ARTICLE pubs.acs.org/JPCB
Hydration of Gelatin Molecules in GlycerolWater Solvent and Phase Diagram of Gelatin Organogels Shilpa Sanwlani, Pradip Kumar, and H. B. Bohidar* Polymer and Biophysics Laboratory, School of Physical Sciences, Jawaharlal Nehru University, New Delhi 110 067, India ABSTRACT: We present a systematic investigation of hydration and gelation of the polypeptide gelatin in waterglycerol mixed solvent (glycerol solutions). Raman spectroscopy results indicated enhancement in water structure in glycerol solutions and the depletion of glycerol density close to hydration sheath of the protein molecule. Gelation concentration (cg) was observed to decrease from 1.92 to 1.15% (w/v) while the gelation temperature (Tg) was observed to increase from 31.4 to 40.7 °C with increase in glycerol concentration. Data on hand established the formation of organogels having interconnected networks, and the universal gelation mechanism could be described through an anomalous percolation model. The viscosity of sol diverged as η ∼ (1 cg/c)k as cg was approached from below (c < cg), while the elastic storage modulus grew as G0 ∼ (c/cg 1)t (for c > cg). It is important to note that values determined for critical exponents k and t were universal; that is, they did not depend on the microscopic details. The measured values were k = 0.38 ( 0.10 and t = 0.92 ( 0.17 whereas the percolation model predicts k = 0.71.3 and t = 1.9. Isothermal frequency sweep studies showed power-law dependence of gel storage modulus 0 00 (G0 ) and loss modulus (G00 ) on oscillation frequency ω given as G0 (ω) ∼ ωn and G00 (ω) ∼ ωn , and consistent with percolation model prediction it was found that n0 ≈ n00 ≈ δ ≈ 0.73 close to gelation concentration. We propose a unique 3D phase diagram for the gelatin organogels. Circular dichroism data revealed that the gelatin molecules retained their biological activity in these solvents. Thus, it is shown that the thermomechanical properties of these organogels could be systematically tuned and customized as per application requirement.
I. INTRODUCTION Polymers gels, and in particular biopolymer gels, continue to attract sustained attention from basic and applied researchers belonging to a variety of disciplines. This is due to the fact that these hydrated three-dimensional interconnected networks constitute a unique state of soft matter associated with a hierarchy of characteristic length and time scales.13 Gels have found applications in food, pharmaceutical, personal care, and cosmetic industries. Recently, smart and stimuli-responsive gels have been used in drug delivery applications for controlled and sustained release.46 Biopolymer gels have generated much interest in the recent past due to their biocompatibility and negligible level of cytotoxicity.711 Most of the work hitherto has been confined to the study and characterization of hydrogels. Organogels have received scant attention regardless of the promise these novel gels offer. There are a few studies devoted to biopolymeric organogels: (i) Duan and Liu11 described L-glutamate-based organogels that had interesting chiroptical property; (ii) organogels from a vitamin C based surfactant was made and characterized by Nostro et al.;12 (iii) Li et al. described the preparation of a reversible heat-set organogel of β-cyclodextrin in N,N-dimethylformamide;13 (iv) physicochemical characterization of poly(acrylic acid) organogels was done by Jones et al.;14 and (v) lecithin organogels were shown to be highly biocompatible and thermodynamically stable.15 These gels have r 2011 American Chemical Society
found use in topical drug delivery applications. Microemulsionbased organogels belong to a special category of soft matter; their special features are extensively reviewed by Atkins et al.16 George and Weiss17 have reviewed the organogels formed of low molecular weight gelators. Herein, we address the issue of formation, phase diagram, and complete characterization of the gelation phenomenon in a most versatile and commonly used polypeptide, gelatin. Gelatin is a denaturated product of collagen which is a supercoiled right-handed triple helix, and is the most abundantly available protein in mammals.18 Depending on the process of recovery, the gelatin molecules bear different physical characteristics. Type A gelatin is acid processed and has an isoelectric pH, pI ≈ 9, whereas the alkali processed type B gelatin has pI ≈ 5.18 We have used gelatin-B in this study. Gelatin is a polyampholyte molecule that makes the net charge on the molecule strongly dependent on pH. Light scattering measurements assigned the following dimensions (radius of gyration, Rg, and hydrodynamic radius, Rh) to the chains of gelatin-B, Rg = 34 ( 3 nm, Rh= 23 ( 3 nm. Thus, one can estimate the chain
Received: February 26, 2011 Revised: April 27, 2011 Published: May 12, 2011 7332
dx.doi.org/10.1021/jp201877d | J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B stiffness from the ratio19 Rh/Rg ≈ 0.67. This clearly attributes a fully flexible conformation to gelatin-B chain.20,21 Gelation of this biopolymer has been exhaustively investigated in studies in the past2224 that attribute a typical gelation concentration, cg ≈ 2% (w/v) and gelation temperature, Tg ≈ 30 °C. Gelatin gels are associated with three major drawbacks: low gelation/melting temperature, low gel strength, and high gelation concentration. Considering the fact that this particular biopolymer, in its gel form, has found applications in pharmacy, food, and cosmetic industry, etc., modifying its thermal and rheological properties to increase its Tg and reduce its cg is called for. This constitutes the primary objective of this work. In a related study, Ferrand et al.25 have used glycerol, a nonsolvent for gelatin, as a plasticizer to improve the elasticity of gelatin films.
II. MATERIALS AND METHODS Gelatin-B (bovine skin extract, 225 bloom) having molecular weight 50 KDa bought from SigmaAldrich, USA, was used to prepare the samples. This was used as received. The Bloom number indicates the amount of pressure, in grams, required to break the surface of a 6.67% (w/v) gel; higher Bloom number indicates higher gel strength. Triple deionized water from Organo Biotech Laboratories, India, was used to prepare the solutions. Seven solvents of water (100 x) glycerol (x) binary mixture were prepared with concentration of glycerol x (% v/v) varying from 0 to 60. The solutions were prepared by dissolving gelatin powder in the required glycerol solutions with concentrations varying from 0.1 to 3.5% (w/v) and these were allowed to homogenize by resorting to continuous stirring for 1 h at 50 °C. A small quantity of sodium azide (NaN3) was added to the samples to prevent bacterial contamination. These samples were allowed to cool slowly through Newtonian cooling and form gels. The gel samples appeared optically clear and transparent to the eye and did not contain air bubbles. These were stored in an airtight chamber at room temperature (20 °C) having relative humidity less than 45%. The gel samples were kept for a week to attain homogeneity. These were later used for rheology studies. Rheology experiments were performed using an AR-500 model stress controlled rheometer (T.A. Instruments, UK) with the objective to interrelate the stiffness and thermal stability of the networks in frequency and temperature sweep modes. The dynamic rheology of the gel samples was measured using parallel-plate geometry of radius 20 mm with a truncation gap of 500 μm with oscillatory stress value as 6.3 Pa. The oscillatory stress value was changed to 1 Pa for samples examined near gelation transition. DSC experiments were performed by using a DSC 4000 (PerkinElmer, USA) instrument. Here, the objective was to determine the melting temperature of the gels and to correlate the same with the results obtained from rheology. In a DSC experiment, typically 10 mg of the sample was taken on an aluminum pan and sealed. Temperature sweep was performed with the heating rate maintained at 10 °C/min. The measurement protocol was as specified by the manufacturer of the instrument. Viscosity values were measured using a vibro viscometer (model SV10, A and D Co., Japan). This instrument uses a matched pair of gold-plated flat electrodes. The mechanical vibrations (frequency ≈30 Hz) set in one of these propagate through the sample and is picked up by the other electrode. The viscoelastic properties of the sample are deduced from the response function through the software provided by the manufacturer. Circular dichroism experiments were carried out with
ARTICLE
an Applied Photophysics Chirascan instrument (USA) to estimate the degree of helicity. The zeta-potential measurements were performed on an electrophoresis instrument (model ZC-2000, Microtec, Japan). The solutions were diluted to 0.25% gelatin-B for all glycerol concentrations to know the surface charge of streaming particles. In the case of the interacting solutions, if one uses the zeta potential (Z) as an approximation of the surface potential j of a uniformly charged sphere, the theory gives26 Z j ¼ 4πðσ=εkÞ
ð1Þ
where σ is the surface charge density of the particle, and ε and k are the dielectric constant and DebyeH€uckel parameter of the solution, respectively. It has been shown to a very good approximation that the surface potential can be determined from the potential existing at the hydrodynamic slip plane, which is called the zeta potential. The relationship between the mobility (μ) and the zeta potential is Z = 4π(μη/ε). Next, μ can be written as μ = σ/ηk, where η is the viscosity of the solution. Because the polyelectrolytes are in random coil conformation, the quantitative application of eq 1 is not expected. Raman spectra of 0.25% (w/v) gelatin-B in all glycerol solutions was recorded on a FT-IR/Raman spectrometer attached with a microscope (Varian 7000 FT-Raman and Varian 600 UMA). We adopted Raman spectroscopy to investigate structure of water in glycerol and gelatin because vibrational spectra are very sensitive to the local molecular environment. Dynamic light scattering (DLS) experiments were performed at a scattering angle of θ = 90° and laser wavelength of λ = 632.8 nm on a 256-channel Photocor-FC (Photocor Inc., USA) that was operated in the multi-τ mode (logarithmically spaced channels). The goniometer was placed on a Newport (USA) vibration isolation table. The solutions were diluted to 0.25% (w/v) gelatin-B for all glycerol concentrations to know the apparent hydrodynamic radius (Rh) of the biopolymer molecule. The time scale spanned 8 decades, i.e., from 0.5 μs to 10 s. Samples were housed inside a thermostated bath, and the temperature was regulated by a PID temperature controller to an accuracy of (0.1 °C. In all the experiments, the difference between the measured and calculated baseline was not allowed to go beyond (0.1%. The data that showed excessive baseline difference were rejected. In this method, the system is physically probed over a length scale q1 where q = (4πn/λ)sin θ/2. The laser wavelength in the scattering medium is λ/n, where n is index of refraction. The diffusion coefficient is related to corresponding hydrodynamic radius through StokeEinstein relation given as D ¼ kB T=ð6πη0 Rh Þ
ð2Þ
where solvent viscosity is η0, kB is Boltzmann’s constant, and T is absolute temperature. Details of DLS are described elsewhere.27
III. RESULTS AND DISCUSSION 1. Hydration of Gelatin in Glycerol Solutions. It was considered imperative to characterize gelatin in glycerol solutions. Gelatin is soluble in water, but not in glycerol. Thus, the aqueous glycerol solution is a marginal solvent for this polypeptide. The measured effective hydrodynamic radius is plotted in Figure 1 as function of glycerol concentration, which indicated that the radius reduced by ≈50%, from its value 48 nm measured in water, as water concentration reduced to 40% v/v. This chain collapse is attributed to the change in the nature of the solvent from a good solvent to a 7333
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B
ARTICLE
Figure 1. Plot of apparent hydrodynamic radius (Rh, circle) and mean zeta potential (square) of gelatin molecules as function of glycerol concentration (gelatin concentration = 0.1% w/v) measured at 20 °C. Note the chain collapses gradually as the glycerol concentration is raised. Note the invariance of the potential with glycerol concentration. Solid line is guide to the eye.
Figure 3. Deconvolution Raman spectra of (A) water, (B) glycerol, and (C) gelatin (0.25% (w/v)) in 30% (v/v) glycerol solution in the 28003800 cm1 region of the spectrum where the signature OH stretching vibrations are located. Water and gelatin in 30%(v/v) glycerol solution spectrum could be resolved into three GaussianLorentzian peaks centered at 3200, 3310, and 3460 cm1 whereas glycerol showed a single peak at 3300 cm1. See text for details.
Figure 2. Raman spectra of solutions recorded at 20 °C (upper) glycerol solutions and (lower) gelatin in glycerol solutions. Note the considerable change in peak heights in the 28003800 cm1 region of the spectrum where the signature OH stretching vibrations are located. These peaks were deconvoluted to resolve the individual peaks; see Figure 3.
marginal one. The corresponding zeta potential data is also depicted in Figure 1 which reveals a mean zeta potential value ≈3 mV invariant of glycerol concentration. The pI of this protein is ≈5 and the glycerol solutions had pH ≈6.5 ( 0.5, invariant of glycerol concentration. Thus, it can be concluded that the protein chain carried an equal amount of protonated and deprotonated amino acids, which gave little net charge to these molecules. Further, this can be attributed to the fact that the glycerol molecules present in the hydration shell located close to the molecular surface did not interact with charged residues through electrostatic forces. Timasheff,28 in his pioneering work involving protein hydration in glycerolwater solvents, has concluded that glycerol is selectively excluded from the nonpolar domains of the protein surface. This causes the water and glycerol molecules to redistribute in the vicinity of protein surface in order to reduce unfavorable contacts. This exchange reaction at hydrobhobic loci is favorable to water molecules. Addition of glycerol raises the chemical potential of the protein, making the situation thermodynamically unfavorable, and as a result the protein molecule
tends to favor a folded state.28 We observed the shrinking of hydrodynamic radius (size) (see Figure 1) with increase in glycerol concentration which is in agreement with the Timasheff formalism. In fact, in this concentration region the dielectric constant of the medium is not sufficiently low to allow ion pair formation or charge neutralization of ionized residues, a conclusion proposed by Singer.29 Such a contention would imply invariance of zeta potential as function of glycerol concentration as shown in Figure 1. The samples were subjected to Raman studies in order to resolve the issue of hydration of the protein in glycerolwater environment. The Raman spectra are shown in Figure 2 for glycerolwater and for gelatinglycerolwater samples. Figure 3 shows the spectra obtained from water (A), glycerol (B), and gelatin in glycerol solutions, and their corresponding deconvoluted peaks, which are very revealing. The Raman peaks are seen to occur in two distinct frequency bands: 6002000 and 28003800 cm1. Interestingly, in Figure 2 for glycerolwater and for gelatinglycerolwater, the first frequency band exhibits identical peaks that could be superimposed exactly. These bands arise from various vibrational, twist, and bending modes associated with COH and CH2 groups present in glycerol.30 The differential part resides only in the frequency range 28003800 cm1 which needs to be discussed in depth. Figures 3 shows similar data and its deconvolution identifies the characteristic vibrational modes. These peaks are located at 3200, 3310, and 3460 cm1 which are identified as follows: 7334
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B
Figure 4. Variation of peak area of different Raman bands as function of glycerol concentration. Note that the area of the 3200 cm1 band (icelike water) grows with increase in glycerol concentration. The other two bands, representing OH stretching (3310 cm1) and liquidlike water (3460 cm1), become weaker. Symbol size represents typical error bar. Lines are guide to eye only. See text for details.
Figure 5. Comparison between the relative area under 3200 cm1 Raman band for glycerol solutions and gelatin in these solutions observed at 20 °C. Note that the presence of gelatin does not alter water structure significantly. This band represents structured water. Symbol size represents typical error bar. Lines are guide to eye only.
ARTICLE
Figure 6. Pictorial depiction of solvation of gelatin molecule in aqueous and glycerol solutions. In the glycerol solution, the hydration sheath mostly comprises amorphous water and there is depletion of glycerol density in the vicinity of biopolymer molecule. See text for details.
Figure 7. Variation of relative viscosity of gelatin solutions as function of glycerol concentration. The arrows indicate gelation concentration (cg); gels formed at lower gelatin concentration as glycerol concentration was raised. Lines are least-squares fitting to data points. The cg values are listed in Table 1.
blue-shifted and resides close to the 3200 cm1 band. We do see an increase in the peak area of this band as a function of increase in glycerol concentration which will be discussed later. It has been argued by Ratajska-Gadomska and Gadomski32 that the peak around 3300 cm1 owes its origin to H-bonds linking ethanol molecules; they recorded no such peak in water spectra. Our data is at variance with their results. We observed a strong peak around 3310 cm1 in water samples and its peak area or height was found to be independent of glycerol or gelatin concentration. Note that this OH band can overlap with the weak NH band (from gelatin molecule). We did not observe any distinctive signature of this in our data. Thus, it cannot be arising from H-bond formation of water with hydroxyl group of glycerol or amine group of gelatin. It can be attributed to the OH stretching of partially structured water molecules (liquidlike water). (iv) The peak around 3460 cm1 arises from poorly H-bonded water molecules (amorphous water). The water spectra could not be fitted to more than three Gaussian Lorentzian functions unlike what was reported by Gadomska and Gadomski earlier.32,33
(i) The peak around 3200 cm1 arises from in-phase vibration of OH stretching mode, structurally arranged water, which is sometimes referred to as icelike structures.31 (ii) The origin of the Raman peak around 3310 cm1 in the water spectra appears intriguing. For glycerol, a peak was seen close to 3300 cm1 identified as COH stretch mode (Figure 3B). Interestingly, in glycerol solutions such a band was not observed. Deconvolution of the raw spectra, to resolve the peaks, was done using Origin 6.1 software. For the peak at 3300 cm1, the chi-squared value obtained was 0.93 for glycerol spectra; and for the peak at 3310 cm1, the same was 0.99 for water, and 0.99 for gelatin B in 30% glycerol solution. Thus, the spectral analysis was robust which made us believe that the two peaks in question were resolvable though they were separated by a small Raman shift. (iii) It is possible that when the hydroxyl groups of glycerol form hydrogen bonds with water molecule, this peak gets 7335
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B
ARTICLE
Figure 8. Gelation temperature (Tg) deduced from the first derivative of the storage modulus for organogels formed in various glycerol solutions. The Tg values are listed in Table 1. Solid lines are guide to eye.
Figure 9. Gelation temperature (Tg) deduced from DSC experiments for organogels formed in various glycerol solutions. The Tg values determined from the onset of the endotherms are listed in Table 1.
A typical spectral deconvolution data is depicted in Figure 3. The relative contributions of the three aforesaid bands to the Raman spectra of water are shown in Figure 4 which indicated that the presence of structured water (3200 cm1) increased by almost 30% as glycerol concentration was increased from 0 to 60% (v/v). At the same time the presence of partially structured water (3310 cm1) and amorphous water (3460 cm1) decreased by about 15% each. This observation concluded that glycerol is a water structure enhancer, which has been reported earlier too.28 This can be explained from the fact that when one glycerol molecule replaces a water molecule, four hydrogen bond acceptors and one hydrogen bond donors are gained. This allows a propensity of hydrogen bonds to prevail leading to the rise in the amount of structured water strongly hydrogen bonded with glycerol. Raman spectra obtained from gelatin dissolved in various solutions are shown in Figure 2. These spectra were deconvoluted and the peak located at 3200 cm1 is of particular interest as this band corresponds to the symmetric stretching of OH bonds present in the structured water in glycerol solutions. Figure 5 depicts this clearly and it is unambiguously inferred that the peak area of the 3200 cm1 band remains invariant of the presence of protein. Such behavior indicates that the hydration of protein molecule does not alter the water structure significantly. The water structure in glycerol solution is mediated by the superior hydrogen-bonding capacity of glycerol molecule. Thus, the inference that water structures did not get altered owes its origin to the fact that very few glycerol molecules attach to the gelatin chain. The hydration sheath mostly comprised water molecules, leading to a depletion of glycerol density in the immediate vicinity of the biopolymer. The inability of glycerol molecules to penetrate the first hydration layer of protein molecules has been reported earlier too.37 This has been demonstrated in a cartoon in Figure 6. 2. Gelation Phase Diagram. Gelatin forms thermoreversible hydrogels where each chain is twisted in a left-handed helix conformation and three such helices supercoil together to form the right-handed triple helix.18,2123 In glycerol solutions, the gelation phase diagram and its pathway will be different. This was probed in a systematic manner. The relative viscosity, ηr = η/η0, where η is the solution and η0 is the solvent viscosity, of the gelling solution is plotted in Figure 7 for various glycerol concentrations. A clear break point in the plots indicated the gelation concentration. In the absence of glycerol, gelatin sol had cg = 1.92% (w/v) whereas 60%
glycerol solution yielded cg = 1.15% (w/v), a reduction of ≈60% which is significant. Thus, it appears that glycerol induces hydrogen bond formation between the constituent gelatin chains. The hot sols were allowed to cool to room temperature and gels were formed typically after 4 h. This was ascertained by tilting the tubes and observing nonflowing meniscus. Ischronal temperature sweep and DSC experiments were performed on these samples after a week. The measured storage moduli (G0 ) values are plotted (as temperature derivative, dG0 /dT) as function of temperature to establish the value of gel melting temperature, Tg. This, and the corresponding DSC data are presented in Figures 8 and 9. The onset of the endotherm in the DSC profile is identified as the gel melting temperature. Two observations were made from the data: (i) presence of glycerol enhanced Tg, implying stronger gels are formed in glycerol solutions, and (ii) the Tg values obtained from rheology and DSC were consistent (see Table 1). The low-frequency gel strength, G0, is plotted as function of glycerol concentration in Figure 10 that supports the above-mentioned observation. In fact, the data shown in Figure 10 could be least-squares fitted to G0 ¼ G00 þ Rcgly
ð3Þ
limω f 0 G0 ðωÞ ¼ G0
ð4Þ
where the glycerol concentration, cgly, is given in % (v/v). The fitting yielded the values G00 = 26 Pa and R = 1.30. The area under the endotherms in the DSC curves represent the enthalpy associated with melting of a gel sample. This was estimated, and it was found that the enthalpy remained almost invariant of glycerol concentration, implying that glycerol molecules did not participate in the interlinking of the helices. Regardless, the storage moduli data implied that stronger gels were formed at higher glycerol concentrations and the gel strength (G0) increased by more than four times as glycerol concentration was raised from 0 to 60% (v/v). The gelation concentration (cg) and temperature (Tg) data, on hand, allowed the construction of a 3D phase diagram which is shown in Figure 11. The values for various physical parameters are listed in Table 1. A pertinent question arises though. What is the effect of glycerol on the biological activity of this protein? In order to answer this, we performed circular dichroism studies on these samples. The circular dichroism data are shown in Figure 12 which exhibited no glycerol-induced conformational change/transition. 7336
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B
ARTICLE
Table 1. Comparison of Critical Exponents Obtained from Measurements Performed on Gelling Sola glycerol (% v/v)
cg (% w/v)
1
Tg (°C)
n0
2
Tg (°C)
n00
δ (calcd)
k
0
1.92
31.4
31.6
0.69
0.73
0.77
0.38
10
1.83
33.6
34.9
0.67
0.75
0.7 (T)
0.70 (P)
20
1.70
35.6
36.3
0.68
0.70
0.75 (Z)
30
1.61
36.7
37.6
0.69
0.72
1.30 (C)
40
1.48
38.7
39.3
0.73
0.76
50
1.28
39.7
40.8
0.75
0.69
60
1.15
40.7
42.3
0.74
0.73
t 0.92
a
Note that there is limited matching between the theoretical predictions of percolation theory and experimental results. The classical and percolation model exponent values were obtained from refs 41 and 45. P, Z, and C represent results from percolation, Zimm, and conductivity models. 1Tg and 2Tg represent gelation temperatures obtained from DSC and rheology studies.
Figure 10. Plot of low-frequency storage modulus (G0) as function of glycerol concentration for a fixed concentration of gelatin, 2% (w/v) sample. Note that stronger organogels were formed at higher glycerol concentration and the dependence was linear. Solid line is straight-line fitting of data to eq 3. See text for details. 37
Identical observations were made earlier by Kozak et al. They argued that protein hydration should not differ significantly as compared to that of water due to the fact that glycerol does not alter the activity of water considerably. Such an observation concludes that gelatin organogels are biologically active. 3. Gelation Mechanism. After establishing the phase diagram, it was felt imperative to base the experimental observations on a gelation model. Let us first peruse some experimental results. In general, the onset of gelation is characteristically signified by the rise in the values of relative viscosity and low-frequency storage modulus.13 The relative viscosity of the gelling sol is plotted as a function of (1 c/cg), for c < cg, in Figure 13 for all the samples and the data was fitted to the power-law function !k c ηr ∼ 1 ; ðc < cg Þ ð5Þ cg that yielded a master plot with mean k = 0.38 ( 0.09. Similarly, for c > cg, a plot of G0 versus (c/cg 1) is shown in Figure 14 which yielded another master plot, and the data was fitted to eq 6 which gave t = 0.92 ( 0.17. !t c G0 ∼ 1 ð6Þ cg
Figure 11. 3D phase diagram of gelatin organogel in glycerol solution. Note that the gelation concentration reduces, and gelation temperature rises, when glycerol concentration is increased in the solvent.
Thus, for |c cg| f 0 one observed universal scaling with characteristic exponents (k and t). This clearly implied, when the gelation transition was approached, the sol viscosity (ηr) and low-frequency shear modulus (G0) exhibited characteristic scaling exponents. Isothermal frequency sweep experiments were carried out to study the frequency-dependent behavior of storage and loss moduli (G0 (ω) and G00 (ω)) close to cg. The data (not shown) was fitted to 0
G0 ðωÞ ∼ ωn 00
G00 ðωÞ ∼ ωn ;
ðc cg Þ
ð7Þ ð8Þ
Figure 15 shows the dependence of n0 and n00 on glycerol concentration. It is interesting to observe that both the exponents remained invariant of glycerol concentration to a large extent. These values are listed in Table 1. Thus, the results on hand suggest strong indication of the existence of a universal gelation mechanism: exponents k, t, n0 , and n00 were found to be associated with values independent of solvent composition for our system. This information needs to be put in proper perspective. 7337
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B
Figure 12. Circular dichorism spectra of gelatin dissolved in glycerol solutions. Note the invariance of spectral profile with glycerol concentration, implying that the biological activity of the molecules remains unaltered.
Figure 13. Relative viscosity data fitted to (1 c/cg) using eq 5 that yielded power-law exponent, k = 0.38. Solid line is the fitted master curve to data points. See text for details.
It is well-known that gelatin forms thermoreversible and physical gel in hydrogen bond friendly environment.18 Though gelatin gels have been known and used extensively for many decades, the intricacies of its gelation kinetics remains elusive. This is due to the fact that the gelatin sol is highly polydisperse which permits several types of inter- and intramolecular interactions persisting over many concentration and time scales. Moreover, only the theoretical understanding of chemical gels can be considered to be mature. In physical gelation processes, parameters like monomer functionality, gel fraction, order parameters, etc. are poorly defined. However, attempts have been made to address this issue in the past. Busnel et al.38 discussed the gelation phenomenon based on a kinetic model using Monte Carlo calculations where random nucleation, propagation, and limited reversion of helical segments governed the gelation pathway. Diffusion-assisted association of gelatin molecules following Smoluchowski aggregation kinetics was proposed as an alternative model in another study.39 In a systematic study performed under controlled conditions, it was found that the gelation process did exhibit percolation type growth.40 This observation was combined with the results obtained from present
ARTICLE
Figure 14. Low-frequency storage moduli data fitted to (c/cg 1) using eq 6 that yielded power-law exponent, t = 0.92. Solid line is the fitted master curve to data points. See text for details.
Figure 15. Comparison of theoretical and experimentally determined 0 00 power-law exponents: G0 ∼ ωn and G00 ∼ ωn defined by eqs 7 and 8. The theoretical exponent, δ, is defined by eq 10. See text for details.
studies which indicated the possibility of applying percolation type model to map the gelation mechanism in this protein in glycerol solutions. Note that eqs 58 introduce characteristic critical exponents (k, t, n0 , and n00 ) with well-defined values. In the percolation model, their origin is soundly based on rigorous theoretical calculations and computer simulations data.4146 In fact, these equations predict the gelation mechanism to be consistent with percolation theory. For instance, close to the gelation transition, the in-phase storage modulus, G0 (ω), and out-of-phase dissipation modulus, G00 (ω), exhibit a dispersion relation given by43 G0 ðωÞ ∼ G00 ðωÞ ∼ ωδ ;
ðc cg Þ
ð9Þ
with k, t, and δ related through the hyperscaling expression δ¼
t kþt
41,45
ð10Þ
Theoretical models have shown that, in a 3D system, k = 0.7 and t = 2 for a conducting network, and k =1.3 and t = 3 for a percolating network with Rouse dynamics. Interestingly, the pairs k = 0.7 and t = 2 and k = 1.3 and t = 3 yield same value for δ ≈ 0.7. Both pairs of exponents have been observed for 7338
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B gelling systems.41 A summary of exponents determined from these studies is provided in Table 1. Several observations can be made from this data: (i) the measured viscosity exponent was k = 0.38 ( 0.10 which is almost half of the predicted value for systems undergoing gelation following Zimm dynamics or conducting percolation (k ≈ 0.7); (ii) the storage modulus exponent was t = 0.92 ( 0.18 as against the predicted value (t ≈ 1.9 for a percolating network); (iii) the hyperscaling exponent determined using eq 10 gave δ ≈ 0.73, not too different from the predicted value δ ≈ 0.7; and (iv) however, δ ≈ n0 and n00 was obtained in the case of some specific samples only. Djabourov40 had obtained t ≈ 1.9 for a 5% (w/v) gelatin hydrogel. Thus, the exponents do not confer consistence with a full percolation type model for the present system. At the same time, it cannot be overlooked that several of these exponents exhibited universal values independent of solvent composition. Since, the percolation-type scaling was universally observed, the gelation mechanism may be treated as anomalous percolation. The superiority of the percolation model over Flory-type treatment to describe gelation in polymer solutions is rather well-known.41 Gelation transition used to be described through the classical theory first conceptualized by FloryStockmayer.4146 Though this model accounted for the gelation phenomenon successfully in various systems, it relied heavily on the geometrical properties and inadequately incorporated any dynamics on its own. The critical growth of connectivity was directly interpreted in terms of percolation transition. The main drawback of this theory was that the model was not based on any periodic lattice structure. Thus, the predicted critical exponents were independent of space dimensionality and monomer functionality. As a result, hyperscaling cannot be applied to these critical exponents.45,46 Many of these shortcomings were addressed in the percolation theory which provides a statistical description of the phenomena of gelation. The detailed theory concludes that a percolation transition (gelation) is like a secondorder phase transition (similar to liquidgas phase transition at the critical point) with well-characterized universal behavior. Results obtained from the present study are in agreement with this conclusion though the critical exponents have different values compared to those predicted.
IV. CONCLUSION The gelation phenomenon of the biopolymer gelatin in glycerol solutions was probed systematically, and a phase diagram depicting the same was produced. The gelation mechanism, though it mimicked percolation-type network growth, did not yield exact critical exponents. Thus, we observed the formation and growth of an anomalous percolating network. Tosh and Marangoni47 observed percolation-type network formation close to the gelation point in gelling gelatin sols. The correlation length (ξ) of the gel network was found to follow ξ ∼ (c cg)υ. The values for υ ranged from 0.71 at 20 °C to 0.57 at 30 °C, decreasing linearly over this temperature range. These results were consistent with the theoretical limits.41 Thus, we conclude that percolation-type gelation of gelatin sol is observed in the glycerolwater system. As the glycerol concentration was raised, higher gel strength and gelation temperature and lower gelation concentration were observed, which implied preferential hydration of the network. A pertinent question arises: glycerol is excluded from the vicinity of gelatin molecules, hence, why does the gel strength increase with glycerol concentration? A possible explanation is as follows: it was observed from DSC measurements that the
ARTICLE
enthalpy remained almost invariant of glycerol concentration, implying that glycerol molecules did not participate in the interlinking of the protein molecules. It is the water that hydrates the protein. With increase in glycerol concentration, the amount of “available glycerol-free water” is reduced due to preferential hydrogen bonding between glycerol and water molecules. This reduces the amount of water available for binding to the protein that provides hydration. Thus, the gelatin molecules compartmentalize into regions of “available glycerol-free water” where the local gelatin concentration is increased considerably, resulting in the formation of stronger gels. We are further probing this behavior through anisotropic light measurements in order to confirm the presence of heterogeneity of these systems. The 3D phase diagram shown in this paper was never established earlier for gelatin system. In addition, we provide a robust and comprehensive percolation model description to explain the gelation kinetics of gelatin in an organic solvent. The Raman data supported the existence of water dominated hydration sheath around the gelatin molecules that had depleted glycerol presence. However, the fraction of structured water increased with increase in glycerol concentration and this remained invariant of the presence of gelatin. Such an observation confirms that glycerol plays the role of water structure enhancer. The biological activity of gelatin was found to be unaffected by the presence of glycerol as a co-solvent. Thus, in summary it can be concluded that the physical properties of gelatin organogels can be sufficiently and adequately tailored in glycerol solutions to customize its application potential.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Fax: þ91 11 2674 1837. Tel: þ91 11 2674 4637.
’ ACKNOWLEDGMENT We are thankful to the Advanced Research Instrumentation Facility of the University for allowing us access to the Raman and CD facility. This work was supported by a research grant received from the Department of Science and Technology, Government of India. ’ REFERENCES (1) Harland, R. S.; Prud’homme, R. K. Polyelectrolyte Gels; American Chemical Society: Washington, DC, 1992; Vol. 480. (2) Bohidar, H. B.; Dubin, P. L.; Osada, Y. Polymer Gels: Fundamentals and Applications; American Chemical Society: Washington, DC, 2003; Vol. 833. (3) Osada, Y.; Khokhlov, A. Polymer Gels Networks; Mercel Dekker: New York, 2001. (4) Lin, D. C.; Horkay, F. Soft Matter 2008, 4, 669–682. (5) Baldock, C.; De Deene, Y.; Ibbot, G.; Lepage, M.; McAuley, K. B.; Oldham, M.; Schreiner, L. J. Phys. Med. Biol. 2010, 55, R1–R63. (6) Pecharroman, C.; Bartolome, J. F.; Requena, J.; Moya, J. S.; Deville, S.; Chegalier, J.; Fantozzi, G.; Torrecillas, R. Adv. Mater. 2003, 15, 507–511. (7) Liu, J.; Lin, S.; Li, L.; Liu, E. Int. J. Pharm. 2005, 298, 117–125. (8) Sriniwas, P.; Kasapis, S.; Tongdang, T. Langmuir 2009, 25, 8763–8773. (9) Boral, S.; Saxena, A.; Bohidar, H. B. J Phys. Chem. B 2008, 112, 3625–3632. 7339
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
The Journal of Physical Chemistry B (10) Nostro, P. L.; Ramsch, R.; Fratini, E; Lagi, M.; Ridi, F.; Carretti, E.; Ambrosi, M.; Ninham, B. W.; Baglioni, P. J. Phys. Chem. B 2007, 111, 11714–11721. (11) Duan, P.; Liu, M. Langmuir 2009, 25, 8706–8713. (12) Nostro, P. L.; Ambrosi, M.; Ninham, B. W.; Baglioni, P. J. Phys. Chem. B 2009, 113, 8324–8331. (13) Li, Y.; Liu, J.; Du, G.; Yan, H.; Wang, H.; Zhang, H.; An, W.; Zhao, W.; Sun, T.; Xin, F.; Kong, L.; Li, Y.; Hao, A.; Hao, J. J. Phys. Chem. B 2010, 114, 10321–10326. (14) Jones, D. S.; Muldoon, B. C. O.; Woolfson, A. D.; Andrews, G. P.; Sanderson, F. W. Biomacromolecules 2008, 9, 624–634. (15) Kumar, R.; Katare, O. P. AAPS PharmSciTech 2005, 6, E298–E310. (16) Atkins, P. J.; Clark, D. C.; Howe, A. M.; Heenan, R. K.; Robinson, B. H. Prog. Colloid Polym. Sci. 1991, 84, 129–132. (17) George, M.; Weiss, R. G. Acc. Chem. Res. 2006, 39, 489–497. (18) Veis, A. Macromolecular Chemistry of Gelatin; Academic Press: New York, 1964. (19) Tiwari, A.; Bindal, S.; Bohidar, H. B. Biomacromolecules 2009, 10, 184–189.See also: Tanford, C. Physical Chemistry of Macromolecules; John Wiley and Sons: New York, 1961. (20) Jena, S. S.; Bohidar, H. B. J. Chem. Phys. 1994, 100, 6888–6895. (21) Pezron, I.; Djabourov, M.; Leblond, J. Polymer 1991, 32, 3201–3209. (22) Djabourov, M.; Leblond, J.; Papon, P. J. Phys. (Paris) 1988, 49, 319–332. (23) Djabourov, M.; Leblond, J.; Papon, P. J. Phys. (Paris) 1988, 49, 333–343. (24) Guo, L.; Colby, R. H.; Lusignan, C. P.; Whitesides, T. H. Macromolecules 2003, 36, 9999–10008. (25) Ferrand, M.; Djabourov, M.; Coppola, M. Macromol. Symp. 2008, 273, 56–65. (26) Ohshima, H. Adv. Colloid Interface Sci. 1995, 62, 189–235. (27) Bohidar, H. B. In Handbook of Polyelectrolytes; Nalwa, H. S., Kumar, S., Tripathy, S. K., Eds.; American Scientific Publishers: Valencia, CA, 2002; Vol. II, pp 117144. (28) Timasheff, S. N. Adv. Protein Chem. 1998, 51, 355–432. (29) Singer, S. J. Adv. Protein Chem. 1961, 17, 1–68. (30) Mudalige, A.; Pemberton, J. E. Vib. Spectrosc. 2007, 45, 27–35. (31) Carey, D. M.; Korenowski, G. M. J. Chem. Phys. 1998, 108, 2669–2775. (32) Gadomski, W.; Ratajska-Gadomska, B. Chem. Phys. Lett. 2004, 399, 471–474. (33) Ratajska-Gadomska, B.; Gadomski, W. J. Chem. Phys. 2010, 133, 234505–17. (34) Gekko, K.; Timasheff, S. N. Biochemistry 1981, 20, 4667–4676. (35) Adamenko, I. I.; Bulavin, L. A.; Ilyin, V.; Zelinsky, S. A.; Moroz, K. O. J. Mol. Liq. 2006, 127, 90–92. (36) Hayashi, Y.; Puzenko, A.; Feldman, Y. J. Phys. Chem. 2005, 110, 6031. (37) Kozak, J. J.; Knight, W. S.; Kauzmann, W. J. Chem. Phys. 1968, 48, 675. (38) Busnel, J. P.; Morris, E. R.; Ross-Murphy, S. B. Int. J. Biol. Macromol. 1989, 119–125. (39) Bohidar, H. B.; Jena, S. S. J. Chem. Phys. 1993, 98, 8970–8977. (40) Djabourov, M. Contemp. Phys. 1988, 29, 273–297. (41) Aharony, A.; Stauffer, D. Introduction to Percolation Theory; Taylor and Francis: London, 1994. (42) Durand, D.; Delsanti, M.; Adam, M.; Luck, J. M. Europhys. Lett. 1987, 3, 297–301. (43) Adam, M.; Delsanti, M.; Durand, D. Macromolecules 1985, 18, 2285–2290. (44) de Gennes, P. G. Scaling Concepts in Polymer Physics; Cornell University Press: Ithaca, NY, 1979. (45) Stauffer, D.; Coniglio, A.; Adams, A. Adv. Polym. Sci. 1982, 44, 103–158. (46) de Arcangelis, L.; Del Gado, E.; Coniglio, A. Eur. Phys. J. E. 2002, 9, 277–282.
ARTICLE
(47) Tosh, S. M.; Marangoni, A. G. Appl. Phys. Lett. 2004, 84, 4242–4244.
7340
dx.doi.org/10.1021/jp201877d |J. Phys. Chem. B 2011, 115, 7332–7340
