needed for selective imaging is of the order of 0.V which is similar to the threshold impedance requiredslide individual atoms on metal surfaces by means ofSTM tip.34,35 It is, therefore, conceivable that by stronenough interactions between the tip and the iodine atomslatter will be forced to temporarily leave the adsorption siunderneath the scanning tip. Due to the high mobility ofiodine atoms, the adlayer may re-order itself quickly alook as if it was unperturbed once the tip has moved aw

IV. CONCLUSIONS

Three distinct iodine adlayer structures on Au~111! havebeen observed under ambient conditions:~i! At low iodinecoverage, a commensurate (A33A3)R30° structure is ob-tained; ~ii ! In the coverage range between 0.33 and 0uniaxially compressed (p3A3) structures were observedThe rarely observed domain-wall structure indicates tharather uniform compression is favored within this phase. Tdomain walls are ordered in a striped array and the phtransition appears to be of second order.~iii ! Near iodinemonolayer saturation coverage, a hexagonally comprestructure is observed, which is found to be incommensuwith the substrate. The adlayer is more uniformly copressed and rotated by a few degrees with respect tocommensurate phase. The coexistence of the moire´-type and(p3A3) structures reveals that the corresponding phase tsition is of first order. The observed sequence of phase tsitions is similar to that found for rare gas monolayers, eXe/Pt~111!, and is entirely consistent with the theoreticprediction: The transition is expected to be second orde



n

-esut

erlygh

in

-

-

g

sel

M

e

hesed.

,

aese

edte-he

n-n-.las

the (A33A3)R30° commensurate phase first transforms ia uniaxially compressed (p3A3) phase and first order as thcompression becomes isotropic in the hexagonally copressed phase. Finally, we note that the ‘‘selective imaginmethod has been successfully used to determine the liodine coverage and the adlayer structure.

ACKNOWLEDGMENTS

L.H. acknowledges the financial support from the Descher Akademischer Austauschdienst~DAAD !. The goldfilms were prepared at the Institut fu¨r Schicht und Ionentechnik ~ISI!, Forschungszentrum Ju¨lich GmbH.

1S. A. Cochran and H. H. Farrell, Surf. Sci.359, 95 ~1980!.2B. G. Bravo, S. L. Michelhaugh, M. P. Soriaga, I. Villegas, D. W. Suggand J. L. Stickney, J. Phys. Chem.95, 5245~1991!.

3X. Gao and M. J. Weaver, J. Am. Chem. Soc.114, 8544~1992!.4N. J. Tao and S. M. Lindsay, J. Phys. Chem.96, 5213~1992!.5W. Haiss, J.K. Sass, X. Gao, and M.J. Weaver, Surf. Sci. 274~1992!L593.

6R. L. McCarley and A. J. Bard, J. Phys. Chem.95, 9618~1991!.7J. Wang, G. M. Watson, and B. M. Ocko, Physica A200, 679 ~1993!.8B. M. Ocko, G. M. Watson, and J. Wang, J. Phys. Chem.98, 897 ~1994!.9M. D. Chinn and S. C. Fain, Jr., Phys. Rev. Lett.39, 146 ~1977!.10E. D. Specht, A. Mak, C. Peters, M. Sutton, R. J. Birgeneau, K.D’Amico, D. E. Moncton, S. E. Nagler, and P. M. Horn, Z. Phys. B69,347 ~1987!.

11K. Kern, R. David, P. Zeppenfeld, R. L. Palmer, and G. Comsa, SoState Commun.62, 391 ~1987!.

12R. Vogel and H. Baltruschat, Ultramicroscopy42–44, 562 ~1992!.13K. Besocke, Surf. Sci.181, 145 ~1987!.14C. E. D. Chidsey, D. N. Loiacono, T. Sleator, and S. Nakahara, Surf.200, 45 ~1988!.

15A. Putnam, B. L. Blackford, M. H. Jericho, and M. O. Watanabe, SuSci. 217, 276 ~1989!.

16J. A. DeRose, T. Thundat, L. A. Nagahara, and S. M. Lindsay, Surf.256, 102 ~1991!.

17M. Hegner, P. Wagner, and G. Semenza, Surf. Sci.291, 39 ~1993!.18W. Haiss, D. Lacky, J. K. Sass, and K. H. Besocke, J. Chem. Phys.95,2193 ~1991!.

19A. Hammiche, R. P. Webb, and I. H. Wilson, Vacuum45, 569 ~1994!.20Ch. Woll, S. Chiang, R. J. Wilson, and P. H. Lippel, Phys. Rev. B39,7988 ~1989!.

21J. V. Barth, H. Brune, G. Ertl, and R. J. Behm, Phys. Rev. B42, 9307~1990!.

22N. J. Tao and S. M. Lindsay, J. Appl. Phys.70, 5141~1991!.23J. V. Barth, R. J. Behm, and G. Ertl, Surf. Sci.302, L319 ~1994!.24K. Kern, Phys. Rev. B35, 8265~1987!.25P. Bak, D. Mukamel, J. Villain, and K. Wentowska, Phys. Rev. B19,1610 ~1979!.

26S. L. Yau, C. M. Vitus, and B. C. Schardt, J. Am. Chem. Soc.112, 3677~1990!.

27P. Bak, Rep. Prog. Phys.45, 587 ~1982!.28R. Dennert, M. Sokolowski, and H. Pfnu¨r, Surf. Sci.271, 1 ~1992!.29P. Zeppenfeld, K. Kern, R. David, and G. Comsa, Phy. Rev. B38, 3918

~1988!.30A. D. Novaco and J. P. McTague, Phys. Rev. Lett.38, 1286~1977!.31J. Chevrier, L. Huang, P. Zeppenfeld, and G. Comsa, Surf. Sci.335, 1

~1996!.32F. Flores, P. M. Echenique, and R. H. Ritchie, Phys. Rev. B34, 2899

~1986!.33G. M. Shedd and P. E. Russell, Nanotechnology1, 67 ~1990!.34D. M. Eigler and E. K. Schweizer, Nature344, 524 ~1990!.35P. Zeppenfeld, C. P. Lutz, and D. M. Eigler, Ultramicroscopy42–44, 128

~1992!.

7, No. 2, 8 July 1997


Determination of iodine adlayer structures on Au(111) by … JChemPhys 107... · and the 223A3...

Documents

Transcript of Determination of iodine adlayer structures on Au(111) by … JChemPhys 107... · and the 223A3...