Nanoporous gold: Novel Catalytic and Sensor ...

Nanoporous gold: Novel Catalytic and Sensor Applications

Dissertation zur Erlangung des Doktorgrades der Naturwissenschaften - Dr. rer. nat. -

Vorgelegt dem Promotionsausschuss des Fachbereichs 2 (Chemie/Biologie) der Universität Bremen von Arne Wittstock Bremen, im April 2010

1. Gutachter: Prof. Dr. Marcus Bäumer (Universität Bremen) 2. Gutachter: PD Dr. Michael Gottfried (Universität Erlangen/Nürnberg)

Hiermit erkläre ich, dass ich die vorliegende Arbeit selbständig angefertigt und keine außer den angegebenen Hilfsmitteln verwendet habe. Diese Arbeit wurde nicht vorher an anderer Stelle eingereicht.

Bremen, den

This thesis is based on the following papers, which will be referred to in the text by their roman numerals: I) Nanoporous Au: An Unsupported Pure Gold Catalyst? Wittstock, A.; Neumann, B.; Schaefer, A.; Dumbuya, K.; Kübel, C.; Biener, M. M.; Zielasek, V.; Steinrück, H.-P.; Gottfried, J. M.; Biener, J.; Hamza, A.; Bäumer, M The Journal of Physical Chemistry C, 2009, 113 (14), 5593-5600, http://pubs.acs.org/doi/abs/10.1021/jp808185v

II) Novel nanoporous gold catalysts for green chemistry: selective gas-phase oxidative coupling of methanol at low temperature Wittstock, A.; Zielasek, V.; Biener, J.; Friend, C. M.; Bäumer, M Science, 2010, 327 (5963), 319-322. http://www.sciencemag.org/cgi/content/full/327/5963/319 III) Surface-chemistry-driven actuation in nanoporous gold Biener, J.; Wittstock, A.; Zepeda-Ruiz, L. A.; Biener, M. M.; Zielasek, V.; Kramer, D.; Viswanath, R. N.; Weissmuller, J.; Baumer, M.; Hamza, A. V. Nature Materials 2009, 8 (1), 47-51 http://www.nature.com/nmat/journal/v8/n1/pdf/nmat2335.pdf

IV) Surface Chemistry in Nanoscale Materials Biener, J.; Wittstock, A.; Baumann, T.; Weissmüller, J.; Bäumer, M.; Hamza, A. Materials 2009, 2 (4), 2404-2428 http://www.mdpi.com/1996-1944/2/4/2404/ V) Effect of surface chemistry on the stability of gold nanostructures Biener, J.; Wittstock, A.; Biener, M.M.; Nowitzki, T.; Hamza, A.V.; and Bäumer, M. Submitted to Langmuir in February 2010

VI) Ultralow loading Pt nanocatalysts prepared by atomic layer deposition on carbon aerogels King, J. S.; Wittstock, A.; Biener, J.; Kucheyev, S. O.; Wang, Y. M.; Baumann, T. F.; Giri, S. K.; Hamza, A. V.; Bäumer, M.; Bent, S. F. Nano Letters, 2008, 8 (8), 2405-2409 http://pubs.acs.org/doi/abs/10.1021/nl801299z

Statement regarding my contribution to the published work: All stated publications are unexceptionally based on the collaboration of several researches. Nevertheless, for publication (I) I was responsible for writing the manuscript, the preparation of samples, and for the catalytic characterization of samples (Björn Neumann as a research student being strongly involved in the experiments). Regarding publication (II) I was strongly involved in writing the manuscript and performed the experimental work. For publication (III) I performed the experimental work and was only to a minor extend involved in the preparation of the manuscript (focus was on artwork). In case of the review type publication (IV) I was responsible for two chapters of the manuscript. For publications(V) I performed the experimental work on the annealing of nanoporous gold and provided some art work to the manuscript. In publication (VI) I performed the catalytic characterization of samples, was involved in the data evaluation (TEM and RBS) and contributed to the preparation of the manuscript. .

Zusammenfassung Nanoporöses Gold (np-Au) wurde untersucht hinsichtlich seiner Anwendungen für die heterogene Gasphasenkatalyse sowie seiner Anwendungen in der Aktorik bzw. Sensorik. Darüber hinaus wurde der Einfluss von Adsorbaten auf die Stabilität der mesoporösen Struktur des Goldes untersucht. Neben der zielgerichteten Herstellung des np-Au lag der Schwerpunkt hierbei auf dem Verständnis der chemischen Prozesse an der Oberfläche des Materials. Nanoporöses Gold kann durch Entlegieren einer Gold-Silber Legierung hergestellt werden. Entlegieren bezeichnet einen Prozess, bei dem selektiv die unedlere Komponente (hier Ag) aus der (bi-) metallischen Legierung fast vollständig herausgelöst wird. Durch Selbststrukturierung entsteht eine selbsttragende mesoporöse Gold-Struktur (> 99 at% Gold), deren Stege und Poren nur wenige zehn Nanometer groß sind. Das hohe Oberflächen-zuVolumen Verhältnis und die stark gekrümmte Oberfläche der Stege dieses Materials führen zu einzigartigen Materialeigenschaften. So kann CO bereits bei Temperaturen von – 20 °C mit molekularem Sauerstoff oxidiert werden. Bei der Untersuchung von Proben mit unterschiedlich hohen Restsilbergehalten stellte sich in diesem Zusammenhang ein deutlicher Einfluss des Silbers heraus, das nach dem Entlegieren im Material verbleibt. Die beobachtete Charakteristik und die Kinetik der katalysierten Oxidation von CO legen eine entscheidende Rolle des Silbers bei der Aktivierung von molekularem Sauerstoff nahe. Die schwache Bindung von Adsorbaten auf Goldoberflächen begründet die katalytische Aktivität bei niedrigen Temperaturen, darüber hinaus kann dieser Umstand zum Erzielen besonders hoher Selektivität genutzt werden. Als Beispiel wurde die katalytische Oxidation von Methanol untersucht. Bereits bei 20 °C konnte die Oxidation von Methanol zum Methylformiat durchgeführt werden. Die Selektivität bei dieser Umsetzung lag bei nahezu 100 %. Die thermodynamisch begünstigte Totaloxidation (Verbrennung) zum CO2 spielte nur eine untergeordnete Rolle, auch bei erhöhtem Umsatz bei 80 °C Reaktionstemperatur machte diese weniger als 3 % des Gesamtumsatzes aus. Die Reaktivität und der Mechanismus der Oberflächenreaktion, insbesondere der Umstand, dass ausschließlich der Ester gebildet wurde, können auf der Grundlage von Modellexperimenten unter UHV Bedingungen verstanden werden. Lediglich die Aktivierung des molekularen Sauerstoffs kann nicht auf die Aktivität des Goldes allein zurückgeführt werden. Jedoch auch hier stellte sich ein deutlicher Einfluss des Silbers heraus, der auf die Aktivierung des molekularen Sauerstoffs zurückzuführen ist.

Das hohe Oberflächen-zu-Volumen Verhältnis von np-Au eröffnet neben der Katalyse viele weitere Anwendungsfelder. So ist der Oberflächenstress von Metallen durch die Adsorption von Substanzen beeinflussbar. Im Falle von np-Au kann der hohe Anteil der Oberfläche am Gesamtmaterial zu einer Beeinflussung der gesamten Materialeigenschaften führen. Die Adsorption von atomarem Sauerstoff (Oads) verdeutlicht dieses beispielhaft. Aufgrund der geringen Dissoziationsrate von molekularem Sauerstoff wurde Ozon als Quelle für atomar auf der Oberfläche adsorbierten Sauerstoff gewählt. Die Bedeckung der Oberfläche des np-Au mit Oads durch Ozon Exposition bzw. dessen Entfernung durch Exposition von Kohlenmonoxid führte zu makroskopisch detektierbaren reversiblen Schrumpfen bzw. Dehnung des Materials von bis zu 0.5 % der Seitenlänge. Dies ist ein Wert, der vergleichbar mit kommerziell erhältlichen Piezo-Keramiken ist. Die Möglichkeit, chemische Energie direkt in mechanische Arbeit umzusetzen, ohne vorher elektrische Energie zu produzieren, stellt ein neuartiges Aktuator-Konzept dar und kann ebenfalls für sensorische Zwecke genutzt werden. Ein weiterer wichtiger Aspekt der Oberflächenchemie bzw. der Adsorbate ist ihre Rolle für die Stabilisierung der Nanostruktur und dessen gezielte Variation. Die Instabilität und Vergröberung der Strukturen ist durch die stark gekrümmte Oberfläche der Stege und die Oberflächendiffusion der Goldatome bedingt. Diese ist stark von der Temperatur abhängig und kann genutzt werden, um die Strukturen durch Tempern bei Temperaturen oberhalb von 150 °C zu vergröbern. Ein entscheidender Parameter hierbei ist der Einfluss der Adsorbate auf die Diffusionsgeschwindigkeit der Oberflächenatome und somit auf die Stabilität der Struktur. Als Beispiel wurde wiederum die Bedeckung der Oberfläche mit atomarem Sauerstoff durch Exposition mit Ozon gewählt. Proben, die unter einer ozonhaltigen Atmosphäre getempert wurden zeigten unterhalb von 500 K eine deutlich erhöhte Strukturstabilität. Im Vergleich zu den unter einer Inertgasatmosphäre (He) getemperten Proben, deren Stege auf die doppelte Größe anwuchsen, blieben die durch Sauerstoff stabilisierten Stege in ihrer Größe unverändert. Des Weiteren können durch Ausnutzung des limitierten Stofftransportes in die Poren des Materials gezielt Gradienten der Poren- und Steggrößen erreicht werden. Das Konzept der adsorbat-induzierten Strukturkontrolle unterstreicht den Einfluss der Oberfläche und der Chemie an dieser auf die Materialeigenschaften.

Content 1

Introduction.......................................................................................................................12

2

Nanoporous Gold ..............................................................................................................14

3

2.1

Creating nanoporosity ...............................................................................................15

2.2

Surface-porosity related phenomenon.......................................................................17

Catalysis by nanoporous gold ...........................................................................................19 3.1

Kinetics of CO oxidation and the influence of residual Ag ......................................20

3.2

Gas phase oxidative coupling of methanol ...............................................................24

4

Actuation by nanoporous gold ..........................................................................................27

5

Tailoring structures of nanoporous gold ..........................................................................29

6

Outlook: Surface engineering and doping of np-Au .........................................................32

7

Summary............................................................................................................................34

8

Experimental .....................................................................................................................35

9

8.1

Synthesis of master alloy and Preparation of np-Au.................................................35

8.2

Catalytic investigation...............................................................................................37

8.3

Measuring actuation of np-Au...................................................................................40

8.4

Annealing Experiments .............................................................................................41

References .........................................................................................................................42

Curriculum Vitae.......................................................................................................................46 Papers........................................................................................................................................48

Introduction

1

Introduction

New materials are sought for applications in catalysis, sensing, optics, electronics and other emerging fields. With the event of nanotechnology, innovations greatly profit from the access to new nano-sized materials with novel properties of materials simply resulting from the reduction of size. As a function of size, the chemical, optical, and mechanical behaviour can be tuned allowing adapting the materials specifically to their application. In recent years the field of such “nano-materials” steadily grew. Examples reflecting their technological importance are nano-particles in catalysis, ultra thin films in semiconductor technology or porous materials, such as metal organic frameworks for energy storage. The ultimate aim is to structure and design materials for a specific purpose on the scale of a few atoms. The length scale determines the electronic, catalytic, optical, and mechanical properties. An example is gold. Despite or due to its value and stability the technological application of gold was for a long time mainly restricted to electrically conductive coatings for electronics, optical coatings, or brazing alloys.(1) However, when in the form of nano-particles with diameters of less than 5 nm gold can show remarkable activity for catalytic applications. This fact was first reported by Haruta and Hutchings in the middle 80th last century.(2-3) Today, gold catalysis is on the very verge to commercial catalytic application.(4) This effect is based on the reduced length scale of nano particles. One reason is the effect of length scale on the surface to volume (S/V) ratio. For example, the S/V ratio of a sphere with a radius of 1 m increases by nine orders of magnitude when the radius is reduced to 1 nm. Hence, when going from “macro” to “nano” several orders of magnitude are crossed. Materials on the nano scale are to a large extent governed by their surface properties. The surface of a material is an interface and atoms at this interface are in a different electronic (chemical) state, owing to their lower coordination than within the bulk. Accordingly, monolithic nanoporous materials recently attracted considerable interest. Examples, are aerogels (5), metal organic frameworks (6), ceramics (7), and nanoporous metals (5, 7-8). The latter group certainly stands for a particularly interesting class of materials in the context of catalysis, electronics, and optics. Today, different metals, such as Pt, Pd, Au, Ag, Cu etc. are available in a nanoporous form (8), thus opening the door for a variety of applications.

Page 12

Introduction

In this thesis I will demonstrate that nanoporous gold follows this trend. It is this delicate balance between gold’s inertness and surface related phenomena which creates a range of new and beforehand unseen applications. I will discuss the surface chemistry of np-Au in terms of catalytic activity and selectivity; this is the impact of surface chemistry on the surrounding gases. In addition, I will discuss the influence of surface chemistry on the surface stress and mobility e.g. surface diffusion of surface atoms; this is the impact of surface chemistry on the material itself. We will see that:

o

np-Au exhibits remarkable catalytic activity for oxidation reactions at low temperatures such as the oxidation of CO and methanol, for example

o

selectivity close to 100 % can be achieved for the oxidation of methanol when using np-Au. This example shows that high selectivity and durability can be achieved taking advantage of the relatively weak interaction of molecules with gold surfaces

o

the surface chemistry of np-Au correlates with that of pure (i.e. single crystalline) gold surfaces. Reactivity and selectivity can be anticipated based on mechanistic understanding of reactivity making np-Au a predictable catalyst

o

(local) tuning of surface properties by residual Ag regulates availability of active oxygen and thus the oxidative power of the material can be tuned

o

the surface chemistry can lead to a change of surface stress. In case of np-Au which is distinguished by its high surface to volume (S/V) ratio this effect can lead to a macroscopic strain response of up to 0.5 %.

o

the structure of np-Au can be tuned for a specific use. Taking advantage of surface chemistry i.e. bonding of adsorbates such as oxygen the structure can be stabilized or even gradients can be induced

Page 13

Nanoporous Gold

2

Nanoporous Gold

Nanoporous gold is generated by dealloying of a gold alloy containing a less noble constituent such as silver or copper. The dealloying process by itself is an ancient technology and was already used by the Incan people some hundred years ago. They dealloyed (etched) and glazed the surface of a gold-copper alloy giving the illusion of shining bulk gold. This technique called depletion gilding (mis-en-colour) was used throughout the centuries in Europe, too.(9-12) The recent focus on “nano” and the development of sophisticated microscopical techniques such as electron microscopy since the middle of the last century shifted the focus on dealloying techniques from corrosion studies towards the material science. Since the 1960s Pickering and Swann investigated the corrosion of Au alloys, presenting first transmission electron micrographs (TEMs) revealing the porous morphology of np-Au. (1314) Later in 1979 Forty and co-workers were able to obtain TEMs of monolithic np-Au generated by etching of a Ag-Au alloy in nitric acid.(12, 15) These images proved the 3 dimensional porous structure of np-Au in the range of only some tens of nanometers. Figure 1: SEM showing the 3 dimensional structure of np-Au generated by etching of a Au-Ag alloy by concentrated nitric acid. The resulting structure can be characterized as (opencell) metal foam with ligaments and pores in the range of a few tens of nanometers. In recent years the number of publications concerning nanoporous gold steadily increased by about 30 percent per year, from 11 publications in 2001 to more than 100 publications in the year 2009 (ISI Web of Knowledge). Today, nanoporous gold is used in the field of catalysis, sensors, actuators, optics, and many more. First review articles appearing in 2009 provided an overview about the various aspects of nanoporous gold. (5, 7, 16-17) The interdisciplinary nature of this material ranging from its outstanding material properties to surface chemistry related phenomena poses a challenge when describing and explaining its exceptional properties. Here, we will discuss np-Au and its potential uses from the perspective of surface chemistry and surface induced phenomena. In the chapters 4 and 5 we will further discuss the impact of surface chemistry on mechanical properties.

Page 14

Nanoporous Gold

2.1 Creating nanoporosity The starting compounds for the generation of np-Au are intermetallics of gold, containing at least one less noble metal. Most prominent materials are the alloys of Au-Ag (10, 12), AuCu (18-19), and Au-Al (8, 20). All three starting materials can be processed resulting in a porous and monolithic gold material. The dealloying procedure and the resulting material though differ with respect to porosity and composition. The reason lies within the different phase behaviour of the various intermetallics and the different tendency to passivate during the dealloying procedure. For example, the intermetallics of Au and Al (Al2Au, AlAu) show different characteristics during etching (21). Most importantly, Al and Cu tend to passivate during dealloying leading to incomplete leaching in the order of more than 20 at% residual Al and Cu within the np-Au, respectively.(8, 18, 20) On the contrary, the alloy of Au and Ag is a so called “solid solution” and solidifies in just one fcc crystal phase, regardless of the composition of the alloy, and can be de-alloyed to > 99 percent.(1, 22) The resulting porous structure of the np-Au is homogenous throughout the entire material. When dealing with npAu it is commonly generated from a Au-Ag alloy owing to the obvious advantages of this preparation route. The driving force for dealloying is the difference in standard potential ΔE of the constituents A and B as described by the change of the free enthalpy ΔG which is a function of the standard potential ΔE0 and the molar fraction x of the particular component A and B1 (additional parameters are the universal gas constant R, the temperature T, the number of electrons transferred z and the Faraday constant F):

ΔG A = ΔG A0 − RT ln( x A ) ⇔ ΔE A = ΔE A0 − and

ΔG B = ΔG B0 − RT ln( x B ) ⇔ ΔE B = ΔE B0 −

RT ln( x A ) zF

RT ln( x B ) . zF

(equ. 1a)

(equ. 1b)

Accordingly, the driving force (potential ΔE) can be expressed by: 0 ΔE dealloying = ΔE A − ΔE B = ΔE AB −

x RT ln( A ) zF xB

(equ. 2)

It is thus the different electrochemical potential of both constituents which is the prerequisite for the dealloying procedure. In this context the component A is the nobler component, such as Au, and the component B is the less noble component, such as Ag.

1

The activity of the particular component within the surrounding electrolyte is assumed to be negligible small i.e. zero

Page 15

Nanoporous Gold

However, it is also clear that the fraction of the less noble constituent cannot become zero and the dealloying cannot proceed to 100 %. Besides thermodynamic considerations two additional parameters are essential for dealloying, the so called “parting limit” (23-24) and the “critical potential” (25-27). Only alloys of a certain range of composition can be dealloyed. The fact that for example the fraction of Ag in the starting alloy has to be more than 55 at% is a consequence of the so called parting limit. For lower fractions of Ag the dealloying does not proceed. This effect can be understood in terms of a coordination effect; for low Ag fractions within the material the coordination by Au atoms increases. A high coordination leads to passivation of the particular atom. Artymowicz et al. recently presented a kinetic Monte-Carlo simulation including percolation theory.(28) They showed that the coordination threshold of 9 leads to the experimentally found parting limit of about 55 at%. Another parameter for de-alloying is the so called critical potential (Ec). This effect can be understood on the bases of the commonly known over-potential in electrochemistry. The theoretical potential as described in equ. 2 does not suffice to induce bulk corrosion. The kinetics i.e. the activation barrier for surface diffusion of the more noble Au atoms prevents corrosion throughout the bulk material. Hence, only at a certain potential (higher than the thermodynamic threshold) the current that is the dissolution of Ag atoms rapidly increases. This potential Ec depends on the composition of the starting alloy, the electrolyte, and further additives like halides.(25-27) None of these parameters can solely explain the formation of a 3-dimensional sponge-like morphology. Different theories based on atomistic models of dealloying have been proposed; Erlebacher and co-workers presented a model which reproduces the morphology of np-Au. (7, 29-30) This model is based on three competing processes, the electrochemical dissolution of the less noble constituent (Ag), the surface diffusion of the noble constituent (Au), and capillary action.(7, 30) A schematic drawing of the mechanism is shown in Figure 2. While the Ag atoms are dissolved in a layer-by-layer fashion, the gold atoms can diffuse along the surface and form islands. Further dissolution of Ag atoms leads to erosion of islands so that ligaments are formed. On the basis of that rather simple model trends in de-alloying characteristics can be satisfactorily explained as for example the tendency to form larger ligaments when halides are added to the electrolyte during dealloying (27) or smaller ligaments are formed when the temperature is lowered (31). Both parameters affect the mobility of the gold atoms and thus the size of islands which are initially formed by surface diffusion.

Page 16

Nanoporous Gold

Figure 2: Atomistic model of de-alloying for a fcc gold-silver alloy (sections (a) and (b) reproduced from ref. (7)). The sections (a) and (b) show illustrations from a kinetic Monte Carlo simulation, the greyish Ag atoms are dissolved in a layer by layer fashion. The diffusion of the Au atoms along the surface leads to formation of islands, as can be seen in section (b). Single steps of the dealloying mechanism are depicted in section (c).

2.2 Surface-porosity related phenomenon When dealing with “nano” sized materials new material properties and applications are discussed and found.(5, 7) But why and to what extend should a material such as gold “change” when becoming nano-sized? The reason lies within the reduced characteristic length and the increased surface2 to volume ratio of the objects. When moving from objects with characteristic length-scales (e.g. diameter, edge length) in the order of centimetres or millimetres to the nanometer scale, the surface to volume ratio is increased by a factor of more than one million. Accordingly, nanosized object are to a large extend determined by their surface chemical properties. In the following, two effects which are related to surfaces and low coordination of atoms are exemplarily discussed. The charge redistribution due to low coordination of surface atoms gives rise to surface stress.(32-36) The excess charge from unsaturated bonds is redistributed into in-plane bonds which are strengthened or weekend, depending on whether the additional charge is distributed into bonding or antibonding states. Compared to bond length of atoms within the bulk the distance between surface atoms can thus be altered leading to surface stress. For compressive stress the surface tends to expand compared to the bulk whereas for tensile stress the surface tends to shrink.(34) A typical consequence of surface stress is a reconstruction of the surface. Giving an example, the Au(111) surface reconstructs in the 2

The term surface is used with the same meaning as the term interface. In reality, surfaces are always interfaces.

Page 17

Nanoporous Gold

commonly known Herringbone reconstruction.(37-38) The tensile surface stress of the Au(111) surface is compensated by incorporation of about 4 % additional atoms into the surface as compared to the bulk. Recently, several investigations showed that in case of nanoporous metals (np-Au, np-Pt) a change of surface stress can lead to macroscopically detectable strain of the entire material, an effect formerly known only for piezo ceramics.(32-33, 39-40) An additional effect is exploited in catalysis. The low coordination of surface atoms enhances the reactivity of (metal) surfaces. The desorption enthalpy of CO on gold surfaces is (amongst others) a function of the roughness that is the coordination number of surface atoms.(41) In general there are two ways in which the coordination of a surface atom can influence the interaction with adsorbates, one is electronic (adsorption enthalpy) and the other one is geometric (activation barrier), in reality both effects are hardly separated. The concept of d-band centre was first introduced from Norskov for the description of electronic effects on the interaction of gases with metallic surfaces.(42) The d-band centre is defined as the first moment of the density of d-states. The position of the d-band centre with respect to the Fermi level affects the ability of a metal surface atom to form a bond with an adsorbate. For example, transition metals tend to have higher d-states in case of low coordination numbers (kink-, edge-sites). Owing to this fact these atoms interact more strongly with adsorbates than atoms in a closed packed surface.(43) Additionally, the geometry of the substrate provides specific adsorption sites which play a crucial role for activation of adsorbed molecules (lowering the activation barrier).(44) Both effects contribute to the catalytic performance of a metal. In case of gold the role of low coordination even determines whether the surface shows any catalytic activity.(45)

Figure 3: (a) The coordination of atoms depends on their location. In general, the stability of atoms decreases with the coordination number (CN) (image courtesy of J. Biener). (b) The fraction of surface atoms rapidly increases when the characteristic length is decreased. This is exemplary shown for a straight ligament (pillar) (numbers are based on a geometric consideration)

Page 18

Catalysis by nanoporous gold

3


The first experimental results regarding the catalytic potential of np-Au were published in 2006 by Bäumer and co-workers (46) and only a short period of time later by Ding and coworkers (47). They investigated the catalytic oxidation of carbon monoxide (CO) with molecular oxygen (O2). Both groups detected independently high catalytic activity at temperatures even below 0 °C. This finding was spectacular and unexpected so far, since all previous catalytically active3 gold catalysts were comprised of nanoparticles on a reducible support, such as TiO2.(48-49) Not only is nanoporous gold catalytically active without any oxidic (reducible) support material but also the size of the ligaments (30 to 40 nm) is about ten times higher than the upper size limit of 5 nm, which is found to be essential for high catalytic activity of nanoparticles. (48-49) In the meantime, additional catalytic applications of np-Au evolved, such as for example, the liquid phase oxidation of glucose (50), the electro-oxidation of methanol (51-52) or oxygen reduction reaction like in fuel cell application. (53-54) Yet, the presented work on the gas-phase oxidative coupling of methanol (22) is the first work showing the potential of np-Au as a catalyst for more complex gas-phase reactions. The origin of the catalytic activity, especially the activation of molecular oxygen remained a matter of discussion. Only recently Ding et al. repeatedly argued that low coordinated gold surface atoms are the source for activated molecular oxygen.(16) However, in the following we will see that most part of the surface chemistry of np-Au can be explained by that of pure gold, while small quantities of Ag remaining within the material after dealloying play a crucial role for the activation of molecular oxygen.

3

With respect to heterogeneous (gas-solid) reactions

Page 19


3.1 Kinetics of CO oxidation and the influence of residual Ag (Relevant papers: I & IV) The catalytic oxidation of CO with molecular oxygen is a relevant reaction from an application as well as from a mechanistic point of view. The oxidation of harmful CO to the nonhazardous CO2 plays an important role in exhaust gas cleaning (e.g. in car catalysts). The superior activity of gold for low temperature oxidation of CO and hydrocarbons compared to typical metals used in car catalysts (Pt, Pd, Rh) initiated intense research efforts.(55-57) Today, first commercial catalysts are available (NS GoldTM from Nanostella Inc.). Additionally, the selectivity of gold to selectively oxidize CO in the presence of H2 gives rise to application in hydrogen fuel cells.(58) Besides its commercial potential, the oxidation of CO was widely used for investigations of the surface chemistry for example under ultra high vacuum (UHV) conditions. Due to its widespread use it is also dubbed as the “fruit fly” of surface chemistry. Especially, the comparatively simple geometry of the CO molecule, the fact that only one product (CO2) can be formed, and the detection/surveillance by IR spectroscopy renders the CO oxidation particularly useful to determine and asses fundamental reactivity of surfaces for oxidation reactions. For these reasons, the CO oxidation was chosen as a first test reaction to investigate the catalytic activity of np-Au . Disks of np-Au (300 µm thick and 4.8 mm in diameter) were investigated under continuous flow conditions (details see ref. (59)). To activate the sample, it was held under elevated temperatures (> 60 °C) under a gas stream containing CO and O2. After 2 to 3 hours the activity steadily increased and eventually reached a steady state. The fact that no activation could be obtained when no CO was present points towards a reaction as the case may be reduction of the surface which leads to the catalytically active state. On the other hand when samples were stored for a longer period of time (several weeks) longer activation periods were necessary, an observation which implies a concurrent cleaning procedure. However, x-ray photoelectron spectroscopy (XPS) at the outer surface of the disks did not reveal a distinct change of chemical state of Au and Ag, respectively, after and before activation. The transfer of the samples from the reactor into the UHV chamber as well as the fact that inherently only the very outer surface of the sample can be investigated by XPS renders this technique difficult to asses the active state of the catalyst surface. After activation of the sample the catalytic performance was investigated in two ways: •

Catalytic activity as a function of the partial pressure of one reactant

•

The overall catalytic activity under stoichiometric conditions (CO:O2 is 2:1)

Page 20


Figure 4: Catalytic activity of np-Au: The oxidation of CO with molecular oxygen at 40 °C. In section (a) the production of CO2 increases with the supply of CO and O2 but is eventually limited by oxygen supply. The inset: XP spectrum of the Au 4d and Ag 3d region proving the presence of Ag on the surface of the catalyst. (b) Activity of samples containing different amounts of residual silver (Ag contents were determined by AAS). The reactants were supplied in a stoichiometric ratio of 2:1 (CO:O2). For elevated Ag contents (here 11.4 at% (blue diamonds)) the catalytic activity is increased by nearly a factor of two as compared to samples containing 0.6 at% (red squares) and 0.7 at% res. Ag (black triangles), respectively.

Exemplarily, the conversion as a function of the CO partial pressure for different fixed oxygen partial pressures is shown in Figure 4 (a). In general, the conversion increased with increasing supply of reactants. No poisoning by either reactant was observed. This fact points towards the absence of competition for adsorption on the surface of the catalyst which implies that the interaction with the surface is weak (low surface coverage) and most importantly the adsorption sites for CO and O2 may be different. To compare and assess the overall catalytic activity of different samples, in particular with respect to the content of residual silver, the reactants were supplied in a stoichiometric ratio. As expected, the activity increased with increasing supply of reactants (see Figure 4 (b)). However, the production of CO2 could be considerably increased when samples were prepared containing higher fractions of residual Ag e.g. up to 11 at%. This indicates that (a) Ag plays a role in the catalytic cycle and (b) its presence is beneficial in case of the oxidation of CO. In order to determine the kinetic parameters of the reaction double logarithmic plots of the activity as a function of the partial pressure where drawn (see Figure 5). The apparent reaction order of O2 was found to be close to 0.5 for all investigated samples, while the reaction order of CO2 varied between 0.7 and 1. When taking mass transport phenomena into account (59-60) this corresponds to an actual reaction order of 0 for O2 and between ~ 0 and 1 for CO. These numbers reflect the interaction of the reactants with the catalyst surface. A reaction order of zero implies a fast saturation of adsorption sites for oxygen, either owing Page 21


to strong interaction (large adsorption enthalpy) or a very limited number of active sites. On the other hand a reaction order close to 1 as was measured for CO, inflicts that the adsorption i.e. the supply of CO from the gas phase is rate limiting, either because of a weak interaction (low adsorption enthalpy) or a large number of available sites for adsorption. However, with respect to different amounts of Ag within the material no substantial change in the kinetics was observed. This in turn implies that even though the overall activity is altered for increasing amounts of residual Ag the reaction mechanism is not noticeably changed.

Figure 5: Kinetics of CO oxidation with np-Au. The partial pressure of the particular reactant was varied under an excess of the co-reactant. A double logarithmic plot reveals the exponent i.e. reaction order of the particular component.

The reactivity of gold and silver surfaces regarding CO and O2 was intensively investigated under ultra high vacuum conditions (UHV).(38, 41, 61-72) In general, the desorption enthalpies of CO on Au single crystal surfaces are higher than those on Ag surfaces (see Table 1). This picture is reversed in case of the adsorption of molecular oxygen. On Au surfaces molecular oxygen is only physisorbed (desorption well below 100 K) and no dissociation is observed (62). The adsorption of molecular oxygen on Ag is considerably stronger and dissociation is observed.(72) Table 1: Desorption enthalpy of CO and O2 on Gold and Silver single crystals.

Au(111) , sputtered Au(111) Ag(111), Ag(110)

CO [kJ/mol] 44-56 27

O2 [kJ/mol] 11 20-30

Reference (41, 61-62, 65) (71-73)

Our findings are in line with investigations using bimetallic particles. Haruta et al. showed that even small quantities of Ag in the range of 1 at% within a bimetallic Au-Ag catalyst have an impact in the catalytic oxidation of CO.(74-75) Although the detailed mechanism and the nature of the active site in these bimetallic catalysts remain a matter of discussion, the activation and adsorption of oxygen is usually ascribed to Ag while the adsorption of CO

Page 22


is ascribed to Au sites.(76-79) This picture fits to our experimental finding, the fast saturation of the surface sites by oxygen (reaction order of zero) is well compatible with the fact that Ag is only the minority on the surface of np-Au. The major part of the surface consists of Au, this may very well offer a large enough number of surface sites so that the supply of CO by the gas phase is rate limiting and thus the reaction order is well above zero, as the case may be even 1.

Figure 6: (a) Influence of water on the rate of formation of CO2. By adding water to the gas stream the catalytic conversion can be enhanced by more than 100 %. The reactant gases were adjusted to a stoichiometric ratio (O2/CO, 1:2, temperature 20 °C). A water content of 5 ppm is typical for dry reactant gases as supplied from the manufacturer; 10,000 ppm corresonends to 0.01 vol% (~0.4 % rel. humidity). (b) Double logarithmic plot revealing the apparent reaction order (here for O2). The reaction order is apparently unaltered (at 20 °C, CO > 90 vol%).

An experiment further highlighting the surface chemistry of np-Au is the influence of additives such as water on the catalytic conversion. Water does only weakly adsorb on metal surfaces, especially on Au surfaces. Only if oxygen is present, the water molecule can be activated to transiently form OH species.(80) This effect is discussed to have an impact on the activity of surface bonded oxygen, as the case may be by disrupting the oxygen layer which leads to higher activity.(67) In case of supported gold catalysts water can amplify the reaction rate by orders of magnitude.(81) Here, it is however mostly the interaction with the oxidic support (such as TiO2) which causes this effect.(82-83) In case of catalysis with npAu and stoichiometric supply of reactants the conversion of CO was increases by about 100 % just by adding 0.01 vol% water to the gas stream (Figure 6 (a)). Neither the apparent reaction orders nor the activation energy was altered when increasing the water content in the gas stream (Figure 6 (b)). When no oxygen was present, the reaction immediately ceased, accordingly, water is not an additional source for oxygen. This finding implies that water works as a co-catalyst; it is not consumed in the process as it is not contained in the products or educts. Nevertheless, it amplifies the total activity of the catalyst. These experimental results are compatible with the observed reactivity of Au surfaces (80, 84) and prove that support free metal catalysts working under ambient conditions are influenced by moisture as well.

Page 23


3.2 Gas phase oxidative coupling of methanol (Relevant papers: II & IV) Besides the activity of gold to catalyze oxidation reactions already at low temperatures an additional potential lies in the field of selectivity. The selectivity of a catalyst refers to its ability to preferentially form the desired product among other possible products. Increasing selectivity promotes the effective use of resources and lowers the energy consumption. Not only the economic viability of a catalyst depends on its selectivity but also the ecological viability of the particular process. Today, the ever increasing demand for “greener” technologies further intensifies the research towards selective and nonhazardous catalysts, working under mild conditions with abundant, cheap, and environmental friendly reactants.(85-89) Gold is handled as one of the most promising catalyst materials in this respect.(22, 90-95)

Figure 7: Scanning electron micrographs of np-Au showing its sponge-like structure. Reactant gases can diffuse from the outer surface through the porous structure so that the entire gold surface is available for reaction. Inset (middle): The thermodynamics of methanol oxidation. The total oxidation i.e. production of CO2 is energetically favoured. The selectivity on gold surfaces is a function of the oxygen coverage.(96)

The selective oxidation of methanol is an exemplary reaction with great industrial relevance. Methanol is the “simplest” alcohol because it contains only one carbon atom (C1). It is the starting compound for a variety of bulk chemicals such as higher alcohols and hydrocarbons, acids and esters.(97) Most prominent bulk chemicals derived from methanol are methyl formate, an intermediate for the production of formic acid, and formamide, both of them being produced in the range of several hundred thousand tons per year.(98-99) The challenge using catalysis is to selectively produce these various chemicals with carbon in different oxidation states. In terms of thermodynamics the undesired total oxidation that is the formation of CO2 is strongly favoured (see Figure 7). Selectivity can be induced by slowing down, as the case may be ceasing the oxidation reaction at a certain intermediate, for example by desorption of the particular intermediate before it is further oxidized. The Page 24


weak interaction (low adsorption enthalpies) of methanol and its oxidation products on gold (Eads < 21 kcal/mol (96)) renders gold surfaces particularly interesting as catalysts for this type of reaction. As np-Au exhibits a large gold surface and thus a large fraction of highly active low coordinated surface atoms it is a promising candidate for the selective oxidation of methanol.

Figure 8: Selective oxidative coupling of methanol to methyl formate using np-Au as catalyst. (a) The production of methyl formate (gray squares) increases with temperature. At 80 °C more than 60 vol% methanol are converted to methyl formate. The total oxidation to CO2 (blue diamonds) is hardly detectable. (b) The selectivity towards methyl formate production decreases for higher temperatures and oxygen partial pressures. For low oxygen content (1 vol%) the selectivity remains close to 100 %, even for maximal conversion, though.

The investigation of the gas-phase oxidation of methanol with np-Au using molecular oxygen revealed remarkable activity (for details see ref. (22)). The methanol is selectively oxidized to the particular ester, the methyl formate (see Figure 8). The reaction already proceeds at 20 °C with nearly 100 % selectivity; the undesired side product CO2 is hardly detectable. When raising the temperature to 80 °C the conversion can be increased to more than 60 % of methanol, the selectivity however is still at about 97 %. The finding that the selectivity more strongly decreased with temperature for higher oxygen partial pressures is in line with the fact that (a) the activation barrier for the formation of CO2 is higher (96) and (b) that the formation of CO2 is promoted with increasing oxygen surface coverage (96). According to experiments on Au(111) single crystals the surface is activated by chemisorbed atomic oxygen.(96, 100-101) In a Brønstedt-type reaction the methanol reacts with surface oxygen and is subsequently bonded as methoxy on the gold surface. Further, deprotonation leads to the formation of a surface bonded aldehyde. In case of gold this aldehyde is highly reactive, and guides the selectivity towards the methyl formate by a coupling reaction with co-adsorbed methoxy. On the contray, elimination of a hydrogen atom from the aldehyde leads to production of CO2. It is thus the activation of the gold surface by atomic oxygen, the high reactivity of the surface bonded aldehyde, and the low desorption barrier for methyl formate which inflicts the observed reactivity with np-Au.

Page 25


To further elucidate the correlation between the molecular model derived from UHV experiments and the observed reactivity under ambient conditions a series of experiments was conducted where a different aldehyde was co-dosed to the reaction mixture. In order to distinguish between formaldehyde which might be generated from reacting methanol, a different aldehyde, i.e., acetaldehyde, was used for this particular experiment. As expected from the molecular model, the selectivity is driven by the activity of the aldehyde. Instead of methyl formate only methyl acetate, the coupling product between the co-dosed aldehyde and methanol was formed. This interesting coupling chemistry was also observed for a variety of different aldehydes with methanol. (100)

Figure 9: Observed selectivity and catalytic activity when using np-Au samples with higher content of residual silver (here, 2.5 at% Ag). (a) Conversion of methanol to methyl formate (grey squares) and CO2 (blue diamonds) as a function of the reaction temperature. The conversion to methyl formate exhibits a maximum at ~ 65 °C while production of CO2 is continuously increasing with temperature. (b) The selectivity towards methyl formate production decreases with temperature and resembles that of sample containing below 1 at% res. Ag at higher oxygen partial pressures (e.g. 50 vol%, see Figure 8)

As the gold surface is activated by oxygen the amount of residual Ag – which was as already shown to be a key element for supplying activated oxygen in the case of CO oxidation - should have an impact on the activity and selectivity of the catalyst. To shed a light on this effect, samples containing different amounts of residual Ag were prepared by varying the preparation route. By reducing the time for leaching the Ag out of the starting alloy the amount of residual silver can be increased while the porous morphology of the material is preserved (see experimental). By this method samples containing elevated amounts of Ag (2.5 at% and 10 at%) could be prepared and investigated for selective oxidation of methanol. As expected from UHV experiments, the selectivity drops with increasing amount of Ag and active oxygen, respectively. When using samples containing slightly elevated amounts of residual Ag (2.5 at%) the selectivity is at about 95 % close to 20 °C and decreases to about 70 % at 80 °C (see Figure 9). This characteristic resembles that of samples containing low fractions of Ag (~ 0.5 at%) in the case of high oxygen partial pressures (50 vol% cp to 1 vol%). This finding strongly suggests that the amount of Ag regulates the supply of activated oxygen on the surface, eventually leading to increased total oxidation and less selectivity. For samples containing about 10 at% residual Ag no activity towards the formation of methyl formate was observed at all. Only total oxidation (CO2) was detected.

Page 26

Actuation by nanoporous gold

4


(Relevant papers: III & IV)

Figure 10: Surface chemistry induced actuation of np-Au. (a) The strain of a cubic sample of np-Au was measured as a function of the surrounding gases. (b) The exposure of ozone leads to coverage of the surface by atomic oxygen which can be removed by CO. A titration experiment of atomic oxygen with CO was carried out using the continuous flow reactor described in chapter 8.2. The amount of oxygen which is detected by this method corresponds to roughly one monolayer of atomic oxygen.

In the last chapters we discussed the catalysis by np-Au, this is the impact of the surface of np-Au on the surrounding gases. In the following two chapters we will discuss how the surrounding gases and their interaction with the surface of np-Au may have an impact on the material properties of np-Au itself. As described in chapter 2.2, the surface of a bulk material is subject to stress, the so called surface stress. The origin of this phenomenon can be understood on the basis of low coordination of surface atoms and the charge redistribution from unsaturated valences into in-plane bonds. Ibach and co-workers showed that the adsorption of molecules alter the surface stress of metals such as gold.(36) Accordingly, the surface stress of np-Au should be influenced by adsorption of molecules in a similar way than in an electrochemical environment, where the surface charge and surface coverage of a nanoporous metal can be changed by applying a potential. The change in surface charge inflicts a change in surface stress. Due to the exceptional high surface to volume ratio of nanoporous materials, this can lead to macroscopically detectable strain amplitudes, as for example, shown for electrodes of np-Au and np-Pt exhibiting strain amplitudes close or even equal to that of piezo ceramics.(5, 33, 40) The same concept – circumventing previous production of electricity - can be applied as well using the adsorption of gases to influence the surface stress in np-Au. Here, oxygen is of high interest as the adsorption of atomic oxygen is known to lift the surface reconstruction.(38, 102) In contrast to molecular oxygen which does not dissociate on a pure gold surface, ozone (O3) can be used to readily create surface bonded atomic oxygen.(103) Page 27


The surface bonded atomic oxygen is known to strongly react with CO (67) so that it can be removed easily (Figure 10(b)). Accordingly, we used the reaction O3 + Au Æ O2 + Au-O and CO + Au-O Æ CO2 + Au to address the surface stress in np-Au in a setup schematically shown in Figure 10. A cubic sample of np-Au was placed in a cell to control the ambient atmosphere. The gases CO and O3 were successfully fed into this cell; to prevent direct reaction the cell was always purged with N2 in between cycles. The uniaxial strain was measured with a quartz rod contacting the outer surface of the cube of np-Au. The displacement of the quartz rod could be precisely measured. For exposure of O3 the sample shrank, this could be almost completely reversed by removal of oxygen with CO (Figure 11). Depending on the exposure time, the amplitude of strain amounts up to ~ 0.5 %. This corresponds to strain amplitudes of commercial piezo materials. The reversible part of the strain amplitude was nearly 100 percent. Freshly prepared samples however showed smaller reversibility which increased with lifetime. Secondly, the reversible part decreased when the total stroke increased. This indicates to a minor extend plastic deformation and to a major extent stress-driven creep which is more consistent with the slow kinetics. It was found however that the preparation of samples is crucial for the effect of the surface chemistry on the surface stress and the sign of the strain. For example, when dealloying conditions are altered (e.g. dealloying at a different potential) the response of the np-Au material can be changed, as for example no shrinkage but expansion in case of ozone exposure can be observed. In an electrochemical environment Weismüller et al. showed that the pre-treatment of np-Au samples determined the sign of strain.(104) Under which conditions the sign of strain is reversed for adsorbate induced change of surface stress, however, remains a matter of future work.

Figure 11: (a) Cycles of ozone and CO exposure lead to a reversible strain. Under an ozone containing atmosphere the sample shrinks, which can be reversed by removal of oxygen by reaction with CO. (b) The stroke amplitude (reversible part) increases with exposure time of O3.

Page 28

Tailoring structures of nanoporous gold

5


(Relevant papers: IV & V) New material properties arise from reducing the length scale. Nevertheless, it is not only important to reduce the length scale but to tune and adjust the length scale for a specific application. In case of np-Au this means that the pore volume and size of ligaments have to be adjusted. As for example the catalytic activity of np-Au is a merit of the high fraction of low coordinated surface atoms, this number strongly depends on the diameter of the particular ligament (see Figure 3 on page 14). (43, 45, 105) Small ligaments therefore are beneficial for the catalytic activity; the mass transport within the pores of the material though is hampered when the pore volume is too small. This is a challenge which renders a bimodal pore structure very attractive. Secondly, applications in optics, e.g. surface enhanced Raman spectroscopy (SERS), are strongly dependant on the size of the ligaments and pores.(106) SERS enhancement factors were found to be maximal when the pores of npAu were close to 250 nm.(107) Furthermore, the mechanical properties of np-Au are strongly dependant on size of the ligaments.(108-109) These examples exemplify that there is a need to tune and control the structural size of np-Au.

Figure 12: Tuning structure of np-Au. (a) Compilation of SEM images of np-Au after 3 hours of annealing under an inert gas atmosphere (He) and an ozone containing atmosphere (~ 6.5 vol% O3 in O2). (b) The evolution of the average ligament diameter as a function of the annealing temperature. Below the desorption temperature of atomic oxygen (indicated by blue column) the structure of np-Au is stabilized in an ozone containing atmosphere; almost no growth of ligaments is observable.

Coarsening in porous systems is typically ascribed to surface diffusion.(110-111) In case of metals this effect is at room temperature very slow (< 10-20 cm2/s) but increases fast with temperature. Owing to this fact, the structure of np-Au is stable at room temperature, but starts coarsening when annealed at higher temperatures.(107-108) Already at temperatures of around 150 °C coarsening of the structures initiates and can be used to increase the Page 29


ligament and pore size.(109) However, two central challenges for designing structures in npAu remain unattained, the stabilization i.e. conservation of the ligament size at elevated temperatures and the introduction of gradients. Both issues can be solved by means of adsorbate controlled surface diffusion of gold atoms. The effect of surfactants on the stability of ligaments during annealing of np-Au was investigated. Samples of np-Au were annealed in an ozone containing atmosphere and, for comparison, in an inert gas atmosphere (He). When using ozone which readily decomposes on Au surfaces, the surface can be terminated by atomic oxygen.(103) Samples were treated for 3 hours in a tube furnace under continuous flow condition at 450 K, thus below the desorption temperature of atomic oxygen on gold, and above at 650 K, respectively. For characterization the samples were broken into two pieces and the particular size of ligaments along this cross section was investigated by SEM. The results for the ligaments close to the outer surface (~ 10 µm) are shown in Figure 12. At 450 K, i.e. below the desorption temperature of atomic oxygen, the structures of np-Au are stabilized as compared to samples annealed under He atmosphere. This is in line with scanning tunnelling microscopy studies in UHV environment which show that oxygen terminated structures – quite similar to those of np-Au - are considerably more persistent to coarsening.(37) Hardly any difference in size is detectable even after 6 hrs of treatment. On the contrary, at 650 K and thus above the desorption temperature of atomic oxygen the coarsening proceeds apparently un-effected from surrounding ozone. The size of the ligaments annealed under ozone atmosphere essentially equals that of ligaments annealed under He atmosphere.

Figure 13: Gradients in ligament size can be induced using mass transport phenomena. A disk of np-Au was annealed for 3 hours at 450 K under an ozone containing atmosphere. (a) While the outer areas of the np-Au disk are stabilized by adsorbed oxygen ligaments in the inner section of the disk grew by about 300 %. (b) The effect of mass transport regulated growth of ligaments can be reproduced on the basis of a simple model where the adsorbate is supplied by the gas phase.

Page 30


Using the stabilizing effect of ozone at lower temperatures, gradients of the ligament and pore size along the cross section of np-Au disks can be induced (see Figure 13). The coverage of the surface with atomic oxygen depends on the sticking time i.e. desorption energy and the supply from the gas phase (impingement rate and sticking probability). It is apparent, however, that the actual impingement rate is a function of the distance from the outer surface.(112) By assuming straight capillaries and based on thermochemistry data (113), this effect can be assessed as can be seen in Figure 13 (b). The inner surface of the material is not evenly covered with atomic oxygen; in fact the coverage exponentially decreases with the distance from the outer surface. Ligaments in the inner sections of the npAu disk are thus not stabilized anymore and coarsening proceeds unhindered. In this way gradients in ligament and pore size can be introduced within the material. Interestingly, the storage of samples had a decisive impact on the coarsening of the ligaments close to the outer surface. For samples stored in membrane boxes for a longer period of time (several days) gradients within the cross section of np-Au disks were observed after annealing at 650 K. After 3 hrs annealing at 650 K the average ligament diameter close to the outer surface (~ 10 µm) was only 200 nm as compared to 400 to 450 nm in the inner section. Presumably, contaminants originating from the contact of the outer surface with the plastic of the membrane box led to the observed effect. In order to further elucidate this effect samples were pre-treated with an ozone containing atmosphere. This treatment is supposed to create active (atomic) oxygen on the surface and thus guarantee good cleaning by oxidation of carbon containing contaminants. Indeed, no gradient along the cross section was observed anymore. This implies that carbon containing contaminants e.g. hydrocarbons initially present on the outer surface can have a stabilizing effect on the ligaments and prevent coarsening. These examples, the stabilization of ligaments by ozone at lower temperatures or surfactants like hydrocarbons at higher temperatures, reflect the importance of adsorbates on the stability of the nanostructure of np-Au. It is also a descriptive example of the impact of adsorbates on the mobility of surface atoms, and thus coarsening.

Page 31

Outlook: Surface engineering and doping of np-Au

6


(relevant papers IV & VI) The use of the inherent material properties of np-Au is certainly limited. As already shown for the activation of molecular oxygen in the chapters 3.1 and 3.2, the presence of residual Ag within the material can play a beneficial role and extends the limitations posed by the intrinsic surface chemistry of gold. It is, however, tempting to further expand the possible applications of np-Au by doping or depositing other materials, such as metals (Pt, Pd) and semiconductors (CeO2, TiO2), in the porous structure. For example the addition of Pt particles on np-Au working as a substrate has been shown to be promising for applications in fuel cells. (54, 114) Furthermore, material systems comprised of TiO2 layers or nanoparticles on Au substrates attracted considerable attention, not only because such systems constitute inverse systems to oxide supported Au nanoparticles (115-116) but also because of their altered electronic properties (117-118) which hold out interesting applications in photo catalysis (51), photovoltaic devices (119) and water splitting reactions (120). Recently, Ding et al. showed that by implementing TiO2 particles in np-Au films the electro-catalytic activity of gold for methanol oxidation can indeed be sensitised to UV irradiation.(51)

Figure 14: Atomic layer deposition (ALD) is based on the self limiting reaction of a precursor on the material surface. Within the first half-cycle of ALD the precursor is dosed onto the surface, the deposition though is restricted to the reaction with active surface sites. After depletion of active sites no further deposition occurs. In a second half cycle, the active sites on the surface have to be regenerated. The deposition within one complete deposition cycle is limited to 1 monolayer. Depending on the growth mechanism layers as well as particles can be grown.

However, the well-directed deposition of material within np-Au poses a challenge due to its mesoporous structure with pores between 20 and 50 nm in size. In case of thin films (~ 100 nm) of np-Au the mass transport limitation plays only a minor role and experimentally facile “bench-top” methods can be used. (51, 114) Applications such as electrodes in fuel cells or catalytic membranes though typically require the use of freestanding and thus much thicker materials in the range of several microns ore even millimetres. The ultra high aspect ratio of those materials in the order of more than several thousands calls for more Page 32


sophisticated methods to guarantee a conformal and precise deposition of the possibly very expensive dopant/deposit. In this context the atomic layer deposition (ALD) showed great potential for depositing nanoparticles in ultra high aspect ratio materials.(121-124) For example, by using ALD, high aspect ratio materials could be functionalized with beforehand unrealised small amounts of Pt creating a highly active and effective catalyst (see Figure 15).(125) In order to further launch applications of np-Au it will be of great importance to draw on the potential of sophisticated doping techniques like the atomic layer deposition.

Figure 15: Catalytic oxidation of CO using a Pt functionalized carbon aerogel. Due to high aspect ratio of the aerogel the ALD method proved to be particularly suitable to deposit unprecedented small amounts of Pt. Upper section: The time series of CO at the inlet of the reactor (see inset) and the corresponding concentration of CO2 detected at the outlet. Lower section: The catalytic activity increases linearly with the supply of CO, this indicates that all CO supplied is converted to CO2. The maximal i.e. total conversion is already achieved by the smallest possible amount of Pt using 2 cycles of ALD. This result underlines the importance of tailoring the deposition of precious Pt as a key element for high economic viability.

Page 33

Summary

7

Summary

The fabrication and applications of np-Au were investigated with the emphasis on catalysis, actuation, and structural control during annealing. The observed characteristics were interpreted and discussed with the perspective on surface related phenomena. First, the catalytic activity of np-Au for gas phase reactions was investigated using the oxidation of CO. After activation of samples the kinetics of the reaction was determined. To further shed light on the catalytic activity of np-Au with respect to residual Ag and the activation of molecular oxygen, samples containing different amounts of Ag were investigated. In general, the activity increased with increasing amount of Ag. These results strongly suggest that Ag participates in the catalytic cycle. A first interpretation of the role of Ag as the crucial element for the activation of molecular oxygen was proposed based on kinetics and model experiments under UHV conditions. To further highlight the catalytic characteristics of np-Au the gas phase oxidative coupling of methanol was investigated. The formation of methyl formate could be detected already at 20 °C and further increased for higher temperatures. Notably, the formation of an ester (and not of the aldehyde or acid) is fully in line with what has been observed for pure gold surfaces so that the general catalytic behaviour is nearly completely understood and anticipated on the molecular level based on UHV model experiments. Solely the activation of oxygen on np-Au is not explainable on the basis of single crystal experiments so far. In this work, it was found that activity and selectivity are strongly dependant on the fraction of Ag within the material. In the case of increasing content, the partial oxidation considerably decreased and total oxidation was increasingly favoured. Finally, at 11 at% Ag the only observed product was CO2. This finding can be explained by an increased supply of activated oxygen and suggest that Ag regulates the availability of reactive oxygen on the surface. In addition, the impact of surface chemistry on the material properties was investigated with respect to actuation and structural control during annealing. The surface of np-Au was reversibly covered with atomic oxygen taking advantage of the decomposition of ozone. In case of actuation experiments, the oxygen coverage was escorted by a macroscopically detectable stroke of the sample which was reversibly lifted by its removal with CO. This experiment exemplifies the importance of surface chemistry induced stress in case of high surface to volume ratio materials. For structural control during annealing of np-Au samples, the impact of coordination (i.e. bonding to oxygen) on the surface diffusion of gold atoms was used to stabilize the structure or even induce gradients in ligament and pore size.

Page 34

Experimental

8

Experimental

8.1 Synthesis of master alloy and Preparation of np-Au The Au and Ag metals were supplied by the department of energy (DOE). The desired composition of the alloy was 70 at% Ag and 30 at% Au, corresponding to a mass ratio of 56.11 wt% Ag and 43.89 wt% Au. Giving an example, 29.7953 g of alloy were produced by mixing 18.3417 g Ag and 13.0760 g Au. Both metals were brought into the form of chippings in order to guarantee good mixing during melting. The chippings were carefully cleaned by sonication for 30 minutes in isopropanol. A graphite crucible filled with the metal chipping was brought into a preheated muffle oven. The oven was then heated to 1100 °C. After 15 minutes the crucible was carefully rotated to guarantee good mixing of both metals. After a total of 40 minutes the cast was poured into a graphite form, rapidly cooling and forming bars of the desired alloy. After weighing of the bar, the alloy was annealed under an Ar atmosphere for 140 hrs at 875 °C. Repeatedly, the alloy was weighed in order to check any loss of material. Subsequently, the bar was cut with a diamond saw into slices of about 5 mm thickness. These slices were rolled into work pieces of about 200 to 300 µm thickness. The pieces were cut into round disks by punching (diameters of 4.8 mm and 6.7 mm, respectively). The alloy disks were again annealed at 800 °C for 4 hours so that residual stress originated from rolling and cutting was released. All samples were cleaned by sonication in isopropanol for about 30 minutes. Finally, the composition of the alloy disks was controlled by energy dispersive X-ray spectroscopy (EDX). The deviation from the set values (Ag 70 at% and Au 30 at%) was less than 1 at% and thus very close to the accuracy limit of EDX. The preparation of np-Au, that means the dissolution of Ag from the Ag-Au alloy, was done either by “free corrosion” or “potentiostatic corrosion”. For free corrosion no external potential was applied, the open-circuit potential of the alloy caused the dissolution of Ag. The alloy disks were immersed in concentrated nitric acid (65 wt%, Fluka Chemical Corp.) on a glass fritt. After 48 hrs the acid was replaced by deionised water, which was replaced at least three times after 1 hour each. For potentiostatic corrosion the sample was place into a basket formed by a gold wire. Using a typical three electrode setup, the sample was immersed in a 5 M solution of nitric acid and dealloyed at a potential of 60 mV (HNO3 from Fluka, dissolved in ultra pure water, > 18 MΩ) (working electrode: gold wire containing the alloy disk, counter- and reference electrode: Pt) (Figure 17). Generally, the residual Ag content was checked by weight analysis. In several cases samples were also characterized by X-ray photoemission spectroscopy (XPS), and energy dispersive X-ray spectroscopy (EDX) (see Figure 16). Although these spectroscopic techniques are non-destructive the detection limit of EDX of about 1 at% for Ag renders – if necessary - atomic absorption spectroscopy (AAS) useful to precisely quantify such low Ag contents.

Page 35

Experimental

Figure 16: Analysis of differently prepared np-Au. Samples as either prepared by free or potentiostatic corrosion. On the very left an XP spectrum showing the Ag 3d and Au 4d region, verifying that Ag is indeed enriched at the surface (about 8 at%) while the EDX analysis shows no clear signal from the residual Ag. This is a consequence of the detection limit of EDX of about 1 at%. For characterization of morphology the samples were broken and the middle of the particular cross section was investigated by SEM. The average ligament diameter and the corresponding histogram are displayed. On the very right, a characterization of a sample is shown which was dealloyed for 32 hours and thus containing a higher fraction of residual silver.

Figure 17: Potentiostatic corrosion of a Ag70Au30 alloy disk in 5 M nitric acid. The sample was placed into a basked formed by a gold wire. On the right: Using a typical three electrode setup the sample was held at constant potential (60 mV vs. Pt). On the left: Based on the current between the sample (working electrode) and the counter electrode the amount of Ag dissolved can be calculated (assuming every electron originated from the dissolution of one Ag metal atom). At about 30 hours the current steeply decreases. This corresponds to a rapid drop in the corrosion rate of Ag.

Page 36

Experimental

8.2 Catalytic investigation

Figure 18: Experimental setup for measuring the catalytic performance of np-Au. Flow and composition of gases were precisely controlled via mass flow controllers (upper right). The stream of gases (highlighted by blue lines) was guided towards the reactor (enlarged on the lower right). For experiments using e.g. methanol an additional He stream was guided through a system of saturator and condenser to precisely adjust the content of the particular liquid in the gas phase.

The catalytic experiments were performed in a continuous flow reactor. A schematic of the setup is shown in Figure 19. An overview of the experimental setup using photographs is depicted in Figure 18. The double walled reactor was completely made of glass, and its inner diameter was around 20 mm with a total volume of 26 ml. The inlet tubing was guided through a circled pipe through the outer compartment of the reactor which was heated and cooled, respectively, by a thermostat. This procedure guaranteed a maximal control over the temperature of the gases. The samples were placed at the lower end of the reactor on a gold plate with an opening in the center. The mass of the gold plate was 2.2 g, thus around 100 times that of the catalyst disk. In this way, a large part of the heat originating from catalytic

Page 37

Experimental

conversion was rapidly dissipated. Isothermal conditions were intended. The inlet gas was directed through the opening in the middle of the gold plate and thus directly guided towards the sample; in consequence it was avoided that large parts of the feed bypassed the sample. The feed gases were He (Linde AG, 5.0), O2 (Linde AG, 4.5), CO (Linde AG, 4.7), and H2 (Linde AG, 5.0). The total flow was always set to 50 sccm. The composition was precisely controlled via mass flow controllers (Maettig Bronkhorst, Netherlands). For experiments using compounds which are liquids under ambient conditions the particular concentration in the gas phase was controlled by its vapour pressure. For this purpose a stream of He was guided through a saturator coupled with a condenser (see Figure 18). In the following the procedure using methanol is described. In principle the same procedure was used for experiments using other liquids such as water or acetaldehyde. First, the gas stream was saturated with methanol at room temperature (around 20 °C), subsequently the amount of methanol in the gas phase was adjusted by directing the stream of saturated gas trough a cooling pipe. Due to lower vapour pressure at lower temperatures, excess methanol condensed. Following this procedure, stable and precise methanol concentrations were achieved. The particular vapour pressure of methanol at different temperatures was calculated using the KDB database (Korea Thermophysical Database, see Figure 20). Prior to measurements, the methanol was dried in order to remove any water. Therefore, methanol (NORMAPUR, 99.8 %, VWR) was mixed with magnesium chipping (8 g per 500 ml) and boiled under reflux for 3 hours. The dried methanol was obtained by distillation under dry gas atmosphere. The stream of gases at the exit of the reactor was monitored online by IR-Gas-Analyzers (URAS 3G, Hartmann und Braun, highly sensitive for CO and CO2). Additionally, the composition of the gas stream was investigated by GCMS. The content of a sample pipe (100 μl) was injected into a gas chromatograph (Fison 8000 series, Column FS-innopeg-1000) coupled with a mass spectrometer (TRIO 1000).

Page 38

Experimental

Figure 19: Schematic of the experimental setup for measuring the catalytic conversion of CO and methanol, respectively. The content of methanol or water in the feed gases were adjusted by their vapour pressure. A stream of carrier gas (He) was saturated, in a second step the precise partial pressure of the particular component was controlled by condensation. The disk of np-Au was placed on a gold plate at the bottom of the reactor. The stream of reactant gases is guided through an opening of the gold plate onto the disk. At the reactor outlet the composition of the gas stream was investigated by IR and GC-MS.

Figure 20: Vapor pressure of methanol and the corresponding concentration (partial pressure) within the gas phase assuming 1 bar total pressure. The pressure was calculated based on the KDB database (coefficients A =-9.372816, B = -7·103, C = 79.09415, D =8·106 )

Page 39

Experimental

8.3 Measuring actuation of np-Au The experiments regarding the measurement of surface chemistry induced actuation of nanoporous gold were performed at the Institut für Nanotechnology (INT) at the Forschungszentrum Karlsruhe. Pictures of the experimental setup are shown in Figure 21. The np-Au samples were supplied by INT. The master alloy of the particular samples was Ag65Au35 which was dealloyed at 55 °C in HClO4 at a potential of 650 mV vs. Ag/AgCl (dealloying was stopped when the anodic current was negligible small). The dimensions of the cubic sample were 1x1x1 mm3. For measurement the sample was placed in a sample compartment to control the composition of the surrounding gases. The stroke of the sample was detected by a quartz rod which was in contact with the surface of the np-Au cube.

Figure 21: Photographs of the experimental setup for measuring surface chemistry driven actuation of np-Au

Page 40

Experimental

8.4 Annealing Experiments The annealing experiments were conducted using an experimental setup shown in Figure 22. At first a disk of np-Au was broken into several pieces (e.g. 4) so that in any case at least one piece remained untreated as reference. For annealing the particular sample piece was placed within a tube furnace made of quartz glass. To prevent direct contact with the glass tube the sample was embedded in quartz wool, a homogeneous distribution of heat through the gas phase was intended. In order to precisely measure the temperature within the tube furnace a thermocouple (Cr/NiCr) was placed close to the sample (about 1 cm distance). The stream velocity of gas (He 5.0 Linde AG and O2 4.5 Linde AG) was adjusted by rotameters (variable area flow meters) to 50 ml/min. Before heating up the furnace the entire compartment was purged with the particular gas for one hour. For experiments using an ozone containing atmosphere the stream of O2 was guided through an ozone generator (BMT Messtechnik, BMT 802 N), the actual content of ozone was controlled by an ozone analyzer (BMT Messtechnik, BMT 964).

Figure 22: Experimental setup for annealing of np-Au under different atmospheres

Page 41

References

9 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

References H. Renner et al., in Ullmann's Encyclopedia of Industrial Chemistry. (Wiley-VCH, Weinheim, 2000). M. Haruta, T. Kobayashi, H. Sano, N. Yamada, Chemistry Letters, 405 (1987). G. J. Hutchings, J. Catal. 96, 292 (1985). C. W. Corti, R. J. Holliday, D. T. Thompson, Top. Catal. 44, 331 (Jun, 2007). J. Biener et al., Materials 2, 2404 (2009). J. Lee et al., Chem. Soc. Rev. 38, 1450 (2009). J. Erlebacher, R. Seshadri, MRS Bull. 34, 561 (Aug, 2009). Z. H. Zhang et al., J. Phys. Chem. C 113, 12629 (Jul, 2009). R. C. Newman, K. Sieradzki, MRS Bull. 24, 12 (Jul, 1999). R. C. Newman, S. G. Corcoran, J. Erlebacher, M. J. Aziz, K. Sieradzki, MRS Bull. 24, 24 (Jul, 1999). Y. Ding, Y. J. Kim, J. Erlebacher, Adv. Mater. 16, 1897 (Nov, 2004). A. J. Forty, Nature 282, 597 (1979). H. W. Pickering, Corrosion Sci. 23, 1107 (1983). H. W. Pickering, P. R. Swann, Corrosion 19, 373t (1963, 1963). A. J. Forty, P. Durkin, Philosophical Magazine a-Physics of Condensed Matter Structure Defects and Mechanical Properties 42, 295 (1980). Y. Ding, M. W. Chen, MRS Bull. 34, 569 (Aug, 2009). E. Seker, M. Reed, M. Begley, Materials 2, 2188 (2009). S. Kameoka, A. P. Tsai, Catal. Lett. 121, 337 (Mar, 2008). S. Kameoka, A. P. Tsai. (Elsevier Science Bv, Seoul, SOUTH KOREA, 2007), pp. 88-92. X. G. Wang, Z. Qi, C. C. Zhao, W. M. Wang, Z. H. Zhang, J. Phys. Chem. C 113, 13139 (Jul, 2009). Z. Qi et al., J. Phys. Chem. C 113, 6694 (Apr, 2009). A. Wittstock, V. Zielasek, J. Biener, C. M. Friend, M. Baumer, Science 327, 319 (January 15, 2010, 2010). G. Tammann, Z. Anorg. Allg. Chem. 107, 1 (Jul, 1919). G. Masing, Naturwissenschaften 11, 413 (1923). A. Dursun, D. B. Pugh, S. G. Corcoran, J. Electrochem. Soc. 152, B65 (2005). A. Dursun, D. V. Pugh, S. G. Corcoran, Electrochemical and Solid State Letters 6, B32 (Aug, 2003). A. Dursun, D. V. Pugh, S. G. Corcoran, J. Electrochem. Soc. 150, B355 (Jul, 2003). D. M. Artymowicz, J. Erlebacher, R. C. Newman, Philos. Mag. 89, 1663 (2009). J. Erlebacher, J. Electrochem. Soc. 151, C614 (2004). J. Erlebacher, M. J. Aziz, A. Karma, N. Dimitrov, K. Sieradzki, Nature 410, 450 (Mar, 2001). L. H. Qian, M. W. Chen, Appl. Phys. Lett. 91, (Aug, 2007). J. Biener et al., Nature Materials 8, 47 (Jan, 2009). J. Weissmuller et al., Science 300, 312 (Apr, 2003). W. Haiss, Reports on Progress in Physics 64, 591 (May, 2001). H. Ibach, Surf. Sci. Rep. 29, 195 (1997). C. E. Bach, M. Giesen, H. Ibach, T. L. Einstein, Phys. Rev. Lett. 78, 4225 (Jun, 1997). J. Biener et al., ChemPhysChem 7, 1906 (Sep, 2006). B. K. Min, A. R. Alemozafar, M. M. Biener, J. Biener, C. M. Friend, Top. Catal. 36, 77 (Aug, 2005). H. J. Jin et al., Acta Mater. 57, 2665 (May, 2009).

Page 42

References

40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78.

D. Kramer, R. N. Viswanath, J. Weissmuller, Nano Lett. 4, 793 (May, 2004). W. L. Yim et al., J. Phys. Chem. C 111, 445 (Jan, 2007). J. Greeley, J. K. Norskov, M. Mavrikakis, Annual Review of Physical Chemistry 53, 319 (2002). J. K. Norskov et al., Chem. Soc. Rev. 37, 2163 (Oct, 2008). J. K. Norskov et al., J. Catal. 209, 275 (Jul, 2002). H. Falsig et al., Angew. Chem.-Int. Edit. 47, 4835 (2008). V. Zielasek et al., Angew. Chem.-Int. Edit. 45, 8241 (2006). C. X. Xu et al., J. Am. Chem. Soc. 129, 42 (Jan, 2007). G. C. Bond, D. T. Thompson, Gold Bull. 33, 41 (2000). G. C. Bond, D. T. Thompson, Catalysis Reviews-Science and Engineering 41, 319 (1999). H. M. Yin et al., J. Phys. Chem. C 112, 9673 (Jul, 2008). C. C. Jia et al., J. Phys. Chem. C 113, 16138 (Sep, 2009). J. T. Zhang, P. P. Liu, H. Y. Ma, Y. Ding, J. Phys. Chem. C 111, 10382 (Jul, 2007). Y. Ding, M. W. Chen, J. Erlebacher, J. Am. Chem. Soc. 126, 6876 (Jun, 2004). R. Zeis, T. Lei, K. Sieradzki, J. Snyder, J. Erlebacher, J. Catal. 253, 132 (Jan, 2008). C. L. Bracey, P. R. Ellis, G. J. Hutchings, Chem. Soc. Rev. 38, 2231 (2009). J. L. Gong, C. B. Mullins, Accounts Chem. Res. 42, 1063 (Aug, 2009). M. Bowker, Chem. Soc. Rev. 37, 2204 (Oct, 2008). N. Bion, F. Epron, M. Moreno, F. Marino, D. Duprez, Top. Catal. 51, 76 (Dec, 2008). A. Wittstock et al., The Journal of Physical Chemistry C 113, 5593 (2009). W. J. T. John Meurig Thomas, Principle and Practice of Heterogeneous Catalysis. (Weinheim, New York, Basel, Cambridge, Tokyo, October 1996). J. M. Gottfried, K. J. Schmidt, S. L. M. Schroeder, K. Christmann, Surf. Sci. 536, 206 (Jun, 2003). J. M. Gottfried, K. J. Schmidt, S. L. M. Schroeder, K. Christmann, Surf. Sci. 525, 184 (Feb, 2003). J. M. Gottfried, K. J. Schmidt, S. L. M. Schroeder, K. Christmann, Surf. Sci. 525, 197 (Feb, 2003). J. M. Gottfried, N. Elghobashi, S. L. M. Schroeder, K. Christmann, Surf. Sci. 523, 89 (Jan, 2003). J. M. Gottfried, K. Christmann. (Elsevier Science Bv, Prague, CZECH REPUBLIC, 2003), pp. 1112-1117. J. M. Gottfried, K. J. Schmidt, S. L. M. Schroeder, K. Christmann, Surf. Sci. 511, 65 (Jun, 2002). B. K. Min, A. R. Alemozafar, D. Pinnaduwage, X. Deng, C. M. Friend, J. Phys. Chem. B 110, 19833 (Oct, 2006). S. Y. Quek, C. M. Friend, E. Kaxiras, Surf. Sci. 600, 3388 (Sep, 2006). J. Pawelacrew, R. J. Madix, J. Stohr, Surf. Sci. 339, 23 (Sep, 1995). M. A. Barteau, R. J. Madix, J. Chem. Phys. 74, 4144 (1981). M. A. Barteau, R. J. Madix, Surf. Sci. 97, 101 (1980). M. Bowker, M. A. Barteau, R. J. Madix, Surf. Sci. 92, 528 (1980). G. McElhiney, H. Papp, J. Pritchard, Surf. Sci. 54, 617 (1976). Y. Iizuka et al., J. Catal. 262, 280 (Mar, 2009). Y. Iizuka et al., Catal. Lett. 97, 203 (Sep, 2004). A. Q. Wang, Y. Hsieh, Y. F. Chen, C. Y. Mou, J. Catal. 237, 197 (Jan, 2006). A. Q. Wang, J. H. Liu, S. D. Lin, T. S. Lin, C. Y. Mou, J. Catal. 233, 186 (Jul, 2005). J. H. Liu, A. Q. Wang, Y. S. Chi, H. P. Lin, C. Y. Mou, J. Phys. Chem. B 109, 40 (Jan, 2005). Page 43

References

79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115.

D. I. Kondarides, X. E. Verykios, J. Catal. 158, 363 (Feb, 1996). R. G. Quiller et al., J. Chem. Phys. 129, 9 (Aug, 2008). M. Date, M. Haruta, J. Catal. 201, 221 (Jul, 2001). L. M. Liu, B. McAllister, H. Q. Ye, P. Hu, J. Am. Chem. Soc. 128, 4017 (Mar, 2006). Z. L. Wu, S. H. Zhou, H. G. Zhu, S. Dai, S. H. Overbury, J. Phys. Chem. C 113, 3726 (Mar, 2009). H. Y. Su, M. M. Yang, X. H. Bao, W. X. Li, J. Phys. Chem. C 112, 17303 (Nov, 2008). N. Dimitratos, J. A. Lopez-Sanchez, G. J. Hutchings, Top. Catal. 52, 258 (Apr, 2009). M. Poliakoff, P. Licence, Nature 450, 810 (Dec, 2007). V. Gewin, Nature 440, 378 (2006 Mar, 2006). M. Toda et al., Nature 438, 178 (Nov, 2005). D. Bradley, Science 300, 2022 (Jun, 2003). C. Marsden et al., Green Chem. 10, 168 (2008). E. Taarning, A. T. Madsen, J. M. Marchetti, K. Egeblad, C. H. Christensen, Green Chem. 10, 408 (2008). B. K. Min, C. M. Friend, Chem. Rev. 107, 2709 (Jun, 2007). T. Ishida, M. Haruta, Angew. Chem.-Int. Edit. 46, 7154 (2007). M. D. Hughes et al., Nature 437, 1132 (Oct, 2005). M. Haruta, Gold Bull. 34, 40 (2001). X. Y. Liu, R. J. Madix, C. M. Friend, Chem. Soc. Rev. 37, 2243 (Oct, 2008). E. Fiedler, G. Grossmann, D. B. Kersebohm, G. Weiss, C. Witte, in Ullmann's Encyclopedia of Industrial Chemistry. (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2000). G. Reuss, W. Disteldorf, A. O. Gamer, A. Hilt, in Ullmann's Encyclopedia of Industrial Chemistry. (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2000). W. Reutemann, H. Kieczka, in Ullmann's Encyclopedia of Industrial Chemistry. (Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, 2000). B. J. Xu, X. Y. Liu, J. Haubrich, R. J. Madix, C. M. Friend, Angew. Chem.-Int. Edit. 48, 4206 (2009). R. J. Madix, C. M. Friend, X. Y. Liu, J. Catal. 258, 410 (Sep, 2008). B. K. Min, X. Deng, D. Pinnaduwage, R. Schalek, C. M. Friend, Physical Review B 72, (Sep, 2005). N. Saliba, D. H. Parker, B. E. Koel, Surf. Sci. 410, 270 (Aug, 1998). H. J. Jin, S. Parida, D. Kramer, J. Weissmuller, Surf. Sci. 602, 3588 (Dec, 2008). J. K. Norskov, T. Bligaard, J. Rossmeisl, C. H. Christensen, Nat. Chem. 1, 37 (Apr, 2009). J. Biener et al., Adv. Mater. 20, 1211 (Mar, 2008). S. O. Kucheyev et al., Appl. Phys. Lett. 89, (Jul, 2006). J. Biener et al., Nano Lett. 6, 2379 (Oct, 2006). A. M. Hodge, J. R. Hayes, J. A. Caro, J. Biener, A. V. Hamza, Adv. Eng. Mater. 8, 853 (Sep, 2006). A. S. Dalton, E. G. Seebauer, Surf. Sci. 601, 728 (Feb, 2007). E. G. Seebauer, C. E. Allen, Prog. Surf. Sci. 49, 265 (Jul, 1995). R. G. Gordon, D. Hausmann, E. Kim, J. Shepard, Chemical Vapor Deposition 9, 73 (Mar, 2003). J. Kirn, E. Samano, B. E. Koel, Surf. Sci. 600, 4622 (Oct, 2006). R. Zeis, A. Mathur, G. Fritz, J. Lee, J. Erlebacher, J. Power Sources 165, 65 (Feb, 2007). M. Chen, Y. Cai, Z. Yan, D. W. Goodman, J. Am. Chem. Soc. 128, 6341 (May, 2006).

Page 44

References

116. 117. 118. 119. 120. 121. 122. 123. 124. 125.

M. S. Chen, D. W. Goodman, Science 306, 252 (Oct, 2004). K. Kamata et al., Science 300, 964 (May, 2003). M. Jakob, H. Levanon, P. V. Kamat, Nano Lett. 3, 353 (Mar, 2003). E. W. McFarland, J. Tang, Nature 421, 616 (Feb, 2003). J. A. Rodriguez et al., Science 318, 1757 (Dec, 2007). S. Ghosal et al., Chem. Mat. 21, 1989 (May, 2009). S. O. Kucheyev et al., Langmuir 24, 943 (Feb, 2008). J. Biener et al., Nanotechnology 18, (Feb, 2007). S. O. Kucheyev et al., Appl. Phys. Lett. 86, (Feb, 2005). J. S. King et al., Nano Lett. 8, 2405 (Aug, 2008).

Page 45

Curriculum Vitae

Arne Wittstock Personal data • Born in Bremen (Germany), March 14th, 1981. • Address: Institute of Applied and Physical Chemistry, University Bremen, Leobener Str. NW2, 28359 Bremen, Germany • E-mail: [email protected]

Education •

2006: Diploma in Chemistry (Dipl.-Chem.), summa cum laude (1.0), University Bremen, Germany

•

2000: Abitur, Altes Gymnasium in Bremen (1.3)

Work experience •

since 2006: PhD Study at the Institute for Applied an Physical Chemistry at the University Bremen, Germany

•

2006: Research assistant, Krüss GmbH, Hamburg, Germany (Diploma thesis)

•

2003 - 2006: Teaching assistant at the University Bremen (mathematics, physical chemistry, and lab training).

•

2002 – 2005: Software development (C++) at the MIR Chem GmbH, Bremen, Germany

Research visits •

2007 - 2008: Research stay at the Lawrence Livermore National Laboratory, Livermore, USA (Dr. Alex Hamza).

. Research Interests Chemistry of surfaces, physical chemistry and surface reactions at nanostructured surfaces, catalytic studies under ambient conditions, innovative catalyst materials (such as nanoporous foams)

Publications (1)

Wittstock, A., Zielasek, V., Biener, J., Friend, C. M. & Baumer, M.; Nanoporous Gold Catalysts for Selective Gas-Phase Oxidative Coupling of Methanol at Low Temperature; Science 327, 319-322, doi:10.1126/science.1183591 (2010).

(2)

Wittstock, A.; Neumann, B.; Schaefer, A.; Dumbuya, K.; Kübel, C.; Biener, M. M.; Zielasek, V.; Steinrück, H.-P.; Gottfried, J. M.; Biener, J.; Hamza, A.; Bäumer, M.; Nanoporous Au: An Unsupported Pure Gold Catalyst?; The Journal of Physical Chemistry C 113, 5593-5600, doi:doi:10.1021/jp808185v (2009).

(3)

Biener, J.; Wittstock, A.; Zepeda-Ruiz, L. A.; Biener, M. M.; Zielasek, V.; Kramer, D.; Viswanath, R. N.; Weissmuller, J.; Bäumer, M.; Hamza, A. V.; Surface-chemistrydriven actuation in nanoporous gold; Nature Materials 8, 47-51, doi:10.1038/nmat2335 (2009).

(4)

Biener, J.; Wittstock, A.; Baumann, T.; Weissmüller, J.; Bäumer, M.; Hamza, A.; Surface Chemistry in Nanoscale Materials; Materials 71, 603-612 (2009).

(5)

King, J. S.; Wittstock, A.; Biener, J.; Kucheyev, S. O.; Wang, Y. M.; Baumann, T. F.; Giri, S. K.; Hamza, A. V.; Baeumer, M.; Bent, S. F. Nano Lett. 8, 2405-2409, doi:10.1021/nl801299z (2008).

(6)

Hass, E. C.; Ottensmeyer, R.; Wittstock, A.; Foam characterization by optical measuring techniques and computer-aided image processing; Tech. Mess. 71, 603612 (2004)

Patents [I]

ACTUATION VIA SURFACE CHEMISTRY INDUCED SURFACE STRESS, IPC8 Class: AC23C2200FI, USPC Class: 148240, Status: pending

[II]

METHOD AND DEVICE FOR MEASURING FOAM, WO/2005/003758, international Filing Date: 24/06/2004

Papers

Paper I: Nanoporous Au: An Unsupported Pure Gold Catalyst? Wittstock, A.; Neumann, B.; Schaefer, A.; Dumbuya, K.; Kübel, C.; Biener, M. M.; Zielasek, V.; Steinrück, H.-P.; Gottfried, J. M.; Biener, J.; Hamza, A.; Bäumer, M The Journal of Physical Chemistry C, 2009, 113 (14), 5593-5600, Link: http://pubs.acs.org/doi/pdf/10.1021/jp808185v

J. Phys. Chem. C 2009, 113, 5593–5600

5593

Nanoporous Au: An Unsupported Pure Gold Catalyst? Arne Wittstock,† Bjo¨rn Neumann,† Andreas Schaefer,† Karifala Dumbuya,§ Christian Ku¨bel,| Monika M. Biener,‡ Volkmar Zielasek,† Hans-Peter Steinru¨ck,§ J. Michael Gottfried,*,§ Ju¨rgen Biener,* Alex Hamza,‡ and Marcus Baümer*,† Institute for Applied and Physical Chemistry, UniVersity of Bremen, Leobener Strasse NW2, 28359 Bremen, Germany, Nanoscale Synthesis and Characterization Laboratory, Lawrence LiVermore National Laboratory, P. O. Box 808, LiVermore, California 94550, Lehrstuhl fu¨r Physikalische Chemie II, UniVersita¨t Erlangen-Nu¨rnberg, Egerlandstrasse 3, 91058 Erlangen, Germany, and AdhesiVe Bonding and Surfaces DiVision, Fraunhofer Institute for Manufacturing Technology and Applied Materials Science (IFAM), Wiener Strasse 12, 28359 Bremen, Germany ReceiVed: September 15, 2008; ReVised Manuscript ReceiVed: January 13, 2009

The unique properties of gold especially in low temperature CO oxidation have been ascribed to a combination of various effects. In particular, particle sizes below a few nanometers and specific particle-support interactions have been shown to play important roles. In contrast, recent reports revealed that monolithic nanoporous gold (npAu) prepared by leaching a less noble metal, such as Ag, out of the corresponding alloy can also exhibit a remarkably high catalytic activity for CO oxidation, even though no support is present. Therefore, it was claimed to be a pure and unsupported gold catalyst. We investigated npAu with respect to its morphology, surface composition, and catalytic properties. In particular, we studied the reaction kinetics for low temperature CO oxidation in detail, taking the mass transport limitation due to the porous structure of the material into account. Our results reveal that Ag, even if removed almost completely from the bulk, segregates to the surface, resulting in surface concentrations of up to 10 atom %. Our data suggest that this Ag plays a significant role in activating of molecular oxygen. Therefore, npAu should be considered a bimetallic catalyst rather than a pure Au catalyst. Introduction In recent years, catalysis by gold became an intensively studied topic in research due to its promising applications, e.g., in low temperature CO oxidation,1-4 selective oxidation of methanol,5,6 epoxidation of propene,7,8 and water gas shift reaction.9,10 Generally, it is believed that catalytic activity of this noble metal only occurs when it is in the form of nanometersized particles supported on a suitable metal oxide support. Numerous studies revealed complex dependencies of the catalytic properties on the size of the particles, the nature of the support, and other factors, such as the presence of water in the feed.11 The use of some supports, including R-Fe2O3 and TiO2 as well as CeO2, results in highly active catalysts, whereas catalysts prepared on others, such as Al2O3 and SiO2, showed usually low or no activity.12 Unfortunately, no consistent picture has arisen so far from the partly deviating results, so the debate regarding the underlying mechanisms still continues. Consequently, some effort was made to prepare and investigate Au catalysts without any support to assess the activity of pure Au without support effects. Dumesic and co-workers, e.g., showed that Au nanotubes embedded in a polycarbonate matrix are active for the oxidation of CO at room temperature.13 Haruta et al. demonstrated that Au powder consisting of particles in the range of 100 nm, prepared by evaporating Au into an inert * Corresponding authors. E-mail: [email protected] (M.B.); [email protected] (J.M.G.); biener2@llnl. gov (J.B.). † University of Bremen. ‡ Lawrence Livermore National Laboratory. § Universita¨t Erlangen-Nu¨rnberg. | IFAM.

gas, exhibits remarkable catalytic activity.14 Interestingly, a careful characterization of the material revealed a strong influence of Ag impurities resulting from the manufacturing process. Another route to unsupported nanostructured Au catalysts only recently proposed in the literature is the preparation of a monolithic nanoporous Au material (npAu) by selectively leaching a less noble metal, such as Ag or Cu, from the corresponding gold alloy.15 The resulting material exhibits a characteristic spongelike open-cell morphology with a uniform feature size on the nanometer length scale. Depending on the particular preparation conditions, the ligament size can be tailored between 10 nm and 1 µm without changing the relative morphology.16 As shown by several groups, such a material can exhibit surprisingly high activity for low temperature CO oxidation.17-21 Based on these results, several authors claimed that this activity reflects the unique properties of Au in the absence of a support.19,20 In particular, it was speculated that this system might help to unravel some of the open questions in Au catalysis. If that was true, however, aspects, such as particle size and the role of the support, would have been misinterpreted in so far as being essential prerequisites in Au catalysis. Rather, they would be special features of supported systems. Putting these new materials in the context of previous research, Haruta pointed out that impurities, such as Ag, left over from the preparation may be important ingredients to understanding the unique properties of npAu.22 In particular, if these components are oxidized under reaction conditions or supply activated oxygen, they could play an analogous role to the supports in the case of supported systems.

10.1021/jp808185v CCC: $40.75  2009 American Chemical Society Published on Web 03/17/2009

5594 J. Phys. Chem. C, Vol. 113, No. 14, 2009 In some of the previous reports on the catalytic activity of npAu, however, residues of the second metal were not taken into consideration when explaining the catalytic properties,19,20 although participation in the catalytic cycle is very likely. While CO adsorbs on Au especially on low coordinated sites,23,24 no adsorption of oxygen was found on single crystal surfaces.25,26 Only in the case of small particles, dissociation becomes feasible.27 In view of these results, dissociation of O2 on pure npAu seems rather unlikely to occur. On the other hand, several investigations of Au-Ag bimetallic particle systems on a mesoporous support showed that catalytic activity for CO oxidation can even be obtained if rather large particles (>30 nm) are used. Moreover, a clear correlation between composition and conversion was detected, suggesting the possibility of tuning the catalytic properties by changing the Au/Ag ratio.22,28-32 This fits into the general trend that alloying Au with other metals can drastically alter its catalytic behavior.33,34 Motivated by these results, we aimed at elucidating whether the catalytic behavior of npAu provides indications for an analogous role of residual Ag left over from the dealloying process for the adsorption and activation of reactants. Using CO oxidation as a probe reaction, we studied the activation of the samples, the reaction kinetics of their steady-state activity, and the chemical state under reaction conditions. Moreover, we aimed at studying the ability of the material to adsorb and bind oxygen after exposing it to molecular oxygen and, for comparison, to ozone, which is known to be able to deliver atomic oxygen to Au surfaces.35 The results of our experiments indicate that npAu is not a pure Au catalyst and imply that Ag is a necessary factor for carrying out oxidation reactions with this material. The remarkable and unexpected high capability of activating molecular oxygen can be easily understood in terms of a bimetallic alloy catalyst in a similar way as was previously reported for bimetallic Au-Ag systems.22,31,36,37 Experimental Section Disks of npAu with a diameter of 5 mm and a thickness of 200 or 300 µm were prepared by etching Ag-Au alloys in concentrated nitric acid (48 h, HNO3, ∼70 wt %, OmniTrace EMD Chemicals). This procedure is very efficient and removes Ag almost quantitatively to concentrations below 1 atom % as determined by EDX (energy dispersive X-ray spectroscopy) or AAS (atomic absorption spectroscopy). Samples with a slightly higher residual silver content were obtained by adding AgNO3 to the nitric acid (0.2 M Ag+). In this way, the silver concentration could be adjusted to concentrations up to 4 atom %, as determined by EDX. The master alloy for all samples was prepared by melting Au and Ag pellets in the desired composition (Ag 70 atom %, Au 30 atom %). After verification of the alloy composition by EDX, the ingots were annealed under an argon atmosphere for 140 h at 875 °C. After this homogenization the ingots were rolled, cut into the desired shape, and again repeatedly annealed in air for 4 h at 800 °C. For structural characterization, the npAu was infiltrated with an epoxy resin (Spoerr) and tempered at 70 °C for 8 h. Afterwards, ultrathin slices with a nominal thickness of 100-200 nm were prepared by ultramicrotomy at room temperature and transferred on carbon coated copper grids. Transmission electron microscopic (TEM) analysis of the infiltrated npAu was performed using a Tecnai G2 F20 STwin instrument equipped with a Schottky field emitter (FEG) and operated at 200 kV. Imaging was performed either in BF-TEM mode or in HAADF-

Wittstock et al. STEM mode with a nominal spot size of 0.5 nm. Analytical work was performed using an EDX detector with a S-UTW window either working in HAADF-STEM mode with a nominal spot size of 1-2 nm or in BF-TEM mode to look at compositional averages over larger areas. X-ray photoelectron spectra (XPS) of selected samples were recorded under ultra-high-vacuum (UHV) conditions (base pressure of the UHV chamber: 10-11 mbar) employing a Mg KR X-ray source at 1253.6 eV and a hemispherical analyzer (Leybold, EA 10 plus). The pass energy was set to 25 eV. The in situ XPS measurements were performed with a highpressure X-ray photoelectron spectrometer, which can be operated in the presence of a reactive gas phase at pressures of up to 1 mbar at the sample; this allows for measurements within the “pressure gap” regime. The apparatus has been described in detail elsewhere.38 Briefly, the experimental setup is based on a modified hemispherical electron energy analyzer (Omicron EA 125), a modified twin anode X-ray source (Specs) (here used; Al KR radiation, hν ) 1486.6 eV), and several differentialpumping stages between the sample region and the electron detector. The vacuum system has a base pressure in the 10-10 mbar range. The reaction gas is provided in situ by either background dosing or beam dosing; the latter uses a directed gas beam from a tube with a small cross section and permits higher local pressures. In this work, only background dosing was employed. For the general calibration of the energy scale of the spectrometer, we used the Fermi edge and the Au 4f7/2 signal (EB ≡ 84.0 eV) of a clean Au(111) sample. The catalytic experiments were conducted in a continuous flow reactor especially designed for catalytic experiments with the npAu disks. A scheme of the reactor is shown as the inset in Figure 2B. The reactor was completely made of glass, and its inner diameter was around 20 mm with a total volume of 26 mL. The inlet tubing and the reactor were embedded in a liquid bath for precise temperature control. The samples were placed at the lower end of the reactor on a Au plate with an opening in the center. The inlet gas was directed through this opening directly toward the sample, thus avoiding that large parts of the feed bypassed the sample. If not stated otherwise, the feed consisted of a mixture of CO (Linde AG, 4.7), N2 (Linde AG, 4.6), and O2 (Linde AG, 4.5). The total flow was always set to 50 sccm. The composition was precisely controlled via mass flow controllers (Maettig Bronkhorst, Netherlands). For experiments using ozone as a supply for atomic oxygen, we used the same apparatus in connection with an ozone generator (Type 802 N, BMT Messtechnik Berlin) and an ozone analyzer (Type 964, BMT Messtechnik Berlin). Results Structural and Spectroscopic Characterization. The morphology of nanoporous Au obtained by dealloying Ag70Au30 alloy samples in nitric acid is demonstrated by the scanning electron microscopy (SEM) and TEM micrographs shown in Figure 1. The material exhibits a characteristic three-dimensional bicontinuous nanoporous structure with a porosity of ∼70% and a uniform feature size of 30-50 nm that can be very reproducibly prepared.39 (Different samples are virtually not distinguishable with respect to their morphology.) In order to determine the amount of silver which is left after dealloying, we performed atomic absorption spectroscopy (AAS) and energy dispersive X-ray spectroscopy (EDX). The residual silver content of material which was dealloyed for 48 h in pure concentrated HNO3 (without added AgNO3) is typically 1 atom

Nanoporous Au Catalyst

J. Phys. Chem. C, Vol. 113, No. 14, 2009 5595

Figure 1. Morphological characterization of npAu by SEM (upper section) and TEM/XPS (lower section).

%. We also performed X-ray photoelectron spectroscopy (XPS) of such a sample which, due to the limited escape depth of the photoelectrons, is suitable to study the composition of the material close to the surface. A quantitative analysis of XP spectra recorded in the Ag 3d and Au 4d regions (see Figure 1) revealed a Ag concentration of 7 atom % (using the integrated signal intensity of the Ag 3d and Au 4d peaks and the appropriate sensitivity factors). Taking the bulk concentration of ∼1 atom % into account, this value points to a distinct enrichment of Ag close to the surface. Reaction Kinetics. As reported previously,18 pristine npAu samples have to be activated before they exhibit catalytic activity for CO oxidation at room temperature. The CO2 signal as recorded at the reactor outlet during activation of a npAu sample at 80 °C is shown in Figure 2A. The increasing CO2 signal after 2-3 h indicates the onset of activation. The activity continues to increase over the next few hours, finally reaching either a plateau or, more often, passing a transient activity maximum after which the CO2 signal decreases again. When the temperature is decreased at this point (for example to 20 °C), the conversion rate usually increases again and is subsequently stable. It is interesting to note, however, that not all samples could be activated by this procedure. In particular, samples which were stored for several weeks after being dealloyed could often not be activated. For samples that could be activated the kinetics of the reaction was studied in detail. For this purpose, one reaction partner (either CO or O2) was kept in excess (constant high partial pressure), whereas the other one was systematically increased

Figure 2. Catalysis: (A) CO2 signal typically detected at the reactor out during activation of a npAu disk; (B) CO2 signal (outlet) for increasing O2 concentration at the reactor inlet (CO as carrier gas, reactor temperature 20 °C); (C) CO2 as a function of the supplied CO concentration for different fixed O2 concentrations (reactor temperature 20 °C).

from very low to very high concentrations. A typical O2 variation experiment (fixed CO concentration) for a sample with 4 atom % Ag is shown in Figure 2B. The signal intensity of the product CO2 follows (delayed by the retention time of the gases in the reactor) the O2 concentration which was increased in a stepwise manner. Note that the concentration was only changed after observing a steady state for at least 5 min. For further analysis of the data, the measured CO2 concentrations at steady state were converted into a conversion rate and referred to the mass of catalyst and the surface area (3.7 m2/g as measured by BET), respectively. Figure 3A shows the results for increasing CO concentrations in an O2 excess at 20 °C. Even though the plot already reveals a linear dependence of the rate on the CO concentration, we replotted the data in log-log coordinates to determine the reaction order in more detail (see inset). As long as the O2 concentration is not limiting the rate (up to ∼8 mol/m3 CO), the total reaction order corresponds to the reaction order of CO under these conditions. Evaluating the slope of the straight line (1.04) confirms that the CO reaction order is 1.

5596 J. Phys. Chem. C, Vol. 113, No. 14, 2009

Wittstock et al.

Figure 4. (A) Temperature dependency of the reaction rate (4 vol % CO, 20 vol % O2 in He, Mcatal ) 21.5 mg). (B) Arrhenius plot: natural logarithm of conversion versus inverse temperature. A linear fit reveals an apparent activation energy of 31 kJ/mol.

Figure 3. Conversion referred to the surface of the catalyst (A) as a function of the CO concentration at the reactor inlet with O2 as carrier gas (CO2 > 66 vol %, Mcatal ) 27.9 mg, T ) 293 K) and (B) as a function of O2 concentration with CO as carrier gas (CCO > 87 vol %, Mcatal ) 19.3 mg, T ) 293 K) Insets: double-logarithmic plots of both curves for determination of reaction orders.

Figure 3B shows the results for the analogous O2 variation experiment (fixed CO concentration). At low concentrations the increase in conversion is larger than for higher concentrations. The overall shape of the curve suggests a parabolic dependence, i.e., a reaction order of 0.5 which is indeed confirmed by the double-logarithmic plot (see inset). Similar experiments were performed for a number of samples. The results clearly demonstrated that the kinetic parameters are not constant, but vary from sample to sample. For the reaction order of CO, in particular, values between 0.5 and 1 were found with a tendency of higher values for the samples with higher silver concentrations. For the reaction order of oxygen, on the other hand, no such large variations were observed. Here, values close to 0.5 were found in all cases. So far we have discussed only experiments where one reaction partner was supplied in large excess. In addition, we performed experiments where one gas concentration was held constant at lower levels while the other concentration was varied. Figure 2C shows results obtained for different constant oxygen concentrations. As expected, the conversion increases linearly at low CO concentrations, whereas for higher CO concentrations the conversion reaches a plateau which increases with increasing O2 concentrations as a result of the limitation of the reaction by the oxygen supply. Apparently, the slope in the linear regime is similar, even though the O2 concentration spans a range from 2 vol % to over 60 vol %, i.e., is increased by a factor of 30. Thus, even large concentrations of oxygen do not lead to a

blocking of CO adsorption sites. This behavior suggests that CO and oxygen do not compete for adsorption sites on the surface. While the reaction shows almost no temperature dependence between 40 and 80 °C (in agreement with our previous study18 and a study of Wang et al. on AuAg nanoparticles on a mesoporous support30,31), a measurable dependence at lower temperatures allowed us to determine the activation energy of the reaction in this temperature regime. For this purpose, the temperature was first lowered to 253 K and then continuously increased to 360 K, while recording the CO2 formed at the reactor outlet with a feed of 4 vol % CO and 20 vol % O2 (Figure 4A). Assuming an Arrhenius type temperature dependence of the rate, the slope of the (natural) logarithm of the rate as a function of the inverse temperature provides the (apparent) activation energy Ea(app). For the sample described above, an Ea(app) of ∼31 kJ/mol was determined in this way (see Figure 4B). Yet, as in the case of the reaction orders, the measured activation energies scattered to some degree from sample to sample and even values below 10 kJ/mol were found. Notably, calculations on small Au and Au-Ag clusters show that Ag has a strong influence on the adsorption29 and activation energies40 so that differences in the distribution of Ag on the surface may be responsible for this result. (Interestingly, also in the case of pure Au nanoparticles deposited on different support materials, the experimental values often scatter over a wide range.41) Role of Mass Transport Limitation. In a previous publication, it was already pointed out that mass transport limitation in the nanoporous Au material may affect the measured kinetics (activity, reaction order) and thus needs to be taken into account when drawing conclusions regarding the actual kinetics of the reaction.17 Usually, the effect of mass transport limitation on the apparent reaction order and activation energy of heterogeneously catalyzed reactions is very difficult to analyze. Here, however, we take advantage of the well-defined structure of npAu to estimate the degree of mass transport limitation by applying a simple diffusion-reaction model based on reasonable assumptions.42-44 We can treat CO and oxygen independently since no indications for competitive adsorption of the two reactants were found. We address the degree of mass transport limitation by using the so-called Thiele modulus, which is a dimensionless number that describes the ratio of reaction rate to the diffusion rate.



Assuming pseudo-first-order reaction conditions, that is, one reaction partner is in large excess, the Thiele modulus Φ can be expressed as

Φ≡a

kcgm-1 D

(1)

Here, a is the aspect ratio of the pores, k is the reaction constant (m/s), Cg is the gas phase concentration (mol/m3), m is the reaction order, and D is the diffusion coefficient. Note that the Thiele modulus is only independent of the gas phase concentration in the case of a first-order reaction.44 The Thiele modulus expresses the degree of limitation by mass transport; the higher this value the more the reaction rate is dominated by mass transport through the pores. One can say that if the Thiele modulus is between 0.2 and 3 the overall reaction rate is influenced by mass transport phenomena; a Thiele modulus of over 3 means that the consumption of reactants is so fast that only mass transport, i.e., supply of reactants into the pores, is rate limiting. The effect of mass transport on the observed kinetic parameters is demonstrated by the examples presented in Table 1. For a reaction order of 1, as determined as an upper bound for CO, the order is not affected by mass transport. As far as the activation energy is concerned, the model predicts that Eact is 2 times larger than the measured (apparent) activation energy if mass transport limits the reaction (Φ > 3).44 If, however, the reaction order is different from 1, the measured apparent order is different from the “true” reaction order of the surface reaction. In particular, an apparent order of 0.5 corresponds to a true reaction order of 0; that is, the rate of the surface reaction is independent of the gas phase concentration. To assess if, and to what extent, our measured reaction kinetics is influenced by mass transport limitations, we calculated the Thiele modulus for both CO and O2. Specifically, we used the following assumptions: First, we approximated the porosity of npAu by capillaries with length L and radius r, in which case the aspect ratio a can be expressed as 2L/r1/2. This yields an aspect ratio of 1.6-2.5 m1/2 if we assume 100-150 µm long (half of the sample thickness) and 30 nm wide capillaries. Furthermore, assuming Knudsen diffusion, the coefficient D in a highly porous material such as npAu can be expressed by45

D ) εp

DK τ

(2)

where DK is the Knudsen diffusion coefficient (m2/s), εp is the porosity of the material (0.7 in our case), and τ is the so-called tortuosity factor (set to 5), which accounts for the tortuous path of a molecule through the porous network (actual distance a molecule must travel divided by the film thickness). Finally, the Knudsen diffusion coefficient DK is calculated by

MT

DK ) 3.067r

(3)

where T is the temperature and M is the molecular mass of the gas molecules. For our specific example, the diffusion coefficients D of oxygen and CO at room temperature are 6.2 × 10-7 and 6.6 × 10-7 m2 · s-1, respectively.59

Figure 5. Calculated profile of CO concentration along the cross section of a npAu disk (0, outer surface; thickness 200 µm; rcapillary ) 15 nm) based on a reaction diffusion model. The plot reflects the situation for high oxygen partial pressures (Figure 3A).

TABLE 1: Dependency of Kinetic Parameters on Mass Transport Limitation44 not limited (Φ ∼ 0.2) reaction order 0 reaction order 1 activation energy

remains order 0 remains order 1 apparent number

limited (Φ > 3) appears as order 0.5 remains order 1 double the apparent number

Based on eqs 1-3 and taking the data set of Figure 3 as an example, a Thiele modulus of 6.3 is obtained for CO; that is, the system is indeed mass transport dominated. In this case, the efficiency η (ratio of reaction velocity of influenced and uninfluenced reaction) of the catalyst is just 0.16; i.e., the (averaged) concentration of reactant gas within the pores is roughly only 1/5 the concentration in the outer gas phase. The calculated cross-sectional concentration profile of CO in a 200 µm thick sample of npAu is shown in Figure 5 and illustrates the depletion of CO in the center of the sample. Note that the effect becomes even more pronounced for ultrafine npAu material prepared by electrochemically dealloying (15 nm pores),20 where η is further decreased by a factor of 4. In the same way, it was confirmed that also for oxygen the Thiele modulus is above 3 in the relevant concentration range of 1-5 mol/m3 (Figure 3B). Thus, for both reactants, mass transport limitation dominates, meaning that for CO the apparent reaction orders between 0.5 and 1 correspond to actual orders between 0 and 1 and the actual activation energy is 2 times larger than the measured one (i.e., 31 kJ/mol corresponds to 62 kJ/mol), whereas the apparent reaction order of 0.5 for O2 corresponds to an actual reaction order of 0. Nature of Oxygen Species on the Surface. An important question in Au catalysis concerns the activation of molecular oxygen. While no chemisorption of molecular oxygen on planar Au surfaces is observed,47,48 low coordinated gold species were proposed to account for the adsorption and activation of O2 on Au nanoparticles.49 In the case of npAu with ligaments in the range of several tens of nanometers, however, this explanation is unlikely. Furthermore, it was speculated that Au oxide clusters, once formed under catalytic conditions, are capable of dissociating molecular oxygen so that a constant steady state coverage of oxygen atoms on the surface can be maintained under reaction conditions.50 In this context, it is interesting to note that Xu et al. discussed a positively charged Au species formed under dealloying conditions.19 Since cationic gold was frequently discussed to be an ingredient for catalytic activity, it could play a role also in the case of npAu. In order to check whether stable Au oxide species can be generated on npAu and whether they can initiate a stable steady

5598 J. Phys. Chem. C, Vol. 113, No. 14, 2009

Figure 6. Titration of adsorbed atomic oxygen after treatment with ozone.

state conversion, we treated a sample with ozone, which is known to decompose readily on Au surfaces and to form Au oxide species.35 In a first step, we exposed the sample to 0.3 vol % ozone in a feed of synthetic air for 30 min. Subsequently, the surface concentration of oxygen was measured by recording the amount of CO2 formed when reacting the oxide species at the surface away with CO. (To prevent direct reaction of ozone with CO, the reactor was thoroughly purged for 40 min after the ozone treatment.) The resulting CO2 peak is shown in Figure 6. Based on the total number of Au surface atoms (specific surface 3.7 m2/g, 1.4 × 1015 atoms/cm2 (Au(111)), Msample ) 29.5 mg) and assuming that every adsorbed oxygen atom reacts with one CO molecule, an oxygen surface coverage of about 1 monolayer (ML) can be calculated, in good agreement with experiments on single crystal surfaces.35 Next, we repeated the experiment using molecular oxygen instead of ozone. Under these conditions, a CO2 peak was not detected, revealing that molecular oxygen is not able to form a stable atomic oxygen species in significant amounts (∼0.1 ML) on npAu. To evaluate the impact of a precoverage of the surface with atomic oxygen on activation of molecular oxygen, we used an inactive sample, i.e., a sample which showed no conversion of CO at room temperature. After ozone treatment and purging, CO and oxygen were simultaneously added to the gas stream. This resulted in a CO2 burst similar to Figure 6 but no continuous activity with respect to CO oxidation. These observations imply that neither an initially present charged Au species nor atomic oxygen previously deposited on the surface is able to initiate a continuous oxidation, i.e., activation of molecular oxygen on npAu. Role of Ag. In order to identify transient changes in the chemical composition of the surface during the reaction, we performed in situ XPS measurements at varying pressures of up to 1 mbar (see Experimental Section for details), focusing in particular on Ag and its chemical state. In the case of the pristine sample, the Ag 3d5/2 signal is found at 367.6 eV, which suggests an oxidized state of the Ag at first sight, because Ag0 shows a typical binding energy of 368.0 eV and oxidized Ag species lead to negatiVe binding energy shifts relative to Ag0.51 It must be noted, however, that binding energy shifts in this direction have been reported for bimetallic Au-Ag particles, where they have been attributed to rehybridization effects in the alloy (reduction of electron density at the silver). In case of the Au 4f signal, both shape and binding energy point toward metallic Au.52

Wittstock et al. The presence of O2, CO, or a mixture of both has only a little effect on the Au 4d and Au 4f signals (data not shown), while significant changes are observed in the Ag 3d region (Figure 7). First, we performed ex situ XPS measurements following exposure to CO and O2: Spectra A, B, and C in Figure 7 show the Ag 3d signal in vacuo, after exposure to 1 mbar of CO for 1 h, and after exposure to 1 mbar of O2 for the same period of time, respectively. The differences between these three XP spectra are obviously very small, indicating that no permanent changes are induced by the presence of the reactive gases. However, spectra D and E in Figure 7, taken in the presence of 1 mbar of O2 and 1 mbar of a 1:1 mixture of O2 and CO, respectively, show significant transient changes in the Ag 3d signal. Most obvious is the broad shoulder in the range of lower binding energies (363-366 eV), which cannot be attributed to chemisorbed oxygen on Ag or to Ag2O or AgO species. These oxidized species show comparatively small Ag 3d5/2 binding energy shifts relative to Ag0 (Ag0, 368.0 eV; Ag2O, 367.7 eV; AgO, 367.2 eV; i.e.,less than 1 eV shift toward lower binding energy53), whereas the shoulder in spectra D and E is shifted by 1.5-4.5 eV relative to the Ag 3d main peak. Apparently, a specific Ag surfaces species (or several different surface species) is formed in the presence of O2 or O2/CO. At the present stage, we can only speculate on its chemical nature. A possible explanation for the large shift toward lower binding energy is a final state effect resulting from an Ag-adsorbate complex which is less efficiently screened than the majority of the Ag atoms embedded in the Au matrix. The appearance of the shoulder is fully reversible (Figure 7, spectrum F, taken after the in situ XPS measurements), indicating that the observed effects are not induced by X-rays or secondary electrons. Note that during and after the in situ XPS measurements (curves D-F in Figure 7, left) a shift of the main Ag 3d5/2 signal toward lower binding energy by ca. 0.5 eV occurs. Besides slow oxidation of the Ag in the subsurface region, beam-induced effects cannot completely be excluded. Discussion The high catalytic activity of nanoporous Au as evidenced by the results of the present work and previous reports in the literature is surprising at first sight. Many criteria that are usually considered to be mandatory for Au catalysis are not fulfilled by this material. Neither is the size of the ligaments in the range of a few nanometers nor is an oxidic support necessary to obtain extremely high conversion rates. In an attempt to assess the overall catalytic activity of the material, we calculated the turnover frequencies (TOF), i.e., the number of product molecules (CO2) formed per surface atom per second. For the maximal conversion of 12 × 10-6 mol/(s · m2) in oxygen excess (see Figure 3A), we obtain a TOF of 0.5 s-1 (1.4 × 1015 atoms/ cm2; Au(111)). Considering that the observed rate is limited by mass transport (η ) 0.16), the actual TOF is ∼3 s-1. This number is in good accordance with reported numbers for Au particles on a TiO2 support.54 As, however, it cannot be expected that every surface Au atom is taking part in the catalytic cycle, the real TOF per active surface site will be significantly higher. Of course, it is an obvious question whether the catalytic activity of npAu is based on the presence of Ag at the surface. It was Haruta who raised questions with respect to the statement of various reports in the literature that npAu is a “supportless” Au catalyst allowing studying the catalytic activity of “pure” Au. In this context, it is important to note that, according to the in situ XPS characterization, Ag is present at the surface in significant amounts and that its chemical state is transiently influenced by the presence of the reactive gas phase.



Figure 7. Left: Ag 3d XP spectra. (A) After activation, (B) after exposure to 1 mbar of CO for 1 h, (C) after exposure to O2 for 1 h, (D) in situ XPS at 1 mbar of O2, (E) in situ XPS at 1 mbar of CO/O2 (1:1), and (F) after the in situ XPS measurements. The arrow marks the position of a transient signal component which appears only in the presence of the gas phase. Right: Signal deconvolution of the in situ XP spectra D and E from the left panel. The peaks under the signal component between 364 and 366 eV are displayed for illustrative purposes; they may not necessarily represent the full complexity of these signals.

Several investigations carried out for particle systems showed that the catalytic activity of Au with respect to CO oxidation is significantly enhanced by alloying Au with Ag.22,28-32 In this case, even comparably large particles in the range of a few tens of nanometers were found to be highly active, whereas pure Au particles in that size regime are known to be inactive.55 Wang and co-workers ascribed the superior catalytic properties to the unique capability of the bimetallic system to adsorb and activate oxygen.29 On the basis of DFT calculations they could show that O2 is distinctly stronger adsorbed on the Au-Ag alloy than on either of the pure metals. In addition, coadsorption of CO and O2 on adjacent Au and Ag sitessoxygen on Ag and CO on Austurned out to be more preferred on the bimetallic system than on pure Au and Ag. Thus, according to the work of these authors, the presence of Ag on the surface of npAu should have a strong impact on the catalytic performance: strong coadsorption and concomitant activation of oxygen by transfer of electron density into the antibonding π* state of the oxygen molecule should cause high reactivity with respect to oxygen atom transfer.29,30 A number of observations substantiate the assumption that the catalytic activity of npAu is indeed caused by the simultaneous presence of Ag. In our investigation all samples had to be activated before catalytic activity was detected. Some samples even showed no activity at all when they were previously exposed to ambient air for a long period of time. Structural characterization of active and inactive samples revealed no distinct differences in the morphology of the samples. Since the morphology determines the number of undercoordinated Au sites, this finding implies that catalytic activity is not a unique feature of the special surface geometry of the material. Most likely it is dependent on the distribution of Ag and its chemical state. Note that the experiments with ozone as a supply of atomic oxygen showed that neither adsorbed hydrocarbons, which can be immediately reacted away with the atomic oxygen forming CO2, nor an oxidized state of gold is a likely explanation for catalytic (in)activity. Moreover, the observed reaction kinetics for O2 and CO are not compatible with most of the data reported for supported

Au particles.41,55 In particular, there is the exceptional reaction order for oxygen, which is zero, if mass transport limitation is properly taken into account. A reaction order of zero implies that adsorption sites are rapidly saturated, meaning that only a minority of the surface sites are able to actively chemisorb molecular oxygen. This is in line with the fact that Ag is significantly less abundant on the surface than Au. The mechanism as proposed by Wang et al. for bimetallic AuAg particles, which assumes oxygen activation on Ag and CO adsorption on Au, is also compatible with the observation that CO and O2 do not compete for adsorption sites. Furthermore, the variation of the reaction order of CO for different samples does not fit the picture of a pure gold catalyst, where CO and oxygen adsorb adjacently on Au and react by a Langmuir-Hinshelwood mechanism.56 Changing reaction orders have been reported, for example, for CO oxidation on Pt where oxygen forms islands with which mobile CO molecules can react at the perimeter.57,58 In a similar way, it can be expected that the distribution of Ag (forming larger or smaller islands with different perimeters) supplying the oxygen affects the kinetics and thus the order. Finally, the in situ XPS measurements reveal that the chemical state of parts of the Ag at the surface undergoes major changes, when the material is exposed to a mixture of O2 and CO at pressures of up to 1 mbar. This is in agreement with titration experiments showing that no stable oxygen species is involved in the reaction. The exact role of the transient Ag species detected by in situ XPS in the reaction is subject to further investigation. In summary, the catalytic behavior of the material in combination with the spectroscopic data and the relevant literature leads to the conclusion that npAu is not a pure, unsupported Au catalyst. Rather, it must be considered as a bimetallic catalytic system. Further studies are under way to clarify the correlation between the catalytic properties of npAu and the specific bimetallic composition in more detail.

5600 J. Phys. Chem. C, Vol. 113, No. 14, 2009 Conclusion We studied the morphology, surface composition, and catalytic properties of nanoporous gold (npAu), which was produced by chemical dealloying of a AuAg alloy. In particular, we focused on the reaction kinetics, the role of mass transport limitation, and transient changes in the Ag core levels under reaction conditions, using in situ XPS. The results strongly suggest that the catalytic activity of npAu is related to the presence of residual Ag, which is enriched at the surface of the nanopores by a factor of up to 10 compared to the bulk concentration. Acknowledgment. M.B. acknowledges a DFG travel grant for a research stay at the Lawrence Livermore National Laboratory. Work at LLNL was performed under the auspices of the U.S. DOE by LLNL under Contract No. DE-AC5207NA27344. The in situ XPS measurements were supported by the Deutsche Forschungsgemeinschaft under Grant GO1812/ 1-1. We thank Martina Hank and Andreas Bachmeier for technical assistance in some of the XPS measurements. References and Notes (1) Haruta, M.; Kobayashi, T.; Sano, H.; Yamada, N. Chem. Lett. 1987, 405. (2) Ishida, T.; Haruta, M. Angew. Chem., Int. Ed. 2007, 46, 7154. (3) Campbell, C. T. Science 2004, 306, 234. (4) Chen, M. S.; Goodman, D. W. Science 2004, 306, 252. (5) Abad, A.; Concepcion, P.; Corma, A.; Garcia, H. Angew. Chem., Int. Ed. 2005, 44, 4066. (6) Chang, F. W.; Yu, H. Y.; Roselin, L. S.; Yang, H. C.; Ou, T. C. Appl. Catal., A: Gen. 2006, 302, 157. (7) Sinha, A. K.; Seelan, S.; Tsubota, S.; Haruta, M. Top. Catal. 2004, 29, 95. (8) Sinha, A. K.; Seelan, S.; Tsubota, S.; Haruta, M. Angew. Chem., Int. Ed. 2004, 43, 1546. (9) Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Science 2003, 301, 935. (10) Rodriguez, J. A.; Liu, R.; Hrbek, J.; Perez, M.; Evans, J. J. Mol. Catal. A: Chem. 2008, 281, 59. (11) Liu, L. M.; McAllister, B.; Ye, H. Q.; Hu, P. J. Am. Chem. Soc. 2006, 128, 4017. (12) Weiher, N.; Bus, E.; Delannoy, L.; Louis, C.; Ramaker, D. E.; Miller, J. T.; van Bokhoven, J. A. J. Catal. 2006, 240, 100. (13) Sanchez-Castillo, M. A.; Couto, C.; Kim, W. B.; Dumesic, J. A. Angew. Chem., Int. Ed. 2004, 43, 1140. (14) Iizuka, Y.; Kawamoto, A.; Akita, K.; Date, M.; Tsubota, S.; Okumura, M.; Haruta, M. Catal. Lett. 2004, 97, 203. (15) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Nature 2001, 410, 450. (16) Biener, J.; Hodge, A. M.; Hayes, J. R.; Volkert, C. A.; ZepedaRuiz, L. ∏.; A Hamza, A. V.; Abraham, F. F. Nano Lett. 2006, 6, 2379. (17) Ju¨rgens, B.; Kubel, C.; Schulz, C.; Nowitzki, T.; Zielasek, V.; Biener, J.; Biener, M. M.; Hamza, A. V.; Baümer, M. Gold Bull. 2007, 40, 142. (18) Zielasek, V.; Ju¨rgens, B.; Schulz, C.; Biener, J.; Biener, M. M.; Hamza, A. V.; Baümer, M. Angew. Chem., Int. Ed. 2006, 45, 8241. (19) Xu, C.; Xu, X.; Su, J.; Ding, Y. J. Catal. 2007, 252, 243. (20) Xu, C. X.; Su, J. X.; Xu, X. H.; Liu, P. P.; Zhao, H. J.; Tian, F.; Ding, Y. J. Am. Chem. Soc. 2007, 129, 42. (21) Kameoka, S.; Tsai, A. P. Catal. Lett. 2008, 121, 337. (22) Haruta, M. ChemPhysChem 2007, 8, 1911. (23) Yim, W. L.; Nowitzki, T.; Necke, M.; Schnars, H.; Nickut, P.; Biener, J.; Biener, M. M.; Zielasek, V.; Al-Shamery, K.; Klu¨ner, T.; Baümer, M. J. Phys. Chem. C 2007, 111, 445. (24) Gottfried, J. M.; Schmidt, K. J.; Schroeder, S. L. M.; Christmann, K. Surf. Sci. 2003, 536, 206.

Wittstock et al. (25) Meyer, R.; Lemire, C.; Shaikhutdinov, S. K.; Freund, H. Gold Bull. 2004, 37, 72. (26) Gottfried, J. M.; Schmidt, K. J.; Schroeder, S. L. M.; Christmann, K. Surf. Sci. 2003, 525, 184. (27) Fernandez, E. M.; Ordejon, P.; Balbas, L. C. Chem. Phys. Lett. 2005, 408, 252. (28) Liu, J. H.; Wang, A. Q.; Chi, Y. S.; Lin, H. P.; Mou, C. Y. J. Phys. Chem. B 2005, 109, 40. (29) Wang, A. Q.; Chang, C. M.; Mou, C. Y. J. Phys. Chem. B 2005, 109, 18860. (30) Wang, A. Q.; Hsieh, Y.; Chen, Y. F.; Mou, C. Y. J. Catal. 2006, 237, 197. (31) Wang, A. Q.; Liu, J. H.; Lin, S. D.; Lin, T. S.; Mou, C. Y. J. Catal. 2005, 233, 186. (32) Lim, D. C.; Lopez-Salido, I.; Dietsche, R.; Kim, Y. D. Surf. Sci. 2007, 601, 5635. (33) Thompson, D. T. Platinum Met. ReV. 2004, 48, 169. (34) Croy, J. R.; Mostafa, S.; Hickman, L.; Heinrich, H.; Cuenya, B. R. Appl. Catal., A: Gen. 2008, 350, 207. (35) Saliba, N.; Parker, D. H.; Koel, B. E. Surf. Sci. 1998, 410, 270. (36) Nakatsuji, H.; Hu, Z. M.; Nakai, H.; Ikeda, K. Surf. Sci. 1997, 387, 328. (37) Kondarides, D. I.; Verykios, X. E. J. Catal. 1996, 158, 363. (38) Pantforder, J.; Pollmann, S.; Zhu, J. F.; Borgmann, D.; Denecke, R.; Steinruck, H. P. ReV. Sci. Instrum. 2005, 76, 9. (39) Hodge, A. M.; Hayes, J. R.; Caro, J. A.; Biener, J.; Hamza, A. V. AdV. Eng. Mater. 2006, 8, 853. (40) Chang, C. M.; Cheng, C.; Wei, C. M. J. Chem. Phys. 2008, 128, 4. (41) Bond, G. C.; Thompson, D. T. Gold Bull. 2000, 33, 41. (42) Petersen, E. E. Chem. Eng. Sci. 1965, 20, 587. (43) Thiele, E. W. Ind. Eng. Chem. 1939, 31, 916. (44) Thomas, J. M.; Thomas, W. J. Principles and Practice of Heterogeneous Catalysis; Wiley-VCH: Weinheim, New York, Basel, Cambridge, Tokyo, 1996. (45) Mizushima, Y.; Hori, M. J. Mater. Sci. 1995, 30, 1551. (46) Norberg, P.; Ackelid, U.; Lundstoem, I.; Petersson, L.-G. J. Appl. Phys. 1997, 81, 2094. (47) Sault, A. G.; Madix, R. J.; Campbell, C. T. Surf. Sci. 1986, 169, 347. (48) Gottfried, J. M.; Schmidt, K. J.; Schroeder, S. L. M.; Christmann, K. Surf. Sci. 2002, 511, 65. (49) Janssens, T. V. W.; Clausen, B. S.; Hvolbaek, B.; Falsig, H.; Christensen, C. H.; Bligaard, T.; Norskov, J. K. Top. Catal. 2007, 44, 15. (50) Quek, S. Y.; Friend, C. M.; Kaxiras, E. Surf. Sci. 2006, 600, 3388. (51) Weaver, J. F.; Hoflund, G. B. Chem. Mater. 1994, 6, 1693. (52) Lim, D. C.; Lopez-Salido, I.; Dietsche, R.; Bubek, M.; Kim, Y. D. Surf. Sci. 2006, 600, 507. (53) Hoflund, G. B.; Hazos, Z. F.; Salaita, G. N. Phys. ReV. B 2000, 62, 11126. (54) Berlowitz, P. J.; Peden, C. H. F.; Goodman, D. W. J. Phys. Chem. 1988, 92, 5213. (55) Bond, G. C.; Thompson, D. T. Catal. ReV.sSci. Eng. 1999, 41, 319. (56) Dekkers, M. A. P.; Lippits, M. J.; Nieuwenhuys, B. E. Catal. Today 1999, 54, 381. (57) Wintterlin, J.; Volkening, S.; Janssens, T. V. W.; Zambelli, T.; Ertl, G. Science 1997, 278, 1931. (58) Gland, J. L.; Kollin, E. B. J. Chem. Phys. 1983, 78, 963. (59) However, one should note that in the case of large adsorption energies (pronounced sticking) pure Knudsen diffusion is no longer an appropriate description of mass transport through nanopores. Indeed, measurements of oxygen diffusion through nanomachined straight pores coated with Pt demonstrated that the effective diffusion through nanopores can be 10 times slower than predicted by pure Knudsen diffusion (see ref 46). In the present case, an influence for CO might be expected, since a large fraction of the surface atoms of npAu should consist of undercoordinated step and kink sites with adsorption energies in the range 39-51 kJ/mol, corresponding to an average residence time between 10-2 and 1 ms (see ref 23).

JP808185V

Paper II: Novel nanoporous gold catalysts for green chemistry: selective gas-phase oxidative coupling of methanol at low temperature Wittstock, A.; Zielasek, V.; Biener, J.; Friend, C. M.; Bäumer, M Science, 2010, 327 (5963), 319-322 Link: http://www.sciencemag.org/cgi/content/full/327/5963/319

REPORTS OMe

OMe

MeO

a

CO2H H

OMe

MeO

CO2H

b–e

O

MeO

CO2t-Bu

C–H Olefination

4d

19

6d2

Reagents and conditions: (a) Pd(OAc)2, BQ, t-butyl acrylate, KHCO3, t-AmylOH, O2 (1 atm), 85 °C, 93%. (b) H2 (balloon), Pd/C, MeOH, rt. (c) CH2N2, Et2O, 0 °C, 93% (two steps). (d) KOt-Bu, Et2O, rt. (e) HCl/HOAc, 110 °C, 69% (two steps).

B HB TIPSO

CO2H HA Me

20

a

TIPSO

CO2H

b–f

CO2Me OH

MeO

CO2Et

PG-Controlled Position-Selective C–H Olefination

Me

Me

21-A

22

Reagents and conditions: (a) Pd(OAc)2, BQ, ethyl acrylate, KHCO3, t-AmylOH, O2 (1 atm), 85 °C, 77%, A:B = 10:1. (b) H2 (balloon), Pd/C, MeOH, rt. (c) Et3N•3HF, THF, rt. (d) MeI, K2CO3, acetone, reflux, 69% (3 steps). (e) KOt-Bu, Et2O, rt, 88%. (f) BrCCl3, DBU, CH2Cl2, rt, 81%.

C HB i-PrO

CO2H

MeO

HA OMe 9

a

Ligand-Controlled Position-Selective C–H Olefination

i-PrO

CO2H

MeO

b, c

i-PrO MeO

CO2R OMe

R = t-Bu

OH CO2R OMe

10-A

R = t-Bu

23

Reagents and conditions: (a) Pd(OAc)2, BQ, t-butyl acrylate, KHCO3, Boc-Ile-OH, t-AmylOH, O2 (1 atm), 85 °C, 86%, A:B = 23:1 (without ligand, A:B = 1.5:1). (b) (COCl)2, CH2Cl2, rt. (c) i-Pr2NEt, CH2Cl2, rt, 87% (two steps).

Fig. 4. (A) Synthesis of 7,8-dimethoxytetalin-2-one. (B) Synthesis of the naphthoic acid component of neocarzinostatin (1). (C) Synthesis of the naphthoic acid component of kedarcidin (3). References and Notes 1. J. F. Hartwig, Nature 455, 314 (2008). 2. A. R. Dick, M. S. Sanford, Tetrahedron 62, 2439 (2006). 3. O. Daugulis, V. G. Zaitsev, D. Shabashov, Q.-N. Pham, A. Lazareva, Synlett 2006, 3382 (2006). 4. L.-C. Campeau, D. R. Stuart, K. Fagnou, Aldrichim. Acta 40, 35 (2007). 5. J.-Q. Yu, R. Giri, X. Chen, Org. Biomol. Chem. 4, 4041 (2006). 6. B. D. Dangel, K. Godula, S. W. Youn, B. Sezen, D. Sames, J. Am. Chem. Soc. 124, 11856 (2002). 7. A. Hinman, J. Du Bois, J. Am. Chem. Soc. 125, 11510 (2003). 8. A. L. Bowie Jr., C. C. Hughes, D. Trauner, Org. Lett. 7, 5207 (2005).

9. H. M. L. Davies, X. Dai, M. S. Long, J. Am. Chem. Soc. 128, 2485 (2006). 10. A. S. Tsai, R. G. Bergman, J. A. Ellman, J. Am. Chem. Soc. 130, 6316 (2008). 11. K. Chen, P. S. Baran, Nature 459, 824 (2009). 12. E. M. Stang, M. C. White, Nat. Chem. 1, 547 (2009). 13. K. C. Nicolaou, P. G. Bulger, D. Sarlah, Angew. Chem. Int. Ed. 44, 4442 (2005). 14. B. M. Trost, S. A. Godleski, J. P. Genet, J. Am. Chem. Soc. 100, 3930 (1978). 15. P. S. Baran, E. J. Corey, J. Am. Chem. Soc. 124, 7904 (2002). 16. N. K. Garg, D. D. Caspi, B. M. Stoltz, J. Am. Chem. Soc. 126, 9552 (2004).

Nanoporous Gold Catalysts for Selective Gas-Phase Oxidative Coupling of Methanol at Low Temperature A. Wittstock,1 V. Zielasek,1 J. Biener,2* C. M. Friend,3* M. Bäumer 1* Gold (Au) is an interesting catalytic material because of its ability to catalyze reactions, such as partial oxidations, with high selectivities at low temperatures; but limitations arise from the low O2 dissociation probability on Au. This problem can be overcome by using Au nanoparticles supported on suitable oxides which, however, are prone to sintering. Nanoporous Au, prepared by the dealloying of AuAg alloys, is a new catalyst with a stable structure that is active without any support. It catalyzes the selective oxidative coupling of methanol to methyl formate with selectivities above 97% and high turnover frequencies at temperatures below 80°C. Because the overall catalytic characteristics of nanoporous Au are in agreement with studies on Au single crystals, we deduced that the selective surface chemistry of Au is unaltered but that O2 can be readily activated with this material. Residual silver is shown to regulate the availability of reactive oxygen.

A

n ever-increasing demand for resources enforces the need of sustainability in all arenas (1). This new challenge has

triggered a growing interest in a “green chemical industry,” especially for the production and processing of commodity chemicals (2, 3), which

www.sciencemag.org

SCIENCE

VOL 327

17. E. M. Beck, R. Hatley, M. J. Gaunt, Angew. Chem. Int. Ed. 47, 3004 (2008). 18. R. J. Phipps, M. J. Gaunt, Science 323, 1593 (2009). 19. Y.-H. Zhang, B.-F. Shi, J.-Q. Yu, J. Am. Chem. Soc. 131, 5072 (2009). 20. For early examples of Pd-catalyzed directed and nondirected arene C–H olefination, see (31–33). For an early example of Ru-catalyzed directed olefination, see (34). 21. T.-S. Mei, R. Giri, N. Maugel, J.-Q. Yu, Angew. Chem. Int. Ed. 47, 5215 (2008). 22. D.-H. Wang, T.-S. Mei, J.-Q. Yu, J. Am. Chem. Soc. 130, 17676 (2008). 23. P. Beak, V. Snieckus, Acc. Chem. Res. 15, 306 (1982). 24. For application of mono-N-protected amino acid ligands for Pd-catalyzed enantioselective C–H activation/C–C cross coupling, see (35). 25. A. G. Myers, Y. Horiguchi, Tetrahedron Lett. 38, 4363 (1997). 26. N. Ji, B. M. Rosen, A. G. Myers, Org. Lett. 6, 4551 (2004). 27. S. Kawata, S. Ashizawa, M. Hirama, J. Am. Chem. Soc. 119, 12012 (1997). 28. F. C. Görth, M. Rucker, M. Eckhardt, R. Brückner, Eur. J. Org. Chem. 2000, 2605 (2000). 29. Á. Gorka et al., Synth. Commun. 35, 2371 (2005). 30. For discussion of the importance of ligand development in Pd chemistry, see (36). 31. M. Miura, T. Tsuda, T. Satoh, S. Pivsa-Art, M. Nomura, J. Org. Chem. 63, 5211 (1998). 32. C. G. Jia et al., Science 287, 1992 (2000). 33. M. D. K. Boele et al., J. Am. Chem. Soc. 124, 1586 (2002). 34. S. Murai et al., Nature 366, 529 (1993). 35. B.-F. Shi, N. Maugel, Y.-H. Zhang, J.-Q. Yu, Angew. Chem. Int. Ed. 47, 4882 (2008). 36. R. Martin, S. L. Buchwald, Acc. Chem. Res. 41, 1461 (2008). 37. Supported by an A. P. Sloan Foundation fellowship ( J.-Q.Y.); predoctoral fellowships from NSF, the U.S. Department of Defense, the Scripps Research Institute, and the Skaggs Oxford Scholarship program (K.M.E.); and the Scripps Research Institute, National Institute of General Medical Sciences grant 1 R01 GM084019-02, NSF grant CHE-0910014, Amgen, and Eli Lilly & Co.

Downloaded from www.sciencemag.org on January 14, 2010

A

Supporting Online Material www.sciencemag.org/cgi/content/full/science.1182512/DC1 Materials and Methods Tables S1 to S8 NMR Spectra References 28 September 2009; accepted 12 November 2009 Published online 26 November 2009; 10.1126/science.1182512 Include this information when citing this paper.

is based on more efficient processes working under mild conditions (low temperatures and ambient conditions) and relying on cheap and abundant feedstock. In this context, gold (Au)– based catalysts have attracted considerable attention in the past decade because of their nontoxic nature and the ability to promote selective reactions at low temperatures. In particular, the potential of Au for partial oxidation reactions, such as the selective oxidation of alcohols (4–6) and hydrocarbons (7, 8), was demonstrated in numerous studies. Model studies on single-crystal Au provided molecular-scale insight into the activity of gold, showing that atomic oxygen is the key species that promotes a range of selective oxida1 Institute of Applied and Physical Chemistry, Universität Bremen, Bremen 28359, Germany. 2Nanoscale Synthesis and Characterization Laboratory, Lawrence Livermore National Laboratory (LLNL), Livermore, CA 94550, USA 3Department of Chemistry, Harvard University, Cambridge, MA 02138, USA.

*To whom correspondence should be addressed. E-mail: [email protected] (M.B.), [email protected] (C.M.F.), [email protected] ( J.B.)

15 JANUARY 2010

319

tive transformations (9). Its presence activates Au surfaces for reactions with organic species, including methanol (10), and also serves as the source of oxygen for other oxidation reactions, such as the transformation of olefins to epoxides. The comparatively weak adsorption of the partial oxidation products on Au allows them to desorb before being further oxidized. This delicate balance between Au’s oxidation power and its relatively weak interaction with partial oxidation products distinguishes Au from other catalytic metals, such as Pd and Pt. Yet, one technological impediment to the use of Au as a catalyst is O2 dissociation. Studies on single-crystal Au surfaces have shown that the dissociation probability for O2 is extremely low [10 µm within the CA support. Excluding some nonideal ALD process close to the surface, the particle size was below 5 nm. RBS analysis showed that the Pt loading in the CAs is extremely low with the areal density of Pt estimated to be ∼0.047 mg/cm2 in the sample prepared with 2 ALD cycles. Remarkably, even at such low loading levels, these materials exhibited high catalytic activity. Carbon monoxide oxidation studies demonstrated near 100% conversion of CO to CO2 at temperatures ranging from 150 to 250 °C with stability up to a power production of 160 Wg1- (6 × 10-4 mol g-1 s-1 at 250 °C). For the case of the 2 ALD cycle Pt-CA sample, the turnover frequency lies in the range of 429 s-1 depending on the assumptions made on mass transport. The technique demonstrated here can be applied to the design of new catalytic materials for a variety of applications, including fuel cells, hydrogen storage, pollution control, green chemistry, and liquid fuel production. Acknowledgment. S.F.B. acknowledges the Global Climate and Energy Project at Stanford University and the donors of The American Chemical Society Petroleum Research Fund for support. Work at LLNL was performed under the auspices of the U.S. DOE by LLNL under Contract DE-AC52-07NA27344. References (1) Bell, A. T. Science 2003, 299, 1688. (2) Campbell, C. T. Science 2004, 306, 234. (3) Raimondi, F.; Scherer, G. G.; Kotz, R.; Wokaun, A. Angew. Chem., Int. Ed. 2005, 44, 2190. (4) Wang, G. X.; Sun, G. Q.; Zhou, Z. H.; Liu, J. G.; Wang, Q.; Wang, S. L.; Guo, J. S.; Yang, S. H.; Xin, Q.; Yi, B. L. Electrochem. Solid State 2005, 8, A12. (5) Liu, H. S.; Song, C. J.; Zhang, L.; Zhang, J. J.; Wang, H. J.; Wilkinson, D. P. J. Power Sources 2006, 155, 95. (6) Du, H. D.; Gan, L.; Li, B. H.; Wu, P.; Qiu, Y. L.; Kang, F. Y.; Zeng, Y. Q. J. Phys. Chem. C 2007, 111, 2040. (7) Saquing, C. D.; Kang, D.; Aindow, M.; Erkey, C. Microporous Mesoporous Mater. 2005, 80, 11. (8) Kucheyev, S. O.; Biener, J.; Wang, Y. M.; Baumann, T. F.; Wu, K. J.; van Buuren, T.; Hamza, A. V.; Satcher, J. H.; Elam, J. W.; Pellin, M. J. Appl. Phys. Lett. 2005, 86, 083108. (9) Baumann, T. F.; Biener, J.; Wang, Y. M. M.; Kucheyev, S. O.; Nelson, E. J.; Satcher, J. H.; Elam, J. W.; Pellin, M. J.; Hamza, A. V. Chem. Mater. 2006, 18, 6106. (10) Biener, J.; Baumann, T. F.; Wang, Y. M.; Nelson, E. J.; Kucheyev, S. O.; Hamza, A. V.; Kemell, M.; Ritala, M.; Leskelä, M. Nanotechnology 2007, 18, 055303. Nano Lett., Vol. 8, No. 8, 2008

(11) Kucheyev, S. O.; Biener, J.; Baumann, T. F.; Wang, Y. M.; Hamza, A. V.; Li, Z.; Lee, D. K.; Gordon, R. G. Langmuir 2008, 24, 943. (12) Elam, J. W.; Xiong, G.; Han, C. Y.; Wang, H. H.; Birrell, J. P.; Welp, U.; Hryn, J. N.; Pellin, M. J.; Baumann, T. F.; Poco, J. F.; Satcher, J. H. J. Nanomaterials 2006, 64501. (13) Leskela¨, M.; Ritala, M. Angew. Chem., Int. Ed. 2003, 42, 5548. (14) Baumann, T. F.; Worsley, M. A.; Han, T. Y.; Satcher, J. H. J. NonCryst. Solids 2008, 354, 3513. (15) Aaltonen, T.; Ritala, M.; Sajavaara, T.; Keinonen, J.; Leskelä, M. Chem. Mater. 2003, 15, 1924.

Nano Lett., Vol. 8, No. 8, 2008

(16) Xue, Z.; Thridandam, H.; Kaesz, H. D.; Hicks, R. F. Chem. Mater. 1992, 4, 162. (17) Doolittle, L. R. Nucl. Instrum. Methods Phys. Res., Sect. B 1985, 9, 344. (18) Santra, A. K.; Goodman, D. W. Electrochim. Acta 2002, 47, 3595. (19) It should be noted that the sample did not fill the area of the frit, thus a portion of the gas stream by-passed the sample; this was the case especially during measurements where smaller amounts of catalyst were used.

NL801299Z

2409