Introduction

Polymer/clay nanocomposites are attractive for research due to the very low cost of clay minerals, relatively simple preparation procedure, and efficient stiffening effect.1 The performance of polymer/clay nanocomposites depends on several factors, such as the clay content and dispersion state.2, 3 Surface modification of nanoclays with organosilane compounds has been studied in the past few decades, and it is known that organosilane modification should enhance filler dispersion in the polymer matrix, resulting in improvement of the mechanical and thermal properties due to the strong interfacial interaction between the clay filler and polymeric matrix.4

Polycarbonate (PC) is an engineering plastic that has been widely used in the fields of transport, construction, packaging and optics because of its outstanding impact resistance, dimensional stability, heat resistance, excellent dielectric properties and transparency.2, 5, 6, 7, 8 Much attention has been given to the preparation of PC/modified clay nanocomposites. Various preparation methods and a series of clays modified with surfactants have been applied to PC nanocomposites. However, the effect of clay modification and the dispersion state of PC/clay nanocomposites remain obscure. Feng et al.9 demonstrated the complicated effect of several clay modifiers on the flame retardancy and mechanical properties of PC nanocomposites and reported decreased Tg due to weaker organoclay/polymer matrix interactions. Hsieh et al.7 observed a depression of Tg by the loading of clays and modified clays,10 while a constant Tg was reported by Chow et al. However, Lee et al.8 reported an increased Tg by differential scanning calorimetry, and Nayak et al.11 also reported an enhanced Tg by both dynamic mechanical analysis and differential scanning calorimetry.

Halloysite nanotube (HNT), with the molecular formula Al2Si2O5(OH)4.2H2O, is a type of aluminosilicate nanoclay with a high length-to-diameter ratio and a hollow tubular structure. Their unique nanotubular structure makes these materials useful as nanocontainers for molecular adsorption,12, 13, 14 molecular storage and transport. The high length-to-diameter ratio is beneficial for their role as a reinforcing nanofiller of nanocomposites.5, 15 However, interfacial modification remains one of the great challenges in the fabrication of HNT/polymer nanocomposites due to the lack of effective functional groups on the outer surface of HNTs.

In this study, HNTs were modified with N-(2-aminoethyl)-3-aminopropyltrimethoxysilane (AEAPS) and n-octadecyltrimethoxysilane (OTMS) through a chemical vapor adsorption process (Scheme 1). The effects of HNT surface modification and loading on the morphology and mechanical and thermal properties of the HNT/PC nanocomposites were investigated.

Experimental Procedure

Materials

Halloysite was provided by Applied Minerals Inc. (New York, NY, USA) and used without further treatment. Poly(bisphenol A carbonate) (Mw=41,500), tetrahydrofuran and ethanol were obtained from Sigma Aldrich Inc. (St Louis, MO, USA) and used without further purification. AEAPS and OTMS were obtained from Shin-Etsu Chemical Co. Ltd (Tokyo, Japan) and used as received.

Preparation of organosilane-modified HNTs

Surface modification of HNTs was performed using a chemical vapor adsorption process. HNTs (0.5 g) were first dried in vacuum at 363 K for 10 h and then exposed to vacuum-ultraviolet rays (λ=172 nm) for 10 min under reduced pressure (30 Pa) three times to decompose organic compounds.16, 17, 18 The HNTs and a glass vessel filled with 1 ml AEAPS were enclosed in a reaction flask in a glove box filled with nitrogen, then the flask was heated in an oven at 383 K for 48 h. The modified HNTs were dispersed and ultrasonicated in toluene. The suspension was centrifuged to remove excess absorbed molecules and washed. This rinsing step was repeated with toluene, ethanol and water. The samples were then dried at 383 K for 5 h. OTMS-modified HNTs were prepared by the same procedure. AEAPS- and OTMS-modified HNTs are denoted as ‘A-HNTs’ and ‘O-HNTs’.

Preparation of PC/HNT nanocomposites

HNTs (3 mg) were dispersed in tetrahydrofuran, and PC (300 mg) was dissolved in tetrahydrofuran; they were then mixed together, followed by precipitation with a large amount of water. The precipitates were dried in vacuum at 363 K for 2 days. Films with thicknesses ∼110 μm were prepared by compression molding with a compression molding machine, TBD-50-2 (Toho Machinery Co. Ltd, Tokyo, Japan) at 523 K under 23 MPa.

Characterization

Fourier-transform infrared spectroscopic measurements were conducted on a Spectrum One (PerkinElmer Inc., Waltham, MA, USA) using KBr pellets. IR data were collected by averaging 64 scans between 4000 and 450 cm−1.

Thermogravimetric analysis (TGA) measurements were performed on a EXSTAR TG/DTA-6200 thermo balance (SII Nano Technology Inc., Chiba, Japan) in a nitrogen stream of 200 ml min−1 at a heating rate of 10 K min−1, ranging from 303 to 850 K.

Dynamic mechanical analysis measurements were performed using a RHEOVIBRON DDV-01FP (Orientec A&D Co., Ltd, Tokyo, Japan). Test specimens of ∼20 mm length, 3 mm width and 200 μm thickness were mounted at a clamping distance of 12 mm. Measurement of storage modulus, E′, and loss modulus, E′′, as a function of temperature was conducted at a frequency of 110 Hz from 123 to 473 K at a heating rate of 2 K min−1.

Scanning electron microscopy observations were accomplished by Real Surface View VE7800 (Keyence Co., Ltd, Osaka, Japan) with applied voltage of 5–15 kV. Samples were sputter-coated with an osmium layer using HPC-1SW Hollow Cathode Plasma CVD (Shinkuu Device Co., Ltd, Ibaraki, Japan).

Optical microscopy specimens of PC nanocomposites with thicknesses of less than 10 μm were prepared by melting on a hot plate. Optical microscopy was performed using a Nikon Eclipse LV 100 POL equipped with a CCD camera.

Thin films of PC nanocomposites were sectioned from bulk specimens at room temperature using an ultramicrotome. The thickness of the thin films is 100 nm; an ∼15 nm carbon layer was subsequently coated on the top surface of the grid. Bright-field transmission electron microscopy images of thin films were acquired at 200 kV using JEM 2000FS (JEOL Ltd, Tokyo, Japan).

Results and Discussion

Characterization of organosilane-modified HNTs by Fourier-transform infrared spectroscopy and TGA measurements.

The grafting of AEAPS and OTMS on the surface of HNTs was confirmed by Fourier-transform infrared spectroscopy. As shown in Figure 1, the IR spectra of A-HNTs and O-HNTs exhibited adsorption bands corresponding to AEAPS and OTMS, respectively. The appearance of alkyl groups at 2930 and 2832 cm−1 (both asymmetric stretching of C-H) indicates that the reaction of methoxysilane with the HNTs has occurred. The typical absorption bands of HNTs at 1032, 912 and 538 cm−1 remained after modification, indicating that the organosilane agents did not attach to the Si-O-Al skeleton of HNTs. The definite assignments of N-H at 3000–3400 cm−1 are difficult to observe due to the various N-H stretching environments. The appearance of alkyl groups at 2930 and 2832 cm−1 on O-HNTs indicates that OTMS has grafted on the HNTs.

Figure 1
figure 1

Fourier-transform infrared spectra of (a) HNTs, AEAPS and A-HNTs, and (b) HNTs, OTMS and O-HNTs. A full color version of this figure is available at Polymer Journal online.

Full size image

The thermal properties of HNTs, A-HNTs and O-HNTs were investigated by TGA. As shown in Figure 2, the original HNTs lose 13.0% of their total mass in two steps. The first region is from 300 to 634 K, with a weight loss of 2.1% that is attributed to the loss of moisture trapped on the HNT surface. The second region is from 634 to 850 K, which corresponds to the dehydroxylation of HNTs. Both A-HNTs and O-HNTs have a similar weight loss profile as HNTs; the weight loss of the moisture in the first region is 1.8% and 1.6%, respectively, whereas in the second region, the weight loss that is related to the decomposition of organic groups and the dehydroxylation of A-HNTs and O-HNTs is 15.5% and 16.5%, respectively. The degradation of HNTs, A-HNTs and O-HNTs starts from 634 K, 599 K and 591 K, respectively, indicating that the degradation of grafted AEAPS and OTMS occurs before the degradation of HNTs. The unreacted silane group on the HNT surface condenses to form covalent bonds and eliminated components. Moreover, the HNTs, A-HNTs and O-HNTs lose 13.0%, 17.3% and 18.1%, respectively, of their initial mass at 873 K, indicating that the AEAPS and OTMS grafting content in A-HNTs and O-HNTs is 4.3 wt% and 5.1 wt%, respectively. This weight loss difference indicates that the long alkyl tails of the OTMS modifiers cause more effective grafting on the HNT surface than AEAPS bearing a shorter alkyl tail with an amino group. (Scheme 2.)

Figure 2
figure 2

TGA curves of (a) HNTs, (b) A-HNTs and (c) O-HNTs in N2 atmosphere at a heating rate of 10 K min−1. A full color version of this figure is available at Polymer Journal online.

Full size image

Microstructure of PC nanocomposites

Generally, nanofiller dispersion in a polymer matrix is one of the key factors for performance of the composites. Scanning electron microscopy observation of the fracture surfaces of nanocomposites can be used for evaluating the adhesion between nanofillers and the polymer matrix. Figure 3 shows scanning electron microscopy images of the fractured surfaces of PC nanocomposites after immersion in liquid nitrogen. HNTs are clearly observed at the rough fracture surfaces of the nanocomposites. HNT aggregates are found in the PC matrix at higher HNT content levels and HNTs without modification. Existence of a considerable number of aggregates in PC nanocomposites indicates insufficient dispersion of HNTs because of weak interfacial interactions between HNTs and the PC matrix and strong filler–filler interaction. However, the aggregation of HNTs was remarkably broken down in the case of modified HNTs.

Figure 3
figure 3

Scanning electron microscopy images for the fracture surfaces of PC composites that were obtained by fracturing in liquid nitrogen. (a) PC/HNTs_0.5 wt%, (b) PC/A-HNTs_0.5 wt%, (c) PC/O-HNTs_0.5 wt%, (d) PC/HNTs_5 wt%, (e) PC/A-HNTs_5 wt% and (f) PC/O-HNTs_5 wt%. A full color version of this figure is available at Polymer Journal online.

Full size image

A fairly homogeneous dispersion of A-HNTs and O-HNTs in the PC matrix across the fractured surface can be observed in Figures 3b and c, indicating that the agglomeration of HNTs is effectively restrained by AEAPS and OTMS modifiers. The improved dispersion of A-HNTs and O-HNTs in PC can be explained by the low surface energy and hindered filler–filler interaction introduced by the organic groups. Furthermore, O-HNT/PC nanocomposites showed a higher degree of dispersion than A-HNT/PC nanocomposites, most likely due to the stronger affinity of hydrophobic alkyl groups of OTMS for the polymer matrix. Compared with the unmodified HNT/PC nanocomposites, a large polymer/filler interface is created by 0.5 wt% modified HNT loading, which leads to a stronger interaction between fillers and PC matrix. The dispersion state could also be demonstrated by optical microscopy and transmission electron microscopy (see the Supporting Information, Supplementary Figures S1 and S2).

Thermal properties of PC/HNT nanocomposites

The thermal stability of PC/HNT, PC/A-HNT and PC/O-HNT composites was analyzed by TGA under a nitrogen atmosphere. The results are shown in Figure 4, and the characteristic weight loss temperatures are summarized in Table 1. The sudden degradation part of PC at ∼780 K corresponds to the carbonization of aromatic polyester (-O-C6H4-CO-).19, 20 All the composites demonstrate higher thermal stability than neat PC. Both HNT loading and HNT surface modification enhanced the thermal stability of the composites. The temperature at 5% weight loss (T−5wt%) for neat PC is 687 K, whereas T−5wt% for the composite filled with 0.5 wt% unmodified HNTs is 725 K, which is 38 K higher than that of neat PC. The T−5wt% values are further increased to 736 K for the PC/A-HNT_0.5 wt% nanocomposite and 743 K for the PC/O-HNTs_0.5 wt% nanocomposite, which are 49 and 56 K higher than that of neat PC, respectively. The nanocomposites show much inferior thermal stability at higher HNT loading. The T−10wt% values of the HNTs, PC/A-HNTs_0.5 wt%, and PC/O-HNTs_0.5 wt% show a similar trend as T−5wt%.

Figure 4
figure 4

TGA curves of (a) PC/A-HNT nanocomposites and (b) PC/O-HNT nanocomposites in nitrogen at heating rate of 10 K min−1. A full color version of this figure is available at Polymer Journal online.

Full size image
Table 1 T−5wt%, T−10wt%, E′, Tαa and ΔHa values determined by TGA, and dynamic mechanical analysis measurements for PC and PC nanocomposites
Full size table

Thermal stability improvement of polymer nanocomposites with a small amount of modified tubular nanomaterials has been reported by several researchers.21, 22 Du et al.22 demonstrate that the enhanced tendency with these HNTs is related to the high volume percentage of the lumen structure. Polymer molecules, which are quite sensitive to water and hydroxyl groups, would initiate a massive hydrolysis or alcoholysis of the carbonate linkages in the case of even a small amount of water or hydroxyl groups. The high loading ability of the lumen of HNTs has recently been reported.17, 18 O-HNT/PC nanocomposites showed higher thermal stability than A-HNT/PC nanocomposites, most likely due to the stronger interaction between hydrophobic alkyl groups of OTMS and the polymer matrix. However, the amine groups of AEAPS are expected to react with the PC matrix at high temperature to increase the interfacial strength between HNTs and the PC matrix. Thus, the Tmax of all the nanocomposites is higher than that of neat PC. However, ambient humidity might influence the hydrolytic degradation of the A-HNT/PC matrix; thus, the Tmax for A-HNT/PC is slightly lower than that of O-HNT/PC.

Dynamic mechanical properties of PC/HNT nanocomposites

To estimate the mechanical performance and molecular aggregation structure of nanomaterials, the temperature dependence of the dynamic storage modulus, E′, and dynamic loss modulus, E′′, of neat PC and the nanocomposites with A-HNTs and O-HNTs from 125 to 475 K at 110 Hz are investigated by dynamic mechanical analysis. As shown in Figure 5, both γ-relaxation and αa-relaxation corresponding to segmental motion were observed in each curve. E′ of PC/O-HNTs_0.5 wt%, which shows the strongest storage modulus, decreased from 2.1 GPa to 5 MPa over the αa-relaxation range of 400–450 K. The γ-relaxation peak was shifted to higher temperatures; observed at 150–250 K for all samples, this peak is assigned to local motion of the methyl groups of amorphous PC. These results indicate that the incorporation of HNTs and modified HNTs into the PC matrix enhanced stiffness, but the load-bearing capability and impact strength23 decreased. HNT loading and surface modification enhanced the storage modulus of the nanocomposites, and modified samples are much more efficient than unmodified HNTs. The enhancement of the E′ of PC nanocomposites can be predicted by the constraint of polymer chains to the filler surface.

Figure 5
figure 5

Temperature dependences of (a) and (c) dynamic storage modulus, E′, dynamic loss modulus, E′′, and αa-adsorption of PC/A-HNT nanocomposites. (b) and (d) E′, E′′ and αa-adsorption of PC/O-HNT nanocomposites. A full color version of this figure is available at Polymer Journal online.

Full size image

As αa-adsorption is associated with the onset of segmental motion in the glass transition process of polymers, the effect of nanofiller-aggregated structure on the glass transition of polymer nanocomposites can be explained by the apparent activation energy (ΔHa) for the αa-adsorption process. The value of ΔHa is determined according to the shift of the αa-adsorption peak temperature (Tmax) and ΔHa of all PC nanocomposites, as shown in Figure 6. The apparent activation energies for the αa-adsorption were evaluated based on the Arrhenius equation and are summarized in Table 1. All the nanocomposites show higher ΔHa than neat PC. Both AEAPS- and OTMS-modified HNT/PC nanocomposites exhibit higher ΔHa than PC/unmodified HNT nanocomposites. These results indicate that modified HNTs are more effectively constrained polymer chains, and higher ΔHa results in stronger confinement between HNTs and the PC chain. The activation energy of αa-absorption of PC/O-HNTs_0.5 wt% is 26 kJ mol−1 higher than PC/A-HNTs_0.5 wt%, suggesting that micro-Brownian motion of PC in O-HNT nanocomposites is more restricted than that of A-HNT nanocomposites. This result indicates that the interfacial interaction at PC/O-HNTs is much stronger than that at PC/A-HNTs.

Figure 6
figure 6

Plots of ln f vs (1/Tmax) for (a) net PC, (b) PC/HNTs_0.5 wt%, (c) PC/HNTs_5 wt%, (d) PC/A-HNTs_0.5 wt%, (e) PC/A-HNTs_5 wt%, (f) PC/O-HNTs_0.5 wt% and (g) PC/O-HNTs_5 wt%. A full color version of this figure is available at Polymer Journal online.

Full size image

Conclusions

HNTs modified by a chemical vapor adsorption process remarkably improves the thermal stability of PC nanocomposites. The use of different modifiers has allowed us to study their effect on nanocomposite performance. Both AEAPS and OTMS were immobilized to the surface of HNTs, which improves compatibility between PC and the modified HNTs. Both the modulus and thermal stability of PC were enhanced by AEAPS- and OTMS-modified HNTs. The glass transition and mechanical properties of nanocomposites are controlled by both the morphology and compatibility of the organoclays with the matrix. The OTMS modifier, bearing longer alkyl tails, shows stronger affinity for the matrix than the AEAPS modifier, bearing shorter alkyl tails with amino groups. These conclusions for organoclay modifiers and PC nanocomposites would be beneficial for the synthesis of high-performance sustainable polymer engineering nanocomposites.

scheme 1

Schematic representation for the modification of HNTs with AEAPS and OTMS by a chemical vapor adsorption process. A full color version of this figure is available at Polymer Journal online.

scheme 2

Schematic representation of PC/HNT nanocomposites preparation through a combination of solution mixing, precipitating and compression molding. A full color version of this figure is available at Polymer Journal online.