Available online at www.sciencedirect.comCOMPOSITES

www.elsevier.com/locate/compscitech

Composites Science and Technology 68 (2008) 1919–1926

SCIENCE ANDTECHNOLOGY

Shape memory and physical properties ofpoly(ethyl methacrylate)/Na-MMT nanocomposites

prepared by macroazoinitiator intercalated in Na-MMT

Min Seok Kim, Jae Kyung Jun, Han Mo Jeong *

Department of Chemistry, University of Ulsan, San-29, Mugeo-dong, Ulsan 680-749, Republic of Korea

Received 26 January 2007; received in revised form 19 September 2007; accepted 19 December 2007Available online 25 December 2007

Abstract

A macroazoinitiator (MAI) containing a poly(ethylene glycol) (PEG) segment was intercalated in the gallery of sodium montmoril-lonite (Na-MMT) and this intercalated MAI was used in the preparation of Na-MMT/poly(ethyl methacrylate) (PEMA) nanocompos-ites via in situ radical polymerization of ethyl methacrylate. The X-ray diffraction pattern and the morphology observed with atransmission electron microscope showed that Na-MMT intercalated with a PEG segment was heterogeneously dispersed in the polymermatrix. Thus Na-MMT intercalated with a PEG segment effectively enhanced the mechanical properties of PEMA. Shape memorybehavior and rheological properties showed that Na-MMT intercalated with a PEG segment performed its role as a physical crosslinkereffectively even with 1.2 wt% of Na-MMT.� 2007 Elsevier Ltd. All rights reserved.

Keywords: A. Polymer–matrix composites; B. Thermomechanical properties; D. Transmission electron microscopy; D. X-ray diffraction; Shape memory

1. Introduction

Shape memory polymer is a smart material that can beutilized in various applications such as medical devices ormechanical actuators. Compared to shape memory alloysor ceramics, shape memory polymers have advantages oflower density, lower cost, higher deformation, and easycontrol of recovery temperature and color variation [1–10].

Thermoresponsive shape memory polymers generallyconsist of two phases, a thermally reversible phase thatbecomes flexible and can be deformed elastically at temper-atures above a thermal transition temperature (TS) and afixed phase for memorizing the original shape. Becausethe thermally reversible phase should show a large dropin elastic modulus and become flexible upon heating aboveTS, the glass transition temperature (Tg) or melting temper-ature (Tm) are generally utilized as TS in the design of shape

0266-3538/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.compscitech.2007.12.015

* Corresponding author. Tel: +82 52 259 2343; fax: +82 52 259 2348.E-mail address: hmjeong@mail.ulsan.ac.kr (H.M. Jeong).

memory polymers [11–14]. In the case where Tg is used asTS, if the shape memory polymer is deformed at the rub-bery plateau above the Tg and subsequently cooled belowthe Tg under constant strain, the deformed shape is fixedbecause chains can barely move at this glassy state. Then,when the polymer is reheated above the Tg, the originalshape can be recovered by entropy elasticity because, atthis rubbery state, it is elastic and soft [2–4]. In this shaperecovery process, the entanglement of polymer chains withtheir neighbors can do its roles as a fixed phase which mem-orizes the original shape, because it can prevent the irre-versible sliding of deformed polymer chains. However, anunrecovered permanent deformation will increase as thenumber of the deformation at rubbery state is increased,because the polymer chains tend to disentangle and to slipoff each other into new positions [2]. So, cross-linkagepoints, crystalline phase, or glassy phase are utilized asfixed phases in shape memory polymers, because the slip-page or flow of polymer chains should be minimized toenhance the shape memory effect.

1920 M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926

A significant drawback of shape memory polymers istheir low stiffness compared to shape memory alloys orceramics. In macroscale composites of shape memory poly-mers, the mechanical weakness can be improved; however,the shape recovery effect can be damaged by the reinforcingfillers [15,16]. Some papers have shown that mechanicalreinforcing can be achieved in nanocomposites reinforcedwith nano-size fillers without any decrease in their shaperecovery effect [16,17]. In a polynorbornene nanocompos-ite, where the inorganic building block polyhedral oligo-meric silsesquioxane (POSS) was included in the mainchain of organic polymer, Mather et al. observed thatshape recovery was improved compared to pristine poly-norbornene [2,18]. Their result was explained as follows:the POSS groups aggregate to a certain extent, and thisseparated phase of inorganic moieties performs its role asa fixed phase to memorize the original shape because thephysical interactions between inorganic moieties hinderthe irreversible sliding of the organic main chain.

The poly(ethylene glycol) (PEG) segment can be easilyintercalated in the sodium montmorillonite (Na-MMT)gallery. Na-MMT intercalated with PEG (Na-MMT/PEG) has high stability, because intercalated PEG cannotbe replaced by organic compounds having a high affinitytoward Na-MMT. And the maximum amount to be inter-calated is quantitative, about 0.3 g-PEG/g-Na-MMT[19,20].

In this study, we intercalated macroazoinitiator (MAI),which has a PEG segment as shown in the chemicalstructure below, into the gallery of Na-MMT, and theNa-MMT intercalated with MAI was used in the polymer-ization of ethyl methacrylate to obtain nanocomposites.We examined the shape memory effect of the nanocompos-ites along with their mechanical and rheological propertiesbecause we anticipated that the Na-MMT/PEG buildingblock linked to poly(ethyl methacrylate) can perform itsrole as a fixed phase in shape memory behavior as well asa moiety to enhance mechanical properties [21].

C CH2

O

C

CH3

CN

N N C

CH3

CN

CH2 C

O

O CH2CH2 O2 2 m

n

2. Experimental

2.1. Materials

Na-MMT (Southern Clay) was used after drying at 60 �Cunder vacuum for 1 day. Macroazoinitiator (MAI, WakoPure Chemical, VPE-0401) was used as received. TheMAI was the condensation polymer of 4,40-azobis(4-cyano-

pentanoic acid) and PEG whose molecular weight was4000. The molecular weight of MAI was in the range of25,000–40,000 and its azo group content was 0.45 mmol/g. Ethyl methacrylate (EMA, Aldrich) was purified by astandard procedure [22]. Acetonitrile (Aldrich), methanol(Aldrich), dimethylformamide (DMF, Aldrich), and 2,20-azobisisobutyronitrile (AIBN, Aldrich) were used asreceived.

Na-MMT intercalated with MAI (Na-MMT/MAI) wasprepared using an acetonitrile/methanol mixture (1/1 byvolume) as a solvent [19,20]. That is, 3 g of MAI was dis-solved in 100 mL of solvent and the solution was stirredwith 7 g of Na-MMT for 1 day at room temperature. Theintercalated compound was separated with a centrifugeand repeatedly washed with acetonitrile and methanol toremove non-intercalated physisorbed MAI [20]. It was thendried at 25 �C for 48 h under vacuum before use. Theamount of MAI intercalated at the gallery of Na-MMT,determined by thermogravimetry, was 0.22 g-MAI/g-Na-MMT.

The recipes for the preparation of poly(ethyl methacry-late) (PEMA)/Na-MMT nanocomposites are shown inTable 1. The DMF, EMA, Na-MMT/MAI, or AIBN werefed into the reactor and polymerization was carried out at60 �C under a N2 atmosphere for 24 h with stirring by amagnetic bar. The prepared PEMA/Na-MMT nanocom-posites were crushed into powders and dried at 80 �C for12 h under vacuum to remove the low molecular weightcomponents. The characteristics of the nanocompositesare shown in Table 2.

2.2. Measurement

The number average molecular weight ð �MnÞ and weight-average molecular weight ð �MwÞ were measured at 43 �Cwith gel permeation chromatography (GPC, WatersM510) and the results are shown in Table 2. The nanocom-posite was dissolved in tetrahydrofuran (THF) and the

solution was filtered with a 0.45 lm membrane filter beforemeasurement. THF was used as an eluent.

To determine the content of Na-MMT in Na-MMT/MAI or nanocomposite, thermogravimetric analysis(TGA) was carried out with a thermogravimetric analyzer(TA Instruments, TGA 2950) at a heating rate of 10 �C/min under an O2 atmosphere using a 5 mg sample in a plat-inum crucible.

4 8 102θ (º)

(a)

(b)

(c)(d)(e)(f)(g)(h)

7.1

5.0

4.9

5.1

5.2

5.3

5.4

2

Arb

itary

Inte

nsity

6

Fig. 1. XRD patterns of (a) Na-MMT, (b) Na-MMT/MAI, (c) NC-0, (d)NC-1,(e) NC-3, (f) NC-5, (g) NC-7, and (h) NC-9.

Table 2Properties of PEMA/Na-MMT nanocomposites

Designation Polymerizationyield (%)

Molecular weighta Content ofNa-MMTb(wt%)

Tgc

�Mn�Mw

NC-0 93 43,400 120,000 0 55NC-1 88 116,500 229,200 1.2 65NC-3 89 179,400 399,300 3.3 65NC-5 91 234,400 434,200 5.7 68NC-7 90 167,200 328,700 7.3 72NC-9 91 148,300 288,700 9.2 72

a Determined by GPC.b Determined by thermogravimetry.c Determined by DSC.

Table 1Recipes for the preparation of PEMA/Na-MMT nanocomposites

Designation Feed (by weight) Amount of azogroup (mmol/100g-EMA)

EMA DMF Na-MMT/MAI

AIBN

NC-0 10.0 10.0 – 0.05 3.05NC-1 10.0 10.0 0.100/0.022 – 0.05NC-3 10.0 10.0 0.300/0.066 – 0.16NC-5 10.0 10.0 0.500/0.110 – 0.26NC-7 10.0 10.0 0.700/0.154 – 0.37NC-9 10.0 10.0 0.900/0.198 – 0.48

M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926 1921

X-ray diffraction (XRD) patterns were obtained with anX-ray diffractometer (Rigaku, RAD-3 C) using CuKa radi-ation (k = 1.54 A) as the X-ray source. The diffractionangle was scanned from 1� at a rate of 1.2�/min.

The morphology of the nanocomposites was examinedwith a transmission electron microscope (TEM, HitachiH-8100) with an accelerating voltage of 200 kV. TheTEM samples were prepared by putting a polymerizednanocomposite powder into an epoxy capsule and curingthe epoxy at 25 �C for 24 h under vacuum. Then, the curedepoxy containing a nanocomposite was microtomed by adiamond knife into 8 nm-thick slices.

Dynamic mechanical properties were determined using adynamic mechanical thermal analyzer (Rheometry Scien-tific DMTA MK I) with a bending mode at a heating rateof 5 �C/min and 1 Hz. Samples were compression moldedat 160 �C with a pressure of 20 MPa.

Differential scanning calorimetry (DSC) was carried outwith DSC-QA10 (TA Instrument) at a heating and coolingrate of 10 �C/min, with 8.5 mg of sample. The samples,stayed at 180 �C for 1 min in the DSC, was cooled downto 25 �C. The glass transition temperature (Tg) was mea-sured in a subsequent heating scan.

Tensile testing was done using a tensile tester (Oriental,OUT-2). Micro-tensile specimens for the test, which werealso prepared by compression molding, had the dimensionsof 25 mm length, 5 mm width, and 1 mm thickness. Thespecimens were elongated at the rate of 10 mm/min at80 �C.

Rheological properties were measured with a cone andplate fixture of an Advanced Rheometrics Expansion Sys-

tem (Rheometrics) at 160 �C. The test fixtures were pre-heated to the measured temperature, samples weresubsequently charged, and the excess material was trimmedfollowed by conditioning for several minutes to relax theresidual stress. The frequency sweep was done with 15%strain. All of the experiments were carried out in a nitrogenenvironment to avoid oxidative degradation.

3. Results and discussion

3.1. XRD and TEM

The XRD patterns of Na-MMT, Na-MMT/MAI, andPEMA nanocomposites are shown in Fig. 1 where wecan see that Na-MMT has a peak at 2h = 7.1� and Na-MMT/MAI has a peak at 2h = 5.0�. This shows that thebasal plane spacing, d001, calculated by Bragg’s law,d001 = k/2 sinh, was increased from 12.4 A to 17.6 A bythe intercalation of MAI. Taking into account that thethickness of the silicate layer itself is about 9.5 A [20,23],these results show that the interlayer distance, the heightof the gap between the layers of silicate, was 8.1 A whenMAI was intercalated. Many researchers have reportedsimilar results with an interlayer distance of about 8 Awhen PEG, whose molecular weight was higher than2000, was intercalated in the gallery of Na-MMT. Twokinds of models, helical conformation or double layer pla-nar zigzag disposition, were suggested to explain the inter-calated structure [19,20,23,24].

The amount of MAI intercalated in the gallery of Na-MMT, measured by TGA, was 0.22 g-MAI/g-Na-MMT.This value is slightly less than the 0.30 g-PEG/g-Na-MMT which was previously reported as the maximumvalue to be intercalated in the case of Na-MMT/PEGnanocomposites [19]. This shows that the other moieties

Fig. 2. TEM micrographs of PEMA/Na-MMT nanocomposite (NC-5) atvarious magnifications.

1922 M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926

linked to the PEG segment in MAI reduces the intercala-tion of MAI compared to PEG itself.

All of the XRD patterns of PEMA nanocompositesshown in Fig. 1 have a peak at the position around2h = 5�. This suggests that the segment intercalated in thegallery of Na-MMT is mainly the PEG segment originatedfrom MAI and the PEMA segment can hardly diffuse intothe gallery of Na-MMT because the Na-MMT intercalatedwith PEG (Na-MMT/PEG) has high stability due to thehigh affinity between Na-MMT and PEG. However, inFig. 1, we can see that the peak position moves slightlyto a higher angle. Although the peak shift is minor, thisshows that the interlayer distance is reduced when the con-tent of Na-MMT/PEG is increased. This suggests the pos-sibility that the aggregation of Na-MMT/PEG, whichexcludes other segments such as PEMA, develops in thematrix of PEMA more evidently as its content is increased.

The TEM micrographs of a PEMA/Na-MMT nano-composite (NC-5) are shown in Fig. 2 with various magni-fications. Fig. 2a shows that Na-MMT is heterogeneouslydispersed in the matrix of PEMA. However, it is observedin Fig. 2b that partially delaminated silicate layers of Na-MMT are dispersed in the PEMA matrix with substantialcurvature. In the TEM morphology of Fig. 2c observedat higher magnification, we can see ordered face–face layerarrangements of several silicate layers. These results sug-gest that small aggregates of silicate layers intercalated withPEG segments are heterogeneously dispersed in the matrixof PEMA as observed in the XRD pattern. The partialdelamination and substantial curvature of silicate layersshow that there exists a partial compatibility between Na-MMT and the matrix polymer due to the compatibilizingeffect of the PEG segment linked to PEMA segment.

3.2. Mechanical and thermal properties

In various nanocomposites, the glass transition temper-ature and modulus were increased compared to the pristinepolymer, because the finely dispersed fillers reduce thechain mobility of matrix polymer and strengthen thematrix [25–28]. The dynamic mechanical thermal analysisis a good method to examine these variations [26].Dynamic mechanical properties of PEMA/Na-MMTnanocomposites measured by DMTA are shown inFig. 3, and the quantitative values of E0 are shown in Table3 where it is seen that the tensile storage modulus, E0,generally increases and the glass–rubber transition temper-ature moves to higher temperatures as the content of Na-MMT is increased. These results show that Na-MMTintercalated with a PEG segment reduces the chain mobil-ity of PEMA segment and reinforces effectively the PEMAmatrix. The rubbery plateau of pristine PEMA (NC-0) isnot evidently manifested in Fig. 3. However, we can seethat the rubbery plateau becomes evident and extends toa much higher temperature when PEMA was reinforcedwith Na-MMT even in the case of NC-1 which contained1.2 wt% of Na-MMT. These results show that Na-MMT

intercalated with a PEG segment performs its role as filleror physical crosslinker which inhibits the slippage or theflow of PEMA segments at the rubbery state. The theoret-ically derived equation on ideal rubber elasticity demon-strated that the modulus is inversely proportional to themolecular weight of the chain segments between crosslinks[29]. So, the increased physical crosslinking density byincreased amount of the Na-MMT intercalated with a

E' (P

a)

106

109

108

107

1010

10

1

0.1

0.01

50 100 150 200 250 300

Tan

δ

Temperature (°C)

Fig. 3. Dynamic mechanical properties of PEMA nanocomposites: (—)NC-0, (�������) NC-1, (- - -) NC-3, and (– � –) NC-7.

M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926 1923

PEG can be a cause of this modulus increase at rubberystate.

In Table 2, we can see that the Tg measured by DSC isincreased when the content of Na-MMT is increased, as inthe glass–rubber transition temperatures measured byDMTA. This also shows that the chain mobility of PEMAis reduced by Na-MMT intercalated with a PEG segment.

Tensile properties of PEMA nanocomposites measuredat 80 �C are shown in Table 3 where it is seen that boththe 10% secant modulus and tensile strength increase asthe content of Na-MMT is increased due to the reinforcingeffect of Na-MMT. The 10% secant modulus and the ten-sile strength of NC-5 were about seven and four timeshigher, respectively, than the corresponding values for pris-tine PEMA (NC-0) indicating that Na-MMT effectivelyreinforced the matrix polymers. In Table 3, we can see thatthe elongation at break is reduced in the nanocompositescompared to that of pristine polymer (NC-0) [27,30]. Thisshows that physical crosslinking by Na-MMT intercalated

Table 3Mechanical properties of PEMA nanocomposites

Sample E0 (Pa) 10% secant m

25 �C 80 �C 130 �C

NC-0 7.91 � 109 2.65 � 108 3.89 � 106 0.15 ± 0.02NC-1 8.55 � 109 3.61 � 108 1.01 � 107 0.50 ± 0.04NC-3 9.11 � 109 5.21 � 108 1.20 � 107 0.82 ± 0.03NC-5 9.34 � 109 6.46 � 108 1.61 � 107 1.05 ± 0.03NC-7 1.03 � 1010 8.54 � 108 2.02 � 107 1.38 ± 0.05NC-9 1.05 � 1010 1.04 � 109 2.19 � 107 1.64 ± 0.04

with a PEG segment hinders the rearrangements of poly-mer chains adapted to external force during tensile test.

3.3. Shape memory behavior

The results of cyclic tensile tests for examining the shapememory effect of PEMA nanocomposites are shown inFig. 4. The sample was elongated at the rubbery state,80 �C, to 50% strain (em) at a constant elongation rate of10 mm/min. While maintaining the strain at em, the sam-ples were quenched to the glassy state (25 �C) for 10 minand unloaded. Upon removing the constraint at 25 �C asmall strain recovery to eu occurs. The sample was subse-quently heated to 80 �C again and remained at that temper-ature for the next 10 min allowing recovery of strain. Thiscompleted one thermomechanical cycle (N = 1) leaving aresidual strain, ep, where the next cycle (N = 2) starts[4,11–13]. We selected 80 �C as a temperature for the elon-gation and recovery in the thermomechanical cycle becausethis temperature was optimum one for carrying out at leastseven thermomechanical cycles without failure, andbecause it was higher than the Tg’s measured by DSC(Table 2), although it was lower than the peak tempera-tures of tand in Fig. 3.

We can see in Fig. 4a that residual strain, ep, of NC-0increased as the test cycle was repeated to a value of about17% when the cycle number was N = 6. This shows that theslipping and disentangling of the polymer chains, whichcauses the permanent deformation, have increased byrepeated deformations. As seen in Fig. 4b and c, onlyminor increases of ep were observed. That is, the ep valueswere less than 3% even when the cycle number was N = 6.These results show that Na-MMT intercalated with a PEGsegment, which is dispersed heterogeneously in the polymermatrix, can perform its role as a fixed phase in the shapememory test. This inhibits the permanent deformationdue to the slippage of a PEMA segment even when the con-tent of Na-MMT was 1.2 wt%. However, in Fig. 4d and e,it is seen that the ep values increased slightly again whencompared to those of Fig. 4b or c. We observed in our pre-vious studies that ep values could be increased when thecontent of the fixed phase was in excess [5,31,32] becausethe probability that the fixed phase can be deformedtogether with a reversible phase by external force increaseswhen the content of the fixed phase is high. Therefore, the

odulus (MPa) Tensile strength (MPa) Elongation at break (%)

0.60 ± 0.04 340 ± 340.95 ± 0.05 282 ± 251.85 ± 0.03 235 ± 92.28 ± 0.05 257 ± 252.33 ± 0.06 222 ± 222.51 ± 0.03 314 ± 24

Strain (%)10 20 30 40 50

0.3

0.6

0.9

1.2

N=1

Stre

ss (M

Pa)

N=2 N=3 N=4 N=5 N=6

Strain (%)10 20 30 40 50

0.3

0.6

0.9

1.2

N=1N=2N=3

N=4N=5N=6

Stre

ss (M

Pa)

Strain (%)10 20 30 40 50

0.5

1.0

2.0

N=1

Stre

ss (M

Pa)

N=2 N=3 N=4N=5

N=6

1.5

Strain(%)10 20 30 40 50

0.5

1.0

2.0

N=1

Stre

ss (M

Pa) 1.5

N=2 N=3N=4

N=5

N=6

Strain (%)10 20 30 40 50

0.5

1.0

1.5

2.0

Stre

ss (M

Pa)

N=1

N=2

N=3

N=4

N=5N=6

a

e

c d

b

εp

Fig. 4. Cyclic tensile behavior of (a) NC-0, (b) NC-1, (c) NC-5, (d) NC-7, and (e) NC-9.

1924 M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926

higher ep values of Fig. 4d or e, compared to those values ofFig. 4b or c, suggest the probability that the fixed phase,Na-MMT intercalated with a PEG segment, can bedeformed by external force is increased at higher contentsof Na-MMT intercalated with a PEG segment. And weobserved that the thermomechanical cycles can be carriedout more than 10 cycles with NC-1 or NC-3, however,can hardly be carried out up to 10 cycles with NC-0,NC-7 or NC-9. This shows that failure can occur more eas-

ily by repeated deformations of 50% when there is toomuch Na-MMT as well as when there is no Na-MMT.

One can see in Fig. 4a that eu is 49% when N = 1 and eu

decreases slightly to 48% when N = 6. Fig. 4d shows that eu

is 48% when N = 1, and is decreased when the thermome-chanical cycle is increased and become 45% when N = 6.Fig. 4e shows that eu is 43% for N = 1 and 39% forN = 6. These results show that the deformed shape wasalmost perfectly fixed by cooling to 25 �C, although minor

0.1 10 100

103

104

105

106

G' (

Pa)

1

ω (rad/s)

Fig. 6. Storage shear modulus versus frequency of (d) NC-0, (N) NC-1,(s) NC-3, (D) NC-5, (h) NC-7, and (e) NC-9.

M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926 1925

shape recovery had been occurred. The result that thisminor shape recovery after cooling increased as the contentof Na-MMT was increased supports that Na-MMT inter-calated with a PEG segment did its role as fixed phasewhich memorized the original shape.

The molecular weight of NC-0 is lower compared tothose of nanocomposites which were prepared with macro-azoinitiator (Table 2). This lower molecular weight of NC-0 may also be a cause of poor shape memory effect of NC-0compared to those of nanocomposites, because the physicalentanglements of polymer chains are reduced at lowermolecular weight.

3.4. Rheological properties

Fig. 5 shows the logarithmic plot of complex viscosity,g�, versus angular frequency, x, at 160 �C where it is seenthat g� increased as the content of Na-MMT increasedand this increase is more evident at a low shear rate showingamplified shear thinning behavior. This peculiar behavior ata low shear rate can also be observed, in the logarithmicplot of the storage shear modulus, G0, versus x as shownin Fig. 6. It is seen in Fig. 6 that the increase of the storageshear modulus, G0, due to the added Na-MMT, is more evi-dent at a low shear rate causing reduced frequency depen-dency of G0 in the presence of Na-MMT. That is the slopeat 0.1 rad/s is 1.14 for NC-0 however slopes are 0.40 forNC-1, and less than 0.20 for NC-3, NC-5, NC-7, and NC-9. This pseudo-solid like behavior at a low shear rate, wherean almost constant G0 exists independent of shear rate, sig-nifies the onset of incomplete relaxation [33–35]. Theincomplete relaxation of dispersed Na-MMT can be causedby the physical jamming of highly anisotropic dispersed sil-icate layers which prevents free rotation from adapting tothe external dynamic shear. This incomplete relaxationand interaction between the silicate layers develops athree-dimensional mesoscopic structure of silicate layersand this causes the silicate layers to not relax independentlywhich results in the pseudo-solid like behavior [33–35]. The

104

105

106

107

0.1 10 1001

ω (rad/s)

η* (P

a·s)

Fig. 5. Complex viscosity versus frequency of (d) NC-0, (N) NC-1, (s)NC-3, (D) NC-5, (h) NC-7, and (e) NC-9.

retarded relaxation of matrix molecules can also occurwhen there is strong interaction between matrix moleculesand Na-MMT [36,37]. Therefore, the pseudo-solid likebehavior observed in Figs. 5 and 6 can be attributed tothe strong interaction between Na-MMT and intercalatedPEG segments which are linked to PEMA segments, as wellas to the development of the three-dimensional mesoscopicstructure of the silicate layer.

It is seen in Fig. 6 that this frequency dependence of G0 issomewhat different from those observed with othernanocomposites [33–41], in the facts that the diminishedfrequency dependence is manifested even at higher frequen-cies where x > 10 rad/s, and frequency dependence isslightly increased at lower frequency compared to that athigher frequency. For example, the slope of NC-3 inFig. 6 is 0.08 at 10 rad/s, however, it is increased to 0.12at 0.1 rad/s. The matrix polymer of our nanocompositesis the multiblock copolymer of PEMA and PEG, and thereexists a strong interaction between Na-MMT and PEG seg-ment. This peculiar phase morphology can be a cause ofthis unusual frequency dependence of G0 [33], however,some more studies will be necessary for exact explanations.

Anyhow, Fig. 6 shows that elasticity of nanocompositesis evidently improved by Na-MMT, which suggests the effi-cient reinforcing effect of Na-MMT.

4. Conclusions

In the nanocomposites prepared by in situ polymeriza-tion of ethyl methacrylate with macroazoinitiator interca-lated at the gallery of Na-MMT, we observed that thePEG segment originating from macroazoinitiator wasintercalated at the gallery of Na-MMT because of the affir-mative interaction between Na-MMT and the PEGsegment.

The mechanical and rheological properties showed thatthe heterogeneously dispersed Na-MMT, which is interca-lated with a PEG segment, effectively acts as filler andphysical crosslinker.

1926 M.S. Kim et al. / Composites Science and Technology 68 (2008) 1919–1926

The shape memory behavior showed that only 1.2 wt%of Na-MMT can perform efficiently as a fixed phase toinhibit the slippage of matrix molecules and excessamounts of fixed phase, more than 7 wt% of Na-MMT,were not conducive for shape recovery.

Because polymers with broad spectrum of Tg can beobtained by the radical polymerization of vinylic mono-mers, and because a fixed phase can be easily incorporatedto the polymers by the method shown in this paper, thismethod can be utilized effectively in the design of shapememory vinylic polymers with various shape recoverytemperatures.

Acknowledgment

This work was supported by the University of Ulsan Re-search Fund of 2006.

References

[1] Kim BK, Lee SY, Xu M. Polyurethanes having shape memory effects.Polymer 1996;37:5781–93.

[2] Lendlein A, Kelch S. Shape-memory polymers. Angew Chem Int Ed2002;41:2034–57.

[3] Feninat FE, Laroche G, Fiset M, Mantovani D. Shape memorymaterials for biomedical applications. Adv Eng Mater 2002;4:91–104.

[4] Jeong HM, Ahn BK, Kim BK. Miscibility and shape memory effect ofthermoplastic polyurethane blends with phenoxy resin. Eur Polym J2001;37:2245–52.

[5] Choi N, Lendlein A. Degradable shape-memory polymer networksfrom oligo [(l-lactide)-ran-glycolide] dimethacrylates. Soft Matter2007;3:901–9.

[6] Sokolowski W, Metcalfe A, Hayashi S, Yahia L, Raymond J.Medical applications of shape memory polymers. Biomed Mater2007;2:S23–7.

[7] Liu C, Qin H, Mather PT. Review of progress in shape-memorypolymers. J Mater Chem 2007;17:1543–58.

[8] Schmidt AM. Electromagnetic activation of shape memory polymernetworks containing magnetic nanoparticles. Macromol Rapid Com-mun 2006;27:1168–72.

[9] Shimamura H, Ishizawa J, Hayashi S. Development of the shapememory polymer composites. Purasuchikkusu 2006;57:124–9.

[10] Behl M, Lendlein A. Shape-memory polymer. Mater Today2007;10:20–8.

[11] Jeong HM, Lee SH, Cho KJ, Jeong YT, Kang KK, Oh JK. Thermaland mechanical properties of the polymers synthesized by thesequential polymerization of propylene and 1-hexadecene. J ApplPolym Sci 2002;84:1709–15.

[12] Jeong HM, Song JH, Chi KW, Kim I, Kim KT. Shape memory effectof poly (methylene-1,3-cyclopentane) and its copolymer with poly-ethylene. Polym Int 2002;51:275–80.

[13] Kim BK, Lee JS, Lee YM, Shin JH, Park SH. Shape memorybehavior of amorphous polyurethanes. J Macromol Sci-Phys2001;B40:1179–91.

[14] Hayashi S. Properties and applications of polyurethane-series shapememory polymer. Int Prog Urethanes 1993;6:90–115.

[15] Ohki T, Ni Q, Ohsako N, Iwamoto M. Mechanical and shapememory behavior of composites with shape memory polymer.Composites 2004;A35:1065–73.

[16] Gall K, Dunn ML, Liu Y, Finch D, Lake M, Munshi NA. Shapememory polymer nanocomposites. Acta Mater 2002;50:5115–26.

[17] Cho JW, Lee SH. Influence of silica on shape memory effect andmechanical properties of polyurethane-silica hybrids. Eur Polym J2004;40:1343–8.

[18] Mather PT, Jeon HG, Romo-Uribe A, Haddad TS, Lichtenhan JD.Mechanical relaxation and microstructure of poly(norbornyl-POSS)copolymers. Macromolecules 1999;32:1194–203.

[19] Wu J, Lerner MM. Structural, thermal, and electrical characteriza-tion of layered nanocomposites derived from a Na-montmorilloniteand polyethers. Chem Mater 1993;5:835–8.

[20] Aranda P, Ruiz-Hitzky E. Poly(ethylene oxide)-silicate intercalationmaterials. Chem Mater 1992;4:1395–403.

[21] Jeong HM, Choi MY, Ahn YT. Morphology and properties ofpolyacrylonitrile/Na-MMT nanocomposites prepared via in-situpolymerization with macroazoinitiator. Macromol Res 2006;14:312–7.

[22] Al-Esaimi. Reaction catalyzed by montmorillonite: polymerization ofmethyl methacrylate. J Appl Polym Sci 1997;64:367–72.

[23] Ruiz-Hitzky E, Aranda P. Polymer–salt intercalation complexes inlayer silicate. Adv Mater 1990;2:545–7.

[24] Billingham J, Breen C, Yarwood J. Adsorption of polyamine,polyacrylic acid and polyethylene glycol on montmorillonite: anin situ study using ATR-FTIR. Vib Spectrosc 1997;14:19–34.

[25] Joshi M, Butola BS. Polymeric nanocomposites-polyhedral oligo-meric silsesquinoxanes (POSS) as hybrid nanofiller. J Macromol Sci-Polym Rev 2004;C44:389–410.

[26] Liu B, Ding Q, He Q, Cai J, Hu B, Shen J. Novel preparation andproperties of EPDM/montmorillonite nanocomposites. J Appl PolymSci 2006;99:2578–85.

[27] Teh PL, Ishak ZAM, Hashim AS, Karger-Kocsis J, Ishiaku US.Physical properties of natural rubber/organoclay nanocompositecompatibilized with epoxidized natural rubber. J Appl Polym Sci2006;100:1083–92.

[28] Chang Y, Yang Y, Ryu S, Nah C. Preparation and properties ofEPDM/organomontmorillonite hybrid nanocomposites. Polym Int2002;51:319–24.

[29] Treloar LRG. The physics of rubber elasticity. 3rd ed. Glasgow:Oxford University Press; 1975.

[30] Kato M, Tsukigase A, Tanaka H, Usuki A, Inai I. Preparation andproperties of isobutylene–isoprene rubber–clay nanocomposites. JPolym Sci Polym Chem 2006;44:1182–8.

[31] Lee HY, Jeong HM, Lee JS, Kim BK. Study on the shape memorypolyamides. Synthesis and thermomechanical properties of polycap-rolactone–polyamide block copolymer. Polym J 2000;32:23–8.

[32] Choi YJ, Kim BK, Jeong HM. A study on the shape memorypolyesters: 1. Shape memory characteristics of aliphatic polyester-rigid chain polyester block copolymers. Polymer (Korea)1998;22:131–9.

[33] Ren J, Silva AS, Krishnamoorti R. Linear viscoelasticity ofdisordered polystyrene–polyisoprene block copolymer based lay-ered-silicate nanocomposites. Macromolecules 2000;33:3739–46.

[34] Pinnavaia TJ, Beall GW, editors. Polymer-clay nanocomposites. NewYork: John Wiley & Sons; 2000.

[35] Hyun YH, Lim ST, Choi HJ, Jhon MS. Rheology of poly(ethyleneoxide)/organoclay nanocomposites. Macromolecules 2001;34:8084–93.

[36] Lim YT, Park OO. Phase morphology and rheological behavior ofpolymer/layered silicate nanocomposites. Rheol Acta 2001;40:220–9.

[37] Krishnamoorti R, Giannelis EP. Rheology and end-tethered polymerlayered silicate nanocomposites. Macromolecules 1997;30:4097–102.

[38] Russo GM, Simon GP, Incarnato L. Correlation between rheological,mechanical, and barrier properties in new copolyamide-based nano-composite films. Macromolecules 2006;39:3855–64.

[39] Zhang M, Sundararaj U. Thermal, rheological, and mechanicalbehaviors of LLDPE/PEMA/clay nanocomposites: effect of interac-tion between polymer, compatibilizer, and nanofiller. MacromolMater Eng 2006;291:697–706.

[40] Loiseau A, Tassin J. Model nanocomposites based on laponite andpoly(ethylene oxide): preparation and rheology. Macromolecules2006;39:9185–91.

[41] Ray SS. Rheology of polymer/layered silicate nanocomposites. J IndEng Chem 2006;12:811–42.