Oxidative, reductive, infrared and catalytic studies of supported rhodium catalysts

supported rhodium catalysts

Citation for published version (APA):

Vis, J. C. (1984). Oxidative, reductive, infrared and catalytic studies of supported rhodium catalysts. Technische

Hogeschool Eindhoven. https://doi.org/10.6100/IR114107

DOI:

10.6100/IR114107

Document status and date:

Published: 01/01/1984

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be

important differences between the submitted version and the official published version of record. People

interested in the research are advised to contact the author for the final version of the publication, or visit the

DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page

numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

attraction, contact-force, catalysis etc., they explain them."

SUPPORTED RHODIUM CATALYSTS

PROEFSCHRIFT

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE

TECHNISCHE

WETENSCHAPPEN AAN

DE

TECHNISCHE

HOGESCHOOL EINDHOVEN,

OP GEZAG VAN DE RECTOR

~~GNIFICUS,

PROF.DR. S.T.M. ACKERMANS,

VOOR

EEN

COf'.1MISSIE

AANGEWEZEN

DOOR

HET COLLEGE

VAN DEKANEN IN HET OPENBAAR TE

VERDEDIGEN OP

VRIJDAG 24 FEBRUARI 1984, TE 14.00 UUR

DOOR

JAN CORNELIS VIS

GEBOREN TE OUD VOSSEMEER

1.2. The importance of metal catalysts 2 1.3. Oxidation and reduction properties of supported

metals

1.4. CO adsorption and I.~.-spectroscopy

1.5. Synthesis gas reactions and their mechanism 1.6. References

2. Reduction and oxidation of Rh/Y-Al

₂

g

₃

~~~~·~-~~

catalysts as studied by Temperature Programmed Reduction and Oxidation

2.1. Introduetion 2.2. Experimental 2.3. Results

2.3.1. Hydragen chemisorption 2.3.2. TPR and TPO of Rh/Y-Al 20 3 2.3.3. TPR and TPO of Rh/Ti0₂ 2.4. Discussion

2.5. Conclusions 2.6. References

3. The and

Y-Al 2

2

3 as studied with Temperature Programmed Reduction-Oxidation and Transmission Electron l"licroscopy 3.1. Introduetion 3.2. Experimental 3.3. Results 3. 3 .1. Hydragen 3.3.2. TPR/TPO 3.3.3. TPR/TPO chemisorption of the RA-series of the RT-series 3.3.4. TEM measurements 3.4. Discussion 3.5. Conclusions 3.6. References 4 4 9 12 15 17 20 23 25 26 31 32 36 39 43 45 47 53 57 60 61

4. 2. 4.3. Results 4. 3 .1. Rh/La₂o 3 4.3.2. Rhjy-Al 2

o

3 4.3.3. Rh/Ti0₂ 4.4. Discussion 4.5. References

5. an in-situ infrared cell and measurements 5.1. Introduetion 68 73 79 82 83 85 5.1.1. General introduetion 88

5.1.2. of a new in-situ I.R.-cell 90 5.1.3. CO on Rh, as studied with I.R.-spectroscopy 91

5.2. Experimental 92

5.3. Results and discussion 93

5.4. Conclusions 101 5.5. References 103 6. 6 .1. Introduetion 106 6. 2. Experimental 109 6. 3. Results

6. 3.1. The structure of rhodium af ter reduction 114 6.3.2. The structure of rhodium af ter

co

admission 121 6.4. Discussion

6.4.1. The structure of rhodium af ter reduction 127 6.4.2. The structure of rhodium af ter

co

admission 130

6.5. Conclusions 133

Acknowledgements Samenvatting Dankwoord Curriculum Vitae List of publications 147 148 151 153 154

Chapter 1. lntrod u ct ion

1.1. Future needs of chemical industry.

The role of chemistry in nowaday society is more than most people think or wish. The chemical industry might be responsible for a major part of environrnental pollution, it also answers to numerous material needs of mankïnd. And last but not least, if solutions have to be found for pollution problems, i t will have to be chemistry

To a great extent this depends on crude oil as a raw material and energy carrier. Through various processes the crude oil is separated in various fractjons of hydrocarbons, which are thereafter mainly used as fuels. The very light fractions (containing 1, 2 or 3 carbon atoms and therefore denoted as Cl, C2 and C3 fractions) serve as raw material for many processes, producing alcohols, aldehydes, polymers and so on.

The of chemical industry upon society, and the vulnerability of this industry through its dependenee on crude oil became very clear during the oil crisis of 1973. Since that time the of crude kept

Only a couple of years ago crude oil prices were predicted to reach $ 100 per barrel in 1990 (1973 $

2.60). The chemical industry was forced to meet this challenge by returning to a cheaper and more abundant raw material: Coal. Coal can be to give a mixture of hydrocarbons resembling the heavy oil fractions. And with known refining and processes they can be

converted to useful bulk raw materials and fuels. 'Coal can also be ~asified. It can react with steam under certain conditions and with certain catalysts to yiel.d a mixture of carbon monoxide and hydrogen, which is in a 1 to 1 mixture known as synthesis gas {syngas). Syngas can be seen as the new raw material that would take the place

of crude oil in a foreseeable future, and the chemistry involving the conversion of syngas to bulk chemieals forms, with some other reactions, the socalled

c

1 chemistry.

Current projections predict however that the price of crude oil will reach only $ 50 in 1990 (current price about $ 30), half of the earlier level. This has narrowed

the economie advantages presented by

c

1 chemistry for the and nineties. In potential i t is still promising, but a more cornprehensive analysis of economie and technical factors is required to decide whether

c

₁ chemistry projects will pay off in the near term or only later. This thesis deals with some of the chemical fundamentals of the conversion of syngas, and of some of the catalysts involved in it.

1.2. The irnportance of roetal catalysts.

Numerous processes in the manufacturing of crude oil involve roetal catalysts, because they are able to dissociate hydrogen and to transfer i t to reacting molecules. Metal catalysts contain mostly noble or transition metals such as platinum (Pt), palladium (Pd), cobalt (Co), rhodium (Rh) or nickel (Ni). The conversion of syngas is an excellent example of a reaction proceding with metallic elements. In fact this conversion proceeds with all the metals mentioned above and some ethersas well, iron being the best known example from the Fischer-Tropsch process.

An important item in syngas conversion is whether the oxygen atom from carbon monoxide is retained in the molecule, leading to products such as methanol, glycol, acetic anhydride etc., or removed in the form of water or carbon dioxide, leading to hydrocarbons. Rhodium then takes a rather special place arnong the catalytic active

metals, since i t seems to be able to catalyse processes in either direction. Recently, interest in rhodium has increased further because i t is an irreplaceable compound in the three-way catalyst, applied in the USA for cleaning exhaust gases of automobiles. Plans exist to introduce this three-way catalyst on cars in Europe too, and this will undoubtedly promate the research on rhodium catalysts very much. However, this thesis is mainly directed towards studies of rhodium catalysts in relation with cl chemistry.

As mentioned above, place on the roetal

many catalytic conversions take surface. So i t is of interest to create as large a surface- to volume ratio as possible for rhodium, especially in view of the availibility and price of Rh. Rh as RhCl3.xHi} casts about f 40/g; as a comparison, the price of gold is f 37/g. The ordinary way to create this Rh surface is to bring the roetal onto a high surface area support, such as silica-alumina's, alumina, silica, etc. These kinds of supports are very porous, so we can apply the roetal as a salt in an aqueous solution, which is suck up into the pores by capillary farces. Once we have established the pare-volume of a certain support, we can dissolve the wanted amount of roetal salt in that particular amount of water and add i t to the support. It can be assumed that after careful drying the roetal salt is present within the pores of the support.

Now we have to bring the catalyst into an active state. This is in certain cases the metallic state, which means we have to reduce the catalyst. In some cases i t may be necessary to heat or to oxidize the catalysts first (calcination), in order to remave for instanee ligands from the roetal salt. This means that the oxidation-reduction behaviour of the catalytically active roetal is of vital interest.

1.3. Oxidation and reduction properties of supported metals.

An excellent way of Temperature Programmed TPR/TPO experiments a

studying these processes is Reduction and Oxidation. In our certain amount of catalyst is placed in a horizontal, cylindrical quartz reactor,, held in an electrical furnace. A mixture of Hz/Ar or

o

₂/He can flow through the catalyst bed, while the temperature of the oven can be varied continuously by a temperature controller. The ingoing gasstream (of known composition) is being compared with the outgoing gasstream. If there is a difference in composition (which means consumption of one of the components in the reactor), this leads to an electrical signal from the detector comparing the gasses, and this signal, integrated over time, is proportional to the amount of gas consumed. This signal is also recorded, and this recording is the so-called TPR- or TPü-profile the reader will encounter many times in the Chapters to come. A schematic representation of the TPR/TPO apparatus is given in Chapter 2. TPR and TPO experiments are the main subject of Chapters 2 and 3.

1.4. CO adsorption and I.R.-spectroscopy.

In hydracarbon and oxygenate synthesis from CO and H2 the bonding between CO and the metal catalyst plays an important role. We will give a short description of this process, following the one given by Coulson (1).

In the CO molecule there are ten valenee electrons, occupying the following molecular orbitals in order of increasing energy: 3v, 4v, lpiy, lpiz and Sa. The bonding in the molecule is provided for by the 3a and the two lpi orbitals. Two electrons occupy the non-bonding 4a orbital, made from the oxygen 2s and 2px orbitals and

a

-QjQ

b

therefore concentrated around the oxygen nucleus. The last two electrons occupy the 5o orbital formed from the carbon 2s and 2px orbital. They form the lone pair on the carbon atom, directed away from the oxygen atom. The charge density contours of these orbitals are shown in Fig. 1. 0.1 0.05 0.15 Z(nm) -0.10

--0.05 X(nm)

t

Figure 1. Charge density contours of the 30(a), 40(b), 1~ or 1~ (c) and So(d) orbitals of CO. From ref. (15). y

z

In the bonding with the metal this

so

lone is donated to an empty metal orbital, forming a dative

a-bond. Then there is back donation from a filled metal d-orbital to the Lowest Unoccupied Molecular Orbital of the carbon monoxide molecule, which is the pi*-antibonding orbital. This back donation to an antibonding orbital causes the carbon-oxygen bond to weaken a little. The amount of weakening will be a measure for the electron donating capacities of the adsorbing metal atom. Since CO has a small dipole moment (0.1 , with the negative charge on the C-atom due to the directional properties of the 5 o orbi tal) the CO stretching vibration is infrared active and so we can study the weakening of the CO bond d

upon adsorption by a metal by means of IR-spectroscopy. Of course under favourable circumstances.

-0----0

\ \... \ \ I I '',\~:'./'

--

::.-:s~0~-

/

.

-c::~---

---

--·~~

-Figure 2. The electrical "image" resulting from a positive charge above-, a dipale vibrating parallel with- and a äipole vibrating perpendicular to a metal surface.

Pearce and Sheppard (2) discuss an important phenomenon first mentioned by Francis and Ellison (3). They argue that in the case of IR-studies of species, adsorbed upon a metal surface, i t is necessary that the IR-radiation makes a large angle with the metal surface (near grazing incidence), because at the metal surface itself there is a knot in the standing-wave field (reflection) and absorption cannot be seen there. Only at grazing incidence the resulting field (in- plus out- going) will have sufficient intensity.

Another point of interest is that a perfect metal surface should be perfectly polarizable and this leads to what is called the metal-surface selection rule (2). Their line of reasoning is that the electric field lines

from a charge above a roetal surface roetal surface, will be directed

going towards this as if there were an opposite charge of equal size under the roetal surface at equal distance.

Likewise, a dipole virtual dipole of the roetal surface.

above a roetal surface will induce a equal size but opposite signs, under As can be seen from Fig. 2 the vibrations of a dipole parallel to the surface will be annihilated by its mirror image under the surface, while for a dipole vibrating perpendicular to the surface its mirror image will reinforce it.

For IR-radiation of for instanee 2000 cm-1, the wavelength is 5 ~m. Since this exceeds sufficiently the thickness of an adsorbed layer, and also the spacing between dipole and image-dipole, the physics described above predict that only dipole vibrations perpendicular to a roetal surface are infrared-active.

In recent years the contribution of infrared studies to the understanding of adsorption processes at solid surfaces has increased substantially, a.o. by the introduetion of Fourier Transform Infrared Spectroscopy (usually indicated as FT-IR spectroscopy). This development had been initiated by the invention of the interferometer by Michelsen in the late nineteenth century (4, 5). Interference of light had been recognized long before that time, but with the Michelson interferometer i t became possible to separate the two interfering beams in such a way that their relative path differences could be varied precisely. Michelsen already realized that the visibility curves of his interference patterns contained speetral information, but i t was Lord Raleigh who recognized that the interferogram was related to the spectrum of the radiation passing through the interferometer through the mathematica! operatien known as the Fourier Transformation (4, 6). The interferometer

played an important role in spectroscopy during the following years, but the technology required for the full Fourier Transformation was lacking at the time. The astronomist Fellgett recognized the possibilities of the interferometer in increasing the intensity of the signal on the detector compared with the grating spectro-photometer, where slits have to be used to create the energy dispersion characteristic for the latter instrument. By an interferometer, data from all speetral frequencies are measured simultanuously. The resulting reduction in measurement time is now known as Fellgett's advantage

(4, 7).

It can be expressed more mathematically if one realises that the signal-to-noise ratio increases for longer measurement times

Fellgett in a lso proportion was the with the first to square root of time.

actually perferm a calculate a spectrum.

numerical Four~er Transformation to

Another advantage of the interferometer is that the "throughput" of radiation,

flow through the apparatus thanks to the absence Jaquinot advantage (4, 8).

a measure for the total energy per second, is also greater of slits. This is known as the

The real break-through spectroscopy came with the

for mid-infrared FT-IR application of the Cooley-Tukey fast Fourier Transferm algorithm to interferometry, and the simultaneous development of fast, dedicated minicomputers. Spectra that took hours to measure and still considerable time to calculate, could now be plotted several seconds after starting the measurement of an interferogram. Because of these reasons, since 1968 FT-IR spectroscopy gained an important place in the study of heterogeneaus catalysis. We started our infrared investigations with the development of an in situ infrared cell, and the first results obtained therewith are reported in Chapter 5.

l.S. Synthesis gas reactions and their mechanism.

We tried to point out in this introduetion so far the importance of the synthesis of certain chemieals from carbon monoxide and hydrogen apart from and together with the broad supply of chemieals from natural sources, whether or not after refining. We explained how these syntheses take place over supported metal catalysts, how these catalysts are made and how one can study their oxidation and reduction behaviour. Subsequently we dealt with the interaction of carbon monoxide with these metal catalysts and_ how we can study this interaction with FT-IR spectroscopy. What is left to introduce is a view on the theories of how the syntheses of chemieals over these catalysts are thought to take place.

For this we refer only to the outstanding review of Biloen and Sachtler on the subject (9) and we restriet us here to their conclusions.

The conversion of synthesis gas to hydrocarbons and alcohols can be described by three overall reactions:

n CO + _{2n H2} CnH2n + n H₂0 [1] m CO _{+ ( 2m+l) H2}

--

_CmH2m+2 + m H 20 [2] p

co

+ 2p H₂

-

c

p 1H2P_ 1CH20H + p-1 H20 [3] In the presence of the product water the watergas shift reaction can occur:

The products are mainly unbranched parafins, olefins and alcohols, and the chain length distribution of the molecules usually obeys the so-called Schulz-Flory

distribution, implying that the molecules are formed by step-wise chaingrowth propagation, followed by a singular termination step. This concept involves the presence on the surface of chains Yn and insertable monomers X,

leading to the following reaction scheme:

x

-

À. etc.

[5]

The subscript of the product P denotes the number of carbon atoms in the product molecule. The Schulz-Flory distribution means that the ratio Pn+l over Pn (denoted as a) is constant. This implies that the insertable monomer X is a Cl species. Biloen and Sachtler suggest it might be a -CH2- carbenic species, a conclusion arrived at after analysis of the literature concerned. However, some writers still prefer a cautious CHx notation. The initiating species YQ (and Yl) are also Cl species. The activation energy for the formation of methane is usually found to be the same as for the formation of higher hydrocarbons, so i t is assumed that the rate-determining step in both processes is the same. Since hydragenation is known to be much faster than the Fischer-Tropsch reaction, the termination step cannot be rate-determining. And since the formation of methane does not involve the propagation step, this one can be ruled out too. Biloen and Sachtler show that in the overall reactions

co

- CH _x -- CH x

-[6]

[7]

that the conversion of -CHx- to -CH2- is rate determining. Biloen and Sachtler propose the Fischer-Tropsch propagation is a cis migration on the metal atom:

[8]

The terminatien steps, hydrogenation and S-H-elimination, lead to parafins and olefins, respectively.

Rhodium catalysts are known to produce alcohols as well as hydrocarbons in the CO hydrogenation. Since insertion of CO in an alkyl group is thermodynamically favourable in the case of rhodium (10), one is tended to formulate a reaction mechanism for the formation of alcohols in which an insertion step of CO plays a key role. However, no definitive mechanism explaining the formation of oxygenated products has been given in literature yet. It should be remarked here that some related reactions are known which make some speculations plausible. For instanc;the Monsanto process, the conversion of methanol to acetic acid (11). This process is catalysed by a mononuclear rhodium complex. The oxidation state of the rhodium in the complex changes from I to III and back through oxidative addition of reactants and reductive eliminatien of products. Polyols can be formed also in a homogeneous process, and not only by rhodium catalysts

(12,13). Watson and Somorjai roughly distinguished three temperature areas in which hydrogenation of CO would lead mainly to the products indicated (14):

T <500 K

co

-

co

-

[int]

-

CH 30H

I

I

[9]

s

s

500-600 K _{CO+ H2 -} CH

-

CH

:...co

-

CH 3CHO [10]

I

x

I

x

s

s

C2H50H

T >600 K CH _...

I

x

s

[11]

It is speculated that steps involving insertion of oxygen containing building blocks take place on oxidic sites, those not being able to dissociate CO because the electrens to donate to the pi*- antibonding orbital of the carbon monoxide. It is generally accepted that of hydrocarbons takes place over metallic formation rhodium ensembles. performance of some Catalytic catalysts studies concerning in the conversion the of synthesis gas are presented in Chapter 4. Some overall kinetic parameters of the processes described above have been determined.

In Chapter 6 we come back to the special example of the interaction between carbon monoxide and a rhodium-alumina catalyst. There has long been a controversy in literature as to what processes took place during this interaction, and we used a number of complementary techniques to elucidate this problem. Since this Chapter will be published as such, we refer to its own introduetion for a review of the relevant literature.

Chapter 7 presents a general discussion of the results presented in this thesis and some suggestions for further research.

1.6. References.

1. coulson, C.A. in "Valence", University Press, London 1961.

2nd Ed., Oxford

2. Pearce, H.A. and Sheppard, N., Surface Sci. 59, 205 {1976).

3. Francis, S.A. and Ellison, A.H., J.Opt.Soc.Am., 49, 131 (1959).

4. Griffiths, P.R., in "Chemical Infrared Fourier Transform Spectroscopy", John Wiley and Sons, New York 1975.

5. Miche1son, A.A., Phil.Mag. Ser. 5, 31, 256 (1891).

6. Lord Ra1eigh, Phil.Mag. Ser. 5, 34, 407 (1892).

7. Fe11gett, P., in "Proceedings of Aspen International Conference

A.T. Stair 1970.

on Fourier Spectroscopy", and D.J. Baker, Eds.),

(G.A. Vanasse, AFCRL-71-0019,

8. Jaquinot, P., 17e Congres du GAMS, Paris, 1954.

9. Bi1oen, P. and Sacht1er, W.M.H., in "Advances in Cata1ysis", Vol. 30, p 165 (D.D. Eley, H. Pines and P.B. Weisz, Eds.), Academie Press Inc., New York 1981.

10. Kuh1man, E.J. and Alexander, J.J., Chemistry Reviews, 33, 195 (1980).

Coordination

11. Pau1ik, F.E. and Roth, J.F., J.C.S.Chem.Oomm., 1578 (1968).

12. De1uzarche, A., Fonseca, P., Jenner, G. and Kienneman, A., Erdoeh1 und Koh1e, 32, 313 (1979).

13. Keim,

w.,

Berger, M. and Sch1upp, J.J., J.Cata1., 61, 359 {1980).

14. Watson, P.R. and Sornorjai, G.A., J.Catal., 74, 282 (1982).

Chapter 2

The reduction-oxidation behaviour of supported Rh/y-Al203 and Rh/Ti02 catalysts as studied with Temperature Programmed Reduction and Oxidation.

Summary.

Careful preparatien of Rh/Al20 3 catalysts leads to ultradisperse systems (H/Rh>l.O). TPR shows that these catalysts are almast completely oxidized during passivation. Identical preparatien of Rh/Ti02 catalysts leads to less disperse systems (H/Rh=0.3), exhibiting two reduction peaks in TPR. These peaks are due to the reduction of smal!, wel! dispersed Rh20 3 particles, and of large, _{bulklike Rh2o 3 particles. In all cases} reduction of Rh 2o 3 is complete above 450 K. Tio 2 is partly reduced by a roetal catalysed proces above 500 K.

2.1. Introduction.

Over the past years i t has become clear catalysts take a special position in

that rhodium the field of supported transition roetal catalysts, because they are able to produce hydrocarbons as wel! as oxygenated products (alcohols, aldehydes, acids) from synthesis gas (1-10). Various workers have tried to influence the selectivity and activity of the rhodium catalysts via a special p .. :eparation (1,2), via additives (3-5) or mixed oxides (8-10), and where this worked out in the wanted direction of higher production of oxygenates at least two of them gave as the reasen the presence of stabilized Rh+

in the surface (8,10). Same authors claimed the presence of isolated Rh+ sites in monometallic Rh catalysts on the basis of IR evidence (Worley et al. (11-13), Primet

(14)), while others metallic rafts with

(15)).

found the Rh to be present as electron microscopy (Yates et al.

In all cases i t seems obvious that the support plays an important role in either bringing or keeping the roetal in a certain state of (un)reactivity. A special example of such a roetal support interaction has been discovered by Tauster et al. (16,17) and is now known as Strong Metal Support Interaction: Supported metals such as Pt and Rh loose their capability for chemisorption of H2 ,

co-

and NO if they have been reduced at high temperatures (e.g. 773 K) on supports such as _{Ti0 2} and _{V20 5.} Normal chemisorption behaviour can be restored by oxidation at elevated temperatures, followed by low temperature reduction, at for instanee 473 K (16). SMSI has been related to the occurrence of lower oxides of the support (18, 19), but the exact nature of the interaction still remains unc1ear.

Many of the above mentioned phenomena have to do with one camman property: The oxidation-reduction behaviour of supported Rh cata1ysts. We decided to study two systems, representative for many of the ones referred to above: 2.3 wt% Rh/Al 2

o

3 and 3.2 wt% Rh/Ti02 . Via sintering (see Experimental) we have induced a variatien in partiele size (dispersion) in order to answer the fo11owing questions:

- how is oxidation-reduction influenced by partiele size - how is oxidation-reduction influenced bu the support

used

-does partiele size show any effect upon SMSI.

Befare we come to the experimental techniques, we have to introduce one last item: Passivation. Since i t is

obvious that reduced systems cannot simply be stared in air, we passivate and stabilize them by applying a layer of oxygen upon the metal particles in a controlled

{see Experimental). Some authors have already way paid attention to the state catalysts are in after starage in air. Thus Burwell jr. et al. used Wide Angle X-ray Scattering, Extended X-ray Absorption Fine Structure (EXAFS), hydragen chemisorption and hydrogen-oxygen titration to characterize their supported Pt and Pd catalysts {20-23).

We will show that a good insight in all these matters can be gained with the aid of Temperature Programmed Reduction and Oxidation (TPR and TPO), supported by chemisorption measurements. TPR as a characterization technique was presented by Jenkins et al. in 1975 {24, 25) and has been used extensively in the past few years, as becomes obvious from a recent review by Hurst et al. (26). The technique allows one to obtain (semi)quan-titative information about the rate and ease of reduction of all kinds of systems, and once the apparatus has been built, the analyses are fast and relatively cheap. We used an aparatus as described by Boer et al. (27), which enabled us to extend the analyses to Ternperature Prograrnmed Oxidation, and to gather information about the rate and ease of oxidation as well.

2.2. Experirnental.

Tio₂ (anatase, Tioxide Ltd., CLDD 1367, surface area 20 m2/g, pore volume 0.5 cm3/g) and y-AlzOJ (Ketjen, 000-1.5E, surface area 200 m2/g, pore volume 0.6 cm3/g) were impregnated with ·an aqueous salution of RhC1 3.xH_{2 0 via}

the incipient wetness technique to prepare a 2.3 wt%

Rh/Al 2o 3 catalyst and a 3.2 wt% _{Rh/Tio 2 catalyst, as was}

catalysts were dried in air at 355, 375 and 395 K for 2 hrs successively, followed by direct pre-reduetion in flowing H₂ at 473, 773 or 973 K for one hour. Prior to removing the catalysts from the reduction reactor they were passivated at room ternperature by replacing the hydrogen flow by nitrogen and subsequently slowly adding oxygen up to 20 %. Then the catalysts were taken out of the reactor and stored for further use.

Programmer

Vent.

Figure 1. Schematic representation of the TPR/TPO apparatus.

The TPR/TPO apparatus used is schematically represented in Fig. 1: A 5% H₂ in Ar or a _{5% 0 2} in He flow (300

ml/hr) can be directed through a microreactor. The temperature of the reactor can be raised or lowered via linear programming. H~ or

o

₂ consumption is being monitored continuously by means of a Thermal Conductivity Detector (TCD). A typical sequence of actions is as fellows:

- the passivated or oxidized sample is flushed under Ar

flow at 223 K

- Ar is replaced by the Ar/H 2 mixture, causing at least an apparent H2 consumption (first switch peak)

- the sample is heated under Ar/H 2 flow with 5 K/min to 873 K

- after 15 min at 873 K the sample is caoled down with 10

K/min to 223 K

- the reduced sample is flushed with Ar

Ar flow is replaced by the Ar/H~ mixture once again, now causing only an apparent H2 consumption (second switch peak).

A similar sequence is followed during TPO. The switch peak procedure deserves sorne closer attention. The strong signa! we cal! the first switch peak is mainly due to the displacement of Ar by _Ar/H2 in the reactor, but in sorne cases real hydragen consumption takes place, even at 223

K. Therefore we repeat the whole procedure after the TPR has been perforrned: In that case the catalyst has been reduced and caoled down to 223 K, and as a consequence is covered with hydrogen. Then we replace the Ar/H_{2 by} pure

Ar,

resulting in a negative TCD signa!. Subsequently we switch back to Ar/H_{2 • Since we do nat expect any hydragen}

consumption frorn the reduced, hydragen covered sample at this time, the resulting switch peak will be due solely to the displacement of Ar by

Ar/H

_2. Sa the difference between the first and second switch peak reveals the real hydragen consumption at 223 K, if there is any.

The reactions that might take place during TPO and TPR are:

4 Rh 0.75) [1]

[2]

The quantities between brackets are the amounts of hydragen or oxygen consumption per rhodium expected for reduction of bulk Rh 2

o

3 or formation of this very materal

(apart from chemisorption of any kind).

Chernisorption measurements were carried out in a conventional volumetrie glass apparatus after reduction

of the passivated catalysts at 473 or 773 K in flowing H for 1 hr followed by evacuation at 473 K for one hr. After H2 admission at 473 K desarptien isotherms were measured at room temperature. As desarptien became only noticeable at pressures below 200 torr we believe that the chemisorption value above that pressure is representative of monolayer coverage (cf. Frennet et al.

(28)). The H/Rh values thus obtained for the various systems are preseuted in Table 1.

H/Rh T(pre-red) RT(LT) (l) (K) RA 473 1. 70 0.37 773 1. 53 0.29 973 1.23 0.12 (1): reduced in situ at 473 K. (2): reduced in situ at 773 K. RT(HT)( 2 ) 0.08 0.05 0.01

Table 1. Hydrogen chemisorption of Rh/y-Al2

o

3 (RA) and Rh/Ti02 (RT) catalysts.

2.3. Results.

2.3.1. Hydragen chemisorption.

The hydragen chemisorption data as given in Table 1 were _{obtained for Rh/Al2}

o

_{3 (RA) after reduction of the} passivated catalyst in situ at 473 K. _{Rh/Ti0 2 (RT) was} also reduced in situ at 773 K, to induce SMSI behaviour. The catalysts will be denoted from now on as RA 773

(Rh/Al

pre-reduced at 973 K), etc., that is de code refers to the temperature of reduction of the dried, impregnated cata1yst (see Experimenta1). The results are represented graphically in Fig. 2, showing the value of H/Rh as a

2.0 1.5 1.0 0.5 473 673 873 1073 Temp. pre-red.(K)

Figure 2. Hydragen chemisorption as a function of pre-reduetion temperature;

a) Rh/Y-Alz03, reduced in situ at 473 K. b) Rh/Ti02 , reduced in situ at 473 K. c) Rh/Ti02, reduced in situ at 773 K.

function of pre-reduetion temperature (i.e. reduction prior to passivation; reduction prior to chemisorption was at either 473 or 773 K, as indicated). The systems discussed here are represented by the separate dots.

Va1ues for Rh/Al

2

o

3 range from 1.70 to 1.00, for non-SMSI Rh/Ti0₂ from 0.37 to 0.12 and for SMSI Rh/TiC2 from 0.10 to

o.oL

a b c d 273 473 673 ~ 873 T(K)

Figure 3. 2. 3 wt% Rh/Y-Alz03 catalyst;

a) TPR of passivated catalyst, pre-reduced at 473 K.

b) TPR of passivated catalyst, pre-reduced at 773 K.

c) TPO following TPR of the catalysts. d) TPR following TPO of the catalysts.

The Temperature Programmed Reduction profiles of the passivated RA 473 and RA 773 catalysts are shown in Fig.

3 a and b. The horizontal axis shows the temperature and the vertical axis the hydragen consumption (in arbitrary units).

RA 473 shows a maximum H2 consumption at 330 K, and some further reduction above 473 K. RA 773 shows a single consumption peak around 273 K, foliowed by desarptien of H₂. The pre-reduetion had apparently been complete, and passivatien of this ultra disperse Rh catalyst had led to dissociative oxygen chemisorption, but nat full oxidation, since the net H2 consumption mounted up to 1.0 H2/Rh. Apparently the pre-reduetion at 473 K (RA 473) had nat been complete, and the oxygen chemisorbed on this system was harder to remave than from the other ones (RA 973 showed an identical TPR profile as RA 773). The subsequent TPO, Fig. 3c, which was identical for all three systems, confirms these observations: 02 consumption starts at 223 K, in the switch peak, continues when the temperature ramp is started, reaches a maximum around 290 K, and then decreases very slowly towards higher temperatures. Integration of the 02 consumption signa! proved to be difficult, due to the small sample sizes (typically 50-75 micromale of metal) and the small thermal conductivity of

o

_{2 but still} mounted up toa rather satisfying 0.6-0.7

o

_{2/Rh (0.75 was}

expected, since chemisorbed H2 had been removed by a heat treatment in between TPR and TPO). The equality of the TPO profiles of the RA 473, RA 773 and RA 973 catalysts is in accordance with the fact that all these catalysts have been brougt up to 873 K during the TPR run and it is also nat surprising to find that the TPR profiles of the

completely oxidized systems (following TPO) are identical too (Fig. 3d). one single peak is observed around 360 K,

corresponding with which agrees with the bulk Rh 2

o

3 showed

a H2 consumption of about 1.5 H2/Rh, reduction of Rh 2

o

₃• Unsupported a reduction peak at 400 K in our apparatus, while bulk Rh roetal only started to become oxidized above 870 K.

R3 473 673

Figure 4. 3.2 wt% Rh/Ti02 catalyst;

~ T{K) a b c 873

a) TPR of ~assivated catalyst, pre-reduced at 473 K.

b) TPO following TPR of the catalysts. c) TPR following TPO of the catalysts.

2.3.3. TPR and TPO of Rh/Ti~.

The TPR profile for the passivated RT 473 catalyst is shown in Fig.' 4a. The most striking feature is that

H2

consumption starts already at 223 K in the switch peak, that is immediately when H 2

/Ar

is being flushed through the reactor. Keeping in mind·the much lower H/Rh value of this catalyst compared to the Rh/Al₂

o

_{3 series, we think} that passivatien here has caused only the formation of an outer layer of oxide on the relatively large roetal particles. If the remaining metallic care could be reached by the hydragen molecules, these molecules could be able to dissociate and provide atomie hydragen for an

easy reduction of the oxide layer at low temperatures. There is also some H₂ consumption just above 473 K, as for the corresponding Rh/Al₂03 catalyst, indicating that also for Rh/Ti0₂ the pre-reduetion at 473 K had not been complete. It can be seen that the Ti02 support is reducible also, leading to H consumption around 573 and 700 K. The H₂ consumption at 273 K amounts to about 0.4

H2/Rh.

The TPR profiles for RT 773 and RT 973 are similar, but since the Rh surface area decreases with increasing pre-reduetion temperature (cf. Table L) the amount of passivatien oxygen decreases also, and so does,

consequently, the H

2 consumption at 223 K in the TPR profiles. Also the consumption just above 473 K has disappeared as a consequence of the higher pre-reduetion temperature.

Fig. 4b shows the TPO profile of a reduced RT catalyst. The three TPO profiles of RT 473, 773 and 973 are alike: Three "areas" of oxygen consumption show up, Which we attribute respectively to chemisorption (which is the only phenomenon on Rh/Alz0

3), corrosive chemisorption and thorough oxidation. We will come back to this assignment in the Discussion part.

The TPR profiles of the oxidized samples of the RT series are identical again (Fig. 4c) and show two clearly divided Hz consumption maxima, at 325 and 385 K. Oonsumption in the first peak is about 1.3 Hz/Rh, in the second one about 0.3 Hz/Rh. Taken together the Hz consumptions come close enough to the expected value of 1.5 Hz/Rh to attribute them both to reduction of Rhz03· The peak at 385 K is assigned to bulklike Rh 2

o

3 particles

(note that the peakmaximum for unsupported Rh zO _{3 is at}

400 K) and the one at 325 K to a better dispersed ~0₃

phase. Furthermore the possibility of some support reduction taking place here as well, in advance of the support reduction around 573 and 700 K, cannot be excluded.

2.4. Discussion.

Hydrogen chemisorption has been used through the years by many workers to characterize metal surfaces (7, 9, 10, 14, 16, 20-23, 29, 30) and when attempts were made to calculate metal surface areas from hydrogen chemisorption data, a hydrogen to metal stoechiometry of one was used. On the other hand, if CO was involved, some authors (31) chose a stoechiometry of one, while others (14) mentioned higher stoechiometries. We are of the opinion that if one accepts a metal atom, such as Rh, to adsorb two or more CO molecules, one should not reject the idea of the same atom adsorbing more than one hydrogen atom. So we think

that although our experimental H/Rh results exceed unity (Table 1), they are real, and that the hydrogen chemisorbed does not exceed a monolayer (cf. Ftennet et al. (28)), and is all bound to the metal. We dit not try to make a distinction between socalled "reversible" and "irreversible" adsorption, because we believe that in the thermodynamic sense there is no such distinction. In following the procedures that lead to that supposed distinction one merely encounters the physical restrictions of ones apparatus, such as pumping speed and conductivity of tubing.

As described in the Experimental section we admitted hydrogen at 473 K. This was simply done to accelerate adsorption but did not effect the ultimate amount of adsorption, as was checked via the measuring of adsorption isotherms at room temperature (32). This leaves us with chemisorption values above unity, and therefore, like some other authors (21, 30), we find i t impossible to calculate a partiele size or a dispersion from these data, since there is no particular stoechiometry value to prefer. We imagine these Rh particles could be raftlike as suggested by Yates et al. (15) (although they were dealing with only 0.5 wt%

Rh/Al 20 3), where the edgeatoms could have the possibility of adsorbi.ng more than one hydrogen atom. "Sintering" of these particles -H/Rh decreases from 1.7 to 1.0 upon reduction at 973 K- would then mean that the number of edgeatoms decreases, for example by growing of the rafts or even formation of (hemi)spherical particles. Still we are dealing with ultradispersed systems. We estimate that in all cases the Rh partiele size does not exceed 1 nm.

Fbr the RT series the H/RH values as established after 473 K reduction are much lower, which was to be expected taking into account the difference in surface area between Al

Here also we see a decrease in the H/Rh values as a result of an increase in the pre-reduetion temperature. But in this case this is simply due to growth of the metal particles, resulting in a decreased metal surface area. The H/Rh values after 773 K reduction show evidence for SMSI behaviour. In this case i t means they are that small that they are of the order of magnitude of the experimental error. Therefore we cannot, at this point, draw any quantitative conclusion about a relationship between SMSI behaviour and partiele size. \'ie knew in advance this would be difficult, since both partiele growth and increasing SMSI behaviour lead to less hydrogen chemisorption, and so the limits of experimental accuracy might prohibit to make a distinction between the two effects. It is evident, though, that RT 773 and RT

973 must have been in the SMSI state after pre-reduction, but they showed normal chemisorption behaviour after passivation

chemisorption following the state.

and re-reduction at 473 K in the apparatus. This means that the passivation, pre-reduction, must have destroyed the SMSI

Our TPR and TPO results also prove that passivation is rather drastic. Fbr the small metal particles on Alzo 3 TPR of the passivated samples differs fram that of the oxidized ones only in the position of the peak, that is the ease of reduction. And from the shape of the TPO pattern we can understand that a long passivation time (like storage in air) comes close to a real oxidation. But i t is very clear that no matter what the initia! state of the Rh was, passivated or oxidized, reduction is complete above 400 K.

Worley and coworkers studied the oxidation state of Rh on various supports, and starting from various Rh precursors, applying IR specroscopy upon CO adsorption (11-13). Apart from IR absorptions attributed to CO

adsorbed on metallic Rh, which they attributed

they to CO

found some absorptions adsorbed on isolated Rh+ sites. Fbr 2.2 wt% Rh catalysts they found these isolated Rh+ _{sites to be more abundant for Al2o 3 than for Ti02 as} a support (13), and more abundant for RhCl3 than for Rh(N0 3) 3 as a precursor (12, 13). They concluded that one gets the worst reduction of Rh (they reduced at 673 K) when using Al 2o 3 as a support, and when starting from RhC1 3 as a precursor. From the TPR evidence presented here, we conclude that the systems they have studied must have been completely reduced prior to admission of 00, and therefore we prefer another explanation. One obtains the best dispersion of metallic Rh _{particles on Al 203,} and when _{starting from RhC1 3 . Upon}

co

adsorption the smaller particles break up and create the isolated dicarbonyl species that were attributed to Rh+l by Worley et al. EXAFS proof for this explanation will be published elsewhere (33).

Our findings for Rh/Ti02 go very well along with the results published by Burwell et al. for Pd and Pt on SiO and Al2o 3 (20-23). Upon passivatien the larger metal particles farm an oxide skin, the formation of which can beseen almast literallyin Fig.4b, the TPO of Rh/TiOz. The small metal particles on Alz03 show only one tailing chemisorption peak, but for Rh/Ti0 2 oxygen chemisorption is followed at higher temperatures by corrosive chemisortion and finally, around 700 K, by thorough oxidation. Apparently oxygen diffusion through the oxide layer is a strongly hindered process. Attribution of the intermediate temperature region of oxidation to corrosive chemisorption is supported by the results of a TPO we did on a 3 wt% _{Rh/Sio 2 (Grace Silica, S.P. 2.-324,382, 290}

m2/g, H/Rh=0.4). This TPO is shown in Fig. Sa, and exhibits two oxygen consumtion areas, one around 273 K which is due to oxygen chemisorption, and one around 500

K due _{to corrosive chemisorption. All of the Rh on Si0 2} is in a well dispersed form, as was confirmed by the subsequent TPR _{of this Rh/Si0 2 catalyst, which showed} only one hydrogen consumption peak at 335 K. This reduction peak corresponds to the first peak observed for Rh/Ti0 2 at 325 K (cf. Results).

The oxidation peak around 700 K in the TPO of Rh/Ti02 is attributed to the formation of rather large, bulklike Rh

20 3 particles, the reduction of which is observed as a separate H

2 consumption maximum at 385 K in TPR (Fig. 4c). That the large particles oxidize at higher temperature than the small particles and reduce at a higher temperature as well, has been proven by a rather

simple experiment, shown in Figs. Sb and c. First a TPO is run with a _{reduced Rh/Ti0 2 system, up to 673 K.} Subsequently a TPR is performed which demonstratee that the reduction peak at 385 K has completely disappeared. Thus the fraction of the Rh which reduces at 385 K oxidizes above 673 K, and vice versa.

/

----, /

~

~

\_

273 473 673

...

T(K} a b c 873

Figure 5. a) TPO of a reduced 3.0 wt% Rh/SiOz catalyst. b) TPO up to 673 Kof a reduced 3.2 wt%

Rh/Ti02 catalyst. c) Subsequent TPR.

That one can distinguish between a well dispersed phase and a _{bulklike phase of Rh2}

o

_{3 on a· support has been} noticed before by Yao et al.

(34),

although they used Al 20 3 as a support and had to oxidize for 12 hrs at 973 K to create the bulk phase.

The fact that part of the Ti02 support is being reduced as well, by a metal-assisted process, is in agreement with findings for Pt/Ti02 (35, 36). Reoxidation of the

support takes place during TPO, exceeds the expected

o

_{2/Rh value of} but is apparently hidden in the

since 02 consumption 0.75 in all cases, TPO profile by the stronger 0 2 consumption caused by the Rh oxidation.

2.5. Cbnclusions.

The 2. 3 wt% Rh/A~

o

₃ catalysts proved to be "ultradisperse" with H/Rh values ranging fran 1. 7 to l.O. The catalysts behaved accordingly in TPR and TPO: Easy reduction and fast oxidation were observed to such an extent that even a mild passivation led to almast complete oxidation.

The 3.2 wt% Rh/Ti0

2 catalysts were much less dispersed (H/Rh 0.37-0.12) and showed evidence of two distinct forms of Rh (and Rh2

o

3 ), appearing as two reduction peaks in TPR (reduction of well dispersed and bulklike Rh₂

o

₃) and three oxidation areas in TPO (oxygen chemisorption and corrosive chemisorption of well dispersed Rh, and thorough oxidation of bulklike Rh).

A start was made in investigating the influence of partiele size upon oxidation-reduction behaviour, but we were not able yet to establish a relationship between SMSI _{behaviour and partiele size. All Rh/Ti0 2 samples} could be brought into the SMSI state.

We shall continue this search dispersion, via the me tal content

by varying the and by varying the

reduction procedure.

This investigation has shawn unambiguously that the TPR-TPO technique is a very powertul tool in discriminating between the various ways in which oxygen can react with a metal, and so TPR-TPO allows a careful analysis of the state of dispersion of the metal on the catalyst to be made.

2.6. References

1. Ichikawa, M., Bull.Chem.Soc.Jap. 51, 2268 (1978).

2. Ichikawa, M., Bull.Chem.Soc.Jap. 51, 2273 (1978).

3. Leupold, E.I., Schmidt, H.-J., Wunder, F., Arpe, H.-J. and Hachenberg, H., E.P. 0 010 295 Al.

4. Wunder, P.A., Arpe, H.-J., Leupold, E.I. and Schmidt, H.-J., Ger. Offen. 28 14 427.

5. Bartley, W.J. and Wilson, T.P., Eur.Pat.Appl. 0 021 443.

6.

7.

s.

9.

Castner, D.G., Blackadar, R.L. and Somorjai, G.A. I

J. Catal. 66, 2 (1980).

Watson, P.R. and Somorjai, G.A. I J.Catal. 72, 347

( 1981).

Watson, P.R. and Somorjai, G.A. I J .Catal. 74, 282

(1982).

Ichikawa, M. and Shikakura, K. in "Proceedings of the 7th International Congres on Catalysis", Tokyo 1980 (T. seyana and K. Tanabe, Eds.), part A, p. 925,

E1sevier, Amsterdam.

10. Wi1son, T.P., Kasai, P.H. and E11gen, P.C., J.Cata1. 69, 193 (1981).

11. Rice, C.A., Wor1ey,

s.o.,

Curtis,

c.w.,

Guin, J.A. and Tarrer, A.R., J.Chem.Phys. 74, 6487 (1981).

12. Wor1ey,

s.o.,

Rice, C.A., Mattson, G.A., Curtis,

c.w.,

Guin, J.A. and Tarrer, A.R., J.Chem.Phys. 76,

20 (1982).

13. Wor1ey,

s.o.,

Rice, C.A., Mattson, G.A., Curtis,

c.w.,

Guin, J.A. and Tarrer, A.R., J.Phys.Chem. 86,

2714 (1982).

14. Primet, M., J.C.S.Farad.Trans.II, 74, 2570 (1978).

15. Yates, O.J.C., Murre11, L.L. and Prestridge, E.B., J.Cata1. 57, 41 (1979).

16. Tauster, S.J., Fung,

s.c.

and J.Am.Chem.Soc. 100, 170 (1978).

Garten, R.L.,

17. Tauster, S.J., Fung,

s.c.,

Baker, R.T.K. and Hors1ey, J.A., Science 211, 1121 (1981).

18. Baker, R.T.K., Prestridge, E.B. and Garten, R.L., J.Catal. 50, 464 (1979) ..

19. Huizinga, T. and Prins, R., J.Phys.Chem. 85, 2156 {1981).

20. Uchijima, T., Herrmann, J.M., Inoue, Y., Burwe11 jr., R.L., Butt, J.B. and Oohen, J.B., J.Cata1. 50, 464

(1977).

21. Kobayashi, M., Inoue, Y., Takahashi, N., Burwell . , R.L., Butt, J.B. and Cbhen, J.B., J.Catal. 64, 74

(1980).

22. Nandi, R.K., Georgopoulos, P., Cohen, J.B., Butt, J.B. and Burwell jr, R.L., J.Catal. 77, 421 (1982).

23. Nandi, R.K., Molinaro, F., Tang, C., Oohen, J.B., Butt, J.B. and Burwell jr., R.L., J.Catal. 78, 289

(1983).

24. Robertson, S.D., McNicol, B.D., de Baas, J.H., Kloet,

s.c.

and Jenkins, J.W., J.Catal. 37, 424 (1975).

25. Jenkins, J.W., McNicol, B.D. and Robertson, S.D., Chem.Tech 7, 316 (1977).

26. Hurst, N.w., Gentry, S.J., Jones, A. and McNicol, B.D., Catal.Rev. Sci.Eng. 24, 233 (1982).

27. Boer, H., Boersma, N.F. and Rev.Sci.Instr. 53, 349 (1982).

Wagstaff, N.,

28. Crucq, A., Degols, L., Lienard, G. and Frennet, A., · Acta Chim. Acad.Sci.Hung. 1982, 111.

29. Sinfelt, J.H. and Yates, D.J.C., J. Catal. 10, 362 (1968).

30. Wanke, S.E. and Dougharty, N.A., J.Catal. 24, 367 (1972).

Vis, J.C., Koningsberger, D.C. and Prins, R., J.Phys.Chem. 87, 2264 (1983).

34. Yao, H.C., Japar,

s.

and She1ef, M., J.cata1. 50, 407 (1977).

Chapter 3

The Morphology of Rhodium Supported on Ti0 2 and Alz 03 as Studied with Temperature Programmed Reduction-Oxidation and Transmission Electron Microscopy.

Supported Rh/Al 2o 3 and Rh/Ti0 2 catalysts with varying roetal loadings have been investigated with chemisorption and temperature programmed reduction and oxidation. Hydragen chemisorption shows that all the rhodium on Al 2o 3 is well (H/Rh > 1 for loadings < 5 wt% and H/Rh > _{0.5 up to 20 wt%), dispersion on Tio 2 is much} lower. TPR/TPO shows this is due to the growth of two different kinds of rhodium /Rh₂o_{3 on Ti02 ; one kind} easily reduced/oxidized, showing high dispersion, the ether kind harder to reduce/oxidize, showing lower dispersion. TEM has shown that the first kind of Rh 2o 3 consists of flat, raftlike particles, the secend kind of spherical particles.

3.1. Introduction.

Over the past years rhodium has been gaining importance in catalytic chemistry. Not only is rhodium widely recognized as the best catalyst to promate the reduction of NO in three way catalysts (1-3), i t also takes a special place in the conversion of synthesis gas, since its product range can include oxygenated products (alcohols, aldehydes, acids) besides hydrocarbons (4-11). Various workers have tried to influence the selectivity

and activity of the supported rhodium catalysts in syngas conversion via a special· preparatien (4,5), via additives (6-8) and via control over the oxidation state of the rhodium in the catalysts (9-11).

Ichikawa deposited rhodium-carbonyl clusters on various supports (4, 5) and after pyrolysis of the clusters, he found a large range of selectivities towards oxygenates in the hydragenation of

co.

He explained this by the acid-bas& properties of the supports (12).

Somorjai c.s. tried to hydrogenate CO over unsupported rhodium and rhodium foil, and found only hydrocarbons, unless they preoxidized the metal (9). _H2

/co

atmosphere led to the reduction to rhodium metal and the production of hydrocarbons, whereas Rh203.5H20 was better resistant towards reduction and produced oxygenates for a long time

( 10).

Where the influence of additives or mixed oxides worked out into the right direction of enhanced oxygenate production, some workers traeed this to the presence of rhodium ions in the surface (11,13). The supposed presence of rhodium ions has been a point of discussion for quite some time. Several authors claimed it to be present in mono-metallic rhodium catalysts. W?rley et al. investigated a.o. a 0.5 wt% Rh/Al203 catalyst via infrared spectroscopy of adsorbed CO (14-16), and ascribed several infrared bands to CO molecules bound to isolated Rh(I) sites. This is in contrast with earlier findings by D.J.C. Yates et al. (18). They investigated some Rh/Al 2

o

_{3 catalysts with electron microscopy and}

found rhodium to be present as metallic rafts. They did notmention any isolated Rh(I) sites, but, as Worley and coworkers stipulated themselves, the two groups used very different methods of preparatien of the samples.

In all cases it seems obvious that the support plays an important role in either bringing or keeping the metal in

a certain state of (un)reactivity. A special example of such an interaction between metal and support has been described by Tauster et al. (19, 20) and is now known as Strong Metal Support Interaction (SMSI). Supported noble and transition metals such as Ft, Rh, Ru etc. are normally capable of chemisorbing a.o. H2 and CO. But if they are supported on oxides as Ti0 2 , V205 and Nb205, and i f they have been reduced at high temperatures (e.g. 773 K), this chemisorption capability is greatly diminished. This SMSI phenomenon has been related to the occurrence of lower oxides of the supports (21), although these are known to be formed at lower temperatures than necessary to cause SMSI (22) and the exact nature of the interaction still remains unclear. The SMSI state can be destroyed according to Tauster c.s. by oxidation at elevated temperatures, followed by

reduction (473 K): this procedure chemisorption behaviour (19).

low ternperature restores normal All of the above rnentioned phenomena have to do with one common property: The oxidation-reduction behaviour of supported rhodium catalysts. We therefore decided to study a nurnber of Rh/Al20 3 and Rh/Ti02 catalysts with varying rnetal loadings (see Experimental). A1203 was chosen because i t is known as a support giving good dispersions and stable catalysts, and Ti02 because i t is known to exhibit SMSI. We varied the metal loading to create a variatien in partiele size, to see if and how oxidation-reduction and SMSI behaviour are influenced by partiele size.

Befare we come to the experimental techniques we used, we want to introduce aother item: passivation. I t is obvious that reduced catalyst systems cannot simply be removed from the reduction reactor and then be stared in air for later u se~ we stabil i ze the me tal surface by applying a layer of oxygen upon the metal particles in a

controlled way (see Experimental): We passivate the catalysts. Although a simple low temperature reduction is sufficient to remove the passivatien oxygen again (as will be shown), some authors have given attention to the state the catalysts are in after storage in air. One can see this as a prolonged passivation, but without the precautions we take to prevent uncontrollable effects upon the first contact between air and the reduced roetal catalyst. So Burwell Jr. et al. used Wide Angle X-ray Scattering, Extended X-ray Absorption Fine Structure, hydragen chemisorption and hydrogen-oxygen titration to characterize their supported Pt and Pd catalysts (23-26), and they found their catalysts to be oxidized to a great extent after prolonged storage in air.

We will show that a good insight in all these matters can be gained with the aid of Temperature Programmed Reduction and of Temperature Programmed Oxidation (TPR and TPO), supported by chemisorption measurements. TPR as a characterization technique was presented by Jenkins, Robertson and McNicoll in 1975 (27,28) and has been used extensively in the past few years. The development has been reviewed by Hurst et al. (29). The technique allows one to get (semi)quantitative information about the rate and ease of reduction of all kinds of systems, and once the apparatus has been built the analyses are fast and relatively cheap. We used an apparatus as described by Boer et al. (30), which enabled us to extend the analyses to Temperature Programmed Oxidation, and to gather information about the rate and ease of oxidation as well.

3.2. Experimental.

Tio 2 (anatase, Tioxide Ltd., CLDD 1367, surface area 20 m2/g, pore-volume 0.5 cm3/g) and Y-Al203 (Ketjen,

wt% Rh 2.3 4.6 8.5 H/Rh 1.53 0.96 0.81 0.67 0.54

Table 1. Hydragen chemisorption of the Rhjy-Al203 (RA) catalysts.

11.6 20.0

H/Rh Table 2. Hydrogen chemi-wt% Rh

LT( 1 J HT(ZJ sorption of the Rh/TiOZ (RT) catalysts. 0.3 1.10 0.00 0.7 0.61 0.01 1.0 0.41 0.01 2.0 0.35 0.01 3.2 0.22 0.02 8.1 0.12 0.04 ( 1 ) : reduced in situ at SZ3 K. ( 2) : reduced in situ at 773 K.

impregnated with aqueous so1utions of RhCl3.xH20 via the incipient wetness technique to prepare the catalysts. Their characteristics are presented in Table 1 and Table 2. The catalysts will be denoted from now on as RT (Rh/Ti0

2) and RA (Rh/Alz03 ) catalysts, foliowed by the metal loading. After impregnation the catalysts were dried in air at 355, 375 and 395 K for 2 hours successively, foliowed by direct pre-reduetion in flowing Hz at 773 K for one hour. Prior to removing the

from the reduction reactor they were

subsequently slowly adding 0 up to 20

%.

Then the catalysts were taken out of the reactor and stared for further use.

In the TPR-TPO apparatus used a 5 % H2 in Ar or a 5 %

o₂ in He flow can be directed through a microreactor, which is connected to a temperature programmer. H2 or 02 consumption is being monitored continuously by means of a Thermal Oonductivity Detector (TCD). A typical sequence of experiments is as fellows:

- the passivated or oxidized sample is flushed under Ar at 223 K

- Ar is replaced by the Ar/H2 mixture, causing at least an apparent H consumption (first switch peak)

the sample is heated under Ar/H 2 flow with 5 K/min to 873 K

after 15 min at 873 K, the sample is caoled down with 10 K/min to 223 K

- the reduced sample is flushed with Ar

- Ar flow is replaced by the Ar/H2 mixture once more, now causing only an apparent H consumption (second switch peak).

An identical sequence is followed during TPO, so the final oxidation temperature in TPO is also 873 K, unless stated otherwise.

The switch peak procedure deserves some closer attention. The streng signal we call the first switch peak is due mainly to the displacement of Ar by Ar/H2 in the reactor, but in some cases real hydragen consumption might take place, even at 223 K. Therefore we repeat the whole procedure after the TPR has been performed: In that case the catalyst has been reduced and caoled down to 223 K, and as a consequence i t is covered by hydrogen. Then we replace the Ar/H 2 by pure Ar. Subsequently we switch back _{to Ar/H2 . Since we cannot expect any hydragen} consumption from the reduced, hydragen covered sample at

this time, the resulting second switch peak will be due solely to the displacement of Ar by Ar/H2· So the difference between the first and second switch peak reveals the real hydragen consumption at 223 K, if there is any.

The reactions that might take place during TPO and TPR are:

4 Rh _0.75) [1]

The quantities between brackets are the hydragen or oxygen consumptions in TPR or TPO expected for reduction of bulk Rh 2

o

3 or formation of this very material (apart from chemisorption of any kind). In a standard experiment a TPR is done on a passivated catalyst, followed by TPO (on the now reduced catalyst), followed by TPR (on the now oxiclized catalyst).

1.2

•

0.8 0.4 4 8 12 16 20 wt'l. Rh~

Figure 1. Hydragen chemisorption of Rh/y-Al20 3 catalysts as a function of roetal loading. T(red) is 773 K.