J Mater Sci (2012) 47:2665–2674 DOI 10.1007/s10853-011-6092-5
Fabrication of transparent polymer-matrix nanocomposites with enhanced mechanical properties from chemically modified ZrO2 nanoparticles Tarik Ali Cheema • Alexander Lichtner • Christine Weichert • Markus Bo¨l • Georg Garnweitner
Received: 18 September 2011 / Accepted: 28 October 2011 / Published online: 11 November 2011 Ó Springer Science+Business Media, LLC 2011
Abstract Optically transparent nanocomposites with enhanced mechanical properties were fabricated using stable dispersions of sub 10 nm ZrO2 nanoparticles. The ZrO2 dispersions were mixed with a commercially available bisphenol-A-based epoxy resin (RIMR 135i) and cured with a mixture of two amine-based curing agents (RIMH 134 and RIMH 137) after complete solvent removal. The colloidal dispersions of ZrO2 nanoparticles, synthesized through a non-aqueous approach, were obtained through a chemical modification of the ZrO2 nanoparticle surface, employing different organic ligands through simple mixing at room temperature. Successful binding of the ligands to the surface was studied utilizing ATR–FT-IR and thermogravimetric analysis. The homogeneous distribution of the nanoparticles within the matrix was proven by SAXS and the observed high optical transmittance for ZrO2 contents of up to 8 wt%. Nanocomposites with a ZrO2 content of only 2 wt% showed a significant enhancement of the mechanical properties, e.g., an increase of the tensile strength and Young’s modulus by up to 11.9 and 12.5%, respectively. Also the effect of different surface bound ligands on the mechanical properties is discussed.
T. A. Cheema A. Lichtner G. Garnweitner (&) Institute of Particle Technology, Technische Universita¨t Braunschweig, Volkmaroder Strasse 5, 38104 Braunschweig, Germany e-mail:
[email protected] C. Weichert M. Bo¨l Institute of Solid Mechanics, Technische Universita¨t Braunschweig, Schleinitzstrasse 20, 38106 Braunschweig, Germany
Abbreviations AFM Atomic force microscopy ATR–FT-IR Attenuated total reflectance–FT-IR DLS Dynamic light scattering DPP Diphenyl phosphate DTG Differential thermal analysis DSC Differential scanning calorimetry FT-IR Fourier-transform infrared spectroscopy MTS Maximum tensile strength NC Nanocomposite NP Nanoparticle rt Room temperature SAXS Small angle X-ray scattering SA Sorbic acid TEM Transition electron microscopy THF Tetrahydrofuran TGA Thermogravimetric analysis
Introduction Cured epoxy resins are probably the most versatile class of thermosetting polymers, with manifold applications in the fields of coatings, automotives, electronics, adhesives, and light-weight construction [1, 2]. They form highly crosslinked, amorphous polymers with favorable mechanical properties for many applications; additionally they possess excellent chemical resistance, and good thermal as well as electrical properties. However, mostly they are brittle and exhibit poor resistance to the initiation and growth of cracks [1, 2]. To overcome these drawbacks, nano-sized fillers are added to the epoxy resins to form nanocomposites with enhanced properties after curing [3–9]. Different
123
2666
methods are known for the fabrication of nanocomposites, the most important being the sol–gel method, in situ polymerization and blending [3, 4, 6, 10–15]. Blending, where nanoparticle powders are added and then mechanically sheared with the polymers or resins, is the most preferred method due to its simpleness and feasibility also on larger scale. However, de-agglomeration of the nanoparticle powders through this method is difficult, which mostly results in opaque or semitransparent nanocomposites [16]. In order to fabricate mechanically enhanced and optically transparent nanocomposites, which are highly attractive for use in the fields of scratch proof surface coatings [17] or optically transparent impact-resistant nanocomposite materials for transport vehicles helicopters and fighter aircrafts [18], it is of utmost importance that the nanoparticles are almost completely de-agglomerated and homogeneously distributed within the matrix [17, 18]. The mechanical properties of the nanocomposites depend to a large extent on the characteristics of the embedded fillers. Giannakopoulos et al. [8] reported in 2011 the use of core–shell rubber fillers in an epoxy-based polymer matrix, showing an increase in the fracture energy but a decrease in the tensile strength and the Young’s modulus. Medina et al. [16] reported in 2008 the use of ZrO2 nano-sized fillers to improve the tensile properties of an epoxy-based thermoset, showing an increase in fracture energy, Young’s modulus, and tensile strength. Commercially available ZrO2 nanoparticles with a primary size of 12 nm were employed without further modification. These particles were dispersed in the epoxy resin using a torus mill with volume fractions of 2 and 8 vol%. However, the dispersed nanoparticles showed agglomerates with average sizes of 20–100 nm, resulting in opaqueness of the nanocomposites. In addition to the characteristics of the embedded filler, the interface between the nanoparticles and the matrix also plays an important role with respect to the properties of the composite. Effects of strong and weak interfacial interactions have shown to directly influence the glass-transition temperature and the Young’s modulus as well as the toughness [12, 19, 20] of the nanocomposites. A full understanding of these effects, however, has not been achieved yet. The focus of this contribution is on the fabrication of epoxy-based optically transparent nanocomposites (NCs) with improved mechanical properties emanating from highly stable nanoparticle dispersions. A key prerequisite thereby was the use of very small ZrO2 nanoparticles (\10 nm). The nanoparticles were synthesized using a nonaqueous approach, which was shown earlier to be a highly promising method for the synthesis of different binary and ternary metal oxide nanoparticles [21–24]. It is of utmost importance that these nanoparticles (NPs) are homogeneously distributed within the matrix, without major
123
J Mater Sci (2012) 47:2665–2674
agglomeration or phase separation. In this article, we focus on the chemical modification of synthesized nanoparticles to achieve a good physical as well as chemical compatibility to the matrix. In order to study the effect of the interfacial interaction between the components, the modifying ligand adsorbed to the particle surface was varied systematically, and the effects on the fabricated NCs investigated by both structural analysis and optical as well as mechanical characterization.
Experimental Materials Zirconium n-propoxide (70 wt% solution in 1-propanol), 2,4–hexadienoic acid (C99%), and diphenyl hydrogen phosphate (99%) were purchased from Sigma-Aldrich; benzyl alcohol (99.9%) was obtained from Roth; tetrahydrofuran (99.5%) and hexane (95%) were acquired from AppliChem. The epoxy resin RIMR 135i and curing agents RIMH 134 and RIMH 137 were obtained from LangeRitter. All materials were used without further purification. Synthesis of ZrO2 nanoparticles and their chemical dispersion The ZrO2 NPs were synthesized using the non-aqueous synthesis as described earlier [25] with slight modifications. 80 mL of precursor solution were dissolved in 500 mL benzyl alcohol (BnOH) and reacted for 4 days at 220 °C in a steel reactor fitted with an additional glass cup (PARR Instruments). The NPs were obtained by centrifuging 35 mL of the suspension at 7500 rpm for 15 min. The supernatant was decanted and the precipitate washed with tetrahydrofuran (THF), followed by centrifugation at 7500 rpm for 15 min. The washing procedure was repeated twice before the precipitate was mixed with 15 mL of a 135 mM solution of either 2,4-hexadienoic acid (sorbic acid, SA) or diphenyl hydrogen phosphate (DPP) in THF and allowed to stir for 24 h. To remove the excessive, unbound ligand fraction from the dispersions, the particles were precipitated by addition into hexane with a volume ratio of 5:1 (hexane: dispersion), shaken for 5 min and then centrifuged at 7500 rpm for 10 min. The precipitate was redispersed in THF with equal dispersion volume as in the earlier step. Preparation of the pre-nanocomposite To fabricate nanocomposites with a ZrO2 content of 2 wt%, 14.15 mL of the dispersion (with a ZrO2 content of 60 mg mL-1) was added to 32 g of the epoxy resin (RIMR 135) and mixed on a magnetic stirrer for 15 min at ambient
J Mater Sci (2012) 47:2665–2674
conditions. Thereafter, THF was removed using a rotary evaporator (Heidolph) set at 55 °C and 150 mbar for 16 h. After removal of THF the pre-nanocomposite was further treated using a vacuum dissolver system (DispermatÒ, VMAGetzmann GmbH) fitted with a 25-mm dissolver disk. The dissolver was operated at 3000 rpm for 2 h under vacuum (100–200 mbar) and an elevated temperature of 50 °C. Curing the nanocomposite The pre-nanocomposite was cured using a mixture of 2.88 g RIMH 134 and 6.12 g RIMH 137 (mass ratio of 30:70). The mixture was added to the pre-nanocomposite by simple stirring, followed by degassing at 10 mbar in a desiccator at room temperature (rt) for 20 min. Finally the mixture was cast into an aluminum-based mold designed according to DIN EN ISO 527-2 with a cross-sectional area of 4 9 4 mm2. The mold was placed in an oven at 65 °C for 3 h. After cooling to room temperature the samples were removed from the mold and tempered for 8 h at 80 °C. Characterization of the nanoparticles and their dispersion To determine the ZrO2 content in dispersion, simple gravimetry was performed in ceramic crucibles using 200 lL of dispersion and heating them to 600 °C. The volumetric particle size distribution of the dispersions (diluted to 40 mg mL-1) was determined using dynamic light scattering on a Malvern Zetasizer Nano ZS device, performing a 173° backscatter measurement with three repeated runs. The optical transmittance of the dispersions (with a ZrO2 content of 60 mg mL-1) was measured using a DU720 UV–Vis spectrophotometer from Beckman Coulter. Transition electron microscopy (TEM) investigations were performed on an Omega 912 instrument (Carl Zeiss) microscope at an acceleration voltage of 120 kV. Diluted ZrO2 dispersions with a concentration of approximately 65 lg mL-1 were deposited on a perforated carbon-coated copper grid and dried under ambient conditions prior to measurement. Attenuated total reflectance fourier-transform infrared spectroscopy (ATR– FT-IR) and thermogravimetric analysis (TGA) were performed using dried powder samples obtained by evaporating the THF from the redispersed ZrO2 nanoparticles under vacuum at rt. ATR–FT-IR was performed on a Bruker Tensor 27 instrument in the range of 4000–500 cm-1. TGA was carried out on a Mettler Toledo TG/SDTA 851 under air in the range of 25–950 °C at 10 °C min-1. Characterization of the nanocomposites TGA of the pre-nanocomposites was performed under N2 atmosphere from 25–250 °C at 5 °C min-1, and TGA of
2667
the cured nanocomposites was performed under air in the range of 25–950 °C at 10 °C min-1. Small angle X-ray scattering was performed on cured nanocomposites on an X’Pert PRO MPD fitted with a PIXcel detector and a transmission-reflection spinner, scanned between -0.1150 and 5.0050° 2h and a step size of 0.01°. Tensile properties were determined according to the DIN EN ISO 527-1, the tests were performed with an initial force of 10 N and a testing speed of 0.1 mm min-1 to determine the Young’s modulus at strain 0, and 1 mm min-1 to determine the maximum tensile strength, using a tensile testing unit from Zwick/Roell. Fracture analysis was performed using atomic force microscopy (AFM) of the fracture surface employing the XE-100 from Park Systems operated at intermittent mode with a cantilever having a spring constant of 7.4 N m-1 at a scan rate of 0.5 Hz. The glasstransition temperature (Tg) of the cured pure polymer and nanocomposite samples was determined employing the differential scanning calorimetry (DSC) on a Netzsch DSC 204 Phoenix performing two heating cycles from 25 to 250 °C at a heating rate of 10 °C min-1 and a cooling rate of 40 °C min-1.
Results and discussion Chemical surface modification and dispersion Owing to the absence of stabilizers in the synthesis reaction medium, the nanoparticles are present as agglomerates several hundred nm in size after the synthesis, resulting in a white turbid suspension in benzyl alcohol [23]. The chemical modification treatment of the ZrO2 nanoparticle surface, when using 35 mL of initial ZrO2 suspension and dispersing the ZrO2 in 15 mL THF, resulted in stable dispersions of ZrO2 in THF, with solid contents of *80 and 85 mg mL-1 for the ligands SA and DPP, respectively. After screening a large number of ligands, SA and DPP were chosen because of their different chemical nature, based on their different anchoring groups binding to the ZrO2-surface, as well as their tail groups, facing away from the particle surface and thus resulting in high dispersibility in the solvent as well as the resin. During the chemical modification of the nanoparticles, the organic ligands anchor on the surface either through covalent or hydrogen bonding [17, 23, 25, 26], thus mostly forming strong interactions the between particle surface and the ligand. To remove unbound or loosely surface-bound ligands from the dispersion, the modified NPs were precipitated using hexane. For ease of discussion the dispersions of the redispersed ZrO2 NPs are referred to as ‘‘washed dispersions’’, whereas the initial dispersions as ‘‘unwashed dispersions’’. The successful modification of the nanoparticles,
123
2668
i.e., strong binding of the ligand to the surface, was verified through the redispersibility of the ZrO2 NPs. The solid contents of the washed dispersions of ZrO2-SA and ZrO2DPP were 77 and 79 mg mL-1, respectively, confirming full redispersibility (when adding lower amounts of solvent, even solid contents of up to 200 mg mL-1 were successfully reached, but were not utilized further). Both the unwashed and washed dispersions were stable for several weeks. The successful colloidal dispersion of the nanoparticles was verified through dynamic light scattering (DLS) and UV–Vis (transmittance) measurements. DLS was employed to determine the particle size distribution (based on the hydrodynamic diameter) of the nanoparticles in dispersion. The analyses of the washed dispersions show that the nanoparticles are monomodally dispersed for both ZrO2-SA and ZrO2-DPP, as can be seen in Fig. 1. The volume median 90 diameters D50 v and Dv of the cumulative distributions of both washed dispersions are also given in Fig. 1, being almost identical for both particle systems, thus proving excellent particle stability in THF even after redispersion. UV–Vis measurements (Fig. 2) show that the samples are almost fully optically transparent at 750 nm (red) and become less transparent at 380 nm (violet); showing a transmittance of 95–39% and 80–22 % from 750 to 380 nm for ZrO2-DPP and ZrO2-SA, respectively. The transparence of the dispersion can also be examined in the photograph of the ZrO2-DPP dispersion with a ZrO2 content of 60 mg mL-1 (Fig. 2). The morphology and stability of the nanoparticles was further verified through TEM measurements. Figure 3 presents the micrograph of ZrO2 NPs dispersed with SA in THF. The image shows that the ZrO2 NPs are edged in shape having primary sizes between 6 and 8 nm, which are in consistence with the DLS measurements. The agglomerates seen in the upper right of the image can be attributed to the drying of the dispersion.
J Mater Sci (2012) 47:2665–2674
Fig. 1 Particle size distribution of the ZrO2 dispersions a ZrO2-SA and b ZrO2-DPP in THF as determined through dynamic light scattering; the solid lines represent the volume distribution and the dotted lines the cumulative undersize
ATR–FT-IR analysis In order to analyze the attachment of the ligand in detail, we employed ATR–FT-IR spectroscopy on the dried nanoparticle powders. Figure 4 shows the IR spectra of SA and ZrO2-SA. In both spectra, characteristic absorption bands of the stretching vibration associated to sp2- and sp3hybridized H–C bonds of the ligand are seen at *3020 and 2960–2850 cm-1, respectively. The absorption band at 993 cm-1 is assigned to the out-of-plane bending vibration of the sp2 H–C bond situated at the C=C bond in conjugation with the C=O bond [27]. The two C=C bonds are visible through the stretching vibrations at 1636 and 1610 cm-1 for the pure SA. These bands are, however, shifted to higher wavenumber for the C=C bonds of the SA ligand bound to the ZrO2 surface, i.e., 1649 and 1616 cm-1. For the pure ligand (SA), absorption bands
123
Fig. 2 Left Transmittance curves obtained though UV–Vis spectrophotometry of two optically transparent ZrO2 dispersions open circle ZrO2-SA and open square ZrO2-DPP, with a ZrO2 content of 60 mg mL-1; right Photograph of a stable ZrO2 dispersion in THF (ZrO2-DPP) with a solid content of 60 mg mL-1
between 2650 and 2515 cm-1 are observed and are assigned to the H–O bond; such absorptions have been reported for various carboxylic acids [28]. The anchoring of SA to the ZrO2 surface through the COOH group is proven by the disappearance of these bands in the spectrum
J Mater Sci (2012) 47:2665–2674
2669
Fig. 5 FT-IR spectra of ZrO2 nanoparticles modified with DPP (above) and of pure DPP (below) recorded from 4000 to 500 cm-1
Fig. 3 TEM image of a highly diluted ZrO2-SA dispersion (at a magnification of 100000 times) showing quasi-monodisperse nanoparticles with a primary size of 6–8 nm
Fig. 4 FT-IR spectra of ZrO2 nanoparticles modified with SA (above) and of pure SA (below) recorded from 4000 to 500 cm-1
of ZrO2-SA. The disappearance of the absorption band of the C=O stretching vibration at 1674 cm-1 (characteristic for a C=O bond in conjugation with a double bond) [27] as well as the appearance of absorption peaks at 1518 and 1420 cm-1 indicate the formation of a carboxylate group, showing asymmetric stretching, and symmetric stretching signals at the respective wavenumbers. Their difference Dm = 98 cm-1 suggests that the carboxyl group acts as a bidentate chelate on the ZrO2 surface, and thus confirms the strong interaction between ligand and surface [26, 29, 30].
The FT-IR spectra of DPP and ZrO2-DPP are shown in Fig. 5. Typical absorption bands at *3065 cm-1 and *750 and 685 cm-1 are assigned to the sp2 H–C stretching vibrations as well as the out-of-plane deformations of the aryl moieties, respectively. The bands of the C=C stretching vibrations at *1590 and 1485 cm-1 are also clearly observed in both spectra. The signals at 1180 and 957 cm-1, in the DPP spectrum, are attributed to absorption from stretching vibrations of the P–O–(C) and (P)–O– C bonds of the P–O–Ar groups, respectively [31]. In case of the modified particles, these signals are shifted to higher frequencies between 1190–1230 cm-1 and 1065–1125 cm-1, being in consistence with the literature [32–35], and are broader. The peak broadening might have been caused by overlapping with absorption bands of the asymmetric and symmetric stretching vibrations of a PO2-grafting group that also absorbs in this region [31]. In addition, a strong absorption at 1020 cm-1 in the DPP spectrum is associated to P–O–(H) stretching vibrations. A similar band is also observed in the spectrum of ZrO2-DPP; however, the band is weak which might account to a small number of free P–OH moieties on the particle surface. The bonding of the phosphate ligand to the ZrO2 surface is thus explained through the diminution of this signal, and by the disappearance of the P=O absorption band at 1267 cm-1, as well as by the appearance of two bands assigned to the PO2-grafting group. This suggests an extensive bidentate and a diminutive monodentate grafting through condensation and coordination of the phosphoryl oxygen to the Lewis sites on the ZrO2 surface, as was reported by Guerro et al. [36] for the anchoring of phenyl phosphinates on TiO2 surfaces. In conclusion, it can be deduced that the DPP anchors to the ZrO2 surface through monodentate as well as bidentate DPP-ZrO2 bonding, as illustrated in Fig. 6. Also the presence of signals of the P–O–Ar groups in both the free ligand as well as the modified particles
123
2670
J Mater Sci (2012) 47:2665–2674
Fig. 6 Models of chemically surface modified ZrO2 nanoparticles a ZrO2-DPP and b ZrO2-SA
proves the preservation of the phenyl moiety on the particle surface. Based on the strong anchoring of both ligands to the particles, i.e., through the chelating of the SA, and the mono- and bidentate binding of the DPP, it is also possible to explain the good redispersibility of the ZrO2 NPs, in contrast to a rather poor redispersibility reported for allylmalonic acid-stabilized ZrO2 nanoparticles [26]. Thermogravimetric analysis TGA was employed to determine the amount of chemisorbed ligands on the ZrO2 NPs. Figure 7 presents the TGA curves as well as their corresponding differential curves for the ZrO2 powders before and after chemical modification with SA and DPP. The TGA curve of unmodified ZrO2 (ZrO2-UF) shows a total mass loss of 13.9 wt%, whereas the curves of ZrO2-SA and ZrO2-DPP show mass losses of 20.2 and 17.1 wt%, respectively. The DTG curve of ZrO2UF suggests two main mass-loss steps: a first step between 25 and 150 °C with a mass loss of *3 wt%, and a second step between 330 and 500 °C with a mass loss of *8.8 wt%. The first step is assigned to physisorbed water as well as volatile organic solvents, whereas the mass loss between 330 and 500 °C is assigned to chemisorbed and surface bound organics such as benzyl alcohol and derivatives [23, 25, 26]. For the chemically modified ZrO2 nanoparticles, the TGA curves present a moderate mass loss between 25 and 100 °C of 0.8 and 1.2 wt%, respectively. For ZrO2-SA, a significant mass loss of 17.1 wt% is observed between 200 and 500 °C, while for ZrO2-DPP the mass loss is further divided in two main steps from 100 to 200 °C (3.9 wt%), and from 300 to 500 °C (8.4 wt%). The mass loss above 100 °C observed for ZrO2-SA and ZrO2DPP is assigned to the chemisorbed ligands. Fabrication of the pre-nanocomposite For successful fabrication of nanocomposites with improved mechanical properties from the ZrO2 dispersions, it is inevitable that the solvent is completely removed from
123
Fig. 7 TGA curves of ZrO2 nanoparticles before and after chemical modification obtained by heating the samples under O2 atmosphere from 25 to 950 °C (curves were cut at 700 °C as no further loss was noted); representing a the mass loss and b their first derivatives
the pre-nanocomposite. This was achieved by evaporating the solvent from the pre-nanocomposite at elevated temperature under vacuum. In addition, to insure homogeneous distribution and good dispersion of the nanoparticles, the pre-nanocomposite was subjected to shearing in a dissolver. The solvent removal was monitored by TGA measurements under N2 atmosphere. It was found out that almost complete removal of the solvent was achieved using a rotary evaporator for 16 h. Nanoparticle distribution within the matrix The highly homogeneous distribution of the nanoparticles within the polymer matrix after the curing process can be observed in Fig. 8, where optically transparent nanocomposites possessing a ZrO2 content of 2–16 wt% are shown. It is notable that for the NC-ZrO2-DPP samples with ZrO2 content up to 8 wt% no turbidity is observed. The NCZrO2-SA samples, however, indicate high optical transparency only up to a ZrO2 content of 4 wt% (Fig. 8b). In
J Mater Sci (2012) 47:2665–2674
Fig. 8 Photograph of polymer samples a ZrO2-SA and b ZrO2-DPP with a content of 0, 2, 4, 8, and 16 wt% ZrO2
2671
nanocomposite comprises of finely distributed nanoparticles with a diameter below 10 nm [38]. The particle size distribution (seen in Fig. 10c) could be determined by employing the SAXS evaluation software GNOM [39] after obtaining a good fit (see Fig. 10b). The corresponding particle size distribution, assuming the particles as solid spheres, shows a D50 v median of 3.9 nm, suggesting that nearly all particles are dispersed as primary particles within the matrix. Tensile properties
Fig. 9 Transmittance curves obtained though UV–Vis spectrophotometry of cured nanocomposite samples with 0, 2, 4, 8, and 16 wt% ZrO2
addition, the transmission of the nanocomposites with a thickness of 4 mm was determined employing UV–Vis spectrophotometry. Figure 9 shows the transmission spectra of five nanocomposite samples with 0, 2, 4, 8, and 16 wt% ZrO2 fabricated from the dispersion ZrO2-DPP, in the visible range (i.e., 380–750 nm wavelength). Interestingly, the transmission of the nanocomposites with 2, 4, and 8 wt% is even higher at 690–750 nm than for the pure polymer (PP) sample, which might be explained through the low absorption indices of ZrO2 as well as its high refractive index [37], because the transparency increased with the ZrO2 content. For the nanocomposite sample with 16 wt%, however, a strong decrease in transparency is visible. At such high ZrO2 content, agglomeration of the particles within the matrix is expected, which would explain the turbidity. Furthermore, small angle X-ray scattering was performed to investigate the distribution of the nanoparticles within the matrix. Figure 10a shows the scattering profiles of the nanocomposite (black) after the scattering of the matrix (red) was subtracted from the raw data (blue). The smooth decay of the curve indicates that the
The tensile tests performed on pure polymer and the nanocomposites NC-ZrO2-SA and NC-ZrO2-DPP give information about the maximum tensile strength (MTS) and the Young’s modulus (see Table 1). The stress–strain curves for the matrix as well as nanocomposites with 2 wt% ZrO2 are displayed in Fig. 11. It is remarkable that the MTS is already increased by *9 and 12% for ZrO2-SA and ZrO2-DPP, respectively, despite the rather low filling of only 2 wt%. Several mechanisms that lead to such an increase in tensile properties are well known. These including crack deflection, crack pinning, debonding, plastic void growth, as well as particle pull-out and microcracking, and have been nicely summarized and studied by Zhao et al. in 2008 [6]. To investigate the increase in the tensile properties, fracture surface analysis was performed by atomic force microscopy (AFM). The results indicate that both nanocomposites possess higher surface roughness than the pure polymer. The surface roughness of the pure polymer, the NC-ZrO2-SA and the NC-ZrO2-DPP were 10, 22, and 30 nm, respectively; determined for an area of 10 9 10 lm2. The increase in roughness of the fracture surface suggests that during the fracture, crack deflection occurs. It is known that during crack deflection, an increase in the surface roughness is observed, which is caused by the growth and the divergence of the crack, resulting in higher energy absorption during deformation and thus causing an increased tensile strength [40]. In case of the Young’s modulus of the nanocomposites, the NC-ZrO2-DPP samples showed a significant increase of 12.5%, whilst the NC-ZrO2-SA samples showed almost no increase (0.4%) as compared to the pure thermoset samples. To explain these results, we considered the different mechanisms that influence the modulus of polymers and composites. For traditional composites, it is known that on addition of high-modulus filler to low-modulus polymer the modulus of the composite increases due to the load transfer from polymer to particles [12]. However, for nanocomposites with very small fillers, comparable to the polymerchain size or even smaller, the load transfer mechanism is difficult to explain. For such nanocomposites the change in
123
2672
J Mater Sci (2012) 47:2665–2674
Fig. 10 SAXS curves of a the nanocomposite (NC-ZrO2-SA), the polymer matrix and the raw signal, as well as b the NC-ZrO2-SA and the fitted curve; c size distribution derived from the fitted curve for NC-ZrO2-SA Table 1 Mechanical properties of the epoxy-based thermoset as well as the nanocomposites NC-ZrO2-SA and NC-ZrO2-DPP Sample
Maximum tensile strength (MPa)
Relative change (%)
Young’s modulus (MPa)
Relative change (%)
Pure polymer
69 ± 0.9
–
2630 ± 80
–
NC-ZrO2-SA
75 ± 0.5
9.3
2640 ± 60
0.4
NC-ZrO2-DPP
77 ± 3.8
11.9
2960 ± 120
12.5
modulus can also depend on other mechanisms, such as a change in degree or type of crystallinity in the polymer [41, 42], or a change in polymer-chain mobility which also affects the glass-transition temperature [12]. These mechanisms, on the other hand, depend on the polymer-particle interactions. It has been reported that strong polymer-particle interactions enhance the Young’s modulus [19] whereas weak interactions do not [20]. In addition, these interactions directly affect the mobility of the polymer chains and can be verified through the glass-transition temperature (Tg) of the composite. Therefore, DSC of the pure polymer and the nanocomposites was performed to determine the Tg of the samples. The Tg of the pure polymer, NC-ZrO2-DPP and NC-ZrO2-SA were determined to
123
91.4, 97.9, and 90.6 °C, respectively, thus indicating an apparent increase in Tg in case of NC-ZrO2-DPP and a slight decrease for NC-ZrO2-SA compared to the pure polymer. This suggests that through the change of the ligand on the particles, the interaction between the particles and the polymer can be positively or negatively altered, even if the particles are well-dispersed within the matrix in both cases. Thus, interaction-defining characteristics of the ligands must be their chemical structure that determines the type of attractive or repulsive forces within the interface. For ZrO2-DPP we assume that the free –OH groups of the phosphate might undergo hydrogen bonds or even covalent bonds (during curing) with the matrix. Also the phenyl moieties present on the surface of the ZrO2-DPP particles may lead to p–p interactions between the particle surface and the thermoset polymer backbone. p–p interactions are well known for aromatic compounds that have multiple aromatic rings adjacent to one another, which undergo p-stacking [43]; it is noteworthy that here we do not suggest a real p-stacking but just an interaction between the aromatic rings. In contrast, for ZrO2-SA there are no such moieties that would assist the interaction between polymer and particle.
J Mater Sci (2012) 47:2665–2674
2673
References
Fig. 11 Stress strain curves for open square pure polymer (PP), open circle NC-ZrO2-SA with 2 wt%, and open triangle NC-ZrO2-DPP with 2 wt% ZrO2. The Young’s modulus was determined from the strain 0 mm
Conclusion The fabrication of epoxy-based nanocomposites through solution processing of pre-fabricated ultra-small ZrO2 nanoparticles has been presented. We showed that through simple mixing of colloidal dispersions after chemical surface modification of the nanoparticles, nanocomposites with both high optical transparency and enhanced mechanical properties could be prepared. During the processing, removal of the THF is a critical point and requires thorough drying, as evidenced by thermogravimetry. The homogeneous distribution of the nanoparticles within the matrix after curing was proven by photographs, transmittance measurements, as well as SAXS, showing that the particles are present within the matrix individually, without a significant fraction of agglomerates. A minimal content of 2 wt% of ZrO2 nanoparticles modified with diphenyl hydrogen phosphate resulted in an increase in the maximum tensile strength of 11.9%, and an increase in the Young’s modulus of 12.5%. The effect of the chemical surface modification of the nanoparticles on the mechanical performance was shown by changing the surface bound ligand to sorbic acid. For these samples, an equable Young’s modulus and a decrease in the glass-transition temperature was shown, indicating a weaker interfacial interaction between the matrix and the filler. Acknowledgements We thank Ms Karin Kadhim at the Institute of Organic Chemistry, TU Braunschweig, for the FT-IR measurements, and Ms Bianca Tiedemann at the Institute of Technical Chemistry, TU Braunschweig, for the DSC measurements. We also thank Ms Rona Pitschke at the Max Planck Institute of Colloids and Interfaces in Potsdam Germany for the TEM measurements. Dr. Ce´leste A. Reiss at PANalytical B.V., Almelo, The Netherlands, is gratefully acknowledged for the SAXS measurements.
1. May CA (1988) Epoxy resins: chemistry and technology, 2nd edn. Dekker, New York 2. Ellis B (1993) Chemistry and technology of epoxy resins, 1st edn. Blackie, London 3. Wetzel B, Haupert F, Zhang MQ (2003) Compos Sci Technol 63:2055 4. Jordan J, Jacob KI, Tannenbaum R, Sharaf MA, Jasiuk I (2005) Mater Sci Eng A 393:1 5. Vaia RA, Maguire JF (2007) Chem Mater 19:2736 6. Zhao S, Schadler LS, Duncan R, Hillborg H, Auletta T (2008) Compos Sci Technol 68:2965 7. Gibson RF (2010) Compos Struct 92:2793 8. Giannakopoulos G, Masania K, Taylor AC (2011) J Mater Sci 46:327. doi:10.1007/s10853-010-4816-6 9. Rafiee MA, Yavari F, Rafiee J, Koratkar N (2011) J Nanopart Res 13:733 10. Kickelbick G (2005) Chem Unserer Zeit 39:46 11. Wen J, Wilkes GL (1996) Chem Mater 8:1667 12. Ajayan PM, Schadler LS, Braun PV (2005) Nanocomposite science and technology, 2nd edn. Wiley-VCH, Weinheim 13. Johnsen BB, Kinloch AJ, Mohammed RD, Taylor AC, Sprenger S (2007) Polymer 48:530 14. Ruiz-Pe´rez L, Royston GJ, Fairclough JPA, Ryan AJ (2008) Polymer 49:4475 15. Ochi M, Nii D, Suzuki Y, Harada M (2010) J Mater Sci 45:2655–2661. doi:10.1016/j.matchemphys.2011.04.034 16. Medina R, Haupert F, Schlarb AK (2008) J Mater Sci 43:3245. doi:10.1007/s10853-008-2547-8 17. Zhou S, Wu L (2008) Macromol Chem Phys 209:1170 18. Rai KN, Singh D (2009) J Compos Mater 43:139 19. Becker C, Mueller P, Schmidt HK (1998) SPIE Proceedings 3469:88 20. Ash BJ, Stone J, Rogers DF, Schadler LS, Siegel RW, Benicewicz BC, Apple T (2001) Mater Res Soc Symp Proc 661:KK2101KK2106 21. Pinna N, Garnweitner G, Antonietti M, Niederberger M (2005) J Am Chem Soc 127:5608 22. Garnweitner G, Niederberger M (2006) J Am Ceram Soc 89(6):1801 23. Garnweitner G, Goldenberg LM, Sakhno OV, Antonietti M, Niederberger M, Stumpe J (2007) Small 3(9):1626 24. Pinna N, Niederberger M (2008) Angew Chem Int Ed 47:2 25. Tsedev N, Garnweitner G (2008) Mater Res Soc Symp Proc 1076:1076-K05-03 26. Zhou S, Garnweitner G, Niederberger M, Antonietti M (2007) Langmuir 23:9178 27. Pavia DL, Lampman GM, Kriz GS (2001) Introduction to spectroscopy: a guide for students of organic chemistry, 3rd edn. Brooks/Cole, South Melbourne 28. Hesse M, Meier H, Zeeh B (2005) Spektroskopische Methoden in der organischen Chemie, 7th edn. Thieme, Stuttgart 29. Tackett JE (1989) Appl Spectrosc 43(3):483 30. Pawsey S, Yach K, Halla J, Reven L (2000) Langmuir 16:3294 31. Thomas LC (1974) Interpretation of the infrared spectra of organophosphorus compounds. Heyden, London 32. Randon J, Blanc P, Paterson R (1995) J Membr Sci 98:119 33. Gao W, Dickinson L, Grozinger C, Morin FG, Reven L (1996) Langmuir 12:6429 34. Schulmeyer T, Paniagua SA, Veneman PA, Jones SC, Hotchkiss PJ, Mudalige A, Pemberton JE, Marder SR, Armstrong NR (2007) J Mater Chem 17:4563 35. Lomoschitz CJ, Feichtenschlager B, Moszner N, Puchberger M, Mu¨ller K, Abele M, Kickelbick G (2011) Langmuir 27:3534
123
2674 36. Guerrero G, Mutin PH, Vioux A (2001) Chem Mater 13:4367 37. Krell A, Klimke J, Hutzler T (2009) Opt Mater 31:1144 38. Guinier A, Fournet G (1955) Small angle scattering of X-rays. Wiley-VCH, New York 39. Svergun DI (1992) J Appl Cryst 25:495 40. Hull D (1999) Fractography: observing, measuring and interpreting fracture surface topography, 1st edn. Cambridge Univ Press, Cambridge
123
J Mater Sci (2012) 47:2665–2674 41. Kojima Y, Usuki A, Kawasumi M, Okada A, Kurauchi T, Kamigaito O, Kaji K (1994) J Polym Sci Part B Polym Phys 32:625 42. Saujanya C, Radhakrishnan S (2001) Polymer 42:6723 43. Tallury SS, Pasquinelli MA (2010) J Phys Chem B 114:9349
