Gradient semi-interpenetrating polymer networks based on polyurethane and poly(2-hydroxyethyl methacrylate) for biomedical applications

Abstract

Gradient semi-interpenetrating polymer networks (gradient semi-IPNs) as well as the traditional semi-interpenetrating polymer networks (semi-IPNs) were synthesized using polyurethane (PU) and poly(2-hydroxyethyl methacrylate) (PHEMA). The materials were characterized with respect to thermodynamic miscibility, NIR imaging, mechanical properties and morphological structure by tapping mode atomic force microscopy (TM AFM). The positive values of Gibbs free energy indicated that polymeric systems were thermodynamically immiscible. The dynamic mechanical analysis as well as TM AFM demonstrated that the systems under investigation were two-phase systems with incomplete phase separation. The gradient semi-IPNs were shown to have unique mechanical properties dependent on the composition and on the degree of microphase separation. The ability to create a layer of biocompatible polymer, such as PHEMA, at the surface, or create nanostructured surface consisting of nanodomains of different polymeric compositions, and engineer the improvements in the mechanical properties of the materials through the use of gradient systems should allow the creation of novel materials for biomedical application through the optimisation of mechanical properties, surface chemistry and biological properties.

---

Introduction

There is significant interest in the development of novel materials with unique properties, including polymeric composite materials which contain gradients of physical or chemical properties.^1–3 These materials include gradient interpenetrating polymer networks (IPNs), whose preparation and characterization have been reported in a number of recent articles.^4–12 Typical functional gradient materials are characterized by a continuous variation in composition, structure or properties from one surface to another, leading to a range of multifunctional properties. Gradient interpenetrating polymer networks can be considered as a series of layers of immiscible polymer mixtures, whose compositions and properties vary gradually from the center of the samples, or from one surface to the other.^4,5,9 These materials are of specific interest for noise- and vibration-dampening,^9,11 for optoelectronics¹ and for biomedical engineering applications.^4,13 The modification of biomaterials using surface treatments has recently become an active area of surface engineering. A number of research groups have focused on the preparation of surfaces with a gradual change in chemical composition in one dimension.¹³ Such a “gradient surface” is of particular interest for basic and applied studies of the interactions between biological species and surfaces. The formation of the gradient layers of a biologically more compatible polymer in the matrix of another one may allow the creation of a more biologically compatible surface and may lead to an increase in the biological compatibility of a material as a whole.^4,8 Such an approach to the creation of biomaterials could be attractive as it provides the ability to change both the chemical properties of the surface and nanodomain topography of a surface both of which can contribute to the biological compatibility of a material whilst allowing the interpenetrating polymer to be used to provide the appropriate mechanical properties for any particular application.¹⁴

Polyurethane (PU) materials have excellent and controllable mechanical properties and are extensively used in blood contacting applications and in organ reconstruction. However the wider use of PUs in other medical applications is limited by the biological compatibility of the materials.¹⁵ They are also known to be prone to biodegradation,¹⁶ stress induced degradation¹⁷ and to surface cracking.¹⁸ Recently, considerable effort has been directed at improving the biocompatibility of polymers used for the fabrication of medical implants.^19,20 This effort includes the preparation of the materials based on traditional interpenetrating polymer networks for different biomedical applications.^21–27 One of the most powerful approaches that can be used to improve biological compatibility, mechanical properties and resistance to degradation of polymers is the creation of the gradient interpenetrating polymer networks with gradient layers of biocompatible polymer.^3–6,13

This paper describes the comparative investigation of a series of traditional interpenetrating polymer networks based on polyurethane and biocompatible polymer poly(2-hydroxyethyl methacrylate) (PHEMA) and a range of gradient semi-IPNs of similar polymeric composition which was undertaken to better understand the potential advantages afforded by the gradient systems.

Results and discussion

Investigation of gradient semi-IPNs by NIR imaging

Near infrared (NIR) imaging was used to verify the synthesis of gradient composites. The samples of gradient composites were scanned along the gradient of concentration of PHEMA. 600 NIR spectra were collected across a transverse section of each material. Fig. 1a shows the full absorbance of a gradient sample containing an average concentration of 20.38% PHEMA, and Fig. 1b shows three NIR spectra showing the composition in the surface (1), the composition at the distance of 1 mm from the surface (2) and in the middle of the sample (3).

Fig. 1 (a) Full absorbance of a sample of gradient semi-IPN containing an average concentration of 20.38% PHEMA and (b) three NIR spectra of gradient semi-IPN measured at the surface (1), 1 mm from the surface (2) and in the middle of the sample (3).

The spectra indicate that the PHEMA is mainly concentrated at the surface of the gradient block, and its concentration decreases with distance from the surface of the sample. Fig. 1b shows the differences in the NIR spectra across the sample. The main differences in the NIR spectra were observed in the range of band 4350–4450 cm⁻¹ (Fig. 1b). For this reason the band ratio at 4350–4450 cm⁻¹ was used to analyse the concentration of PHEMA in the matrix of PU (Fig. 2a).

Fig. 2 Main part of NIR spectra of gradient semi-IPN used for PHEMA concentration analyses (a) and image of gradient composite in false colours (b).

Thermodynamic state of gradient semi-IPNs

The creation of the gradient semi-IPNs starts from the non-equilibrium swelling of the polyurethane network with monomer (2-hydroxyethyl methacrylate) (HEMA) which creates the concentration profile through the thickness of the sample. During the photopolymerization of the monomer HEMA in the matrix of polyurethane the conditions for transition from the one-phase state to the two-phase state vary throughout the sample. For kinetic reasons, during the polymerization, the “freezing” of the non-equilibrium state occurs rather than separation of the polymer components into two phases. Non-equilibrium states of the system at different depths from the surface differ from one to another as the initial conditions for the phase separation will vary.^28,29

The degree of deviation of the system from the equilibrium state in the gradient semi-IPN layers at different distances from the surface will also vary. It was interesting to investigate the changes in the thermodynamic miscibility of polymer components spaced at different depth from a surface of a sample of gradient semi-IPN. For this purpose the sorption of low-molecular weight compound vapours by samples of individual polymers and by the layers of gradient blocks was investigated and the calculations of thermodynamic parameters were carried out.

In Fig. 3 the isotherms of dichloromethane vapour sorption at 298 K by polyurethane (1), poly(2-hydroxyethyl methacrylate) (2) and by layers of gradient block of semi-IPN (3–6) are presented. The isotherm for polyurethane (1) has the shape of a typical isotherm of sorption by elastomers. For poly(2-hydroxyethyl methacrylate) (2) the sorption of dichloromethane is very small in the range of the relative pressure from 0 up to 0.6, and then increases. Such a behavior is typical for polymers in glassy state.²³ For layers of gradient semi-IPN we observed the isotherms of sorption similar to that observed for polyurethane (Fig. 3, curves 3–6). The isotherms for two surface layers of the gradient block (Fig. 3, curves 5, 6) are located lower relative to the two internal layers (Fig. 3, curves 3, 4). This suggests, as might be expected, that the surface layers of the gradient block contain more poly(2-hydroxyethyl methacrylate) than the internal ones.

Fig. 3 The isotherms of dichloromethane vapours sorption at 298 K by samples of polyurethane (1), poly(2-hydroxyethyl methacrylate) (2), and layers of gradient semi-IPN containing 20.38% PHEMA (3–6). The indexing of layers varies from the interior (3) up to the surface (6).

The calculation of the thermodynamic properties of the materials using the experimental data from the dichloromethane vapour sorption by the layers of the gradient semi-IPN allows a quantitative comparison of the different materials. Using the method described by A. A. Tager,³⁰ the changes in free energy of mixing between the polyurethane and poly(2-hydroxyethyl methacrylate) in the layers of the gradient semi-IPN located at different distances from a surface and containing a different amount of PHEMA were calculated.

The results of calculations of the free energy of mixing Δg_x of polyurethane and poly(2-hydroxyethyl methacrylate) in the layers of gradient semi-IPN are presented in Table 1.

Table 1 Variation of the free energy of mixing Δg_x of polyurethane and poly(2-hydroxyethyl methacrylate) in the layers of gradient semi-IPNs with average amount of 20.38% PHEMA

Layers of gradient semi-IPN	Amount of PHEMA in the layers (%)	Δ gx (J g^−1)
First layer (surface layer)	53.5 ± 0.1	+3.70 ± 0.11
Second layer	20.6 ± 0.1	+1.25 ± 0.04
Third layer	8.1 ± 0.1	+0.69 ± 0.02
Fourth layer (core of sample)	0.2 ± 0.1	+0.06 ± 0.01

---

Table 1 shows that the free energy of mixing of PU and PHEMA is positive. Therefore, the polymer components of semi-IPN are immiscible. However, there is the difference in value of a free energy of mixing for layers at different depth in the gradient block, the value of a free energy of mixing decreases with the distance from the surface to the center of the gradient block.

As shown previously,³¹ the value of a free energy of mixing depends on the conditions of polymer mixture preparation. First of all these conditions influence the value of entropy of mixing ΔS, which is part of the value of free energy of mixing ΔG:

ΔG = (ΔH − TΔS)

The free energy of mixing (the Gibbs energy) for a two-phase polymer–polymer system, which contains the interphase layer, was theoretically calculated previously.²³ This demonstrated that the free energy of mixing depends on the thickness of the interphase layer and the specific two-phase system formed. In the case of a gradient semi-IPN, mixtures with different structures are formed at different distances from a surface. They differ in the thickness of the interphase layer, the sizes of the phase domains, the diffusion of the phase boundaries and the degree of phase segregation.^3,4,32 So, the energy of interphase boundary formation for such mixtures should be theoretically different and consequently, they should differ by the values of free energy of mixing. By calculating the free energy of mixing for our gradient samples (Table 1) we have shown that not only theoretically but practically these values are different for different layers of the samples. This reflects the formation of different semi-IPN structures at different distances from a surface.

Dynamic mechanical analysis of gradient semi-IPNs

The viscoelastic properties of the gradient semi-IPNs, the native PU and PHEMA polymers and traditional semi-IPNs of the same chemical composition were compared experimentally.

Fig. 4 shows the variations of tanδ versus temperature for PU, PHEMA and for six traditional semi-IPNs. Using a temperature range from 120 K to 473 K, allowed examination of the α-relaxation temperature domains of the two polymer components. It was noted that the amplitude of the α-relaxation of PU (243 K at 10 Hz) decreases in the semi-IPNs with an increase in PHEMA content but the temperature range of this relaxation is about the same as that for native PU. Similar behavior of a polyurethane network has been observed for other immiscible IPNs.^3–8,14

Fig. 4 Dynamic mechanical measurements of Tan Delta with temperature for polyurethane, PHEMA and for semi-IPNs containing different concentrations of PHEMA.

The amplitude of the α-relaxation of PHEMA (406 K at 10 Hz) decreases with the increasing PU content in the semi-IPNs and the maximum of this relaxation shifts to lower temperatures (Fig. 4). This supports the conclusion that the investigated semi-IPNs are two-phase systems but their phase separation is not complete as the position of the α-relaxation peak of PHEMA in the semi-IPNs is not the same as in the native polymer. The dynamic mechanical behaviour of the series of traditional semi-IPNs is described in detail in a previous paper.³²

The dynamic mechanical properties of gradient semi-IPN with overall concentration of PHEMA 20.38% and gradient semi-IPN with overall concentration of PHEMA 11.23% were investigated layer-by-layer along the gradient of concentration. Fig. 5 and 6 show the variation of tanδ with temperature for the different layers within the gradient semi-IPN with overall concentration of PHEMA 20.38% and 11.23%.

Fig. 5 Dynamic mechanical measurements of tanδ with temperature for layers of gradient semi-IPN with overall concentration of PHEMA 20.38% showing parallel sections gradient of concentration of PHEMA. For each of the following samples a surface layer of 0.6 mm thickness was removed: ■ - polyurethane matrix; ▲ - 1 layer removed; □ - 2 layers removed; [hexagon filled, flat-side down] - 3 layers removed from the PHEMA 20.38% gradient semi-IPN.

Fig. 6 Dynamic mechanical measurements of tanδ with temperature for layers of gradient semi-IPN with overall concentration of PHEMA 11.23% showing parallel sections gradient of concentration of PHEMA. For each of following samples a surface layer of 0.6 mm thickness was removed: ■ - polyurethane matrix; ▲ - 1 layer removed; □ - 2 layers removed; [hexagon filled, flat-side down] - 3 layers removed from the PHEMA 11.23% gradient semi-IPN.

Analysis of the curves provides further confirmation of the PHEMA concentration gradient with depth from a surface of the samples. Thus, for the sample of gradient semi-IPN with overall concentration of PHEMA 20.38%, which represents a complete set of all layers (Fig. 5, curve 1), the intensive wide maxima of PHEMA overlapping the interval of temperatures 298–448 K, as well as a small maxima of polyurethane are observed in its glass transition domains. On removing the surface layers from the sample which contains the highest concentration of PHEMA (Fig. 5, curve 2), the intensity of the tanδ peak for PHEMA is reduced and the intensity of the tanδ peak of polyurethane is increased. In the third sample, representing a core of gradient block of semi-IPN with overall concentration of PHEMA 20.38% (Fig. 5, curve 3), we observed an intense tanδ peak for polyurethane and only a shoulder in the region of the glass transition of PHEMA. This confirms the data obtained using NIR imaging (Fig. 1 and 2), showing that the PHEMA is present at the core of the gradient block in small amounts and increases in concentration towards the surface of the material.

We observed a similar pattern for the samples of gradient semi-IPN with overall concentration of PHEMA 11.23% (Fig. 6); with the tanδ peak of PHEMA in a sample, which represents the complete set of all layers (Fig. 6, curve 1), is much lower reflecting the smaller average amount of PHEMA in this gradient semi-IPN. Thus already for the second sample (Fig. 6, curve 2) we observe only a shoulder for the PHEMA in the domain of its glass transition. The third sample, the core of gradient block of semi-IPN with overall concentration of PHEMA 11.23%, consists of only the polyurethane (Fig. 6, curve 3). Thus, we could conclude, that in the gradient semi-IPN with overall concentration of PHEMA 11.23%, this PHEMA is only found in the outermost layers of the material.

It is important to emphasise, that the presence of wide intensive maxima for PHEMA on the tanδ versus temperature curves of the layers of gradient semi-IPN does not provide evidence of the miscibility of the polymer components; it is the result of a superposition of a major number of relaxation maxima in the layers with gradually varying composition, that we also observed for the gradient IPNs of other chemical compositions.^3–8,28,29

In Fig. 7 the storage modulus versus temperature data for polyurethane (1), for PHEMA (2) and for series of traditional semi-IPNs based on these two polymers are presented. The curves of storage modulus reflect the trends of the loss factor data. The monotonous increase of the storage moduli with increasing amount of PHEMA in the systems and mainly monotonous decrease of storage moduli with the temperature could be observed for these samples. The curves of storage moduli are typical of two-phase systems with incomplete phase separation as we do not observe two sharp decreases in storage moduli with temperature or a plateau region between the two glass temperature transitions of two components. This reflects the reinforcement of polyurethane by the high T_g PHEMA polymer in these semi-IPNs.

Fig. 7 Log storage modulus versus temperature data for polyurethane, for PHEMA and for a series of traditional semi-IPNs with different amounts of PHEMA.

In Fig. 8 and Fig. 9 the storage modulus for layers of gradient semi-IPN with overall concentration of PHEMA 20.38% and of gradient semi-IPN with overall concentration of PHEMA 11.23% accordingly are presented. The differences in modules of surface layers and internal layers of the gradient systems are most clearly displayed above the glass transition temperature of polyurethane (ca. 298 K). For example, at a temperature of 298 K, where the polyurethane is in the elastic state and PHEMA is in a glassy state, the modulus of the first and third layers differ by 500 MPa for gradient semi-IPN with overall concentration of PHEMA 20.38% (Fig. 8) and differ by 260 MPa for the gradient semi-IPN with overall concentration of PHEMA 11.23% (Fig. 9).

Fig. 8 Dynamic mechanical measurements of Log storage modulus with temperature for layers of gradient semi-IPN with overall concentration of PHEMA 20.38% showing parallel sections gradient of concentration of PHEMA. For each of the following samples the surface layer with a thickness of 0.6 mm was removed: ■ - polyurethane matrix; [hexagon filled, flat-side down] - 1 layer removed; ▲ - 2 layers removed; △ - 3 layers removed from the PHEMA 20.38% gradient semi-IPN.

Fig. 9 Dynamic mechanical measurements of Log storage modulus with temperature for layers of gradient semi-IPN with overall concentration of PHEMA 11.23% showing parallel sections gradient of concentration of PHEMA. For each of the following samples the surface layer with a thickness of 0.6 mm was removed: ■ - polyurethane matrix; [hexagon filled, flat-side down] - 1 layer removed; ▲ - 2 layers removed; △ - 3 layers removed from the PHEMA 11.23% gradient semi-IPN.

The comparison of the storage moduli of the series of semi-IPNs of definite compositions (Fig. 7) and the storage moduli of the layers of gradient systems (Fig. 8 and Fig. 9) allow to make a conclusion that PHEMA is mainly concentrated in the surface layers of the gradient systems. This is in accordance with the results of NIR imaging investigation of these samples (Fig. 1 and Fig. 2).

Hence, the investigation of the viscoelastic properties of these material has shown a modification of material properties with depth from the surface; the storage modulus of the surface layer of the gradient systems is 5–10 times higher than the storage modulus of the internal layer. The gradient semi-IPNs are characterized by the wide temperature range of glass transition of polymer, arising from the composition gradient. This wide overall temperature range for the glass transition of the material arises from a superposition of changing glass transition across the material due to the changing PHEMA content. Gradient semi-IPNs also demonstrate anisotropy of viscoelastic properties in different directions, which can have significant impacts particularly when the direction of deformation of the samples is parallel or perpendicular to the direction of gradient of concentration. This not only provides interesting possibilities for the creation of noise- and vibration-dampening materials using gradient semi-IPNs but offers potential for the development of novel biomedical materials.

Atomic force microscopy

Phase imaging is a powerful extension of tapping mode atomic force microscopy that provides nanometer scale information about surface structure often not revealed by other techniques. By mapping the phase of the cantilever oscillation during the tapping mode scan, phase imaging goes beyond simple topographical mapping to detect variation in composition, viscoelasticity, adhesion and other properties. This method allows the identification of different components in composite materials, and different regions of high and low surface adhesion or hardness at very high resolution. Phase images of the semi-IPNs based on crosslinked PU and PHEMA as well as of the gradient composites were detected by the tapping mode AFM. The results of the tapping mode AFM phase images for PU, PHEMA and semi-IPNs are presented in Fig. 10.

Fig. 10 Tapping mode AFM phase images for semi-IPNs based on PU and PHEMA: native PU (a); native PHEMA (b); 16.15% PHEMA (c); 31.89% PHEMA (d); 40.35% PHEMA (e); 57.70% PHEMA (f). The images represent fields of 500 nm × 500 nm.

Pure PU demonstrated the segregation of hard and soft segments within the overall surface structure (Fig. 10a); the native PHEMA displays no phase separation with the surface of the sample appearing homogeneous and flat (Fig. 10b). In the semi-IPN samples a distinct phase separation is observed at the nanometer scale (Fig. 10c–f). Phase domains of 15–25 nm could be observed in the semi-IPNs with PHEMA content 16–32% but with an increasing amount of PHEMA in the semi-IPNs up to 40%, the size of the domains increased to 25–50 nm (Fig. 10d–e). A bi-continuous phase structure could be observed in the semi-IPN with PHEMA content 57% and an inversion of phases could be seen for this sample (Fig. 10f).

The results of gradient composite investigation, as well as of native PU and PHEMA, by tapping mode atomic force microscopy are presented in Fig. 11 and Fig. 12.

Fig. 11 Tapping mode AFM, topography, scan size 500 nm × 500 nm, polyurethane (a); polyHEMA (b).

Fig. 12 Tapping mode AFM, topography, scan size 500 nm, surface of gradient composite with 20% PHEMA (a), 1.5 mm from the surface of gradient composite with 20% PHEMA (b); middle of gradient composites with 20% PHEMA (c).

Topography scans for PU and PHEMA are presented in Fig. 11, and topography scans of layers of gradient composite with a whole concentration 20.38% of PHEMA are presented in Fig. 12a–c. The gradient composites appear to be nanostructured materials with the surface of gradient samples enriched with PHEMA. The comparison of the structure of the layers of gradient composites (Fig. 12a–c) have shown that the surface of the first layer is rougher, the second layer is smoother and includes less PHEMA than first one, and third layer of gradient block looks more like a native PU. As the surface of first layer of gradient composite has a honeycomb nanostructure we might expect bacterial adhesion to such a material to be lower than to native polymers because of the lowering in surface free energy which may influence the bacterial adhesion.

Mechanical testing

The physico-mechanical properties of the gradient systems were examined by compression method as described below. The comparative tests for the pure PU, the semi-IPNs and the samples of gradient composites with the same amount of PHEMA were carried out. The results of testing are presented in Fig. 13 and Fig. 14.

Fig. 13 Mechanical properties of PU, semi-IPN with concentration of PHEMA 11.31% (SIPN) and of gradient composite with overall concentration of PHEMA 11.23% (GRAD) investigated by compression method.

Fig. 14 Mechanical properties of PU, semi-IPN with concentration of PHEMA 20.34% (SIPN) and of gradient composite with overall concentration of PHEMA 20.38% (GRAD) investigated by compression method.

We observed (Fig. 13) that for the same deformation value l/l₀ the compressive stress is lowest for the PU sample. The formation of a semi-IPN with 11.31% of PHEMA results in an increase of the compressive stress of the samples. From examination of the mechanical properties of the semi-IPN and gradient semi-IPN containing the same amount of PHEMA (11.23%), it is evident that the gradient semi-IPN has a slightly higher energy of fracture. The bulk moduli for gradient composites are also higher compared with the same parameter for the semi-IPN samples with the same amount of PHEMA (Table 2).

Table 2 Bulk moduli of PU, semi-IPNs and gradient composites investigated by compression method

Sample	Bulk modulus K, MPa
PU	131.0
SIPN with 11.31% PHEMA	149.2
Gradient IPN with 11.23% PHEMA	175.7
SIPN with 20.34% PHEMA	197.6
Gradient IPN with 20.38% PHEMA	243.0

---

As the amount of PHEMA in the samples increases, this effect becomes more evident (Fig. 14, Table 2); the gradient composite with 20.38% of PHEMA have higher physico-mechanical parameters compared with the semi-IPN containing the same amount of PHEMA and the gradient composite with 11.23% of PHEMA (Fig. 13). The bulk modulus of gradient composites (Table 2) is also higher than the bulk modulus of the semi-IPNs containing the same amount of PHEMA.

This data suggests that it is possible to use the synthesis of gradient systems to improve the mechanical characteristics of semi-IPNs and composites. A number of different hypotheses have been proposed to explain this phenomena.^33,34 The most likely model treats the gradient semi-IPNs as a sequence of infinite number of layers of semi-IPN, whose composition and elastic (or bulk) moduli vary gradually from the surface to the core of the samples. When the strain on a sample comprised of all layers is extended to the same degree, the stress in each layer must be directly related to its modulus. This distribution of stresses promotes a plastic deformation, instead of brittle failure, and resulting in an increase in the energy of fracture.

Experimental

Preparation of semi-IPNs and gradient semi-IPNs of polyurethane and poly(2-hydroxyethyl methacrylate)

The semi-IPNs and gradient semi-IPNs were prepared on the basis of polyurethane network and poly(2-hydroxyethyl methacrylate) using a sequential method. The polyurethane network was synthesized in two steps. The first step involved the preparation of the adduct of trimethylol propane (TMP) and toluylene diisocyanate (TDI), both supplied by Bayer. TDI was distilled on a vacuum line. TMP was dried at 35 °C under vacuum, and then was dissolved in ethyl acetate supplied by Aldrich. The TMP/TDI adduct was prepared by reacting 1 equivalent of TMP with 1.5 equivalents of TDI at 65 °C. The reaction was carried out until the theoretical isocyanate content was reached, which was determined by the di-n-butylamine titration method.

The second step involved the synthesis of the crosslinked PU. The films of PU network were obtained from the mixtures of poly(oxypropylene)glycol (PPG) with M_w = 2000 g mol⁻¹ and TMP/TDI adduct (ratio 1 : 2 g-eqv.). Before use, PPG was degassed under vacuum, for 8 h at 70 °C. The films were cured under nitrogen atmosphere for 48 h at 80 °C. Unreacted materials were extracted from the PU network films by Soxhlet method using ethyl acetate as a solvent.

The semi-IPNs and gradient semi-IPNs were obtained by the sequential method. The polyurethane network was swollen with a freshly distilled 2-hydroxyethyl methacrylate (HEMA), supplied by Sigma, containing Irgacure 619 as the initiator to the equilibrium state. The obtained second polymer is poly(2-hydroxyethyl methacrylate).

To produce the semi-IPNs, the swollen polyurethane network was put in the hermetic box to reach the equilibrium state of the monomer HEMA. Then the monomer was cured by photopolymerization. The photopolymerization of the monomer was carried out in the temperature-controlled chamber during 1 h. The wavelength of UV light was 365 nm.

The gradient semi-IPNs were formed using non-equilibrium swelling of the polyurethane network with monomer HEMA containing Irgacure 619 as the initiator. The swelling was terminated at a certain stage before equilibrium is established and then the polymerization of monomer was carried out to produce the gradient semi-IPNs.

The semi-IPNs and gradient semi-IPNs were held in vacuo at 10⁻² Pa at 373 K for 36 h to reach the constant weight. A range of semi-IPNs with compositions containing 0 to 57% of PHEMA and gradient semi-IPNs with an average PHEMA content of 20.38% and 11.23% were prepared using these methods.

Vapour sorption

The dichloromethane vapour sorption by semi-IPN samples and by gradient semi-IPN samples was studied using a vacuum installation and a McBain balance.^21,22 The changes in partial free energy of dichloromethane by sorption (dissolution) were determined from the experimental data using eqn (1)

Δμ₁ = (1/M)RT ln(P/P₀), (1)

where M is the molecular mass of dichloromethane and P/P₀ is the relative vapour pressure. The value Δμ₁ changes with solution concentration from 0 to −∞.

To calculate the free energy of mixing of the polymer components with solvent, the changes in partial free energy of the polymers (native polymers, semi-IPNs, gradient semi-IPNs) needs to be determined. This requires the calculation of the difference between the polymer chemical potential in the solution of a given concentration and in pure polymer under the same conditions (Δμ₂). Δμ₂ for the polymer components were calculated using the Gibbs–Duhem equation:

ω₁d(Δμ₁)/dω₁ + ω₂d(Δμ₂)/dω₁ = 0, (2)

where ω₁ and ω₂ are the weight fractions of a solvent and of a polymer. This can be rearranged to give eqn (3)

∫d(Δμ₂) = −∫(ω₁/ω₂)d(Δμ₁) (3)

Which allows the determination of Δμ₂ for each polymer from the experimental data by integration over definite limits. The average free energy of mixing of solvent with the individual components, semi-IPNs of various compositions for the solutions of different concentration, was then estimated using eqn (4) using computational analysis.

Δg^m = ω₁Δμ₁ + ω₂Δμ₂ (4)

Dynamic mechanical analysis

The dynamic mechanical analysis (DMA) measurements were carried out using a Dynamic Mechanical Thermal Analyzer Type DMA 2980 from TA Instruments over the temperature range from −140 to +200 °C and at fixed frequencies (5, 10, 15, 20, 30 Hz) with a heating rate of 3 °C min⁻¹. The experiments were performed in the tension mode on rectangular specimens (35 mm × 5 mm × 1 mm). As poly(2-hydroxyethyl methacrylate) is a hydroscopic polymer, all samples were dried at 80 °C for 48 h under vacuum before measurements. The samples were subsequently subjected to the following thermal cycle during DMA measurements: a first run from 20 °C up to 100 °C, then second run from −140 up to +200 °C. The second run was used for analysis of the results.

Near infrared (NIR) imaging

A Perkin Elmer 400N FT-NIR Imaging System was used to investigate the distribution of PHEMA in the gradient composite using a Linear Detector Array. 600 NIR spectra were obtained for the gradient composite by scanning from one surface to the other one across the samples. The characteristic absorbance band at 4350–4450 cm⁻¹ was used to analyse the concentration of PHEMA in the matrix of PU across the sample.

Atomic force microscopy

Surface topography and phase images were obtained using a multimode Nanoscope IIIa atomic force microscope. Tapping mode was employed in air using a 125 μm silicon tip. Topographic and phase images were obtained simultaneously.

Mechanical testing

Mechanical properties of the semi-IPNs and gradient composite samples were measured by the compression method using Computerized Universal Testing Machine (Hounsfield H25KS). The samples were cut into the blocks with a gauge length of 20 mm, widths of 10 mm, and sample thickness of 6 mm. The samples were held between the two platforms and the strain rate was fixed at 10 mm min⁻¹.

Conclusions

Gradient semi-interpenetrating polymer networks (gradient semi-IPNs) as well as the traditional semi-interpenetrating polymer networks (semi-IPNs) were synthesized using polyurethane (PU) and poly(2-hydroxyethyl methacrylate) (PHEMA). The materials were characterized with respect to thermodynamic miscibility, NIR imaging, mechanical properties and morphological structure by TM AFM. The positive values of Gibbs free energy indicated that the polymeric systems were thermodynamically immiscible.

Using NIR imaging the distribution of the constituent polymers in the gradient systems across the samples were investigated and it was shown that PHEMA concentration in the samples varied within a 1.5 mm zone from both surfaces. The content of PHEMA in the layers of gradient sample with overall concentration of 20.34%, varied from 53.5% in the surface layer to 0.2% in the core of the sample.

The investigation of the viscoelastic properties of gradient semi-IPNs has shown that there was anisotropy in the viscoelastic properties within the samples. In particular, the properties of layers of the gradient semi-IPN varied with distance from the surface of samples resulting in the elastic modulus of the surface layer of the gradient systems being 10 times higher than the elastic modulus of internal layer (core of the sample). The gradient semi-IPNs are characterized by the broad glass transition over a 100 °C range as result of superposition of a large number of relaxation maxima within the layers arising from the gradual variation in composition. This offers the possibility of creating novel noise- and vibration-dampening materials using both gradient IPNs and gradient semi-IPNs.

The investigation of the traditional and gradient semi-IPN by TM AFM have shown that semi-IPN samples have nanostructured surfaces consisting of PU and PHEMA nanodomains. The surface of gradient semi-IPNs are enriched by PHEMA nanodomains and this surface has a honeycomb morphology. Such results might suggest that the interactions with biological cells, such as bacteria, may be reduced on the surface of gradient samples relative to the surface of native polymers.

The comparison of the physico-mechanical properties of the traditional semi-IPNs and gradient semi-IPNs with the same content of PHEMA has shown that gradient systems have higher energy of fracture. The bulk modulus for the gradient semi-IPNs was also higher than for traditional semi-IPNs. These results suggest that appropriate designed gradient systems may be used to improve the mechanical properties of the materials.

The ability to create layers of polymers with enhanced biological compatibility, such as PHEMA, at the surface of a material or nanostructured surface consists of nanodomains of different materials alongside the ability to engineer the internal mechanical properties of these types of materials through the use of gradient systems should allow the fabrication of new families of novel range of materials for biomedical application by allowing the optimisation of mechanical properties, surface chemistry and biological properties.

Acknowledgements

The authors gratefully acknowledge the FP7-PEOPLE-IRSES-230790 COMPOSITUM Grant and the Royal Society Exchange Visit Grant (Dr L. V. Karabanova) for financial support of the collaboration between Biomedical Materials Research Group at the University of Brighton and the Institute of Macromolecular Chemistry of the National Academy of Science of the Ukraine. The authors thank Prof. H. Sautereau (UMR CNRS 5627, France) for assistance with the compression testing of the materials.

References

1. B. Ilshner, J. Phys. IV, 1993, 3, 763–772.
2. M. Kryszewski, Polym. Adv. Technol., 1997, 9, 244–259.
3. Y. S. Lipatov and L. V. Karabanova, in Advances in Interpenetrating polymer Networks, ed. D. Klempner and K. Frisch, Technomic Publishing, Lancaster, Basel, 1994, V.4, p.456.
4. L. Karabanova, S. Mikhalovsky, A. Lloyd, G. Boiteux, L. Sergeeva, T. Novikova, E. Lutsyk and S. Meikle, J. Mater. Chem., 2005, 15, 499–507.
5. L. V. Karabanova, T. I. Novikova, L. M. Sergeeva and E. D. Lutsyk, Russ. J. Appl. Chem., 2002, 75(5), 800–804.
6. Y. S. Lipatov and L. V. Karabanova, J. Mater. Sci., 1995, 30, 1095–1104.
7. L. V. Karabanova, P. Pissis, A. Kanapitsas and E. D. Lutsyk, J. Appl. Polym. Sci., 1998, 68, 161–171.
8. L. V. Karabanova, A. W. Lloyd, S. V. Mikhalovsky, G. Boiteux, T. I. Novikova, L. M. Sergeeva and E. D. Lutsyk, Ukrainien Polymer Journal, 2004, 26, 18–25.
9. N. Desilles, L. Lecamp, P. Lebaudy and C. Bunel, Polymer, 2003, 44(20), 6159–6167.
10. N. Desilles, L. Lecamp, P. Lebaudy, B. Youssef, Z. Lebaudy and C. Bunel, Polymer, 2006, 47, 193–199.
11. H. Nakanishi, N. Namikawa, T. Norisuye and Q. Tran-Cong-Miyata, Macromol. Symp., 2006, 242, 157–164.
12. C. L. Qin, W. M. Cai, J. Cai, D. Y. Tang, J. S. Zhang and M. Qin, Mater. Chem. Phys., 2004, 85(2–3), 402–409.
13. M. S. Kim, G. Khang and H. B. Lee, Gradient polymer surfaces for biomedical applications, Prog. Polym. Sci., 2008, 33, 138–164.
14. L. V. Karabanova, A. W. Lloyd, S. V. Mikhalovsky, M. Helias, G. J. Phillips, S. Rose, L. Mikhalovska, G. Boiteux, L. M. Sergeeva, E. D. Lutsyk and A. Svyatyna, J. Mater. Sci.: Mater. Med., 2006, 17(12), 1283–1296.
15. M. D. Lehah and S. L. Cooper, Polyurethane in Medicine and Surgery, CRC Press, Boca Raton, FL, 1986, p.158.
16. R. W. Paynter, H. Martz and R. Guidoin, Biomaterials, 1987, 94, 8.
17. R. S. Thoma, J. Biomater. Appl., 1987, 1, 449–486.
18. M. Szycher, D. Dempsey and V. L. Poirier, Trans. Am. Soc. Biomater., 1984, 24, 7–21.
19. A. W. Lloyd, R. G. A. Faragher and S. P. Denyer, Biomaterials, 2001, 22, 769–785.
20. A. B. Lowe, M. Vamvakaki, M. A. Wassall, L. Wong, N. C. Billingham, S. P. Armes and A. W. Lloyd, J. Biomed. Mater. Res., 2000, 52, 88–94.
21. Z. D. Gomurashvili, R. Katsarava, G. E. Katsarava, G. Chumburdze, N. Mumladze and D. Tugushi, U.S. Pat. 7,649,022 (2010), “Bioabsorable Elastomeric Polymer Networks, Cross-linkers and Methods of Use”.
22. R. S. Langer and A. Andreas, U.S. Pat. 6,720,402, (2004), “Shape Memory Polymers”.
23. J. Turner and Yu-L. Cheng, U.S. Pat. 6,331,578 (2001), “Process for Preparing Ingterpenetrating Polymer Networks of Controlled Morphology”.
24. V. R. Shastri, R. S. Langer, and P. J. Tarcha, U.S. Pat. 5, 837, 752 (1998), “Semi-interpenetrating Polymer Networks”.
25. M. E. Dillon, U.S. Pat. 4,832,009 (1989), “Semi-interpenetrating Polymer Network Backsheet Bandage.”.
26. M. E. Dillon, U.S. Pat. 5,980,923 (1999), “Semi-interpenetrating Polymer Network Scar Treatment Sheeting, process of Manufacture and Useful Articles Thereof.”.
27. L. H. Sperling, C. J. Murphy, and V. Mishra, U.S.Pat. 5,786,425 (1998), “Process for Producing Biocompatible Polyisoprene-Polyurethane Interpenetrating Polymer Network Compositions for Use in Medical Devices and the Like”.
28. Yu.S. Lipatov, L. V. Karabanova, L. A. Gorbach, E. D. Lutsyk and L. M. Sergeeva, Polym. Int., 1992, 28, 99–107.
29. L. V. Karabanova, Ukrainien Polymer Journal, 1998, 64(N3), 24–36.
30. A. A. Tager, Vysokomolek. Soed., 1977, 19, 1654–1667.
31. A. A. Tager, Phisicochemistry of Polymers, ed. G. M. Mednikova, Chemistry, Moscow, 1978, p.544.
32. L. V. Karabanova, G. Boiteux, O. Gain, G. Seytre, L. M. Sergeeva and E. D. Lutsyk, Polym. Int., 2004, 53, 2051–2058.
33. J. Jang and C. Lee, J. Mater. Sci., 1998, 33, 5445–5450.
34. G. C. Martin, E. Enssani and M. Shen, J. Appl. Polym. Sci., 1981, 26, 1465–1473.

---