Bowker M., Davies P.R. (Eds.) Scanning Tunneling Microscopy in Surface Science, Nanoscience and Catalysis

Подождите немного. Документ загружается.

conditions, lead to a retraction of the STM tip, which results in a larger diameter and

height in the STM images. A very similar electronic state is found in the same energy

range for bulk iron [46] proving that the presence of this state does not dramatically

change with cluster size. We can therefore explain the apparent changes in the STM

images in region II by a combination of electronic effects stemming from the Fe

clusters (1–2 V) and the alumina ﬁlm (2–3 V). The observed changes in cluster size

anddiametercanclearly be related tothecomplexelectronicstructureof the tunneling

gap.Therealclusterheightinthepresentcaseismostlikelyrepresentedbythelimiting

values of the bias voltage around 0 and 5 V and corresponds roughly to 0.2–0.3 nm

(see Figure 2.14), thus to a single atomic layer.

As a result, we want to stress that extreme care has to be exercised when using STM

as a tool for determining the size of oxide-supported metal clusters. Reliable results

can be expected only if the electronic structure of the system is well known and bias

voltage ranges can be avoided where localized electronic states determine the

tunneling process. It seems that the most reliable results can be obtained at rather

high bias voltage corresponding to the conduction band of the oxide ﬁlm and that bias

voltages in the vicinity of the lower conduction band edge of the oxide ﬁlm have to be

avoided.

2.5

Template-Controlled Growth of Model Catalysts

2.5.1

Oxides as Templates

In the previous chapter, we have explored the application of STM for the character-

ization of thin oxide ﬁlms and supported clusters. We now focus our attention on

the template-controlled growth of metal particles on oxide ﬁlms. As we have already

mentioned in Section 2.1, lithographic techniques can be employed to create a

regular array of nucleation sites for cluster growth but only at the expense of a limited

lateral resolution. This does not limit the minimum size of the particles, which

depends only on the amount of metal that is deposited per nucleation site. However,

the resolution limitation is important for the number density of particles on the

support that can be reached. If the unit all size of the cluster array is, for example,

100 nm 100 nm, a number density of only n

¼10

10

2

is obtained, which is

approximately ﬁve orders of magnitude lower than the number density of atoms in

a metallic plane. This number density is much lower than the typical sensitivity of

standard surface science techniques, which is on the order of 1% of a monolayer

(10

13

2

). Consequently, such techniques cannot be applied. It may, however, be

large enough to conduct kinetic measurements of the catalytic reactions on such a

surface at higher pressures. Since our prime goal is – as stated in the introduction – to

study the catalyst surface with atomic precision and correlate this with the reactivity,

we have to conceive systems of much higher island density. The nucleation and

growth of metal particles on random defects on bulk oxides [47] and oxide ﬁlms [4] can

2 The Template Route to Nanostructured Model Catalysts

provide such large number density of clusters, however, at the expense of increased

complexity. This can be overcome only by using ordered nanoscopic reconstructions

of oxide surfaces and ﬁlms as growth templates, which provide a periodicity of a few

nanometers at most.

Fortunately, an increasing number of such surfaces have been found over the

years and promising results in terms of metal particle growth have been achieved.

One of the best examples of template-controlled growth is the thin alumina ﬁlm

on Ni

Al(1 1 1), which has already been discussed in Section 2.3. This ﬁlm is

characterized by a hexagonal superstructure of 4.16 nm lattice constant, which can be

related to the presence of an electronic state in the bandgap of the oxide (Figure 2.9).

High-resolution STM images of this ﬁlm by Schmid et al. suggest that this structure

is caused by oxygen vacancies (holes) in the ﬁlm (Figure 2.15) [30].

Previous AFM measurements did not show these holes [48, 49] so that the origin of

the superstructure is still under debate. However, the STM measurements proved,

in agreement with the AFM measurements, that the corners of the unit cell of the

super lattice formed by the ﬁlms are high-symmetry points and these points can be

related to the so-called dot structure (bright dots in Figure 2.7c) [50]. In addition, a

second type of high-symmetry points was found on the ﬁlm with a periodicity of

2.4 nm, which has previously been referred to as network structure (dark holes in

Figures 7d and 15) [50]. Interestingly, both structures can be employed as templates

for the growth of metallic nanostructures. These high-symmetry points are obviously

high binding energy sites for metals on the alumina ﬁlm due to high coordination

and/or speciﬁc electronic properties.

The growth of a large variety of metals on the alumina ﬁlm on Ni

Al(1 1 1) has been

studied. Initial studies were conducted for the coinage metals Cu, Ag, and Au [51].

For these metals, a rather strong inﬂuence of the dot structure on the nucleation and

growth of clusters was observed at low coverage. As can be seen in Figure 2.16a, the

few Ag clusters that are visible are located on the dot structure, which is also imaged

as protrusions. However, as the coverage is increased, the template control does not

Figure 2.15 High-resolution STM image (a) [30] and AFM image

(b) [49] of the alumina film on Ni

Al(1 1 1). The high-symmetry

sites marked by triangles (circles) and the hexagons correspond

to the network and dot structure, respectively.

2.5 Template-Controlled Growth of Model Catalysts

work to that perfection any more. Au (not shown here) tends to grow to granular

ﬁlms [51] and Ag shows large agglomerates (Figure 2.16b). Only Cu seems to keep

a rather good order showing signs of a regular arrangement with a lattice constant

corresponding to that of the dot structure (Figure 2.16c). In these cases, the growth

temperature was 300 K,and it can be expected that different temperatures, which lead

to different growth kinetics, could signiﬁcantly change the order. Furthermore, it

seems that a stronger interaction of the metal with the substrate, evidenced, for

example, by a higher energy of adhesion [53], leads to a better order of the ﬁlm at high

coverage. This assumption was in part veriﬁed by the growth of Pd clusters on the

alumina ﬁlm on Ni

Al(1 1 1).

In this case, the dot structure of the alumina ﬁlm acts as an efﬁcient template for

the growth at 300 K (Figure 2.17). A nearly perfect array of Pd clusters can, thus, be

obtained by deposition of Pd vapor onto the Al

/Ni

Al(1 1 1). Moreover, the Pd

clusters show a narrow size distribution, which suggests that this surface is close to

the ideal model catalyst of low complexity we are striving for. Unfortunately, for the

reasons given in Section 2.4, it is not easy to determine the actual cluster size by STM

in this case. It can also be seen in Figure 2.17 that the growth kinetics, as explained

in this chapter, plays an important role for the order. By raising the substrate

temperature during growth, the order of the Pd clusters is gradually lost and the

Figure 2.16 STM images of Ag clusters (a, b) and Cu clusters

on Al

/Ni

Al(1 1 1) (c). The image sizes and bias voltages

are (a) 60 nm 60 nm, 2.1 V; (b) 80 nm 80 nm, 0.7 V; and

Ref. [51].)

Figure 2.17 STM images of Pd clusters grown on

/Ni

Al(1 1 1) at different substrate temperatures. The

image size is 100 nm 100 nm and the bias voltage is 0.7 V.

(Reproduced with permission from Ref. [54].)

2 The Template Route to Nanostructured Model Catalysts

number density of the Pd islands decreases as it is expected for heterogeneous

nucleation on the template sites (compare Figure 2.4). The growth of the clusters at

300 K is indeed best described by heterogeneous nucleation and, thus, template-

controlled, as can be rationalized by looking at Figure 2.18. Here, the number

density of islands n(t) is plotted as a function of the deposition time t (deposited Pd

amount). The experimental points are nicely described by the curve for heteroge-

neous nucleation kinetics n(t) ¼n

tanh(t/t), which was introduced by Robinson and

Robins [52]. In this equation, t is a typical timescale, which accounts for the inﬂuence

of substrate temperature and ﬂux of the impinging atoms, and n

is the number

density of the nucleation sites. From Figure 2.18, it can also be seen that initially only

one- and two-layer islands are found. Only after the saturation number density is

reached – that is, all template sites are covered – do these islands start to increase in

height. This indicates that the interaction of the template with the Pd is very speciﬁc

and leads to the almost perfect nucleation behavior. As can be seen from Figure 2.17,

the favorable temperature range for perfect heterogeneous nucleation does not

extend far beyond 300 K for Pd. Already at 400 K substrate temperature, some of

the template control is lost and the island size starts to become inhomogeneous.

A further increase in the growth temperature leads to an additional loss of order and a

reduction in the number density of the Pd clusters. Finally, at 600 K it can be seen that

a number of the smaller Pd clusters are still positioned on the dot structure (template

sites) and that the morphology is dominated by the presence of large ﬂat islands. The

former observation, again, indicates that for this particular system the interaction

between the nucleation sites and the Pd clusters is relatively strong.

0.05

0.04

0.03

0.02

0.01

0.00

number density [nm

-2

]

300250200150100

deposition time [s]

4444444

layer clusters

total

Figure 2.18 Plot of the number density of clusters of different

height as a function of deposition time. The line is a fit to the

nucleation equation given in the text [54].

2.5 Template-Controlled Growth of Model Catalysts

As was mentioned before, not only the dot structure corresponds to high-sym-

metry sites of the oxide ﬁlm but also the network structure may serve as a template

for nanostructured growth for a different choice of metal. Indeed, it has been found

that metals, which show a stronger interaction with the alumina ﬁlm than the metals

discussed above, possess a stronger tendency to nucleate and grow on the sites of

the network structure. This has been observed for Mn [51, 55] and Fe [56], but neither

of these metals shows a regular arrangement of the clusters at higher coverage

much likethe case of Ag and Au on the dot structure. By contrast, for Va nearly perfect

array of clusters has been found for a deposition temperature of 550 K. The lattice

constant of this array is 2.4 nm, which corresponds to the periodicity of the network

structure and, thus, indicates that the network structure can also be used as a growth

template. As can be seen in Figure 2.19, even at 300 K the order of the V cluster is

already very good; however, their size distribution is not as narrow as for a deposition

temperature of 550 K.

The results for Pd and V lead to the conclusion that the high-symmetry sites of the

alumina ﬁlm on Ni

Al(1 1 1) can act as template for the growth of nanostructured

model catalysts. They also prove that kinetic control of the growth is of utmost

importance.

The two cases of Pd and V show the enormous potential of the Al

/Ni

Al(1 1 1)

system for the template-controlled growth of model catalysts. However, a number

of other nanostructured oxide surfaces have also been found to act as growth

templates. Among these examples, we ﬁnd the TiO

(1 1 0)-(1 2) cross-linked

reconstruction, which provides a rectangular arrangement of nucleation sites. On

this surface, a nanostructured growth of Pd clusters is obtained by deposition at 373 K

(Figure 2.21) [57–59]. As in the case of Al

/Ni

Al(1 1 1), the deposition of Au does

not lead to ordered cluster arrays on the TiO

(1 1 0)-(1 2) cross-linked surface [60].

Gold nanoparticles can, however, be grown in ordered arrays on the so-called z

-TiO

ﬁlms on Pt(1 1 1) (Figure 2.20). Again, the template-controlled growth leads to nicely

ordered arrays of Au clusters. In this case, a particular intriguing phenomenon is

encountered. Postannealing of these ﬁlms leads to a change in the symmetry of the

Figure 2.19 STM imagesof V clustersgrown on Al

/Ni

Al(11 1)

at different substrate temperatures. The image size is

80 nm 80 nm and bias voltage is 0.7 V. (Reproduced with

permission from Ref. [51].)

2 The Template Route to Nanostructured Model Catalysts

system, which is interpreted in terms of a transformation of the structure of the

TiO

ﬁlm from the zigzag-like z

-phase to the hexagonal w-phase (wagon wheel

phase) [61]. Iron oxide surfaces can also be used for template-controlled growth as has

been demonstrated in the case of the Fe

(1 1 1) and FeO ﬁlms on Pt(1 1 1), which

have a common reconstructed close-packed oxygen surface layer with long-range

order. On Fe

ﬁlms on MgO(1 0 0), nicely ordered arrays of Fe clusters (Figure 2.22)

have been grown, and similar results were obtained for Fe on FeO/Pt(1 1 1) [62].

Furthermore, the template-controlled nucleation of Cr on Fe

(1 1 1) [62] and V on

FeO/Pt(1 1 1) [62] was demonstrated.

Figure 2.21 STM image of Pd nanoparticles on the cross-linked

(1 2) TiO

(1 1 0) surface [57, 59]. The image size is

100 nm 100 nm.

Figure 2.20 (a) Au on z

-TiO

and (b) Au on w-TiO

after

annealing at 600 K [61]. The image sizes are (a) 100 nm 100 nm

and (b) 110 nm 80 nm.

2.5 Template-Controlled Growth of Model Catalysts

Most of these examples show that it is apparently easier to grow ordered arrays of

more reactive metals, for example, Pd, than unreactive metals such as Au and Ag.

This is most likely due to the low interaction energy of the latter metals with the

oxide surfaces and, consequently, a lower speciﬁcity of the traps. This calls for a

modiﬁcation of the template to increase the interaction energy of the template sites

(traps) with the deposited metals.

2.5.2

Modified Templates

As discussed in the previous section, the huge potential of alumina ﬁlms on

Al(1 1 1) for the growth of model catalysts has been exploited. However, only

in the case of Pd and V reasonably well-ordered nanostructures have been found.

For the other metals that were investigated, either the growth conditions chosen

were not such that nanostructuring was accomplished or the interaction between

the metal and the nucleation sites was not sufﬁcient. To overcome these limitations,

an approach can be used that relies on the modiﬁcation of the alumina ﬁlm on

Al(1 1 1) in such a way that the interaction of the deposited metal and the

nucleation sites is increased. This can be done by taking advantage of the perfect

nucleation behavior of Pd on this ﬁlm: The deposition of a small amount of Pd,

which nucleates on the dot structure, will create an array of small Pd clusters. We

will call this novel surface (Pd/Al

/Ni

Al(1 1 1)) a second-generation template

since it was already produced using a template. On this modiﬁed template, ordered

structures of metals could, indeed, be grown that cannot be produced by nucleation

of the metal on the bare alumina ﬁlm. So far, two cases have been found that

illustrate the potential of this new second-generation template. Hamm et al.

Figure 2.22 STM of Fe nanoparticles on an Fe

film on MgO(1 0 0) [62].

2 The Template Route to Nanostructured Model Catalysts

reported on the room-temperature growth of Au clusters [63] and Schmid et al. on

the growth of Fe clusters [30] on Pd/Al

/Ni

Al(1 1 1). In both cases, nicely ordered

cluster arrays have been observed on the dot structure (Figure 2.23). One should,

however, bear in mind that by this route only bimetallic cluster arrays can be grown

but this does not necessarily represent a big limitation since bimetallic surfaces are

often interesting model catalysts. It should also be mentioned that Au and Fe show a

different nucleation behavior on the bare alumina ﬁlm. Only the action of the

second-generation template leads to the similar nucleation behavior observed.

The situation that is encountered in these two systems is closely related to the

reasoning of Venables [23] who claimed that in order to overcome coarsening a high

chemical speciﬁcity of the nucleation sites and low diffusion lengths are required.

The former is apparently provided by the modiﬁcation of the original template with

Pd atoms, which increases the interaction strength with the second metal. This

suggests that the second-generation template will most probably work for the

nucleation of many other metals as well. This has to be exploited in the near future

and will pave the way for a large variety of nanostructured bimetallic model catalysts.

2.6

Conclusions

Oxide surfaces, and in particular oxide ﬁlms, are versatile substrates for the

preparation of model catalysts. Quite a few of these systems show nanoscale

reconstructions, which can be employed as templates for the growth of ordered

model catalysts of reduced complexity. In order to efﬁciently control the growth of

nanostructured metal particle arrays, two conditions have to be met. First, the

template must provide sites of high interaction energy that trap the deposited metals.

Second, the kinetics of the growth process must be carefully controlled by choosing

Figure 2.23 STM images of bimetallic AuPd [63] (a) and FePd [30]

(b) clusters grown on the Pd/Al

/Ni

Al(1 1 1) second-

generation template. The image sizes are (a) 100 nm 100 nm

and (b) 80 nm 69 nm.

2.6 Conclusions

the right substrate temperature and ﬂux of impinging atoms during deposition.

While the latter point is in general easy to control by the experimentalist, the former

is essentially a function of the materials employed. This limitation can, however,

be overcome by modifying the template such that the interaction energy of the traps

is increased, thereby creating a second-generation template.

References

1 Vang, R.T., Lægsgaard, E., and

Besenbacher, F. (2007) Phys. Chem. Chem.

Phys., 9, 3460.

2 Somorjai, G.A., York, R.L., Butcher, D.,

and Park, J.Y. (2007) Phys. Chem. Chem.

Phys., 9, 3500.

3 B

€

aumer, M., Libuda, J., Neyman, K.M.,

€

osch, N., Rupprechter, G., and Freund,

H.-J. (2007) Phys. Chem. Chem. Phys., 9,

3541.

4 B

€

aumer, M. and Freund, H.-J. (1999)

Prog. Surf. Sci., 61, 127.

5 Bowker, M., Bowker, L., Bennett, R., Stone,

P., and Ramirez-Cuesta, A. (2000) J. Mol.

Catal. A: Chem., 163, 221.

6 Rousset, S., Croset, B., Girard, Y., Pr



evot,

G., Repain, V., and Rohart, S. (2005) C. R.

Physique, 6, 33.

7 Jacobs, P.W., Ribero, F.H., Somorjai, G.A.,

and Wind, S.J. (1996) Catal. Lett., 37, 131.

8 Hla, S.-W. (2005) J. Vac. Sci. Technol. B, 23,

1351.

9 Dutta, J. and Hofmann, H. (2003)

Encyclopedia of Nanoscience and

Nanotechnology, vol. X (ed. H.S. Nalwa),

American Scientiﬁc Publishers, pp. 1–23.

10 F

€

orster, S. (2003) Top. Curr. Chem., 226,1.

11 Kung, M.C. and Kung, H.H. (2005)

Top. Catal., 34, 77.

12 Gehl, B., Fr

€

omsdorf, A., Aleksandrovic, V.,

Schmidt, T., Pretorius, A., Flege, J.-I.,

Bernstorff, S., Rosenauer, A., Falta, J.,

Weller, H., and B

€

aumer, M. (2008) Adv.

Funct. Mater., 18, 2398.

13 Brune, H., R

€

oder, H., Boragno, C., and

Kern, K. (1994) Phys. Rev. B, 49, 2997.

14 Gestermann, S., Nohlen, M., Schmidt, M.,

and Wandelt, K. (1997) Surf. Rev. Lett., 6,

1179.

15 A

€

ıt-Mansour, K., Rufﬁeux, P., Xiao, W.,

€

oning, P., Fasel, R., and Gr

€

oning, O.

(2006) Phys. Rev. B, 74, 195418.

16 G

€

unther, C., Vrijmoeth, J., Hwang, R.Q.,

and Behm, R.J. (1995) Phys. Rev. Lett., 74,

754.

17 Meunier, I., Tr



eglia, G., Legrand, B.,



etot, R., Aufray, B., and Gay, J.M. (2000)

Appl. Surf. Sci., 162, 219.

18 Bendounan, A., Cercellier, H., Fagot-

Revurat, Y., Kierren, B., Yu Yurov, V., and

Malterre, D. (2003) Appl. Surf. Sci., 212, 33.

19 Bendounan, A., Cercellier, H., Fagot-

Revurat, Y., Kierren, B., Yu Yurov, V., and

Malterre, D. (2003) Phys. Rev. B, 67,

165412.

20 Brune, H. and Kern, K. (1997) Growth

and Properties of Ultrathin Epitaxial Layers

(eds D.A. King and D.P. Wopdruff),

Elsevier, pp. 149–206.

21 Wang, C. and Bai, C. (2006)

Applied Scanning Probe Methods IV

(eds B. Bhushan and H. Fuchs), Springer,

pp. 159–181.

22 Venables, J.A., Giordano, L., and Harding,

J.H. (2006) J. Phys.: Condens. Matter, 18,

S411.

23 Venables, J.A. (1997)

Physica A, 239, 35.

24 Pr



evot, G., Guesmi, H., and Croset, B.

(2007) Surf. Sci., 601, 2017.

25 Ellmer, H., Repain, V., Sotto, M., and

Rousset, S. (2002) Surf. Sci., 511, 183.

26 Sabiryanov, R.F., Larsson, M.I., Cho, K.J.,

Nix, W.D., and Clemens, B.M. (2003)

Phys. Rev. B, 67, 125412.

27 Bromann, K., Giovannini, M., Brune, H.,

and Kern, K. (1999) Eur. Phys. J. D, 9, 25.

28 Besenbacher, F. (1996) Rep. Prog. Phys., 59,

1737.

2 The Template Route to Nanostructured Model Catalysts

29 Tersoff, J. and Hamann, D.R. (1983)

Phys. Rev. Lett., 50, 1998.

30 Schmid, M., Kresse, G., Buchsbaum, A.,

Napetschnig, E., Gritschneder, S.,

Reichling, M., and Varga, P. (2007) Phys.

Rev. Lett., 99, 196104.

31 Degen, S., Krupski, A., Kralj, M.,

Langner, A., Becker, C., Sokolowski, M.,

and Wandelt, K. (2005) Surf. Sci., 576, L57.

32 Viernow, J., Petrovykh, D.Y., Kirakosian,

A., Lin, J.-L., Men, F.K., Henzler, M., and

Himpsel, F.J. (1999) Phys. Rev. B, 59,

10356.

33 Bertrams, T., Brodde, A., and

Neddermeyer, H. (1994) J. Vac. Sci.

Technol. B, 12, 2122.

34 Rosenhahn, A., Schneider, J., Kandler, J.,

Becker, C., and Wandelt, K. (1999) Surf.

Sci., 433–435, 705.

35 Bardi, U. and Rovida, G. (1990) Surf. Sci.,

239, L511.

36 Maroutian, T., Degen, S., Becker, C.,

Wandelt, K., and Berndt, R. (2003) Phys.

Rev. B, 68, 154414.

37 Sebastian, I. and Neddermeyer, H. (2000)

Surf. Sci., 454, 771.

38 Hagendorf, C., Shantyr, R., Meinel, K.,

Schindler, K.-M., and Neddermeyer, H.

(2003) Surf. Sci., 532, 346.

39 Sebastian, I., Bertrams, T., Meinel, K., and

Neddermeyer, H. (1999) Faraday Discuss.,

114, 129.

40 Reiss, G., Vancea, J., Wiltmann,H., Zweck,

J., and Hoffmann, H. (1989) J. Appl. Phys.,

67, 1156.

41 Barbet, J., Garvin, A., Thimonier, J.,

Chauvin, J.-P., and Rocca-Serra, J. (1993)

Ultramicroscopy, 50, 355.

42 Klyachko, D. and Chen, D.M. (2000) Surf.

Sci., 446, 98.

43 Nilius, N., Kulawik, M., Rust, H.-P., and

Freund, H.-J. (2004) Surf. Sci., 572, 347.

44 Degen, S. (2005) Ph.D. thesis, Universit

€

Bonn.

45 Zein, N.E. and Antropov, V.P. (2002) Phys.

Rev. Lett., 89, 126402.

46 Castro, M. (1997) Int. J. Quant. Chem., 64,

223.

47 Henry, C.R. (1998) Surf. Sci. Rep., 31, 231.

48 Hamm, G., Barth, C., Becker, C., Wandelt,

K., and Henry, C.R. (2006) Phys. Rev. Lett.,

97, 126106.

49 Gritschneder, S., Becker, C., Wandelt, K.,

and Reichling, M. (2007) J. Am. Chem. Soc.,

129, 492.

50 Wiltner, A., Rosenhahn, A., Becker, C.,

Pervan, P., Milun, M., Kralj, M., and

Wandelt, K. (2001) Thin Solid Films,

400, 71.

51 Becker, C., Rosenhahn, A., Wiltner, A.,

von Bergmann, K., Schneider, J.,

Pervan, P., Ilun, M., Kralj, M., and

Wandelt, K. (2002) New J. Phys., 4, 75.1.

52 Robinson, V.N.E. and Robins, J.L. (1974)

Thin Solid Films, 20, 155.

53 Campbell, C.T. (1997) Surf. Sci. Rep.,

27,1.

54 Degen, S., Becker, C., and Wandelt, K.

(2004) J. Chem. Soc., Faraday Discuss.,

125, 343.

55 Becker, C., von Bergmann, K., Rosenhahn,

A., and Wandelt, K. (2001) Surf. Sci., 486,

L443.

56 Lehnert, A., Krupski, A., Degen, S.,

Franke, K., Decker, R., Rusponi, S.,

Kralj, M., Becker, C., Brune, H., and

Wandelt, K. (2006) Surf. Sci., 600, 1804.

57 Bennett, R.A., Newton, M.A., Smith, R.D.,

Evans, J., and Bowker, M. (2002) Mater. Sci.

Technol., 18, 710.

58 Bennett, R.A., Tarr, D.M., and

Mulheran, P.A. (2003) J. Phys: Condens.

Matter, 15, S3139.

59 Bowker, M. (2007) Phys. Chem. Chem.

Phys., 9, 3514.

60 Maeda, Y., Fuitani, T., Tsubota, S., and

Haruta, M. (2004) Surf. Sci., 562,1.

61 Sedona, F., Agnoli, S., Fanetti, M.,

Kholomanov, I., Cavaliere, E., Gavioli, L.,

and Granozzi, G. (2007) J. Phys. Chem. C,

111, 8024.

62 Berdunov, N., Mariotto, G., Balakrishnan,

K., Murphy, S., and Shvets, I.V. (2006)

Surf. Sci., 600, L287.

63 Hamm, G., Becker, C., and Henry, C.R.

(2006) Nanotechnology, 17, 1943.

References