N-Phenylaminomethyl hybrid silica, a better alternative to achieve reinforced PU nanocomposites

Abstract

A novel N-phenylaminomethyl silica (NPAM-silica) was firstly synthesized by the one-step emulsion polymerization of N-phenylaminomethyltriethoxysilane (ND-42). The hybrid silica could be dispersed in a variety of organic solvents including THF, acetone, DMF and DMAC. The NPAM-silica was mainly composed of fully condensed silsesquioxanes and a small quantity of partially condensed silsesquioxanes, which was confirmed by ²⁹Si NMR and MALDI-TOF MS. GPC showed that the polydispersity of the NPAM-silica was 1.08. It was then introduced into polyurethane (PU) as a chemical cross-linker. The high reactivity of NPAM-silica was investigated by means of solid state Nuclear Magnetic Resonance (SSNMR). TEM and AFM showed that the NPAM-silica was nano-dispersed in the hard segments region of PU. The mechanical properties of the NPAM-silica/PU nanocomposites in the rubber state were enhanced when NPAM-silica was added, which was reflected in the DMA and tensile tests at constant T_test − T_g. It is also concluded that the increase of the modulus was derived from both the nanofiller contribution and entropic contribution which were tested by the temperature dependence of the modulus at the rubber region. NPAM-silica/PU possessed a higher storage modulus at the rubber plateau than that of N-phenylaminomethyl POSS/PU when the loading was below 26 wt%. The homogeneity of NPAM-silica/PU was better than that of N-phenylaminomethyl POSS/PU, which was characterized by TEM and small angle X-ray scattering (SAXS). The NPAM-silica showed greater advantages in practical applications compared with N-phenylaminomethyl POSS.

---

Introduction

Polymer nanocomposites have been intensively studied in the past decades because nanofillers offer the promise of polymer composites with enhanced properties.^1,2 The key points for the enhancement are the way the nanofiller is integrated into the polymer, and the dispersion degree of the nanofiller. The favorable effects are evident when nanofillers are integrated into a polymer via a chemical reaction because chemical bonding can enhance the compatibility between the filler and the polymer matrix. An adequate amount and type of filler such as clay,³ carbon black^4,5 and silica^6–8 have been added into polymer matrices to tune the properties. However, due to the incompatibility between the filler and polymer, phase separation may be induced and the amount of the additives is limited,⁹ hindering the application of the composites in coatings. Much work has been done to improve the compatibility of polymers and nanofillers.^10–14 Organic silanes are often used as precursors to chemically modify inorganic fillers. However, excessive silane during the modifying process often necessitates extensive cleaning, and the particle surfaces can be only partially modified which results in the inhomogeneity of the functional fillers. Besides, aggregation would occur in the polymer matrix during the solvent-evaporating process. These will result in the poor degree of dispersion of the filler in the polymer matrix. Therefore, it is meaningful to find novel economic additives that can not only improve the compatibility between the filler and polymer but also enhance the properties of the composite.

Silsesquioxanes are an important class of organic–inorganic hybrids containing organic functional groups inherently and uniformly. Recently, there have been considerable literature reports relating to polyhedral oligomeric silsesquioxanes (POSS), with which the mechanical and thermal properties can be enhanced.^15–18 However, because of their cost, their practical use is limited. A much easier and economical route to synthesize silsesquioxanes with appropriate properties is still needed. Though POSS may enhance the mechanical and thermal properties of POSS/polymer nanocomposites, it is hard to disperse POSS in polymers at the molecular level due to the immiscibility and crystallization of POSS.¹⁹ In physical blending systems such as PLA/POSS, the diameters of POSS ranged from 100 nm to 200 nm.²⁰ In random POSS copolymers, the POSS moieties had been found to aggregate into nano-crystalline domains of 10–100 nm.²¹ For multifunctional POSS/polymer composites, when the amount of POSS was above a certain value, it would also aggregate²² or even form nanocrystals^23,24 due to POSS–POSS interactions. The crystallization will sacrifice the dispersibility, and the inner functional groups of POSS will not take part in the reaction.

One of the possible ways to overcome the problems may be the application of a mixture of reactive silsesquioxanes instead of mono-dispersed POSS. Mori and coworkers^25,26 reported a series of silsesquioxane-based silica nanoparticles with very high yields, which were derived from the hydrolytic condensation of organic silane. Bliznyuk²⁷ used the same method to synthesize the silsesquioxane-based nanoparticles and found that the 2–4 nm nanoparticles were dispersed in a polyurethane matrix, providing a new idea to prepare similar POSS mixtures. Silsesquioxanes were also prepared by environmentally friendly methods.^28,29 Lee³⁰ used a simple emulsion process to prepare an attractive hybrid silica with high yield. It could be used in polymer nanocomposites at lower cost. Though hybrid silica has been used in several fields such as ceramics, magnets,³¹ semiconductors³² and chromatographic adsorbents,³³ hybrid silica applied in polymer nanocomposites is still less-reported.³⁴

PU is one of the most commercial elastomers that can be used as adhesives, flexible films, coatings or hard plastics.³⁵ Despite its high wear resistance and good toughness, the low modulus and poor heat resistance of PU limit the scope of its application.⁹ Effective methods to improve the mechanical and thermal properties of PU are to introduce fillers into the PU matrix³⁶ and increase the crosslinking density of PU.³⁷

In our previous work, we reported a simple route to synthesize N-phenylaminomethyl POSS, which was then used as a cross-linker to incorporate conventional PU.³⁸ In this work, we report the synthesis of new N-phenylaminomethyl hybrid silica by a facile route, and its structure was confirmed by NMR, MALDI TOF-MS and GPC. The as-prepared hybrid silica was then introduced into polyurethane. The morphology of the NPAM-silica/PU nanocomposites was investigated by TEM and AFM. The thermal–mechanical properties of the nanocomposites were investigated by DMA and tensile tests. Besides, the thermal–mechanical properties and morphology of NPAM-silica/PU were compared with those of N-phenylaminomethyl POSS/PU. The purpose of this paper is to provide a new alternative polymer nanocomposite for practical applications.

Experimental section

Materials

N-Phenylaminomethyltriethoxysilane (ND-42) and phenyltriethoxysilane were acquired from Liyang Mingtian Chemical Corporation, China. Ammonium hydroxide and sodium dodecylbenzene sulfonate (SDBS) were supplied by Jiangsu Yonghua Fine Chemistry Co. Ltd, China. All chemicals were used as received without further purification. Water was obtained from a Sartorius arium 611DI water purification system. Stannous octoate was obtained from Shanghai Guoyao Corporation, China. Polytetramethylene glycol (PTMG, M_n = 1000), isophorone diisocyanate (IPDI) and 1,4-butanediol (BD) were supplied by Sigma-Aldrich (USA) and dried under vacuum before use. N,N-Dimethylacetamide (DMAC, Shanghai Guoyao Corporation, China) was dried by BaO and vacuum transferred.

Synthesis of NPAM-silica and phenyl-silica

The synthetic route of NPAM-silica is shown in Scheme 1. 5.40 g of ND-42 and 0.013 g of SDBS were added into 30 mL of water under vigorous stirring until an emulsion formed. 0.5 mL of NH₃·H₂O was added dropwise to the emulsion (pH = 11.5), and the reaction was kept at 50 °C for 48 h. A homogeneous emulsion was obtained. The concentration of SDBS used in the preparation of the sample was 0.2 critical micelle concentrations (CMC).

Scheme 1 Synthetic route of NPAM-silica by the hydrolytic condensation of ND-42.

The synthetic route of phenyl-silica was the same as that of NPAM-silica. 5.40 g of phenyltriethoxysilane and 0.013 g of SDBS were added into 30 mL of water under vigorous stirring until an emulsion formed. 0.5 mL of NH₃·H₂O was added dropwise to the emulsion (pH = 11.5), and the reaction was kept at 50 °C for 48 h. A homogeneous emulsion was obtained.

Synthesis of NPAM-silica/PU nanocomposites

The nanocomposites were synthesized according to the literature.¹⁸ The details are described below: The reference prepolymer was synthesized by dissolving polytetramethylene ether glycol (PTMG) (3.00 g, 0.003 mol) and isophorone diisocyanate (IPDI) (1.33 g, 0.006 mol) in a 100 mL three-necked flask equipped with a mechanical stirrer under a nitrogen atmosphere. Stannous octoate (0.1 wt%) was added as the catalyst and the reaction was stirred at 85 °C for 4 h to obtain the prepolymer which was a viscous liquid. The prepolymer was then dissolved in 10 mL of DMAC. Various amounts of NPAM-silica were dissolved in DMAC (10 mL), and the solutions were added into the prepolymer solution and stirred for 10 minutes at 95 °C. BD dissolved in 10 mL of DMAC was then added to the mixture. The amount of BD was accordingly adjusted in order to keep NCO/(OH + NH) = 1 for PU11. The amount of NH of PU18 was equivalent to NCO, so BD was not added. BD was not added for PU26 and PU40 either, because NH/NCO exceeded 1. After stirring for 5 min, the mixture was poured into a PTFE plate. After the majority of the solvent was evaporated under vacuum at 60 °C, the plates were put in an oven and kept at 95 °C for 12 h to complete the reaction. Finally the NPAM-silica/PU films were obtained (Scheme 2). PU11 + 7 (UnR) was synthesized to compare the temperature dependence of the modulus with PU11 and PU18. 11 wt% NPAM-silica and 7 wt% phenyl-silica were added to the PU prepolymer and stirred for 10 minutes at 95 °C. BD dissolved in 10 mL of DMAC was then added to the mixture. The amount of BD was also accordingly adjusted in order to keep NCO/(OH + NH) = 1. With stirring for 5 min, the mixture was poured into a PTFE plate. The post-process method was the same as for PU11. The compositions of the neat PU and NPAM-silica/PU nanocomposites are listed in Table 1.

Scheme 2 Synthesis of the NPAM-silica/PU nanocomposites.

Table 1 Compositions of the neat PU and NPAM-silica/PU composites

Sample	NPAM-silica	Ph-silica	IPDI	PTMG	BD
	wt%	wt%	wt%	wt%	wt%
Neat PU	0	0	28.9	65.2	5.9
PU11	11.4	0	26.6	60.0	2.2
PU11 + 7(UnR)	11.4	7	26.6	60.0	2.2
PU18	18.0	0	25.2	56.8	0
PU26	26.0	0	22.8	51.2	0
PU40	40.0	0	18.5	41.5	0

---

Characterization

Solid state ¹³C NMR and ²⁹Si NMR spectra were recorded on a Bruker Avance 400 NMR spectrometer at 399.7 MHz using a 7 mm T3 double resonance CPMAS probe and a zirconia PENCIL rotor. All of the experiments were carried out at room temperature. Solution ¹³C NMR and ²⁹Si NMR spectra were recorded on a Bruker DRX NMR spectrometer at 500 MHz, and tetramethylsilane (TMS) was used as an internal standard.

Wide-angle X-ray diffraction experiments (XRD) with NPAM-silica was carried out on a Bruker D8 Advance with Cu KR radiation (1.541 Å), scanning from (2θ) 2° to 50° with a step size of 0.02 and a time per step of 4 s.

Gel permeation chromatography (GPC) was performed at room temperature on a Waters 515 instrument equipped with a Wyatt Technology Optilab Rex refractive index detector. The columns were Styragel HR3, HR4 and HR5 (300 × 7.8 mm) from Waters. Tetrahydrofuran (THF, HPLC grade, Aldrich) was used as the solvent at a flow rate of 1 mL min⁻¹. Sample solutions in THF were filtered over a filter with a pore size of 0.45 μm (Nylon, Millex-HN 13 mm Syringes Filters, Millipore, USA). The columns were calibrated by using polystyrene standards with molecular weights in the range of 900 and 1.74 × 10⁶ g mol⁻¹ (with NMD for 1.02–1.11).

Matrix Assisted Laser Desorption Ionization Time of Flight Mass Spectrometry (MALDI-TOF MS) was performed using a Bruker Daltonics Autoflex TOF/TOF in the linear and reflectron modes. The 2,5-dihydroxybenzoic acid (DHBA) matrix was dissolved in THF (10 mg mL⁻¹), and mixed with the sample solution (1 mg ml⁻¹ in THF) in a 1 : 1 v/v ratio. Samples were spotted onto the target and dried in air.

The Fourier transform infrared spectra (FTIR) were recorded on a Nicolet Nexus 870 spectrophotometer. The attenuated total reflection (ATR) accessories were used to measure the FTIR spectra of all the specimens on fresh surfaces. In all cases, 64 scans at a resolution of 2 cm⁻¹ were used to record the spectra.

The morphology of the silica/PU composites was measured by HRTEM. The materials were frozen to −100 °C, and ultrathin sections were made using a Leica Ultracut UC6 with an approximate thickness of about 70 nm. The sections were collected onto copper grids. The grids were then imaged using a JEM-2100 transmission electron microscope at 200 kV. Surface topographies were also characterized by atomic force microscopy (AFM) in the tapping mode using silicon tips with an Olympus OMCL-AC160TS and controller from Veeco Corporation. The samples were dropped onto the silicon wafer surfaces with DMAC solutions of polymer films (1 wt%) and cured in a vacuum oven at 95 °C for 12 h.

The tensile tests were carried out on an Instron Series IX Automated Materials Testing System (Instron 4466, UK). The mechanical properties were measured according to the ASTM D1708 standard method. The Young's modulus was obtained by calculating the slope of the initial part (<5% strain). Tests were performed in triplicate to give the mean values. The dynamic mechanical tests were carried out on a Dynamic Mechanical Analyzer (DMA) (METRAVIB DMA+450, France). The viscoelastic properties were measured under a nitrogen atmosphere, at a heating rate of 2 K min⁻¹ from −120 °C to 200 °C and a frequency of 10 Hz. The experiments were carried out until the samples became too soft to be tested.

SAXS was performed on an Anton Paar SAXSess instrument at room temperature. The experiment was carried out by simultaneously recording the data on an imaging plate (IP) which was extended to the high-angle range (the q range covered by the IP was from 0.1 to 29 nm⁻¹, q = 4π sin θ/λ, where the wavelength λ of the Cu-Kα radiation was 0.1542 nm and 2θ is the scattering angle) at 40 kV and 40 mA for 5 min.

Results and discussions

Synthesis of novel NPAM-silica

The synthetic route was facile and the emulsion was stabilized by ammonia. The NPAM-silica is in the form of monodisperse nanospheres, as characterized by TEM and SEM (Fig. 1). The DLS results also confirmed the monodispersity of NPAM-silica (Fig. S1, ESI†). The size of the nanospheres could be controlled between 40 nm and 200 nm by adding different amounts of surfactants with 1.0 CMC to 0.2 CMC. Here, we chose 0.2 CMC of the surfactant, and the size of the nanospheres was about 200 nm.

Fig. 1 TEM (left, scale bar 400 nm) and SEM (right, scale bar 2.0 μm) images of NPAM-silica.

After the water and surfactant was removed, the product turned to a white powder with a yield of 97%. Table 2 clearly shows the differences between the syntheses of hybrid silica and POSS. The solvent used for the synthesis of NPAM-silica is water , while a large amount of organic solvent and hydrochloric acid is needed for the synthesis of POSS. Two steps are needed to synthesize N-phenylaminomethyl POSS: hydrolytic condensation and then neutralization by Et₃N, while only one step is needed for the synthesis of NPAM-silica. The yield of NPAM-silica is about 97%, which is much higher than that of POSS (64%). From Table 2, it can be concluded that the route to synthesize NPAM-silica is economical and environmentally friendly.

Table 2 Comparisons between the synthesis of NPAM-silica and POSS

Sample	Solvent	Steps	Yield
NPAM-silica	Water	One step	97%
POSS	Methanol, HCl, DMSO, Et3N	Two steps: hydrolytic condensation and neutralization	64%

---

There were five major peaks in the solid state ¹³C NMR spectrum corresponding to the chemical shifts of –CH₂– (30 ppm) and the phenyl ring (110 ppm–150 ppm) respectively. The solid state ²⁹Si NMR spectrum showed a strong peak at about −70.9 ppm, meaning that the product was mainly the T³ structure. It also indicated that the hydrolytic condensation of ND-42 in the emulsion process was sufficient. Other silanes, with the same synthetic route, such as vinyltriethoxysilane, usually form T² and T³ structures.³⁸ Therefore, we can abbreviate the molecular formula of the product as (PhNHCH₂SiO_3/2)_n (Fig. 2).

Fig. 2 Solid state NMR and solution NMR of NPAM-silica. Solid state ¹³C NMR (top left); solid state ²⁹Si NMR (top right); solution ¹³C NMR (bottom left); solution ²⁹Si NMR (bottom right).

XRD experiments were performed in order to further characterize the structure of NPAM-silica. As shown in Fig. 3, there were two major peaks. One broad amorphous halo at about 20° was ascribed to different monomers, known as T₈, T₁₀, T₁₂, T₉(OH), etc. (which could be found in MALDI-TOF MS), in NPAM-silica being haphazardly packed at the molecular level.^39,40 The other sharp peak at about 5.5° corresponded to the intramolecular distance between NPAM-silica,^41,42 which was in accordance with some other multifunctional polysilsesquioxanes.^39,43,44 This indicated that the molecules in the system were aggregated to form ordered structures to some extent.⁴¹

Fig. 3 XRD of the NPAM-silica powder.

Solution of NPAM-silica

NPAM-silica was sphere-like, of size 200 nm, when the NPAM-silica emulsion was spin-coated on the silicon wafer (Fig. 4, left). Interestingly, the NPAM-silica could be readily dispersed in THF, acetone, DMF and DMAC, and was presented as a transparent dispersion. After the powder was dispersed in organic solvents (1 mg L⁻¹) and spin-coated on the silicon wafer, the NPAM-silica turned into nanoparticles of size 50 nm (Fig. 4, right). The dispersion was concentrated and NPAM-silica was then aggregated on the silicon wafer during the spin-coating process. The good solubility in organic solvents was ascribed to the multiple organic functional groups of NPAM-silica, which was similar with that of N-phenylaminomethyl POSS.^18,41

Fig. 4 AFM of NPAM-silica samples spin-coated from the emulsion (left) and spin-coated from the DMAC dispersion (right).

The results of the solution ¹³C NMR showed that the major chemical shifts of NPAM-silica did not change compared with those in the solid state. Meanwhile, the result of the solution ²⁹Si NMR showed that the product in solution was mainly the T³ structure (Fig. 2). According to the results of the solution NMR, it was concluded that the chemical structure of NPAM-silica did not change during the solution process.

The molecular weight of the NPAM-silica in THF was 1327 (M_w), as characterized by GPC. The polydispersity (M_w/M_n = 1.08) remained low, and the result was consistent with other hydrolytic condensed silsesquioxane systems.⁴¹

MALDI-TOF MS was used to analyze the molecular mass of the NPAM-silica. It showed that there were strong peaks at 1266, 1604 and 1921 which corresponded to the fully condensed silsesquioxanes [T₈ + H]⁺, [T₁₀ + Na]⁺ and [T₁₂ + Na]⁺, respectively. However, there were also some peaks at 1458 [T₉(OH) + Na]⁺ and 1771 [T₁₁(OH) + H]⁺ which corresponded to the incompletely condensed silsesquioxanes. Fig. 5A–E show the mass spectra of NPAM-silica in different molecular mass ranges. The possible structures are also inserted in the figures. They indicated that the product was a mixture of fully condensed and a small quantity of incompletely condensed silsesquioxanes. Meanwhile, from the results of the solution ²⁹Si NMR (Fig. 2, bottom right), the signal related to silanol structures did not appear. This was ascribed to the content of silanol groups in the product being lower than the detection limit.²⁶ The product was directly employed for further use, because the exclusion of time-consuming processes offered the possibility of large-scale production and practical applications. Briefly, although there were different monomers in NPAM-silica, it could be more easily used in practice due to its facile synthesis in contrast to N-phenylaminomethyl POSS.

Fig. 5 MALDI-TOF MS of NPAM-silica in the range of 1000–2500 m/z (top). MALDI-TOF MS of NPAM-silica in the range of 1255–1295 m/z (A), 1400–1480 m/z (B), 1575–1625 m/z (C), 1745–1800 m/z (D) and 1890–1930 m/z (E). The corresponding structures are inserted in each figure from A to E; R represents the phenylaminomethyl group.

Reactivity of NPAM-silica with PU

Solid state ¹³C NMR was performed to test the extent of reaction of the NPAM-silica when introduced into PU. The major change of the chemical shifts was in the phenyl ring. For unreacted phenylaminomethyl groups, there were four main chemical shifts corresponding to four types of carbon atom in the phenyl ring. When the phenylaminomethyl group was reacted with isocyanate, the circumstance of the carbon atoms would change. The ortho-position (a) and para-position (b) corresponding to the peaks at 110–120 ppm in NPAM-silica (Fig. 6 black line) would move to 130 ppm (a′ and b′) in the NPAM-silica/PU composites (Fig. 6 red line). The shift of the carbon atom at the meta-position would remain constant after reaction. Therefore, the degree of reactivity could be monitored by the residual peak at 110 ppm. From the figure, it is suggested that almost all of the NPAM-silica had reacted with isocyanate when the loading reached 18 wt%, because the peaks at 110–120 ppm had almost disappeared.

Fig. 6 Solid state ¹³C NMR of NPAM-silica and the NPAM-silica/PU composites (left). Representative chemical structure of the phenyl rings of NPAM-silica and the composites (right).

Morphology of the NPAM-silica/PU composites

Fig. 7 exhibits a typical image of the NPAM-silica/PU composite (PU40), which appeared transparent. HRTEM was carried out to analyze the microstructure of the NPAM-silica/PU composites. PU18 was chosen to reveal the distribution of nanoparticles in the composites. Fig. 8A showed alternately dark and bright features, which was attributed to the phase regions of the hard and soft segments of the POSS/PU nanocomposites. Interestingly, among these dark domains, it could be seen that considerable amounts of much darker spherical particles with the size of around 3 nm were uniformly dispersed in the NPAM-silica/PU networks (Fig. 8B). The dark nanoparticles could be attributed to NPAM-silica because of the high electron density of the silicon atoms. It revealed that the silica was nano-dispersed in the hard phase region. The good compatibility could be attributed to the following two reasons: (1) there were physical interactions between multiple functional groups of NPAM-silica and PU,^45,46 and (2) almost all of the active groups of NPAM-silica were reacted with diisocyanate, and NPAM-silica was then introduced into the hard phase. This was consistent with the solid state ¹³C NMR result as mentioned above (Fig. 6).

Fig. 7 Typical image of the NPAM-silica/PU composite (PU40).

Fig. 8 HRTEM images of the NPAM-silica/PU composite. (A) Scale bar 100 nm; (B) scale bar 20 nm. Much darker particles are labeled by red circles.

AFM also suggested the good compatibility of the NPAM-silica in PU. In Fig. 9, PU18 presented a featureless surface in the height image, and the roughness average (R_a) of the film was 0.229 nm which indicated that the film was flat. PU18 showed alternately dark and bright features in the phase contrast image. It was revealed that the PU18 film had a two-phase heterogeneous structure with bright domains of about 10 nm in diameter, which could be attributed to the microphase separation of the hard and soft segments in the PU. Meanwhile, a number of particles with much brighter contrast could be found in the phase images. It was reasonable to suggest that these particles were nano-dispersed NPAM-silica particles, considering the AFM tip effect.

Fig. 9 AFM (height image and phase image) of PU18 with different scales (1 μm and 3 μm).

Mechanical properties of the NPAM-silica/PU composites

DMA. DMA tests were performed to analyze the dynamic mechanical properties of the composites. The enhanced effect of the reactive NPAM-silica can be seen in Fig. 10. When the NPAM-silica concentration was less than 18 wt%, the molar ratio of [NCO]/[NH] < 1, and the storage modulus of the composites increased with the increase of NPAM-silica concentration over almost all of the test temperature range. Further description is provided for two distinct regions. (1) In the rubber plateau region, the length and the height of the rubber plateau of the composites increased regularly, which meant that the cross-linking densities of the nanocomposites increased as the concentration of NPAM-silica increased. Because the molecular weight between the cross-linking junctions decreased, the composites became more and more rigid. (2) In the wake of the rubber plateau region, the storage modulus of PU11 began to decrease at about 100 °C. When the NPAM-silica concentration reached 18 wt%, the rubber plateau could be preserved up to 150 °C. This indicated that the addition of NPAM-silica could greatly enhance the service temperature of polyurethane.

Fig. 10 DMA of neat PU and the NPAM-silica/PU composites.

When the silica loading reached 18 wt%, the amount of NPAM-silica was exactly satisfied by [NCO]/[NH] = 1. From the ¹³C NMR results of the NPAM-silica nanocomposites, all of the functional groups of NPAM-silica had reacted. On one hand, the modulus increased with the addition of fillers. On the other hand, the increase of the amount of NPAM-silica would result in the decrease of the effective extent of reaction per unit of NPAM-silica, and thus lead to the decrease of the crosslinking density. As a result, the storage modulus at the rubber plateau region did not change too much with slightly excessive NPAM-silica (PU26) compared with PU18. This could be attributed to the combined effects of the crosslinking density contribution and the nanofiller addition contribution. When the NPAM-silica concentration increased further, such as for PU40, the molar ratio of [NCO]/[NH] exceeded 1 by too much, and the crosslinking density of the composites would decrease. Therefore, the storage modulus of PU40 at the rubber plateau would decline compared with PU18.

The temperature dependence of tan δ was also plotted, as shown in Fig. 10. In order to obtain the value, loss moduli vs. temperature was also plotted (Fig. S2, ESI†), and T_g of both the soft and hard segments are summarized in Table S1.† On one hand, the T_g of the soft segments did not change much when the NPAM-silica loading was below 26 wt%, while the T_g of the soft segments of PU40 increased by about 10 K compared with the other nanocomposites. On the other hand, the T_g of the hard segments increased gradually as the NPAM-silica loading increased. Because NPAM-silica reacted with the diisocyanate, it is natural to believe that NPAM-silica entered into the hard segment phase. Interestingly, when the silica loading was too much in excess ([NH]/[NCO] > 1), some of the NPAM-silica would enter into the soft segment phase and hinder the motion of the soft segment, which resulted in the increase of T_g of the soft segments of PU40. The amounts of the total hard segments (which were composed of IPDI, NPAM-silica and BD) increased as the NPAM-silica loading increased, hence the T_g of the hard segments increased.⁴⁷

Tensile test at constant T_test − T_g. Tensile tests were performed to investigate the enhanced mechanical properties of the composites. Considering the states of the different samples, the test temperature of the different samples was performed at a constant T_test − T_g (here, T_g refers to that of the hard segment). The constant T_test − T_g was chosen as 35 K, because all of the composites were in the rubber plateau under this situation. The stress–strain curves are expressed in Fig. 11A.

Fig. 11 (A) Tensile tests at constant T_test − T_g; (B) Young's modulus (E) versus NPAM-silica loading; (C) Young's modulus/T_test (E/T) versus NPAM-silica loading; (D) storage modulus versus T_g + T according to the DMA tests.

According to the typical rubber equation, the Young's modulus at low strains follows:

where E represents the Young's modulus, and M_c is the average molecular weight between the crosslinking junctions. R is the gas constant, ρ means the density, T is the temperature, and M_n means number average molecular weight of polymer. The value of the Young's modulus could be obtained from the slope of the initial part of the tensile curve. The Young's modulus versus the NPAM-silica loading is plotted in Fig. 11B. The Young's modulus of the NPAM-silica/PU nanocomposites increased with the increase of the NPAM-silica loading. Based on the DMA results, the excess NPAM-silica did not further enhance the Young's modulus of PU26 compared with that of PU18 in the temperature range of the measurements.

Nevertheless, from the equation, we found that the Young's modulus was affected by the test temperature. Because the samples were measured at different temperatures, the Young's modulus would be different. In order to avoid the effect of temperature, E/T was also plotted versus the NPAM-silica loading (Fig. 11C). This showed almost the same trend as the Young's modulus. Therefore, under the same T_test − T_g conditions, the enhanced effect of the introduction of NPAM-silica was obvious without regard to T_g.

We also plotted the storage modulus versus T_g + T in order to check the results of the tensile tests at constant T_test − T_g (Fig. 11D). The modulus was enhanced with the increase of NPAM-silica loading up to 18 wt%. When the content of NPAM-silica reached 26 wt%, the modulus of PU26 decreased a little, which showed the same trend as E/T. Because the number of active groups of NPAM-silica was in excess, this would lead to the decrease of the crosslinking density.

Entropic effect and nanofiller effect. In polymer nanocomposites, the reinforcement of the modulus often comes from two aspects. One is the modulus increase simply caused by the addition of fillers. The other is the entropic effect of the building network if the filler acts as a cross-linker in composites. However, these two contributions are seldom investigated separately.

The entropic effect of the network is described by the well-known rubber equation, where the proportionality factor, (ρR/M_c), between the modulus and temperature corresponds to the crosslinking density (eqn (1)). The temperature dependence of the Young's modulus in the plateau region of the composite systems was then observed to investigate the entropic effect in the nanocomposites. All of the samples were tested at their rubber regions (323 K, 333 K, 343 K and 353 K). As shown in Fig. 12, all of the samples showed a slight increase in the Young's modulus when the temperature increased. The Young's modulus increased with temperature because of the increased thermal or Brownian motion, which caused the stretched molecular segments to tug at their “anchor points” and try to assume a more probable coiled-up shape.⁴⁸ Moreover, the cross-linking density of PU18 was higher than that of PU11 since it showed a steeper slope, indicating that the entropic effect was a contribution of the enhancement.

Fig. 12 Temperature dependence of the Young's modulus of the NPAM-silica/PU composites at the rubber plateau.

In order to separate the entropic contribution from the effect of the addition of the filler, we also tested the composite of PU11 + 7(UnR). The sample PU11 + 7(UnR) was composed of 11 wt% NPAM-silica and 7 wt% phenyl-hybrid silica (a similar hybrid silica without reactive groups), where the “UnR” part had a similar chemical structure to NPAM-silica but could not act as a cross-linker. Our observations showed that the trends of the modulus change of PU11 and PU11 + 7(UnR) were almost the same, indicating that the crosslinking densities of these two samples were nearly the same, but the Young's modulus increased as the filler addition increased. So the nanofiller addition was the other contribution of the enhancement effect. This indicated that the reinforcement effect was derived from both the entropic contribution and the nanofiller contribution.

Besides, from the dashed line of PU18, it can be seen that the trend of the modulus change was almost the same with PU11 and PU11 + 7(UnR) at low test temperatures, while at high test temperatures, the trend increased. This suggested that the entropic contribution was obvious at high test temperatures while nanofiller contribution was more obvious at low test temperatures.

Comparison of the NPAM-silica/PU composites with POSS/PU composites

In our previous work, we introduced N-phenylaminomethyl POSS into the PU network and obtained enhanced POSS/PU nanocomposites.¹⁸ It is worth comparing the DMA results of the NPAM-silica/PU nanocomposites with those of the POSS/PU nanocomposites (Fig. 13). For the additive loadings of 11 wt% and 18 wt%, the storage moduli were almost the same in the glassy state. Meanwhile, in the rubber state, a great difference was found. There was a rubber plateau for the NPAM-silica/PU nanocomposites, but not for the POSS/PU nanocomposites. The storage modulus of NPAM-silica/PU 18 wt% at 100 °C was 6.9 MPa, which was 3 times larger than that of POSS/PU 18 wt% (1.7 MPa). Besides, a lower loading was needed for the NPAM-silica/PU composites to reach the rubber plateau than for the POSS/PU composites. This indicated the great advantage of NPAM-silica, reducing the amount of raw material needed to achieve the same reinforced properties.

Fig. 13 Comparison of the dynamic mechanical properties of POSS/PU and NPAM-silica/PU composites (loadings of 11 wt% (left) and 18 wt% (right)).

The results were intuitively confirmed by HRTEM. Fig. 14 shows high resolution images of the crosslinking junctions in two typical nanocomposites. In POSS/PU, POSS aggregated orderly to form nanocrystals of size ca. 5–10 nm, while in NPAM-silica/PU, the NPAM-silica was presented as amorphous nanoparticles about 3 nm in size, and no lattice fringes appeared. This revealed different crosslinking junctions in the two nanocomposites. In the POSS/PU composites, POSS tended to form nanocrystals in the PU network due to the interactions between POSS, and then formed an ordered aggregate. This para-crystallization structure formation has previously been reported in other POSS/polymer systems.^27,49 The crystallization would hinder the ability of the inner functional groups to fully react with isocyanate. As for NPAM-silica, because it had a smaller bulk and did not form crystals easily, many more active groups would participate in the reaction. Therefore, when the loading was high, the NPAM-silica could be well dispersed and incorporated in PU uniformly, while POSS tended to form the inhomogeneous network to some degree.

Fig. 14 HRTEM images showing a comparison of the POSS (left) and NPAM-silica particles (right) in the POSS/PU and NPAM-silica/PU composites.

SAXS curves for the POSS/PU and NPAM-silica/PU composites are displayed in Fig. 15. They show the dependence of the logarithm of scattered X-ray intensity on the wave vector q. The curves reveal that there were microphase separations in the composites. The nano-sized domains were enriched with the inorganic phase (POSS or NPAM-silica). The peak position (q_max) is indicative of the mean inter-domain spacing d, by d = 2π/q_max. It can be seen that the scattering curve of NPAM-silica/PU was characterized with a well pronounced maximum corresponding to the wave vector 1.70 nm⁻¹, which is a signature of a spatial periodicity of 3.7 nm. On the other hand, the peak of POSS/PU had a much higher intensity and shifted towards a lower wave vector value of 0.75 nm⁻¹, corresponding to a periodicity of 8.3 nm. This indicates that POSS formed nanocrystal-enriched domains in the composites which had a much higher scattering density. The mean inter-domain spacing of the POSS/PU composites was longer than that of the NPAM-silica/PU composites.

Fig. 15 Desmeared small angle X-ray diffractograms of the POSS/PU and NPAM-silica/PU nanocomposites.

Finally, three points were suggested to describe the differences of the POSS/PU and NPAM-silica/PU nanocomposites. (1) The size of the crosslinking junctions of POSS/PU was larger than that of NPAM-silica/PU. For POSS/PU, it was about 5–10 nm, and about 3 nm for NPAM-silica/PU. The spatial distance of POSS or NPAM-silica-enriched domains was different. For POSS/PU, it was 8.3 nm, and 3.7 nm for NPAM-silica/PU. (2) The states of the crosslinking junctions were different. POSS tended to form nanocrystal domains in the network, while the NPAM-silica was well dispersed in the network in the form of nanoparticles. (3) The effective numbers of the reactive groups were different. Because of the formation of crystals in POSS, it would hinder the inner functional groups from participating in the reaction, and the effective amount of reactive groups in the reaction in POSS/PU was less than that in NPAM-silica/PU. These factors would result in the different crosslinking densities of the composites; hence the moduli at the rubber plateau were different.

Conclusions

In summary, we provided an economical route to synthesize NPAM-silica. The NPAM-silica could be easily dispersed in organic solvents. The NPAM-silica in solution was mainly of the T³ structure, and the weight average molecular weight was 1327 with low polydispersity. The MALDI-TOF MS and NMR results suggested that NPAM-silica was mainly composed of fully condensed silsesquioxanes and less partially condensed silsesquioxanes.

NPAM-silica was then chemically incorporated into polyurethane, which acted as a chemical cross-linker. It showed excellent compatibility with PU and was nano-dispersed in hard domains. The mechanical properties were greatly enhanced with the increase of the NPAM-silica amount. The storage modulus and glass transition temperature increased regularly with the increase of the amount of NPAM-silica when the loading of NPAM-silica was below 26 wt%. The Young's modulus of the NPAM-silica/PU composites also increased with the increase of the NPAM-silica amount without regard to the T_g. The enhancement in the modulus came from both the nanofiller addition and the entropic contribution, which was tested by the temperature dependence of the Young's modulus in the rubber region.

Comparing POSS/PU with NPAM-silica/PU, it can be seen that the storage modulus of the NPAM-silica/PU was higher than that of POSS/PU at the rubber plateau. The size of the crosslinking junctions in POSS/PU was larger than those in NPAM-silica/PU. The spatial distance between the POSS-enriched domains was also larger than those between NPAM-silica-enriched domains. POSS formed crystals of size 5–10 nm in the composites, while NPAM-silica formed nanoparticles of size 3 nm. This resulted in the higher degree of reactivity and higher crosslink densities for NPAM-silica/PU. Finally, the NPAM-silica provided a new alternative for polymer nanocomposites in practical applications, with great advantages compared with POSS.

Acknowledgements

The authors gratefully acknowledge support from the Natural Science Foundation of Jiangsu Province (Grant SBK 201340307). The authors would like to thank the Program for Changjiang Scholars and Innovative Research Team in University for financially supporting this research.

Notes and references

1. L. Schadler, Nat. Mater., 2007, 6, 257–258.
2. M. R. Bockstaller, R. A. Mickiewicz and E. L. Thomas, Adv. Mater., 2005, 17, 1331–1349.
3. V. Mittal, Polym. Nanocompos., 2009, 143–152.
4. M. Qu, F. Deng, S. M. Kalkhoran, A. Gouldstone, A. Robisson and K. J. Van Vliet, Soft Matter, 2011, 7, 1066–1077.
5. Y. Konishi and M. Cakmak, Polymer, 2005, 46, 4811–4826.
6. W. Zhang, R. S. Blackburn and A. Dehghani-Sanij, Scr. Mater., 2007, 57, 949–952.
7. D. W. Janes, J. F. Moll, S. E. Harton and C. J. Durning, Macromolecules, 2011, 44, 4920–4927.
8. J. S. Meth, S. G. Zane, C. Z. Chi, J. D. Londono, B. A. Wood, P. Cotts, M. Keating, W. Guise and S. Weigand, Macromolecules, 2011, 44, 8301–8313.
9. C.-J. Chen, M.-H. Tsai, I. H. Tseng, A.-W. Hsu, T.-C. Liu and S.-L. Huang, RSC Adv., 2013, 3, 9729.
10. S. H. Stelzig, M. Klapper and K. Muellen, Adv. Mater., 2008, 20, 929–932.
11. N. Bitinis, M. Hernandez, R. Verdejo, J. M. Kenny and M. A. Lopez-Manchado, Adv. Mater., 2011, 23, 5229–5236.
12. V. Mittal, Polym. Nanocompos., 2009, 1–14.
13. D. L. Green and J. Mewis, Langmuir, 2006, 22, 9546–9553.
14. C. K. Wu, K. L. Hultman, S. O'Brien and J. T. Koberstein, J. Am. Chem. Soc., 2008, 130, 3516–3520.
15. K. Tanaka and Y. Chujo, J. Mater. Chem., 2012, 22, 1733–1746.
16. Y.-J. Lee, S.-W. Kuo, C.-F. Huang and F.-C. Chang, Polymer, 2006, 47, 4378–4386.
17. Y. D. Zhang, S. H. Lee, M. Yoonessi, K. W. Liang and C. U. Pittman, Polymer, 2006, 47, 2984–2996.
18. Q. H. Zhang, H. He, K. Xi, X. Huang, X. H. Yu and X. D. Jia, Macromolecules, 2011, 44, 550–557.
19. W. Zhou, S. Yuan, L. Tan, Y. Chen and Y. Huang, Polym. Eng. Sci., 2012, 52, 2063–2070.
20. H. Pan and Z. Qiu, Macromolecules, 2010, 43, 1499–1506.
21. O. Monticelli, A. Fina, A. Ullah and P. Waghmare, Macromolecules, 2009, 42, 6614–6623.
22. G. M. Kim, H. Qin, X. Fang, F. C. Sun and P. T. Mather, J. Polym. Sci., Part B: Polym. Phys., 2003, 41, 3299–3313.
23. Y. Zhang, S. H. Lee, M. Yoonessi, H. Toghiani and C. U. Pittman Jr, J. Inorg. Organomet. Polym. Mater., 2007, 17, 159–171.
24. P. Chen, X. Huang, Q. Zhang, K. Xi and X. Jia, Polymer, 2013, 54, 1091–1097.
25. H. Mori, A. H. E. Muller and J. E. Klee, J. Am. Chem. Soc., 2003, 125, 3712–3713.
26. H. Mori, M. G. Lanzendorfer, A. H. E. Muller and J. E. Klee, Macromolecules, 2004, 37, 5228–5238.
27. V. N. Bliznyuk, T. A. Tereshchenko, M. A. Gumenna, Y. P. Gomza, A. V. Shevchuk, N. S. Klimenko and V. V. Shevchenko, Polymer, 2008, 49, 2298–2305.
28. T. Matsumoto, Y. Takayama, N. Wada, H. Onoda, K. Kojima, H. Yamada and H. Wakabayashi, J. Mater. Chem., 2003, 13, 1764–1770.
29. F. Dong, W. Guo, S.-S. Park and C.-S. Ha, J. Mater. Chem., 2011, 21, 10744–10749.
30. D. Lee, Y. A. Kim, Y. B. Kim, J. K. Kim and Y. K. Han, Macromol. Res., 2008, 16, 353–359.
31. Z. H. Zhou, J. M. Xue, H. S. O. Chan and J. Wang, Mater. Chem. Phys., 2002, 75, 181–185.
32. W. Zhang, R. S. Blackburn and A. A. Dehghani-Sanij, Scr. Mater., 2007, 56, 581–584.
33. C. Perruchot, M. M. Chehimi, M. Delamar, S. F. Lascelles and S. P. Armes, J. Colloid Interface Sci., 1997, 193, 190–199.
34. G. Y. Bae, J. Jang, Y. G. Jeong, W. S. Lyoo and B. G. Min, J. Colloid Interface Sci., 2010, 344, 584–587.
35. D. J. Liaw, C. C. Huang and B. Y. Liaw, Polymer, 1998, 39, 3529–3535.
36. H. S. Xia and M. Song, Soft Matter, 2005, 1, 386–394.
37. N. Dolmaire, F. Mechin, E. Espuche and J. P. Pascault, J. Polym. Sci., Part B: Polym. Phys., 2006, 44, 48–61.
38. Z. Meng, C. Y. Xue, Q. H. Zhang, X. H. Yu, K. Xi and X. D. Jia, Langmuir, 2009, 25, 7879–7883.
39. P. S. G. Krishnan and C. B. He, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 2483–2494.
40. J. Zhang, R. W. Xu and D. S. Yu, Eur. Polym. J., 2007, 43, 743–752.
41. H. Mori, C. Sada, T. Konno and K. Yonetake, Polymer, 2011, 52, 5452–5463.
42. R. M. Laine and M. F. Roll, Macromolecules, 2011, 44, 1073–1109.
43. J. Zhang, R. Xu and D. Yu, Eur. Polym. J., 2007, 43, 743–752.
44. W. Li, F. Liu, L. Wei and T. Zhao, J. Appl. Polym. Sci., 2007, 104, 3903–3908.
45. N. Hao, M. Boehning and A. Schoenhals, Macromolecules, 2007, 40, 9672–9679.
46. N. Hao, M. Bohning, H. Goering and A. Schonhals, Macromolecules, 2007, 40, 2955–2964.
47. A. K. Nanda, D. A. Wicks, S. A. Madbouly and J. U. Otaigbe, Macromolecules, 2006, 39, 7037–7043.
48. D. W. Van Krevelen and K. Te Nijenhuis, Properties of polymers: their correlation with chemical structure; their numerical estimation and prediction from additive group contributions, 2009.
49. B. Alvarado-Tenorio, A. Romo-Uribe and P. T. Mather, Macromolecules, 2011, 44, 5682–5692.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra01409g

---