Hydration studies in polymer hydrogels



Polymer hydrogels have attracted much interest in recent years based on numerous applications mainly in biotechnology and medicine. For the knowledge-based design and development of new materials for these and similar applications, it is essential to understand better the hydration properties of hydrogels and of polymers in general. With this term, we mean the particular organization of water in the hydrogel, which determines the properties of the water component, typically different than those of bulk water, and the impact of water on the properties of the polymer matrix itself. In this review, we focus on recent work with hydrogels based on poly(hydroxyethyl acrylate), mostly copolymers with a second hydrophobic polymer and silica nanocomposites. The combination of water sorption/diffusion, thermal and dielectric studies, by fully exploiting the capabilities of each individual technique, proves essential in providing significant information on particular aspects of hydration, such as water uptake, water organization, and diffusion coefficients; glass transition and plasticization; water and polymer dynamics; protonic conductivity, and in revealing interesting correlations between these particular aspects. In the outlook similarities and differences to other related systems, such as protein-water and polymer solutions in non-polar solvents, are stressed in the perspective of a broader study. © 2012 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys, 2013


Polymer hydrogels are hydrophilic macromolecular networks which are able to absorb large amounts of water, because of being hydrophilic, but swell rather than dissolve in water, because of being cross-linked.1, 2 Their hydrophilicity is the consequence of the existence of both polar groups included in their chemical structure, which form hydrogen bonds with water molecules, and network expansion. Good biocompatibility and water permeation properties, in addition to several other good properties arising from their polymeric nature, form the basis for several applications of polymer hydrogels, in particular biomedical applications,1–7 however also applications in other diverse fields, such as ion-conducting membranes,8 water retention in agriculture,9 and sensors and actuators.10–12 The scientific and technological interest in hydrogels is reflected also in the large number of recent reviews on various aspects of hydrogels and related topics, next to those on applications mentioned above.13–18 For designing new materials with superior properties for these and similar applications, it is essential, among others, to investigate in detail and understand better their hydration properties. With this term, we mean the particular organization of water in the hydrogel, which determines the properties of the water component, typically different than those of bulk water, and the impact of water on the properties of the polymer matrix itself.19

Two different concepts may be followed, and have in fact being followed, to discuss hydration properties of polymers, including polymer hydrogels. In the first, water absorbed in a polymer is classified into different states (classes), depending on the particular experimental technique used, such as freezable and non-freezable water; mobile, immobile, and clustered water; free and bound water. Jhon and Andrade20 were the first to introduce a three-state model in hydrogels based on the observation of structured water in the vicinity of a solid surface and in natural macromolecular gels. In their classification, “bound” water consists of water molecules interacting strongly with specific primary hydration sites such as hydroxyl or ester groups, so that it behaves dynamically and thermodynamically as a part of the polymer chains; “intermediate” water is formed by water molecules with weaker interaction to polymeric chains or preferentially structured around the polymer network; finally, “free” water is formed by water molecules with negligible interaction to polymeric chains.20 The properties of water, determined to a large extent by specific polymer-water interactions21 and by geometrical confinement in the pores of the polymer,22 and the impact of water on the properties of the polymer matrix are different for the different states of water. Using this more traditional concept, for example, Lu et al.23 resolved different populations (classes) of water in Nafion membranes by means of dielectric spectroscopy; Ruiz et al.24 discussed the results of differential scanning calorimetry (DSC), 1H nuclear magnetic resonance (NMR), and wide-angle X-ray diffraction (XRD) measurements in equilibrium-swollen poly(vinyl alcohol) (PVA) derivative hydrogels in terms of different classes of water, their number depending on the particular technique used. In a series of articles, McBrierty et al.25 (and references therein) investigated the complex behavior of water and buffered saline in soft contact lenses hydrogels based on various polymers and in wide ranges of water contents, mimicking the ocular environment, which consists essentially of saline solution (tears), using DSC and NMR. The results by the two techniques were correlated to each other and were discussed in terms of three types of water in general: loosely bound water, more tightly bound and highly dispersed water, and near-normal water. The behavior of these different types of water was followed as a function of temperature. The trend of relaxation times as a function of temperature was interpreted in terms of complex water behavior and onset of motion of “glasslike” water at 180 K.25

Using the second concept, hydration properties are discussed in terms of phase diagrams26 and chemical exchange processes between water protons and hydroxyl protons of polymer chains,27 whereas no resource is made of the concept of different classes of water. Figure 1 shows a schematic representation of such a phase diagram with the characteristic dependences of crystallization (Tc) and melting (Tm) temperatures and of the glass transition temperature (Tg) on the fraction of water in the hydrogel (ω, defined by the ratio of water in the hydrogel and the total mass of the hydrogel).28 Roorda argued that data previously discussed in terms of different classes of water, such as DSC melting endotherms and exotherms showing structure, reflect merely the development of non-equilibrium conditions in the hydrated system.29 Salmeron et al.30 showed by DSC the existence of non-crystallizable solvent in a system of non-polar solvent swelling a hydrophobic polymer and argued that polar water–polymer interactions cannot be the sole factor responsible for the existence of “non-freezable” water. Molecular dynamics (MD) simulations in a PVA hydrogel with a water volume fraction of 40% gave evidence for two states of water only at a temperature below the freezing point of bulk water and showed that water molecules in contact with the polymer are less mobile than free water even below the freezing transition of bulk water.31 McBrierty and coworkers32 pointed out that, obviously, there are elements of truth in both concepts, and it is hard to rule out the presence of different types of water as specified by their characteristic behavior, in particular if several experimental techniques are used and a broad temperature range of measurements is covered.

Figure 1.

Schematic representation of the phase diagram of a polymer hydrogel. ω is the fraction of water in the hydrogel, Tc and Tm are the crystallization melting temperatures of water in the hydrogel, and Tg is the glass transition temperature of the swollen hydrogel. (Reproduced from Ref. 28, with permission from [Elsevier Science]).

Several experimental techniques have been used to study hydration properties of polymers. These include water sorption/diffusion measurements to follow water uptake at equilibrium and as a function of time (dynamic);33, 34 DSC to follow thermal transitions of water and of the polymer matrix itself;28, 35 spectroscopic techniques, in particular NMR (including the various new spectroscopy versions36) and dielectric relaxation spectroscopy (DRS) to follow water and polymer dynamics,19, 33–42 in addition to MD simulations;43 dynamic mechanical analysis (DMA) to follow effects of water on mechanical properties and mechanical stability and on glass transition of the polymer matrix (plasticization);34, 44 infrared measurements to monitor polymer-water interactions.45, 46 We would like to stress here the necessity of combining several, complementary techniques, covering together broad ranges of spatial and time scale (e.g., frequency in DRS) and of temperature to gain a complete picture of hydration. For example, NMR can routinely detect the onset of mobility in glass-like water at low temperatures, where DSC is largely insensitive.9 In addition, NMR enables independent analysis of the water by using 1H, 2H, 17O, and of the polymer chains with 1H, 13C, and 15N nuclei, so that information can be extracted from the NMR measurements on MD (from relaxation dynamics) and on hydrogen bonding (from the induced chemical shifts).36 MD simulations, on the other hand, can be obviously more selective in the study of interaction of water molecules with specific groups in the polymer structure or with solid surfaces having specific properties. Thus, by using a set of idealized models of solid surfaces, the effects of particular parameters, which characterize a polar surface, such as density, distribution, orientation, and flexibility of the hydroxyl groups, on the density of water in the interfacial layer and the ability to hydrogen bond formation to other water molecules in further layers could be separated in MD simulations.16, 47 To give an example of results, it was found in these simulations that increase of the surface density of hydroxyl groups increases the wetting coefficient and more water molecules come closer to the surface which results in a decreased reduction of the density of water in the interfacial region.47 In another study, MD simulations were performed for three polymeric hydrogel models focusing on the effects of different polar groups on structure and dynamics of water.48 The dynamics of hydrogen bonds was investigated through the lifetime distribution and the hydrogen-bond correlation function. The effects of polymers on the properties of water molecules were found to vary with the species of polar groups. The mobility of water molecules was found to be highly reduced around polymer chains for both translational and rotational motions, because of both hydrogen bond formation with hydrophilic groups and structuralization of water around the hydrophobic groups.48

In this review, we focus on the hydration properties of hydogels based on poly(hydroxyethyl acrylate) (PHEA) by means of water sorption/diffusion measurements, thermal and dielectric techniques. The chemical structure of PHEA monomeric unit is shown in Figure 2. We take the specific system studied as an example to illustrate the theoretical concepts used and the methodology of combining several, partly complementary techniques to study in detail hydration properties in a polymeric system. Water sorption measurements, dynamic and at equilibrium, from the vapor phase and by immersion in liquid water, provide information on water uptake and diffusion coefficients. Measurements by DSC and by a special dielectric technique in the temperature domain, thermally stimulated depolarization currents (TSDC),19 at several levels of water content, provide information on thermal transitions, in particular glass transition and crystallization/melting of water. Finally, DRS measurements in wide ranges of frequency and temperature, again at several levels of water content, provide, after thorough, sophisticated and critical analysis, detailed information on water and polymer dynamics. Correlations between results obtained by different techniques shed new light on the hydration properties of the specific system and open new perspectives, in relation also to similar studies on other systems.

Figure 2.

Schematic diagram of the structure of PHEA monomeric unit.

Next to the fundamental and methodological interest of the present review, there is also a strong practical interest based on the many, mostly biotechnological, applications of the concrete PHEA-based hydrogels. PHEA is less known than poly(hydroxymethyl methacrylate) (PHEMA), the material of contact lenses.2 A main difference between the two polymers, of advantage for PHEA in many applications, is the much lower glass transition temperature (Tg) of PHEA (around 20 against around 70 °C for the dry polymers), arising from the replacement of the bulky [BOND]CH3 group close to the polymer backbone in PHEMA by the smaller H atom in PHEA. Moderately cross-linked PHEA can accommodate water up to 30–40% of its dry mass when exposed to humid atmosphere,19 becomes however very soft. For most of the applications envisaged, such as materials for scaffolds for tissue engineering and for implantation, mechanical stability should be improved. This has been done in previous work by combining PHEA with a second hydrophobic component in the form of interpenetrating polymer networks (IPNs)19, 34, 49 and statistical copolymers19, 44, 50 and/or by reinforcing the polymer matrix with silica nanoparticles using the concept of polymer nanocomposites.46, 51 Of particular interest are nanocomposite biopolymer hydrogels, which form an emerging class of materials with potential applications in biotechnology and biomedicine.1, 14, 15 In the case of copolymers and IPNs, mechanical stability is improved at the expense of water uptake capability. In the case of microphase-separated IPNs, the hydrophilic PHEA domains behave essentially like pure PHEA, that is, water uptake increases in proportion to PHEA fraction in the IPN.19, 49 Results are less predictive in the case of homogeneous statistical copolymers.19, 33, 52 Finally, an advantage of using silica nanoparticles for polymer matrix reinforcement is that these are hydrophilic, making their own contribution to water uptake.53 It is interesting to mention in this connection that, next to improving mechanical stability, the addition of a second hydrophobic and biodegradable polymer, such as poly(l-lactide)54 and poly(ε-caprolactone) (PCL),40 to PHEA, which is not biodegradable itself, has been used to make biodegradable hydrogels. Complexity of the system increases in this way, however, at the same time more options are available for designing new materials with improved properties.55 A prerequisite for that is to understand better structure-property relationships of these complex polymer matrices, including also their hydration properties.

The hydration properties of PHEA-based hydrogels have been reviewed in 2005.19 Thus, emphasis in the present work is given to work done and published within the last 7–8 years. In addition, the review focuses mostly on using PHEA copolymers, as well as PHEA/silica and copolymer/silica nanocomposites, as hydrogel matrices (xerogels). The review is organized as follows. The next section is devoted to water uptake and water diffusion coefficients. The following three sections are then devoted to thermal transitions, dynamics, and electrical conductivity, respectively, studied at various levels of water content. In the last section, perspectives, we discuss correlations to other related systems and outlook to future work.


In this section, we review results on water uptake and diffusion coefficients obtained with PHEA-based hydrogels including neat PHEA, copolymers, and nanocomposites. Figure 3 shows results of water sorption measurements from the vapor phase in random copolymers of PHEA and the hydrophobic polymer poly(ethyl acrylate) (PEA) at different compositions: water content h, defined as the ratio of the weight of water in the hydrogel to the weight of the dry sample (dry basis), determined by weighing, against water activity, αw [relative humidity (RH)]. For these equilibrium sorption isotherm (ESI) measurements, samples were allowed to equilibrate to constant weight over saturated salt solutions in sealed jars at controlled RH at room temperature,50 whereas dry weights were determined by drying to constant weight in vacuum (5 × 10−2 Torr) for at least 24 h at 80 °C. The P(HEA-co-EA) copolymers were obtained by simultaneous photopolymerization of the corresponding HEA and EA monomers. They are coded in Figure 3 and in the following by XX/YY, where XX and YY are the weight fractions of HEA and EA components, respectively. Ethyleneglycol dimethacrylate, 2 wt % with respect to monomer, was used as cross-linking agent. Details of preparation have been given elsewhere.44

Figure 3.

Water content h against water activity αw for the PHEA-co-PEA samples indicated on the plot at 25°C. Two different data sets are represented for neat PHEA. The lines are fittings of eq 1 to the experimental data. The inset shows the Brown plot for the analysis of water clustering in neat PHEA. (Reproduced from Ref. 50, with permission from [Elsevier Science]).

The water sorption isotherms of Figure 3 are of class III in the Brunauer classification,56 that is, they describe adsorption onto adsorbents with weak adsorbate–adsorbent interactions. Regarding effects of copolymer composition, we observe that at each water activity, h decreases with increasing PEA content, reflecting the decreasing hydrophilicity of the sample. Gravimetric measurements and NMR imaging studies in PHEMA and its copolymers with other hydrophobic methacrylate monomers gave lower values for the copolymers for both water diffusion coefficient and maximum water uptake.57 For pure PHEA and 70/30 PHEA-co-PEA, a steep increase of h is observed in Figure 3 at αw larger than 0.5–0.6, reflecting the completion of the first hydration layer and the formation of clusters around the primary hydration sites.58 For analyzing these and similar ESI data, an equation derived by a model, typically in terms of different classes of water, is fitted to the experimental data.19, 49 Here, the Guggenheim–Anderson–DeBoer (GAB) equation was used49, 58

equation image(1)

In this equation, hm is the first monolayer sorption capacity, c is the ratio of the chemical potential of water molecules in the first monolayer to that of water in the second and successive layers, and f is the ratio of the chemical potential of water molecules in the second and successive layers to that of liquid water at the same temperature and pressure. The results show that hm, the most significant parameter in eq 1, decreases with increasing PEA content, from 0.07 in pure PHEA to 0.035 in 70/30 PHEA-co-PEA and to 0.007 in 10/90 PHEA-co-PEA. From the value of hm for pure PHEA, we calculate the number of primary hydration (sorption) sites per each HEA segment by49

equation image(2)

where MHEA and Mw are the molecular weights of the repeating unit of the PHEA network and of water, respectively, to 0.41. Previous work by several investigators19, 21, 32, 58 has provided evidence that every adsorbed water molecule in PHEA is associated with two hydroxyl groups in adjacent side chains (Fig. 2). Thus, the calculated value of 0.41, smaller than 0.50, indicates that a fraction of hydroxyl groups of about 20% in neat PHEA are not available as primary sorption sites, presumably being compromised in intermolecular or intramolecular hydrogen bonds forming micelles.50, 59

A different method of analysis has been proposed by Brown,60 which combines conventional solution theory (Flory–Huggins) and cluster theory (Zimm and Lundberg61, 62). The inset to Figure 3 shows for the neat PHEA hydrogel, the plot of reciprocal water content against reciprocal water activity. From the slope and the intersection of the straight line with the 1/h axis, we can calculate53, 58 the critical water content where formation of clusters sets into hc = 0.11, which corresponds to a water activity of about 0.65, and the mean number of water molecules in a cluster Nc at the highest water content of about 0.30 to about 4.50

Figure 4 shows results of water sorption measurements from the vapor phase, that is, similar to those of Figure 3, but now for PHEA/silica nanocomposites.53 A different procedure of measurements was used in that work by uing an IGASorp Moisture Sorpion Analyzer (Hiden Analytical, Warrington, England), which enables to perform both dynamic and equilibrium sorption measurements at various levels of water activity.53 Thus, in addition to ESIs (Fig. 4), water diffusion coefficients at various levels of water activity, corresponding to various levels of water content, could be determined. The nanocomposites of Figure 4, designated in the following as PHEA/x%silica, where x denotes the weight content of silica, were prepared by the simultaneous polymerization of HEA and tetraethoxysilane in a sol-gel process as described elsewhere.63 Measurements by a variety of experimental techniques52, 63 showed that silica is porous, the pores being filled with polymer. The silica phase consists of elementary particles, 30–80 nm large, which form larger aggregates up to about 400 nm in size. With increasing silica content, the isolated silica aggregates start to connect themselves through silanols present on their surfaces and finally, at the percolation threshold of about 15 wt % silica, become a continuous inorganic network. Similar to those in Figure 3 for the copolymers, the water sorption isotherms of Figure 4 are of class III in the Brunauer classification.56 The inset to Figure 4 shows the sorption of a silica sample, prepared also by sol-gel, which is of type I in the Brunauer classification.56 The data for silica were analyzed according to the Langmuir equation (fit in Fig. 4), we refer to Pandis et al.53 for more details.

Figure 4.

Equilibrium sorption isotherms of neat PHEA, neat silica, and PHEA/silica nanocomposites measured at room temperature. The solid lines are fittings of eq 1 to the data. The inset shows a fitting of the Langmuir equation49 to the neat silica data. (Reproduced from Ref. 53, with permission from [John Wiley & Sons, Inc.]).

We observe in Figure 4 that for the PHEA/5%silica nanocomposite, that is, below the percolation threshold of about 15 wt % silica, the isotherm is practically indistinguishable from that of pure PHEA. The PHEA/25%silica nanocomposite, on the other hand, with silica content above the percolation threshold, shows a lower hydrophilicity, in particular at high αw values. This and similar results have been analyzed and discussed in detail in terms of constraints imposed by the silica network on the expansion of PHEA matrix at higher water activities.53 Analysis by means of the GAB eqs 1 and 2 gave for pure PHEA a value of 0.34 for the number of primary hydration sites per HEA segment,53 in rather good agreement with the value of 0.41 obtained from the data in Figure 3.50

The constraints to swelling of the PHEA matrix imposed by the silica network become much more pronounced in the case of sorption by immersion in liquid water. The results in Figure 5 show that water uptake decreases continuously and significantly as silica content increases up to 15 wt % and becomes then almost independent of silica content. In the same figure, the water content obtained from sorption measurements from the vapor phase at αw = 0.98 is also depicted for comparison. In contrast to above, only a slight decrease of water content is observed as the amount of silica increases. The difference in water uptake by the two modes of sorption decreases with increasing silica content up to the percolation threshold and stabilizes then. These findings have been explained taking into consideration that in the case of pure PHEA, the hydrogel has the ability to swell freely, leading to high water uptake. Contrarily, in the nanocomposites the addition of silica particles into the PHEA matrix affects the liquid water uptake ability of the hybrid material, by partially restricting the swelling of PHEA, the restrictions becoming more pronounced at 15 wt % and higher silica contents when a silica network has been formed.53 Please note that in addition to the silica content dependent difference in water uptake from the liquid and the vapor phase in Figure 5 and the explanation given above, there is a general phenomenon of deviation between the water uptake from the liquid phase and that from the saturated vapor phase, both of unit water activity, observed in many gels and ionomers, commonly referred to as Schroeder's paradox, the origin of which is still not fully understood.8, 64

Figure 5.

Liquid water and water vapor uptake of PHEA/silica nanocomposites against silica content obtained from immersion into liquid water and from sorption measurements at αw = 0.98, respectively. (Reproduced from Ref. 53, with permission from [John Wiley & Sons, Inc.]).

Figure 6 shows results for the water diffusion coefficient D against water content h for the PHEA/silica nanocomposites of Figures 4 and 5. D-values were calculated according to Fick's second law for diffusion in one dimension by53, 65

equation image(3)

where (Δm)t is the water uptake or loss at time t (sorption or desorption, respectively), (Δm) is the corresponding limiting value at equilibrium, and l is the thickness of the sample presumed constant over the whole sorption or desorption process. By plotting (Δm)t/(Δm) against equation image D-values were obtained from the initial slope for (Δm)t/(Δm) < 0.5. Gravimetric measurements and NMR imaging studies in PHEMA and its copolymers with other hydrophobic methacrylate monomers were analyzed also in terms of Fick's equation.57 We observe in Figure 6 that D-values corresponding to each water activity are similar for PHEA and PHEA/5%silica, that is, below the percolation threshold, whereas lower values (by a factor of up to two) are obtained for PHEA/25%silica, which is above the percolation threshold. These results correlate well with those of DSC measurements on the same materials showing an increase of the glass transition temperature in the nanocomposites, as compared to the neat PHEA matrix, and immobilization of a polymer fraction resulting from strong polymer-filler interactions.63 Thus, stiffening of the polymer matrix is the reason for the reduced diffusion coefficients rather than constraints imposed by the presence of the silica nanoparticles, which would increase the tortuosity of diffusion paths. Regarding the effect of water content h, an initial increase of D is observed up to a maximum, followed by a decrease at higher h values. The same trend was observed for all the compositions studied and was explained by Pandis et al.53 in terms of polymer-water and water-water interactions. Interestingly, a similar h dependence of D was observed in Nafion membranes.8 Diffusion coefficients refer to non-equilibrated states and Shapiro36 stressed the necessity to establish a correlation between the motions in equilibrated state and the diffusion in the real-time non-equilibrated dynamic situations in hydrogels, where the swelling and the dissolution of the gel network, the compatibility of the solvent, solute, and polymer should all be considered.

Figure 6.

Diffusion coefficient of water D against water content h for the PHEA/silica nanocomposites indicated on the plot. (Reproduced from Ref. 53, with permission from [John Wiley & Sons, Inc.]).


In this section, we review results for the thermal transitions of the hydrogels obtained by DSC and by TSDC. The former is the classical technique for thermal transition studies, whereas the latter is a special dielectric technique in the temperature domain, a thermodielectric technique. It consists of studying the thermal release of stored dielectric polarization.66 In terms of traditional dielectric spectroscopy in the frequency domain, TSDC corresponds to measuring dielectric loss as a function of temperature at a low-equivalent frequency in the range 10−2 to 10−4 Hz. The technique is characterized by high sensitivity and high peak resolving power.66

Figure 7 shows results for the glass transition temperature of a PHEA hydrogel as a function of water fraction in the hydrogel, defined as the ratio of the weight of water in the hydrogel to the weight of the sample (wet basis). Tg is the glass transition temperature measured by DSC and TmII is the temperature of current maximum (peak temperature) of the TSDC peak corresponding to the α relaxation associated with the glass transition.66 The original DSC and TSDC thermograms at several levels of water fraction can be found in ref.58. The first observation, of particular interest from the methodological point of view of studying glass transition, as pointed out by Kyritsis et al.,58 is that, at each hydrogel composition, Tg and TmII are close to each other, providing evidence that TmII is a good measure of Tg, obviously due, at least partly, to similar equivalent frequencies of the two techniques.66 Both Tg and TmII decrease steeply with increasing water fraction x2, expressing the strong plasticization of the glass transition by water. Tg can be unambiguously determined by DSC only up to x2 ∼ 0.2, due to superimposed crystallization/melting peaks of water at higher water fractions. Interestingly, again from the methodological point of view of studying glass transition in hydrogels, this limitation does not apply for TSDC. Here, we focus on the glass transition and we refer to ref.58 for a detailed study of crystallization/melting of water in PHEA hydrogels.

Figure 7.

Peak temperature of the TSDC α peak, TMII, (⊙) and glass transition temperature, Tg, (•), measured by DSC, against water fraction x2 in neat PHEA hydrogels. The lines are fits of eq 4 to the data, details in text. (Reproduced from Ref. 58, with permission from [Elsevier Science]).

The lines in Figure 7 are fits of the Couchman–Karasz equation58, 67

equation image(4)

to the experimental DSC (solid line) and TSDC results (dotted line). In this equation, x1 and x2 are the weight fractions of PHEA and water, respectively, in the hydrogel, Δcp1 and Δcp2 the corresponding heat capacity increments at the glass transition, and Tg1 and Tg2 the corresponding glass transition temperatures. We refer to ref.58 for values of the fitting parameters and details of fitting. Here, we would like to stress that the results in Figure 7 strengthen the idea that the PHEA/water system is a homogeneous one for water contents lower that about 0.2–0.3, whereas phase separation occurs at higher water contents, in agreement with the observation of crystallization/melting events.58 The critical water content of about 0.3 for the appearance of a separate water phase corresponds to about two water molecules per repeating unit (eq 2). We will come back to this homogeneous/heterogeneous picture and the corresponding critical hydration values in the next section with respect to plasticization of secondary relaxations. Interestingly, similar values of critical water contents for the formation of a separate water phase of 0.2–0.3, corresponding to about two water molecules per aminoacid residue are obtained for the globular proteins lysozyme and bovine serum albumin (BSA) by a combination of thermal and dielectric techniques.68, 69

Fornasiero et al.70 used temperature modulated DSC to investigate plasticization of the glass transition in three commercial soft contact lens hydrogels based on different polymers and/or cross-linking agents. Strong plasticization was observed, similar to the results shown in Figure 7, different for the three types of hydrogels. The experimental results were satisfactorily fitted by the Gordon–Taylor equation.70 Measurements were limited to rather low water contents, so the saturation of the plasticization and the leveling off of Tg observed in Figure 7 could not be followed in that work. By extending measurements to higher water contents Hodge et al.71 followed clearly the stabilization of Tg in PVA hydrogels at high water contents and postulated that non-freezing water is responsible for the majority of plasticization in the hydrogel studied. The mechanism of plasticization was investigated by combining positron annihilation lifetime spectroscopy, 13C solid state NMR and DMA. The authors attributed plasticization to the increasing free volume (measured by positron annihilation lifetime spectroscopy) and lubrication provided as the water swells the polymer and disrupts polymer-polymer hydrogen bonding.

Figure 8 shows results similar to those in Figure 7, obtained by DSC, for a more complex PHEA-based hydrogel, PCL–PHEA copolymer network at a composition PCL/PHEA 30/70. We refer to ref.40 for details of the preparation of the networks and their investigation by a combination of DSC/TSDC/DRS techniques. The copolymer is microphase-separated, so two glass transitions are observed at each water fraction. The one at lower temperatures originates from the hydrophobic PCL phase and the corresponding Tg does not change with water fraction, whereas that at higher temperatures is strongly plasticized by water, indicating that it originates from the hydrophilic PHEA phase. These results suggest that water is practically absorbed only in the hydrophilic phase. More information about the microphase separation of the hydrogel matrix has been obtained by fitting the Couchman–Karasz eq 4 to the experimental data. The two components in this equation are now the PHEA fraction and the water fraction absorbed in the PHEA microphase (solid line in Fig. 8). The fit is not satisfactory and becomes so (dotted line) only if it is assumed that microphase separation is incomplete and water is absorbed in a PHEA-rich phase, consisting of the whole PHEA fraction mixed at the nanometer scale with some PCL units (see ref.40 for more details).

Figure 8.

Glass transition, Tg, of the PCL phase (•) and of the PHEA phase (▴) against water fraction, w, in PCL/PHEA 30/70 hydrogels. The lines are fits of eq 4 to the data, details in text. (Reproduced from Ref. 40, with permission from [Springer Science]).

If the hydrogel matrix is phase-separated into hydrophilic and hydrophobic microphases and water is absorbed only in the hydrophilic one, then we can predict the properties of each phase, in particular the water content dependence of the two Tgs, as mentioned in Introduction. Experimental TSDC results in sequential IPNs of PHEA and the hydrophobic PEA, with practically complete microphase separation, show that in a plot similar to those in Figures 7 and 8, Tgs for the different IPN compositions fall into a single curve if water content is normalized to the PHEA fraction.59 The situation is completely different for a homogeneous hydrogel matrix of the same two components, the random PHEA-co-PEA copolymers of Figure 3. Figure 9 shows results for the dependence of the TSDC peak temperature of the α relaxation in nanocomposites of copolymers with various PHEA/PEA weight fractions and fixed at 20 wt % silica on water content h divided by the PHEA fraction of the xerogel.52 The materials were prepared by simultaneous copolymerization of the HEA and EA monomers and silica generation by the sol-gel method and are designated as XXYYsZZ, where XX and YY are the fractions of HEA and EA, respectively, in the organic phase and ZZ the fraction of silica in the nanocomposite xerogel.52 Please note that the silica content, which is fixed anyway in Figure 9, has no significant effect on Tg in these nanocomposites.52, 72 The results show that, in contrast to the microphase-separated IPNs mentioned above, Tg depends sensitively on copolymer composition, increasing significantly with increasing PEA fraction, for the same normalized water content. In addition to their theoretical interest, these results are also of practical interest with respect to applications of hydrogels based on various combinations of their hydrophilic and hydrophobic components, such as in scaffolds for tissue engineering and as implantation materials.34, 40, 44, 46

Figure 9.

Peak temperature of the TSDC α peak (cop) against water content divided by the HEA content in PHEA/PEA nanocomposites of the compositions given on the plot and fixed 20 wt % silica. The lines are guides to the eye. (Reproduced from Ref. 52, with permission from [Elsevier Science]).

In the results presented above, DSC was combined with TSDC to follow in details plasticization of the glass transition. Other aspects of hydration can be studied by combining DSC to follow thermal transitions with other techniques. Daniliuc and David73 combined DSC and Fourier transform infrared spectroscopy (FTIR) to study water–polymer interactions in water-vapor saturated PVA, poly(acrylic acid) and their blends. Water–polymer interactions which are formed in the presence of water and polymer-polymer interactions which are destroyed at the same time were identified by focusing on the OH stretching, C[DOUBLE BOND]O stretching, and HOH bending regions in the FTIR spectra. The results showed that polymer-bound water (non-freezing water) contains two types of molecules: weakly and strongly hydrogen bound to the chains absorbing in different regions of the FTIR spectra and vaporizing at different temperatures. A controversial question of general interest refers to the relative significance of such specific polymer–water interactions (and, thus, of the chemical structure of the polymer) for the mobility of water and of the polymer, and of the final properties of the hydrogel. Temperature-dependent Raman spectroscopy measurements in dried glassy poly-N,N,-dimethylacrylamide indicated structural changes of bound water at about 310 K.74 The authors concluded that these changes have to be considered together with the chemical structure of the polymer to understand the nature of hydrated polymer systems. Also, DSC and 1H NMR T2 measurements in proton conducting sulfonated copolymer membranes for fuel cells at various levels of water content showed saturation of the glass transition in good correlation with freezing and melting events of water.75 The authors concluded that the state of water within the polymer rather than the total water content may be a stronger indicator of the water and methanol transport properties in these membranes. Conversely, by investigating a variety of contact lens hydrogels over wide ranges of water content at temperatures well above the freezing temperature of water, Pope and co-workers37, 76 proposed a simplified exchange model for the complex 1H NMR T1 and T2 relaxation, which can be applied to a range of materials using the same set of parameters. The model suggests that the water in these materials is only slightly perturbed relative to bulk water and can be described by either the presence of free and bound water in fast exchange or a single water species, that is slightly modified in its behavior relative to the behaviour of free water. Moreover, the average mobility of both water and polymer are largely dependent on water content. No significant dependence of this mobility on polymer composition was found that could not be related to water content.37, 76


In this section, we review results for the water and polymer dynamics of the hydrogels obtained by two dielectric techniques, broadband DRS and TSDC. Dielectric studies in PHEA-based hydrogels show, in agreement with other techniques, such as NMR21, 27, 32, 36, 38, 44 and DMA34, 44 two secondary relaxations at lower temperatures, the γ and the βsw, and the segmental α relaxation associated with the glass transition (dynamic glass transition) at higher temperatures. The γ process is attributed to the motion of the polar group [BOND]CH2[BOND]CH2[BOND]OH in the side chain of the HEA monomer (Fig. 2) and the βsw process is then assigned to the association of one water molecule with the polar groups of two adjacent side chains.49, 58, 59, 66, 77 All three relaxation processes are significantly affected by water and their detailed investigation has provided in the past important information on polymer–water interactions, phase morphology, and molecular mobility.19 It should be noted that a similar picture is obtained with hydrogels based on other polymers, a main difference originating from the number of the secondary relaxations in the corresponding xerogel, in some cases related with its semi-crystalline nature.39, 78 Thus, five relaxations were reported in cellulose and other polysaccharide materials by Einfeldt et al.79 the one of particular interest with respect to water, similar to the βsw process in the PHEA-based hydrogels, being the βwet relaxation.

Figure 10 shows results for the frequency dependence of dielectric loss (imaginary part ε″ of the complex dielectric permittivity) for PHEA hydrogels at selected values of RH and temperature to illustrate the evolution of the three processes, γ, βsw [Fig. 10(a)], and α [Fig. 10(b)], with hydration level. The βsw process, at lower frequencies than γ, practically absent in the nearly dry sample, emerges with the first traces of water and becomes significant in magnitude already at low water contents (comparable to γ already at RH = 11%, corresponding to h = 0.01–0.02) and also faster, whereas the γ process is less sensitive to water content [Fig. 10(a)]. The two processes merge at higher hydration levels and we observe then only the βsw process, because of its higher strength. The α process shifts to higher frequencies with increasing RH [Fig. 10(b)], much faster than the secondary relaxations, it cannot be, however, resolved at higher hydrations, because of overlapping with the rapidly increasing electrical conductivity, which gives rise to the steep wing at low frequencies. Interestingly, this limitation does not hold in the case of TSDC measurements,19, 66 as demonstrated also by the results in Figure 7. Bermejo and Mijangos Ugarte43 used fully atomistic MD simulations focusing on the effect of water on polymer dynamics in aqueous solutions of PVA over wide ranges of composition, in particular on segmental dynamics. It was found that PVA dynamics is strongly plasticized by water and different mobilities are observed for the different PVA hydrogens, this dynamic heterogeneity decreasing with increasing water fraction of the solution.

Figure 10.

Dielectric loss against frequency, ε"(f), in neat PHEA hydrogels at −80 °C (a) and at 50 °C (b), at several values of RH indicated on the plot. (Reproduced from Ref. 50, with permission from [Elsevier Science]).

For a systematic analysis of the dielectric data of Figure 10 and other similar data in terms of time scale, strength and shape of each relaxation, model functions are typically fitted to the data. For the results shown in the next figures, the Havriliak–Negami (HN) model function19, 50, 80 was used

equation image(5)

In this expression Δε is the dielectric strength, Δε = εs − ε, where εs and ε are the low- and high-frequency limits of ε′, respectively, τ is the relaxation time, τ = 1/2πfHN, where fHN is a characteristic frequency closely related to the loss peak frequency fmax, and α, β are shape parameters describing the shape of the ε′′(ω) curve, 0 < α ≤ 1 and 0 < β ≤ 1. Depending on the shape of the data at each temperature and RH of measurements, a more complex expression may be used, consisting of the appropriate sum of HN terms, one for each relaxation, plus one term for conductivity.

Figure 11 shows results for the time scale of the three relaxations in the form of the Arrhenius plot (log of loss peak frequency fmax against reciprocal temperature, activation diagram) for 70/30 PHEA-co-PEA at selected relative humidities, characteristic also for other PHEA-based hydrogels.19, 50, 72 Included in the plot are also TSDC data for the βsw and the α processes, peak temperatures at the equivalent frequency of 1.6 mHz, corresponding to a relaxation time of 100 s,66 generally in good agreement with the DRS data. We observe that all three processes are plasticized by water, each at a different degree. The lines in Figure 11 are fittings of the Vogel–Tammann–Fulcher (VTF) equation81

equation image(6)

to the data for the cooperative α process and of the Arrhenius equation81

equation image(7)

to the data for the local, secondary γ, and βsw processes. In these equations A, B, To (Vogel temperature), Eact (activation energy), and fo are temperature independent empirical constants.

Figure 11.

Arrhenius plot for the α and βsw (a) and the γ (b) relaxation processes observed in 70/30 PHEA-co-PEA hydrogels at various values of RH indicated on the plot. Included in Figure 11(a) are also TSDC data for the copolymer at the equivalent frequency of 1.6 mHz, as well as DRS data by Sugimoto et al.85 obtained with cellulose at RH = 93% (---). The continuous lines are fittings of the VTF eq 6 to the data for the α relaxation in the dry copolymer and of the Arrhenius eq 7 to the data for the βsw and the γ relaxations. The inset to Figure 11(b) shows the dielectric strength Δε of γ and βsw, obtained from eq 5, in pure PHEA against RH. (Reproduced from Ref. 50, with permission from [Elsevier Science]).

A question of general interest in hydrated systems, including hydrogels and proteins,51, 68, 69 refers to the potential contribution of water molecules themselves, organized in water clusters, by their reorientation, to the dielectric response in the region of the βsw process, probably coupled with this local process.82, 83 Related to this question is also the controversial discussion as to whether the so-called βwet-relaxation in polysaccharides originates from the combined motion of water–polymer complexes79 or from the local motion of hydrogen-bonded water molecules.84 The inset to Figure 11(b) shows data for the RH dependence of the dielectric strength Δε of γ and βsw, obtained from eq 5 in pure PHEA, representative also for those obtained with other PHEA-based hydrogels.50 Of particular interest is the significant, overproportional increase of Δε of βsw at RH values higher than about 54%, that is, after the completion of the first hydration layer (ESI results, Fig. 3), in qualitative agreement also with TSDC results not shown here,19, 50 indicative of such an additional contribution arising from water. It is interesting in that respect that Sugimoto et al.85 followed by DRS in hydrated cellulose at high RH values and analyzed a process in the temperature/frequency region of the βsw process [dotted line in Fig. 11(a)] and discussed that process in terms of contributions of both a cellulose relaxation and motion of adsorbed water molecules. Sutokawa and Shikata41 used high-frequency DRS measurements up to 20 GHz to follow dynamics in aqueous solutions of PVAs at 25 °C. They analyzed the experimental data into three contributions to the dielectric response arising from the rotational relaxation of bulk state water molecules, the exchange of hydrated water molecules to the OH groups of the PVA monomer units, and the local motion of the main chain of PVAs. The results were discussed in terms of intramolecular hydrogen bonding between mainly adjacent OH groups of PVA against intermolecular hydrogen bonding to water molecules. A change in the water fraction dependence of the dielectric strength, similar to that shown in the inset to Figure 11, suggesting two different contributions to the dielectric response, was also observed by Cerveney et al.86 in water intercalated in graphite oxide. A main difference to the results in the hydrogels presented above, as well as in protein-water mixtures,68, 69 is that both contributions were assigned to water, rather than one to the polymer (substrate) and the other to water with an interplay between them. The authors used DRS, DSC, XRD, and FTIR focusing on water structure and dynamics. The DRS data indicated the existence of two types of water molecules at low hydration levels, strongly and weakly bound water, with different rotational dynamics and a change of behavior at about 7% of water. At higher hydration levels, the dielectric strength was found to increase more rapidly. In agreement with that, FTIR spectra in the OH stretching region showed a marked shift in the position of the peaks indicating that water is progressively highly coordinated with increasing water content, supporting the presence of water clusters in the samples at high hydration levels.86

Even more striking are the changes of the temperature dependence of the dielectric strength, Δε(T), of the βsw relaxation with RH following the completion of the first hydration layer, shown in Figure 12 at the example of 70/30 PHEA-co-PEA. These results may suggest an additional contribution to the βsw process at high RH values, with a different temperature dependence of Δε, likely the relaxation of water molecules themselves, as mentioned above. It is interesting to note in this connection that in various water systems the relaxation of un-crystallized water was found to increase in magnitude with increasing temperature.87–89 The inset to Figure 12 shows results for the water content dependence of the dielectric strength, Δε(h), for the same composition, 70/30 PHEA-co-PEA, at the temperature of −63 °C, marked by the vertical, dotted line in the main figure. We observe a change in the slope of Δε(h) at RH = 54%, similar to that shown in the inset to Figure 11 for neat PHEA, in consistency with the assumption that an additional process, the reorientation of water molecules in water clusters, contributes to the βsw relaxation at higher RH values. Similar results were obtained also with the other copolymer compositions, as indicated also by the results in pure PHEA in the inset to Figure 11(b). It is interesting that at the two highest h values in the inset to Figure 12 the slope of Δε(h) changes again, now to smaller values, a point worth to be followed by measurements at higher water contents, for example, immersion experiments, in future work.

Figure 12.

Dielectric strength Δε of the βsw process against temperature at several values of RH indicated on the plot in 70/30 PHEA-co-PEA hydrogels. The straight lines through the experimental points are guides to the eye. The vertical arrows indicate Tgs determined by TSDC. The inset shows Δε against water content h at the temperature of −63 °C, marked by the vertical, dashed line in the main figure. The dotted line in the inset is a guide to the eye. (Reproduced from Ref. 50, with permission from [Elsevier Science]).

A coupling between polymer and water dynamics was observed also in other polymer–water systems by using a variety of experimental techniques. Diffusion measurements by means of 1H NMR in a polymer hydrogel at various levels of hydration provided evidence for anomalous diffusion behavior indicating the presence of bulk water diffusion and diffusion of molecules linked to the polymer chain with interchange of molecules between the two domains.90 The strength of the hydrogen bond and the average time, the water molecule remains linked to the polymer chains was found to be strongly affected by the chain dynamics, which is temperature dependent. Matyushov and Morozov91 proposed a theoretical model for the coupling between the protein atomic motions and the protein-water interface in protein-water solutions and analyzed two mechanisms of that coupling, viscoelastic deformation of the global protein shape plasticized by the hydration shell and electrostatic fluctuations coupled to the atomic charge.

Also changes of the activation parameters of the βsw relaxation with hydration level provide support for an additional contribution in the region of that relaxation arising from water molecules organized in clusters. The activation energy Eact of the βsw process of the PHEA-co-PEA hydrogels was found to increase with increasing RH up to the completion of the first hydration layer and then to decrease again.50 Moreover, in a so-called compensation plot, that is, a plot of activation energy Eact against pre-exponential factor fo with the RH of the measurement as parameter, a correlation between the two activation parameters was observed, and an increase of both with increasing RH, followed by a decrease at high RH values, after completion of the first hydration layer.50 Similar results with respect to changes of the activation parameters with hydration level and a compensation law were observed by Sugimoto et al. in hydrated cellulose.85 Also, in DRS measurements in water-filled nanoporous silica Spanoudaki et al.92 followed and analyzed a process in the temperature/frequency region of the βsw process. They attributed that process to water and found an increase of its activation energy with increasing water content, which they explained in terms of the build-up of the hydrogen bonding network.92 The dynamical properties of hydration water at elevated water fractions in various host matrices, for example, synthetic polymeric hydrogels and biomacromolecules, are the main subject in numerous research studies (refs68, 69, and references therein).


Hydrogels exhibit relatively high values of protonic conductivity, rapidly increasing with increasing RH/water fraction,8, 19 which may be used in various applications.93 Hickner8 stressed the importance of understanding the coupling of the adsorbed water with the polymer matrix and the dynamics of water inside the polymer network for tuning the transport and mechanical property tradeoffs in hydrated membranes used in various electrochemical devices. In this section, we review briefly protonic conductivity in PHEA-based hydrogels in relation to xerogel composition and water content dependence, focusing on understanding the transport mechanism at the microscopic level. As an example, Figure 13 shows the Arrhenius plot for dc conductivity σ for pure PHEA at several RH values indicated on the plot. Values of σ may be obtained directly from dc conductivity measurements, either two- or four-electrode measurements, depending on the level of conductivity.94 The results shown in Figure 13 were obtained from DRS measurements. To that aim, ac conductivity σac (actually real part of the complex conductivity) was calculated from the measured dielectric loss ε″(ω) by93, 95

equation image(8)

where εo is the permittivity of free space and σ values were obtained as the plateau (frequency independent) values of σac(ω), or σac(f), at low frequencies.93 The temperature dependence of conductivity in Figure 13 follows, at each RH value, the VTF behavior, similar to eq 6 for the segmental α relaxation, this dependence indicating that proton conductivity is governed by the motion of the polymeric chains.93 In a recent work, it was shown that approaching Tg from high temperatures a decoupling of conductivity from the α process is observed, with the decoupling index R96 depending on hydration level and taking intermediate values, as compared to other ionically conducting polymeric systems.50 The inset to Figure 13 shows the same data as in the main figure in the Tg-scaled form,50 using Tg values determined by TSDC. Now the points for the various RH values come closely together, suggesting that the main contribution to the increase of σ with increasing RH arises from the concomitant acceleration of segmental dynamics rather than the increase of charge carrier (proton) concentration.77, 93 It is interesting to note in this connection that Al Kobaisi et al.97 reported identical hysteresis behavior of the βwet-relaxation and ionic conductivity in chitosan hydrogels with both degrees of chemical cross-linking and water contents and a power law correlation between the two processes.

Figure 13.

Arrhenius plot of dc conductivity, σ, in neat PHEA hydrogels at several values of RH indicated on the plot. The inset shows the Tg-scaled Arrhenius plot for the same compositions. (Reproduced from Ref. 50, with permission from [Elsevier Science]).

Figure 14 shows results similar to those of Figure 13, now for the PHEA/PEA 70/30 copolymer.72 The main interest here is the silica content dependence of conductivity. At each silica content of measurements, 0 and the highest one of 20%, which is above the percolation threshold for the formation of a continuous silica network, σ increases significantly with RH, whereas at each RH, σ is by about two orders of magnitude smaller in the nanocomposite hydrogel. This large deviation cannot be due to any differences in the corresponding water sorption isotherms (small anyway, compare Fig. 4), as it is observed also in the xerogels. In addition, σ becomes more Arrhenius-like in the nanocomposite xerogel and this behavior changes back to VTF-like (concave temperature dependence) with increasing RH. These and similar results have been discussed in terms of stiffening of the polymer matrix imposed by the interpenetrated silica network in the xerogels and then the relaxation of the structure in the hydrogels.72 It has been noted that the reduction of conductivity at high silica content resembles the reduction of water diffusion coefficient in PHEA/silica nanocomposites at high silica contents (Fig. 6), in terms of silica content dependence. Both processes, electrical conductivity and water diffusion, are characterized by macroscopic spatial scales and can be better explained by the formation of a continuous silica network rather than immobilization in an interfacial polymer layer around the nanoparticles, similar to swelling (Fig. 5) and the significant increase of elasticity modulus in ref.44. In contrast to that, the dipolar γ, βsw, and α relaxations are characterized by local spatial scale and, as a result, they are less affected by the formation of the continuous silica network.72

Figure 14.

Arrhenius plot of dc conductivity, σ, in the nanocomposite hydrogels with 70/30 PHEA-co-PEA matrix and 0 or 20 wt % silica at various levels of RH indicated on the plot. (Reproduced from Ref. 72, with permission from [Springer Science]).


In this review on hydration studies in polymer hydrogels, we attempted to illustrate the various concepts used in describing the hydration properties of this emerging class of materials in terms of organization, structure, and dynamics of water and of its impact on the properties of the polymeric matrix. To do that more effectively we focused on PHEA-based hydrogels, mostly neat PHEA, random copolymers with hydrophobic PEA, and silica nanocomposites. The combination of water sorption/diffusion measurements providing information on water uptake and water diffusion coefficients, DSC and TSDC for studying thermal transitions, in particular the glass transition, and broadband DRS for a detailed study of water and polymer dynamics with gradual variation of RH/water fraction of the hydrogel proved a powerful methodology. Of particular interest are correlations between results obtained by the various techniques used, such as the correlation between the critical water content for the completion of the first hydration layer determined from ESI measurements and the dielectric strength of the βsw relaxation, which provide deeper insight into the organization of water and the water and polymer dynamics in the hydrogel.

The methodology and the concepts described in this review for the investigation of the hydration properties of the PHEA-based hydrogels, as well as many of the results presented, apply also for the hydration properties of polymers in general. Moreover, the same methodology and concepts have been successfully used for the investigation of similar issues in related systems, namely, protein-water systems and mixtures of polymers with non-polar solvents, revealing interesting correlations between results obtained with the different systems. In the outlook of this review, we would like to present very briefly a couple of characteristic results from these areas of research, pointing to similarities and correlations with results obtained with hydrogels.

The hydration properties of proteins have been intensively studied, both from the materials science and the biological point of view. In the context of this article, the studies which use methodologies similar to those used for the investigation of hydration properties of synthetic polymers are of great interest. The transfer of accumulated experience and knowledge on such relatively simple systems to the studies of complex biological systems has been proved successful.98–101 For instance, Sartor et al.99 found that the calorimetric features of hydration water in metmyoglobin, methemoglobin, and lysozyme are similar to those of water imbibed in the pores of a synthetic hydrogel but very different from those of glass bulk water. It is worth pointing out that the aim of all these efforts is to describe and understand the biological functions in physical terms.

Concerning the dynamics of protein macromolecules, it is well established that it is strongly interrelated with water molecules dynamics. It has been suggested that protein dominant conformational motions are slaved by the hydration shell and the bulk solvent,102 that a symbiotic relation exist between the secondary relaxation of water molecules and the segmental relaxation of hydrated proteins100 or that the protein salvation shell exhibits a regular glass transition.103 To explore the specific features of this interrelationship various studies were recently performed, where the protein dynamics is studied in a broad hydration range spanning the range from (almost) dry samples (powders) up to protein solutions.68, 83, 97, 104–110 This approach allows the direct comparison between hydration studies in synthetic hydrogels and in proteins, revealing local relaxation processes at the protein surface and the secondary relaxation process of water molecules themselves, at low water fractions.

A key point for the investigation of macromolecular dynamics is the segmental relaxation process related with the thermal glass transition of the material. The thermal glass transition of the hydrated proteins has been the subject of many investigations and, in general, it has been found that it is much broader than the glass transitions in synthetic macromolecules.101, 111–116 Regarding the relaxation process associated to the thermal glass transition, there is a discrepancy in the literature concerning the corresponding dynamical responses detected by the various techniques and their assignment either to a glass transition or to a “dynamical transition.”103 Here, the term “dynamical transition” refers to transitions which are related to dynamical crossovers with respect to the temperature dependence of the relaxation times of macromolecular processes where water molecules are involved. More specifically, dielectric and NMR studies, methods sensitive to molecular reorientation, revealed the segmental relaxation process of hydrated proteins, linked to a dynamical glass transition around −110 to −80 °C for globular proteins,100, 103, 117 whereas quasielastic neutron scattering (QENS) experiments (mainly) revealed a dynamical transition of the mean square displacements of hydrated protein motion at about −50 °C connected with a fragile to strong transition for hydration water existed at the same temperature region.102, 118, 119. However, the existence of such a crossover for the hydration water has been questioned by other experimental studies on hydrated proteins.108, 109, 120–123

In a recent work, Panagopoulou et al.68 used the same combination of experimental techniques presented above for the hydrogels (water sorption/diffusion, DSC, TSDC, and DRS) to study glass transition and water and protein dynamics in mixtures of water and a globular protein, BSA, in extremely wide ranges of water content, both solutions and hydrated solid samples. Figure 15 presents characteristic TSDC and DSC results in terms of glass transition (Tg), melting (Tm), and crystallization temperatures (Tc) against water fraction hw. Two α processes have been identified by TSDC, corresponding to two glass transition temperatures, Tg1 and Tg2, the second one being related with the formation of a separate phase of water clusters containing initial nuclei of water crystals,68 against only one by DSC. The significant plasticization of Tg and the stabilization within the temperature range of −80 °C to −90 °C for water fractions where a part of the water crystallizes during cooling, is evident for both TSDC and DSC, and in agreement with various results reported in the literature for fully hydrated proteins.68, 83, 111, 114 Analysis of the data, including the observation that the DSC data are fitted very well by the Fox mixing formula,68 which does not take into account any interactions between its components, along with the fact that the glass transition seems to be absent in completely or nearly dry proteins, may lead to an important conclusion relevant to the origin of the glass transition in hydrated proteins. It is likely that bound water molecules on the primary hydration sites of the protein, being a structural component, are the condition for the existence of segmental dynamics in proteins. In other words, we can imagine the protein as a polymer matrix, and primary hydration water molecules as structural part of the latter. In that case, the plasticization of the glass transition is caused by the additional uncrystallized water of the hydration shell or of water molecules confined in other nanosized areas of the protein. This assumption is supported by the value estimated by the Fox equation for the glass transition temperature of water in the case of the DSC data (Tg,w = −165 °C), which is the glass transition of the uncrystallized water, that is, uncrystallized water of the primary hydration shell.68 Therefore, this study reveals, on one hand, the existence of the segmental relaxation process strongly related with the thermal glass transition and, on the other hand, the vital role of uncrystallized water molecules for the existence of the segmental relaxation process and its strong plasticization. It is obvious that water molecules in the primary hydration shell of the protein molecule play the crucial role for proteins conformation and probably for proteins functionality. Based on similar conclusions, many researchers are focused on the dynamical features of uncrystallized water molecules in hydrated proteins. The transfer of the theoretical understanding gained in the study of un-crystallized water in hydrogels and other water containing systems to the more complex systems of proteins may help towards a satisfactory explanation of protein's hydration.85, 99, 100, 124–128

Figure 15.

Characteristic DSC and TSDC temperatures, indicated on the plot, against water fraction hw obtained with BSA samples. The solid lines are fits of the Fox equation68 to the experimental data. The dotted lines are used to estimate the fraction of uncrystallized water according to the phase diagram.68 (Reproduced from Ref. 68, with permission from [Elsevier Science]).

In macroscopically homogeneous glass-forming polymer/solvent mixtures, the main topics of research, both experimental and theoretical, are similar to those in hydrogels, namely effects of solvent to polymer plasticization, polymer segmental dynamics, secondary polymer relaxations and characteristic features, such as the dynamical heterogeneity.129, 130 Of particular interest are partially crystallized mixtures, including as a prominent category the polymer/water mixtures, that is, polymer/solvent mixtures, where the polymer component does not spontaneously crystallize, whereas the thermodynamics allows the phase separation and crystallization of the solvent. Interestingly, DSC results indicated that the phase diagram of a polymer mixture with a non-polar solvent can be similar to those of hydrogels.131 Figure 16 shows results from a work in progress, the Arrhenius plot of the α process in cross-linked PEA swollen in p-xylene in wide ranges of composition, based on DRS and TDSC data. Dielectric studies on such systems do not suffer from solvent relaxations, dc conductivity and interfacial polarization masking the polymeric dipolar relaxations in hydrogels. In agreement with the phase diagram of the system, constructed on the basis of DSC and TSDC data,131 the mixtures can be classified into three qualitatively different groups, each one corresponding to different solvent content, cpx, region. More specifically, for p-xylene contents cpx < 0.20, the mixtures are homogeneous in the whole temperature range studied and the polymer is monotonously and strongly plasticized. For intermediate p-xylene contents, 0.20 ≤ cpx ≤ 0.30 the mixtures undergo phase separation at relatively low temperatures and exhibit crystallization (cold crystallization) only on heating at temperatures higher than the respective polymer glass transition temperature.131 The polymer component is still plasticized; however, the temperature dependence of the time scale becomes specifically intriguing reflecting the microphase separation. Cold crystallization was observed and studied also in hydrogels45 and in protein-water mixtures.68 In an attempt to better understand the mechanism behind this phenomenon and identify the fraction of solvent involved in that, Ide et al.45 studied in detail cold crystallization of water in hydrogels based on acrylates and methacrylates by combined DSC and FTIR techniques. It was concluded that cold crystallization might be generated by caging the water molecules in a small space by the polymer chain with a small hydration area and that cold crystallisable water is different than intermediate water and non-freezable water. For the highest p-xylene concentration region studied, 0.30 < cpx ≤ 0.85, the mixtures exhibit phase separation during cooling and a fraction of p-xylene crystallizes.131 In addition, the glass transition of the gel is increased and stabilized at a higher value (regulation effect26, 28, 131) corresponding to the reduced solvent content in polymer/solvent phase. In consistency with that, the time scale of the α process becomes independent of composition and, moreover, similar to the time scale of mixtures in the range of original p-xylene contents between cpx = 0.11 and 0.15 (Fig. 16). This result directly indicates that the p-xylene content in the non-crystallized phase (gel phase) is between 0.11 < cpx < 0.15, in agreement with the critical content for non-crystallized solvent estimated by DSC.131

Figure 16.

Arrhenius plot (in terms of peak frequencies, fmax, and mean relaxation times, τ, τ = 1/2πfmax) for the α process in PEA swollen in p-xylene at various solvent fractions indicated on the plot. The lines are fittings of the VTF eq 6 to the experimental data. Included in the plot are also TSDC data at the equivalent frequency fmax = 1.6 mHz, corresponding to τ = 100 s (dotted horizontal line).

The representative results in protein-water systems and in mixtures of polymers with non-polar solvents reported briefly in this section illustrate that the main topics of research are common in the three classes of systems, hydrogels, hydrated proteins, and mixtures of polymers with non-polar solvents, including among others solvent dynamics in relation to interactions and confinement and effects of solvent on polymer dynamics. Thus, much can be learned from comparative studies in these and in similar systems. For example, by proper choice of the polymer and the solvent, the strength of interaction between the two can be tuned and effects of interaction can be better resolved against effects of mixing. In fact, a significant progress in understanding the hydration properties of foods were achieved by adopting the “polymer concept,” that is, by learning from the more advanced area of hydration studies of synthetic polymers.132 An additional benefit of such comparative studies, this time from the experimental point of view, is that shortcomings of a specific technique in a particular system can be overcome by properly selecting a different system. For example, polymer dynamics can be much more clearly followed by DRS in mixtures with non-polar solvents than with water, as the dielectric response is in the first case free from solvent relaxations, dc conductivity, and interfacial polarization.131


P. Pissis acknowledges support by the Basic Research Program PEBE 2010 of NTUA.

Biographical Information

original image

Polycarpos Pissis received his Diploma in Physics in 1973 and his PhD in Physics in 1977 both by the University of Goettingen, Germany. He is Professor at the Physics Department, National Technical University of Athens—NTUA, Greece. He teaches several subjects, both at under-graduate and post-graduate level, in particular in the field of materials science. He is main coordinator/partner in international and national projects. He has published more than 230 journal papers, more than 80 papers in conference proceedings, and 10 book chapters.

Biographical Information

original image

Apostolos Kyritsis received his Diploma in Physics in 1988 and his PhD in Materials Science—Physics in 1995 both by the University of Athens, Greece. He has published more than 60 scientific papers, more than 20 papers in conference proceedings, and three book chapters. His scientific interests include dielectric, calorimetric, and vapor sorption studies in ionic crystals, ceramics, polymers and complex polymeric systems and structure-property relationships in polymers, biopolymers, nanocomposites. He is also strongly interested in the metrology aspects of experimental measurements.