Keywords

1 Introduction

Hydrophilic gels called hydrogels are cross-linked materials absorbing large quantities of water without dissolving. Softness, smartness, and the capacity to store water make hydrogels unique materials (Tanaka 1981; Shibayama and Tanaka 1993). The ability of hydrogels to absorb water arises from hydrophilic functional groups attached to the polymer backbone while their resistance to dissolution arises from cross-links between network chains. Water inside the hydrogel allows free diffusion of some solute molecules, while the polymer serves as a matrix to hold water together. Another aspect of hydrogels is that the gel is a single polymer molecule, that is, the network chains in the gel are connected to each other to form one big molecule on macroscopic scale. It is natural to expect that the conformational transitions of the elastically active network chains become visible on the macroscopic scale of hydrogel samples. The gel is a state that is neither completely liquid nor completely solid. These half liquid-like and half solid-like properties cause many interesting relaxation behaviours that are not found in either a pure solid or a pure liquid. From the point of view of their mechanical properties, the hydrogels are characterized by an elastic modulus which exhibits a pronounced plateau extending to times at least of the order of seconds, and by a viscous modulus which is considerably smaller than the elastic modulus in the plateau region (Almdal et al. 1993).

Hydrogels may exhibit drastic volume changes in response to specific external stimuli, such as the temperature, solvent quality, pH, electric field, etc. (Dusek and Patterson 1968; Tanaka 1978). Depending on the design of the hydrogel matrices, this volume change may occur continuously over a range of stimulus level, or, discontinuously at a critical stimulus level. The volume transition behaviours of hydrogels received considerable interest in the last three decades and large parts of the work have been collected in different reviews (Shibayama and Tanaka 1993; Khokhlov et al. 1993).

Polymeric hydrogel networks may be formed by various techniques, however the most common synthetic route is the free-radical cross-linking copolymerization of a hydrophilic non-ionic monomer such as acrylamide (AAm) with a small amount of a cross-linker, e.g., N, N’-methylenebis(acrylamide) (MBAAm). In order to increase their swelling capacity, an ionic comonomer is also included into the reaction mixture. Since the monomers for hydrogel preparation are usually solid at the usual polymerization temperature, it is necessary to carry out the polymerization reactions in an aqueous solution. Hydrogel structure and, thus, the hydrogel properties are closely related to the conditions under which the hydrogels are formed, i.e., the cross-linker concentration, the initial degree of dilution of the monomers and the chemistry of the units building the network structure. The understanding of the formation mechanism of hydrogels under various experimental conditions is of great interest to predict their physical properties.

2 Swelling and Elasticity of Hydrogels

The equilibrium swelling degree and the elastic modulus of hydrogels depend on the cross-link and charge densities of the polymer network as well as on the cross-linked polymer concentration after the gel preparation. Although the theories predict the swelling properties and the elastic behavior of hydrogels formed under various conditions, the agreement between theory (see Chap. 3 Sect. 1.2) and experiments is only qualitative. Figure 1 illustrates the characteristic features of poly(acrylamide) (PAAm) hydrogels prepared from AAm and MBAAm in aqueous solutions (Kizilay and Okay 2003a; Orakdogen and Okay 2006a). In Fig. 1a, the modulus of elasticity after the gel preparation, G o , is plotted against the cross-linker (MBAAm) content for three series of gels prepared at various initial monomer concentration C o . Hydrogels exhibit elastic moduli in the range of 0.01 to 10 kPa, which are much smaller than the calculated values from their cross-linker contents. The initial period of the G o versus MBAAm % plots can be used to estimate the lower limit of the cross-linker concentration required for the onset of gelation. The best-fit curves through the G o versus % MBAAm data intersect with the positive abscissa at 0.03, 0.19, and 0.55 mol % MBAAm for C o = 7, 5, and 3 %, respectively (Orakdogen and Okay 2006a). Thus, the larger the dilution degree of the reaction system, the higher is the threshold concentration of MBAAm for the formation of an infinite network. Figure 1b shows cross-linker concentration dependence of the cross-linking efficiency ε xl of MBAAm, that is the fraction of MBAAm forming effective cross-links. ε xl is less than 20% and, it further decreases below 1% as the initial monomer concentration is decreased. This is a consequence of the increase of probability of cyclization and multiple cross-linking reactions as the initial monomer concentration decreases (Funke et al. 1998).

Fig. 1
figure 1

The elastic modulus G o of PAAm hydrogels after preparation (a) and the cross linking efficiency ε xl (b) shown as a function of MBAAm concentration. Initial monomer concentration C o = 3 (filled circle), 5 (open circle), and 7 w/v % (filled triangle). Reprinted from Orakdogen and Okay (2006a) with kind permission of Springer Science + Business Media

Full size image

The polymer network concentration at the state of gel preparation (index o), represented by the cross-linked polymer volume fraction \( \varphi_2^0 \), also alters significantly the hydrogel structure and, in turn, alters the hydrogel properties. The effect of \( \varphi_2^0 \) on the hydrogel properties is illustrated in Fig. 2 for polyacrylic acid (PAAc) hydrogels prepared at various \( \varphi_2^0 \) (Yazici and Okay 2005). In Fig. 2a, the modulus of elasticity G o and the effective cross-link density νc of PAAc hydrogels are plotted against \( \varphi_2^0 \). Figure 2b shows \( \varphi_2^0 \) dependence of the swelling ratio of PAAc gels in terms of the volume swelling ratio Q v (volume of swollen gel in water / volume of dry gel). G o increases from 1.4 kPa to 50 kPa as \( \varphi_2^0 \) is increased. The inset to Fig. 2a shows that the modulus data can be described by a power law \( {G_0}\; \propto \;{\left( {\varphi_2^0} \right)^x} \)where x = 2.1 \( \pm \) 0.1. The exponent is much larger than the linear dependence (x = 1) predicted by the theory of rubber elasticity (Flory 1953; Treloar 1975), and indicates existence of non-idealities during the gel formation process. Increasing number of wasted MBAAm molecules in cycles on raising the dilution of the reaction solution explains this discrepancy (Naghash and Okay 1996). Indeed, νc is an increasing function of \( \varphi_2^0 \) (Fig. 2a), that is, the higher the initial monomer concentration, the larger the effective cross-link density of the hydrogels and the smaller their swelling capacity (Fig. 2b).

Fig. 2
figure 2

The modulus of elasticity G o , the effective crosslink density ν c (a) and the volume swelling ratio Q v of PAAc hydrogels in water (b) as function of the polymer network concentration \( \varphi_2^0 \); 1.2 mol % MBAAm. The inset to Fig. 2a shows a double logarithmic G o vs. \( \varphi_2^0 \) plot. Reprinted from Yazici and Okay (2005) with kind permission from Elsevier

Full size image

Increasing number of ionic groups in hydrogels is known to increase their swelling capacities. This is mainly due to the simultaneous increase of the number of counterions inside the gel, which produces an additional osmotic pressure that swells the gel (Flory 1953). The excess swelling over the swelling of the corresponding non-ionic hydrogels can be suppressed with increasing salt concentration in the external solution, which decreases the concentration difference of the counterions between the inside and outside the gel phase. Figure 3 illustrates the typical swelling behaviour of ionic PAAm hydrogels of various charge densities in water and in aqueous NaCl solutions (Durmaz and Okay 2000). The ionic comonomer used in the hydrogel preparation is sodium salt of 2-acrylamido-2-methylpropane sulfonic acid (AMPS Na). AMPS Na units dissociate completely over the entire pH range so that AMPS Na containing hydrogels exhibit pH-independent swelling. Increase of the AMPS Na content from 0 to 80 mol % results in a 27-fold increase in the hydrogel volume in water. In 1.0 M NaCl solution, the swelling ratio is almost independent on the ionic group content due to screening of charge interactions within the hydrogel.

Fig. 3
figure 3

Equilibrium volume swelling ratio Q v of ionic PAAm hydrogels shown as function of the NaCl concentration in the external solution;1.2 mol % MBAAm. AMPS Na mole fraction x i in the comonomer mixtures is indicated. Reprinted from Durmaz and Okay (2000) with kind permission from Elsevier

Full size image

Since ionic hydrogels are highly swollen in water, their swelling equilibrium is mainly determined by the mixing entropy of the counterions, which is balanced by the gel´s rubberlike elasticity. According to the theory of rubber elasticity of Gaussian chains (Flory 1953), the Gibbs free energy of elastic deformation \( \Delta {G_{el}} \) scales with the deformation ratio as

$$ \Delta {G_{el}}\; \approx \;{N_s}^{ - 1}\,{\alpha^2}, $$
(1)

where N s is the number of segments between two successive cross-links, and α the linear deformation ratio. α is related to the normalized gel volume V r by the equation \( \alpha = \left( V /V_o \right)^{1/3} = {V_r}^{1/3}\), where V is the gel volume at a given degree of swelling and V o is the gel volume in the reference state, i.e., at the state after preparation. On the other hand, the existence of fixed ions on the network chains results in an unequal distribution of mobile counterions between the inside and outside of the gel. The ionic contribution to the Gibbs free energy \( \Delta {G_{ion}} \) may be written as

$$ \Delta {G_{ion}}\; \approx f\ln \left( {{{\,f\,\varphi_2^0} \mathord{\left/{\vphantom {{f\,\varphi_2^0} {{\alpha^3}}}} \right.} {{\alpha^3}}}} \right), $$
(2)

where \( f\; \)is the effective charge density of the network (Flory 1953). Balancing the two opposite free energy contributions represented by \( \Delta {G_{el}} \) and \( \Delta {G_{ion}} \) by minimizing their sum with respect to α, one obtainsFootnote 1

$$ {V_{eq}}\; \approx \;{\left( {\,f\,{N_s}} \right)^{{3 \mathord{\left/{\vphantom {3 2}} \right.} 2}}}\,, $$
(3)

which indicates a scaling parameter of 3/2 between the equilibrium swollen normalized gel volume \( {V_{eq}} \) and the number of charges per network chain \( f\;{N_s} \).

Figure 4a shows the double-logarithmic plot of \( {V_{eq}} \) against f N s . Experimental data are for poly(N-isopropyl acrylamide) (PNIPAAm) hydrogels prepared in the presence of the ionic comonomer AMPS Na (Gundogan et al. 2002). The dotted curve in the Figure represents the prediction of (3), i.e., Flory–Huggins (FH) theory with a scaling parameter of 3/2. The solid curve is the best fitting curve to the experimental swelling data, which gives a scaling relation \( {V_{eq}}\; \approx \;{\left( {f\,{N_s}} \right)^{{3 \mathord{\left/{\vphantom {3 4}} \right.} 4}}}\, \). The scaling parameter 3/4 is much smaller than the predicted value of 3/2 of the FH theory. An exponent between 0.6 and 0.8 has been reported for both weak and strong polyelectrolyte hydrogels equilibrium-swollen in water (Durmaz and Okay 2000; Silberberg-Bouhnik et al. 1995; Bromberg et al. 1997; Melekaslan and Okay 2000).

Fig. 4
figure 4

The equilibrium swollen normalized gel volume \( {V_{eq}} \) of ionic PNIPAAm hydrogels as function of the number f N s of charges per network chain (see (3)) (a) Reduced modulus \( {G_r} \) of ionic PNIPAAm hydrogels as function of the normalized gel volume V r (b) The mole fractions x i of AMPS Na are indicated in the figure. Reprinted from Gundogan et al. (2002) with permission of American Chemical Society

Full size image

The discrepancy between theory and experiment is related to the non-Gaussian behaviour of fully swollen hydrogels in water. The theory (3) assumes that the polymer network is a collection of Gaussian chains, which can be extended to infinity. However, the network chains in the equilibrium swollen ionic hydrogels as given in Fig. 4a are three to nine times as elongated as in the dry state. At such high swelling ratios, deviation from the Gaussian statistics may appear due to the finite extensibility of the network chains. A further evidence for the non-Gaussian behaviour of the network chains in the swollen hydrogels comes from the elasticity data. In Fig. 4b, the dependence of the reduced modulus \( {G_r} \) of ionic PNIPAAm hydrogels is shown as a function of the normalized gel volume V r (Gundogan et al. 2002). The reduced modulus \( {G_r} \) is defined as the ratio of the elastic modulus of the gel at a given degree of swelling \( {Q_v}\; = \;1/{\varphi_2} \) to that one of the same gel after its preparation. \( {G_r} \) is given for a network of Gaussian chains by (Flory 1953)

$$ {G_r}\; = \;\;G\left( {{\varphi_2}} \right)/G\left( {\varphi_2^o} \right)\; = \;\;{V_r}^{ - {1 \mathord{\left/{\vphantom {1 3}} \right.} 3}}. $$
(4)

However, Fig. 4b shows that the dependence of the reduced modulus on the gel volume cannot be described by a single scaling exponent. For the gel volumes V r of less than 0.4, the reduced modulus \( {G_r} \) decreases sharply with increasing volume V r . The rapid decrease of \( {G_r} \) with increasing gel volume V r in the first regime is usually interpreted as the transition of the polymer from the glassy to the rubbery state by addition of solvent (Gundogan et al. 2002). In the range of the gel volume V r between 0.4 and 3.5, the slope of \( {G_r} \) versus V r plot is −0.32, close to the theoretical value of −1/3. Thus, PNIPAAm hydrogels in this regime behave as Gaussian. For gel volumes larger than 3.5, the reduced modulus \( {G_r} \) starts to increase with increasing gel volume with a slope of 0.22 which is an indication of the limited extensibility of the network chains and is connected with the high stretching of the network chains.

Swelling behaviour of hydrophobically modified hydrogels has also received considerable attention due both to fundamental and to technological interests (Hirotsu 1993). Such hydrogels generally exhibit a temperature sensitivity, which is associated with the temperature dependence of hydrogen bonding and hydrophobic interactions (Hirotsu 1993; Arndt et al. 2001). A phenomenon called reentrant swelling transition was also observed in hydrophobically modified hydrogels immersed in aqueous solutions of organic solvents or linear polymers (Katayama et al. 1984; Melekaslan and Okay 2001; Okay and Gundogan 2002). In such a transition, the gel first collapses and then reswells if a particular external parameter such as the organic solvent or linear polymer concentration is continuously varied. As a consequence, the organic solvent (or linear polymer) first flows from the gel to the solution phase but then reenters the gel phase at higher concentrations. Examples of such transitions are illustrated in Fig. 5 for PNIPAAm, poly(N, N-dimethylacrylamide) (PDMAAm), and poly(N-t-butyl-acrylamide-co-AAm) (TBA/AAm) hydrogels immersed in aqueous solutions of poly(ethylene glycol) of molecular weight 300 g mol−1 (PEG-300), acetone, and dimethylsulfoxide (DMSO), respectively (Melekaslan and Okay 2001; Orakdogen and Okay 2006b; Ozmen and Okay 2003). The competing attractive forces between the gel components are responsible for the reentrant transition behavior of hydrophobically modified hydrogels.

Fig. 5
figure 5

Variation of the volume ratio V sol /V w (equilibrium swollen gel volume in solution / equilibrium swollen gel volume in water) of PNIPAAm, PDMAAm, and TBA/AAm (60/40 by mole) hydrogels ( filled symbols) and PAAm hydrogels (open symbols) with the volume fraction Φ of PEG-300, acetone, and DMSO in the outer aqueous solution. (a) reproduced from Melekaslan and Okay (2001) with permission from Wiley-VCH Verlag GmbH & Co. KGaA; (b, c) reproduced from Orakdogen and Okay (2006b) and Ozmen and Okay (2003) with permissions from Elsevier

Full size image

3 Inhomogeneity of Hydrogels

Another non-ideal feature of hydrogels is the so-called spatial gel inhomogeneity (Shibayama 1998; Bastide and Candau 1996). In contrast to ideal gels with a homogeneous distribution of cross-links, hydrogels always exhibit an inhomogeneous cross-link density distribution, known as the spatial gel inhomogeneity. The spatial inhomogeneity is undesirable because it dramatically reduces the optical clarity and strength of hydrogels. Since the gel inhomogeneity is closely connected to the spatial concentration fluctuations, scattering methods such as light scattering, small angle X-ray scattering, and small angle neutron scattering have been employed to investigate the spatial inhomogeneities. The gel inhomogeneity can be manifested by comparing the scattering intensities from the gel and from a semi-dilute solution of the same polymer at the same concentration (Lindemann et al. 1997). The scattering intensity from gels is always larger than that from the polymer solution. The excess scattering over the scattering from polymer solution is related to the degree of the inhomogeneities in gels.

In general, the spatial inhomogeneity increases with the gel cross-link density due to the simultaneous increase of the extent of network imperfections producing regions more or less rich in cross-links. On the other hand, the inhomogeneity decreases with the ionization degree of gels due to the effects of the mobile counter ions, electrostatic repulsion and the Donnan potential (Kizilay and Okay 2003b). The degree of swelling of gels subjected to scattering measurements also affects the scattering intensities (Kizilay and Okay 2004; Gundogan et al. 2004; Orakdogen et al. 2005). The scattering intensity at low scattering vectors is enhanced as the swelling degree is increased. This behaviour was interpreted as the enhancement of the difference of polymer concentration between the more and the less cross-linked regions. The initial monomer concentration used in the gel preparation significantly affects the scattering intensities (Kizilay and Okay 2003a, 2004; Gundogan et al. 2004). An inflection point was observed in the excess scattering versus monomer concentration plot, at which the inhomogeneity attained a maximum value.

Figure 6 shows excess scattering intensity R ex, q plotted as a function of \( \varphi_2^0 \) for PAAm hydrogels with 1.5 mol % MBAAm cross-linker (Kizilay and Okay 2004). R ex, q significantly increases as the gel swells beyond its swelling degree after preparation. Moreover, PAAm gels at both states exhibit a maximum scattering intensity at a critical polymer network concentration. As the monomer concentration is increased, the effective density of cross-links also increases (Fig. 1), so that the spatial inhomogeneity becomes larger. Opposing this, increasing monomer concentration, i.e., decreasing the degree of swelling of the gels after preparation reduces progressively the concentration difference between densely and loosely cross-linked regions of gel, so that the apparent inhomogeneity decreases. The interplay of these two opposite effects determines the inhomogeneity in PAAm gels and results in the appearance of a maximum gel inhomogeneity at a critical monomer concentration.

Fig. 6
figure 6

Excess scattering intensities from PAAm hydrogels R ex, q measured at the scattering vector q = 1 × 10−2 nm −1 shown as function of \( \varphi_2^0 \). The filled and open symbols represent results of measurements after gel preparation and after equilibrium swelling in water, respectively; 1.5 mol % MBAAm. Reprinted from Kizilay and Okay (2004) with permission from Elsevier

Full size image

4 Hydrogels with Improved Properties

The design of hydrogels with a good mechanical performance is of crucial importance in many existing and potential application areas of soft materials. Several attempts, such as topological gels and double network gels, have been made in recent years to design hydrogels with even better mechanical performance (Tanaka et al. 2005). The nanoscale dispersion of layered silicates or clays in polymer networks is one of the techniques offering significant enhancements in the material properties of hydrogels. Haraguchi et al. prepared such nanocomposite hydrogels starting from AAm-based monomers together with Laponite as a physical cross-linker, replacing the traditional chemical cross-linkers (Haraguchi and Takehisa 2002). Laponite, a synthetic hectorite clay, when suspended in water, forms disc-like particles with a thickness of 1 nm, a diameter of about 25 nm, and a negative surface charge density stabilizing dispersions in water. Formation of a cross-linked polymer network using a small amount of Laponite indicates that these nanoparticles act as a multifunctional cross-linker with a large effective functionality (Okay and Oppermann 2007).

A fast response of hydrogels to the external stimuli is also a requirement in many application areas of these materials. However, the kinetics of hydrogel volume change involves absorbing or desorbing solvent by the polymer network, which is a diffusive process. This process is slow and even slower near the critical point of volume phase transition (Shibayama and Tanaka 1993). Increasing the response rate of hydrogels has been one of the challenging problems in the last 25 years (Arndt Schmidt et al. 2004). In order to increase their response rate, several techniques were proposed (see also Chap. 3 Sect. 3.2):

  • Submicrometer-sized gel particles: Since the rate of response is inversely proportional to the square of the size of the gel (Shibayama and Tanaka 1993), small hydrogel particles respond to the external stimuli more quickly than bulk gels (Oh et al. 1998).

  • Gels having dangling chains: Attachment of linear polymer chains on the gel particles is another approach to increase the response rate of hydrogels (Yoshida et al. 1995). Dangling chains in a gel easily collapse or expand upon an external stimulus because one side of the dangling chain is free.

  • Macroporous gels: Another technique to obtain fast-responsive hydrogels is to create voids (pores) inside the hydrogel matrix, so that the response rate becomes a function of the microstructure rather than the size or the shape of the gel samples (Okay 2000). For a polymer network having an interconnected pore structure, absorption or desorption of water occurs through the pores by convection, which is much faster than the diffusion process that dominates the non-porous hydrogels.

The basic technique to produce macroporous hydrogels involves the free-radical cross-linking copolymerization of the monomer-cross-linker mixture in the presence of an inert substance (the diluents), which is soluble in the monomer mixture (Okay 2000). In order to obtain macroporous structures, a phase separation must occur during the course of the network formation process so that the two-phase structure formed is fixed by the formation of additional cross-links. After polymerization the diluent was removed from the network, leaving a porous structure within the highly cross-linked polymer network. Thus, the inert diluent acts as a pore-forming agent and plays an important role in the design of the pore structure of cross-linked materials. Another technique to create a macroporous network structure is the use of inert templates in the preparation of hydrogels. By this technique, the polymer formation reactions are carried out in the presence of templates; a macroporous structure in the final hydrogel matrix appears after extraction of template materials. For example, by the cryogelation technique, the polymer formation reactions are carried out below the bulk freezing temperature of the reaction system (Lozinsky 2002). The essential feature of such reaction systems is that the monomers and the initiator are concentrated in the unfrozen microzones of the apparently frozen system. The polymerization and cross-linking reactions proceed in the unfrozen microzones of the reaction system. A macroporous structure in the final material appears due to the existence of solvent crystals acting as a template for the formation of the pores. The advantage of these so-called “cryogels” compared to the macroporous hydrogels obtained by phase separation is their high mechanical stability (Dinu et al. 2007). They are very tough and can withstand high levels of deformations, such as elongation and torsion; they can also be squeezed under mechanical force to drain out their solvent content. A typical SEM image of such materials in their dried state is shown in Fig. 7a illustrating their honeycomb morphology. These materials respond against the external stimuli such as the solvent composition change immediately (Fig. 7b).

Fig. 7
figure 7

(a) SEM of PAAm network prepared by cryogelation at −18 C. C o = 3 w/v %; 1.2 mol % MBAAm; magnification 50; scaling bar 100 μm. Reprinted from (Ozmen et al.2007) with permission from Taylor & Francis Group. (b) Swelling and deswelling kinetics of PAAm hydrogel in water and in acetone, respectively, shown as variation of the relative weight swelling ratio m rel with time of swelling or deswelling. C o = 3 w/v %; 1.2 mol % MBAAm; polymerization temperature − 18 °C ( filled circle) and + 21 °C (open circle). Reprinted from Dinu et al. (2007) with permission from Elsevier

Full size image