ORIGINAL PAPER

Synthesis of Supported Catalysts by Impregnationand Calcination of Low-Temperature PolymerizableMetal-Complexes

Tiejun Zhao • Sara Boullosa-Eiras •

Yingda Yu • De Chen • Anders Holmen •

Magnus Ronning

Published online: 23 August 2011

� Springer Science+Business Media, LLC 2011

Abstract The low-temperature polymerizable metal

complex solution has been used as an active component

precursor to prepare supported (structured) catalysts by a

straight-forward sequence of impregnation, drying and

calcination. Two typical samples, 5 wt% ZrxCe1-xO2/

Al2O3 nanocomposite and structured carbon nanofiber

supported Cu–CeO2 catalyst, are prepared to explore the

potential of this method in the controlled synthesis of

catalysts and catalyst supports. The interaction between the

active component precursor and the surface of the solid

matrices during impregnation and drying is investigated by

infrared spectroscopy (IR) and transmission electron

microscopy (TEM), demonstrating that the in situ poly-

merization process is crucial for the deposition of the active

component on the surface of solid matrices. The evolution

of the phase transformation and the structure of the pro-

duced materials during the calcination step is studied by

coupling thermal gravimetric analysis–differential thermal

analysis–mass spectroscopy (TGA–DTA–MS), TEM and

X-ray diffraction (XRD) measurements, indicating that

higher thermal stable particles or smaller particles can be

obtained with the presence of the Al2O3 or structured

carbon nanofibers (SCNF), respectively, after calcination.

This method combines the advantages of sol–gel and

impregnation, representing a promising route for preparing

supported catalysts and catalyst supports. The limitations

of the method are also discussed.

Keywords Polymerizable metal complex

catalyst precursors � Cu–CeO2 catalyst �Carbon nanofiber support � ZrO2

1 Introduction

Solid catalysts play an important role in diverse fields such

as energy production, chemical transformation and envi-

ronment abatement. Precise synthesis of the advanced bulk

and supported catalysts is a fundamental base to enhance

their activity, selectivity and stability in a specific process.

Various synthesis methods [1] such as ion exchange,

homogeneous deposition–precipitation, co-precipitation,

sol–gel, chemical vapour deposition, etc. are conducted to

obtain the specific spatial distribution of active components

in the final catalyst, thus obtaining excellent performance.

By far, impregnation [2, 3] is still the most used route for

synthesis of supported metals (metal oxides). The

impregnation method contains two main steps. The first

step is to deposit the active component precursor, mostly

using aqueous solutions, onto a support, and the second

step is to transform this precursor into the final active

phase, which often involves drying and heat treatment

(calcination or reduction). Although the execution of the

impregnation is simple, the complexity of the interaction

between the metal precursor and the support, the solvent

evaporation or the phase transformation during the prepa-

ration (i.e. impregnation, drying and calcination) may lead

to a poor dispersion of the active species in the final cat-

alyst [4]. When several active components are required in

the final catalyst, sequenced or co-impregnation is

T. Zhao � S. Boullosa-Eiras � D. Chen � A. Holmen �M. Ronning (&)

Department of Chemical Engineering, Norwegian University

of Science and Technology (NTNU), 7491 Trondheim, Norway

e-mail: ronning@chemeng.ntnu.no

Y. Yu

Department of Material Science and Technology, Norwegian

University of Science and Technology (NTNU),

7491 Trondheim, Norway

123

Top Catal (2011) 54:1163–1174

DOI 10.1007/s11244-011-9738-2

performed. In some cases, this more complicated procedure

will lead to undesired distribution of active components in

the catalyst due to the different interaction between the

precursor and support in the preparation procedures.

Forming a stable active component complex with a suitable

precursor–support interaction is a promising strategy to

avoid the drawbacks of the impregnation method.

The polymerizable metal complex method [5–7], often

referred to as the Pechini route, is an elegant variation of

the sol–gel method to prepare multi-compound oxides

(nanocomposites) with outstanding homogeneity. This

method often contains two steps: first, formation of the

three-dimensional polymer resin of a metal-complex; sec-

ond, calcination of this gel-like solid at elevated tempera-

tures to obtain the oxides. The key step of the Pechini

approach is the in situ polymerization between citric acid

(CA) and ethylene glycol (EG) or poly-ethylene glycol

(PEG). It has been found that the chemical interaction

between the CA and EG in the solution already occurs to

form the esters at room temperature without any treatment

as shown by the 13C NMR spectrum [8]. The esterification

process between CA and EG can be catalyzed by acidic or

basic media [9]. This reaction can also shift towards the

polyester by increasing the reagent concentration and

continuously removing water, leading to an increase in the

viscosity of the solution. The strongly chelating ability of

CA with most metal ions often provides the formation of a

metal citrate complex during the above polymerization

process. Due to its simplicity and ease of controlling, this

polymeric metal complex method has been extensively

used to synthesize ferroelectric [10, 11], superconducting

[12], fuel cell [13], electrode [14, 15], catalytic and other

types of materials [16–18]. Concerning the applications of

this method in heterogeneous catalysis, most reports focus

on the preparation of bulk oxides or the formation of cat-

alytic films [19–24]. However, this method has obvious

disadvantages such as violent reactions during the calci-

nation, extensive use of organics (CA and EG), difficulties

for controlling the particle size, etc. [7, 25]. These disad-

vantages might be circumvented if solid matrices with high

surface area are employed to disperse the polymeric pre-

cursor during the preparation. With this methodology, the

amount of organics, such as CA and EG, can be signifi-

cantly reduced if large fractions of the substrate are used

during the synthesis. Thus, the violent reactions taking

place during the calcination may also be prevented.

In addition, only adding the CA to the metal precursor

solution is effective to obtain highly dispersed active par-

ticles in the final catalysts. For instance, higher thermal

stability of CeO2 is observed in the CeO2/Al2O3 system

prepared by the CA mediated impregnation method [26]. It

has been reported that the addition of EG or PEG will

further increase the homogeneity in the multi-component

system [9]. Recently, a series of ZrxCe1-xO2/Al2O3 nano-

composites have been prepared by impregnation of these

metal–CA–PEG complexes. High thermal stability of these

composites is illustrated by X-ray diffraction (XRD) and

Raman spectroscopy, and an oxide-coated nanostructure of

these composites is found by combining visible and UV

Raman spectroscopy [27].

In this contribution, we investigate the evolution of the

solid during impregnation, drying and calcinations of the

ZrxCe1-xO2 (x = 0, 0.5, 0.75, 1)/Al2O3 nanocomposites

(mainly focusing on 5 wt% CeO2/Al2O3 here). In addition,

this method is extended to prepare a structured catalyst

using structured carbon nanofiber (SCNF) supported

Cu–CeO2 as an example. For the ZrxCe1-xO2(x = 0, 0.5,

0.75, 1)/Al2O3 nanocomposites (as a promising catalyst

support for catalytic partial oxidation of methane [28–30]),

an excessive solvent impregnation (ESI) method is used;

For the SCNFs supported Cu–CeO2 catalysts (as a prom-

ising catalyst for preferential oxidation of CO in hydrogen-

rich gases [31, 32]), incipient wetness impregnation is

used. IR, TEM, XRD, TGA-DTA-MS experiments are

performed to investigate the chemical reaction and struc-

ture evolution during the impregnation, drying and calci-

nation procedures, in order to understand the interaction

between the active components and the surface of the solid

matrices. The limitations in applying the impregnation of

polymeric metal complex solutions are also discussed.

2 Experimental

2.1 Al2O3 Based Nanocomposites

Sasol Puralox SCCa Al2O3 (surface area about 170 m2/g,

pore volume about 0.5 cm3/mL), is used as a solid matrix

for the preparation of the nanocomposite. An ESI method is

used to produce the CeO2–Al2O3 nanocomposites. Typi-

cally, a certain amount of Ce(NO3)3�6H2O and CA

(C6H8O7) with the molecular ratio of 1:2 are dissolved into

a certain amount of deionized water to form a complex

solution, followed by adding PEG (0.1 g/mL). Afterwards,

the weighed alumina powder is gradually added into the

complex solution under vigorous stirring to produce a

suspension. This suspension is subsequently evaporated at

358 K for 2 days to form a dried solid. A similar procedure

is carried out to prepare the multi-component complex

solution, 5 wt% ZrxCe1-xO2/Al2O3 (x = 0, 0.25, 0.5,

0.75.1), using ZrO(NO3)2�xH2O, Ce(NO3)3�6H2O, CA

(C6H8O7). More details can be found elsewhere [27, 33].

Nanocomposites are produced after these dried solids are

calcined in air at 1,173 K for 5 h or 1,373 K for 24 h.

1164 Top Catal (2011) 54:1163–1174

123

2.2 SCNFs Supported Cu–CeO2 Catalysts

SCNFs are composed of the carbon felt (CF) matrix and

carbon nanofibers. Carbon nanofibers can be catalytically

grown on the interwoven macro carbon fiber in the CF disk

(1 m2/g) supported 2 wt% Ni with a thickness of 6 mm and

a diameter of 25 mm. After reducing the Ni/CF catalyst,

the SCNFs can be obtained by decomposition of a C2H6/H2

gas mixture at 923 K for 16 h. Details of the synthesis of

this structured material can be found elsewhere [34, 35].

The formed SCNFs solids are treated in the HNO3 solution

at 373 K for 3 h to introduce the oxygen-containing groups

on the surface of the carbon nanofibers, before being

washed and dried overnight at 393 K.

The Cu–CeO2/SCNF catalysts are prepared by the

impregnation method. A complex water solution con-

taining Cu(NO3)2�3H2O, Ce(NO3)3�6H2O, CA and EG

(the molecular ratio between Cu:Ce:CA:EG is 1:7:8:8) is

immediately impregnated on the SCNFs after forming the

complex solution (by ultrasonic treatment) until the liquid

appears on the surface of the SCNF. Afterwards, excess

liquids in the structured materials are purged in com-

pressed air. Then, the sample is dried at room temperature

for 16 h. A repeated impregnation procedure is carried out

after drying to increase the loading of Cu–CeO2 in the

final structured catalyst. Calcination of the dried solid is

performed at 673 K for 10 min with a ramp rate of 5 K/

min in 10% O2/N2 flow (100 mL/min). To investigate the

effect of the polymerization process at room temperature,

the aqueous complex solution is kept at room temperature

for 16 h, before a similar impregnation procedure is car-

ried out. For comparison, the Cu–CeO2 composite is

obtained by drying the Cu–Ce–CA–EG complex solution

at 353 K overnight, followed by calcination at 673 K for

10 min. The preparation procedure for the structured

catalyst and the nanocomposite support is illustrated in

Scheme 1.

2.3 Characterization of the Nanocomposites

and Structured Catalysts

Powder XRD measurements were performed in a Bruker

AXS D8 Focus (40 kV, 50 mA) X-ray diffractometer with

Cu Ka radiation (1.541 A) at room temperature from 20 to

80�. TGA–MS analysis was carried out on a TA Instru-

ments Q50 TGA coupled with a Pfeiffer Vacuum Ther-

moStar mass spectrometer. The evolved gases are fed into

the mass spectrometer via a capillary tube kept at 393 K.

The flow of the sample gas is 90 mL/min. The TGA–DTA

profiles are obtained on a Netzsch STA-449 C instrument

in air flow (90 mL/min) from 300 K to high temperature

(10 K/min). A FTIR study of the solid was performed

on a Bruker Tensor 27 instrument with 400–4,000 cm-1 at

room temperature. Raman spectra were collected on a

Horiba Jobin-Yvon LabRAM HR800 spectrometer. The

emission lines at 633 nm from a He–Ne laser are focused

on the sample with a 1009 objective and 325 nm from

He–Cd are focused on the sample with a 40NUV objective.

The morphology of the solids were examined by HRTEM

(JOEL JEM 2010). The powdered samples were suspended

in absolute ethanol for 2 min with ultrasonication, and a

droplet of this suspension is dipped on to a holey carbon-

coated copper grid. For the structured catalyst, the sample

was initially ground into small pieces, before treated as the

powder samples.

3 Results and Discussion

3.1 Impregnation and Drying of the Metal-Complex

Solution on the Al2O3 and SCNF

Typically, 5 wt% CeO2/Al2O3 is prepared by the ESI

method using a powdered Al2O3 suspension containing the

Ce–CA–PEG complex solution as the precursor through an

Al2O3

CalcinationSupported

catalyst

Citric acid

PEG

COOH

COOH

COOH

HO

HO OH

CNF

-OH

-OH

-OH

-OH

-Al 3+

-Al3+

-Al3+

-OH-OH

-COOH

-COOH

-CO

OH

-CO

OH

-OH

-OH

Polymeric metal complex

Impregnation

Zr4+ Ce4+

Drying

Coatednanocomposites

Ce3+, ZrO2+

Al2O3 or CNFHO OH

Scheme 1 Dried CeO2/Al2O3

preparation procedure: the

Ce–CA–PEG impregnated

Al2O3 is prepared by

impregnation of the

Ce(NO3)3�6H2O, CA and EG

complex solution in Al2O3, and

then dried at 358 K for 2 days.

Dried Cu-CeO2/SCNF

preparation procedure: the

Cu-Ce-CA-EG complex

solution is immediately

impregnated on to the SCNF,

and dried at room temperature

for 16 hours

Top Catal (2011) 54:1163–1174 1165

123

evaporation–drying procedure. An IR study of the CA,

Al2O3, dried 5 wt% CeO2/Al2O3 sample, respectively, is

performed at room temperature to follow the evolution of

the structure of the solid during the preparation. Figure 1

gives the IR spectra of the solids between 1,000 and

2,000 cm-1. In the CA sample, peaks in the range of 1,150

and 1,210 cm-1 are ascribed to the -OH groups, whereas a

broad peak at 1,700 cm-1 reflects the saturated carboxylic

groups. In the dried CeO2/Al2O3 sample, the two broad

peaks appearing at 1,120 and 1,310 cm-1, respectively, are

ascribed to the asymmetric stretching vibration of =C–O–C

and C–O–C groups [36], clearly indicating the formation of

an ester due to the reaction between the CA and PEG. A

larger amount of ester groups can be expected if the EG is

used instead of PEG [37]. In addition, the peak at

1,600 cm-1 in dried CeO2/Al2O3 is related to the formation

of carboxylate salts, which indicates that Ce and Al ions

react with CA to form the complex in the dried sample. It

has been found that the CA can bond with the surface Al3?

to form Al citrate [38].Therefore, after the evaporation–

drying procedure, the metal ions probably combine with

the carboxylic groups in the CA to form the metal citrate

salt, instead of the metal nitrate.

The morphology of the dried Ce–CA–PEG impregnated

Al2O3 by the ESI method is shown in Fig. 2a, b. From

Fig. 2a, a dense nanorod-like morphology of Al2O3 can be

observed, whereas an amorphous phase appears on the

edge zones of this agglomerated particle as shown by the

arrows. Figure 2b shows a HRTEM image of the specific

zone in Fig. 2a, where the crystalline structure of Al2O3

and the amorphous phase in contact with it can be seen.

Figure 2c presents a typical image of powdered Al2O3

support without impregnated metal complexes, where both

the crystalline structure and the amorphous phase can also

be observed with a similar morphology as shown in

Fig. 2a. Therefore, the amorphous phase on the crystalline

structures in the solids may be related to both the amor-

phous phase on the surface of Al2O3 itself and the in situ

produced polymer. Alternatively, a well-crystallized cata-

lyst support may give a better evidence of the interaction

between the metal–CA complex solution and the surface of

the catalyst support.

SCNFs are used as a typical sample to investigate the

interaction between the metal-complex solution and

the surface of the carbon nanofibers, and to synthesize the

Cu–CeO2/SCNF catalyst. Carbon nanofibers are predomi-

nantly graphitic carbon. The SCNFs can be produced on

the CF supported Ni catalyst by decomposition of C2H6/H2

mixed gases at 923–973 K. The morphology of this SCNF

is shown in Fig. 3a, b, where the porous CNF layers are

uniformly covering the surface of the macro carbon fibers

in the CF. The HRTEM image of the carbon nanofibers

collected from the SCNFs in Fig. 3c shows that the

stacking of the graphene sheets are fishbone-like. It is

observed that the as-grown carbon nanofibers are highly

hydrophobic, and that the aqueous solution cannot effec-

tively disperse on the surface of the SCNF [39]. However,

the concentrated HNO3 treatment at 383 K for 2 h can

significantly enhance the hydrophilicity of the SCNF by

introduction of the oxygen-containing groups on the sur-

face [40]. Therefore, the aqueous solution containing

copper nitrate, cerium nitrate, CA and EG can be impreg-

nated on to the hydrophilic SCNF. The morphology of the

dried Cu–Ce–CA–EG impregnated solid examined by

TEM is shown in Fig. 4a, b. It can be seen that an amor-

phous layer is uniformly coating the surface of carbon

nanofibers in Fig. 4a. The tip of the carbon nanofibers is

covered by the amorphous layer as observed in Fig. 4b,

which also reflects the removal of the active metal particles

located on the tip of the fishbone carbon nanofiber by

oxidation in HNO3. In this case, immediate impregnation

of the metal complex solution to the SCNF is achieved.

However, when this metal–CA–EG complex solution is

kept at room temperature for 16 h, and then impregnated

on the SCNF, the morphology of the dried SCNF

impregnate shown in Fig. 5a, b. Figure 5a, gives an

amorphous gel indicated by the arrows, which is not coated

on the surface of the carbon nanofibers. The surface of the

carbon nanofibers is clean as shown in Fig. 5b. It is well

known that the carboxylic or hydroxyl groups can be

introduced on the surface of the carbon nanofibers by

oxidation in concentrated HNO3. These groups can interact

with the hydroxyl groups in CA or EG, and carboxylic

groups in CA, respectively. If the esterification reaction

between the CA and EG occurs, the amount of carboxylic

0.0

2.5

1700

13101120

1210

Dried CeO2/Al

2O

3

Al2O

3

CA1150

1610

1000 1200 1400 1600 1800 2000

Wavenumber, cm-1-1

Fig. 1 IR spectra in the range of 1,000–2,000 cm-1 of CA, Al2O3

and dried CeO2/Al2O3 samples

1166 Top Catal (2011) 54:1163–1174

123

groups and hydroxyl groups will be consumed. It is noticed

that the deprotonated H? from CA can recombine with

NO3- from the metal precursor to increase the acidity in

the complex solution (the pH is less than one at room

temperature for this complex solution), thus promoting the

esterification reaction. The lower amount of these active

groups in the complex solution will prevent the interaction

between the metal complex and the surface of carbon

nanofibers. Therefore, the immediate impregnation of the

metal complex solution on to the catalyst support is needed

to promote the occurrence of the in situ polymerization.

However, in the CeO2/Al2O3 case, PEG is used instead of

EG, thus significantly reducing the consumption of car-

boxylic groups in CA. Free carboxylic groups can still react

with the surface Al3? or hydroxyl groups in Al2O3. The

polymerization is often accompanied by an increase of

the viscosity in the complex, which may contribute to the

anchoring of the active components to the surface of the

catalyst support [3, 4].

IR-sensitive Al2O3 is used as a solid matrix to study the

interaction between the metal complex and the surface of

Al2O3 in the dried CeO2/Al2O3. The well-crystalline car-

bon nanofibers as the substrate are used to directly observe

the polymerization process on the surface of carbon

nanofibers.

3.2 Calcination of the Dried Metal-Complex

Impregnated SCNF and Al2O3

3.2.1 TGA–DTA–MS Study for CeO2–Al2O3

and Cu–CeO2/SCNF

After the impregnates are dried at room temperature for

Cu–CeO2/SCNF or at 363 K for CeO2–Al2O3 nanocom-

posites, the solids are treated at high temperature to obtain

the corresponding oxides or oxide supported catalysts. The

evolution of this heat treatment by TGA–DTA–MS gives

information on the chemical transformation of the

Fig. 2 Morphology of the dried Ce–CA–PEG impregnated Al2O3

samples (a, b) and Al2O3 sample (c). The Ce–CA–PEG impregnated

Al2O3 is prepared by impregnation of the Ce(NO3)3�6H2O, CA and

EG complex solution in Al2O3, and then dried at 358 K for 2 days.

The Al2O3 sample is calcined at 773 K for 3 h to remove the residues

Top Catal (2011) 54:1163–1174 1167

123

impregnated precursors, whereas the XRD and TEM

studies give structural information on the calcined solid.

Figure 6a shows the TGA–MS profiles for the 5 wt%

CeO2/Al2O3 sample in air flow from 300 to 873 K. From the

TGA profile, the weight loss mainly occurs at temperatures

less than 700 K. In the MS signal, the water release starts

from 300 K, gradually increasing with temperature, whereas

only a peak at 550 K appears for the NO profile. The NO

signal is responsible for the nitrate in the dried composites. A

more complicated pattern for released CO2 is found: No CO2

is released at temperatures below 475 K, and afterwards, the

signal intensity of CO2 increases with further increasing

temperature, where two peaks can be identified at 550 and

620 K. Simultaneous release of NO, H2O and CO2 at 550 K

could be related to the reaction between the organics, nitrate

and oxygen. The weight loss of the solid in the range of

Fig. 3 Morphology of the structured carbon nanofibers: a blobal morphology, b porous carbon nanofiber layer, c graphene sheet orientation.

Synthesis conditions catalyst, 2 wt% Ni/carbon felt; gas mixture, C2H6/H2 = 90/150 mL/min; temperature, 923 K; time, 16 h

Fig. 4 Morphology of dried Cu–Ce–CA–EG impregnated SCNF:

a polymeric metal complex solid are uniform coated on the surface of

carbon nanofibers by fast impregnation of the Cu–Ce–CA–EG metal

complex solution; b detailed morphology of the polymeric solid

coated on the tip of carbon nanofibers

1168 Top Catal (2011) 54:1163–1174

123

550–700 K can be due to the decomposition of polymeric

organics in the composites. The released CO2 and H2O at

further higher temperatures could be ascribed to the burn-off

of the polymeric residues, since no further obvious weight

loss takes place in this period. Another dried CeO2–Al2O3

impregnated sample is studied in a TGA–DTA instrument

from room temperature to 1,800 K, showing two heat

release peaks at 550 and 620 K as shown in Fig. 6b. This

observation is consistent with the previous TGA–MS pro-

file. In addition, two broad peaks in the DTA signal appear at

temperatures between 1,200 and 1,800 K due to the phase

transformation of Al2O3, as reported elsewhere [27]. In that

report, a higher temperature of the phase transformation to a-

Al2O3 for Al2O3 nanocomposite was observed when the CA

and PEG were added during the preparation. Introducing the

CeO2 on the surface of Al2O3 will lead to a higher energy

barrier for the phase transformation to a-Al2O3.

For comparison, the Cu–CeO2 composite prepared by

drying the complex solution at 358 K overnight is also

treated with the same procedure. In this case, the dominant

compound in the final solid is CeO2 (95 wt%). Therefore, we

can expect that this solid can be treated as the pure CeO2 for

Fig. 5 Morphology of the polymerized metal complex impregnated SCNF: a dried SCNF impregnate where the metal CA–EG complex solution

is kept at room temperature overnight, b no gel-like solid can be observed on the surface of the carbon nanofibers

H2O

NO

CO2

70

80

90

100 Mass of solid, w

t %

[1]

[1]

0

10

20

30

40

50

DT

A,uW

/mg

NO, 30

H2O, 18

CO2, 44

30

40

50

60

70

80

90

100

110

Mass of solid, w

t%

300 400 500 600 700 800 900

1E-14

1E-13

1E-12

Temperature, K

MS

sign

al i

nten

sity

, a.u

.

a

400 600 800 1000 1200 1400 1600 1800

Temperature /K

0.0

0.2

0.4

0.6

0.8

1.0

1.2

DTA /(uV/mg)

65

70

75

80

85

90

95

100TG /%

↑ exo

b

c d

400 600 800 100030

40

50

60

70

80

90

100

Temperature, K

Mas

s pe

rcen

t of

solid

,%

350 400 450 500 550 600 6501E-15

1E-14

1E-13

1E-12

1E-11

1E-10

Temperature, K

Inte

nsity

of

MS

sign

al, a

.u.

Fig. 6 TGA–DTA–MS profiles

for the CeO2/Al2O3, Cu–CeO2

and Cu–CeO2/SCNF: a TGA–

MS profiles for the dried 5 wt%

CeO2/Al2O3 sample; operating

conditions, ramp rate 10 K/min,

air flow, 90 mL/min; sample

weight, 14.8 mg; 300–873 K;

b TGA–DTA profiles for the

dried 5 wt% CeO2/Al2O3

nanocomposite; operating

conditions, ramp rate 10 K/min;

air flow, 90 mL/min; sample

weight, 15.8 mg; 300–1,800 K;

c TGA–DTA profiles for the

dried Cu–CeO2 nanocomposite;

operating conditions, ramp rate

10 K/min; air flow, 90 mL/min;

sample weight, 15.3 mg;

300–1,000 K; d TGA–MS

profiles for the dried Cu–CeO2/

SCNF; operating conditions,

ramp rate 10 K/min; air flow,

90 mL/min; sample weight,

15.8 mg; 300–670 K

Top Catal (2011) 54:1163–1174 1169

123

qualitative analysis. Note that during the preparation of this

dried solid, only EG instead of PEG was used. Due to the

absence of solid matrices such as Al2O3, organics and

nitrates are dominating in this dried Cu–CeO2 solid. The

TGA profiles in Fig. 6c present a sharp weight loss at around

550 K, similar to the TGA profiles in Fig. 6a. This sharp

weight loss is accompanied by a large heat release shown in

the DTA profiles for this Cu–CeO2 nanocomposite. Distur-

bance of the temperature profile is also observed in this

period. The violent reaction could be due to the reaction

between the nitrate, oxygen and organics in the dried solid. It

is observed there is no obvious weight loss and heat release

after this sudden weight loss-heat release period even when

the temperature is increased up to 1,000 K. This observation

is different from the situation with the 5 wt% CeO2/Al2O3. In

the latter, another heat release peak appears at 620 K in

Fig. 6a. This can be due to the following reasons: The

organics formed by EG and CA are easily burned off due to

the higher porosity of the formed polymer compared to the

one with PEG. The violent reaction promotes the complete

combustion of the organics in the Cu–CeO2 nanocomposite.

The dried Cu–CeO2/SCNF was heat treated in air in a

TGA–MS coupled instrument to investigate the evolution

of the polymerizable metal complex deposited on the sur-

face of carbon nanofibers. The results are shown in Fig. 6d

for TGA–MS profiles when the dried impregnated CNF

sample is heated from 318 to 650 K. The weight loss of the

dried sample occurs below 373 K, probably due to the

release of adsorbed water as evidenced by the intensity

increase of the MS signal of water. Subsequently, more

water, NO and CO2 are simultaneously evolved, reaching

the maximum at a temperature around 373 K, accompanied

by a sharp weight loss of the solid sample, which may be

due to oxidation of the organics by NO3-. This observation

does not appear in the heat treatment of 5 wt% CeO2/Al2O3

samples. Higher temperature (358 K) during the prepara-

tion of 5 wt% CeO2/Al2O3 samples can lead to oxidation of

organics by the nitrate [41]. Subsequently, a major weight

loss takes place at around 550 K, also accompanied by the

release of NO, CO2 and H2O as shown in the MS signal.

Note that the amount of CO2 in the MS signal is still

continuously increasing. This observation can be due to the

combustion of the organics by oxygen or nitrate, similar to

the observation for the 5 wt% CeO2/Al2O3 sample. It is

also observed that the temperature profile is disturbed in

this period due to the strong exothermal oxidation reaction

during the calcination. Increasing the temperature further

does not lead to an obvious weight loss. However, small

peaks corresponding to CO2 and H2O in the evolved gases

around 640 K can still be observed probably due to the

remaining polymeric species. The origin of these evolved

gases can not be determined here since the decomposition

of oxygen-containing groups on the surface of carbon

nanofibers can not be excluded at high temperatures. Car-

bon nanofibers are stable at temperatures below 673 K, but

increasing the temperature in air will lead to combustion

the carbon nanofibers. Only 32.8 wt% of the solid remains

when the sample is treated in air at a temperature close to

673 K after the removal of water, organics (polymers) and

nitrate. The loading of Cu–CeO2 in this catalyst is about 20

wt%, which can be determined by oxidising all the carbon

in the catalyst in air in the TGA (data now shown here). In

summary, calcination of the dried impregnate in air will

lead to a highly exothermal reaction between the organics,

nitrate and oxygen taking place at 550 K. Combustion of

the organics leads to formation of oxides in the calcined

solid.

3.2.2 TEM, IR, XRD and Raman Study

Figures 7a–d present the morphology of 5 wt% CeO2/

Al2O3 by calcination at 1,173 K for 5 h and 1,373 K for

24 h. Compared to the morphology of the dried 5 wt%

CeO2/Al2O3 in Fig. 1c, no obvious particle growth is

observed after calcination at 1,173 K for 5 h. A HRTEM

image of this calcined composite is presented in Fig. 7b,

where the different nanophases of less than 10 nm are

stacked together as shown by the boundary grain between

the particles. The amorphous phase in the sample can still

be observed in the specific region indicated by the arrows.

Furthermore, when the calcination is carried out at 1,373 K

for 24 h, the particle size of the composite is similar to the

sample calcined at 1,173 K as shown in Fig. 7c. However,

from the HRTEM image in Fig. 7d, almost all of the

amorphous phase disappears and a faceted crystalline

structure appears.

The IR spectra between 600 and 1,000 cm-1 for the

calcined CeO2/Al2O3 samples is shown in Fig. 8a. The IR

spectra below 1,000 cm-1 usually reflect the vibration of

Al–O bond as presented by Morterra et al. [42]. The range

600–750 cm-1 gives the stretching vibration of Al–O for

the octahedrally coordination of Al ions, while the range

750–900 cm-1 gives the Al–O vibration in tetrahedrally

coordinated Al ions [43]. From Fig. 8a, a broad asym-

metric band between 600 and 620 cm-1 (indicated by the

band around 610 and 617 cm-1, respectively) and a sharp

band at 668 cm-1 appears in the 600–750 cm-1 for these

two composites. Furthermore, two broad bands at 760 and

826 cm-1 appear for these two calcined nanocomposite

samples as indicated in the inserted in Fig. 8a. Sharper

bands are observed in the range between 600 and 700 cm-1

in the CeO2/Al2O3 calcined at 1,373 K for 24 h compared

with the one calcined at 1,173 K for 5 h, whereas the later

has the sharper bands between 750 and 900 cm-1. It is

known that the thermodynamically stable a-Al2O3 only

has octahedrally coordinated Al ions, while both the

1170 Top Catal (2011) 54:1163–1174

123

tetrahedrally and octahedrally coordinated Al ions appear

in the transition Al2O3 [26]. Therefore, the decrease of the

bands in the range of 750–900 cm-1 reflects the formed

solid with higher crystallinity. The fact that these peaks

(750–900 cm-1) can still be clearly observed implies that

the formation of a-Al2O3 is limited.

The XRD profiles for the 5 wt% ZrxCe1-xO2/Al2O3

calcined at 1,373 K for 24 h are shown in Fig. 8b. For

comparison, the profile for the Al2O3 powder calcined

under the same conditions is also shown. It is found that

the c-phase has been completely transformed into a-phase

in the powdered Al2O3 sample after this heat treatment.

However, the introduction of 5 wt% ZrxCe1-xO2 by the

ESI method can significantly prevent the formation of the

thermodynamically stable a-phase in Al2O3-based nano-

composites: only small peaks corresponding to a-Al2O3

phase are identified. The dominant phase in these nano-

composites is transition Al2O3. This is consistent with the

IR results in Fig. 8a. It is also found that the deposited

ZrxCe1-xO2 on the Al2O3 has higher thermal stability

compared with bulk ZrxCe1-xO2 oxides, indicating a

synergism stabilization mechanism [26, 27]. The forma-

tion of a-Al2O3 in the catalytic process is often accom-

panied by a decrease of surface area, and sometimes,

encapsulation of the active phase. The delay of the Al2O3

phase transformation can be achieved by lowering the

surface energy or by introducing the highly dispersed

oxides on the Al2O3 surface. Therefore, the formation of a

stable and dispersed phase on the Al2O3 surface is a

necessary step to prepare such thermally stable nano-

composites. Polymeric metal complexes deposited on the

Al2O3 surface seem to be effective for obtaining thermally

stable nanocomposites.

Li et al. [37] have pointed out that changing the

wavelength of the optical detector can influence the

detection depth of materials, thus giving the possibility to

obtain the bulk structure of materials using a visible light

detector and the surface region of materials using a UV

Fig. 7 Morphology of 5 wt% CeO2/Al2O3 calcined at different temperatures for 5 or 24 h: a, b calcined at 1,173 K for 5 h; c, d calcined at

1,373 K for 24 h. From HRTEM images in (b, d), a more crystalline structure appears on the sample calcined at 1,373 K for 24 h

Top Catal (2011) 54:1163–1174 1171

123

detector. Near UV and visible Raman spectra for calcined

samples are shown in Fig. 8c. The UV Raman spectrum of

Al2O3 sample calcined at 1,373 K for 24 h in Fig. 8c

clearly shows strong characteristic peaks (green) for the

a-phase (378.6, 416.2, 579.0, 646.3, 752.7 cm-1), whereas

only the peak (blue) at 463.1 cm-1 appears on the 5 wt%

CeO2/Al2O3 calcined at 1,373 K for 24 h. This peak can

be ascribed to the CeO2. From XRD results shown in

Fig. 8b, small amount of a-Al2O3 can be observed.

Therefore, completely disappearance of peaks responsible

for a-Al2O3 in 5 wt% CeO2/Al2O3 can not be due to the

absence of a-Al2O3. UV Raman is highly sensitive to the

presence of a-Al2O3 [27]. On the other hand, visible

Raman spectrum for 5 wt% CeO2/Al2O3 is also shown in

Fig. 8c, where a red trace is responsible for CeO2

(463.3 cm-1) in the range 200–700 cm-1 and an blue

trace shows Al2O3 in the range 1,000–1,900 cm-1. The

peaks at 1175.3 and 1242.6 cm-1 in this visible Raman

spectra can be due to the h-Al2O3, while the peaks at

1367.8 and 1399.1 cm-1 are related to a-Al2O3. From this

observation, it can be concluded that the Al2O3 in the

calcined 5 wt% CeO2/Al2O3 is most likely covered by a

layer of CeO2, consistent with the previous report focusing

on 20 wt% ZrxCe1-xO2/Al2O3 [27].

Higher stability of these nanocomposites gives a higher

stability of the 0.5 wt% Rh/5 wt% ZrxCe1-xO2/Al2O3

catalyst in partial oxidation of methane to synthesis gas.

More details about the structure and the phase transfor-

mation have been reported elsewhere [27, 33].

The TEM examination results of structured Cu–CeO2/

CNF after calcined at 673 K for 10 min are shown in

Fig. 9a, b. Small particles are deposited on the surface of

carbon nanofibers as shown in Fig. 6a for the sample with a

loading around 20 wt%. Some agglomerated particles can

be observed in this sample due to the high loading of

Cu–CeO2 in this structured catalyst. Agglomerated crys-

talline particles of 3–5 nm are clearly seen in Fig. 9b. For

comparison, HRTEM pictures of Cu–CeO2 nanocompos-

ites prepared by the ESI procedure without CNF matrices

are presented in Fig. 9c, d. From Fig. 9c, the rod-like

nanostructures of Cu–CeO2 composites are produced by

drying the complex solution. The distribution of CuO in the

Cu–CeO2 nanocomposites and the SCNF supported cata-

lyst is not identified at this stage. A HRTEM image of the

rod-like nanostructures is shown in Fig. 9d, and it is found

that this rod-like structures are composed of the nano-

crystalline Cu–CeO2 with the size around 10 nm. The

presence of carbon nanofibers can effectively promote the

600 650 700 750 800 850 900 950 1000

4

750 800 850 9004.0

4.2

4.4

4.6

4.8

5.0

5.2826760

Wavenumber, cm-1

CeO2/Al

2O

3-1173

617610

668

Wavenumber, cm-1

617

CeO2/Al

2O

3-1373

200 400 600 800 1000 1200 1400 1600 1800

463

.1

579

.0 378

.6 4

16.2

646

.3

752

.7

466

.3

139

9.1

117

5.3

124

2.6

136

7.8

a b

c

Fig. 8 a IR spectra of the

calcined CeO2/Al2O3 samples:

the dried CeO2/Al2O3 are

calcined at 1,173 K for 5 h and

1,373 K for 24 h, respectively.

b XRD patterns of the 5 wt%

ZrxCe1-xO2/Al2O3 and Al2O3

calcined at 1,373 K for 24 h.

a, Al2O3; b, x = 0; c, x = 0.25;

d, x = 0. 5; e, x = 1. Samples

are prepared by the Ce and

Zr–CA–EG. Data collected at

0.02�/step in the range 2h =

20–70� (d a-Al2O3, v d-Al2O3,

u t-ZrO2, c-CeO2).

c UV–Vis Raman spectroscopy

for the calcined samples: greenUV Raman spectrum for Al2O3;

blue UV Raman spectrum for 5

wt% CeO2/Al2O3; red visible

Raman spectrum for CeO2 in 5

wt% CeO2/Al2O3

(200–700 cm-1); light bluevisible Raman spectrum for

Al2O3 wt% CeO2/Al2O3

(1,000–1,900 cm-1)

1172 Top Catal (2011) 54:1163–1174

123

formation of smaller Cu–CeO2 with smaller particles. The

XRD patterns of the composites and Cu–CeO2/SCNF are

shown in Fig. 10. Broad peaks for CeO2 in Cu–CeO2/

SCNF are identified, whereas relatively sharp peaks of

CeO2 are detected in the Cu–CeO2 nanocomposite. This

observation is consistent with the TEM examination. The

CuO phase is amorphous to XRD in both the SNCF sup-

ported and nanocomposite samples.

4 Conclusions and Remarks

The nanocomposite and structured catalysts can be pre-

pared by impregnation with a low-temperature polymeric

precursor, followed by calcination to remove the polymeric

solid on the support by the reaction between the organic

material, oxygen and nitrate. The alumina-based nano-

composite was prepared through an ESI method using a Ce

Fig. 9 Morphology of Cu–CeO2/SCNF (a, b) and Cu–CeO2 nano-

composites calcined at 673 K for 10 min in 10% O2/N2 (c, d):

a global morphology, showing the agglomerated particles on the

surface of the carbon nanofibers, b Cu–CeO2 nanoparticles (3–5 nm)

is deposited on the surface of the carbon nanofibers, c rod-like

morphology of Cu–CeO2 and d Cu–CeO2 nanoparticles

20 30 40 50 60 70 802

4 CNF/CF

Inte

nsity

, a.u

.

2 Theta

CeO2

Fig. 10 XRD patterns of Cu–CeO2/SCNF (a) and Cu–CeO2 nano-

composites (b) calcined at 673 K for 10 min in the 10% O2/N2

Top Catal (2011) 54:1163–1174 1173

123

and Zr–CA–PEG solution, followed by calcination at high

temperatures. TEM and XRD studies confirmed that the

formed Al2O3-based composites represent improved ther-

mal stability, whereas the Raman spectroscopy indicates

that a coating layer of ceria is formed on the alumina

surface in the composites. In addition, SCNF supported

Cu–CeO2 is successfully prepared by rapid impregnation of

the Cu–Ce–CA–EG complex solution on to the SCNF. The

in situ polymerizable precursor solution can be effectively

deposited on the surface of carbon nanofibers, and highly

dispersed Cu–CeO2 nanoparticles are achieved by calci-

nation at 673 K for 10 min. The TGA–DTA–MS study of

these composites indicates that a highly exothermal reac-

tion between the organics, nitrate and oxygen takes place at

about 550 K during the calcination, promoting the forma-

tion of dispersed nanoparticles. Due to the pronounced

chelating ability with metal ions and the high solubility of

CA in the aqueous solution, the active component precur-

sor is easy to prepare, especially for multi-component

systems. However, the addition of EG and CA will occupy

the extra pore volume of the support, thus limiting the

loading of active components in the catalyst. Therefore, in

some cases, if a high loading of the active phase is

required, successive impregnation steps are necessary.

Acknowledgment The Research Council of Norway (NFR) is

acknowledged for their financial support via the KOSK programme.

References

1. DeJong KP (2009) Synthesis of solid catalysts. Wiley-VCH,

Weinheim

2. Marceau E, Carrier X, Che M, Clause O, Marcilly C (2008)

In: Gerhard E, Helmut K, Ferdi S, Jens W (eds) Handbook of

heterogeneous catalysis, vol 1. Wiley-VCH, Weinheim, p 467

3. van Dillen AJ, Terorde R, Lensveld DJ, Geus JW, de Jong KP

(2003) J Catal 216:257

4. Geus JW (2007) In: Regalbuto J (ed) Catalyst preparation science

and engineering. CRC press, London, p 341

5. Kakihana M, Yoshimura M (1999) Bull Chem Soc Jpn 72:1427

6. Lin J, Yu M, Lin CK, Liu XM (2007) J Phys Chem C 111:5835

7. Cushing BL, Kolesnichenko VL, O’Connor CJ (2004) Chem Rev

104:3893

8. Fang TT, Wu MS, Tsai JD (2002) J Am Ceram Soc 85:2984

9. Petrykin V, Kakihana M (2005) In: Sakka S (ed) Handbook

of sol–gel science and technology (sol–gel processing), vol 1.

Kluwer Academic Publishers, New York, p 77

10. Kakihana M, Arima M, Nakamura Y, Yashima M, Yoshimura M

(1999) Chem Mater 11:438

11. Paris EC, Leite ER, Longo E, Varela JA (1998) Mater Lett 37:1

12. Chao CH, Ownby PD (1995) J Mater Sci 30:6136

13. D’Epifani A, Fabbri E, Di Bartolomeo E, Licoccia S, Traversa E

(2008) Fuel Cells 8:69

14. Hernan L, Morales J, Sanchez L, Santos J (1997) Solid State Ion

104:205

15. Liu W, Kowal K, Farrington GC (1998) J Electrochem Soc

145:459

16. Yamashita Y, Yoshida K, Kakihana M, Uchida S, Sato T (1999)

Chem Mater 11:61

17. Yoshimura Y, Sato T, Shimada H, Matsubayashi N, Imamura M,

Nishijima A, Higo M, Yoshitomi S (1996) Catal Today 29:221

18. Isupova LA, Alikina GM, Tsybulya SV, Salanov AN, Boldyreva

NN, Rusina ES, Ovsyannikova IA, Rogov VA, Bunina RV,

Sadykov VA (2002) Catal Today 75:305

19. Frolova-Borchert YV, Sadykov VA, Alikina GM, Lukashevich

AI, Moroz EM, Kochubey DI, Kriventsov VV, Zaikovskii VI,

Zyryanov VV, Uvarov NF (2006) Solid State Ion 177:2533

20. Sadykov V, Mezentseva N, Alikina G, Bunina R, Pelipenko V,

Lukashevich A, Tikhov S, Usoltsev V, Vostrikov Z, Bobrenok O,

Smirnova A, Ross J, Smorygo O, Rietveld B (2009) Catal Today

146:132

21. Zyryanov VV, Uvarov NF, Sadykov VA, Frolova YV, Alikina

GM, Lukashevich AI, Ivanovskaya MI, Criado JM, Neophytides

S (2005) Catal Today 104:114

22. Kim N, Kim BH, Lee D (2000) J Power Sources 90:139

23. Quinelato AL, Longo E, Perazolli LA, Varela JA (2000) J Eur

Ceram Soc 20:1077

24. Sadykov VA, Kumetsova TG, Alikina GM, Frolova Y,

Lukashevich AI, Potapova YV, Muzykantov VS, Rogov VA,

Kriventsov VV, Kochubei DI, Moroz EM, Zyuzin DI, Zaikovskii

VI, Kolomiichuk VN, Paukshtis EA, Burgina EB, Zyryanov VV,

Uvarov NF, Neophytides S, Kemnitz E (2004) Catal Today

93(5):45

25. Delmon B, Devillers M (2008) In: Gerhard E, Helmut K, Ferdi S,

Jens W (eds) Handbook of heterogeneous catalysis, vol 1. Wiley-

VCH, Weilheim, p 295

26. Di Monte R, Fornasiero P, Kaspar J, Graziani M, Gatica JM,

Bernal S, Gomez-Herrero A (2000) Chem Commun 2167

27. Boullosa-Eiras S, Vanhaecke E, Zhao T, Chen D, Holmen A

(2010) Catal Today 166:10

28. Silva FD, Ruiz JAC, de Souza KR, Bueno JMC, Mattos LV,

Noronha FB, Hori CE (2009) Appl Catal A Gen 364:122

29. Tanaka H, Kaino R, Okumura K, Kizuka T, Tomishige K (2009)

J Catal 268:1

30. Donazzi A, Michael BC, Schmidt LD (2008) J Catal 260:270

31. Avgouropoulos G, Ioannides T, Matralis H (2005) Appl Catal B

Environ 56:87

32. Gamarra D, Munuera G, Hungria AB, Fernandez-Garcia M,

Conesa JC, Midgley PA, Wang XQ, Hanson JC, Rodriguez JA,

Martinez-Arias A (2007) J Phys Chem C 111:11026

33. Boullosa-Eiras S, Zhao T, Vanhaecke E, Chen D, Holmen A

(2011) Catal Today 171:104

34. Pham-Huu C, Vieira R, Louis B, Carvalho A, Amadou J, Dintzer

T, Ledoux MJ (2006) J Catal 240:194

35. Ledoux MJ, Vieira R, Pham-Huu C, Keller N (2003) J Catal

216:333

36. Socrates G (2002) Infrared and Raman characteristic group

frequencies. Wiley, New York

37. Li C, Li MJ (2002) J Raman Spectrosc 33:301

38. Hidber PC, Graule TJ, Gauckler LJ (1996) J Am Ceram Soc

79:1857

39. Ros TG, van Dillen AJ, Geus JW, Koningsberger DC (2002)

Chem A Eur J 8:1151

40. Toebes ML, van Heeswijk EMP, Bitter JH, van Dillen AJ, de

Jong KP (2004) Carbon 42:307

41. Kazakov AI, Rubtsov YI, Lempert DB, Manelis GB (2003) Russ

J Appl Chem 76:1214

42. Papageorgiou P, Price DM, Gavriilidis A, Varma A (1996)

J Catal 158:439

43. Di Monte R, Fornasiero P, Kaspar J, Rumori P, Gubitosa G,

Graziani M (2000) Appl Catal B Environ 24:157

1174 Top Catal (2011) 54:1163–1174

123