Biomacromolecules 2005, 6, 14-23
14
Bridging the Divide between the High- and Low-Solid Analyses in the Gelatin/K-Carrageenan Mixture Stefan Kasapis* and Insaf M. Al-Marhoobi Department of Food Science & Nutrition, College of Agricultural & Marine Sciences, Sultan Qaboos University, P.O. Box 34, Al-Khod 123, Sultanate of Oman Received May 4, 2004; Revised Manuscript Received October 1, 2004
Over the past few years, a considerable amount of work has been done in several laboratories on the measurement of structural properties of low-solid biopolymer mixtures or high-solid materials of a single biopolymer in the presence of co-solute. The main objective of this work has been to establish a correlation between the two types of systems and extend it to a binary mixture in a high-solid environment. In doing so, it employed well-characterized κ-carrageenan and gelatin samples in an aqueous preparation or in the presence of glucose syrup and sucrose. The phase behavior of the composite gel was ascertained using small-deformation dynamic oscillation, differential scanning calorimetry, and light microscopy. Experimental observations were built into polymer blending laws that argued for an explicit phase topology and distribution of solvent between the two networks. A working hypothesis was formulated and applied to high-solid mixtures thus identifying phase or state transitions in the time/temperature function. This led to the development of a mechanical glass transition temperature as the threshold of two distinct molecular processes governing the “rubber-to-glass” transformation. A stage was reached at which the predictions of the hypothesis were found to be in good agreement with the experimental development of viscoelasticity in the high-solid κ-carrageenan/gelatin mixture ranging from the rubbery plateau and the transition region to the glassy state. Introduction In many food and pharmaceutical applications, proteins are used in mixture with polysaccharides to provide the required structure, mouthfeel, processability, storage stability, etc.1-3 Applied science serving the needs of the industry has demonstrated how a mixture of two (or more) macromolecules can achieve synergisms and enhance performance beyond the utility of a single component.4 In food technology, for example, molecular interactions in binary systems can guide the development of low-fat dairy spreads, confections, and processed fish products.5-7 In pharmaceuticals, mixtures of biopolymers and co-solute have been used as drug/capsule matrixes to control the mobility transition temperatures of residual water below the glass transition temperature (Tg) and the specific interactions between the active compound and the glassy solid.8 It is an observation of ours that a dividing line has emerged, which is quite rigorous, with researchers in the structure-function relationships of macromolecules opting to address issues solely in high or lowsolid systems. The latter commonly refer to an aqueous environment containing nonstarch polysaccharides, gelatin or globular proteins, and starch hydrolysates typically up to 2, 10, and 20% solids, respectively. Mixing two macromolecules can result in no interactions, phase separation due to thermodynamic incompatibility, and “synergistic” interactions due to * To whom correspondence should be addressed. Fax: 00 968 513 418. E-mail:
[email protected]. Address (after Jan 1, 2005): Food Science and Technology Programme, Department of Chemistry, National University of Singapore, Singapore 117543, Singapore.
conformational or electrostatic compatibility.9,10 A cursory exploration of the recent literature on these mixtures using a scientific search engine downloads a large number of documents and there is no question that the subject has reached a certain degree of sophistication. In phase-separation phenomena, for example, physical theories from the realm of synthetic polymers have been utilized to rationalize the multitude of applications of binary systems. Several novel concepts have been developed including (i) the partition of solvent between the phases of two incompatible biopolymers at equilibrium;11 (ii) the deswelling of a continuous network, which possessed “permanent” cross-links, in the presence of a gradually structuring polymer;12 (iii) the effect of phase inversion on the textural properties of the mixture;13 and (iv) the use of blending models to relate structure to network morphology and composite topology.14 Today, further opportunities for the application of aspects of steric incompatibility to product development are intensively pursued following guidelines from our industrial partners whose moving spirit and finances make much of this research possible. High-solid systems are mainly biopolymer and sugar based formulations with moisture content between 5 and 30%. Most of the work has been carried out in partially frozen biomaterials utilizing the concepts of the glass transition temperature and the state diagram.15,16,51 Recently, fundamental investigations on high sugar/biopolymer mixtures that find application in confections revealed significant scope for innovation.17,18 At the moment, it seems that there is a gap between the voluminous literature on basic studies and a clear pathway for processing, preservation, and
10.1021/bm0400473 CCC: $30.25 © 2005 American Chemical Society Published on Web 12/03/2004
Biomacromolecules, Vol. 6, No. 1, 2005 15
Communications
innovation in high-solid products. State diagrams have been effective tools in mapping out the physical behavior of pure ingredients, but it is high time to be tested in more complex samples. In such systems, rationalization of physicochemical stability on the basis of a single glass transition curve as a function of a total (agglomerate) level of solids is questionable. Furthermore, one feels compelled to note that structural studies in high-solid systems have been carried out largely using thermal analysis, which is not the technique of choice in synthetic polymer research. A new concept of “network Tg” has been introduced to the literature and mechanical analysis in combination with valid application of the free volume theory should be utilized to complement the DSC results.19,20,51 It is our view that fundamental interpretation of processes in low-solid systems is well advanced and this understanding should be extended to resolving issues in the high-solid counterparts. The purpose of the present communication is to initiate a link between the theoretical frameworks used to support analysis in the two types of materials opting to work in the gelatin/κ-carrageenan mixture. Experimental Section Materials. The sample of κ-carrageenan was a gift from Hercules, Lille Skensved, Denmark (batch ×6960). 1H NMR analysis showed that ι-carrageenan-like segments (i.e., with a sulfate group at position 2 of the 3,6-anhydride residue) constitute about 8% of the polymer. An Amberlite IR-120 exchanging resin from BDH was used to prepare the polysaccharide in the potassium form. κ-Carrageenan in the potassium form was characterized with intrinsic viscosity measurements [η] at a constant ionic strength (0.01 M KCl) and at 40 °C yielding a [η] value of 10.5 ( 0.2 dL/g.52 The gelatin sample was prepared especially for research from Sanofi Bio-Industries, Baupte, Carentan, France. It was the first extract from a single batch of cowhide produced by alkaline hydrolysis of collagen (type B). Compared to the acidic process on pig skin, cow hides require the longer and more drastic lime treatment before extraction, as the skins are much older. In this case, the treatment may last several months and the long soak converts many of the basic side chains into acidic groups thus reducing considerably the resultant gelatin’s isoelectric point (pI ) 4.5 in our case). Table 1 reproduces analytical characteristics of the sample, which were determined by the manufacturer. The isoelectric point was measured by completely deionising the sample on a bed of ion-exchanged resin and then measuring the pH of the eluant. Gel permeation chromatography was used to identify the number average molecular weight (Mn) of the sample and the percent weight of 10 molecular mass classes. Pullulan with a number average molecular weight ranging from 6 to 788 kD was used as a standard. The Bloom value is proportional to the elastic modulus of the gelatin gel and it decreases with decreasing Mn. The glucose syrup used was a product of Cerestar, Vilvoorde, Belgium. The dextrose equivalent of the sample was 42 and gel permeable chromatography provided the relationship between degree of polymerization and surface
Table 1. Data on the Physicochemical Characterization of Our Gelatin Sample sample
gelatin
bloom (g)a isoelectric point (pI) % moisture (wwb) calcium (ppm) sulfate (%) chloride (%) phosphate (ppm) Mnb M. Wt. > 106 kD M. Wt. > 540 kD tetra + penta gamma beta alpha subunits 1 subunits 2 subunits 3 subunits 4
259 4.5 8.5 80 < 0.1 0.16 53 98 700 8.6 9.5 8.7 8.8 17.2 29.3 8.0 5.0 2.0 2.9
a Bloom is the weight in grams required to push a piston of strictly defined shape 4 mm into a gelatin gel matured for 16-18 h at 10 °C. b The alpha, beta and gamma fractions of gelatin are well characterized and monodisperse with characteristic masses. The latter two are respectively a dimer and a trimer of the alpha fraction. The tetra and penta are higher order but less well defined fractions. The low molecular weight side of the GPC spectrum is divided arbitrarily into four different fractions of subunits. The percent weight of the GPC spectrum in each of the 10 molecular mass classes is quoted.
area (%) of the spectrum of the material.21 The total level of solids was 83% and glucose syrup compositions in this work refer to dry solids. Sucrose was AnalaR grade from Sigma. Methods. Polysaccharide solutions at various concentrations were prepared by dissolving at 90 °C with stirring for 20 min. The temperature was dropped to 70 °C for addition of potassium chloride to the low-solid preparation. The hydration temperature of the protein did not exceed 70 °C and low-solid solutions were readily prepared within 15 min. In both cases, 50:50 mixtures of glucose syrup and sucrose were added to produce high-solid samples. Besides the single systems, binary mixtures of κ-carrageenan and gelatin were also made in the presence of potassium chloride and sucrose/ glucose syrup for the high-solid materials. Low amplitude oscillatory measurements of the real (G′; storage modulus) and imaginary (G′′; loss modulus) parts of the complex shear modulus (G* ) G′ + iG′′), and tan δ (G′′/G′) were performed with the Advanced Rheometrics Expansion System (ARES), which is a controlled strain rheometer (Rheometric Scientific, Piscataway, NJ). Lowsolid samples were loaded onto the preheated platen of the rheometer (70 °C), their exposed edges covered with a silicone fluid to minimize water loss, and cooled to 0 °C at a rate of 1 °C/min. This was followed by a frequency sweep from 0.1 to 100 rad/s at 0 °C and a heating scan at 1 °C/min to 90 °C. In high-solid samples, the experimental temperature range was extended from +70 to -65 °C thus accessing molecular motions, which cover the glassy state, the softening dispersion (glass transition region), “rubbery plateau” and the flow region. The scan rate was 1 °C/min, the frequency of oscillation for each isothermal profile was between 0.1 and 100 rad/s, and the applied strain varied from 0.00075 in
16
Biomacromolecules, Vol. 6, No. 1, 2005
the glassy state to 1% in the rubbery plateau to accommodate the huge changes in the measured stiffness of the sample. Differential scanning calorimetry measurements were performed on MDSC Q1000 with autosampler (TA Instruments Ltd, Leatherhead, UK). The instrument used a refrigerated cooling system to achieve temperatures down to 0 °C and a nitrogen DSC cell purge at 25 mL/min. Hermetic aluminum pans were used. The DSC heat flow was calibrated using a traceable indium standard (∆Hf ) 28.3 J g-1) and the heat capacity response using a sapphire standard. At the beginning of each experiment, gelatin was heated to 70 °C and κ-carrageenan to 95 °C to eliminate erroneous phenomena due to thermal history during sample preparation and loading. Then samples were cooled to 0 °C and heated to 90 °C at 1 °C/min to reproduce the rheological thermal regime. Samples of 7-10 mg were analyzed at (0.53 °C temperature amplitude of modulation and a 40 s period of modulation. The reference was an empty hermetically sealed aluminum DSC pan. Three runs were recorded, and the average of essentially overlapping traces is reported. For the microscopy work, samples were allowed to set up over 12 h at 5 °C. Pieces of dimension 10 × 10 × 0.5 mm were cut from the single and composite gels using a sharp scalpel. Unstained samples were placed on a microscope slide with the cover slip being lowered gently onto the gel surface, excluding any trapped air, and they were examined straight away. For the stained specimen, a drop of 0.05% w/w aqueous Sirius red (contains sulfonic acid groups that can react with basic groups of collagen) was placed on the upper surface of gelatin gels. Toluidine blue (a metachromatic dye that stains polysaccharides) was used for the κ-carrageenan samples. Ten minutes were allowed for staining to take place and then materials were processed like their unstained counterparts. Images were acquired on an Axioskop 2 plus microscope with transmitted-light brightfield and interference contrast facilities from Zeiss, Gottingen, Germany. Digital image processing was carried out with AxioVision software for microscopy, a digital camera and a Printpix Digital Photo Printer CX-400 from the Fuji Co., Japan.
Communications
Figure 1. Temperature variation of G′ for 1.5% [cool (O); heat (b)], 1.9% [heat (9)], 2.3% [heat (2)], 2.8% [heat ([)], and 3.5% [heat (+)] κ-carrageenan in the presence of 25 mM KCl (scan rate: 1 °C/ min; frequency: 1 rad/s; strain: 0.007%).
Figure 2. Temperature variation of G′ for 10% [cool (O); heat (b)], 15% [heat (9)], 20% [heat (2)], 25% [heat ([)], and 30% [heat (+)] gelatin gels (scan rate: 1 °C/min; frequency: 1 rad/s; strain: 1%).
Results and Discussion Single κ-Carrageenan and Gelatin Gels. The literature on the structural and thermal properties of κ-carrageenan and gelatin is extensive. Perhaps the best-known aspect of the gelation of the polysaccharide is the formation of double helical structures, with three disaccharides per turn of helix, which come together to form aggregates.22 Potassium ions are the main stabilizing counterions of the aggregates, and according to the “domain model”, they form an array between adjacent helices in the gel.23 The purpose of the present exercise is to provide a series of data for the follow up with the gelatin mixtures. Figure 1 illustrates the mechanical profiles of κ-carrageenan gels ranging in concentration from 1.5 to 3.5% in the presence of 25 mM KCl (scan rate: 1 °C/min). There is a ten-degree thermal hysteresis between the cooling and heating traces of the polysaccharide at 1.5% solids. All traces
are sharp, and the onset of network formation as a function of polymer concentration is confined within the temperature range of 40-46 °C (most of the date are not shown), whereas melting broadens from 50 to 65 °C in the gels under investigation (Figure 1). This extends the thermal hysteresis of the most concentrated preparation to nineteen degrees thus arguing that the onset of gelation is governed largely by the coil-to-helix transition, whereas network disintegration requires melting of the increasingly dense polymeric agglomerates. Gelatin, on the other hand, is a linear molecule whose gelation involves the formation of a triple helix, consisting of three left-handed helices (R-chains) wound around each other into a right-handed super-helix.24 Figure 2 reproduces the mechanical profiles of the protein at various solid contents as a function of temperature. The onset of gelation is confined to temperatures below 30 °C (most of the date
Communications
are not shown) leading to the formation of ordered junctions zones which constitute the structural knots of the network being interlinked with “soluble”, disordered regions. In contrast to the results recorded in Figure 1, thermal hysteresis in gelatin remains constant between six and seven degrees throughout the examined concentration range. This pattern of behaviour argues that helices, as opposed to aggregation, are the main drive of structure formation in gelatin. It has been proposed that gelation is initiated with a β-bend, bringing together two strands of the same molecule. Nuclei are formed when this “hairpin” structure collides with a third strand which at low concentrations is likely to be an intramolecular collision, whereas at higher concentrations, intermolecular collisions are more probable.25,26 A recent proposal by te Nijenhuis and Ross-Murphy that gelatin gels show elements of macroscopic flow under stress at long times without appearing to have an equilibrium modulus may cause some controversy.27,28 Recording of the structural properties in single polymer preparations of Figures 1 and 2 will serve as a quantitative reference for characterization of the interactions between the two components in low and high-solid blends. Experimental Observations on Low- and High-Solid κ-Carrageenan/Gelatin Mixtures. The conformational characteristics and ionic groups of a macromolecule, the presence of counterions, and the pH of the solution will determine the nature of physical interactions in a binary system. Our mixture has been designed to eliminate the development of a heterotypic structure where positive interactions dominate. This is due to the absence of conformationally compatible sites of gelatin and κ-carrageenan, the addition of potassium ions that neutralizes the sulfate groups thus negating possible electrostatic interactions between the two constituents, and the pH of 6.7 that renders both polymers primarily negatively charged. The experimental evidence of this section aims to provide a firm footing as to the nature of the topology of our polymeric composite. Figure 3 illustrates typical exothermic peaks for single and mixed preparations of our systems obtained at a scan rate of 1 °C/min. A relatively sharp peak with a maximum heat flow temperature (Tmax) of 38 °C denotes the cooperative conformational transition of the polysaccharide segments (Figure 3a). By comparison, development of the gelatin network is relatively gradual culminating at lower temperatures (Tmax is about 17.5 °C in Figure 3b). Upon mixing, the temperature sequence and the overall peak form for each transition was maintained (Figure 3c), a result which argues that the two components gel independently in the absence of direct interactions. In contrast, heterotypic associations usually manifest themselves by distorting the peaks of the individual gels and generating a new thermal event in the DSC spectrum.29,30 However, the temperature band of the κ-carrageenan gel in the mixture is shifted to 46 °C in accordance with the increased thermal stability of concentrated preparations of the polysaccharide in Figure 1. The last point will be further addressed next. To determine the two-dimensional arrangement of the blend, we used interference contrast and bright field microscopy depicted in Figure 4. Single gelatin preparations appear
Biomacromolecules, Vol. 6, No. 1, 2005 17
Figure 3. DSC exothermic peaks for (a) 1.5% κ-carrageenan (25 mM KCl), (b) 10% gelatin and c) a mixture of 1.5% κ-carrageenan (25 mM KCl) plus 10% gelatin (cooling rate: 1 °C/min).
as a featureless background, which undoubtedly constitutes the isotropic phase of a nonaggregated material (image is not shown), as argued on the basis of the rheological evidence in Figure 2. κ-Carrageenan gels on the other hand exhibit a heterogeneous structure with areas of low and high density
18
Biomacromolecules, Vol. 6, No. 1, 2005
Communications
Figure 5. Temperature variation of G′ for 1.5% κ-carrageenan (25 mM KCl) plus 10% gelatin in the presence of 0% [cool (O); heat (b)], 28.5% [cool (0); heat (9)], and 56.5% [cool (4); heat (2)] sugar, which is a 50:50 mixture of glucose syrup and sucrose (scan rate: 1 °C/ min; frequency: 1 rad/s; strain: 0.007%).
Figure 4. Bright field micrograph for 1.5% κ-carrageenan with 25 mM KCl added (a), and interference contrast micrograph for 1.5% κ-carrageenan (25 mM KCl) plus 10% gelatin (b), which were obtained at magnification of 200×.
of the histochemical dye (Figure 4a). Mixtures of the two components show clearly the protein as a flat area and the polysaccharide in an agglomerate form of considerable size (in the order of 5-10 µm in Figure 4b). Therefore, gelatin is excluded from the domain of the aggregated polysaccharide sequences thus leading to phase separation and the creation of concentrated phases in accordance with the increased thermal stability of the κ-carrageenan gel in Figure 3c. Although calorimetry and microscopy argued convincingly for the noninteractive nature of the two polymers, the phase behavior of the mixture is not entirely clear. We set about exploring this as a function of increasing solids content. Figure 5 shows the cooling and heating profiles of 1.5% κ-carrageenan (25 mM KCl added) with 10% gelatin obtained using dynamic oscillation at a scan rate of 1 °C/ min. Samples were cured within the range of 60-0 °C which unveiled a thermally stable transition followed at lower temperatures by a second wave of structure formation. Subsequent heating reproduces the bimodal profile with the prolonged melting trace resulting in thermal hysteresis. Considering the data of Figures 1-4, it is obvious that the gelation (tgel) and melting (tmel) temperatures of 42 and 57
°C relate to the κ-carrageenan network. Those of 22 (tgel) and 29 °C (tmel) in Figure 5 are identified with the phase transition of the protein. κ-Carrageenan gels first and should form a continuous network. Similar steps of gel formation are recorded in the presence of 28.5 and 56.5% sugar but the networks become more thermally stable. Samples were cooled to 0 °C and then heated to complete the curing cycle. In a second experiment, identical formulations were cooled to subzero temperatures thus monitoring the gradual development of additional molecular processes. Clearly, pictorial rheology of the lowsolid mixture can serve as an indicator of thermal events of the two constituents in the presence of cosolute (Figure 5). This concept will be taken up later on in Figures 8 and 12 while developing a predictive framework linking the structural properties of the two types of systems. Regarding the changes in elasticity at subzero temperatures, ice formation is observed in the 28.5% cosolute mixture at -16 °C by a mounting of the values of the storage modulus. Soon after that (-26 °C), increasing crystallinity makes the sample slippery, proper adhesion to the surface of the measuring geometry is lost, and the experiment is abandoned. At 56.5% sugar content, the dependence of elasticity on temperature is quite spectacular, and it develops unabated down to -54 °C in Figure 5. Recent work in single high-sugar/biopolymer mixtures has identified this kind of mechanical response as the transition zone between glasslike and “rubberlike” behavior.31,32 The molecular process will be discussed further in relation to the vitrification characteristics of our systems in Figures 9-12. At temperatures below -54 °C, some crystallinity was again evidenced by the sudden jump in the values of G′. Quantitative Analysis of the Phase Behavior in Lowand High-Solid κ-Carrageenan/Gelatin Mixtures. To make headway from pictorial rheology to a quantitative consideration of phase phenomena in a binary system, it is essential
Biomacromolecules, Vol. 6, No. 1, 2005 19
Communications
to determine the solvent (water) partition between the two polymers.33 The approach allows estimation of the effective concentration of a polymer within its phase, as opposed to the original composition in the formulation, and hence of the final “rigidity” of each structure in the mixture. This is achieved with a straightforward calculation that recasts findings into the relative amount of water held at each polymer phase using the so-called p factor11 p ) (SG/xG)/(SC/xC)
(1)
where SG and SC refer respectively to the amount of solvent in the phases of gelatin and κ-carrageenan (SG + SC ) 1), with xG and xC being their original (nominal) concentrations. A point that has been debated extensively is the requirement for complete phase separation in the gel since this is not, of course, the case in phase separated solutions.34 Nevertheless, there is evidence that this is a reasonable assumption, as polymer-entrapping gelation is minimized and extensive phase separation is allowed to develop by slowing down the rate of cooling of a binary solution from about 30 to 1 °C/ min.35 This has been demonstrated using microscopy and observing a 10-fold increase in the mechanical strength of the slow scanned mixture at which the chain segments of the protein and polysaccharide networks link effectively. The overall rigidity of the composite gel can be estimated by the phase volume (φ) and the storage modulus of each component. In synthetic polymer research, this has been addressed to a certain extent by the so-called “isostrain and isostress blending laws”, which for our system recast in the following numerical form14 GU ) φGGG + (1 - φG)GC
(2)
1/GL ) φG/GG + (1 - φG)/GC
(3)
For GG > GC and gelatin forming the supporting matrix, this blending gives an isostrain or upper bound behavior (GU in eq 2), whereas phase inversion in the system, with the gelatin being now the discontinuous filler, results in the socalled isostress or lower bound model (GL in eq 3). The approach found utility in the assessment of the effect of relative polymer concentration and conformational characteristics on the “ergonomics” of phase separation in several binary mixtures.36,37 In biomaterials, this is facilitated by the formation of a discontinuous phase in the form of spherical inclusions, the “rough” interfacial surfaces between the continuous matrix and these inclusions that enhance adhesion, and the relatively small strains of dynamic oscillation or transient testing employed in most of the studies. It is expedient to implement a computation that spans every possible distribution of solvent between the two phases yielding each time the corresponding effective polymer concentrations.38 These concentrations can be used in the linear relationships of Figure 6 to provide modulus estimates for each polymer in each own phase. Our modeling assumes a uniform distribution of potassium ions in the “water lattice”. Thermodynamic analysis argues that gelation in κ-carrageenan is the result of binding of specific cations on various sites of the helices leading to the formation of a “dynamic
Figure 6. Concentration-storage modulus relationships for κ-carrageenan and gelatin gels obtained at the end of the cooling runs (0 °C) in Figures 1 and 2, respectively.
Figure 7. Variation of calculated moduli (G′gel and G′car) and the resulting upper (G′U) and lower (G′L) bounds for the 1.5% κ-carrageenan (25 mM KCl) plus 10% gelatin as a function of the solvent fraction of the gelatin phase. The experimental value of the composite gel is shown as a dashed line parallel to the abscissa (modeling and experimental values refer to 0 °C).
shield” that reduces the charge density.39 Sulfate groups are not critical in this process with the density of the cation binding sites being equal for furcellaran and κ-carrageenan.40 The conditions of our experiment also support an electrostatic attraction between the negatively charged gelatin chains and the potassium ions. Furthermore, it has been documented that these levels of counterions have little effect on the mechanical gelling and melting characteristics of the protein.41 The modulus traces and the resulting upper and lower bounds calculated according to eqs 2 and 3 have been plotted against the solvent fraction in the gelatin phase in Figure 7. Calculations were made on the assumption that both com-
20
Biomacromolecules, Vol. 6, No. 1, 2005
ponents were restricted to their individual phases at the time of gelation. Partial phase separation becomes apparent earlier, i.e., upon mixing the two clear solutions, which yields a translucent mixture. Steric exclusion phenomena were also inferred from the discussion of Figures 1-5. For example, the classic definition of a noninteractive, interpenetrating network would assume a final modulus for the mixture equal to the sum of the contributions of the single polysaccharide and protein gels (φ ) 1 for each constituent).42 In the case of 1.5% κ-carrageenan and 10% gelatin, that would be equal to 26.7 kPa (Figures 1 & 2), whereas the actual value of the mixture at 0% cosolute in Figure 5 amounts to 86.2 kPa. In Figure 7, the curve corresponding to the κ-carrageenan continuous network runs from bottom-left to top-right of the diagram and vice-versa for the gelatin continuous system. The respective shear moduli for the protein and polysaccharide phases that determine the overall rigidity of the composite are also shown. The experimental network strength of our mixture remains well below the values of the upper bound and intersects the traces calculated for the isostress arrangement. This corresponds to gelatin or κ-carrageenan continuous blends but the experimental evidence discussed in Figure 5, for example, demonstrated that the polysaccharide forms a continuous phase. Based on this outcome, we considered a single value of solvent distribution between the two phases from the κ-carrageenan continuous trace (Sgel ) 0.29 and Scar ) 0.71). This type of quantitative analysis argues that the polysaccharide component captures higher levels of water, a pattern of behaviour which is supported by previous explorations of binary blends. Thus, it was proposed that the continuous phase topology of a network, as opposed to that of a filler, and its balanced cross-linking morphology, as opposed to a precipitated gel, are essential determinants which allow a favorable redistribution of solvent in the polymeric phase.36,37 The final (effective) concentrations of the gelatin and κ-carrageenan phases become 28 and 2.33%, respectively. Our water partition analysis in Figure 7 argues that at these solid levels the protein network should be stronger than the polysaccharide. This was confirmed and extended to highsolid materials considering that their structural characteristics in terms of phase separation, temperature band, and bimodal mechanical or thermal profile originate from the aqueous systems (Figures 1-5). The cooling profile for a sample of 2.33% κ-carrageenan made in the presence of 35.34% glucose syrup and 35.33% sucrose is illustrated in Figure 8. It exhibits a rubbery plateau, followed by the glass transition region (G′′ > G′ from -16 to about -40 °C), and the glassy state at the lower range of experimental temperatures. Next, the gelatin mixture was tailor-made containing 28% of the protein with 22.5% glucose syrup and 22.5% sucrose at 73% total solids, thus affording a direct comparison with the κ-carrageenan sample. As predicted, the high protein content produces a strong network compared to the κ-carrageenan gel at the rubbery plateau (temperatures above -10 °C in Figure 8). However, the magnitude of the glass transition region of the protein in terms of temperature range (G′′ > G′ from -27 to about -40 °C) and modulus increase is limited in relation to the
Communications
Figure 8. Temperature variation of shear modulus for a mixture of 28% gelatin with 22.5% glucose syrup and 22.5% sucrose [G′ (b), G′′ (O)], and a mixture of 2.33% κ-carrageenan with 35.34% glucose syrup and 35.33% sucrose (25 mM KCl added) [G′ (2), G′′ (4)] using the experimental conditions of Figure 5.
polysaccharide trace. Both patterns of vitrification converge at about -40 °C, a result which makes small additions of κ-carrageenan to the sugar mixture an efficient accelerator of vitrification phenomena; a more than 10-fold increase in gelatin concentration is required for a similar outcome in Figure 8. Direct determination of phase composition is an obvious major target for future research on biopolymer blends. The present work, however, demonstrated that a sufficient protocol could be developed to provide a “safe passage” through the low-solid co-gels to high-solid materials. Further applications of this reasoning to biopolymer/cosolute samples will be discussed next. Rationalization of Structural Properties in High-Solid κ-Carrageenan and Gelatin Gels. Our work dealt with highsolid materials that exhibit aspects of temperature-induced vitrification phenomena. Determination of the glass transition temperature constitutes the first step of quantifying these phenomena. In synthetic polymer research, this is accomplished through the dissociation of the contributions of time and temperature to the overall mechanical response in the form of a basic function of time alone and a basic function of temperature alone.43,44 The latter has been discussed in the temperature profiles obtained at a fixed frequency of 1 rad/s in Figures 5 and 8. To examine the effect of the former, there is a need to go beyond the normal experimental procedure that covers a limited three to four decades of frequency () 1/time). To achieve a wider frequency window, the preparation of 2.33% κ-carrageenan (25 mM KCl) with 35.34% glucose syrup and 35.33% sucrose was scanned between +30 and -65 °C at a rate of 1 °C/min thus obtaining frequency sweeps from 0.1 to 100 rad/s at constant temperature intervals of three degrees centigrade. Figure 9 reproduces the mechanical spectra for the storage and loss modulus that cover almost 6 orders of magnitude of stress (from 104.5 to ∼ 1010 Pa). Then, a reference
Communications
Biomacromolecules, Vol. 6, No. 1, 2005 21
Figure 10. Composite curve of reduced storage (G′p; b, 2) and loss (G′′p; O, 4) modulus for a mixture of 2.33% κ-carrageenan with 35.34% glucose syrup and 35.33% sucrose (25 mM KCl added). The reference temperature is -29 °C.
Figure 9. Real (a) and imaginary (b) parts of the complex shear modulus, plotted logarithmically against frequency for a sample of 2.33% κ-carrageenan with 35.34% glucose syrup and 35.33% sucrose (25 mM KCl added). Bottom curve is taken at 30 °C (9); other curves successively upward, -13 (0), -17 (b), -21 (O), -25 (2), -29 (4), -33 ([), -37 (]), -41 (-), -45 (×), -49 (+), -53 (/), -57 (-), -61 (0), -65 (O) °C, respectively.
temperature was chosen arbitrarily (-29 °C) within the glass transition region and the remaining mechanical spectra were shifted horizontally along the log frequency axis until they fell into a single curve. Figure 10 reproduces a rather spectacular transition zone between glasslike and rubberlike consistency for the reduced storage (G′p) and loss (G′′p) modulus. Above 105 rad/s, the reduced frequency range corresponds to the glassy zone; G′p is quite high, around 109.5 Pa and does not change much with frequency. The transition zone makes its appearance between 10° and 105 rad/s with the viscous component dominating over the elastic component of the network. Below 10° rad/s, the behavior appears to correspond to the rubbery region; development of both moduli is characteristic of a very soft rubberlike solid.
The scheme for construction of the composite curve in Figure 10 by empirical shifts of data is known in synthetics as the method of reduced variables or the time-temperature superposition principle.45 Its applicability requires exact matching of the shapes of experimental curves, an outcome which would argue that during a change in state all relaxation times of a molecular process depend identically on temperature. Vitrification is in the nature of a second order thermodynamic transition,46 and it appears that the superposition is successful in the κ-carrageenan/cosolute mixture while it undergoes a change in state. However, it is not possible to match empirically adjacent curves of loss modulus in the rubbery and flow zones where the time-temperature superposition failed in Figure 10. The reason for the lack of reduction should be attributable to the gradual dissociation of the junction zones of the polysaccharide with increasing temperature. A reduction failure in the rubbery plateau due to changing entanglement concentration with a change in temperature was observed in synthetics for poly (n-octyl methacrylate).47 The effect of changing the temperature in Figure 9 is to shift the frequency scale of the mechanical spectra in the manner discussed in Figure 10 thus generating a shift factor (aT), which is a fundamental descriptor of the temperature/ time dependence of viscoelastic functions. The utility of the factor aT can be demonstrated by plotting its progression versus the temperature range that covers the process of vitrification in Figure 11. Clearly, there is a change of pace in the logarithmic development of factor aT as a function of temperature, which occurs at about - 39 °C. This temperature is in the vicinity of the conjunction of the glass transition region and the glassy state in the cooling profile of Figure 8. Its significance is bolstered further by quantifying the kinetics of vitrification using the combined framework
22
Biomacromolecules, Vol. 6, No. 1, 2005
Communications
Figure 11. Temperature variation of the factor aT within the glass transition region and the glassy state (b) for a mixture of 2.33% κ-carrageenan with 35.34% glucose syrup and 35.33% sucrose (25 mM KCl added), with the solid lines reflecting the WLF and modified Arrhenius fits of the shift factors in the glass transition region and the glassy state, respectively (dashed line pinpoints the Tg prediction).
of the theory of free volume and the Williams, Landel, and Ferry equation (WLF)48 log aT ) -
C01(T - T0) C02 + T - T0
(4)
where, T0 is the reference temperature and the parameters C01 and C02 relate to the free volume theory. The framework has been used extensively to interpret glassy phenomena in terms of molecular processes,49 although the recently introduced ‘coupling model’ appears to be of promise in the synthetic polymer research.50 As shown in Figure 11, treatment of the present system with the WLF equation provides a good fit of the experimental shift factors in the glass transition region thus making free volume the overriding mechanism behind molecular mobility. However, progress of the mechanical properties in the glassy state (
