Epoxy networks reinforced with Ti[O.sub.2] generated by nonhydrolytic sol-gel process: a comparison between in situ and ex situ syntheses to obtain filled polymers.
The field of organic-inorganic hybrid materials has become of great interest thanks to the possibility to combine some peculiar mechanical and functional properties of both organic polymers and inorganic materials in a synergistic way by joining the different phases at a nanoscale. Organic-inorganic hybrid materials represent not only a new and exciting field of basic research, but also offer perspectives for several new applications in different technological fields. The very large set of accessible hybrid materials span a wide spectrum of properties which yield the emergence of innovative industrial applications in various domains such as optics, microelectronics, transportation, health, energy, housing, and the environment among others .
Concerning the specific area of thermoset or thermoplastic polymers reinforced/filled with metal oxide nanoparticles, it is well known that nanoparticles arising from flame pyrolysis or aqueous syntheses cannot be easily dispersed in organic oligomers or polymers by using conventional mechanical mixing processes, due to their hydrophilic character and strong tendency to agglomeration.
In this respect, organic-inorganic hybrid materials based on in situ generation of nanoparticles could represent an interesting alternative approach to design new composite materials. Among the different synthetic procedures , the sol-gel chemistry represents one of the preferred ways for the preparation of organic-inorganic hybrids according to the so-called bottom-up approach. The sol-gel process is a chemical method used for the synthesis of inorganic metal oxides, initially employed to synthesize high purity inorganic networks as glasses and ceramic materials. This method can be applied at room temperature or slightly higher, becoming strategic when thermallydegradable organic materials are involved in the process. The most used method is based on the so-called hydrolytic (or aqueous) sol-gel process, which is generally based on metal alkoxides as precursors and can be schematically divided into two steps: the first one named hydrolysis, which produces hydroxyl groups, and the second one named condensation, that involves the polycondensation of hydroxyl groups and residual alkoxyl groups to form a three-dimensional network , The presence in the reactive system of an organic oligomer or polymer (having or not suitable groups reactive toward to the sol-gel process) leads to the formation of organic-inorganic hybrid structures in which the inorganic oxide and the organic phases are intimately mixed each other.
Alternatively to the aqueous route, the so-called nonhydrolytic (or nonaqueous) sol-gel (NHSG) process can be used to obtain very pure and crystalline metal oxides . As well as the aqueous route, the NHSG is divided in two steps. The first step involves the reaction of a metal halide or a metal alkoxide with an organic oxygen donor (such as alcohols, ether, etc.). The second step (condensation) can follow different pathways depending on the employed oxygen donor. In principle, the NHSG route offers access to a wide range of organic-inorganic hybrid materials similar to those obtained by using the hydrolytic route. In practice, differences arising from the nature of the precursors, possible solvent choice and the different reaction mechanisms may dictate the type of hybrid, which can be prepared by either route. With respect to the hydrolytic sol-gel, it is recognized that NHSG process is potentially solvent-free, without problems with hydrophobic substances and particularly suitable for water-sensitive species. On the other hand, the formation of alkyl halide and/or alkyl ethers as by-products and the potential incompatibility with oxygen-containing species have to be taken into account as possible negative aspects .
The preparation of polymer-matrix nanocomposites exploiting the NHSG process for the generation of metal oxide nanoparticles was already proposed in the literature as described in the follows.
The preparation of polyfmethyl methacrylate) (PMMA)/titania composites was recently suggested by Messori and coworkers [6, 7] exploiting the reaction of Ti[Cl.sub.4] and benzyl alcohol or tert-butanol in the presence of preformed PMMA. The preparation of PMMA/silica composites was also proposed with a very similar approach through the reaction between Si[Cl.sub.4] and ethanol . In comparison with the samples prepared by traditional sol-gel route, these hybrids exhibit higher optical transparency and better thermal stability.
Transparent PMMA nanocomposites containing silica and titania with enhanced thermal stability were also prepared starting from methyl methacrylate and 3-(trimethoxysilyl)propyl methacrylate as organic monomers and tetraethoxysilane and titanium ethoxide as inorganic oxides precursors . A similar synthetic approach was also proposed by Du and coworkers [10-12] for the preparation of PMMA/silica/litania and PMMA/ silica/zirconia ternary nanocomposites with improved thermal and thermooxidative stabilities for potential applications in optical devices.
Polypropylene/silica , polyphenylsulfone/silica , polyimide/titania , and polyimide/silica/titania  are other examples of thermoplastic-matrix nanocomposites in which the filler nanoparticles were obtained through NHSG process.
Concerning thermoset organic matrices, silica-based organic-inorganic hybrid resins were synthesized by NHSG from 3-glycidoxypropyltrimethoxysilane and diphenylsilanediol (DPSD) at a fixed amount of phenyltrimethoxysilane (20 mol%) using barium hydroxide as a catalyst for potential use in encapsulating commercial phosphor powder for white light emitting devices (LEDs) , Also cycloaliphatic epoxy oligosiloxane resins were proposed as UV-curable resins for application in encapsulation of organic LEDs ,
The NHSG condensation reaction of methacryloxypropyl trimethoxysilane and DPSD was exploited in order to synthesize nano-structured units for the preparation of organic/inorganic hybrid coatings in the presence of suitable thermal and photoinitiators .
In a recent work, we reported an innovative procedure based on the preparation of titania nanoparticles suspended in benzyl alcohol and the mixing of this suspension with UV-curable cycloaliphatic epoxy resin . The cationic photopolymerization produced a three-dimensional network in which the suspending medium benzyl alcohol was covalently linked to the epoxy network according to the so-called "activated monomer" mechanism  during the propagation step in the cationic ring-opening polymerization. The presence of titania resulted in a reinforcing and stiffening effect due to both the presence of inorganic nanofillers and, most importantly, a higher crosslinking density of the composite material with respect to the pristine epoxy matrix.
In the present work, the same approach was used in order to verify the possibility to incorporate titania suspensions in benzyl alcohol in a cycloaliphatic epoxy resin by means of a thermally activated cationic polymerization instead of the UV-curing process. Furthermore, a comparison between thermally cured composites prepared by in situ generation of titania within the epoxy resin and by ex situ synthesized titania powder subsequently mechanically dispersed is reported.
3,4-Epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate epoxy resin (CE), titanium(IV) chloride (Ti[Cl.sub.4]), benzyl alcohol (BzOH), propylene carbonate (PC), ytterbium(III) trifluoromethanesulfonate hydrate [Yb[(OTf).sub.3]], chloroform, acetone, ethanol, and diethyl ether were purchased by Sigma Aldrich (Milan, Italy). All materials were high purity reactants and were used as received without any further purification.
Preparation of Titania Suspensions and Titania Powders
According to a previously reported  synthetic procedure, a given quantity of Ti[Cl.sub.4] (see composition details in Table 1) was added drop-wise to 1.50 g of BzOH at room temperature under vigorous stirring. The reaction was left stirring at room temperature for 15 min and then heated at 70[degrees]C for 24 h, by oil bath. A stable suspension was obtained after reaction. In order to prepare samples for the inorganic phase characterizations, a part of the obtained suspension was centrifuged at 4000 rpm for 15 min. The powders were carefully washed twice using chloroform, twice by acetone and once by diethyl ether, sonicating for 10 min every time that fresh solvent was added. Finally powders were dried at 60[degrees]C for 8 h under dynamic vacuum.
Preparation of Epoxy-Titania Composites
A typical thermally-curable formulation was prepared by mixing titania suspension, CE epoxy resin and the solution of Yb[(OTf).sub.3] in PC (1:3 wt/wt ratio) as thermal cationic initiator (see composition details in Table 2) by using a T18 UltraTurrax[R] lka disperser (5 min mixing time) followed by 15 min treatment in an ultrasonic bath. The formulations were cast into silicone molds (having cavity dimensions of 4 X cm 1 X cm 0.4 cm) and subsequently degassed by dynamic vacuum in order to avoid bubbles formation in the final specimens. All the formulations were cured at 120[degrees]C for 25 min and as postcuring 20 min long at 160[degrees]C was subsequently applied. The materials were coded as CE_Tx in which x represents the nominal amount of titania [expressed as part per hundreds part of resin (phr)] calculated from stoichiometry by assuming the complete conversion of Ti[Cl.sub.4] to Ti[O.sub.2].
According to ex situ synthesis procedure, a further set of composites were also prepared by mechanical mixing of a suitable amount of titania powders (obtained from S25 suspension) with BzOH (1.50 g), CE (5.0 g), and Yb[(OTf).sub.3]/PC solution (0.15 g). Titania powders were used both as obtained and after calcination in air at 500[degrees]C for 5 h (the calcined powder was coded as T_S25_C) in order to eliminate the organic substances and/or hydroxide groups present on the particles' surface. The same curing conditions of CE_Tx materials were applied. The final materials were coded as CE_T5_ex and CE_T5_ex_C, respectively.
Characterization of Titania Powders
The synthesized powders were analyzed by computer-assisted conventional Bragg-Brentano diffractometer using the Ni-filtered CuK[alpha] monochromatic radiation ([delta] = 1.5418 [Angstrom]; X'Pert PRO, Panalytical) to identify the crystalline phase. The X-ray diffraction (XRD) patterns were collected at room temperature in a 20 range of 10-90[degrees], with a scanning rate of 0.005[degrees] x [s.sup.-1] and a step size of 0.02[degrees]. Crystallites size was estimated from X-ray line broadening measurements; it was calculated by using the Ti[O.sub.2] (1,0,1) diffraction line according to the Scherrer formula (Eq. 1) :
D = 0.9 x [delta]/[beta] x cos [theta] (1)
where D is the crystallite diameter, [delta] is the wavelength of the X-ray, [theta] is the diffraction angle, and [beta] is the full width at half maximum.
The particles morphologies were examined by transmission electron microscopy, JEM 2010 TEM (Jeol), operating at 200 kV. A drop of the so-obtained reaction mixture was diluted in ethanol (2 mL) and sonicated (15 min); the final suspension was centrifuged for 1 min. A drop of the supernatant was placed on a copper grid (Carbon Film 200 Mesh Cu; Agar Scientific) followed by drying at 40[degrees]C. Moreover, the obtained images were analyzed and processed by ImageJ open-source software in order to obtain an average particles size.
The amount of organic fraction was evaluated by gravimetric analysis measuring the percentages of weight loss after a treatment at 500[degrees]C for 3 h in an electric furnace (Optolab--FS5), up to constant mass was reached.
The qualitatively investigation of the residual organic groups and/or free -OH chemically bonded on the particles surfaces was performed by Fourier transform infrared spectroscopy (FTIR) analysis on the obtained powders. The analysis was carried out using an FT-IR VERTEX 70 spectrometer (Bruker) in ATR mode from 4000 to 500 [cm.sup.-1], equipped with a diamond crystal.
The specific surface area (SSA) and density ([rho]) of the powders were determined by the BET method (Gemini 2360 apparatus, Micromeritics) and by a picnometer (Accupic 1330 apparatus, Micromeritics), respectively. The samples were carefully dried using the before described procedure.
Characterization of Epoxy-Titania Composites
The actual titania contents in the composites were obtained by gravimetric analysis, evaluating the amount of titania after the treatment at 500[degrees]C for 4 h in an electric furnace (Optolab--FS5), up to constant mass was reached.
FT-IR spectroscopy was performed using an Avatar 330 FT-IR Thermo Nicolet spectrometer, equipped with a diamond crystal, operating in the ATR mode from 4000 to 500 [cm.sup.-1] (32 scans and resolution of 4 [cm.sup.-1]). Epoxy groups conversion ([alpha]) values were determined on the basis of the signal at 910 [cm.sup.-1], corresponding to epoxy groups of CE, normalized with respect to the carbonyl signal at 1730 [cm.sup.-1], corresponding to carbonylic groups of CE.
[alpha] was calculated by the following equation:
[alpha] = 100 - ([A.sub.t]/[A.sub.0] x 100) (2)
where [A.sub.0] is the normalized area of the signal at 910 [cm.sup.-1] of the mixture before the curing process and A, is the normalized area of the signal at 910 [cm.sup.-1] after the curing process.
Extraction tests were carried out by the immersion of specimens (about 0.30 g each), wrapped in a fine net, in 50 ml of chloroform at room temperature for 24 h. The samples were then dried to a constant mass (i.e., the dried mass [m.sub.d]), and the absolute extractable fraction (f) was determined as follows:
f = [m.sub.0] - [m.sub.d]/[m.sub.0] x 100 (3)
where [m.sub.0] is the mass of the sample before its immersion in chloroform.
The values of f were also normalized to the actual organic phase content (i.e., the extractable phase in the case of incomplete cross-linking), and these values were called [f.sub.ORG]. Their values were calculated as follows:
[f.sub.ORG] = f/[w.sub.ORG] (4)
where [w.sub.ORG] is the weight fraction of organic phase present in the composite.
The prepared samples were broken in liquid nitrogen and the cross-sections were analyzed by field emission scanning electron microscopy (FESEM, FEI STRATA DB235M Dual beam FIB-SEM Instrument) by the application of an accelerating voltage of 15 kV. The samples (cross-sections) were previously coated with gold (thickness 10 nm) by an electro-deposition method to impart electrical conduction. The obtained images were analyzed and processed by ImageJ open-source software.
Dynamic-mechanical thermal analyses (DMTA) were carried out on an MK III Rheometrics Scientific Instrument at 1 Hz frequency in tensile configuration employing 5[degrees]C [min.sup.-1] heating rate. The storage modulus, E', and the loss factor, tan [delta], were measured from 0[degrees]C up to the temperature at which the rubbery state was attained. The glass transition temperature ([T.sub.g,DMTA]) was assumed as the maximum of the loss factor curve.
Differential scanning calorimetry (DSC) was carried out by Thermal Analysis TA2010 Instruments at a scanning rate of 5[degrees]C x [min.sup.-1] from 0 to 200[degrees]C under nitrogen flow. The glass transition temperature ([T.sub.g,DSC]) was assumed as the mean value of the energy jump of the thermogram (average value between the onset and the endpoint of the glass transition range).
The filler volume fractions of the cured nanocomposites were calculated taking into account the experimentally determined density values of pristine epoxy "CE-TO" (1.2021 g x [cm.sup.-3]) and titania particles under the assumption of density additivity.
RESULTS AND DISCUSSION
Characterization of Titania Powders
XRD analyses (Fig. 1) show that titania powders, obtained after washing and drying the corresponding suspensions, are characterized by the presence of a crystalline phase (anatase ICPDS file 01-075-1537). The sample T_S50 shows, centred at 18[degrees] 2[theta], a reflection probably due to an unknown phase, attributable to the presence of an organic phase on the particles' surface, which decreases by increasing the BzOH/Ti[Cl.sub.4] molar ratio.
The average crystals size of the T_S15, T_S25, and T_S50 powders (Table 3), calculated by the Scherrer's equation (Eq. 1), is around 10 nm showing an almost negligible increase when the BzOH/Ti[Cl.sub.4] molar ratio decreases. As expected, the T_S25_C powder (calcined sample) showed an increase in the average crystals size, due to the treatment at high temperature.
The shape and dimensions of the primary particles were also evaluated by TEM. The obtained images of T_S15, (as representative example) and T_S25_C (Fig. 2a and b) show that the synthesized powders are nanostructured having rather uniform shape and sizes. The evaluated average particle diameters by image analysis is 10 [+ or -]4, 10 [+ or -] 3, and 20 [+ or -] 5 nm for T_S15, T_S50, and T_S25_C powders, respectively; these results are consistent with those calculated by Scherrer's equation (Eq. 1).
Taking into account that the adopted washing step to obtain the titania powders should ensure the complete elimination of unlinked organic substances, the presence of an organic and/or volatile fraction chemically linked and/or strongly adsorbed to the titania particles was quantified by gravimetric analysis. The inorganic residue values reported in Table 3 show that the powders obtained from S15 and S25 suspensions were comprised of about 70 wt% of inorganic (titania) phase while the powder obtained from S50 suspension contained a significantly higher organic content (about 76 wt%).
The presence of titanol groups (Ti-OH) and organic phase was also confirmed by FT-IR analysis (Fig. 3), which are probably due to the uncondensed -OH and benzyl groups. The FT-IR spectra of the T_S15 and T_S25 show a large band centred around 3200 [cm.sup.-1] due to the stretching vibration of the -OH on the particles surface and related bending at 1600 [cm.sup.-1], whereas below 900 [cm.sup.-1] the typical broad band due to Ti-O-Ti vibrations was observed . Moreover, Fig. 3 clearly shows that T_S50 is functionalized by residual benzyl groups presenting the common peaks such as C-H aromatic stretching between 3100 and 3000 [cm.sup.-1], C-H aliphatic stretching 2950-2800 [cm.sup.-1] and the related bending vibration in the range 1500-1400 [cm.sup.-1] that match perfectly with the benzyl alcohol spectrum.
Whereas the calcined sample (T_S25_C) showed only the broad band due to Ti-O-Ti vibrations starting at 900 [cm.sup.-1]. As expected, intensity of the signals, above described, increases when the BzOH/ Ti[Cl.sub.4] ratio decreases, consolidating the XRD and gravimetric analysis data that show, in the case of T_S50, the presence of unknown phase (probably due to the uncondensed benzyl groups) and the highest mass loss, respectively.
On the basis of these evidences, the experimental density ([[rho].sub.exp]) values of the powders determined by picnometer are relatively low, ranging from 2.60 to 1.39 g x [cm.sup.-3], and they can be considered as deriving from hybrid materials comprised of titania and organic substances presumably covering the surface of the inorganic particles. In order to determine the actual density of the titania particles without the contribution of the organic component, density values were corrected by assuming BzOH as main component of the organic phase (with a density [[rho].sub.BzOH] = 1-045 g x [cm.sup.-3]) and the density additivity rule, according to the following equation:
[[rho].sub.exp] = [[rho].sub.TiO2] x [w.sub.TiO2] + [[rho].sub.BzOH] x [w.sub.BzOH] (5)
in which [w.sub.TiO2] and [w.sub.BzOH] are the weight fraction of titania and BzOH, respectively. These corrected values were used for the calculation of the volume composition of the prepared epoxytitania nanocomposites.
SSA values range from a maximum of 118 [m.sup.2] x [g.sup.-1] for titania powders obtained from S15 suspension to a minimum of 26 [m.sup.2] x [g.sup.-1] for titania powders obtained from S50 suspension. This could be caused by the residual organic substance that creates a real capping on the particles, which is able to decrease SSA making less available the particles surface. That is consistent with results found by XRD, FT-IR, and gravimetric analyses that showed an increase in this organic layer when the BzOH/ Ti[Cl.sub.4] ratio decreases.
Characterization of Epoxy-Titania Composites
The gravimetric analyses showed actual titania contents (values in Table 2) are rather closed to the nominal ones proving that the employed procedure is an effective alternative to produce composites with respect to the conventional mixing methods.
FTIR analysis was carried out in order to evaluate the conversion degree of epoxy groups (Table 4). A relative high conversion degree was obtained for all samples indicating the efficiency of the curing process based on the cationic ring opening polymerization mechanism thermally initiated by Yb[(OTf).sub.3]. Epoxy groups conversion values ranged from 75 to 79% almost independently on the curable investigated formulation, suggesting that the presence of titania did not significantly influence the curing process. These maximum values of epoxy group conversion were presumably limited by the vitrification of network which inhibited any further progress of the reaction.
Extractable fraction values were determined in order to evaluate the effect of the presence of the titania suspension on the cross-linked structure (see data reported in Table 4). Taking into account that the conversion of epoxy groups was not quantitative, f values were relative low with values ranging from 7 to 13 for the CE_Tx series. In this case it is interesting to observe that the extractable fraction was always significantly lower than the nominal amount of BzOH present in the formulation (in the range 21-22 wt% depending on the titania content) suggesting the formation of a relatively well developed three-dimensional network even in the case of high content of BzOH. This result suggested that both "active chain end" and "activated monomer" mechanisms (see Scheme 1) occurred during the propagation step in the cationic ring-opening polymerization leading to a covalent incorporation of BzOH (and corresponding by-products of NHSG reaction) within epoxy network.
It is also interesting to observe that both f and [f.sub.ORG] values systematically decreased by increasing the amount of titania. In this respect, the presence of in situ generated titania particles resulted in a positive contribution to the development of a network structure, presumably due to the formation of a stronger interphase between filler and organic matrix.
The significance of the filler in situ generation can be clearly underlined considering the composites CE_T5_ex and CE__T5_ex_C prepared by the ex situ synthesized powders T_S25 and T_S25_C (not calcined and calcined, respectively). In this case the epoxy groups conversion values (75-79%) are very similar to those obtained for the CE_Tx series indicating a negligible effect on the kinetics of epoxy groups reaction. On the contrary, a significantly higher amount of extractable phase was observed for both samples suggesting that the addition of ex situ generated particles did not contribute to the development of a cross-linked structure.
FESEM images of the composites permitted to evaluate the distribution and aggregation of titania particles in the matrix. Typical FESEM micrographs of fractured surface epoxy/titania composites (CE_5T, CE_10T, and CE_5Tex_C as representative example) are reported in Fig. 4. All samples showed the presence of a dispersed phase rich in titanium, as indicated by EDX analysis (here not reported), attributable to titania. The samples prepared by in situ generated Ti[O.sub.2] showed the formation of small aggregates of some titania particles, but presenting anyway very homogeneous filler distribution. The sample CE_5T (Fig. 4a) showed an average particles size of 95 [+ or -] 22 nm, whereas bigger particles, ranging from 580 nm to 1.3 [micro]m, were detected in the case of CE_10T (Fig. 4b), phenomenon that was probably due to the usual tendency of titania nanoparticles to aggregate . The sample prepared embedding ex situ synthesized calcined powder (CE_5Tex_C in Fig. 4c) shows that the average particles size increases compared to CE_5T, but the most visible change is the filler dispersion, which is less homogeneous leading to the formation of big aggregates. This is probably due to the calcination that eliminates the organic phase on the particles surface that enhances the interaction between organic and inorganic phase and partially avoids the particles agglomeration improving the dispersion.
Glass transition temperature ([T.sub.g]), damping (expressed as maximum value of loss factor, tan [[delta].sub.max]), and storage modulus (E') values obtained by DMT A are reported in Table 4.
The obtained results showed that the dynamic-mechanical properties of composites were significantly affected by the titania content. It is usually expected that, in the case of polymers filled with particles, the incorporation of rigid fillers into a polymeric matrix made difficult the movement of the polymer chains leading to a damping decrement and a shift of [T.sub.g] values to higher temperatures. Also in the present case, a significant reduction of damping (quantified as tan [[delta].sub.max]) was observed by increasing the titania concentration in the composites, together with a significant increment in [T.sub.g] values. The glass transition temperature of unfilled epoxy (95[degrees]C) increased up to 114, 106, and 119[degrees]C for CE_T3, CE_T5, and CE_T10, respectively. As already mentioned, such increases in Tg can be attributed to the constrained chain mobility by well-dispersed fillers. A rather similar trend (Table 4) was also observed by DSC analysis that confirmed the increase in [T.sub.g] above described.
[FORMULA NOT REPRODUCIBLE IN ASCII]
The general increase in [T.sub.g] and decrease in loss factor tan [[delta].sub.max] by increasing the titania content can be also considered as an indirect evidence of the presence of a strong filler-matrix interface. Kim at al.  reported the preparation and the characterization of epoxy-based composites containing inorganic nanoparticles functionalized or not functionalized with suitable coupling agents and reported that composites with weak filler-matrix interface exhibit essentially no modification in [T.sub.g] and damping with filler contents, opposite to composites with strong filler-matrix interface which show a systematic increase in [T.sub.g] and decrease in loss factor with filler content. Accordingly, it is interesting to observe that CE_T5_ex (ex situ composite with a weaker filler-matrix interface) showed a tan [[delta].sub.max] of 0.904, very similar to the value of unfilled resin CE_T0 (0.843) and much higher than the values of in situ composites CE_T3, CE_T5 and CE_T10 (tan [[delta].sub.max] in the range 0.410-0.440). The storage moduli E' measured in the glassy state (that is at a temperature of 40[degrees]C) were influenced by the presence and the amount of titania. [E'.sub.T=40[degrees]C] values for composites were always higher than the unfilled one CE_T0 (0.63 GPa) with the maximum value of 1.89 GPa for composite with intermediate filler content (CE_T5). A stronger influence due to the presence of titania was observed in the rubbery state. In this respect, it is very interesting to note that the storage moduli E' measured above the glass transition temperature (that is at a temperature of [T.sub.g] + 50[degrees]C) increased significantly by increasing the titania content. [E'.sub.T=Tg+50[degrees]C] values of CE_T0 (0.78 MPa) raised up to 5.19, 12.59, and 8.45 MPa for CE_T3, CE_T5, and CE_T10, respectively.
The reduced storage modulus (that is the ratio between the moduli of the composite and of the matrix, respectively) measured in the rubbery region as a function of titania volume fraction is reported in Fig. 5. In order to compare the experimental results with data predicted by the well-known models for composites materials, the generalized Kerner equation for the reduced modulus of filled polymers was applied in order to evaluate its dependence on the nanoparticles content:
E'/[E'.sub.1] = 1 + A x B x [[phi].sub.2]/1 - B x [PSI] x [[phi].sub.2] (6)
in which E' and [E'.sub.1] are the storage moduli of composite and unfilled epoxy, respectively.
The constant A, in the case of spherical filler particles and for any Poisson's ratio v of the matrix, is defined as:
A = 7 - 5v/8 - 10v. (7)
The constant B depends on the ratio between filler and matrix moduli and it can be approximated to 1 for very large moduli ratios. The parameter [PSI] is a reduced concentration term, which depends on the maximum packing fraction of the particles ([[phi].sub.m]) according to the following definition:
[PSI] = 1 + [1 - [[phi].sub.m]/[[phi].sup.2.sub.m]] x [[phi].sub.2] (8)
in which [[phi].sub.2] is the filler volume fraction.
In the present study, the Poisson's ratio v of the matrix and the maximum packing fraction of the particles [[phi].sub.m] were taken equal to 0.5 (epoxy in the rubbery state) and to 0.601 (random loose packing, non agglomerated packing configuration), respectively .
Data reported for CE_Tx series indicated a very important increment of reduced storage modulus with respect to the values predicted by the generalized Kerner equation (Eq. 6). Taking into account that, in the rubbery region temperature range, the modulus values are mainly governed by the cross-linking density of the network, a considerable increase in this last parameter has to be considered in the presence of in situ generated titania. In other words, titania nanoparticles generated by NHSG process acted not only as rigid reinforcing filler but also as cross-linking points, increasing the cross-linking density of the composite material with respect to the pristine epoxy matrix.
Once again, the ex situ composites CE_T5_ex and CE_T5_ex_C (in Fig. 5, unfilled triangle and X, respectively) showed a completely different behavior with a reduced storage modulus very similar to the predicted value.
In order to further underline the differences between in situ and ex situ composites, the storage modulus (E') as a function of the temperature is reported in Fig. 6 for CE_T5, CE_T5_ex, and CE T5_ex_C. Data showed a dramatic difference in storage modulus in the rubbery region for in situ and ex situ composites even if containing the same amount of filler. Remembering once again that the modulus in the rubbery region is mainly determined by the cross-linking density, dynamic-mechanical analysis supported the hypothesis that titania nanoparticles in situ generated increased the cross-linking density of the composite with respect to both the unfilled epoxy matrix and the ex situ composites.
In order to further investigate this aspect, the network densities (that is the average molecular weight between two adjacent cross-linking points, [M.sub.C]) were determined in the rubbery region from the storage modulus values at [T.sub.g] + 40[degrees]C by using the following relationship:
[E'.sub.Tg+40[degrees]C] = 3 x q x n x RT = 3 x q x ([rho]/[M.sub.c]) x RT (9)
where q is the front factor (usually equal to 1), n is the apparent cross-linking density, R is the gas constant (R = 8.314 J x [K.sup.-1] x [mol.sup.-1]), [rho] is the density of the material, and T is the absolute temperature in K , Originally, Eq. 10 was used for single phase materials, but it also gives very good results for biphasic systems . However, the values obtained by this formula should be intended only for comparing the network density of the samples investigated in the present work.
The calculated values of polymer molecular weight between two cross-linking points, [M.sub.C], are reported in Table 4. The network density increases and [M.sub.C] decreases for CE_Tx series with respect to the unfilled epoxy resin CE_T0. Accordingly to the above discussed data, the ex situ composites CE_T5_ex and CE_T5_ex_C showed significantly higher [M.sub.C] values.
Suspensions of titania nanoparticles in benzyl alcohol were synthesized by means of nonhydrolytic sol-gel process and mixed with a cycloaliphatic epoxy resin and subsequently cross-linked through a thermally activated cationic polymerization.
The extraction tests and the dynamic-mechanical analysis showed a significantly different behavior between the composites obtained by in situ generation of titania or by simple mechanical mixing of the same purified and dried titania powders with or without a calcination treatment. The in situ generation of Ti[O.sub.2] and subsequent polymerization of the suspending medium resulted in a very strong filler-matrix interaction with a significant improvement in the cross-linking density due to the titania contribution, which in turn incremented the glass transition temperature and the storage modulus values in the rubbery region with the filler content. Moreover, both samples prepared by embedding ex situ synthesized filler showed worse mechanical properties than equivalent sample prepared by in situ filler generation, supporting the hypothesis that the particles in situ generated are able to act not only as a rigid filler but also a "cross-linking densifier," leading to an increase in cross-linking density.
[1.] C. Sanchez, P. Belleville, M. Popall, and L. Nicole, Chem. Soc. Rev., 40, 696 (2011).
[2.] C. Sanchez, B. Julian, P. Belleville, and M. Popall, J. Mater, diem., 15, 3559 (2005).
[3.] C.J. Brinker and G.W. Shrerer, "Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing," Academic Press, San Diego (1990).
[4.] I. Bilecka and M. Niederberger, Electrochim. Acta, 55, 7717 (2010).
[5.] J.N. Hay and H.M. Raval, Chem. Mater., 13, 3396 (2001).
[6.] D. Morselli, F. Bondioli, M. Fiorini, and M. Messori, J. Mater. Sci., 47, 7003 (2012).
[7.] D. Morselli, M. Messori, and F. Bondioli, J. Mater. Sci., 46, 6609 (2011).
[8.] X.M. Song, X.X. Wang, H.T. Wang, W. Zhong, and Q.G. Du, Mater. Chem. Phvs., 109, 143 (2008).
[9.] H.C. Kuan, S.L. Chiu, C.H. Chen, C.F. Kua, and C.L. Chiang, J. Appl. Polym. Sci., 113, 1959 (2009).
[10.] H.T. Wang, S. Meng, P. Xu. W. Zhong, and Q.G. Du, Polym. Eng. Sci., 47, 302 (2007).
[11.] H.T. Wang, P. Xu, S. Meng, W. Zhong, W.C. Du, and Q.G. Do, Polym. Degrad. Stab., 91, 1455 (2006).
[12.] H.T. Wang, P. Xu, W. Zhong, L. Shen, and Q.G. Du, Polym. Degrad. Stab., 87, 319 (2005).
[13.] J. Qian, H.G. Zhang, G.C. Cheng, Z.J. Huang, S.Y. Dang, and Y.S. Xu, J. Sol-Gel Sci. Technol., 56, 300 (2010).
[14.] S. Licoccia, M.L. Di Vona, A. D'Epifanio, Z. Ahmed, S. Bellitto, D. Marani, B. Mecheri, C. de Bonis, M. Trombetta. and E. Traversa, J. Power Sources, 167, 79 (2007).
[15.] M.A. Saeed, Z.H. Lodhi, A.-u. Khan and W. Asghar, Adv. Mater. Res. (Durnten-Zurich, Switz.), 326, 88 (2011).
[16.] H. Wang, W. Zhong, P. Xu, and Q. Du, Composites Part A, 36, 909 (2005).
[17.] S. Jana, M.A. Lim, I.C. Baek, C.H. Kim, and S. II Seok, Mater. Chem. Phys., 112, 1008 (2008).
[18.] K. Jung, J.Y. Bae, S.J. Park, S. Yoo, and B.S. Bae, J. Mater. Chem., 21, 1977 (2011).
[19.] S. Dire, V. Tagliazucca, G. Brusatin, J. Bottazzo, I. Fortunati,
R. Signorini, T. Dainese, C. Andraud, M. Trombetta, M.L. Di Vona, and S. Licoccia, J. Sol-Gel Sci Technol., 48, 217 (2008).
[20.] D. Morselli, F. Bondioli, M. Sangermano, and M. Messori, Polymer, 53, 283 (2012).
[21.] P. Kubisa, J. Polym. Sci. Part A-Polym. Chem., 41, 457 (2003).
[22.] P. Klug and L.E. Alexander, X-Ray Diffraction Procedure, Wiley, New York (1954).
[23.] M.J. Velasco, F. Rubio, J. Rubio, and J.L. Oteo, Thermochim. Acta, 326, 91 (1999).
[24.] S. Kang, S.I. Hong, C.R. Choe, M. Park, S. Rim, and J. Kim, Polymer, 42, 879 (2001).
[25.] L.E. Nielsen and R.F. Landel, Mechanical properties of polymers and composites, 2nd ed., Marcel Dekker Inc., New York (1994).
[26.] E. Vassileva and K. Friedrich, J. Appl. Polym. Sci., 89, 3774 (2003).
[27.] G. Levita, S. Depetris, A. Marchetti, and A. Lazzeri, J. Mater. Sci., 26, 2348 (1991).
Davide Morselli, (1,2) Federica Bondioli, (2,3) Marco Sangermano, (2,4) Massimo Messori (1,2)
(1) Department of Engineering "Enzo Ferrari", University of Modena and Reggio Emilia, Via Vignolese 905/A, 41125 Modena, Italy
(2) Italian Interuniversity Consortium on Materials Science and Technology (INSTM), Via G. Giusti 9, 50121 Firenze, Italy
(3) Department of Industrial Engineering, University of Parma, Parco Area delle Scienze 181 /A, 43124 Parma, Italy
(4) Department of Applied Science and Technology, Polytechnic of Turin, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
Correspondence to: Massimo Messori; e-mail: email@example.com
TABLE 1. Composition of the prepared suspensions. Ti[O.sub.2] BzOH/Ti suspension Ti[Cl.sub.4] BzOH [Cl.sub.4] Nominal code (g) (g) molar ratio Ti[O.sub.2] (g) S15 0.36 1.50 7.3 0.15 S25 0.59 4.5 0.25 S50 1.19 2.2 0.50 TABLE 2. Composition of the prepared thermally- curable composite formulations. Composite Yb[(OTf).sub.3]/ code CE (g) PC (g) Ti[O.sub.2] suspension CE_T0 5.0 0.15 only BzOH (1.50 g) CE_T3 SI5 CE_T5 S25 CE_T10 S50 CE_T5ex only BzOH (1.50 g) CE_T5ex_C only BzOH (1.50 g) Composite Added Ti[O.sub.2] Nominal Ti[O.sub.2] code powder (g) Ti[O.sub.2] (phr) (phr) (a) CE_T0 -- -- -- CE_T3 -- 3 2.6 CE_T5 -- 5 4.1 CE_T10 -- 10 7.6 CE_T5ex 0.385h 5 4.4 CE_T5ex_C 0.270b 5 4.8 (a) Actual titania content experimentally determined by gravimetric analysis. (b) Ti[O.sub.2] powders were added in different amount considering that S25 powder and T_S25_C powder have different amount of organics due to the calcination. TABLE 3. Properties of titania powders obtained from the corresponding suspension; mean crystallite size as determined by XRD pattern elaboration, specific surface area (SSA), residue after organic phase combustion and experimental and corrected densities ([[rho].sub.exp], [[rho].sub.TIO2]) corrected on the basis of the actual inorganic content. Residue at Ti[O.sub.2] Average crystal 500[degrees] [[rho].sub.exp] powder code size by Eq. 1 (nm) C (wt%) (g x [cm.sup.3]) T_S15 9 70.4 2.60 T_S25 10 69.6 2.01 T_S50 11 24.1 1.39 T_S25_C (a) 26 -- 3.35 [[rho].sub.TiO2] SSA Ti[O.sub.2] by Eq. 5 ([m.sup.2] x powder code (g x [cm.sup.-3]) [g.sup.-1) T_S15 3.25 118 T_S25 2.44 48 T_S50 2.49 26 T_S25_C (a) 3.35 47 (a) Powder calcined at 500[degrees]C for 3 h. TABLE 4. Properties of the thermally cured composites: epoxy groups conversion ([alpha]), absolute and normalized extractable fractions (f and [f.sub.ORG]), glass transition temperature from DSC ([T.sub.g,DSC]), glass transition temperature from DMTA ([T.sub.g,DMTA]), damping (expressed as maximum value of tan S, tan <$max), storage modulus measured at T = 40[degrees]C and T = [T.sub.g] + 50[degrees]C ([E'.sub.T] = 40[degrees]C and [E'.sub.T = Tg + 50[degrees]C]), and network density (average molecular weight between two adjacent crosslinking points, [M.sub.C]). Composite [f.sub.ORG] [T.sub.g,DSC] code [alpha] (%) f (wt%) (wt%) ([degrees]C) CE_T0 80.9 13.4 13.4 51 CE_T3 74.9 11.5 11.8 71 CE_T5 78.4 8.0 8.3 87 CE_T10 76.8 6.5 7.0 101 CE_T5_ex 74.9 25.5 26.7 52 CE_T5_ex_C 78.8 18.2 19.0 50 Composite [T.sub.g,DMTA] tan [[delta] [E'.sub.T = 40 code ([degrees]C) .sub.max] [degrees]C] (Pa) CE_T0 95 0.843 6.33E+08 CE_T3 114 0.440 1.78E+09 CE_T5 106 0.410 I.89E+09 CE_T10 1 19 0.412 I.26E+09 CE_T5_ex 107 0.904 I.65E+09 CE_T5_ex_C 115 0.690 I.63E+09 Composite [E'.sub.T = Tg + 50 [M.sub.C] code [degrees]C] (Pa) (g [mol.sup.-1]) CE_T0 7.85E+05 3600 CE_T3 5.19E+06 660 CE_T5 I.26E+07 260 CE_T10 8.45E+06 430 CE_T5_ex I.62E+06 1920 CE_T5_ex_C I.66E+06 2010
|Printer friendly Cite/link Email Feedback|
|Author:||Morselli, Davide; Bondioli, Federica; Sangermano, Marco; Messori, Massimo|
|Publication:||Polymer Engineering and Science|
|Date:||Jul 1, 2015|
|Previous Article:||Effect of blend ratio and elongation flow on the morphology and properties of epoxy resin-poly(trimethylene terephthalate) blends.|
|Next Article:||Process-induced phase and crystal morphologies in water-assisted injection molded polypropylene/polymeric [beta]-nucleating agent blend parts.|