Bowker M., Davies P.R. (Eds.) Scanning Tunneling Microscopy in Surface Science, Nanoscience and Catalysis
Подождите немного. Документ загружается.
lies on the periphery of the chain [62]. Chiral recognition occurs because there is also
a repulsive interaction between the amine groups of the two molecules and this is less
for the favored enantiomer than for the other [62].
1.3.4
Prochiral Molecules Interacting with Chiral Surfaces
In enantioselective catalysis, the key problem to overcome is the fact that the Gibbs
energy change from gas-phase prochiral reagent to gas-phase product is identical for
each enantiomeric product molecule. Hence, in the absence of any chiral influence
on the reaction, a racemic mixture of products should always ensue. By providing a
reaction pathway to one product that has a much lower activation barrier, the
selectivity can be skewed to give an enantiomeric excess of one product. One of
the most heavily researched examples of heterogeneous enantioselective catalysis is
the hydrogenation of b-ketoesters over Ni catalysts [28]. The simplest b-ketoester is
methylacetoacetate (MAA). This molecule is approximately planar and can adsorb via
either molecular face with equal probability. The hydrogenation reaction is believed
to occur via dissociative adsorption of H
2
on the metal surface and attack by H from
underneath the molecular plane of MAA. The stereochemistry of the chiral center
thus produced would be determined by which face of the prochiral reagent lies down
on the surface. To skew the reaction in an enantioselective direction, a clear
requirement seems to be to restrict the adsorption geometry to exclusively one
enantioface. In this respect, the coverage of chiral modifier is thought to be crucial. If
the coverage is too low, the formation of 1 : 1 complexes between modifier and
reactant could induce some enantioselectivity, but the adsorption of MAA on bare
metal sites would be expected to occur racemically. If the coverage is too high, there
Figure 1.17 The left-hand STM image shows homochiral adenine
rows aligned in low symmetry but mirror image azimuths on a
Cu(1 1 0) surface. On the right, adenine rows in the (1,2) direction
are decorated with double rows of S-phenylglycine molecules,
while no such interaction occurs with (1, 2) rows. (Adapted
with permission from Ref. [60]. Copyright 2000, Macmillan
Publishers Ltd.)
24
j
1 Chirality at Metal Surfaces
may be insufficient space for MAA to adsorb on the surface. Indeed, the sticking
probability ofMAA on Ni{1 1 1} coveredby high coveragesof (R,R)-tartaric acid [63] or
(S)-glutamic acid [64] is essentially zero. At intermediate tartaric acid coverages, Jones
and Baddeley showed that the adsorption of MAA caused a restructuring of the
tartrate adlayer and the formation of an ordered array of 1 : 1 tartrate:MAA complexes
(Figure 1.18). Interestingly, the geometry of each MAA molecule in the array
appeared identical and corresponding to the geometry required for the formation
of (R)-methyl-3-hydroxybutyrate – the product observed in excess in the catalytic
reactions [63].
1.4
Conclusions
Chirality at surfaces can be manifested in a number of forms including the intrinsic
chirality of the surface structure and even the induction of chirality via the
adsorption of achiral molecules onto achiral surfaces. The ability of STM to probe
surfaces on a local scale with atomic/molecular resolution has revolutionized the
understanding of these phenomena. Surfaces that are globally chiral either due to
their intrinsic structure or due to the adsorption of chiral molecules have been shown
by STM to establish control over the adsorption behavior of prochiral species. This
could have profound consequences for the understanding of the origin of homo-
chirality in life on Earth and in the development of new generations of heterogeneous
chiral catalysts that may, finally, make a substantial impact on the pharmaceutical
industry.
Figure 1.18 STM image (4 nm 4 nm) showing the 2D
cocrystalline structure consisting of an ordered array of 1 : 1
H-bonded complexes of (R,R)-tartrate and methylacetoacetate
species on Ni{1 1 1} giving a chiral h31|34istructure. (Adapted
with permission from Ref. [62]. Copyright 2002, Elsevier.)
1.4 Conclusions
j
25
References
1 Cahn, R.S., Ingold, C.K., and Prelog, V.
(1966) Angew. Chem. Int. Ed. Engl., 78, 385.
2 Knowles, W.S. (2002) Angew. Chem. Int.
Ed., 41, 1999.
3 Noyori, R. (2002) Angew. Chem. Int. Ed., 41,
2008.
4 Sharpless, K.B. (2002) Angew. Chem. Int.
Ed., 41, 2024.
5 Bradshaw, A.M. and Richardson, N.V.
(1996) Pure Appl. Chem., 68, 457.
6 Weckesser, J., De Vita, A., Barth, J.V., Cai,
C., and Kern, K. (2001) Phys. Rev. Lett., 87,
096101.
7 Bohringer, M., Morgenstern, K.,
Schneider, W.D., Berndt, R., Mauri, F.,
De Vita, A., and Car, R. (1999) Phys. Rev.
Lett., 83, 324.
8 Bohringer, M., Morgenstern, K.,
Schneider, W.D., and Berndt, R. (1999)
Angew. Chem. Int. Ed., 38, 821.
9 Blum, M.C., Cavar, E., Pivetta, M., Patthey,
F., and Schneider, W.D. (2005) Angew.
Chem. Int. Ed., 44, 5334.
10 Barlow, S.M., Kitching, K.J., Haq, S., and
Richardson, N.V. (1998) Surf. Sci., 401, 322.
11 Booth, N.A., Woodruff, D.P., Schaff, O.,
Giessel,T., Lindsay, R., Baumgartel, P., and
Bradshaw, A.M. (1998) Surf. Sci., 397, 258.
12 Chen, Q., Frankel, D.J., and Richardson,
N.V. (2002) Surf. Sci., 497, 37.
13 Rankin, R.B. and Sholl, D.S. (2004) Surf.
Sci., 548, 301.
14 Toomes, R.L., Kang, J.H., Woodruff, D.P.,
Polcik, M., Kittel, M., and Hoeft, J.T. (2003)
Surf. Sci., 522, L9.
15
Haq, S., Massey, A., Moslemzadeh, N.,
Robin, A., Barlow, S.M., and Raval, R.
(2007) Langmuir, 23, 10694.
16 Sayago, D.I., Polcik, M., Nisbet, G.,
Lamont, C.L.A., and Woodruff, D.P. (2005)
Surf. Sci., 590, 76.
17 Rankin, R.B. and Sholl, D.S. (2005) Surf.
Sci., 574, L1.
18 Lackinger, M., Griessl, S., Heckl, W.M.,
and Hietschold, M. (2002) Anal. Bioanal.
Chem., 374, 685.
19 England, C.D., Collins, G.E., Schuerlein,
T.J., and Armstrong, N.R. (1994) Interfac.
Des. Chem. Sensing, 561, 202.
20 Richardson, N.V. (2007) New J. Phys., 9,
395.
21 Bennett, R.A., Pang, C.L., Perkins, N.,
Smith, R.D., Morrall, P., Kvon, R.I., and
Bowker, M. (2002) J. Phys. Chem. B, 106,
4688.
22 Lackinger, M., Griessl, S., Heckl, W.M.,
and Hietschold, M. (2002) J. Phys. Chem. B,
106, 4482.
23 Gyarfas, B.J., Wiggins, B., Zosel, M., and
Hipps, K.W. (2005) Langmuir, 21, 919.
24 Yoshimoto, S., Narita, R., and Itaya, K.
(2002) Chem. Lett., 356.
25 Richardson, N.V. (2007) New J. Phys., 9,
395.
26 Schrock, M., Otero, R., Stojkovic, S.,
Hummelink, F., Gourdon, A., Laegsgaard,
E., Stensgaard, I., Joachim, C., and
Besenbacher, F. (2006) J. Phys. Chem. B,
110, 12835.
27 Webb, G. and Wells, P.B. (1992) Catal.
Today, 12, 319.
28 Izumi, Y. (1983) Adv. Catal., 32, 215.
29 Lorenzo, M.O., Haq, S., Bertrams, T.,
Murray, P., Raval, R., and Baddeley, C.J.
(1999) J. Phys. Chem. B, 103, 10661.
30 Lorenzo, M.O., Baddeley, C.J., Muryn, C.,
and Raval, R. (2000) Nature, 404, 376.
31 Barbosa, L. and Sautet, P. (2001) J. Am.
Chem. Soc., 123, 6639.
32 Fasel, R., Wider, J., Quitmann, C., Ernst,
K.H., and Greber, T. (2004) Angew. Chem.
Int. Ed., 43, 2853.
33 Humblot, V., Haq, S., Muryn, C., Hofer,
W.A., and Raval, R. (2002) J. Am. Chem.
Soc., 124, 503.
34 Hofer, W.A., Humblot, V., and Raval, R.
(2004) Surf. Sci., 554, 141.
35 Hermse, C.G.M., van Bavel, A.P., Jansen,
A.P.J., Barbosa, L., Sautet, P., and van
Santen, R.A. (2004) J. Phys. Chem. B, 108,
11035.
26
j
1 Chirality at Metal Surfaces
36 Hazen, R.M. and Sholl, D.S. (2003) Nat.
Mater., 2, 367.
37 Halasyamani, P.S. and Poeppelmeier, K.R.
(1998) Chem. Mater., 10, 2753.
38 McFadden, C.F., Cremer, P.S., and
Gellman, A.J. (1996) Langmuir, 12, 2483.
39 Zhao, X.Y. and Perry, S.S. (2004) J. Mol.
Catal. A – Chem., 216, 257.
40 Sholl, D.S., Asthagiri, A., and Power, T.D.
(2001) J. Phys. Chem. B, 105, 4771.
41 Jenkins, S.J. and Pratt, S.J. (2007) Surf. Sci.
Rep., 62, 373.
42 Attard, G.A. (2001) J. Phys. Chem. B, 105,
3158.
43 Horvath, J.D. and Gellman, A.J. (2001)
J. Am. Chem. Soc., 123, 7953.
44 Kuhnle, A., Linderoth, T.R., and
Besenbacher, F. (2006) J. Am. Chem. Soc.,
128, 1076.
45 Bowker, M., Poulston, S., Bennett, R.A.,
and Stone, P. (1998) J. Phys: Condens.
Matter, 10, 7713.
46 Chen, Q., Perry, C.C., Frederick, B.G.,
Murray, P.W., Haq, S., and Richardson,
N.V. (2000) Surf. Sci., 446, 63.
47 Chen, Q., Frankel, D.J., and Richardson,
N.V. (2001) Langmuir, 17, 8276.
48 Chen, Q. and Richardson, N.V. (2003) Prog.
Surf. Sci., 73, 59.
49 Pascual, J.I., Barth, J.V., Ceballos, G.,
Trimarchi, G., De Vita, A., Kern, K., and
Rust, H.P. (2004) J. Chem. Phys., 120
,
11367.
50 Zhao, X.Y., Gai, Z., Zhao, R.G., Yang, W.S.,
and Sakurai, T. (1999) Surf. Sci., 424,
L347.
51 Wang, H., Zhao, X.Y., and Yang, W.S.
(2000) Acta Phys. Sin., 49, 1316.
52 Wang, H., Zhao, X.Y., Zhao, R.G., and
Yang, W.S. (2001) Chin. Phys. Lett., 18,
445.
53 Zhao, X.Y. (2000) J. Am. Chem. Soc., 122,
12584.
54 Humblot, V., Lorenzo, M.O., Baddeley,
C.J., Haq, S., and Raval, R. (2004) J. Am.
Chem. Soc., 126, 6460.
55 Parschau, M., Romer, S., and Ernst, K.H.
(2004) J. Am. Chem. Soc., 126, 15398.
56 Green, M.M., Reidy, M.P., Johnson, R.J.,
Darling, G., Oleary, D.J., and Willson, G.
(1989) J. Am. Chem. Soc., 111, 6452.
57 Weigelt, S., Busse, C., Petersen, L.,
Rauls, E., Hammer, B., Gothelf, K.V.,
Besenbacher, F., and Linderoth, T.R.
(2006) Nat. Mater., 5, 112.
58 Kuhnle, A., Linderoth, T.R., Hammer, B.,
and Besenbacher, F. (2002) Nature, 415,
891.
59 Lingenfelder, M., Tomba, G., Costantini,
G., Ciacchi, L.C., De Vita, A., and
Kern, K. (2007) Angew. Chem. Int. Ed., 46,
4492.
60 Chen, Q., Frankel, D.J., and Richardson,
N.V. (2002) Langmuir, 18, 3219.
61 Chen, Q. and Richardson, N.V. (2003) Nat.
Mater., 2, 324.
62 Blankenburg, S. and Schmidt, W.G. (2007)
Phys. Rev. Lett., 99, 196107.
63 Jones, T.E. and Baddeley, C.J. (2002) Surf.
Sci., 519, 237.
64 Jones, T.E., Urquhart, M.E., and Baddeley,
C.J. (2005) Surf. Sci., 587, 69.
References
j
27
2
The Template Route to Nanostructured Model Catalysts
Conrad Becker and Klaus Wandelt
2.1
Introduction
Even though the majority of industrially produced chemicals are based on processes
that include at least one heterogeneously catalyzed step, the basic physical and
chemical properties that govern the action of a particular catalyst are far from being
understood. Surface science can play a major role in the elucidation of the molecular
principles that govern reactions on catalyst surfaces. This approach was brought to
attention of a wider public when the Nobel Prize for Chemistry in 2007 was awarded
to Gerhard Ertl for his studies of chemical processes on solid surfaces. The
investigation of the basic physical and chemical properties ofheterogeneous catalysts
down to the molecular and atomic level is, indeed, one of the prime goals of surface
science. On the route to ultimate understanding of heterogeneously catalyzed
reactions, two major obstacles need to be overcome, and these are commonly
referred to as the pressure gap and the materials gap. The pressure gap stems from
the fact that industrial catalysts are usually run at rather high pressures, whereas
surface science studies are generally undertaken in ultrahigh vacuum. In the recent
past, however, considerable progress has been made to close this gap by making
standard surface science techniques work at elevated pressures (see Refs [1–3]).
Therefore, it seems likely that the pressure gap is not the major obstacle for
understanding heterogeneously catalyzed reactions any more provided the structure
of the catalytic surface is well known. Here the materials gap enters the scene. It is
related to the fact that industrial catalysts possess a highly complex structure and
topology, which makes them inaccessible for surface science tools. This problem can
be overcome by the use of model catalysts of reduced complexity, that is, designed
model catalysts that mimic the nanoscopic structure of the industrial catalyst.
For many years, metal single-crystal surfaces have been – and are still – used for
this purpose. However, since industrial catalysts are often composed of oxide-
supported metal particles, extended metal surfaces are – even though of very low
complexity – rather unrealistic models. More realistic model catalysts are provided by
j
29
metal-on-oxide systems of low complexity. In this respect, it has turned out that thin
oxide films on metal substrates can be used as support for the preparation of such
systems. By depositing metal vapor on these films, model catalysts can be prepared,
which possess the big advantage over bulk oxide materials that they can be
investigated by surface science methods that rely on charged particles [4]. The big
drawback of such systems is, however, that in general their complexity is still rather
high. Owing to the nucleationprocess of metals on oxide surfaces, which is in general
controlled by the defects of the surface, it is quite dif ficult to reach a uniform particle
size and distribution since the defects are randomly distributed over the surface. This
renders investigation of particle size effects difficult because only the ensemble
average over a particular size distribution is experimentally accessible. The statistical
arrangement of defects on the surface not only leads to different capture zones, and,
as a consequence, to metal particles of different sizes (Figure 2.1a), but, in turn, also
results in a different reverse spillover from the support to the individual particles
during a catalytic reaction. This influences the reaction kinetics, which is closely
connected to the arrival rate of reactants on the particle [5]. One way to overcome all
these limitations is to use an ordered nanostructured model catalyst. The schematic
representation of such a model catalyst is provided in Figure 2.1b. It will provide
a uniform particle size and distribution, thus, enabling measurements on an
ensemble of identical particles.
The fabrication of such a system can be accomplished only by nanofabrication,
and different routes can be imagined in this context. We will focus in the following
section on the template-controlled growth of metal clusters on thin oxide films, which
has proven to give excellent results in terms of low complexity. This approach has
been successfully employed for metal-on-metal systems (for a comprehensive review
see [6]) and has recently been extended to metal growth on oxide films.
The key factor in this approach is the provision of suitable growth templates.
Before we turn to this point, we will discuss in which way surfaces may be thought of
as templates.
(a) (b)
Figure 2.1 Schematic representation of a random model catalyst
surface (a) and an idealized model catalyst (b). The black dots
correspond to metal particles, dotted circles to their capture
zones.
30
j
2 The Template Route to Nanostructured Model Catalysts
2.2
Surfaces as Two-Dimensional Templates
The concept of templates is well known in our contemporary life. Templates are
generally used to shape a product. This can be, for example, the mold of a church bell
or style sheet in office applications in computing. One common feature of templates
is that the information, which is used to shape a product, is intrinsically encoded in
the template. A good example in this context is a strand of DNA that sets the genetic
sequence of new strands. If we want to use surfaces to act as templates, for example,
in nanostructuring, we have to find a way to encode the information into the surface
structure. This approach is commonly used in the fabrication of semiconducting
devices, where lithography is employed to pattern the surface of a wafer. The pattern
thus created is then used as a template for the following manufacturing steps. This
approach, generally referred to as top-down approach, relies on the active pattern-
ing of the structure of the surface by external intervention. Even though this approach
is frequently and successfully applied in a variety of areas, it is subject to a number of
limitations in the context of model catalyst fabrication. First, if a lithographic process
is used, spatial resolution is limited by either the physical properties of the radiation
source employed for patterning (e.g., light, electron beams) or the chemical prop-
erties of the resist, which contains the pattern. In general, these limitations impose a
resolution limited to a few tenths of nanometers. However, very well ordered model
catalysts have been prepared in the past using this approach (see Figure 2.2).
To fully exploit the nanoscopic properties of materials, for example, in catalysis,
this structure size is much too large since it corresponds to a regime where the bulk
properties of materials still dominate. An alternative approach can be the patterning
of a surface by direct manipulation of atoms or molecules with the scanning
tunneling microscopy (STM) [8], which has been successfully employed in the past
Figure 2.2 STM image of platinum particles on an oxidized silicon
wafer. (Reproduced with permission from Ref. [7].)
2.2 Surfaces as Two-Dimensional Templates
j
31
in many cases. In this case, only the size of the objects, which are manipulated,
determines the resolution limit. However, it is still a top-down approach, which
requires the intervention of the operator. Furthermore, the speed of the assembling
process is rather slow and the patterning of surface areas on the micrometer scale
may be tedious or even impossible. We, therefore, have to conceive an alternative
pathway for creating nanopatterned templates. Fortunately, nature points the way for
that providing us with nanopatterned surfaces that are derived from the intrinsic
physical and chemical properties of the system itself. One possibility is to use the
self-organization of colloidal particles to form regular nanostructured arrays [9, 10].If
applied to model catalyst design, this approach has, however, the disadvantage that
the ligand shell, which stabilizes the particles, has to be removed without destroying
the particles [11, 12]. Nanostructuring by self-organization is also found to occur
in many metal surfaces. A particularly beautiful example is the herringbone or
chevron reconstruction of the Au(1 1 1) surface. This Au(1 1 1)-(
p
3 22) recon-
struction, which has been observed and characterized by a large number of surface
science techniques, provides us with a regular array of specific sites – the elbows of
the reconstruction – that can act as nucleation sites [6]. The fabrication of a template
using self-organization can be referred to as bottom-up approach since it relies on
the inherent physical and chemical properties of the system. The beauty of this
approach is that nature is doing the job; by choosing the components of the system,
the self-organization process is controlled by the intrinsic properties of the system.
The major obstacle that we are facing in this context is to set the controlling
parameters of the self-assembly process in such a way that a desired nanopattern
is generated. In the case of a pure metal surface such as Au(1 1 1), this of course
cannot be done since it is a one-component system with no control parameters.
However, other routes to nanostructured surfaces can be envisioned that provide an
experimental control on the size and symmetry of the pattern. Among these, thin
films provide the most promising route. This has been shown in a number of cases
for metal films, oxide films, and molecular films. In the case of metal films, the
difference in lattice parameters of substrate and film leads in many cases to the
formation of strain relief pattern, to compensate for the lattice mismatch. This
has been found for systems such as Ag/Pt(1 1 1) [13–15], Cu/Ru(0 0 0 1) [16],
Ag/Cu(1 1 1) [17–19], and many others [20]. Likewise, oxide films can form regular
nanopattern depending on the accommodation of the film to the substrate, and we
will come back to this later. Finally, molecular films provide access to a whole new
class of nanopatterned surfaces since in this case the intermolecular interactions of
the molecules along with the molecule –substrate interaction play a decisive role
in the pattern formation [21]. The latter case will thus provide us with interesting
properties for the growth of nanostructures on surfaces but will have only a limited
application for the fabrication of model catalysts because these should be, in general,
metal-on-oxide systems.
The nanopatterning of a surface is, however, only the first step toward a surface
template. We will further need a specific property of the surface, which is encoded
into the pattern. Only this will assure that the template will control the subsequent
nucleation or growth processes. In many cases, this specific property will be
32
j
2 The Template Route to Nanostructured Model Catalysts
a particularly strong interaction (adsorption energy) at certain points of the sur-
face [22]. This will provide us with a given density of nucleation sites that is
determined by the nanostructure of the template. If we want to use these sites as
a growth template, we have to control the kinetics of the process to reach nucleation
on the template sites. Besides the adsorption energy of the particles on the traps, the
flux of the impinging particles and the temperature of the substrate play a decisive
role [23]. The latter has an important impact on the mobility of the particles and, thus,
their ability to reach the traps. This has been illustrated in the work of Pr
evot et al.
on the nucleation and growth of Au on the reconstructed N/Cu(1 0 0) surface [24].
This surface provides a square lattice of nucleation sites, which constitute a template
for the ordered growth of Au clusters [25]. The kinetic Monte-Carlo (KMC) simula-
tions shown in Figure 2.3 illustrate the importance of growth temperature for the
action of the template. Only at intermediate temperatures (e.g., 240 K) is the growth
perfectly template controlled. At this temperature, the diffusion of Au atoms on the
surface is fast enough so that the nucleation sites are actually reached. At higher
temperature also, the template is operative, all islands are found on template sites,
but not all template sites are covered by islands, which is due to higher mobility of
Au atoms. At lower temperature, the number density of islands exceeds that of the
template sites indicating that besides the template-controlled (heterogeneous)
nucleation, homogeneous nucleation also takes place.
A more detailed picture of the temperature dependence of the growth is given in
Figure 2.4, where the island density is plotted as a function of temperature. It can be
seen that only in the temperature range from 207 to 288 K the growth is perfectly
template controlled and the number of islands matches the number of available
nucleation sites. This illustrates the importance of kinetic control for the creation of
ordered model catalysts by a template-controlled process. Obviously, there has to be
a subtle balance between the adatom mobility on the surface and the density of
template sites (traps) to allow a template-controlled growth. We will show more
examples of this phenomenon below.
That not only an increased interaction energy at the traps can be responsible for
a template-controlled growth but also an anisotropy of the surface diffusion
Figure 2.3 Kinetic Monte-Carlo simulations of Au growth on
N/Cu(1 0 0) for three different temperatures. From left to right,
T ¼180 K, T ¼240 K, and T ¼300 K. The coverage is 0.11 ML.
(Reproduced with permission from Ref. [24].)
2.2 Surfaces as Two-Dimensional Templates
j
33