WO2013038188A1

WO2013038188A1 - Optimisation of catalyst alloy surfaces

Info

Publication number: WO2013038188A1
Application number: PCT/GB2012/052264
Authority: WO
Inventors: Paul Alexander John BAGOT; Tong Li; Shik Chi Edman Tsang; George David William Smith; Emanuelle MARQUIS
Original assignee: Isis Innovation Limited
Priority date: 2011-09-13
Filing date: 2012-09-13
Publication date: 2013-03-21
Also published as: GB201115783D0

Abstract

The present invention provides methods for altering the surface composition of an alloy, which comprise the following steps: a)providing an alloy comprising at least two metals; b)exposing said alloy to an oxidising agent whereby to cause oxidation of at least one of said metals to form at least one metal oxide on a surface of the alloy; and c)exposing the resulting alloy to a reducing agent whereby to reduce said metal oxide; wherein said metals are each independently selected from groups 6, 7, 8, 9, 10 and 11 of the Periodic table and include at least one metal selected from platinum (Pt), palladium (Pd), rhodium (Rh) and ruthenium (Ru), and further wherein said alloy is not a platinum-rhodium (Pt-Rh) alloy. Surface- modified alloys obtainable by such methods are also provided by the invention and find particular use as catalysts and in catalytic converters and fuel cells.

Description

Optimisation of Catalyst Allov Surfaces

This invention relates to a method for controlling the surface chemical composition of metal alloys, particularly catalytic nanoalloys, to alloy materials produced by the method and their use.

In particular, the invention relates to a method capable of altering the surface composition of an alloy in which one or more metals (e.g. catalytica!ly active metals) can be driven to the alloy surface from within the bulk structure. The resulting alloy materials, enriched in at least one metal at their surface, find use in a wide range of applications, especially as catalysts. These find particular use in catalytic converters and In. fuel cells.

Platinum (Pt) and palladium (Pd) based alloy catalysts are used in applications such as fuel cells, automobile exhaust pollution reduction and steam reforming. With the addition of rhodium (Rh) and ruthenium (Ru) the catalytic properties are dramatically enhanced due to either synergistic effects or changes in electronic properties. However, the platinum group metals are extremely expensive, especially rhodium. Optimal ways of utilising each metal are thus needed in order to maximise catalytic efficiency at relatively low cost.

Current methods of producing automobile exhaust catalysts have very little control over the final surface structure, using metal salt solutions to permeate all the pores of the support material before heating to decompose the salts into nanometre sized particles.

Attempts have been made to 'engineer' the surface structure of catalyst nanoparttoles in order to maximise the presence of the expensive catalytic metals at the active surface. For example, chemical synthesis methods have been used to create core-shell structures in which a shell material including a catalytlcally active metal surrounds a less expensive core material. Typically, such methods involve two stages in which a core is Initially produced followed by coating using known coating methods. Examples of such coating methods include chemical vapour deposition (CVD), plasma methods, physical vapour deposition (PVD), reactive sputtering, electrodeposition, laser pyrolysis, sol gel techniques, redox displacement processes, and co-deposition with metal salts. However, there can be problems with reproducibility and such methods require careful attention and control of the process throughout. Due to the complexity of such methods, their costs are also high.

Mu et al. (Applied Surface Science 255: 7296-7301 , 2009) and Tao et al. (Science 322: 932-934, 2008) describe the use of oxidation/reduction cycles to drive reversible structural changes to the surface of Pt-Ni and Rh-Pd nanoparticles, respectively. Oxidative conditions result in the formation of a core-shell structure enriched at the surface in one metal oxide (either NiO or RhO_y). Subsequent reduction under the same temperature conditions reduces the metal oxide whilst also reversing the observed structural change, restoring the original surface composition. No permanent "fixing" or "freezing" of a surface enriched in elemental Ni or Rh is observed.

More recently, the inventors have found that a sequential oxidation and reduction treatment carried out on a platinum-rhodium (Pt-Rh) alloy is capable of producing a permanent core-shell structure (see Tong Li et al., "Characterisation of oxidation and reduction of a platinum-rhodium alloy by atom-probe tomography", Catalysis Today, published online 22 April 201 1

(http://dx.doi.Org/10.1016/j.cattod.201 1 .03.046). Exposure of the surface of the Pt-Rh alloy to oxygen at elevated temperatures was found to favour creation of rhodium oxide which was then drawn up to the surface from the bulk of the alloy. This resulted in the formation of a complete layer or 'shell' of rhodium oxide on the alloy surface. Following the oxidation treatment, reduction of the alloy in hydrogen was found to remove all oxide, but retain a thin, almost pure, layer of the element rhodium on the surface. Up until now, it had been assumed that the higher heat of formation of the rhodium oxide played a key role in favouring creation of this oxide and driving this up to the surface. However, the inventors' latest research, which is the subject of this application, suggests that this is not the only factor which influences the surface structure. As such, the findings which are presented herein are unexpected in light of this earlier work.

In working with other metal alloys, the inventors have now surprisingly found that an oxidation process is capable of producing spatially separated 'islands' of at least one metal oxide on the surface of the alloy. Following reduction, this 'island' structure induced under oxidation is retained to give a material (e.g. a catalytically active particle) which differs from the core-shell structure previously produced when using a platinum-rhodium (Pt-Rh) alloy. The 'island' structures retained following reduction are in the form of islands of elemental metal, rather than islands of metal oxide.

In particular, the inventors have found that separate 'island'-like structures or nanoregions which are locally rich in one or more of the different alloying elements can be produced when carrying out oxidation/reduction treatment on other metal alloys. As will be appreciated, the ability to drive more than one distinct active species (e.g. catalytically active species) to the surface during the same overall reaction cycle is a significant advance in seeking new methods to 'engineer' the surface structure of alloy materials using simple treatment methods.

The methods which are herein described have significant potential in producing catalytic nanoparticles having unique surface structures. By varying the initial alloying elements, along with fine-tuning the oxidation/reduction exposure conditions, it is possible to produce 'nano-engineered' catalyst structures from initially homogeneous alloys, with each structure optimised for use in a specific catalytic application. To the extent that these drive expensive, highly active metals to the alloy surface without having to increase the bulk concentration of these in the original alloy, such methods also serve to reduce the cost of catalytic materials.

In a first aspect, the invention provides a novel method for controlling the surface chemical composition of an alloy. This method comprises the following steps: a) providing an alloy comprising at least two metals;

b) exposing said alloy to an oxidising agent whereby to cause oxidation of at least one of said metals to form at least one metal oxide on a surface of the alloy; and

c) exposing the resulting alloy to a reducing agent whereby to reduce said metal oxide; wherein said metals are each independently selected from groups 6, 7, 8, 9, 10 and 1 1 of the Periodic table and include at least one metal selected from platinum (Pt), palladium (Pd), rhodium (Rh) and ruthenium (Ru),

and further wherein said alloy is not a platinum-rhodium (Pt-Rh) alloy.

Other than Pt-Rh, any metallic alloy containing two or more metals may be treated in accordance with the method herein described. Binary and ternary alloys are preferred. The choice of metal elements will depend on the desired end use of the surface-modified alloy. A preferred use is as a catalyst material and so the alloy will preferably comprise at least one catalytically active metal. Catalytically active metals include platinum, palladium, ruthenium, rhodium, iridium, osmium, molybdenum, tungsten, iron and nickel. Typically, the catalytically active metal will be either platinum or palladium.

In a preferred embodiment, the alloy will comprise at least one platinum group metal selected from palladium, platinum, rhodium, iridium, ruthenium and osmium. Binary and ternary alloys of platinum group metals are particularly preferred.

Especially preferred alloys are those which include platinum and/or palladium optionally in combination with one or more of rhodium and ruthenium.

Examples of suitable alloys include Pd-Rh, Pt-Ru, Pd-Ru, Pd-Pt, Pd-lr, Pt-lr, Pt-Co, Pd-Co, Rh-Ru, Pt-Pd-Rh, Pt-Pd-Ru, Pt-Rh-Ru, Pt-Rh-lr and Pd-Rh-Ru. Particularly preferably, the alloy is selected from Pd-Rh, Pt-Pd-Rh and Pt-Ru.

In the same way that the metallic elements present in the alloy may be chosen according to the intended end use of the material, their relative atomic amounts may similarly be varied as desired. Typically, in order to reduce the cost of the alloy, it is envisaged that the more expensive metals such as rhodium and/or ruthenium will be provided in lower atomic amounts. For example, these may be present in an amount of 30at.% or less, preferably 25at.% or less, more preferably 20at.% or less. Rhodium and ruthenium will each typically be present in an amount of less than 15at.%, e.g. less than 10at.%. Metals such as palladium and/or platinum will generally comprise the bulk (i.e. major component) of the alloy composition. For example, these metals will generally be provided in an amount of at least 70at.%, more preferably at least 80at.%, e.g. at least 90at.%. Alloy materials suitable for treatment in accordance with the method herein described are readily available commercially, for example from Aldrich, Sigmund Cohn, Johnson Matthey and Alfa Aesar.

Specific examples of embodiments of the alloy composition include:

Pd - 6.4at.% Rh;

Pd - 17at.% Rh;

Pt - 23.9at.% Rh - 9.7at.% Ru;

Pt- 17.4at.%Rh - 13.9at.% Ir;

Pt - 8.9at%Ru;

Pt - 9at.%Ru;

Pt - 10at.%lr; and

Pt - 9at.%Pd - 9at.%Rh.

The alloy composition may be crystalline. Where this is crystalline, this may be subject to treatment according to the method herein described at one or more of its crystal surfaces. The surface chemical composition of at least the {001 } and/or the {1 1 1 } crystallographic faces may be altered by such treatment. Alternatively, multiple different crystallographic faces may simultaneously be subjected to the oxidation/reduction method.

Typically, the alloy will be provided in particulate form, e.g. having nano

dimensions. Particulates ranging in size up to 1 μm may be employed, however those having a mean particle diameter of less than 1 μm are generally preferred, especially for catalytic applications. Nanoparticulates are particularly preferred. Typically, these will have a mean particle diameter of less than about 500 nm, preferably less than about 200 nm, more preferably less than about 100 nm, yet more preferably less than about 50 nm, e.g. less than about 15 nm or less than about 10 nm. Particulates having a mean particle diameter in the range of from 5 to 50 nm are particularly preferred, e.g. for use as catalyst materials. When provided in particulate form, the alloy particles may be polydisperse or monodisperse, but preferably will be monodisperse.

The method of the invention may give rise to alloys in which the surface has discrete regions rich in at least one metal, e.g. comprising two or more metals. Methods in which two or more metal oxides (e.g. two) are formed on the surface of the alloy during the oxidation step are preferred. Following reduction, these give rise to alloy structures having separate nanoregions or 'islands' of two or more different metals.

Any substance which yields atomic oxygen on contact with the alloy surface under the conditions employed in carrying out the method herein described may be used as an oxidising agent. Suitable oxidising agents may readily be determined by those skilled in the art and include oxygen (0₂), air, nitric oxide (NO), nitrous oxide (N₂0) and combinations thereof. In a preferred embodiment, the oxidising agent will be employed in the form of a gas (under the temperature and pressure conditions employed). Suitable gaseous oxidising agents include oxygen and nitrous oxide. Oxygen is especially preferred.

Gaseous oxidising agents may be employed at a pressure of from 5 mbar to 5 bar, preferably from 10 mbar to 1 bar. Suitable flow rates of from 1 to 100 cm³/minute, preferably 10 to 50 cm³/minute, may be used.

The oxidation step is preferably carried out at a temperature of up to 1500 K, preferably at a temperature in the range of from 350 K to 1000 K, more preferably in the range of from 700 K to 1000 K, e.g. in the range of from 800 K to 900 K.

The duration of the oxidation process will vary according to the choice of alloy material and the surface changes required, but will generally range from about 5 minutes to about 8 hours, preferably from about 10 minutes to about 5 hours, e.g. from about 15 minutes to about 4 hours. The minimum oxidation time needed to achieve the desired 'island' structures on the alloy surface will vary depending on the constituent elements of the alloy, but can readily be selected by those skilled in the art. At relatively short oxidation times, the desired oxide-rich 'islands' may not form and in some cases oxidation may need to be carried out for a minimum period (e.g. about 10 minutes) to achieve the desired results.

Suitable reducing agents may readily be determined by those skilled in the art and include hydrogen, carbon monoxide, ammonia and combinations thereof, as well as citrates (e.g. potassium citrate, sodium citrate, ammonium citrate), hydrazine, sodium hypophosphite, alkali metal borohydrides, formaldehyde, formates, acetates, oxalates, sulfanilates such as sodium hydroxymethanesulfinate, and monohydric or dihydric alcohols such as ethanol and ethylene glycol. In a preferred embodiment, the reducing agent will be employed in the form of a gas (under the temperature and pressure conditions employed). Suitable gaseous reducing agents include hydrogen, carbon monoxide and ammonia. Especially preferred is hydrogen.

Gaseous reducing agents may be employed at a pressure of from 5 mbar to 5 bar, preferably from 10 mbar to 1 bar. Suitable flow rates of from 1 to 100 cm³/minute, preferably 10 to 50 cm³/minute, may be used.

Reduction may be carried out at a temperature of up to 1000 K, preferably at a temperature of from 450 K to 750 K, e.g. from 600 K to 700 K. Typically, reduction will be carried out at a lower temperature than the oxidation step. For example, the reduction step may be carried out at a temperature at least 100 K lower than the oxidation step, e.g. from 100 K to 300 K lower, preferably about 200 K lower.

Particularly preferred conditions may involve oxidation at about 873 K followed by reduction at about 673 K.

The duration of the reduction process will be dependent on the extent of oxidation and may readily be varied according to need. Reduction should be carried out such that this essentially removes all oxides, but importantly retains the segregation induced under oxidation. Typically, the reduction step will be shorter than the oxidation step. Reduction may be carried out for a period of from 30 minutes to 4 hours, preferably for a period of from 45 minutes to 2.5 hours, more preferably from 1 hour to 2 hours.

Prior to carrying out the oxidation step, the alloy may be subject to additional treatment steps, for example to clean the surface, remove adsorbates and/or remove any residual surface oxides which may have formed and which might otherwise impede effective oxidation. Appropriate cleaning methods include, for example, electropolishing (e.g. in molten salt mixtures, preferably molten salt mixtures of NaNO₃ and NaCI, more preferably a 4:1 molten salt mixture of NaN0₃ and NaCI); dipping in warm dilute nitric acid; and field-evaporation (for example, using a Field Ion Microscope, e.g. under a He atmosphere).

The materials which are produced as a result of carrying out the methods herein described are physically different from those of the prior art. Such materials accordingly form a further aspect of the invention.

In a further aspect, the invention thus provides alloy materials (herein also referred to as surface-modified alloys) obtained or obtainable by a method as herein described. Such materials will typically have at least one surface comprising 'island' structures or nanoregions of at least one metal. These islands will preferably comprise at least one metal selected from Pt, Pd, Rh and Ru.

The 'island'-structures herein described comprise discrete nanoregions on the surface of the alloy. Such nanoregions may comprise mono-, bi- or multilayers of at least one elemental metal (e.g. a single metal species). Typically, the nanoregions will have an average diameter of up to 20 nm, e.g. between 2 and 10 nm.

Surface-modified alloy materials which are provided in the form of particles, especially nanoparticles are particularly preferred.

Also provided in accordance with the invention is a surface-modified alloy comprising two or more metals,

wherein said alloy has at least one surface comprising island structures (e.g.

nanoregions) of at least one metal,

wherein said metals are each independently selected from groups 6, 7, 8, 9, 10 and 1 1 of the Periodic table and include at least one metal selected from platinum (Pt), palladium (Pd), rhodium (Rh) and ruthenium (Ru),

and further wherein said alloy is not a platinum-rhodium (Pt-Rh) alloy.

The surface-modified alloys produced by the method of the invention can be used for many different purposes, but primarily are of interest in relation to their catalytic properties. In particular they may be used in a number of different heterogeneously catalysed reactions. These include, for example, aminations, hydrogenations, dehydrogenations, dehydrocyclisations, dehydroxylations, oxidations and epoxidations of organic molecules. They are particularly suitable for use as catalysts in hydrogenation and oxidation reactions, especially selective

hydrogenation reactions and partial oxidations. Catalyst materials comprising a surface-modified alloy as herein described form a preferred aspect of the invention.

Other uses for the catalytic materials are as electrocatalysts in electrochemical cells, such as fuel cells (e.g. low-temperature fuel cells). The catalysts may be used in all types of fuel cell, including hydrogen-containing fuel cells and those which employ liquid hydrocarbon fuels, such as methanol. When used in methanol fuel cells, these may be employed as methanol reformation catalysts. Those alloys containing platinum and/or palladium are preferred for use in fuel cell applications.

The catalytic materials also find use in catalytic converters in which these are used to breakdown the harmful by-products present in exhaust gases, including:

reduction of nitrous oxides to nitrogen and oxygen; oxidation of carbon monoxide to carbon dioxide; and oxidation of unburned hydrocarbons. Materials which comprise a platinum group metal, especially platinum or palladium, optionally in combination with rhodium are particularly suitable for this purpose. Platinum is widely used as a catalyst for the oxidation of carbon monoxide and hydrocarbons. The addition of rhodium is beneficial due to its unique properties for NO_x conversion. Palladium is very active for CO and hydrocarbon oxidation and also for NO_x reduction (although less active for the latter than rhodium).

In any of the uses herein described, the catalytic materials may be provided on a suitable support material. The choice of support material will depend on the desired end use. Support materials are well known in the art and will preferably have a high surface area which is provided by an inert, porous, open structure in which the catalytic particles may be embedded, including, but not limited to, inorganic materials such as silicon dioxide (silica), silica gel, aluminium oxide (alumina), titanium dioxide (titania), zirconium dioxide (zirconia), aluminium silicates, carbon (e.g. carbon black, graphitized carbon, graphite or activated carbon), silica-alumina, calcium carbonate, barium sulphate, zeolites, interstitial clay, metal carbides, boron doped diamond, kieselguhr, diatomaceous earth, bentonite, zeolites, and combinations thereof. For use as electrocatalysts, the catalytically active particles will preferably be bound to or otherwise embedded in an electrically conductive support, such as carbon black, graphitized carbon, graphite or activated carbon. Carbon, particularly carbon black, is a preferred support material due to its high surface area and porosity as well as its electrical conductivity. Other suitable support materials include conductive metals and metal oxides such as Sr-doped LaMn0₃ (commonly known as LSM), alumina, zirconia, titania, and zeolites.

Other uses for the surface-modified alloys are as catalysts for automobile exhaust emission control, ammonia oxidation, methanol oxidation, methanol reformation, or in Fischer-Tropsch catalysis.

In a yet further aspect the invention provides a catalytic converter or fuel cell comprising the catalytically active material as herein described, especially in nanoparticulate form. In the fuel cell, the nanoparticulates may be bound to an electrically conductive support and may form either the cathode and/or the anode of the cell. For example, in a hydrogen fuel cell, a Pt-containing catalyst may be used to oxidise hydrogen gas at the anode of the cell. Alternatively, at the cathode of the cell, a Pt-containing catalyst may be used to catalyse the reduction of oxygen. Pt-based alloy catalysts containing rhodium and ruthenium are particularly suitable for use as an anode in a fuel cell.

The invention is illustrated further in the following non-limiting examples and in the attached Figures, in which:

Figure 1 shows APT reconstructions of oxidized Pd-Rh specimens showing spatial distributions of PdOx, RhOx, Pd and Rh ionic species for varying oxidation times in 1 bar O₂;

Figure 2 shows a proximity histogram obtained from a specimen of Pd-Rh oxidized at 873K for 10 minutes;

Figure 3 shows iso-concentration surfaces, proximity histograms, and

concentration profiles obtained from various Pd-Rh specimens; Figure 4 shows a) an APT reconstruction of a Pd-Rh specimen oxidized at 873K for 20 mins in 1 bar O₂ and reduced at 623K for 2 hrs in 1 bar H₂; and b) a proximity histogram of a Rh isosurface of the specimen;

Figure 5 shows the top view (X-Y plane) of APT reconstructions of atom

probe Pd-Rh specimens (Z = 3 nm) treated under various oxidation- reduction conditions;

Figure 6 shows APT reconstructions showing overall, oxide and metal ionic species of a) Pt-Rh and b) Pd-Rh alloys;

Figure 7 shows APT reconstructions of Pt-Rh-Ru specimens after oxidation at

873K in 1 bar O₂ for varying oxidation times;

Figure 8 shows a proximity histogram and 1 d concentration profiles obtained from Pt-Rh-Ru specimens;

Figure 9 shows APT reconstructions and a 1 d concentration profile of a

Pt-Rh-Ru specimen; and

Figure 10 shows the top view of APT reconstruction of Pt-Rh-Ru specimens oxidized at 873K in 1 bar O₂ for varying oxidation times.

In the following examples, atom probe tomography (APT) is used for the analysis of the various alloy materials. APT is ideally suited for the atomic-scale analysis of metal/metal oxide systems. The geometry of the specimen is advantageous, as the needle-shaped apex has a radius of curvature of <100nm. The apex region is therefore a good model for a catalyst nanoparticle; approximately double its size and with a similar curved structure.

Example 1 - Preparation of surface-modiifed Pd-Rh alloy

Materials and Methods

Pd-6.4at%Rh alloy in the form of as-drawn wire (0.1 mm diameter) was obtained from Johnson Matthey Materials Limited. APT specimens were prepared in an electrolyte solution of 10% perchloric acid and 90% acetic acid in a first stage. The specimens were further sharpened in a mixture of 2% percholoric acid and 98% butoxyethanol under a second electropolishing stage. The resulting specimens were cleaned of any residual surface oxide by field evaporation of the first few atomic layers before being exposed to any oxidizing treatment. For all specimens, the [100] direction was parallel to the specimen axis. After cleaning, specimens were oxidized at 873K for 10, 30 or 180 mins in 1 bar oxygen at a flow rate of 35 cc/minute. In a separate set of experiments, several specimens were oxidized at 873K for 20 mins in 1 bar oxygen, and then reduced at 623K for 120 mins in 1 bar of hydrogen.

Specimens were characterized by APT using a Cameca LEAP 3000X-HR™ instrument. Several specimens were analysed for each unique thermal/gas exposure. The analyses were carried out in laser pulsing mode with a specimen temperature of 55K, target evaporation rate of 2 ions per 1000 pulses, pulse rate of 160kHz, pulse energy of 0.5nJ and a laser spot size of approximately 10 μm. The standing voltage was automatically increased in order to maintain a constant evaporation rate. Compositional analyses of the as-received materials by APT yielded a Rh content of 6.4±0.3at.%, in excellent agreement with the nominal composition of 6.4at.%. The different reconstruction parameters (image

compression factor, evaporation field, radius and field factor) can have a significant influence on the scale and aspect ratio of the 3D reconstructions and therefore any quantitative measurement of distances. In this study, the values of the image compression factor and the field factor were kept as the default values (1 .65 and 3.3) of the reconstruction procedure.

Results

Morphology observations

Pd-Rh specimens were oxidized at 873K for different times in order to investigate the nature of the oxides forming in this system. Two types of oxidized regions were observed on Pd-Rh specimens oxidized for 10 mins, as shown in Figure 1 a): a thin surface at the apex of the specimen layer and oxide islands growing in from the shank. More severe oxidation was observed on Pd-Rh oxidized for 30 mins. The corresponding atom maps in Figure 1 b) clearly show that the increased oxidation time enhances the formation of oxides. A separate Rh-oxide rich island is also apparent within the dominant PdO oxide layer. As the oxidation time increases to 180 mins, the analyzed region of the specimen demonstrates extensive oxidization. Several Rh-rich regions were again observed, as apparent in the corresponding PdO_x/RhO_x/O atom maps in Figure 1 c).

Compositions of oxide phases

Detailed analyses of the oxide phases and oxide/metal interfaces were mainly performed using proximity histograms or 1 D concentration profiles. The number of atoms exported from the proximity histogram or 1 D concentration profile was used to calculate an average concentration of each element in the surface layer.

The composition of the thin surface layer on Pd-Rh oxidized for 10 mins as shown in Figure 1 a) was analyzed using a ladder diagram by placing a 8nm analysis cylinder parallel to the surface. The measured concentrations from this were 3.9±0.2at.% Rh, 92.7±1.0at.% Pd, and 3.3±0.2at.% O. The oxygen concentration is low in the surface Oxide' and it is therefore probably a simple oxygen-chemisorbed layer. This layer is slightly depleted in Rh with a Rh/(Rh+Pd) value of 4.1 %.

Conversely, the region directly underneath is slightly enriched in rhodium with a Rh/(Rh+Pd) value of 7.8% (nominal bulk composition is 6.4at.% Rh). The composition of the relatively larger shank oxide region is analysed within the black rectangle in Figure 1 a) by proximity histogram, using an atomic oxygen iso- concentration surface of 30at.%. The measured composition in the oxide (marked as 1 ^a in Figure 1 a)) is 46.6±0.2at.% Pd - 2.7±0.1 at.% Rh - 50.7±0.2at.% O, determined as an average over the black arrowed region in Figure 2. The ratio between oxygen and metal is approximately 1 .02:1 , very close to the stoichiometry of PdO oxide.

As the oxidation time increases to 30 mins, Figure 1 b), a proximity histogram (with an oxygen isosurface of 27.0at.%) also determines the oxide to have PdO stoichiometry, with an overall composition of 47.2±0.3at.% Pd - 2.0±0.1 at.% Rh - 50.8±0.3at.% O. It shows that Rh is depleted in the PdO oxide layer and in the transition layer. The 15.0at.% atomic Rh iso-concentration surfaces were plotted in Figure 3a) and the proximity histogram of the atomic Rh isosurface in the dashed line region is also shown in Figure 3b). The composition in this Rh-rich oxide island (marked as 1 ) is measured in the black arrowed region in Figure 3b) to be 21.5±0.7at.% Pd - 25.1 ±0.8at.%Rh - 53.4±1.1 at.% O, consistent with a

stoichiometry for the Rh-rich oxide island as (Rh₁Pd₁)O₂.

For oxidation over 180 mins, detailed analysis of the atom maps revealed localised enrichment of Pd-oxides in addition to the Rh-oxides seen for shorter oxidation times. In order to distinguish these, which are not immediately apparent in Figure 1 c), 1 1 .0at.% Rh iso-concentration surfaces and 57.0at.% Pd iso-concentration surfaces are plotted in Figure 3c). A 5nm 1 D analysis cylinder was placed through the distinct regions, cutting through Pd oxide (3*), Rh-rich oxide (1 *), Pd oxide (3*), Pd-rich oxide (2*) and again into the Pd oxide region (3*). It was calculated in the corresponding 1 D composition profile in Figure 3d) that the Rh concentration in the Rh-rich oxide (1 *) is approximately 25.0±1 .1 at.%, substantially higher than the Rh content of 2.8±0.2at.% seen in the PdO phases (3*) (both calculated from the black arrowed regions under the (RhPd)0₂ and PdO regions in the concentration profile of Figure 3b)). The ratio of the total number of Rh and Pd atoms is approximately 1 .18:1 in the 1 * region, therefore the stoichiometry of the Rh-rich oxide is again (RhPd)O₂. In the Pd-rich oxide regions (2*), the Pd concentration was measured to be 68.8±2.2at.% with an oxygen concentration of 29.1 ±1 .5 at.% (averaged over the black arrowed region in the "Pd₂O" band in Figure 3b)).

Oxidation/reduction cycle

As Pd-Rh was found to oxidize rapidly at 873K, only a short oxidation of 20 mins in 1 bar O₂ was firstly applied to the specimens in this separate study. The subsequent reduction treatment was at 623K for 2 hrs in 1 bar H₂. Following reconstruction of the data, Figure 4a), it is clear that some regions remain enriched in Rh following the reduction, as shown using 26.0at.% Rh iso-concentration surfaces (red). The Rh-rich metallic clusters have an average composition of 54.1 ±2.0at.% Rh -45.9±1 .8at.% Pd content of, as calculated over the right black arrowed region from the proximity histogram in Figure 4b). The ratio of Rh:Pd in both regions(l ') is nearly 1 :1 , equivalent to the ratio seen in the (RhPd)O₂ oxide clusters (labelled 1 *) in Figure 3c). Furthermore, the Rh:Pd ratio in the metal matrix (labelled 2') is calculated within the left black arrowed region in Fig. 4b) to be approximately 0.06:1 , which is comparable with the Rh:Pd ratio in the stable PdO phase (1 ^a in Fig.1 a)) noted following oxidation at 873K for 10 mins. Consequently these results show that reduction at 623K for 2 hours does not affect the distribution of metallic species, retaining the separate Rh-rich and Pd-rich regions which formed under oxidation.

Discussion

Oxidation behaviour

The heats of formation reported in the literature are 1 16.24±0.41 kJ/mol for PdO and 365-405 kJ/mol for Rh₂0_3,. Therefore, in the case of an oxidizing environment, the surface should be enriched in the Rh oxide, as this is the more

thermodynamically stable species. However, the proximity histogram in Figure 2 shows that following oxidation at 873K the surface is enriched with Pd forming Pd oxide with the expected PdO stoichiometry. For the longest time oxidation used in this study (180 mins), the Rh-rich oxide phase (Rh-|Pdi)O₂ was observed next to the dominant PdO oxide phase, as shown in the concentration profiles in Figs 3c) and d).

Surface effects

The surface segregation of the different metals was measured carefully by taking the top 3nm slice of the APT data and creating iso-concentration surfaces within this small volume, as seen in Figure 5. No Rh-rich or Pd-rich regions were found on the surface Pd-Rh oxidized for 10 mins (Fig. 5a)). The measured composition of this surface oxide layer is 3.9±0.2at.% Rh, 92.7±1 .0at.% Pd, and 3.3±0.2at.% O. It can be seen that the oxygen concentration is low and is probably an oxygen chemisorbed layer. The 53. Oat. % Pd and 7.5at.% Rh iso-concentration surfaces were plotted on the dataset from oxidation for 30 mins in Fig 5b), while 67. Oat. % Pd and 7.5at.% Rh iso-concentration surfaces were plotted on the dataset for 180 mins oxidation in Fig. 5c). Finally, the 25.0at% Rh iso-concentration surface is plotted in Fig. 5d) for the top 3 nm surface of Pd-Rh after the oxidation and reduction cycle outlined above. For Pd-Rh oxidized over 30 mins, a Rh-rich oxide island is observed on the PdO surface. For the Pd-Rh specimen oxidized for longer times (180 mins), the Rh-rich island segregation is more obvious. A large region of Rh- oxide forms with a length of around 20nm, while Pd-rich oxide also grows on the surface, as seen in Figure 5c). A Rh-rich metallic region with a concentration higher than 25. Oat. % was observed in Figure 5d), which suggests that separate Rh- rich oxides form following oxidation for only 20 mins, and that these cluster-like Rh- rich metallic regions are also present on the surface of the alloy following an oxidation/reduction cycle, which is of particular importance for catalytic applications of this material.

Bulk effect

The oxidation of Pd-Rh at 873K reveals that PdO forms readily after 10 mins along the specimen shank. This oxide is slightly depleted in Rh, with a Rh/Rh+Pt+O value of 2.7±0.1 at.%. All the measured concentrations are summarised in Table 1 below:

A transition region is observed between the PdO oxide and the metal matrix. This region is depleted in Rh and with low oxygen concentration, insufficient for the formation of a stable oxide. Longer oxidation times (30 and 180 mins) yield thicker PdO oxides. The development of Rh rich oxide islands is within the PdO oxide phase. The stoichiometry of these Rh-rich oxide islands is (RhPd)0₂. Following oxidation at 180 mins, several Pd₂0 islands were also found to accompany the (Rh- rich) RhPd0₂ islands. The ratio of metal and oxygen in the Rh-rich oxide islands is 1 :1 , however, the Pd₂0 islands are deficient in oxygen. The mechanisms of the formation of the Pd₂0 and Rh-rich oxide islands (RhPd)0₂ is not yet fully clear. The formation of the Rh-rich oxide phase, RhPd0_2, at 873K for longer oxidation times has not been observed previously. It was suspected that the initial prevailing formation of Pd oxide prevents the diffusion of Rh to the surface. However, as the oxidation time increases this layer becomes saturated with oxygen, and Rh atoms start diffusing into the PdO from the metal and ultimately to the surface. It is this diffusion of Rh into the Pd-oxide which is believed to induce the formation of the Rh-rich oxide, i.e

Oxidation/reduction treatment

After the oxidation/reduction treatment, the Rh concentration in the enriched clusters is 49.0at.%, as shown in the region labelled V in Fig 4b). The Rh concentration dramatically increases in these compared to the nominal

concentration of 6.4at.%, due to the formation of Rh-rich oxide phases at this oxidation temperature. The Rh-rich regions are formed during the oxidation treatments and the resulting Pd:Rh ratios in the separate regions of the

oxidised/reduced samples (Fig. 4) are similar to those for the oxidised alloy (Figs. 2 and 3), indicating very little metal atom diffusion during the reduction stage.

Example 2 - Comparison with oxidation/reduction treatment of Pt-Rh alloy

As Pd-Rh and Pt-Rh are often used in similar catalytic devices, a comparison of their oxidation is of interest. This is shown in Fig. 6 for Pd-Rh oxidized at 873K for 10 mins and Pt-Rh oxidized at 973K for 1.25 hrs. The different exposure conditions were necessary in order to produce reliably measurable changes following oxidation in each material. The large differences however effectively serve on a basic level to demonstrate significant behavioural differences between the alloys. The surface of Pd-Rh is enriched with Pd oxide after 10 mins oxidation at 873K, Fig 6b), while the surface of Pt-Rh alloy is conversely enriched with Rh oxide. In Pd-Rh, although Rh oxide is more stable than Pd oxide, the oxidation of the major element Pd (93at.%) prevails. In the Pt-Rh alloy on the other hand (Fig. 6a), the oxide of the major element Pt is volatile, so as Rh oxide is the relatively stable species at this oxidation temperature the formation of Rh oxide dominates.

Turning to the changes in the surface and bulk after oxidation, at the same temperature (873K), APT reconstruction data of Pt-Rh oxidized at 873K for 5 hrs shows only a thin oxide layer formed on Pt-Rh, while for Pd-Rh alloy oxidized at 873K but for a shorter time (3hrs), Fig. 1 c) shows that the Pd-Rh specimen was heavily oxidized into the bulk. The proximity histogram of Pt-Rh reveals that the surface oxide has a stoichiometry of Rh₂O₃ with a Rh concentration 35.6±0.7at.% and a Pt content of 6.7±0.3at.%. Comparatively, the structure of Pd-Rh in Fig. 1 c) reveals separate Rh-rich and Pd-rich regions, as also shown on the alloy surface in Fig. 5c).

Along with the different oxidation behaviour of the two alloys, oxidation/reduction treatments also produce different final structures for both. For Pt-Rh, the formation of a Rh-rich metallic shell layer is observed after an oxidation/reduction treatment. The Rh concentration on the surface reaches nearly 100at.%. In the case of Pd-Rh, a complex Rh-rich metallic island structure was observed to form on the surface, Fig. 5d), after an oxidation/reduction treatment with an average Rh concentration within the islands of 54.1 at.%. Therefore, oxidation/reduction cycle treatment yields a core-shell structure on Pt-Rh alloy and an island structure on the Pd-Rh alloy. However, both surfaces are dramatically enriched in Rh metal which diffuses from the bulk during the oxidation treatment, and this is retained after reduction at relatively lower temperatures.

Example 3 - Preparation of surface-modified Pt-Rh-Ru and Pt-Ru alloys Materials and Methods

A Pt-24at.%Rh-9at.%Ru alloy in the form of as-drawn wire (0.035mm diameter) was purchased from Johnson Matthey Materials Limited. APT specimens were then prepared by electropolishing using a molten salt mixture of 4:1 NaNO₃:NaCI, and the residual surface impurities and surface oxide were removed by field evaporation prior to any oxidizing environment. APT specimens were then oxidized at 873K for 10 minutes, 30 minutes, 90 minutes and 300 minutes in 1 bar of oxygen at a flow rate of 35 cc/minute. In another set of experiments several APT specimens were oxidized at 873K for 1 hour in oxygen, and then reduced at 673K for 2 hours in 1 bar of hydrogen.

A Pt-9at.%Ru alloy in the form of as drawn wire (0.1 mm in diameter) was obtained from Sigmund Cohn Wire. The APT Pt-Ru specimens were prepared and cleaned by the same methods as discussed above. Afterward, the Pt-Ru specimens were oxidized at 873K for 5 hours in 1 bar oxygen.

Specimens were characterized by APT using a Cameca LEAP 3000X-HR™ instrument. Several specimens were analysed for each unique thermal/gas exposure. The analyses were carried out in laser pulsing mode with a specimen temperature of 55K, target evaporation rate of 2 ions per 1000 pulses, pulse rate of 160kHz, pulse energy of 0.5nJ and a laser spot size of approximately Ι Ομηη. The standing voltage was automatically increased in order to maintain a constant evaporation rate. Compositional analyses of the as-received materials by APT yielded a Rh and Ru content of 23.6±0.2at.% and 9.3±0.6at.% respectively compared with the nominal composition of 23.9at.% and 9.7at.% for Rh and Ru. The measured Ru composition in Pt-8.9at.%Ru alloy was 8.6±0.4at.%. The different reconstruction parameters (image compression factor, evaporation field, radius and field factor) can have a significant influence on the scale and aspect ratio of the 3D reconstructions and therefore any quantitative measurement of distances. In this study, the values of the image compression factor and the field factor were kept as the default values (1.65 and 3.3) of the reconstruction procedure.

Results

Oxidation of Pt-Rh-Ru alloy

The evolution of oxides of APT Pt-Rh-Ru specimens oxidized at 873K for various time were examined by APT as seen in Figure 7. A shorter time oxidation, 10 minutes, was applied on the Pt-Rh-Ru specimen in order observe the initial stage of the oxidation reaction. The dashed line in Figure 7a) represents the grain boundary across the specimen. The APT reconstruction shows that oxidation occurs at the grain boundary as well as the top surface. The oxide at the grain boundary grows exactly along the contour of the grain boundary. A longer time oxidation for 30 minutes was conducted on Pt-Rh-Ru specimens as seen in Figure 7b). The dashed line in Figure 7b) represents a grain boundary again. Similarly to the 10- minute oxidation state, the oxide particles continue to form along the grain boundaries, although they do not form a smooth continuous oxide film but rather exhibit a complex topology. To compare with the shorter time oxidation, a relatively longer treatment time, 90 minutes, was applied on the Pt-Rh-Ru specimen. The APT reconstruction in Figure 7c) illustrates that the oxide continued to grow in a very complicated shape and the RhO_x-containing species (pink-colour)

concentrates at several regions while RuCycontaining oxide species (green-colour) appears nearby the RhCycontaining species. As the oxidation time increases to 300 minutes, the Pt-Rh-Ru specimen was heavily oxidized as seen in Figure 1 d). This shows very clearly that only a small volume of metallic region remains. The oxide layer consists mainly of RhO_x and RuO_x containing species, with the two different oxide species distributed separately.

Composition of oxide phases

The detailed composition analysis in oxide was conducted by proximity histogram or 1 D concentration profiles and summarized in Table 2 below:

For the Pt-Rh-Ru APT specimen oxidized for 10 mins in Figure 7a), the

concentrations in the oxide at the surface in Figure 7a) were determined using an analysis cylinder perpendicular to the surface. The atomic Rh, Ru and Pt concentrations in the oxide at the free surface are 25.7 + 1 .9at.%, 4.3 + 0.8at.% and 6.5+ 1 .0at.% respectively. These results demonstrate that the surface oxide is a mixture of Rh, Ru and a trace of Pt oxide species and it is mainly Rh-enriched oxide species. Moreover, the ratios of total number of metal atoms/atomic oxygen atoms on the surface and grain boundary are 0.54 + 0.01 :1 and 0.57 + 0.06:1 . To analyse the oxide formed on the grain boundary in Figure 7a), an iso-concentration surface at 1 1.2at.%0 was used to isolate the oxidised and alloyed regions. The

composition of the oxide formed in the grain boundary is 27.8 ± 0.2at.% Rh, 6.6 ± 0.1 at.% Ru and 0.74 ± 0.04at.% Pt calculated from the black arrowed region in Figure 8a). This indicates that the oxide formed at the grain boundary is enriched with Rh and slightly deficient in Ru with a negligible amount of Pt. This might be due to the stable phase of Ru oxide (RuO₂) starting to form the unstable oxide phase (RuO₄) which is volatile at relatively low temperature (973K). Although the dissociation temperature is higher than the oxidation treatment temperature which was used in the experiment, it is suspected that the Ru may start to vanish at around 873K. The composition of the grain boundary beneath the oxide in Figure 7a) was obtained using the analysis cylinder and the 1 D concentration profile is shown in Figure 8b). The Rh and Ru concentrations on the grain boundary away from the oxide are approximately 17.1 + 1 .Oat.% and 2.9 + 0.4at.% respectively calculated in the black arrowed region in Figure 8b). These values are lower than the nominal values of Pt-24at.%Rh-10at.%Ru for the bulk alloy. Thus the depletion of Rh and Ru on the grain boundary away from the oxide might caused by the formation of oxide.

An atomic oxygen iso-concentration surface (40.0at.%) was plotted including the whole oxide layer grew for 30 mins as seen in Figure 7b). The proximity histogram is not shown here, but the oxide consists mainly of Rh oxide species at a concentration of 26.2 + 0.1 at.% and a small amount of Ru oxide species with a concentration of 6.80 + 0.02at.%. Again the oxide is enriched with Rh and slightly Ru depleted compared to the bulk. The composition of Pt in the oxide is negligible. The ratio of the total number of metal atoms over total oxygen atoms remains 0.51 + 0.01 , which confirms that for this relatively short time treatment the oxide stoichiometry is roughly M0₂. Hence the (Rh, Ru)0₂ oxide phase may be formed. An analysis cylinder passing through the two different oxide species regions in Figure 7d), from which the 1 D concentration profile was plotted in Figure 8c). In the Ru-enriched oxide species region, it was found that the Rh and Ru concentrations are 20.9 + 0.1 at.% and 12.3 + 0.1 at.% over the left black arrowed region in Figure 8c). The ratio of the total number of metal atoms to the total oxygen is

0.50 + 0.01 :1. The ratio of metal/oxygen implies that the Ru-enriched oxide might be formed as Ru0₂, which has the rutile structure. Regarding the pink regions which represent Rh oxide species, it was calculated that the Rh and Ru

concentrations were 38.3 + 0.4at.% of Rh and 1.4 + 0.1 at.% of Ru respectively, calculated from the right black arrowed region in Figure 8c). The ratio of metal/oxygen is 0.66 + 0.01 :1 , suggesting that the Rh-enriched oxide is a phase with a similar structure to corundum Rh₂O₃. In both regions, the amount of Pt is negligible. Oxidation/reduction treatment

When the oxides of Pt-Rh-Ru alloy are reduced to the metallic state, the island structure remains on the surface as shown in Figure 9a). This APT specimen was oxidized at 873K for 1 hour in 0₂ and reduced at 673K for 2 hours in H_2. The cross- sectional APT reconstruction shows apparently that there is still a difference in the distribution of metallic Rh and Ru. There is a very small amount of oxygen left in the specimen. An analysis cylinder with a diameter of 2nm as the red arrow pointed is oriented across the matrix, Rh enriched, Ru enriched and Rh enriched regions, in Figure 9a) so as to measure the concentration of each metallic species. The 1 , 2 and 3 regions marked in Figure 9a) and b) reveals the region and the

concentrations in corresponding regions respectively. In the Rh enriched region the concentration of Rh and Ru is 90.7 + 10.7at.% and 8.7 + 2.5at.% respectively while the concentration of Ru in the Ru-enriched region is 42.7 + 6.4at.% with a Rh concentration of 53.3 + 7.4at.%.

Discussion Oxidation behaviour

Based on the heats of formation (167 kJ/mol for Pt0₂, 142kJ/mol for PtO, 365- 405kJ/mol for Rh₂0₃, and 300-310 kJ/mol for Ru0₂) the oxidation of Pt-Rh-Ru alloy should induce the Rh segregation. Regarding the dissociation temperatures of each metal, it is concluded that Pt oxide is not stable above 923K, while Rh oxide is stable below 1413K. The Ru0₂ itself is relatively stable, however, it was reported that the volatile oxide, Ru0₄ forms at around 923K. Therefore the formation of Rh oxide (Rh₂0₃) and the stable Ru oxides (Ru0₂) is most favourable compared with Pt oxide formation in the presence of oxygen below the dissociation temperatures of the stable oxides. This prediction for the oxidation of Pt-Rh-Ru holds below the lowest dissociation temperature of Ru and Rh oxides around 993K. The segregation behaviour of the oxides in bulk was shown in Figure 7. For short times oxidation, the grain boundaries are more easily oxidized and the oxide was observed to grow along the contour of the grain boundary. It is known that diffusion along the grain boundary is more rapid than diffusion in the lattice (short-circuit diffusion). This is reflected in the high diffusion coefficient (normally 10⁴-10⁶ times the lattice diffusion coefficient) and low activation energy (0.5-0.7 times for lattice diffusion). The region along the grain boundary beneath the oxide formed after 10- minute oxidation is deficient in Rh and Ru and enriched with Pt, indeed suggesting diffusion of these species along the grain boundary as the oxides grow. Not all specimens contained grain boundaries and the few specimens without these exhibited only surface oxidation in a thin layer without any evidence for formation of thick oxide in the bulk.

At short times, the measured compositions are consistent with the stoichiometry of the (Rh,Ru)O₂ oxide, with the remaining Pt possibly coming from positioning uncertainty in thin surface oxide layer. After 30 minutes, the Pt concentration has become negligible and the observed oxides are still (Rh,Ru)O₂ type. As the oxidation time is increased to 90-300 minutes, separate Rh-rich and Ru-rich regions form with compositions consistent with that of the (Rh, Ru)₂O₃ and (Ru, Rh)O₂ phases respectively. These two phases may be associated with the stable oxide phases of Rh (Rh₂O₃) and Ru (RuO₂). These have the corundum and rutile structures respectively. As indicated by the measured compositions in Table 2 the partitioning of Rh and Ru to these two oxide phases is different, with little Ru in Rh₂O₃ but significant amounts of Rh in RuO₂.

Surface effect

Segregation of Rh-rich and Ru-rich regions was observed on the shank and internal structure of the oxidized Pt-Rh-Ru specimens. It is also of interest to investigate the segregation on the surface, as seen in Figure 10. This shows the top 5-1 Onm regions selected from Pt-Rh-Ru specimens oxidized for 10 minutes, 30 minutes, 90 minutes and 300 minutes in Figure 7. The APT rescontructions in Figure 10 are orientated at an angle to show the x-y plane (top surface) with Rh and Ru iso- concentration surfaces. No obvious Rh or Ru-rich regions were found on the Pt- Rh-Ru specimen oxidized for 10 minutes, as observed in Figure 10a). The

32.0at.% atomic Rh iso-concentration surface was plotted within the dataset as seen in Figure 10b). It illustrates that Rh-segregation is already present after 30- minute oxidation, however, no obvious Ru segregation occurs during this period. As the oxidation time increases, Ru-rich regions start to appear in the sub-surface regions with a concentration of 1 1 .0at.% atomic Ru, higher than the nominal Ru concentration in the alloy shown in Figure 10c). Again Rh-rich regions were also observed after 90-minute oxidation treatments. By increasing the oxidation time to 300 minutes, Ru-rich regions separate from the Rh-rich islands, which are observed very obviously in Figure 10d). The Ru-rich region has a concentration higher than 23. Oat. % and forms an oxide island of approx. 10nm in length. The Rh-rich oxide island is close to the Ru-rich region, again the Rh-rich region precipitates with a concentration of 33.0at.%. Therefore it can be concluded that the Rh-rich oxide firstly precipitates from the bulk oxide after 30-minute oxidation, while the Ru-rich oxide precipitates after 90-minute oxidation. The extent of separation into Rh-rich and Ru-rich oxide regions either on the surface or in the bulk increases with longer oxidation times.

Oxidation/reduction treatment

This separation of isolated oxide islands was observed on specimens with a longer oxidation treatment (longer than 30 minutes) shown in Figure 10c) and d).

Consequently a nano-island structure surface with two separated catalytically active species could be induced by oxidation and reduction. The cross-section APT reconstruction in Figure 9a) also includes different iso-concentration surfaces with atomic Rh (75.0at.%)(red-colour) and atomic Ru (25.0at.%)(black-colour) were plotted to show the Rh-rich and Ru-rich metallic regions. As discussed in the previous section, bulk segregation was observed clearly for Pt-Rh-Ru oxidized for 90 minutes, but surface segregation of Rh and Ru was still minor. Therefore it can be expected that surface segregation would not be obvious or cannot be induced after oxidation for only 60 minutes.

From the concentration profile in Figure 9b) which measured the concentration change along the red arrowed region, it was found that, in the Rh enriched region the concentrations of Rh and Ru were 90.7 + 10.7at.% and 8.7 + 2.5at.%

respectively. While the content of Ru in the Ru-enriched region is 42.7 + 6.4at.% with a Rh concentration of 53.3 + 7.4at.% measured in Figure 9b). In both regions, the amount of Pt is negligible. According to the phase diagram of Rh-Ru, the maximum solubility of Rh in Ru is 60.0at.%, while highest Ru concentration in Rh is 34.5at.%. Therefore it indicates that, in the present alloy the Rh concentration nearly reaches the upper limit in the Ru-rich region according to the phase diagram.

Claims

Claims:

1. A method for altering the surface composition of an alloy, said method comprising the following steps: a) providing an alloy comprising at least two metals;

and further wherein said alloy is not a platinum-rhodium (Pt-Rh) alloy.

2. A method as claimed in claim 1 , wherein said alloy is a bimetallic or trimetallic alloy.

3. A method as claimed in claim 1 or claim 2, wherein said alloy comprises at least one catalytic metal.

4. A method as claimed in any preceding claim, wherein said alloy comprises Pt and/or Pd.

5. A method as claimed in claim 1 , wherein said alloy is selected from Pd-Rh, Pt-Ru, Pd-Ru, Pd-Pt, Pd-lr, Pt-lr, Pt-Co, Pd-Co, Rh-Ru, Pt-Pd-Rh, Pt-Pd-Ru, Pt-Rh-Ru and Pd-Rh-Ru.

6. A method as claimed in any preceding claim, wherein said alloy is provided in nanoparticulate form.

7. A method as claimed in any preceding claim, wherein in step (b) two metal oxides are formed.

8. A method as claimed in any preceding claim, wherein said oxidising agent is gaseous, preferably O₂ or NO.

9. A method as claimed in any preceding claim, wherein said reducing agent is gaseous, preferably H₂.

10. A method as claimed in any preceding claim, wherein said oxidising agent is provided at a pressure of from 5 mbar to 5 bar.

1 1 . A method as claimed in any preceding claim, wherein said reducing agent is provided at a pressure of from 5 mbar to 5 bar.

12. A method as claimed in any preceding claim, wherein at least one of said oxidising and said reducing agent is gaseous and provided at a flow rate of from 1 to 100 cm³/minute.

13. A method as claimed in any one of claims 1 to 1 1 , wherein at least one of said oxidising and said reducing agent is gaseous and provided at a flow rate of from 10 to 50 cm³/minute.

14. A method as claimed in any preceding claim, wherein step (c) is performed at a lower temperature than step (b).

15. A method as claimed in any preceding claim, wherein step (b) is performed at a temperature of up to 1500 K.

16. A method as claimed in any one of claims 1 to 14, wherein step (b) is performed at a temperature of from 600 K to 1200 K.

17. A method as claimed in any preceding claim, wherein step (c) is performed at a temperature of up to 1000 K.

18. A method as claimed in any one of claims 1 to 16, wherein step (c) is performed at a temperature of from 450 K to 750 K.

19. A method as claimed in any preceding claim, wherein step (b) is carried out for a period of from 5 minutes to 8 hours.

20. A method as claimed in any one of claims 1 to 18, wherein step (b) is carried out for a period of from 5 minutes to 5 hours.

21 . A method as claimed in any preceding claim, wherein step (c) is carried out for a period of from 30 minutes to 4 hours.

22. A method as claimed in any one of claims 1 to 20, wherein step (c) is carried out for a period of from 45 minutes to 2.5 hours.

23. A surface-modified alloy obtainable by a method as claimed in any preceding claim.

24. A surface-modified alloy as claimed in claim 23 having at least one surface comprising island structures of at least one metal.

25. A surface-modified alloy as claimed in claim 24, wherein said island structures comprise at least one metal selected from Pt, Pd, Rh and Ru.

26. A surface-modified alloy as claimed in any one of claims 23 to 25 which is a nanoparticle.

27. A surface-modified alloy comprising two or more metals,

wherein said alloy has at least one surface comprising island structures of at least one metal,

and further wherein said alloy is not a platinum-rhodium (Pt-Rh) alloy.

28. Use of a surface-modified alloy as claimed in any one of claims 23 to 27 as a catalyst.

29. The use as claimed in claim 28, wherein said catalyst is a heterogeneous catalyst.

30. The use as claimed in claim 28 or claim 29, wherein said catalyst is for automobile exhaust emission control, ammonia oxidation, methanol oxidation, methanol reformation, or Fischer-Tropsch catalysis.

31 . A catalytic converter or fuel cell comprising a surface-modifed alloy according to any one of claims 23 to 27, e.g. in nanoparticulate form.

32. A fuel cell as claimed in claim 31 , wherein said fuel cell is a low-temperature fuel cell.

33. A fuel cell as claimed in claim 31 or 32, wherein said surface-modified alloy forms an electrode of said fuel cell.