Notice TP III

Analyses Thermiques

• Thermogravimétrie (TG) • Analyse thermique différentielle (DTA) • Calorimétrie Différentielle Programmée (DSC)

Définition : Group de techniques pour lesquelles une propriété physique d’un matériau est mesurée en fonction de la température, pendant la mesure le matériau étant soumis à un programme contrôlé en température.

.

Assistant S.C. Sandu, 2004 Responsable: R. Sanjinés, 09-2014

NOTICE: Travaux pratiques III Analyses thermiques Section de physique – FSB

2

2

I. Introduction

Les techniques d’analyse thermique sont appliquées dans tous les domaines de la science, de l’archéologie à zoologie et pour tous les types de matériaux.

Tableau 1. Exemple de techniques d’analyse thermique

Technique Paramètre mesuré Instrument utilisé

Thermogravimétrie (TG) Masse Thermo balance

Analyse Thermique Différentielle (DTA)

Différence de températures Appareil DTA

Calorimétrie Différentielle Programmée (DSC)

Différence de flux de chaleur Calorimètre DSC

Analyse Thermomécanique (TMA)

Volume ou longueur Dilatomètre

Thermoluminescence (TL) Emission de la lumière Photo détectrice

Pour d’autres exemples de technique d’analyse thermique consulter la Ref. [1].

Les premières trois méthodes mentionnées auparavant (TG, DSC et DTA) peuvent fournir des informations concernant la stabilité thermique des matériaux et les changements d’enthalpies pendant la décomposition thermique ou pendant les changements de phase. Ces sont ces techniques qu’on va employées au cours de ce TP.

II. Rappels théoriques

2.1. L’analyse thermogravimétrie

C’est une technique dans laquelle la variation de masse d’un échantillon est mesurée en fonction du temps ou de la température lorsque la température de l’échantillon change de façon contrôlée (dite aussi en mode programmée) dans un atmosphère déterminée. Il y a trois modes possibles :

a) mode isotherme: la mesure se fait à température constante

NOTICE: Travaux pratiques III Analyses thermiques Section de physique – FSB

3

3

b) mode quasi-isotherme: l’échantillon est chauffé seulement quand sa masse n’est varie pas. Lorsque la masse change la température est maintenue constante.

c) mode dynamique (le plus utilisé): l’échantillon est chauffé dans un environnement contrôlé (sous vide, gaz neutre, etc.). La montée en température est réglée de sorte à chauffer l’échantillon progressivement, de préférence d’une manière linéaire.

Quelques applications pour la thermogravimétrie :

1. Décomposition thermique des matériaux organiques et inorganiques (polymères, metaux, etc.)

2. Corrosion de métaux dans différentes atmosphères 3. Détermination des températures et des vitesses de sublimation ou

d’évaporation 4. Mesures de désorption, absorption et adsorption 5. Calcination des matériaux

L’appareille de mesure est la thermo balance. L’instrument doit être capable

d’enregistrer les variations de masse avec une précision meilleure de 0.01% est les variations de température avec une précision de 1%. Dans nos expérimentations nous utilisons une microbalance électronique [ref. 1, p 89]. La relation proportionnelle qui lie l’intensité du courant à la force d’équilibrage électromagnétique associe la mesure des variations de courant à la mesure des variations de masse [2].

2.2. L’analyse thermique différentielle (DTA)

C’est une technique dans laquelle la différence de température entre l’échantillon et la référence est mesurée en fonction du temps ou de la température lorsque la température de cet ensemble est programmée dans une atmosphère contrôlée.

2.3. La calorimétrie différentielle programmée (DSC)

C’est une technique dans laquelle le flux de chaleur (puissance thermique) de l’échantillon et la référence est mesuré en fonction du temps ou de la température lorsque la température de cet ensemble est programmée dans une atmosphère contrôlée. En pratique, on mesure la différence de flux de chaleur entre un creuset contenant l’échantillon et un creuset de référence.

2.4 Dispositifs expérimentaux

Les changements de température (en DTA) ou de flux de chaleur (en DSC) sont la conséquence de transitions ou de réactions endothermique (∆Q < 0) ou exothermique (∆Q > 0) [Ref.1, p 214] comme celles déterminées par les changements de phases, fusion, cristallisation, transitions vitreuses, dissociations, oxydations, polymérisation, etc. Ces changements sont détectés par une méthode différentielle (voir figure 1).

NOTICE: Travaux pratiques III Analyses thermiques Section de physique – FSB

4

4

Fig 1. Représentation schématique d’un dispositif expérimental de DTA

Fig.2 Courbe DTA en fonction de la température de chauffage : I – transition de deuxième ordre (ex : transition vitreuse) II et III – pics endothermes (ex : fusion ou décomposition)

NOTICE: Travaux pratiques III Analyses thermiques Section de physique – FSB

5

5

Les courbes DTA ou DSC en fonction de la température de chauffage contiennent des séries de pics orientés négativement (vers le bas) ou positivement (vers le haut) de l’axe de flux de chaleur (Fig. 2). Les positions des pics sur l’axe de température et leur nombre permettent l’identification qualitative du matériau analysé. Quant aux aires des pics, elles permettent les calcules des chaleurs de réaction, transition, fusion, ou cristallisation. Parfois, des informations sur les cinétiques de réactions peuvent être obtenir à partir des courbes DTA/DSC.

La différence de principe entre DTA (b) et DSC (a) est illustrée dans la figure 3.

Fig. 3 Représentation schématique des dispositifs DSC et DTA

2.5 Transition de phase et réactions chimiques

Au sens thermodynamique, la phase caractérise l’état physique d’une substance chimique déterminée ou d’un mélange homogène de telles substances. Un système homogène est un système dont les propriétés restent identiques en tout point. De cette façon on distingue les phases gazeuses, liquides et solides, et dans le cas de la matière condensée, les modifications allotropiques éventuelles concernant la structure cristalline, magnétique, électrique, etc.

NOTICE: Travaux pratiques III Analyses thermiques Section de physique – FSB

6

6

Une transition de phase est donc le passage de la substance d’une phase à une autre qui se traduit par une modification de la structure interne de la substance. Tous les changements de phase s’accompagnent d’une absorption ou d’un dégagement de chaleur. Il faut toujours fournir de la chaleur à un solide pour le fondre ou le sublimer. Inversement, la solidification d’un liquide ou la condensation d’une vapeur dégage de la chaleur.

On appelle chaleur latente de changement de phase d’un corps pur la quantité de chaleur L qu’il faut fournir ou soutirer à un corps d’une certaine masse pour la transformer d’une phase en une autre dans des conditions d’équilibre isotherme/ isobare. Ainsi donc, expérimentalement l’existence d’une chaleur latente facilite la détection d’un changement de phase. Evidemment, la quantité de chaleur présente lors d’un changement de phase dépend non seulement de la chaleur latente mais aussi de la masse total de la substance.

Les réactions chimiques d’oxydation ou de réduction mettent en jeu de l’énergie, le plus souvent sous forme de chaleur. Lorsqu’une réaction thermochimique dégage de l’énergie la réaction est dite exothermique. Au contraire, si la réaction adsorbe de l’énergie, elle est dite endothermique. La détection de ces chaleurs de réaction ou enthalpies de formation permet d’étudier certains mécanismes d’oxydation, de nitruration, d’hydrolyse, etc.

2.6 Transfert de chaleur

Un transfert de chaleur peut s’opérer suivant les trois processus suivants : par conduction, par convection, et par rayonnement thermique. Dans le cas de la conduction thermique, l’énergie thermique s’échange par contact direct de deux corps. C’est ce type de transfert de chaleur qui est utilisé pour la mesure de la température entre l’échantillon et le four. Par contre les processus d’échange de chaleur par convection et par rayonnement thermique sont utilisés pour chauffer l’échantillon. A basses et moyennes températures (4K-600K) le chauffage se fait principalement par convection en utilisant comme gaz d’échange un gaz neutre (He, Ar, N2) ou un gaz réactive (O2, CO, NOx, H2,.. ). A hautes températures ou sous vide, le rayonnement thermique assure un chauffage efficace.

III. Buts de l’expérience

1. Acquérir une connaissance de techniques d’analyses thermiques et de leurs applications.

2. Connaître quels sont les effets physiques et chimiques sur un échantillon soumis à un échauffement prolongé et ou programmé.

3. Investiguer et comparer les propriétés physico-chimiques des matériaux. 4. Calibrer un DSC en utilisant de standards pour la mesure des chaleurs

spécifiques de fusion et de cristallisation des différents matériaux. 5. Etudier la cinétique de la décomposition thermique des composés.

NOTICE: Travaux pratiques III Analyses thermiques Section de physique – FSB

7

7

IV. Exemples d’applications:

1. Transition vitreuse, cristallisation et fusion du PET. 2. Décomposition thermique d’oxalate de calcium hydraté. Calcul des vitesses de

réaction chimique. 3. Cristallisation d’un alliage amorphe Ni-Ti. 4. L’étude des alliages eutectiques (Ex. Sn-Pb). Déterminer la composition

chimique d’un alliage eutectique à partir de données de ATD. 5. Comportement des substances organiques (sucre, pâte de biscuits) lors de leur

cuisson (Attention : adapter la quantité de matière à analyser et le domaine de température !).

6. Décomposition thermique et formation de phases d’une poudre de YBaCuO. 7. Observer le comportement et identifier les changements de phase des

polymères soumis au chauffage contrôlé (lire par ex. ref. 7). 8. Caractérisation de composés ternaires ou quaternaires (poudres) synthétisés par

la méthode SOL-GEL (exemple ref. 8) .

Bibliographie:

1. Thermal Analysis (3rd Edition), W. Wendlandt, John Wiley and Sons (USA), 1986.

2. Notices SETARAM 3. Thermal Analysis, B. Wunderlich, Academic Press (USA),1990. 4. Thermal Characterization of Polymeric Materials, E.A. Turi, Academic Press

(USA), 1981. 5. Thermal Analysis of Foods, V.R. Harwalkar, Elsevier (UK), 1990. 6. Thermal kinetic TG-analysis of metal oxalate complexes, Li Jun &co,

Thermochimica Acta, 406 (2003) 77-87. 7. « Study of the early deactivation in pyrolysis of polymers in the presence of

catalysts », A. Marcilla, A. Gomez-Siurana, D. Berenguer, J. Anal.Appl. Pyrolysis (2007).

8. Synthesis and characterization of Ba1-xSrxTiO3 nanopowders by citric acid gel method », Z. Wang et al, Ceramics International (2006).

+ Models

JAAP-2061; No of Pages 7

Study of the early deactivation in pyrolysis of polymers in

the presence of catalysts

A. Marcilla *, A. Gomez-Siurana, D. Berenguer

Dpto. Ingenierıa Quımica, Facultad de Ciencias, Universidad de Alicante, Apartado 99, 03080 Alicante, Spain

Received 20 June 2006; accepted 2 February 2007

Abstract

In this work, the early degradation step of the pyrolysis of some polymers in the presence of certain catalysts has been studied using

thermogravimetric analysis (TGA). Three commercial polymers (PE, PP and EVA) and three catalysts were studied (ZSM-5, MCM-41a, and

MCM-41b), and the MCM-41a catalyst has been selected for the analysis of the earlier steps of the pyrolysis process carried out in the presence of

catalysts. Several cycles of heating–cooling were performed using a thermobalance, in order to analyze the influence of the first stages of

decomposition on the activity of the more accessible active sites involved. In this way, the behaviour of the polymer–catalyst mixtures (20% (w/w)

of catalyst) was studied and compared with that observed in the corresponding thermal degradation as well as in the pyrolysis in the presence of

catalysts, in a single heating cycle.

The results obtained clearly show the existence of an early degradation step. For a polymer–catalyst system with low steric hindrances such as

PE-MCM-41, this early degradation step causes a noticeable decrease of the catalyst activity for the main decomposition step (i.e., cracking of the

chain). The decrease of the catalytic activity is lower for a polymer–catalyst system with higher steric restrictions, as occurs in the EVA-MCM-41

degradation process. However, in this case, the catalyst activity in the first decomposition step (i.e., the loss of the acetoxi groups) is noticeable

decreased after one pyrolysis run, thus reflecting that the active sites involved are mainly the most accessible ones.

# 2007 Elsevier B.V. All rights reserved.

Keywords: Pyrolysis; Catalysts; Deactivation; PP; EVA; HZSM-5; MCM-41

www.elsevier.com/locate/jaap

J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx

1. Introduction

The rapid growth of the range of applications of plastics:

packaging, agriculture, construction, household, etc., has

turned these materials into a serious environmental problem

since plastics comprise 10 wt% (>30 vol.%) of municipal solid

waste.

Pyrolysis is one of the promising alternatives to dumping or

incineration for the treatment of plastic wastes, where the

polymer sample is heated in an inert atmosphere causing the

cracking of the polymer backbone. The thermal cracking of

polymers produces a very complex mixture of products, which

may be used as fuel but has low value as a source of chemicals.

However, by plastic cracking over different solid catalysts

(pyrolysis in the presence of catalysts) it is possible to obtain

* Corresponding author. Fax: +34 965903826.

E-mail address: antonio.marcilla@ua.es (A. Marcilla).

0165-2370/$ – see front matter # 2007 Elsevier B.V. All rights reserved.

doi:10.1016/j.jaap.2007.02.002

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Py

products with higher commercial value and with a wider

number of applications [1–3].

Modern cracking uses zeolites as catalysts [1–5], which

cause an important reduction in the decomposition temperature

and allow more valuable gaseous products to be obtained.

However, the pore size of the zeolites is limited to a maximum

value of about 1.0 nm, which hinders the access of bulky

molecules to the acid sites located inside the channels. MCM-

41 is a catalyst, member of the relatively new family of

mesoporous materials, first synthesized in 1992 by Beck et al.

[6], which possesses highly regular arrays of uniform size pore

channels ranging from 1.5 to 10 nm and a large surface area,

thus improving the accessibility to the active sites located in the

inner of the mesopores [6].

Coke formation and its retention inside the pores is the main

cause of deactivation of zeolite and aluminosilicates catalysts

in hydrocarbon processing. Deactivation of these catalysts in

the pyrolysis of polymers in the presence of catalysts is mainly

due to active site poisoning and pore blockage by coke

rol. (2007), doi:10.1016/j.jaap.2007.02.002

Table 1

Characteristics of the commercial polymers

Material Density

(g/cm3)

Melt index

(g/10 min)a

Vicat

point

(8C)

Tensile

modulus

(MPa)

LDPE 780R 0.923 20 (190/16) 93 115

PP Novolen 1100L 0.910 8 (230/2.2) 70 1500

EVA Escorene UL15028CC 0.950 150 (125/0.3) 64 13

a 8C/kg.

A. Marcilla et al. / J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx2

+ Models

JAAP-2061; No of Pages 7

deposition [7]. Both coking rate and coke composition depend

on a variety of parameters such as the characteristics of

the catalyst pore structure, the characteristics of the acid

sites, the nature of the reactant and the operating conditions

[8,9].

The study of the pyrolysis of polymers in the presence of

catalysts using thermogravimetric analysis (TGA) shows a

early degradation step of the process for some polymer–catalyst

systems. This process causes a widening of the corresponding

peaks in the derivative of TGA curves (DTG curves), and in

some cases, the appearance of a shoulder [10]. This early

degradation step of the decomposition reactions is related to the

acid sites located on the surface of the material [11,12]. In fact,

according to bibliography [13], the influence of the acid sites

located on the external surface and in the pore mouth region of

zeolites cannot be ignored, although the concentration of these

acid sites is assumed to be smaller, generally 3–5% of the total

acid sites.

As an example, previous works on the pyrolysis of EVA in

the presence of catalysts [12,14] showed that, despite the

presence of some catalyst advances noticeably at the beginning

of the pyrolysis of the first step of EVA pyrolysis, the

corresponding temperature of maximum reaction rate – i.e., the

peak temperature in the DTG curves – remains almost

unaltered, thus suggesting a possible simultaneous deactivation

of the active sites involved in the early degradation step, and

consequently the overall process ends at the same temperature

as in the absence of catalyst.

In a previous work [10], a comparative study of the

behaviour of different solid acid catalysts in the pyrolysis of

different polymers was performed. In that work, a possible

influence of some catalysts in the initial degradation steps of

some polymers was observed. A hypothesis was suggested

that the active sites located on the external surface of the

catalyst or, in the case of mesoporous materials, the more

easily accessible active sites, would be involved in these

processes. On the other hand, as previously mentioned, some

studies about the pyrolysis of EVA over different catalysts

[12,14] also suggested that these active sites would undergo

deactivation processes simultaneous with their participation

in the above-mentioned earlier degradation steps. Therefore,

this work is focused on the study of the initial stages of the

pyrolysis in the presence of catalysts of three selected

polymers, with the main objective of validating the

hypothesis previously suggested. In this way, TGA studies

were carried out, and four cycles of heating–cooling were

performed in order to analyze the influence of the first stages

of decomposition on the activity of the active sites involved.

Three polymers (Polyethylene, PE, polypropylene, PP and an

ethylene–vinyl acetate copolymer, EVA) were selected, and

three catalysts were studied (ZSM-5, MCM-41a, and MCM-

41b). Then, MCM-41a was selected to study this phenom-

enon in more detail. The behaviour of polymer-catalyst

mixtures (20% (w/w) of catalyst) was studied and compared

with that observed in the corresponding thermal degradation

as well as in the pyrolysis in the presence of catalysts, in a

single heating cycle.

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Py

2. Experimental

2.1. Materials

Table 1 shows the main characteristics of the commercial

polymers (polyethylene, PE, polypropylene, PP, and an

ethylene-vinyl acetate copolymer, EVA) selected in this work.

The nominal vinyl acetate (VA) content of the EVA copolymer

was 27% (w/w). The catalysts used were ZSM-5, MCM-41a

and MCM-41b, prepared in accordance with the method

reported in literature [15,16]. The main characteristics of these

catalysts are summarized in Table 2. As can be seen, in the case

of the MCM-41a, the low value of the Si/Al ratio produces a

material with high acidity. This material also presents a high

pore size and a large surface area, typical of this type of material

[10].

Pure polymer or copolymer samples as well as the

corresponding mixtures with 20% (w/w) catalyst were

pyrolysed under the experimental conditions described in the

next section.

2.2. Equipment and experimental conditions

The pyrolysis experiments were carried out in a Netzsch TG

209 thermobalance controlled by a PC under the Windows

operating system. The atmosphere used was nitrogen with a

flow rate of 45 STPml/min. The balance and the oven

thermocouple of the equipment were calibrated following the

manufacturer’s instructions. Pure polymers as well as mixtures

of powdered polymers and a catalyst of around 4 mg (3.2 mg of

the polymer with 0.8 mg of the catalyst) were pyrolysed.

In the pyrolysis in the absence of catalyst as well as in the

pyrolysis runs in the presence of catalyst in a single heating

cycle, the temperature was raised to 550 8C with a heating rate

of 10 8C/min. On the other hand, for studying the influence of

the initial decomposition steps in the overall process, four

cycles of heating–cooling were applied to the polymer + ca-

talyst system studied. The heating–cooling cycles have been

performed under N2 atmosphere, in the thermobalance

equipment, which is provided with a cooling system, and

involve a heating rate of 10 8C/min and a cooling rate of 50 8C/

min. The initial and final temperatures were 50–300 8C for the

PE mixtures, 50–245 8C for the PP mixtures, and 50–235 8C for

the EVA mixtures, respectively. In each heating run,

temperature was raised to a value below the temperature

corresponding to the apparent beginning of the thermal

decomposition of each polymer. All the heating–cooling runs

rol. (2007), doi:10.1016/j.jaap.2007.02.002

Table 2

Characteristics of the three catalysts used

Property HZSM-5 MCM-41a MCM-41b

Pore size (nm)a 0.51 � 0.55 1.7 2.4

BET area (m2/g)b 334 956 1136

External surface area (m2/g)c 67 126 383

Pore volume (cm3/g)d 0.27 1.00 1.34

Si/Al ratioe 24 7 47

Acidity (mmol/g)f 1.36 2.08 0.60

Acidity weak acid sites f 0.82 1.39 0.60

Acidity strong acid sitef 0.54 0.69 –

Tdesorption NH3 (8C)f weak

acid sites

165 155 138

Tdesorption NH3 (8C)f strong

acid sites

385 550 –

Aloct/Altetrag 0.12 0.70 0.37

a BJH.b N2 adsorption isotherms; BET method.c N2 adsorption isotherms; t method.d N2 adsorption isotherms; measured at P/P0 = 0.995.e XRF.f TPD of NH3.g NMR.

Fig. 2. DTG and TGA curves from NH3 TPD of the three catalysts studied by

[10].

Fig. 1. TGA and DTG curves obtained in the pyrolysis of PP in the absence of

catalyst and in the pyrolysis in the presence of the different catalysts [10].

A. Marcilla et al. / J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx 3

+ Models

JAAP-2061; No of Pages 7

have been performed under the same experimental conditions

above described.

In order to evaluate the influence of the heating–cooling

cycles on the activity of the catalyst, an experiment was

performed, where the pyrolysis run was stopped when the

fourth cycle was ended and more fresh polymer was added in

order to once more achieve a 20% (w/w) catalyst mixture; then

the pyrolysis in the presence of catalysts was carried out in the

thermobalance, from 50 to 550 8C at 10 8C/min.

3. Results and discussion

According to bibliography [17,18], the first step in the

pyrolysis of a polymer in the presence of catalysts consists in

the cracking of the polymeric chains over the catalyst surface,

leading to oligomers, with higher possibilities for access to the

active sites located in the inner part of the catalyst pores, where

the complete degradation of the oligomers occurs. However,

most articles – referring to [17,19] or not referring to [18,20,21]

this initial step – do not report any noticeable loss of weight

associated to the oligomer formation, and the references are few

where a two-step process is considered [22]. In fact, according

to the results obtained in the present work, the existence of an

initial step involving the formation of volatile compounds

seems to be very dependent on the chemical and structural

characteristics of the polymer and the catalyst.

Fig. 1 shows the TGA and DTG curves obtained for the

pyrolysis of PP in the presence of the two different MCM-41

samples as well all in the presence of HZSM-5 zeolite. The

behaviour of these systems was discussed and compared

elsewhere [10], and here only the relative importance of the

initial decomposition step is emphasised. According to Table 2,

the main differences between the two MCM-41 samples

(MCM-41a and MCM-41b) are related to the respective BET

area and pore size (BET area 956 and 1136 m2/g and pore size

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Py

of 1.7 and 2.4 nm, respectively) and the acidity (2.08 and

0.60 mmol/g, respectively). The main characteristics of the

HZSM-5 sample were BET area of 334 m2/g, pore size of

0.51 nm � 0.55 nm and acidity of 1.36 mmol/g [10]. Accord-

ing to these values, the higher specific surface and pore size of

MCM-41b favours the increase of the activity of this sample,

but the higher acidity of MCM-41a results in a higher activity,

as can be seen in Fig. 1, and as discussed elsewhere [10]. On the

other hand, the lower specific surface of HZSM-5 leads as a

consequence to the lower activity of this material in comparison

with the two MCM-41 samples.

According to the results of Fig. 1, the tail appearing at

temperatures below the peak temperature would suggest, for the

three catalysts, the existence of an initial step, and this is

specially marked in the presence of MCM-41a. This behaviour

reflects a clear influence of the catalyst characteristics, and

suggests that this can be only clearly observed if the catalyst

possesses a certain combination of pore size, specific surface

and acidity which lead to the existence of a high enough number

of very accessible active sites with also high enough acidity.

Fig. 2 shows the TGA and DTG curves obtained in programmed

thermal desorption of NH3, in the conditions described

elsewhere [10] for the catalysts reported in Fig. 1. The first

peak – i.e., the peak at lower temperatures – corresponds to the

NH3 desorption from the weak acid sites, whereas the second

rol. (2007), doi:10.1016/j.jaap.2007.02.002

Fig. 3. TGA and DTG curves obtained for the pyrolysis in the absence of

catalyst and in the presence of MCM-41a catalytic pyrolysis at 10 8C/min of (a)

PE, (b) PP and (c) EVA.

Table 3

Temperature of maximum decomposition rate for the main decomposition steps

of the different systems studied

Sample Experiment T (8C)

PE Thermal 473

PE-MCM-41 Conventional pyrolysis 358

PE-MCM-41 Pyrolysis after 4 cycles-polymer addition 386

PP Thermal 462

PP + MCM-41 Conventional pyrolysis 340

PP + MCM-41 Pyrolysis after four cycles-polymer addition 356

EVA Thermal 353–468

EVA + MCM-41 Conventional pyrolysis 353–406

EVA + MCM-41 Pyrolysis after 4 cycles-polymer addition 353–405

A. Marcilla et al. / J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx4

+ Models

JAAP-2061; No of Pages 7

peak corresponds to NH3 desorption from stronger acid sites.

The respective peak temperatures indicate the relative strength

of the acid sites. As can be seen, the strength of the weak acid

sites is similar for the three catalysts; but MCM-41b does not

possess strong acid sites, and the strength of the strong acid sites

of MCM-41a is noticeably higher for MCM-41a than for

HZSM-5. Table 2 shows the acidity associated to each type of

acid sites for each material. As can be seen MCM-41a possesses

the higher acidity. Therefore, the two main characteristics of

this material – i.e. high acidity and high accessibility – support

its role in the earlier stage of the pyrolysis in the presence of

catalysts. The behaviour observed for MCM-41b shows that the

only presence of accessibility is not enough for enhancing the

early step of the pyrolysis process. On the other hand, neither

high acidity nor strong acid sites are enough, as the behaviour of

HZSM-5 shows, and the simultaneous presence of high

accessibility properties is also needed.

Therefore, according to the previous comments, MCM-41a

has been selected for the detailed study of the early degradation

step of the pyrolysis in the presence of catalysts process. As

shown in Table 2, this material has high acidity and very

accessible active sites. Fig. 3 shows the TGA and DTG curves

corresponding to decomposition in the absence of catalyst and

the MCM-41a-pyrolysis of PE, PP and EVA. As can be seen, the

DTG curves corresponding to the pyrolysis of PE and PP in the

presence of catalysts (Fig. 3a and b) exhibit a shoulder at

temperatures below the temperature of maximum reaction rate

– i.e., the peak temperature – indicating the existence of an

initial reaction step with some weight loss, i.e. involving the

generation of volatile compounds, which is different to the

main reaction step, despite the fact that both are clearly

overlapped. This shoulder does not appear in the curves

corresponding to the thermal processes curves and, therefore, it

can be concluded that the catalyst must be involved in the

reactions generating the shoulder. As the comparison between

the curves corresponding to the pyrolysis in the absence and in

the presence of catalysts of each polymer shows (Fig. 3a and b),

and in good agreement with bibliography [23–25], the presence

of the catalyst causes a noticeable decrease in the temperature

of maximum reaction rate for the main decomposition step (see

Table 3). Therefore, despite the clear appearance of an early

degradation step, no conclusion can be extracted in relation to

the above-mentioned eventual deactivation processes from the

curves shown in Fig. 3a and b and only the role of the catalyst in

the early beginning of the decomposition can be emphasised.

Fig. 3c shows the TGA and DTG curves corresponding to

the decomposition of EVA in the absence and in the presence of

catalysts. As in the previous case, a shoulder appears in the

DTG peak, corresponding to the first decomposition step and

associated to the participation of the catalyst in the early

degradation step of the process. However, some differences

with respect to the PE and PP cases can be observed. It is well

known that whereas the decomposition of PE and PP in the

absence and in the presence of catalysts show a single main

decomposition step [26], the decomposition of EVA proceeds

through two main decomposition steps [12,27–29], which

involve, respectively, the vinyl acetate groups (VA groups) loss

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Pyrol. (2007), doi:10.1016/j.jaap.2007.02.002

A. Marcilla et al. / J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx 5

+ Models

JAAP-2061; No of Pages 7

and the cracking of the polymeric chain resulting from the

previous step. These two steps are reflected in the weight loss of

the TGA curves and in the peaks of the corresponding DTG

curves of Fig. 3c. According to bibliography [12], a shift of the

second decomposition step towards lower temperatures appears

as a consequence of the presence of the catalyst, whereas no shift

in the temperature peak of the first decomposition step can be

observed (see Table 3). As in the cases of PE and PP, the early

beginning also appears in the degradation of EVA in the presence

of catalysts. However, in this case, the reaction step affected by

the early beginning is not the cracking of the chain, but the loss of

VA groups. Therefore, it can be concluded that the catalyst also

acts in the first step of EVA degradation, provoking a widening of

the corresponding interval of reaction temperature. Moreover,

the temperature of maximum rate for the first step results almost

unchanged, thus suggesting that the active sites involved in the

process undergo some deactivation process. Nevertheless,

despite this eventual deactivation, related to the participation

of the catalyst in the VA groups loss, the catalyst maintains

enough activity to reduce the temperature of the second

decomposition step, where the polymeric chain is cracked

(see Table 3). Considering the voluminous nature of the VA

groups, the hypothesis that the active sites involved are the more

accessible ones can be suggested.

Therefore, there are two aspects to be studied and assessed:

(a) the influence of the external active sites in the early

beginning of the decomposition process and (b) the eventual

deactivation of these active sites. The first question will be

clarified through the comparison among the behaviours of

different polymer + catalyst systems, and applying the meth-

odology described in the previous section for performing

several consecutive heating–cooling cycles. The second

question will be solved considering the behaviour during these

cycles together with the study of the case where fresh polymer

was added, and then the pyrolysis is continued until 500 8C, in

comparison with the pyrolysis runs in the presence of catalyst,

in a single heating cycle.

3.1. Influence of the external active sites in the early

beginning of the decomposition process

The comparison between Fig. 3a and b shows that the initial

degradation step in the pyrolysis in the presence of catalysts is

more pronounced for the PP decomposition than in the case of

PE. This fact could be related to the high number of tertiary

carbon atoms in the PP chains, which results in a more reactive

polymer [10,30]. However, the case of the pyrolysis of EVA in

the presence of catalysts, also with high number of tertiary

carbons, deserves a separate discussion. Unlike PE and PP,

whose degradation occurs in a main single step, as mentioned

above, the EVA pyrolysis involves two steps. Moreover,

according to bibliography [12], the peak temperature of the first

step is almost not affected by the presence of the acid solid.

Therefore, there are two possibilities:

(a) The chemical processes involved in the early degradation

step of EVA pyrolysis are similar to those occurring in the

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Py

cases of PE and PP. In other words, the weight loss observed

before the DTG peak related to the first step of EVA

decomposition is a result of the same chemical reactions in

both cases. However, this hypothesis does not seem very

feasible, because it implies that the catalytic cracking of the

chain could start before the VA groups loss, and this is not

supported by experience [14,27,29].

(b) The active sites of the catalyst are also involved in the step

of VA groups loss. Therefore, the catalytic phenomenon

observed in this case is somewhat different than those

involved in the cases of PE and PP, not corresponding to the

cracking reactions of the main chain, but also involving the

availability of very accessible active sites with enough

acidity. According to the bibliography, this seems to be the

more probable hypothesis, as shown by the composition of

the gas evolved during the first step of EVA pyrolysis (i.e.,

mainly acetic acid). Therefore, the results obtained show

that the process of loss of the VA group’s loss also

undergoes the early degradation step.

3.2. Eventual deactivation of the active sites in the early

beginning of the pyrolysis process in the presence of

catalysts

Fig. 4 shows the TGA curves obtained in the heating–

cooling cycles performed for each polymer + catalyst system.

For each polymer, the final temperature of the heating cycles is

different (300 8C for the PE mixture, 245 8C for the PP mixture,

and 235 8C for the EVA mixture), and has been selected below

the temperature where the pyrolysis in the absence of catalyst

starts, in order to ensure that the weight loss observed could

only be attributed to the catalyst presence. As can be seen, for

the three polymers, the weight loss obtained in the first heating

cycle is higher than that obtained in the next cycles. This fact

can be explained considering that the active sites involved in

this process deactivate quickly, resulting in a material with less

capacity for its catalytic role in the following cycles. In other

words, considering that the existence of deactivation simulta-

neous to the initial step occurs.

The analysis of the different heating–cooling cycles shown

in Fig. 4 indicates that the weight loss associated to the initial

step starts at lower temperatures in the first cycle than in those

following. Table 4 shows the temperatures of the starting point

of the weight loss associated to the first and to the following

heating cycles. Despite the fact that the low temperature

observed for the first cycle could be related to the elimination

of some adsorbed water on the catalyst, a clear decrease of the

starting temperature for the degradation process can be

observed, thus suggesting that deactivation readily occurs.

Moreover, the weight loss observed in each cycle is noticeably

lower in the case of EVA than in the cases of PE and PP. This

behaviour is in good agreement with the fact that the

degradation step affected by the early degradation step

involves lower weight loss in the case of EVA, where only

the 27% corresponding to the VA groups is pyrolysed, in

contrast to the case of PE and PP, where 100% of polymer may

react.

rol. (2007), doi:10.1016/j.jaap.2007.02.002

Fig. 4. TGA curves corresponding to the heating–cooling cycles of (a)

PE + MCM-41, (b) PP + MCM-41 and (c) EVA + MCM-41.

Table 4

Temperatures corresponding to the starting point of the weight loss in the first

and in the following heating cycles for the different polymers studied

Polymer T (8C)

Cycle 1 Cycle 2 Cycle 3 Cycle 4

PE 89 208 217 225

PP 90 174 176 181

EVA 70 162 167 175

Fig. 5. TGA and DTG curves corresponding to the pyrolysis in the absence and

in the presence of catalyst experiments performed in different conditions for (a)

PE, (b) PP and (c) EVA.

A. Marcilla et al. / J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx6

+ Models

JAAP-2061; No of Pages 7

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Py

Fig. 5 shows the DTG curves corresponding to the following

pyrolysis runs: (i) pyrolysis in the presence of catalysts, in a

single heating cycle, (ii) pyrolysis in the presence of catalysts

after the four heating–cooling cycles, with addition of fresh

polymer, and (iii) pyrolysis in the absence of catalyst, performed

as described in Section 2. As can be seen, for the three polymers,

the shoulder appearing before the peak of the DTG curves (the

only peak in the cases of PE and PP and the first peak in the case of

EVA), which has been related to the early degradation step, is

clearly diminished after the heating cycles. This behaviour again

indicates that simultaneous deactivation processes occur,

provoking a noticeably loss of the activity of the catalyst after

rol. (2007), doi:10.1016/j.jaap.2007.02.002

A. Marcilla et al. / J. Anal. Appl. Pyrolysis xxx (2007) xxx–xxx 7

+ Models

JAAP-2061; No of Pages 7

the heating–cooling cycles, when the fresh polymer is added and

the pyrolysis process is completed.

In the cases of PE and PP, when the pyrolysis in the presence

of catalysts is completed until the complete degradation of the

polymer, the catalyst losses part of its activity after the heating–

cooling cycles, as compared with the pyrolysis in a single

heating cycle (see the corresponding peak temperatures in

Table 3), thus corroborating the previous comments. However,

in the case of EVA, the capacity of the catalyst to decrease the

temperature of the second decomposition step results almost

unmodified by the successive heating–cooling cycles. PE and

PP could access the most of the active sites, i.e. those located on

the external surface, as well as those located at the opening of

the pores or mesopores and in the inner parts of the catalyst.

However, before the VA group loss, the EVA chains can only

access the more accessible active sites (simplifying, the

‘‘external’’ active sites), but leaving unaltered the less

accessible active sites, which therefore remain untouched for

the main decomposition step.

In good agreement with the previous comments, the

decrease of the activity, or in other words, the importance of

the deactivation process, is noticeably higher for PE than for PP.

As has been previously suggested, this behaviour can be

explained considering that the lower steric hindrances

associated to the PE chains could favour the access to the

polymer chains to less accessible active sites, perhaps some

located in the inner part of pores, thus resulting in an increase of

the deactivation of these acid sites. On the other hand, in the

case of EVA the higher steric hindrances associated to the VA

groups make access to the ‘‘less exposed’’ of the more

accessible active sites more difficult, resulting in a lower loss of

activity. For the second decomposition step, (which according

to bibliography [12,14], and in agreement with Figs. 3c and 5c

and Table 3, proceeds through a catalytic mechanism), these

steric effects would be enhanced by the crosslinking

phenomena, suggested by some authors [31], occurring before

the second step of EVA decomposition.

4. Conclusions

The pyrolysis of polymers in the presence of catalysts could

involve the existence of an initial degradation step, depending

on the chemical and structural characteristics of the polymer

and the catalyst. The behaviour observed suggests that this step

can only be clearly observed if the catalysts possesses a certain

combination of pore size, specific surface and acidity which

leads to the existence of a high enough number of very

accessible active sites and with a high enough acidity. On the

other hand, the relative importance of the early beginning of the

process seems to be increased as the reactivity of the polymer

chains increase: i.e., when the number of tertiary carbon atoms

increases. In the case of the pyrolysis of EVA in the presence of

catalysts, the active sites of the catalyst seem to be also involved

in the VA groups loss process.

The results obtained suggest that deactivation processes

occur simultaneously to the initial step. The more accessible

active sites, involved in such processes, are quickly deactivated,

Please cite this article in press as: A. Marcilla et al., J. Anal. Appl. Py

resulting in a noticeable effect in the activity of the catalyst

versus the main decomposition step. On the other hand, the role

of these active sites, as well as the corresponding deactivation

process seems to be increased as the steric hindrances

associated to the polymer structure decreases.

Acknowledgements

Financial support for this investigation has been provided by

the Spanish ‘‘Comision de Investigacion Cientıfica y Tecno-

logica’’ de la Secretarıa de Estado de Educacion, Universi-

dades, Investigacion y Desarrollo and the European

Community (FEDER refunds) (CICYT CTQ2004-02187),

and by the Generalitat Valenciana (project ACOMP06/162).

References

[1] K. Obin, D. Koumantaropoulus, G. Manos, Stud. Surf. Sci. Catal. 135

(2001) 4989.

[2] J. Aguado, D.P. Serrano, J.M. Escola, Polym. Deg. Stab. 69 (2000) 11.

[3] C. Vasile, H. Pakdel, B. Mihai, P. Onu, H. Darie, J. Anal. Appl. Pyrol. 57

(2001) 287.

[4] V. Ramdas, J.-S. Kim, S.-B. Lee, M.-J. Choi, J. Mol. Catal. A: Chem. 222

(2004) 133.

[5] J. Weitkamp, Sol. Stat. Ionics 131 (2000) 175.

[6] J.S. Beck, J.C. Vartuli, G.J. Kennedy, C.T. Kresge, S.E. Schramm, Chem.

Mater. 6 (1994) 1816.

[7] P.D. Hopkins, J.T. Miller, B.L. Meyers, G.J. Ray, R.T. Roginski, M.A.

Kuehne, H.H. Kung, Appl. Catal. A: Gen. 136 (1996) 29.

[8] M. Guisnet, P. Magnoux, in: E.G. Derouane, F. Lemos, C. Naccache, F.R.

Ribeiro (Eds.), Zeolite Microporous Solids. Synthesis, Structure and

Reactivity, NATO ASI Series, p. 437.

[9] M. Guisnet, P. Magnoux, Stud. Surf. Sci. Catal. 88 (1994) 53.

[10] A. Marcilla, A. Gomez-Siurana, D. Berenguer, Appl. Catal. A: Gen. 301

(2) (2006) 222.

[11] A. Gartforth, S. Fiddy, Y.-H. Lin, P.N. Sharratt, J. Dwyer, Thermochim.

Acta 294 (1997) 65.

[12] A. Marcilla, A. Gomez, S. Menargues, J. Garcıa-Martınez, D. Cazorla-

Amoros, J. Anal. Appl. Pyrol. 68–69 (2003) 495.

[13] S. Zheng, Tesis Doctoral, Universitat Munchen, 2002.

[14] A. Marcilla, A. Gomez, S. Menargues, Pol. Degr. Stab. 89 (2005) 454.

[15] IZA, International Zeolite Association. http://www.iza-synthesis.org/.

[16] D.P. Serrano, J. Aguado, J.M. Escola, Micropor. Mesopor. Mater. 34

(2000) 43.

[17] D.P. Serrano, J. Aguado, J.M. Escola, Ind. Eng. Chem. Res. 39 (2000) 1177.

[18] K. Gobin, G. Manos, Polym. Degrad. Stab. 86 (2004) 225.

[19] G. Manos, A. Garforth, J. Dwyer, J. Ind. Eng. Chem. Res. 39 (2000) 1198.

[20] J. Schirmer, J.S. Kim, E. Klemm, J. Anal. Appl. Pyrol. 60 (2001) 205.

[21] G.J.T. Fernandes, V.J. Fernandez, A.S. Arujo, Catal. Today 75 (2002) 233.

[22] J. Aguado, D. Serrano, Spain (1999) 140.

[23] A. Marcilla, A. Gomez, A.N. Garcıa, M.M.J. Olaya, Anal. Appl. Pyrol. 64

(2002) 85.

[24] A. Marcilla, A. Gomez, J.A. Reyes-Labarta, Kinetic Model. Polym. 42

(2001) 8103.

[25] A. Marcilla, A. Gomez, J.A. Reyes-Labarta, A. Giner, Polym. Degrad.

Stab. 80 (2) (2003) 233.

[26] H. Bockhorn, A. Hornung, U. Hornung, J. Anal. Appl. Pyrol. 50 (1999) 77.

[27] B.J. McGrattan, Appl. Spectrosc. 48 (12) (1993) 1472.

[28] A. Marcilla, A. Gomez, S. Menargues, J. Anal. Appl. Pyrol. 74 (1-2)

(2005) 224.

[29] A. Marcilla, A. Gomez, S. Menargues, Polym. Degrad. Stab. 89 (1) (2005)

145.

[30] A. Marcilla, A. Gomez, S. Menargues, R. Ruiz, Polym. Degrad. Stab. 88

(3) (2005) 456.

[31] B.-A. Sultan, E. Sorvik, J. Appl. Polym. Sci. 43 (1991) 1761.

rol. (2007), doi:10.1016/j.jaap.2007.02.002

+ Models

CERI-2415; No of Pages 5

Synthesis and characterization of Ba1�xSrxTiO3

nanopowders by citric acid gel method

Zheng Wang a, Shenglin Jiang b, Guangxing Li a, Mingpeng Xi b, Tao Li a,*a Department of Chemistry, Huazhong University of Science and Technology, Wuhan 430074, China

b Department of Electronic Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, China

Received 2 February 2006; received in revised form 20 February 2006; accepted 20 March 2006

Abstract

Stoichiometric and monophasic Ba1�xSrxTiO3 (x = 0.3) nanopowders were successfully prepared by the citric acid gel method using barium

nitrate, strontium nitrate and tetra-n-butyl titanate as Ba, Sr, Ti sources and citric acid as complexing reagent. Thermogravimetric analysis (TGA),

differential scanning calorimetry (DSC), infrared (IR) spectroscopy, X-ray diffraction (XRD) and scanning electron microscopy (SEM) were used

to characterize the thermal decomposition behavior, the crystallization process and the particle size and morphology of the calcined powders. The

results indicated that single-phase and well-crystallized Ba1�xSrxTiO3 (x = 0.3) nanopowders with particle size around 80 nm could be obtained

after calcining the dried gel at 950 8C for 2 h.

# 2006 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: Ba1�xSrxTiO3; Sol–gel synthesis; Nanopowders

www.elsevier.com/locate/ceramint

Ceramics International xxx (2006) xxx–xxx

1. Introduction

Ba1�xSrxTiO3 (BST) ceramics, one of the most interesting

materials due to its pyroelectric, ferroelectric and piezoelectric

characteristics, has been widely used in the preparation of high

dielectric capacitors, PTC resistors, transducers and ferro-

electric memories [1–4]. BST powders are usually prepared by

solid-state reaction [5,6] with calcination temperatures in the

range of 1300–1450 8C. Other preparation routes, including the

hydrothermal method [7], the precipitation method [8], the

metalorganic solution deposition (MOSD) technique [9] and

the sol–gel method [10,11], have been reported. It is widely

recognized that chemical synthesis methods are able to provide

high chemical purity, precise composition, uniform micro-

structure and a lower formation temperature of the perovskite

phase based on molecular scale mixing in the preparation of the

precursor [12]. The sol–gel route is proved to be a very

commonly applied chemical method for fabricating uniform

large area thin films and synthesizing powders of micrometer,

sub-micrometer, or nanometer size with high purity and

* Corresponding author. Tel.: +86 27 87544432; fax: +86 27 87544532.

E-mail address: taoli@mail.hust.edu.cn (T. Li).

0272-8842/$32.00 # 2006 Elsevier Ltd and Techna Group S.r.l. All rights reserve

doi:10.1016/j.ceramint.2006.03.015

Please cite this article in press as: Z. Wang et al., Synthesis and charac

Ceram. Int. (2006), doi:10.1016/j.ceramint.2006.03.015

homogeneity [13]. However, the sol–gel process based on

hydrolysis of metal alkoxides needs long refluxing time.

Furthermore, the expensive and unstable metal alkoxides or

mixtures of metal salts and metal alkoxides must be dissolved

in large quantities of organic solvents, implying a high cost. As

an alternative to methods reported previously, Liu and co-

workers [14] synthesized BST powder using a citrate gel

method firstly. The process has the advantages of simplicity and

no use of any contaminative and expensive reagents.

Furthermore, the morphology of the particles could be easily

controlled by changing the synthesis parameters.

In this paper, we focus on a novel sol–gel synthesis method

of BST nanopowders involving a new stable aqueous precursor

based on a metal–citrate complex system reported recently. Our

method herein offers an easy route to shorten the procedures,

save energy, easily control the final stoichiometry of BST

composition and lead to uniform, fine powders.

2. Experimental details

Ba1�xSrxTiO3 (x = 0.3) powders were synthesized by a

complex precursor route as summarized in Fig. 1. Citric acid

(99.5+%), tetra-n-butyl titanate, Ti(n-OC4H9)4 (98+%), H2O2

d.

terization of Ba1�xSrxTiO3 nanopowders by citric acid gel method,

Fig. 1. Flowchart for the preparation of BST powders.

Z. Wang et al. / Ceramics International xxx (2006) xxx–xxx2

+ Models

CERI-2415; No of Pages 5

(30 wt.% in H2O), Ba(NO3)2 (99.5+%), Sr(NO3)2 (99.5+%) and

ammonia solution (25–28 wt.% in H2O) were employed as the

starting ingredients.

Ti(n-OC4H9)4 was added to stirred deionized water, the

precipitation and turbidity were observed gradually, then the

precipitate was filtered and washed for several times with

deionized water. Citric acid and H2O2, both in a 2:1 molar ratio

against Ti(n-OC4H9)4, were added. With the addition of these

ingredients, the solution became orange in color and viscous. The

pH value of the solution was adjusted to 6 using ammonia

solution. A water-soluble crimson precursor for Ti(IV) was

prepared with continuous stirring and refluxing at 60 8C. After

stoichiometric amounts of Ba(NO3)2 and Sr(NO3)2 were

dissolved in citric acid solution (5 M) of pH 6 adjusted by

ammonia solution with molar ratio of CA:M=2:1 (CA, citric acid,

M, Ba, Sr), Ba(II) and Sr(II) precursors were mixed with Ti(IV)

precursor. During the process, the cations molar ratio of Ba:Sr:Ti

was 0.7:0.3:1. The whole mixture was refluxed at 60 8C for 1–2 h

with continuous stirring and then a porous yellow resin was

formed after drying in a furnace at 60 8C overnight under flowing

air. This solid resin precursor was pulverized and then calcined at

various temperatures for 2 h to obtain the BST powders.

The FT-IR spectrum was recorded with an EQUINOX55

(BRUKER) spectrometer by using KBr pellet. The XRD

patterns of the powders were recorded on a Model x’Pert PRO

of PANalytical B.V. diffractometer using Cu Ka radiation

(l = 1.5406 A) in the range from 108 to 808 (2u) to examine the

crystallization and structural development of BST powders.

Thermogravimetric analysis (TGA) and differential scanning

calorimetry (DSC) were performed on a TG/DSC apparatus

Please cite this article in press as: Z. Wang et al., Synthesis and charac

Ceram. Int. (2006), doi:10.1016/j.ceramint.2006.03.015

(Model NETZSCH STA 409 PC) with a heating rate of 10 8C/

min under flowing dry air. The particle size and morphology of

the samples were determined with a scanning electron

microscope (SEM, Model FEI Sirion 200).

3. Results and discussion

The synthesis method is an easy way of obtaining an

aqueous water-soluble precursor without large quantities of

organic solvents used in the synthesis process. In general, the

process consists of three major steps [15]: (1) the metal cations

react with carboxylate ligands to form metal carboxylate

complexes; (2) the formation of metal carboxylate gel by cross-

linking metal carboxylate complexes into a three-dimensional

structure; (3) the decomposition of the precursor gel at

moderate temperature. Note that the compositional homo-

geneity of the final BST powders is no longer ensured at

molecular level mixing once insoluble precipitates have formed

during the precursors preparation. Therefore, a key to success in

the preparation is to find out an experimental condition,

wherein the formation of insoluble precipitates does not occur

throughout sol and gel preparation, so that powders with high

chemical purity and precise composition can be synthesized.

The flowchart presented in Fig. 1 is based on the synthesis of

Ba(II) precursor and Sr(II) precursor reported by Liu and co-

workers [14], however, the preparation of Ti(IV) precursor is

modified: an easy way to synthesize clear Ti(IV) precursor by

the addition of H2O2 and ammonia, which can effectively avoid

the separation of an oil-like two-layer liquid to get Ti(IV)–

citrate complex in the works of Xu et al. [16] and Shen et al.

[17]. Moreover, the formation of a porous drying gel at lower

temperature in furnace does not need higher temperature

(120 8C in vacuum [17]), which is significant in potential

industrial application. During the synthesis process, when the

molar ratio of citric acid to the Ba2+, Sr2+ metal cations was less

than 1.5, white precipitates were observed in the sol due to

relatively low solubility of barium nitrate and strontium nitrate

without insufficient complexation with citric acid. Further-

more, the pH value of the precursor solution is also important

to obtain a clear gel without precipitation. Considering

the stability constants for BaH2Cit+ ðlog10KðBaH2CitþÞ=ðmol�1 dm3Þ = 0.6) and BaCit� ðlog10KðBaCit�Þ=ðmol�1 dm3Þ= 2.95) [14], it was appropriate that the optimum pH value

range was 6.0–7.0.

Fig. 2 shows the TGA and DSC curves of the Ba1�xSrxTiO3

(x = 0.3) precursor. The TG curve shows a drastic weight loss at

162–219 8C, followed by two small weight loss at 219–305 8Cand 305–462 8C; there is a big weight loss in the temperature

range 462–491 8C and no further weight loss beyond 619 8C.

The endothermic peak at 187.4 8C is due to the evaporation of

volatile components, corresponding to a weight loss of 27.6%.

Two successive faint exothermic peaks at 252.9 and 292.2 8Care caused by the burnout of organic species, corresponding to a

total weight loss of 16.2%. Most of organic components are

removed below 488 8C. Exothermic peaks at 479.1 8C,

accompanied by a weight loss of 15.7%, might be due to the

thermal decomposition of citrate complex [16].

terization of Ba1�xSrxTiO3 nanopowders by citric acid gel method,

Fig. 2. Combined TG-DSC curves of uncalcined BST precursor with x = 0.3.

Fig. 4. FT-IR spectra of the gel (a) and the powders calcined at (b) 600 8C; (c)

800 8C and (d) 950 8C.

Z. Wang et al. / Ceramics International xxx (2006) xxx–xxx 3

+ Models

CERI-2415; No of Pages 5

Fig. 3 shows the FT-IR spectra of the Ba1�xSrxTiO3 (x = 0.3)

precursor. For the precursor with CA/M = 1.0 and pH 6.0, two

bands, possibly the asymmetric and symmetric stretching

vibrations of carboxylate [18], are observed at 1581 and

1398 cm�1. For the precursor with CA/M = 2.0 and pH 6.0,

there are two strong absorption peaks at 1574 and 1400 cm�1,

corresponding to the asymmetric stretching mode and the

symmetric stretching mode of COO� for a bridging complex,

respectively [19]. The band at 1078 cm�1 is observed, which is

attributed to the organic network or the COH groups [20]. The

broad absorption band around 3400 cm�1 is due to the O–H

stretching modes for the intermolecular hydrogen bond or

molecular water. The FT-IR spectrum of Ba1�xSrxTiO3

(x = 0.3) precursor with CA/M = 1.5 and pH 6.0 is similar to

that of the precursor with CA/M = 2.0 and pH 6.0. The

characteristic intensity increased for the metal carboxylate

complexes (Fig. 3c) can be ascribed to the enhanced interaction

between metal cations and carboxylate ligands. Fig. 4 is the FT-

IR spectra of Ba1�xSrxTiO3 (x = 0.3) gel powders heat-treated

at various temperatures, showing the evolution of the materials

Fig. 3. FT-IR spectra of (a) CA/M = 1; (b) CA/M = 1.5; (c) CA/M = 2.0 in the

condition of pH 6.0.

Please cite this article in press as: Z. Wang et al., Synthesis and charac

Ceram. Int. (2006), doi:10.1016/j.ceramint.2006.03.015

with the increase in calcinations temperature. For the spectrum

of gel powders, the strong and sharp absorption bands at 1399

and 1572 cm�1 may be ascribed to the nsym(COO�) and

nasym(COO�) of acetyl groups, respectively. Broadened

absorption band at 3440 cm�1 associates with O–H stretch

of intermolecular hydrogen bonds or molecular water. The

HOH bending mode characteristic of water appears at

1630 cm�1. After firing more than 800 8C, the band intensity

of the carbonate group at 1443 cm�1 was reduced significantly,

but a large broad band appeared at 555 cm�1. Music et al. [21]

have assigned that this band is caused by the metal-oxygen

stretching vibration of Ti–O bond. It revealed that the formation

of a large amount of BST.

The phase formation starting from the BST precursor was

studied by means of XRD. Fig. 5 shows the XRD patterns of

the Ba1�xSrxTiO3 precursor with x = 0.3 and powders

calcined at various temperatures. Fig. 5a shows the precursor

to be amorphous. Some small diffraction peaks appeared for

the powders calcined at 500 8C, indicating that incipient

crystallization starts from around 500 8C. The perovskite-like

phase crystallization occurred at 600 8C. However, there was

still an impurity phase of barium carbonate (JCPDS 01-0506)

with the diffraction peaks at 24.08 and 34.28 (2u 8). The

intensity of the diffraction peaks of BaCO3 decreased with the

increase in calcination temperature, which agreed with FT-IR

results (Fig. 4). At 950 8C, the carbonate impurity phase

disappeared and the tetragonal perovskite monophase of

Ba1�xSrxTiO3 (x = 0.3) was obtained (JCPDS 89-0274).

Obviously, this temperature was much lower than for the

solid-state reaction method [22]. The crystallite size of the

particles calcined at various temperatures could be calculated

by the Scherrer’s equation: D = kl/b cos u, where D is the

crystallite size, k is a constant (0.9, spheres), l is the

wavelength of the X-ray radiation, b is the line width obtained

after correction for the instrumental broadening and u is the

angle of diffraction. The crystallite size obtained from XRD

data (950 8C) is 86 nm.

terization of Ba1�xSrxTiO3 nanopowders by citric acid gel method,

Fig. 6. SEM micrograph of BST powders calcined at 950 8C for 2 h.

Fig. 5. Powder X-ray diffraction patterns of Ba1�xSrxTiO3 (x = 0.3) precursor

(a) and powders calcined at (b) 500 8C; (c) 600 8C; (d) 800 8C; (e) 900 8C and

(f) 950 8C.

Z. Wang et al. / Ceramics International xxx (2006) xxx–xxx4

+ Models

CERI-2415; No of Pages 5

From the SEM image (Fig. 6) of the Ba1�xSrxTiO3 (x = 0.3)

sample calcined at 950 8C, the particle size was estimated to be

around 80 nm, a value consistent with the one calculated by the

XRD data.

4. Conclusions

A novel synthesis process was designed and stoichiometric

and monophasic Ba1�xSrxTiO3 (x = 0.3) nanopowders were

Please cite this article in press as: Z. Wang et al., Synthesis and charac

Ceram. Int. (2006), doi:10.1016/j.ceramint.2006.03.015

successfully prepared by the citric acid gel method. The process

is an easy way for obtaining an aqueous water-soluble

precursor. Single-phase and well-crystallized Ba1�xSrxTiO3

(x = 0.3) nanopowders could be synthesized at 950 8C with

particle size of around 80 nm.

Acknowledgements

This work was supported by the National 973 Project of

China (2004CB619300) and the HUST Postgraduate Innova-

tion Foundation (2005).

References

[1] P.K. Sharma, V.V. Varadan, V.K. Varadan, Porous behavior and dielectric

properties of barium strontium titanate synthesized by sol–gel method

in the presence of triethanolamine, Chem. Mater. 12 (2000) 2590–

2596.

[2] I.P. Selvam, V. Kumar, Synthesis of nanopowders of (Ba1�xSrx)TiO3,

Mater. Lett. 56 (2002) 1089–1092.

[3] B. Wu, L. Zhang, X. Yao, Low temperature sintering of BaxSr1�xTiO3

glass-ceramic, Ceram. Int. 30 (2004) 1757–1761.

[4] X. Li, F. Qiu, K. Guo, B. Zou, J. Gu, J. Wang, B. Xu, Synthesis and

humidity sensitive properties of nanocrystalline Ba1�xSrxTiO3 thick films,

Mater. Chem. Phys. 50 (1997) 227–232.

[5] O.P. Thakur, C. Prakash, D.K. Agrawal, Microwave synthesis and sinter-

ing of Ba0.95Sr0.05TiO3, Mater. Lett. 56 (2002) 970–973.

[6] O.P. Thakur, C. Prakash, D.K. Agrawal, Dielectric behavior of

Ba0.95Sr0.05TiO3 ceramics sintered by microwave, Mater. Sci. Eng. B

96 (2002) 221–225.

[7] M. Niederberger, G. Garnweitner, N. Pinna, M. Antonietti, Nonaqueous

and hilide-free route to crystalline BaTiO3, SrTiO3, and (Ba,Sr)TiO3

nanoparticles via a mechanism involving C–C bond formation, J. Am.

Chem. Soc. 126 (2004) 9120–9126.

[8] Y.B. Khollam, S.V. Bhoraskar, S.B. Deshpande, H.S. Potdar, N.R.

Pavaskar, S.R. Sainkar, S.K. Date, Simple chemical route for the

quantitative precipitation of barium-strontium titanyl oxalate precursor

leading to Ba1�xSrxTiO3 powders, Mater. Lett. 57 (2003) 1871–

1879.

[9] M.W. Cole, R.G. Geyer, The dependence of dielectric properties on

compositional variation for tunable device applications, Mech. Mater.

36 (2004) 1017–1026.

[10] H. Shiibashi, H. Matsuda, M. Kuwabara, Low-temperature preparation of

(Ba,Sr)TiO3 perovskite phase by sol–gel method, J. Sol–Gel Sci. Technol.

16 (1999) 129–134.

[11] T. Hu, H. Jantunen, A. Uusimaki, S. Leppavuori, BST powder with sol–gel

process in tape casting and firing, J. Eur. Ceram. Soc. 24 (2004) 1111–

1116.

[12] A. Hardy, D. Mondelaers, M.K. Van Bael, J. Mullens, L.C. Van Poucke, G.

Vanhoyland, J. D’Haen, Synthesis of (Bi,La)4Ti3O12 by a new aqueous

solution-gel route, J. Eur. Ceram. Soc. 24 (2004) 905–909.

[13] C. Liu, B. Zou, A.J. Rondinone, Z.J. Zhang, Sol–gel synthesis of free-

standing ferroelectric lead zirconate titanate nanoparticles, J. Am. Chem.

Soc. 123 (2001) 4344–4345.

[14] C. Shen, Q.F. Liu, Q. Liu, Sol–gel synthesis and spark plasma sintering of

Ba0.5Sr0.5TiO3, Mater. Lett. 58 (2004) 2302–2305.

[15] Y. Narendar, G.L. Messing, Mechanisms of phase separation in gel-based

synthesis of multicomponent metal oxides, Catal. Today 35 (1997) 247–

268.

[16] Q. Xu, S. Chen, W. Chen, S. Wu, J. Zhou, H. Sun, Y. Li, Synthesis and

piezoelectric and ferroelectric properties of (Na0.5Bi0.5)1�xBaxTiO3 cera-

mics, Mater. Chem. Phys. 90 (2005) 111–115.

[17] C. Shen, Q.F. Liu, Q. Liu, Ba0.5Sr0.5TiO3 dielectric materials prepared by

citrate-nitrate combustion, J. Inorg. Mater. 19 (2004) 681–685.

terization of Ba1�xSrxTiO3 nanopowders by citric acid gel method,

Z. Wang et al. / Ceramics International xxx (2006) xxx–xxx 5

+ Models

CERI-2415; No of Pages 5

[18] R.A. Nyquist, R.O. Kagel, Infrared Spectra of Inorganic Compounds,

Academic Press, Inc., New York, 1971.

[19] W. Cao, Y. Xu, S. Wang, P. Lu, G. Huang, C. Xu, Preparation of

La(Zn0.5Ti0.5)O3 powders via citric acid precursor, Mater. Lett. 59

(2005) 1914–1918.

[20] Y. Xu, P. Lu, G. Huang, C. Zeng, Synthesis of La(Ma1/2Ti1/2)O3 via citric

acid precursor, Mater. Chem. Phys. 92 (2005) 220–224.

Please cite this article in press as: Z. Wang et al., Synthesis and charac

Ceram. Int. (2006), doi:10.1016/j.ceramint.2006.03.015

[21] S. Music, M. Gotic, M. Ivanda, S. Popovic, A. Turkovic, R. Trojko, A.

Sekulic, K. Furic, Chemical and microstructural properties of

TiO2 synthesized by sol–gel procedure, Mater. Sci. Eng. B 47 (1997)

33–40.

[22] F.D. Flaviis, N.G. Alexopoulos, O.M. Stafsudd, Planar microwave inte-

grated phase-shifter design with high purity ferroelectric material, IEEE

Trans. Microw. Theory Tech. 45 (1997) 963–969.

terization of Ba1�xSrxTiO3 nanopowders by citric acid gel method,