SiO2 catalysts

SiO2 catalysts

Applied Catalysis, 12 (1984) 309-330 Elsevier Science Publishers B.V., Amsterdam -Printed INFLUENCE OF METAL-SUPPORT REDUCIBILITY M. MONTESa, INT...

3MB Sizes 0 Downloads 8 Views

Applied Catalysis, 12 (1984) 309-330 Elsevier Science Publishers B.V., Amsterdam -Printed

INFLUENCE

OF METAL-SUPPORT

REDUCIBILITY

M. MONTESa,

INTERACTIONS

AND CATALYTIC

Ch. PENNEMAN

ACTIVITY

309 in The Netherlands

ON THE DISPERSION,

OF Ni/SiO2

DE BOSSCHEYDE,

DISTRIBUTION,

CATALYSTS

B.K. HODNETT,

F. DELANNAY,

P. GRANGE

and

B. DELMON Groupe Place

de Physico-Chimie

Croix du Sud, 1, 1348 Louvain-la-Neuve,

a0n leave from Dpto. San Sebastidn,

(Received

Universitg

Miner-ale et de Catalyse,

Fat. Qufmica

Q.T&nica,

Catholique

de Louvain,

Belgium. de San Sebastiin

(UPV/EHU),

Apdo

1072,

Spain.

23 March

1984, accepted

29 June 1984)

ABSTRACT The influence of metal-support interactions on the dispersion, distribution, reducibility and catalytic activity for benzene hydrogenation of Ni/Si02 catalysts, prepared by impregnation or deposition-precipitation, was studied. Weak interactions were observed for the former; the latter technique gave rise to strong interactions. Nickel contents were from 6-9 wt%. The impregnated precursor featured nickel evenly distributed between the internal (pores) and external surface of the support. Direct reduction did not noticeably change the distribution and a Ni/SiO catalyst formed in which the active phase was finely divided and evenly distri.guted. Calcination of this precursor caused migration of the nickel to the external surface with the formation there of large crystallites of NiO. The precursor prepared by deposition-precipitation, consisted of nickel hydro-silicate deposited uniquely on the external surface and in the pore mouths. Calcination or reduction of this material did not noticeably change the distribution, but high dispersion was observed in both cases. The stability of catalysts prepared by this technique is discussed in terms of interaction between nickel and silica.

INTRODUCTION It has been recognised by a variety the metal

that Ni/Si02

which

result

and the support.

Three

preparation

1) impregnation; studies

for some years

of techniques

2) ion-exchange;

have focused

and whereas

these

have often

full characterisation characterisations over the whole

methods

can be prepared

of interaction

of catalysts

and compared

been achieved. in which

between

have been most widely

A large number

prepared

the active

of

by these techniques

two of the above

Especially

used:

neglected phase

techniques, in previous

is distributed

of the support. of the active

dependent

out in an oxidizing from the latter

examined

has rarely

catalysts

extents

3) deposition-precipitation. aspects

has been the manner

The dispersion ion is notably

on diverse

in varying

upon whether

or reducing

[I]. Several

0166-9834/84/$03.00

phase

in Ni/Si02

catalysts

its initial

thermal

atmosphere.

studies

Better

of calcined

prepared activation

dispersion

catalysts

0 1984 Elsevier Science Publishers B.V.

by impregnatis carried

generally

prepared

results

by impregnation

310

indicate oxide;

that NiO is usually

dispersion

and only weak

interactions

By contrast, that a surface to the metal particle

reports

dispersed

catalysts,

a strong

urea. These

of the latter ions, either conditions

can be carefully

by injection

of up to 30 wt*: nickel

as solution

Several achieved

is formed

Despite

of 40 mole the above

ised. As indicated the internal Ni/Si02 silica

above,

precipitation

are present

This publication ence of preparation to deactivation investigation support a number

for other

parameter

parameters

and poisoning of the influence

on the dispersion, of Ni/SiO2

treatment

on the stability LIZ]. Here details of interaction

between

In order to appreciate

that pore

increase

catalysts

[Ill.

the influ-

with

respect

of a physico-chemical

the active

phase and the

and catalytic fully

hints

by deposition-

to determine

are reported

reducibility

conditions

[IO]. Some

loading

of Ni/SiO2

and

in air or

et al. who showed

programme

for any

has not been studied.

prepared

time or nickel

between

nickel

of preparation

catalysts

part of a wider

distribution,

catalysts.

between

[Z].

character-

or pellets

such as Ni/v-A1203

for Ni/Si02

as preparation

constitutes

of the carrier

of nickel

particles

thermal

in the work of Richardson

develops

the whole

is con-

up to the

be said to be fully

as a function

systems

can be

interaction

loadings,

of interaction

after

[9].

that a nickel-layer-silicate

by deposition-precipitation

catalytic

in this parameter

constriction

in catalyst

the influence

prepared

in the distribution

as to variation

zones

Magnetic

can however

increase

With higher with

of

1 and 3 nm

that good dispersion

is known of the distribution

surface

the final dispersion

of the catalysts

Changes

little

between

loadings

cannot

The release

of a homogeneously

temperatures

occurs

these catalysts

so as to

by the decomposition

size distribution

low nickel

% NiO, interaction

In addition,

have been observed

with

As

hydroxide.

of the carrier.

have confirmed

layers of the support.

studies

in determining

hydrogen

and reduction studies

[S].

by the controlled

a range

[7,8] and XPS has indicated

and external

catalyst.

indicate

[9]. Particle

[Z]. For samples

to the outermost

equivalent

mouth

microscopy

by this technique

structure fined

concentrations

electron

on the surface

size distribution

process

can be modified

of base or often

condit-

technique.

and nickel

determined

and metal

to reduction

particles

to lead to the formation

hydrosilicate

for loadings

nickel

the support

to reduce

of reduction

in the reduction

of small

12-41.

indicate

is difficult

that the degree

conditions

between

of particle

and the support

forms which

formed

reduced

by ion exchange

the deposition-precipitation

are reported

layer of nickel

phase

to the bulk

is easily

are very sensitive

vapour

reaction

interaction

measurements

widen

nickel

the formation

have developed

to Ni/SiO2

of hydroxide

of water

and favours

Geus et al. [6-81

concentration

dispersed

oxide

prepared

et al. have shown

removal

is very similar

the active

catalysts

for this type of catalyst

sintering

it applies

occur between

of NilSi

[3,4]. Martin

ions. The efficient

bring about

in a form which

very poor; the supported

layer of highly

diameter

prevents

present

is usually

activity

interaction

of

effects,

311

samples

of Ni/SiO2

the metal

and the support,

precipitation various

method,

thermal

by impregnation,

prepared

are compared

interaction

in which

treatments

with no interaction

with samples should

in air or hydrogen

prepared

expected

between

by the deposition-

be maximised.

The influence

investigated

for each tYPe

were

of

of preparation.

EXPERIMENTAL Preparation

of catalysts

Silica

supported

technique

nickel

surface

distribution

fraction

between

maximum

The dry impregnated

content

be referred

heated

washed

of I and G were reduced

(heating

treating

in air.

ed by exposure

1. Catalysts

and the

with distilled

content either

calcined

prepared

by this method

were mixed

water

measurements,

with

The will

80 grams of

to 3.5 by addition

at which

of nitric

then added.

time the suspension

at 90°C and dried

in Table

The

was

at 120°C

1.

in air at 450°C

for 16 h (subscript

(190 ml min -' per gram of catalyst)

for 16 h (subscript

C) at

R).

it was necessary

a passivation

at 25°C

of

by the deposition-precipitation

is indicated

in a flow of hydrogen

catalysts

a solution

I.

for 6 hours,

In these cases,

the reduced

adding

to 90°C; 50.4 grams of urea were

rate 8°C min-')

For some physico-chemical catalysts

was ground

to 99 grams of SiO2, with constant

The pH was adjusted

to proceed

the filtrate

or directly

in Table

water.

for 16 h. The final nickel

500°C

In both cases,

at 12O'C for 16 h under vacuum.

with ca. 8 wt: nickel

acid and the suspension

Batches

by slowly

G) 81.4 grams of Ni(N03j2.6H20

Si02 in 2 1 of distilled

was al lowed

was dried

by the abbreviation

catalysts

(abbreviation

was prepared

in 79.2 ml of water

obtained

is indicated

to below

To obtain

use, the carrier

at 8 nm. Before

catalyst

The precursor

filtered,

[S-S].

50 and 100 urn was used.

42.8 grams of Ni(N03)2.6H20

reaction

method

wetness

by Kali Chimie, grade AF125 (99.8:, Si02), with 3 -1 2 -1 and with a pore area of 275 m g , pore volume 0.80 cm g

specific

stirring.

by the incipient

used was supplied

diameter

method

were prepared

[12] or by the precipitation-deposition

the silica

nickel

catalysts

procedure

to manipulate

was used which

reduced involved

in a flow of 1% O2 in Ar for 30 min follow-

to a flow of 54: 0

to this treatment

will

in Ar for a further 30 min. Catalysts subjected 2 be identified below by the subscript p, (e.g. IRp or GRp,

etc.,)

Characterisation

techniques

After dissolving determined

Of

by atomic

Metallic

surface

hydrogen

at

the catalyst absorption

in hydrofluoric

areas were measured

250~~ after

acid, the nickel

content

was

spectroscopy, in a volumetric

prior reduction

system

of the catalysts

in

by

chemisorptio,,

hydrogen

at

500~)~.

312

314 the sensitivity models (i)

factors

determined

Homogeneous

Solid Model according

INT 1 -==__x

i 1

CONC 1

INTZ

i2

CONC 2

where

INT is the intensity

determined

by Wagner

concentrations (ii)

Uniform

Layer Model,

i 1

INT2

12

i is the sensitivity

= 3 and iSi2p = 0.27)

following

and C the atomic

assumes

that a solid

(component

I) is covered

II. Thus

1 _ EXp-b/h*2

x

EXP-b'xL'

CnNC 2

n. In this work,

the method

For some samples comprised

outlined

by Szajmann

using a Hewlett

trial and error

Catalytic

activity

prepared

Fluka "purum"

(thiophene

Packard

technique

system

with a tubular

was under

rates were used, preliminary

to zero between

of solid

the catalysts

hydrogen

and analysis

being present

that the activation

180 and 200°C;

studied,

thus the latter

with

the C6H6 partial

ed with the extreme

exothermicity

pressure

literature

was 47 Torr.

of this reaction

Raney

of "dry Na" tests

consisted gas

tng was used was always

two heating reports

for this reaction

temperature

of

silica.

for 16 h at 500°C;

energy

with

by on-line

in the reactor

The total flow rate was 340 ml min -' and the total

1 atmosphere;

batch

20-200

with additional

(100 ml min-')

A

used was

by refluxing

over a fresh

peak

line shapes.

for the activity

used for the catalytic

i.e. 8 and 30°C min -I. Consistent

tests indicated

gaussian

The benzene

purified

Pyrex reactor

upon the catalyst

to 400 mg by diluting

Activation

catalysts.

it was stored

arrangement

for each run but the total amount adjusted

Ni/Si02

free) and was further

Depending

assumed

was used as a test reaction

The experimental

chromatography.

Mod 9835B + HP 1351 A micro-computer. was used which

and activated

nickel for 2 h. After distillation

of a dynamic

et al. [16].

reduced

of benzene

the variously

IPb-Na alloy).

of peak m

were calculated

and passivated, the observed Ni 2p3:2 2+ from Ni" and Ni . The individual contributions were

graphical

The hydrogenation

the mean free paths of the electrons

of Ni/Si02,

contributions

deconvoluted

factor

being studied.

which

x CONC 1

peak;

b is the layer depth and x z the mean free path of the electrons

in component

tests.

with 2

to which

et al. Cl41 (iNi2p3,2

layer of component

INT 1 -=-

was

et al. Cl41 in combination

of a given X.P.S.

of the elements

by a uniform

where

by Wagner

[157:

was chosen pressure To avoid

[17],

was close

for standard

(H2 + CBHB) problems

associat-

~~~~~~~~~ = 51.2 kcal mole-'),

315

FIGURE

1

standards.

X-ray diffraction

patterns

of various

Ni/Si02

catalysts

and some

31’6 the conversion

was never allowed

of Satterfield

and Sherwood

Delmon

[19] for diffusion

not diffusion To avoid adopted

switching

to an argon

ed. The temperature added

cooling

on the models

and that proposed

that the reaction

at the start of the reaction,

the activated

catalyst

flow for 15 min to desorb

was then further

to the system.

the argon-benzene

diffusion

indicated

based

kinetics

by were

in the test conditions. run-away

involved

10%. Calculations

[I81 for external

in pores,

controlled temperature

which

to exceed

reduced

The temperature

any hydrogen

raised

taken as zero time on stream).

An increase

after the addition

but the temperature

of hydrogen

min at less than 200°C;

flow.

in temperature

the final temperature

remained

usually

adsorbwas

stabilised,

(This point was

was usually stabilised

was then quickly

and

point benzene

to 180°C and, once

by a hydrogen-benzene

was

in hydrogen

which

to 25°C at which

was again

flow was replaced

a protocol

to 200°C

registered within

regulated

10

to this

value. Values

for catalytic

conversion

activity

was established,

values were

reproducible

reported

which,

below were measured

generally,

to within

required

after

steady

2-3 h on stream.

state

These

+ 5%.

RESULTS The specific are presented lined above

surface

in Table

areas,

and subjected

The lower nickel

loading

In addition,

this material

surface

prepared

areas.

Nickel

1. The precursor, the supoort.

istic peaks were lower degree

was highly

hygroscopic

less intense

ed for ICRp, but the shoulder

less pronounced. dispersed

its surface

and broader,

indicating

Peaks characteristic

Far the

very similar except

1 are presented

indicates

lc.

in Figure

oeaks associated

of Ic clearly

with

the presence

for

1 but its characterRP a smaller particle size or

of metallic

on the peak at 28 = 44.5"

nickel

indicates

were detect-

the presence

of NiO. of the samples

crystalline

prepared

of G and GC indicated

materials

peaks for NiO were detected

by deposition-precipitation

of unknown

the oresence

structure,

for the corresponding

whereas

reduced

was much of well poorly

and passivated

samples. Further

indications

I.

to outgassing

area.

for all samples

in Table

only the broad background

The diffractograms

or poorly

listed

the diffractogram

of crvstallinity.

by the precursor,

i.e. I or G, exhibited

was verv similar

treatments.

Ic, may be explained

and its sensitivity

of this oxide was again indicated

The crystallinity

size and dispersion

and passivation with

retained

in determining

of the samples

I, exhibited

also of small amounts

particle

reduction

and water

by the same method,

Bv contrast,

of NiO. The presence

resolved

of nitrate

dispersion

X-ray diffractograms

nickel

for I by comparison

aave rise to a large error

rest, samples

areas,

to calcination,or observed

in terms of the larger amount

conditions

metallic

1 for catalysts which were prepared by the two methods out-

of the state of dispersion

of nickel

in these catalysts

317 TABLE X.P.S.

2 binding

energies

and intensities

Binding

of Ni/SiO,

energy/eV

catalysts

Percentage

Sample

f

103.4

IC I R"a, IRP,

Ni+;p

Niopp

si2p

Ni"

Ni++

INi:Isi

855.8

0

100

0.81

103.2

852.9

855.8

20

80

0.34

103.5

853.0

855.9

19

81

0.26

103.5

853.1

856.2

49

51

0.33

103.5

853.1

856.0

34

66

0.16

103.5

852.9

855.7

69

31

0.14

856.7

0

100

8.49

856.1

23

77

4.19

103.4

'RP a I 'CRpb 'CRpc I CRP

0.25

103.3

Gc d GRpb

-

103.3

GRp d GCRPb G CRP

852.9

103.5

853.1

856.0

23

77

2.16

103.6

853.1

855.8

32

68

3.81

103.5

853.0

855.7

26

74

1.86

852.9

856.0

103.7

GRpe

3.53

103.7

GCRpe

18

82

3.29

aUnground sample mounted with the aid of double sided adhesive tape. b Sample ground and mounted with hydrolic pressure (2500 kg cm -2 for 1 minute) 'As b, but hydrogen treated in situ to remove passive NiO layers d Unground sample mounted by applying light pressure with a spatula eAs d, but hydrogen treated in situ to remove passive Ni02 layers f The symbol (-) denotes that the indicated entry was not measured.

and the crystallinity study.

C.T.E.M.

for I, either

did not distinguish

The micrograph ranging

in the range

between

bigger

Electron

between

2a,

recorded

agglomerates

nm, with some particles pore diameter

micrographs

of nickel

surface

and silica or due to a

shows

the presence

5 and 50 nm, with the majority

10-25 nm. The micrograph

on the external

the presence

microscopy

and the support.

for IRp, Figure

of large

from the electron

small size of the crystallites

nitrate

in size between

than the mean

must reside

nickel

recorded

and shows the presence range 80-200

phase was obtained

due tn the extremely

lack of contrast

particles

of the active

for IC is presented

of cubic shaped

as small as 20 nm. These

of the support

of nickel

of the oarticles in Figure

particles

in the

large particles

(8 nm), which

implies

2b

are

that they

or in the macropores.

of ICRp showed

similar

features

to IC, but better

contrast

318

FIGURE

2

Electron

micrographs

of a) IRp, b) Ic.

FIGURE

2 (continued)

Electron

micrographs

of c) GC and d) GCRp.

320

0

FIGURE

3

T.P.R.

100 200 300

profiles

Electron latter

pure support covered

micrographs

exhibited

reduction

but an additional 2d). X.R.D. measurements

recorded

in Fiaure

by a compound

After

of metallic

bv the corresponding

is presented

feature

texture.

spread out

appeared,

of the crystall-

pattern

(Figure

Otherwise

the support

i.e. small crvstallites the presence

that a small amount

The

be detected aopeared

in the form of fine sheets

indicated

1).

identical.

free SiO2 could

to GCRp the basic background

of this sample

(Table 2) indicated

in the interior

for G and GC were essentially

a granulous which

nickel

X-ray diffraction

2~. In places where

and passivation

analysis

TI

of I, IC, G and GC and bulk NiO.

was noted due to the presence ites as indicated

LOO 500 600

the

to be

or as filaments.

features

remained

3-6 nm in size

(Fiaure

of NiO but the X.P.S.

of Ni remained

after

passivat-

ion. T.P.R. Figure

profiles

for I, IC, G and GC and for unsupported

3. The profile

peak emerged The maximum featured

rate of reduction

a shoulder

Reduction

for I exhibited

at a lower temperature

at 370°C;

was observed reduction

of G and GC occurred

for the imoregnated

samoles.

2 small peaks below

than the corresponding

NiO are presented 200°C;

its principal

peak for bulk NiO.

at 330°C for IC, but its profile

of this sample was complete

at considerably

Thus a maximum

higher

reduction

in

temperatures

also

at 450°C. than required

rate was observed

for G at

321

-?A c

r

850

FIGURE

4

CompositeNi2p3,2X.P.S.

ionsdueto

47O"C,

individual

oxidation

with a shoulder

in these conditions

a single

peak centred

For subsequent, catalytic

testing,

by treating

for the Ni

865

peak for ICRp showing

below 600°C.

treatments,



r



I

eV

870

deconvolution

of contribut-

side of the peak;

The T.P.R.

profile

continued

physico-chemical

reduction

to occur

up to 650°C.

characterisation

that all these catalysts

was not

for GC comprised

could

and

be fully

reduced

at 500°C for 16 h.

of the X.P.S.

shows the Ni2p3,2

,



states.

it was assumed

study are summarised

in Figure

peak for ICRp; the individual

Ni2+ are shown deconvoluted

measured

,

860

at 510°C but reduction

thermal

in hydrogen

The results

.

on the low temperature

complete

figure

-.

055

below

the composite

4 and Table

contributions

peak. The binding

2. The

by Ni" and

energy

measured

312 peak for metallic nickel was 853 i 0.1 eV. The corresponding value 2P for NiO was 855.9 * 0.3 eV, except for GC, for which a value of 856.7 eV

activities

109

for benzene -1 . rates of 8 or 30°C min

Catalytic

with heating

FIGURE 5

5u

-_

~~--~--r-~~__-____-_

-_-----

hydrogenation

t ime/min



366

x lo?

as a function

150

f’

‘I



mol m6

g-‘NI

gNI_’

of time for IR, ICR, GR and GCR activated

X-X-X’.

&G,, I' 370 x lo-’ AG, '8 36 z x lo-3

01,



31 L x 16’

182 Y 10.’



304 x10-3





K 5 x 10-X



mol ml6

125 x10?

\c eq

at 5OO'C

323 TABLE

3

Specific

activities

of Ni/Si02

catalysts

Catalytic

Sample

activitya -1 m-*Ni) min

x IO4

(mol benzene 1.2

ICR

2.8

IR

2.5

GR

2.7

GCR aHeating

rate during

reduction

was 8°C min

-1

was recorded. As summarised ably higher

in Table

2, the Ni:Si

for the samples

prepared

it may be noted that reduction

addition,

had a detrimental characterised

effect

noticeably

by lower Ni:Si ratios.

reduced

ratio measured

increased

rates were corrected

activation

pretreatments

IR was sensitive

as large when

The catalytic heating

activities

rate used and were

Specific

the very

activities

except

to a much

I in this case the Rp' extent than the factor

smaller

The proportion

of Ni++ detected

by X.P.S.

of grinding.

tests are presented the varying

before

for ICR,

rate,

were

Grinding

to this parameter

the lower heating

catalysts to remove

in Figure

amounts

5. The reaction

of catalysts

per gram of nickel.

R in the subscripts

in hydrogen

were recorded

IC and GC,

all reduced treatment

rates are expressed

the final

catalysts,

In

this ratio.

to take into account

activities,

Low activities

activity

consider-

method.

did not influence

of the severity

of catalytic

by X.P.S,were

An in situ hydrogen

had changed

for all the others.

for each test and the final to catalytic

of the calcined

the Ni:Si ratio for all samples

as a function

The results

samples

after grinding

of ca. 2 observed

as measured

in each case and, in general,

NiO layer from the passivated

Only

ratios,

by the deposition-precipitation

used

When referring

to I and G refer to the

testing.

irrespective

of the heating

and exhibited a reaction -1 ) was used.

rate used.

rate about twice

(8°C min

of the samples

GCR and GR were

insensitive

to the

similar.

per unit area of metallic

nickel

are presented

in Table

3.

DISCUSSION As the object of this work was to study the influence actions nickel

on the reducibility, supported

the evidence

on silica,

for interaction

dispersion,

distribution

this discussion to be found

will,

of metal-support

and catalytic

in the first

in the physico-chemical

inter-

activity

place,

of

evaluate

characterisation.

324

The influence

of interaction

on the parameters

of interest

will

then be considered.

Interaction Neither

X.R.D.

of the supported nitrate other

nor Electron nickel

or hydroxide,

impregnated

Microscopy

the latter

(Figure 2b) and X.R.D.

analysis literature

several

that interaction a)

The B.E.T.

formed

during

indicated

drying were

the nature

it was in the form of a

in vacua

at 120°C. For the

distinguished

by C.T.E.M.

NiO (Figure

I), in

for this type of preparation of evidence

[2-41.

may be cited which

indicate

occurred: surface

area for these

catalysts

Since nickel alone could not be present this increase,

in determining that

that those were

reports

G and GC, four pieces

For the samples

useful

IC, large crystallites

sample

agreement

with

were

it is assumed

in I. However,

this finding

implies

was

10% higher

in a sufficiently

the formation

than for the support.

porous

of a complex

fnrm to explain

between

nickel

and

the silica. b)

The binding

energy

for the Ni

for IC, i.e. NiO. Assuming the formation c)

latter d)

of a different

E.M. showed

composed

3/ peak was 0.9 eV greater than that recorded 2P 2 the same oxidation state, i.e. +2, these data point to compound

the presence

uniquely

of silica

in this case.

of two zones for these and the second

part must thus have been very porous

The X.R.D.

data

which

indicate

catalysts;

composed

of nickel

to explain

the presence

one granulous and silica.

the increased

of poorly

The

surface

crystalline

area.

materials

or small particles. These findings

taken together

by the same method, silicate

indicate

[6]. The structure

strong chemical

interaction

to that of nickel

hydroxide

layer of tetrahedral silicate between

platelets

or rolled

Influence

texture

literature

formed

arises

hydroxide

It has already

[13].

because

Individual

that filaments in Figures

hydrc-

as the result

of a

corresponds

ions is replaced

been postulated

arrays

prepared

was nickel

and silica

a layer of hydroxyl

2-dimensional

shown

data for catalysts

which

which

the nickel

in which

of this compound

of interaction

Numerous reduction

between

to give the impression

is the turbostratic

the literature

of this compound

silica.

can form in almost

with

that the compound

that nickel

by a

hydro-

of very poor adhesion

platelets

had formed

can become

folded

C221. The overall

effect

2c and 2d.

on the reducibility reports

of NiO so that

indicate

its reduction

that nucleation rate is expected

is the slow step in the to decrease

as the particle

size diminishes. In view of the X.R.D. NiO, the difference simply

be attributed

data

which

in reducibility to the smaller

indicated

that nickel

of this catalyst particle

was present

vis-a-vis

size which

in IC as

the bulk oxide may

prevails

on the support.

325 The shoulder

observed

tentatively must

be noted

al artefact

to some nickel

associated

The situation

respect

cases can an analogy was never

detected

The process

be drawn with

occurred

Literature

can be assigned

to an improvement

of interaction

The most

reliable

from hydrogen techniques, valuable

reduced

insight

present

indicated

greater

on the basis

which

its presence

of a nickel is stable

between

occurred

G and GC

during

calcinat-

to reduce.

structure;

Thus,

the former

the proportion

particles

were covered

X-ray analysis

results

with

with

size

being measured Cl53 which

a layer of NiO, of the oxide. by Coenen

and

[I33 and the value of 3 mono-

is in agreement

measurements.

that 91 ~01% of the nickel with

the X.R.D.

for I in which metallic nickel alone was detected. CRP For this sample calculation of mean particle size from hydrogen at 45 nm is in good agreement

for

the depth

particle

reported

chemisorption

it was estimated

of ICRp was Ni". This calculation

vis-a'-

of each oxidation

was applied

3 and 4 monolayers

that of 2 monolayers

additional

information

IC, the mean

layer model

from other

Ni" detected

free path of the electron

by Dell et al. [Zl] from oxygen

On the basis of these

give Ni++:

of the percentage

to between

of quantitative

those provide

can be used to calculate

the homogeneous

with

with

can however

for the sample

than the mean

size iS that calculated

data combined

hence an estimation

is in good agreement

reported

observed

particle

and X.P.S.

by the latter

This corresponds

layers

content

samples

layer,

Linsen

LCi

1). These

and shape;

In this case,

0.71 nm in depth.

I emerged or hydroxide.

In none of these

hydrosilicate

it more difficult

of nickel

microscopy

that for ICRp the nickel

This value

in crystallinity

(Table

in these samples.

(0.69 nm) [16].

that nickel

rendered

into the catalyst

(45 nm) was much

with

nitrate

of bulk NiO since

in reducibility

measure

size distribution

and passivated

of nickel

on the dispersion

overall

chemisorption

of the passivation state

which

such as electron

vis particle

control

for G and GC was the reduction

indicate

[Zl]. The difference

Influence

it

samples.

in T.P.R.

reports

at 5OO"C,

the peaks associated

and reduction

the reducibility

in air to 560°C

ion of the latter

to thermodynamic

to G and GC is more complicated.

for any of these

which

hydrosilicate.

at which

to decomposition

with

from kinetic

However,

of ar experiment-

[ZO].

view of the low temperatures

they may be assigned

IC may be

with the support.

may have been the result

with a changeover

profile

side of the peak for

in interaction

that this type of behaviour

of the desorption In

on the high temperature

assigned

the range of particle

sizes

results

chemisorption

(20-200

nm) observed

by C.T.E.M. _I:

Electron

nm, whereas

microscopy

hydrogen

this apparent

indicated

chemisorption

discrepancy,

it must

the presence

gave an average be recalled

of particles value

in the range

5-50

of 5.2 nm. To explain

that as supported

nickel

particles

326 diminish

in size,

as confirmed particles

the proportion

by XPS

limiting

resolution

G and GC;

of the electron

implied

By contrast

with catalysts

a loss of dispersion

reduced,

the opposite

1)

[53. The calcination

treatment

though

during

calcination

effect

[23],

of water

lattice

it could

ity with

respect

Influence

of nickel

greater

was observed

for G and GC.

this observation:

sintering

for Ni/SiO2

the nickel

catalysts

a part of the structural effects

also provoke thermal

water

[23]. a phenomenon

treatments

of nickel

crystallites

similar

which

towards

involve

the interior

and increase

entered their

the

stabil-

on the distribution

greatly.

unground

this finding surface

IR and ICR, as measured

in the samples

1) differed

for the corresponding

on the external

could

i.e. whereby

calcination

to sintering.

(Table

To explain

any sintering

particles

were directly

Even if only a small part of the nickel

to anchor

of interaction

Dispersion isorption

promotes

where

which

could explain

could cause the migration

lattice.

serve

which

NiO from SiO2

individual

by impregnation catalysts

may have eliminated

2)

The loss of water

with

of water

so minimising

of the hydrosilicate

larger,

data.

less pronounced,

of the hydrosilicate,

to the Hoffman-Klemen

to identify

prepared

in the literature

It is known that the presence

the elimination

was much

for distinguishing

chemisorption

by comparison

behaviour,

are known

microscope

it very difficult

by the hydrogen

provoked

Two phenomena

passivation

with IRp of Table 2). It was calculated that I CRP 5.2 nm would consist of 62 ~01% NiO after passivation. The

of 5 nm, making

in the size range

after

(compare

of diameter

was of the order

of NiO present

However,

samples,

it must

the X.P.S.

by hydrogen

measured

IRp (a) and ICRp (a), were almost

be accepted

of the support,

that the proportion

hence

detectable

for ICRp than in the case of I RP' models can be visualised: in the first,

Two limiting

of nickel

by X.P.S.,

nickel,

chem-

Ni:Si ratios identical. present

was much

in large crystall-

ites such as with

IC or I is deposited uniquely on the external surface of CRp' in the second, nickel is deposited on the external surface and within

the support; the pores. The sample

ICRp could

of the second. is offset

by its lower dispersion

for the two samples The influence diminution since

'CRp Conversely,

be visualised

The high external

would

of nickel Ni:Si

ratio measured

if the dispersion

of nickel of Table

in terms of this model;

Ni:Si ratio would

by this treatment

IRp should

face.

Examination

such that the final

can also be explained

measured

the pores exposed grinding

concentration

by X.P.S.

be similar.

of grinding

of the X.P.S.

in terms and I RP in the former

in terms of the first model

surface

have little

was similar 2 indicates

would

be predicted not contain

or no influence

inside

after

a strong

grinding

any nickel.

on the X.P.S.

ratio,

the pores and on the external

that these

limiting

situations

sur-

were closely

327

followed

bv the experimental

upon grinding, that the reached

whereas

but that the model

nickel

for

the Ni:Si ratio for ICRn diminished

this ratio for

case of equal

limiting

These models

results;

only slightly

I

RP internal and external

surface

gives a good approximation

I and IC are consistent

particles

on IC were

generally

diminished

StrOVlY

indicating

dispersion

of the actual

with the EM results

which

Was not

distributionindicate

too big at ca. 45 nm to enter

that

the Silica

pores.

The fact

ion of the Ni crease

these models

that 2+

:Ni" ratios

represent

in this ratio observed

been because

the limited

limiting

of nickel

is emphasized

by considerat-

grinding. The in.2+ to 66%, must have i.e. from 51% Nl

after grinding,

number

cases

for ICRp before

observed

particles

and after

situated

within

the pores

were much smaller than those on the external surface. 'CRp It has already been noted that calcination of I to yield

IC provoked

of dispersion

in distribution

above

of the active

for the reduced

upon to explain

complex

I, contained

moved

toward

of I and

to our knowledge, The electron

reduction

been shown

hydrated

where

that the INi:Isi

The value measured

toward

surface

by nickel

on with

the external

Calculations

ratio by X.P.S. for a sample

should

be as a result

sizes were was an order

indicate

that with

similar

be 16.8 for pure nickel nickel

of magnitude

greater prepared

may be explained from considerations 2+ of Ni or OH- can be expected

the micropores

from the pore mouth

inhibition

hydro-

hydrosilicate

diminished

coverage

within

the catalysts

this effect

consumption

partially

solid model

was

further

of the silica

the pores

by comparis-

for G, GC and I, but the Ni:Si

of the support.

The concentration

1)

phase.

surface.

particle

did not enter

2)

in

has never,

were

on the homogeneous

of nickel

ion, nickel

distance

pores

that both samples

based

of synthesised

and a lower concentration

by X.P.S.

findings

been cited

before.

of less than complete

must

this

calcination

of the active

the mouth

for GC, i.e. 8.49, which

grinding,

The

of NiO formed.

has already

16.2 [24]. The lower value measured

as measured

elaborated

nitrate.

During

after

These

noted

have transformed

surface.

with the dispersion

of G and GC suggested

indicated

nickel

must

large crystallites

by its migration

hydrosilicate.

silicate.

The nickel

a loss

IC can be further

calcination

of the precursor

in relation

of nickel

micrographs

by nickel

during

did not wet the silica

but always

distribution

a semidecomposed

the pore mouths

of direct

the literature,

covered

1). The difference

analogs

produced

into a liquid which

The advantage

Changed

and passivated

and atmosphere

this liquid

(Table

this loss of dispersion.

The precursor, temperature

phase

of

of diffusion of reactives

in the former

ratio

two cases.

by deposition-precipitatAs outlined

of the method to decrease

in Figure

of preparation

as a function

6, used.

of the

for 2 reasons: within

the pores

at the pore mouths

as the result

of an auto-catalytic

328

FIGURE

6 Schematic

reduced effect more

IR and GR and calcined whereby

rapidly

nickel

Influence

image

the reaction

reaction

lower specific

on of catalytic preparation

rate should

phase

rates

tests

surface

precipitation

a nucleus

only slightly

area,

important

of Si02 covered

to the metallic

as presented with

of supported samples

by Boudart surface

of Ni/Si02

in Table

the exception

been reported

parameter.

proposed

preparations

rate used during

for the various

is another

proceeds

into the pores.

in the sense

be proportional

have already

sintering

,

activity reaction

for the various

the heating

to which

activities

method

is overcome,

IG and GC

catalysts.

at the pore mouths.

3, show that

of IG. However,

for Ni/SiOp

in the form of large crystallites

previously,

the extent

IGR and GCR, Ni/Si02

rates are very similar

reaction

the active

As mentioned influence

is a facile

of catalytic

I and S, calcined,

G and GC involves enters

on the catalytic

per unit of metallic

the measured

barrier

to deposit

which

of benzene

The results

expressed

and reduced

of the catalysts

of interaction

[25] in which

having

ions tending

hydroxylsilicate

Hydrogenation

area.

of the precursors,

once the nucleation

with

The final with

representation

catalysts

[26]. activation

nickel

studied

can strongly

can occur,

but comparis-

here indicates

Thus some loss

that

in activity

was

329

recorded

for IC after activation

even more

pronounced

sintering

when

changes

the higher

observed

The latter reduction

heating

observations

where

Nickel

to be eventually

by inhibiting manifested

supposes

occurs

by recalling

of

for G and GC any

that the mechanism

the migration

[27]. A similar

implies

through

model

This type of structure

surface

migration

to pretreatment

the platelets

of nickel would

can be postulated

of nickel.

conditions

were

stabilise

activity

networks

were

broken. partially

embedded

G and GC against to sintering

sintering

would

also be

of G and GC being relatively

as was the

case

for

and leave the lattice,

particles

Resistance

of

of iron to the edges of

this structure

at points where the formation

in terms of the catalytic

sensitive

This effect was

of a large amount

In the cases of G and GC the hydrosilicate could migrate

reduced,

This mechanism in the support.

ratio.

rate was used. By comparison,

may be explained

reduction

hydrosilicate.

not perfect.

heating

due to the occurrence

were minimal.

of iron hydrosilicates

the platelets nickel

using the higher

for I, presumably

in Figure

in-

5.

CONCLUSIONS The final in Figure

images of the catalysts

evenly

distributed

nickel

hydrosilicate

studied

here are presented

schematically

I and G, it is clear that whereas

6. For the precursors

in the case of G

in the former,

of turbostratic

texture

interaction

which

covered

nickel was

gave rise to a

the external

SUrfaCe

of the support. Calcination vis-a-vis surface

brought

about

of the support.

upon calcination support

a drastic

I; large particles

Dispersion

which

changes

preparation

occurs

between

high heating

was fairly

evenly

GR and the support

rates were

distributed

No changes

in distribution

almost

of nickel

exclusively

in IC

on the external

in dispersionor distributionoccurred is associated the metal

with the strong metal-

and support

for the deposition-

technique.

of IR was sensitive

occur when

support.

No major

of G and this stability

interaction

precipitation

change

of NiO formed,

used.

between

in dispersion retained

to reduction

conditions;

IR was the only sample

the internal

or distribution

the original

sintering

and external occur

turbostratic

tended

in which surface

upon reduction

texture

to

nickel of the

of G to

of the hydrosilicate

precursor.

ACKNOWLEDGEMENTS We wish ments.

to thank M. Breysse

M.M. thanks

Programmation Lema?tre

UPV/EHU

for carrying

and GV/EJ

de la Politique

Scientifique

for use of the apparatus

out the metallic

for a fellowship. (Belgium)

used for testing

W. Stone for help with the EM and X.P.S.

surface

area measure-

We thank the Service for financial

catalysts

support;

and M. Genet

de J.

and

330 REFERENCES

9 IO 11

12 13

14

1:

1: 19 20 21 22 23 24 25 26 27

D.J.C. Yates, W.F. Taylor and J.H. Sinfelt, J. Amer. Chem. Sot., 86 (1964) 2996. P. Lorenz, J. Finster, G. Wendt, J.V. Salyn, E.K. Zumadilov and V.I. Nefedov, J. Electron Spectrosc. Relat. Phenom., 16 (1979) 267. M. Houalla, F. Delannay, I. Matsuura and B. Delmon, J.C.S. Faraday I, 76 (1980) 2128. M. Houalla and B. Delmon, Surface and Interface Analysis, 3 (1981) 103. G.A. Martin, G. Mirodatos and H. Praliaud, Appl. Catal., 1 (1981) 367. J.W. Geus, Preparation of Catalysts III, (G. Poncelet, P. Grange and P.A. Jacobs, eds.), Elsevier, Amsterdam, (1983) p.1. L.A.M. Hermans and J.W. Geus, Preparation of Catalysts II, (B. Delmon, P. Grange, P.A. Jacobs and G. Poncelet, eds.), Elsevier, Amsterdam, (1983) p.113. J.A. van Dillen, J.W. Geus, L.A.M. Hermans and J. van der Meijden, Proc. 6th Intl. Congr. Catalysts, (G.C. Bond, P.B. Wells, F.C. Tompkins, eds.), The Chemical Society, London (1976), p. 677. J.T. Richardson and R.J. Dubus, J. Catal., 54 (1978) 207. M. Houalla, F. Delannay and B. Delmon, J. Phys. Chem., 85 (1981) 1704. J.T.Richardson, R.J. Dubus, J.G. Crump, P. Desarc, N. Osterwalder and T.S. Cale, Preparation of Catalysts II, (B. Delmon, P. Grange, P.A. Jacobs and G. Poncelet, eds.), Elsevier, Amsterdam (1979) p.131. M. Montes, Thesis, Universite Catholique de Louvain, Louvain-la-Neuve, Belgium, (1984). J.W.E. Coenen and B.G. Linsen, Physical and Chemical Aspects of Adsorbents and Catalysts, (Linsen, Fortuin, Okkerse and Steggerda, eds.), Academic Press, London (19701, p.495. C.D. Wagner, L.E. Davis, M.V. Zeller, J.A. Taylor, R.H. Richmond and L.H. Gale, Surface and Interface Analysis, 3 (1981) 211. C. Defosse, J. Electron Spectrosc. Relat. Phenom., 23 (1981) 157. J. Szajmam, J. Liesengang, J.G. Jenkins and R.C.G. Leckey, J. Electron Spectrosc. Relat. Phenom., 23 (1981) 97. R.Z.C. van Meerten and J.W.E. Coenen, J. Catal., 37 (1975) 37. C.N. Satterfield and T.K. Sherwood, The Role of Diffusion in Catalysis, Addison-Wisley Ph Co., Reading, Massachusetts, (1963) p.31. B. Delmon, Introduction a la cinetique h&terogPne, Ed. Technip, Paris, (1968) p.198. J. Lemaitre in Characterization of Heterogeneous Catalysts, F. Delannay, Ed. Marcel Dekker, New York, 1984, p. 29. R.M. Dell, D.F. Klemperer and F.S. Stone, J. Phys. Chem., 60 (1956) 1586. G. Dalmai-Imelik, C. LeClerc and A.J. Maubert-Muguet, Solid State Chem., 16 (1976) 129. R. Clavet and R. Prost, Clays and Clay Miner., 19 (1971) 175. G.A. Martin, B. Imelik and M. Prettre. C.R. Acad. SC. Paris. S&ie C. 264 (1967)

186.

M. Boudart, Proc. 6th Intl. Congr. Catalysis, (G.C. Bond, P.B. Wells, F.C. Tompkins, eds.) The Chemical Societv. London (1976) 0.1. J.W:E. Coenen, R.Z.C. Van Meerten and H.T. Rijnten; broc. 5th Intl. Congr. Catalysis, (J.W. Hightower, ed.), North-Holland, Amsterdam (1973) p.671. W.E. Addison and J.H. Sharp, Clay Miner. Bull., 5 (1962) 73.