Poly(ε-caprolactone)-based nanocomposites: Influence of compatibilization on properties of poly(ε-caprolactone)–silica nanocomposites

COMPOSITES

www.elsevier.com/locate/compscitech

Composites Science and Technology 66 (2006) 886–894

SCIENCE ANDTECHNOLOGY

Poly(e-caprolactone)-based nanocomposites: Influence ofcompatibilization on properties of poly(e-caprolactone)–silica

nanocomposites

Maurizio Avella a, Federica Bondioli b, Valeria Cannillo b, Emilia Di Pace a,Maria Emanuela Errico a,*, Anna Maria Ferrari c,

Bonaventura Focher b, Mario Malinconico a

a Institute of Chemistry and Technology of Polymers (ICTP)-CNR, Via Campi Flegrei, 34 c/o Comprensorio Olivetti, 80072 Pozzuoli (NA), Italyb University of Modena and Reggio Emilia, Department of Materials and Environmental Engineering, Via Vignolese 905/a, 41100 Modena, Italy

c University of Modena and Reggio Emilia, Department of Engineering Science and Methods, Viale Allegri, 15 I-42100 Reggio Emilia, Italy

Received 29 April 2005; received in revised form 26 August 2005; accepted 28 August 2005Available online 12 October 2005

Abstract

In the present paper, results about preparation and characterization of poly(e-caprolactone) (PCL) based nanocomposites filledwith silica nanoparticles are reported. In order to promote polymer/inorganic nanofiller compatibility and to increase the interfacialadhesion between the two components, silica nanoparticles surface has been functionalised by grafting a Mw = 10,000 Da PCL onto it.Successively, PCL based nanocomposites have been prepared by extrusion process. The relationships among size, amount of the nano-filler, organic coating and the final properties have been investigated. The morphological analysis has revealed that the silica function-alization can provide a useful method of preparation of the nanocomposites with the achievement of a fine, a good dispersion and astrong adhesion level. Thermal characterization has shown an improved thermal stability due to the presence of the silica nanopar-ticles, especially in the case of modified nanofillers. Finally mechanical tests revealed an increase of the Young�s modulus in the PCLbased nanocomposites.� 2005 Elsevier Ltd. All rights reserved.

Keywords: Polymers; Hybrid compounds; Coupling agent; Interface; Mechanical properties

1. Introduction

The recent developments of new materials based onnanometer sized filler particles in polymeric matrices repre-sent a radical alternative to conventional-filled polymers orpolymer blends resulting thus in a disruptive change incomposite technology.

Polymeric nanocomposites combine the excellent flexi-bility, low density and easy processability of polymers withhigh strength, rigidity, heat resistance of inorganic materi-

0266-3538/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.compscitech.2005.08.014

* Corresponding author. Tel.: +39 818675061; fax: +39 818675230.E-mail address: [email protected] (M.E. Errico).

als, and may become the most important and practicalmaterials [1–8].

Uniform dispersion of these nanoscopically sized fillerparticles produces ultra-large interfacial region per volumebetween the nanoelement and host polymer due to their highspecific surface area. This enormous interfacial region repre-sents the peculiar characteristic of the polymer based nano-structured materials and differentiates them from traditionalcomposites and filled plastics containing micrometric rein-forcements and fillers in the forms of fibres or powders(whiskers, talc, mica, calcium carbonate, etc.) [9,10].

In particular an interphase layer forms at the interfacebetween organic and inorganic phases. The interphase

M. Avella et al. / Composites Science and Technology 66 (2006) 886–894 887

polymer layer near the inorganic surface dramatically dif-fers from the bulk polymer [11]. Because of this interphaselayer, nanocomposites show enhanced properties with re-spect to constituents bulk properties.

As a result of the large surface area of nanoscopic fillers,significantly smaller amounts of the particles (1–6% byweight) compared to conventional composites (up to 60%by volume) can induce dramatic changes in host matrixproperties [12,13]. In this way, it is possible to preparematerials characterized by a lower density and a betterprocessability with respect to the use of conventionalmicrofiller.

The different nature of the nanofiller (inorganic) with re-spect to that of the polymeric matrix (organic), and theirhigh adsorption surface energies are responsible for a strongnanofillers tendency to form aggregates. At the same time,enhanced performances nanomaterials are correlated to anhomogeneous nanoparticles dispersion and to a stronginterfacial adhesion between filler/polymer. Hence, moreversatile synthetic and/or preparative approaches areneeded to obtain polymer based nanocomposites with con-trolled composition and microstructures. The key to any ofthese fabrication processes is the engineering of the poly-mer–nanoparticle interface.

Developing an understanding of the characteristics ofthis interphase region, its dependence on nanoelement sur-face chemistry, the relative arrangement of constituentsand, ultimately, its relationship to the nanomaterials prop-erties, is a current research frontier in nanocomposites.Equally important is the development of a general under-standing of the morphology-property relationships formechanical, barrier, and thermal response of these systems.As matter of fact, researchers currently focus most effortson developing interfacial tailoring and on compatibiliza-tion promotion [14–19].

In the present research, results about PCL based nano-composites filled with silica nanoparticles are reported.PCL is a synthetic semicrystalline polymer, which is char-acterized by low glass transition temperature (Tg), lowmodulus, high elongation at break and, overall, it is biode-gradable. PCL is frequently found as a component instarch based formulations of biodegradable commodityfilms for packaging, but it is also proposed for nursery potsand transplantation bags [20]. In the area of biomedical de-vices, it is studied for suture filaments and as a componentin polylactide based blends, where its resistance to hydroly-sis, compared to Poly(lactic acid) (PLA), allows a longerpermanence of the article within the body [21,22]. The maindisadvantages of PCL reside in its low melting temperature(Tm � 60 �C) and, mainly, in its low modulus and abrasionresistance. The addition of silica nanoparticles should, inprinciple, be beneficial for all the above aspects, providedthat an efficient control of the interfacial tensions in thecomposite formation is achieved [23]. In the present paper,a preliminary investigation on the chemistry of PCL graft-ing onto silica nanoparticles and on the influence of addi-tion of modified nanoparticles in high molecular weight

PCL matrix (Mw = 60,000 Da) on thermal, mechanicaland morphological behaviour of nanocomposites isreported.

2. Experimental

2.1. Materials

Poly(e-caprolactone) [molecular weight Mw = 60,000 Da]has been supplied by Solvay. Poly(e-caprolactone) [molec-ular weight Mw = 10,000 Da] has been supplied by Poly-Sciences, Inc. c-aminopropyltriethoxysilane (APTEOS),tetraethoxysilane (TEOS), Aldrich reagent-grade product,and hexamethylendiisocyanate (HMDI), Fluka reagentgrade product, have been used without further purification.

2.2. Preparation of SiO2 nanoparticles

Monodispersed silica particles have been prepared bythe hydrolysis and polycondensation reaction of tetraeth-oxysilane (TEOS). The reaction has been carried out in atightly screw-capped container by adding TEOS solution0.025 M to a EtOH solution containing NH4OH 0.5Mand H2O 8 M at 40 �C. After an equilibration time of10 min, the suspension has been cooled, filtered, washedand dried in a vacuum oven at 100 �C for 24 h.

2.3. Modification of SiO2 nanoparticles surface

Four hundred mg of c-aminopropyltriethoxysilane (20%by weight of silica) have been dissolved in 40 ml of waterand the aqueous solution (1% wt/wt) has been kept at roomtemperature for 2 h under mechanical mixing to hydrolysisreaction. A fine dispersion of 2 g of silica nanoparticles into50 ml of ethanol has been added to this solution and kept at70 �C for 24 h under stirring. After the reaction, the treatedsilica nanoparticles have been dried at 80 �C.

2.4. Preparation of coupling agent (SiO2 CA)

Two g of silica nanoparticles modified with aminosilanehave been dispersed into 40 ml of chloroform. To this dis-persion, 1% by weight solution of HMDI in chloroformhas been added with a separator funnel and kept for 24 hat room temperature under mixing. The amount of HMDIin mol is the same of c-aminopropyltriethoxysilane,(2.23 · 10�3 mol).

This dispersion has been added dropwise to a solutioncontaining 22 g of hydroxyl terminated PCL (Mw =10,000 Da) dissolved in chloroform (HMDI/PCL molar ra-tio 1:1); the mixture was kept for 24 h at reflux under mixing.

2.5. Preparation of PCL/SiO2 nanocomposites

PCL based nanocomposites have been prepared bymixing PCL (Mw = 60,000 Da) and SiO2-CA into a

888 M. Avella et al. / Composites Science and Technology 66 (2006) 886–894

miniextruder operating at 110 �C and 32 rpm. The saidminiextruder (SCAMIA) is a single screw extruder (L/D = 20) with a particular screw design that allows an inti-mate mixing of the components. PCL filled with 1% and2.5% by weight of SiO2-CA have been prepared. To com-parison a PCL filled with 2.5% by weight of non-modifiedsilica has been also prepared in the same experimentalconditions.

3. Techniques

3.1. FT-IR analysis

The infrared spectra have been obtained by using a Per-kin–Elmer Paragon 2000 FT-IR spectrometer. The sampleshave been investigated both in pressed disc mixed withpowdered KBr and on deposited films from solution. Thespectra have been recorded at room temperature using 64scans, 2 cm�1 resolution.

3.2. Transmission electron microscopy (TEM)

Morphological analysis of the prepared silica nanoparti-cles has been carried out by TEM. Samples for TEM wereprepared by dispersing the obtained powders in ethanoland then placing a drop of suspension on one side of thetransparent polymer coated 200 mesh copper grid. Sampleswere successively coated by carbon to increase the thermaland the electric conductivities of powders. TEM imageshave been taken with a Model JEM 2010 (Jeol, Tokyo,Japan) instrument equipped with an EDS analyser.

3.3. Scanning electron microscopy (SEM)

Morphological analysis of compression-molded sam-ples, obtained in a common heated press operating atT = 100 �C, has been carried out by a SEM Philips XL20 series microscope. The samples have been kept in liquidnitrogen for 5 min and fractured. Before the electronmicroscopy observation, the surfaces have been coatedwith Au–Pd alloy with a SEM coating device (SEM Coat-ing Unit E5 150-Polaron Equipment Ltd.).

3.4. Differential scanning calorimetry (DSC)

The thermal properties have been measured with a dif-ferential scanning calorimeter Mettler DSC-30. The appa-ratus has been calibrated with pure indium, lead and zincstandards at various scanning rates. Each sample has beenheated from 30 to 150 �C at a scanning rate of 10 �C/min,cooled to �100 �C at a scanning rate of 10 �C/min and thenre-heated to 150 �C at 10 �C/min.

The melting temperatures (Tm) and the crystallizationtemperatures (Tc) have been measured at the maxima ofthe endothermic and exothermic peaks of the DSC curves,respectively. The crystalline fractions (Xc) have been calcu-lated by integration of the melting endotherms, using the

literature data for the enthalpy of fusion of PCL in the fullycrystalline state of 146 J/g.

3.5. Thermogravimetrical analysis (TGA)

The thermal stability of the samples has been evaluatedby means of thermogravimetrical analysis (TG) with a TC10 A Mettler TG equipped with a M3 analytical thermo-balance, by recording the weight loss as a function of tem-perature. Each sample has been heated from 40 to 700 �Cat a scanning rate of 20 �C/min in air atmosphere.

3.6. Dynamic-mechanical analysis (DMTA)

Dynamic-mechanical data have been collected at 1 Hz ata heating rate of 3 �C/min from �100 to 50 �C under nitro-gen with a Dynamic Mechanical Thermal Analyser MKIII, Polymer Laboratories, configured for automatic dataacquisition. The experiments have been performed in bend-ing mode.

3.7. Tensile tests

Tensile tests have been performed by using an Instronmachine on dumb-bell specimens, having thickness of1 mm, prepared in a common heated press operating atT = 100 �C. The deformation rate was of 10 mm/min.The procedure used for the calculation is in accordanceto the ASTM (D256) standard, (average of 10 samplestested).

4. Results and discussion

4.1. Preparation of compatibilizing agent

In the present paper results about PCL based nanocom-posites filled with silica nanopowders are reported.

Monodispersed silica particles have been prepared bythe hydrolysis and polycondensation reaction of tetraeth-oxysilane (TEOS) according to the Stober et al. method[24].

The dried particles have been examined by transmissionelectron microscopy (TEM) and X-ray diffraction analysis.TEM analysis has revealed that the silica nanoparticles arenearly perfect with no agglomerated spheres and with anaverage diameter of 100–200 nm (Fig. 1). X-Ray analysishas shown that silica nanopowders are almost amorphous.

PCL-based nanocomposites have been prepared by atwo steps preparation methodology: in a first step silicananoparticles are reacted with PCL by a surface graftingreaction, then in a second step a melt mixing process be-tween modified silica and PCL is performed.

Generally speaking, the grafting of polymers onto thenanoparticles surface as well on classical microparticles,could remarkably improve the dispersion of the ultrafineparticles in solvents and in polymer matrices. This is con-sidered to be due to the fact that grafted chains on the sur-

M. Avella et al. / Composites Science and Technology 66 (2006) 886–894 889

face obstacle particles aggregation by increasing the affinityof particle surface towards organic solvents and/or poly-meric matrices. Moreover, the presence of organic mole-cules onto the surface also promotes polymer/inorganicnanofiller compatibility increasing in this way the interfa-cial adhesion between the two nanocomposite components.

For these reasons, silica nanoparticles surface has beenfunctionalised by grafting a moderate molecular weight hy-droxyl terminated PCL (Mw = �10,000 Da) onto it. Inparticular c-aminopropyltriethoxysilane (APTEOS) hasbeen used as coupling agent in order to introduce func-tional organic molecules onto silica nanopowders, beingknown the reactivity of the silica silanol groups with the si-lane coupling agent after the hydrolysis of the alkoxygroups of APTEOS has occured.

The amount of silane used is about 20% by weight withrespect to dried silica nanoparticles. This percentage hasbeen chosen on the basis of previous experience, where itwas demonstrated that the amount of grafted silane in-creases with increasing amount of silane used and then lev-els off [25].

The reaction has been monitored by IR spectroscopy.The FT-IR spectra of silica nanopowders and of treated sil-ica have been compared. The relative ratio between the

intensity of the band at �3430 cm�1 (stretching of the silicahydroxyl groups) and that of the peak at �1100 cm�1

(stretching of Si–O–Si bond) decreases in the silane-modi-fied silica spectrum, indicating that the above describedreaction occurs. It has been estimated that about 32% ofthe hydroxyl groups have reacted and that there are about1.5 silanol groups per molecule of APTEOS.

The silane modified silica have been successively treatedwith HMDI in order to build up bonded polyester chainsonto silica.

The reaction has been carried out at 70 �C for 24 h inchloroform and the molar amount of HMDI added hasbeen the same of APTEOS one. In this way, it has beenpossible to obtain an isocyanate-end capped modified sil-ica, as following reported.

In order to avoid possible cross-links formation, theHMDI solution has been added dropwise in the solutionof aminated silica.

The carbonyl group of the HMDI is particularly suscep-tible to a nucleophilic addition of the terminal aminegroups present onto silica surface. The above reportedreaction is a typical hydrogen displacement mechanismwith formation of the ureic bond.

The infrared spectra of the reaction product show anabsorption band around 1690 cm�1, corresponding to thestretching of the –C–O belonging to the (NH–CO–NH)ureic groups, an absorption band at 2273 cm�1 corre-sponding to the stretching of the residual terminal isocya-nate groups, and the absorption bands corresponding tothe bending of the N–H ureic group (1570–1630 cm�1)and to the stretching of C–N group (1400 cm�1), Fig. 2.

Finally, the solution containing the isocyanate-endcapped modified silica has been added dropwise to a lowmolecular weight hydroxyl end-capped PCL (Mw =10,000 Da) in chloroform. Reaction of isocyanate with hy-droxyl is a typical reaction of polyurethane chemistry. Inthis way, a chemical bond between silica nanoparticlesand polyester has been created obtaining, in this way aproper coupling agent as following illustrated.

The dropwise addition avoids the formation of PCLbridges on two silica particles. PCL of Mw = 10,000 Dahas been used because higher values of molecular weightwould give some kinetic problems related to the efficiencyof coupling reactions, while for lower values of molecularweight the compatibilization process to the polymeric ma-trix could be unfavourable according to thermodynamicconsiderations; infact, the PCL homopolymer used inblend has a Mw = 60,000 Da.

The above reaction has been followed by IR spectros-copy and DSC analysis.

The IR spectra show the absorption bands attributableto PCL groups while the isocyanate absorption band at2273 cm�1 disappears thus indicating that the reactionbetween the residual terminal isocyanate groups of the

Fig. 1. TEM micrograph of silica nanoparticles.

Fig. 2. Infrared spectra of APTEOS-silica nanoparticles after reactionwith HMDI.

890 M. Avella et al. / Composites Science and Technology 66 (2006) 886–894

modified silica and the hydroxyl terminal groups of PCLoccurs.

The DSC thermograms of the coupling agent show ashift of the PCL glass transition value of about 9 �C com-pared to neat PCL Mw = 10,000 Da, (Table 1). Moreover,the crystallinity content of neat PCL decreases after the

Table 1Thermal parameters of PCL (Mw = 10,000 Da) and PCL grafted ontosilica nanoparticles: glass transition temperature (Tg), crystallizationtemperature (Tc), melting temperature (Tm) and crystallinity (Xc)

SAMPLE Tg (�C) Tc (�C) Tm (�C) Xc (%)

PCL (Mw � 10,000) �60 29 58 45PCL-g-SiO2 �51 30 58 36

coupling reaction with the silica nanoparticles. This modi-fication of chemico-physical parameters can be attributedto the fact that the polyester chains are anchored to the sil-ica surface thus reducing the PCL chains mobility.

4.2. Preparation and morphology of PCL basednanocomposites

PCL based nanocomposites have been prepared by anextrusion process. In particular materials containing 1%and 2.5% by weight of silica modified nanoparticles havebeen prepared. PCL filled with 2.5% by weight of non mod-ified silica has been also prepared to evaluate the influenceof the organic coating on the structure and properties ofthe nanocomposites.

The process of extrusion has been carried out at 150 �Cin a monoextruder with a mixing screw. Morphological,thermal, and mechanical characterizations have been per-formed on nanocomposites films obtained by compressionmolding.

To evaluate the state of dispersion of silica nanoparticleswithin the polymeric matrix, the nanocomposites sampleswere cryogenically fractured after immersion in liquidnitrogen and the fractured surfaces have been observedby scanning electron microscope (SEM).

In Figs. 3–5, the SEM micrographs of the nanocompos-ites fractured surfaces are shown. From these figures it canbe easily observed that silica particles are covered andfirmly bonded, quite welded, to the PCL matrix. Moreover,the fractured nanocomposites surface appears completelyfilled by the nanoparticles and no pull out phenomenaare evidenced (absence of voids). In the case of the nano-composite containing modified silica, SEM micrographsshow a fine and homogeneous dispersion of the nanoparti-cles into polymeric matrix (see Figs. 3 and 4). In fact a veryhigh number of small discrete nanoparticles is observable.As far as the nanocomposite containing no modified silicais concerned, aggregation phenomena seem to occur. As

Fig. 3. SEM micrograph of compatibilized PCL/SiO2 (1 wt%) fracturedsurface.

Fig. 4. SEM micrograph of compatibilized PCL/SiO2 (2.5 wt%) fracturedsurface.

Fig. 5. SEM micrograph of non-compatibilized PCL/SiO2 (2.5 wt%)fractured surface.

Table 2Thermal parameters of PCL (Mw = 60,000 Da) and PCL based nano-composites: melting temperature (Tm) crystallization temperature (Tc),and crystallinity (Xc)

Sample Tm (�C)II RUN

Tc

(�C)Tc onset–Tc

endset (�C)Xc

(%)

PCL 56 26 33–19 35PCL/SiO2 (2.5%) non-modified 57 24 27–17 28PCL/SiO2 (1%) 57 24 32–20 34PCL/SiO2(2.5%) 58 25 33–18 34

M. Avella et al. / Composites Science and Technology 66 (2006) 886–894 891

shown in Fig. 5, even though the nanofillers are still embed-ded into the matrix, several nanoparticles agglomerates ofabout 1 lm in diameter are visible, together with individualnanoparticles.

This result suggests that the compatibilization promotesstrong interaction between silica surface modifier molecules(low Mw PCL) and polymeric matrix chains decreasing thesilica-silica interactions and as a matter of fact the nano-particles agglomeration phenomena.

Therefore, the morphological analysis has then re-vealed that the silica functionalization can provide auseful method of preparation of the nanocomposites withthe achievement of a fine, good dispersion and strongadhesion levels.

4.3. Properties of PCL based nanocomposites

In order to test the influence of silica nanoparticles andthe effect of their functionalization on PCL properties,

thermal, thermogravimetrical and mechanical analysishave been performed.

Thermal parameters, (melting point, crystallization tem-perature, crystallinity) have been valuated by DSC analy-sis. It has been seen that in presence of 2.5% by weight ofnon modified silica nanoparticles a slight decrease (about7%) in crystallinity content and a slight decrease (about5 �C) in crystallization onset temperature, measured duringthe cooling scan, have been recorded. The same effect hasnot been observed in the compatibilized nanocomposites.Finally, no influence on melting point with compositionhas been recorded, see Table 2.

These results could be explained on the basis of emulsi-fying effect of grafted PCL chains onto nanospheres. Infact, in the compatibilized nanocomposites we have alreadyseen that the dispersion of nanoparticles into the polymericmatrix appears finer and more homogeneous, while in theother case nanoparticles aggregation has been evidenced.As a matter of fact, the dimensions of the individual nano-particles are lower than the growth radius of the PCLspherulite. In the case no aggregation of nanoparticles oc-curs, the overall crystallization process of the polymericmatrix should not be disturbed by the presence of nanopar-ticles. When nanoparticles aggregate, as in the case of notfunctionalised nanosilica, the size of the clusters may belarge enough to obstacle nucleation and growth of PCLspherulites, hence resulting in a lower onset crystallizationtemperature, and in lower crystallinity content.

Mechanical dynamic analysis (DMTA) of nanocompos-ites is reported in Figs. 6. Neat PCL is characterized by adamping in the modulus (E 0) centred at �41.3 �C followedby a quite distinct rubbery plateau until the occurrence ofthe melting process (Tm � 60 �C), Fig. 6(a). The additionof as much as 1% by weight of modified silica nanoparticlesis responsible of an increase in the damping temperature ofabout 9 �C, (from �41 to �32 �C) as consequence of astrong enhancement in the rigidity of the amorphousphase, with the nanofillers working as physical cross-links,Fig. 6(b). In a classical composite with silica microbeadsdispersed in a rubbery matrix, such effect would requireat least 10% by weight of filler. The much larger surfaceto volume ratio available in a nanocomposite is henceresponsible of the increased rigidity of the matrix.

The addition of 2.5% by weight of nanoparticles doesnot lead to any further change in the behaviour of the com-posite so indicating that already 1% by weight of nanofiller

Fig. 6. DMTA curves of: (a) neat PCL, (b) compatibilized PCL/SiO2 (1 wt%), (c) compatibilized PCL/SiO2 (2.5 wt%), (d) non-compatibilized PCL/SiO2

(2.5 wt%).

892 M. Avella et al. / Composites Science and Technology 66 (2006) 886–894

may be sufficient to fill completely the available surface,Fig. 6(c).

The addition of 2.5% by weight of non modified nano-particles leads to similar behaviour, Fig. 6(d). Probably,this occurs because 2.5% by weight is an amount of nano-filler, which is overestimated, thus even in the presence oflarge amounts of agglomerated particles, the behaviourof the nanocomposites level off. On the other hand, wehave already evidenced that non-modified silica nanoparti-cles appear well welded to the polymeric matrix, althoughagglomerated.

Moreover, a second peak (shoulder) in tan d is evi-denced in the DMTA curves, at a temperature lower thanthat of the starting high molecular weight (Mw) PCL, asshown in the Figs. 6(b) and (c). This shoulder is absentin the neat PCL, as well as in the not-compatibilized nano-composite, and can be tentatively attributed to the lowerMw PCL grafted onto the silica surface.

The above findings prompted us in the future, to inves-tigate the behaviour of PCL/SiO2 nanocomposites in thecomposition range of less than 1% by weight.

The mechanical performances of polymer compositesare generally associated to the interfacial strength betweenthe matrix and the reinforcing phase. If the interface be-tween the two components is good, the external load willbe transfer from the polymer matrix to the reinforcementthrough the interface and the mechanical performances ofthe composite increase. On the contrary, if the interface it

is not strong enough, it will become a weak point of thematerial and the mechanical behaviour will result poor.The same concept can be applied in the case of nanocom-posites. Moreover, in this case, the formation of an inter-phase region due to a strong polymeric matrix/nanofillersinteractions drastically affects the properties of the nanom-aterials allowing to reach higher performances with respectnot only to those experimentally shown by microcompos-ites, but also to those theoretically calculated applying ana-lytical equations that are not able to take into account thepresence of the interphase layer [26,27].

In our case, as it can be seen from the values reported inFig. 7, the PCL/SiO2nanocomposites have shown an in-crease of Young�s modulus with respect to that found forPCL homopolymer of about 25%. On the contrary theYoung�s modulus, calculated for the non modified PCL/SiO2 sample is similar to that of neat PCL. These resultscan be easily attributed to the action of compatibilizingagent that allows a better dispersion of nanoparticles witha good interfacial adhesion to the matrix, due to the strongpolymeric matrix/nanoparticles surface modifiersinteractions.

The thermal stability of examined samples was investi-gated by using the technique of thermogravimetric analysis(TGA).

In Table 3, the samples weight loss as function of thetemperature is reported. As it is possible to observe, thepresence of silica nanoparticles raises the starting tempera-

Table 3Thermogravimetrical parameters of PCL and PCL based nanocomposites:decomposition temperatures at different weight loss

Sample Td (start)

(�C)Td (40%)

(�C)Td (50%)

(�C)Td (80%)

(�C)

PCL 343 410 414 428PCL/SiO2 (2.5%) non-modified 354 410 414 430PCL/SiO2 (1%) 355 422 427 440PCL/SiO2 (2.5%) 367 425 430 443

Fig. 7. Effect of silica content and of silica modification on tensilemodulus of PCL based nanocomposites: (a) compatibilized PCL/SiO2,(b) non-compatibilized PCL/SiO2.

M. Avella et al. / Composites Science and Technology 66 (2006) 886–894 893

ture of the degradation process of the material. It can beoutlined that only in the case of compatibilized nanocom-posites this increasing is function of composition and it iskept during all the degradation process.

This improved thermal stability can be attributed to arestricted thermal motion of PCL and to the hindered dif-fusion of volatile decomposition products within the nano-composite due to homogeneous and fine dispersion of silicananoparticles.

5. Conclusions

Nanocomposites based on PCL filled with silica nano-powders. The influence of the addition of an emulsifierhas been studied.

In particular silica nanoparticles has been functionalisedby grafting a moderate molecular weight PCL. Amin-opropylltriethoxysilane (APTEOS), as coupling agent,has been used in order to introduce organic molecules ontosilica nanoparticles surface. Successively PCL nanocom-posites obtained by adding to the PCL matrix the modifiedsilica nanoparticles in an amount of 1% and 2.5% byweight were processed by extrusion. As comparison aPCL filled with 2.5% by weight of non-modified silica hasbeen also prepared.

SEM analysis on fractured surfaces of compatibilizednanocomposites has shown a fine distribution of the nano-particles in the matrix together with a good interfacialadhesion between the two phases, while large aggregation

phenomena are present in the case of material filled withnon-modified silica. The mechanical properties of compat-ibilized nanocomposites display an improvement ofYoung�s modulus attributable to the action of compatibi-lizing agent that allows a good distribution of nanoparti-cles and interfacial adhesion. On the contrary the Youngmodulus calculated for non modified PCL/SIO2 sample isquite similar to that of neat homopolymer.

Moreover also the thermal stability of nanocompositesprepared with modified silica particles is improved, proba-bly owing to the fact that the homogeneous and fine disper-sion of nanoparticles produces a restricted thermal motionof PCL molecules hindering also a fast diffusion of volatiledecomposition products in the material.

In conclusion, we can assess the important role playedby the compatibilized agents to generate nanocompositeswith good performances.

A more detailed investigation on the potential biomedi-cal applications of these materials will be the topic of aforthcoming paper.

Acknowledgements

The authors thank Dr. L. Calandrelli, Mr. G. Orselloand Mr. G. Romano for the unvaluable technicalassistance.

The authors gratefully acknowledge the financial sup-port of ‘‘Centro di Competenza Nuove Tecnologie per leattivita produttive’’ of Regione Campania, Italy.

References

[1] Novak BM. Hybrid nanocomposite materials – between inorganicglasses and organic polymers. Adv Mater 1993;5(6):422–33.

[2] Lu S, Melo M, Zhao J, Pearce EM, Kwei TK. Organic–inorganicpolymeric hybrids involving novel poly (hydroxymethylsiloxane.Macromolecules 1995;28(14):4908–13.

[3] Ahmadi SJ, Huang YD, Li JW. Synthetic routes, properties andfuture applications of polymer-layered silicate nanocomposites. JMater Sci 2004;39(6):1919–25.

[4] Usuki A, Kawasumi M, Kojima Y, Fukushima Y, Okada A,Kurauchi T, et al. Synthesis of nylon 6-clay hybrid. J Mater Res1993;8(5):1179–84.

[5] Zuiderduin WCJ, Westzaan C, Huetink J, Gaymans RJ. Tougheningof polypropylene with calcium carbonate particles. Polymer2003;44(1):261–75.

[6] Lan T, Pinnavia TJ. Clay-reinforced epoxy nanocomposites. ChemMater 1994;6(12):2216–9.

[7] Messersmith PB, Giannellis EP. Synthesis and barrier properties ofpoly(e-caprolactone)-layered silicate nanocomposites. J Polym Sci1995;33(7):1047–57.

[8] Alexandre M, Dubois P. Polymer-layered silicate nanocomposites:preparation, properties and uses of a new class of materials. MaterSci Eng Rep 2000;28(1–2):1–63.

[9] Henglein A. Small-particle research: physicochemical properties ofextremely small colloidal metal and semiconductor particles. ChemRev 1989;89(8):1861–73.

[10] Stucky GD, MacDougall JE. Quantum confinement and host/guestchemistry: probing a new dimension. Science 1990;247:669–71.

[11] Schmidt D, Shah D, Giannelis EP. New advances in polymer/layeredsilicate nanocomposites. Curr Opin Solid State Mater Sci2002;6:205–12.

894 M. Avella et al. / Composites Science and Technology 66 (2006) 886–894

[12] Ulrich DR. Prospects for sol–gel processes. J Non-Cryst Solids1990;121(1-3):465–79.

[13] Avella M, Errico ME, Rimedio R. PVA/PTFE nanocomposites:thermal, mechanical and barrier properties. J Mater Sci2004;39(19):6133–6.

[14] Sanchez C, Ribot F. Chemical design of hybrid organic–inorganicmaterials synthesized via sol–gel. New J Chem 1994;18(10):1007–47.

[15] Avella M, Errico ME, Martuscelli E. Novel PMMA/CaCO3 nano-composites abrasion resistant prepared by an in situ polymerizationprocess. Nanoletters 2001;1(4):213–7.

[16] Di Lorenzo ML, Errico ME, Avella M. Thermal and morphologicalcharacterization of poly (ethylene terephthalate)/calcium carbonatenanocomposites. J Mater Sci 2002;37(11):2351–8.

[17] Glasel HJ, Bauer F, Ernst H, Findeisen M, Hartmann E, LangguthH, et al. Preparation of scratch and abrasion resistant polymericnanocomposites by monomer grafting onto nanoparticles, 2. Char-acterization of radiation-cured polymeric nanocomposites. MacrChem Phys 2000;201(18):2765–70.

[18] Avella M, Cosco S, Errico ME, Rimedio R, Martuscelli E.Recent preparation methodologies to obtain polymer basednanocomposites. Pandalai SG, editor. Recent research develop-ments in materials science, vol. 3. Trivandrum, India: ResearchSign Post; 2003.

[19] Yu YY, Chen CY, Chen WC. Synthesis and characterization oforganic–inorganic hybrid thin films from poly (acrylic) and monodi-spersed colloidal silica. Polymer 2003;44(3):593–601.

[20] Takagi S, Koyama M, Kameyama H, Tokiwa Y. Development ofpolycaprolactone/gelatinized starch blends and their enzymatic deg-radation. In: Doi Y, Fukuda K, editors. Biodegradabile plastics andpolymers (studies in polymer science, 12). New York: ElsevierScience; 1994.

[21] Lewis DH. Controlled release of bioactive agents from lactide/glycolide polymers. Biodegradable polymers as drug delivery systems.In: Langer R, Chasin M, editors. New York: Marcel Dekker Inc;1990.

[22] Pitt CG, Marks TA, Schindler A. Biodegradable drug deliverysystems based on aliphatic polyesters: application of contraceptivesand narcotic antagonists. In: Baker R, editor. Controlled release ofbioactive materials. New York: Acadamic Press; 1980.

[23] Tian D, Dubois P, Grandfils C, Jerome R, Viville P, Lazzaroni R,et al. A Novel biodegradable and biocompatible ceramer preparedby the sol–gel process. Chem Mater 1997;9(4):871–4.

[24] Stober W, Fink A, Bohn EJ. Controlled growth of monodispersesilica spheres in the micron size range. Colloid Interface Sci1968;26(1):62–9.

[25] Bourgeat-Lami E, Espiard P, Guyot A. Poly(ethyl acrylate) latexesencapsulating nanoparticles of silica: 1. Functionalization anddispersion of silica. Polymer 1995;36(23):4385–9.

[26] Barbero EJ. Introduction to composite materials design. NewYork: Taylor & Francis Inc; 1998.

[27] Nielsen LE, Landel RF. Mechanical properties of polymer andcomposites. New York: Marcel Dekker Inc; 1994.