Revue Roumaine de Chimie, 009, (-), 98 986 Dedicated to the memory of Professor Ecaterina Ciorănescu-Nenitzescu (909 000) INTRINSIC VISCOSITY OF AQUEOUS POLYVINYL ALCOOL SOLUTIONS Irina-Elena LĂMĂTIC, Maria BERCEA * and Simona MORARIU Petru Poni Institute of Macromolecular Chemistry, -A Grigore Ghica Voda Alley, 70087 Iaşi, Roumania Received May 9, 009 The behavior of polyvinyl alcohol aqueous solutions as investigated by viscometry. In region of very lo polymer concentration, the reduced viscosity as a function of concentration deviates from linearity in the uggins plots. The analysis of the concentration dependence of polymer solutions in terms of ln( η ) vs. concentration provides the possibility to obtain accurately the intrinsic viscosity values. The ationship beteen the intrinsic viscosity and the molecular eight for polyvinyl alcohol in ater at 0 C as examined and discussed. INTRODUCTION A parameter of great importance for polymer characterization is the intrinsic viscosity, [η], hich depends on the concentration and size of the dissolved macromolecules, as ell as on the solvent quality and temperature. Usually, it is evaluated by extrapolation of experimental data to zero polymer concentration or, for fast evaluations, from measurements performed at a single concentration. Deite a considerable number of theoretical and experimental studies (as for example several studies reported during last ten years - ), the current understanding of the viscosity behavior of neutral polymer solutions at lo concentrations is still far from being complete. Systematic errors appear in the evaluation of the intrinsic viscosity, being originated from the improper technique by hich the viscosity of polymer solution is determined at very lo concentrations. In some cases, belo a definite concentration, the curves of reduced viscosity plotted against concentration sho either an upard or a donard turn as a function of concentration. This can be due to the adsorption of the polymer chains on the glass all of the viscometer, but also to an expansion of the individual coils, conformational changes or others effects. Yang et al. 7,8 considered that the anomalous viscosity behavior of neutral polymer solutions at lo concentrations can be avoided if the flo time of polymer solution at zero concentration is considered to determine the intrinsic viscosity instead of the measured flo time of the pure solvent. This is not valid for all polymer-solvent systems, our previous studies on polyacrylonitrile solutions in dimethylformamide, shoed that the turn of the η / c curves could be due in part to the adsorption of polymer chains on the glass all of the viscometer, but also in part to an expansion of the individual coils or conformational changes of macromolecules in solution. The adsorption of the polymer onto capillary all surface usually occurs in the glass viscometer and determines a reduction of the capillary diameter, as ell as a loss of polymer and a change of the equilibrium concentration in the viscometer. Many efforts are continuously carried out in order to obtain a iable equation for determining the intrinsic viscosity by using the viscometric data Correonding author: bercea@icmpp.ro
98 Irina-Elena Lămătic et al. obtained at different concentrations. Recently, - a ne phenomenological approach as proposed as alternative method for the determination of the intrinsic viscosity for the polyelectrolytes, in presence and in the absence of salts. The value of the intrinsic viscosity is determined from the initial slope of the dependence of ln η (here η is the ative viscosity) as a function of concentration at sufficiently lo shear rates and polymer concentrations. This method as also successfully applied to polyacrylonitrile solutions in dimethylformamide. 6 In this study, e apply this concept to aqueous solutions of polyvinyl alcohol (PVA) ith different molecular eights in order to calculate the intrinsic viscosity values and then to determine the viscosity molecular eight ationship. RESULTS AND DISCUSSION Intrinsic viscosity of PVA solutions The intrinsic viscosity, [η], can be evaluated from the viscosity measurements hich provide data at finite concentration. The most general ationship beteen the intrinsic viscosity and dilute solution viscosity takes the form of poer series in concentration (c): η = c The [η] values for a polymer solution can be obtained by using extrapolation methods to infinite dilution, such as uggins equation: η c = [ η] + k [ η] c + k [ η] c + k [ η] c + [ η] + k [ η] c () k is referred to as the uggins dimensionless constant and ates to the size and shape of polymer segments, as ell as to hydrodynamic interactions beteen different segments of the same polymer chain; η represents the ecific viscosity ( η = η ) hich is the fractional change in viscosity produced by adding the solute and η is the ative viscosity shoing the change in viscosity, usually expressed as a ratio beteen the viscosity of the polymer solution and the viscosity of the pure solvent. Table () Figure presents the viscometric dependences according to the method of uggins obtained for PVA samples (Table ) in ater at 0 C. The values of [η] and k obtained for PVA samples in the dilute regime (for hich. < η <.9) are given in Table. It can be observed that k is higher than 0. for lo molecular eight sample. This can be due to stronger polymer-polymer interactions in competition ith polymer-solvent interactions. Also, in the region of lo polymer concentrations (for hich the values of the ative viscosity are less than.) the reduced viscosity ( η / c ) deviates from the linear dependence given by equation () and this made more difficult the correct evaluation of the intrinsic viscosity through the extrapolation to zero polymer concentration. Weight average molecular eights of PVA samples and the viscometric data obtained in ater at 0 C, according to different approaches Sample M x 0 - (g/mol) 0.9.0 6.0 8.9 6.60 % hydrolized 80 [η] 0.8 0.9 0. 0.7. k eq. () 0.688 0.70 0.7 0.8 0. [η] W 0. 0.98 0.9 0.78.8 B * 0.6 0. 0.7 0.7 [η] 0. 0.6 0. 0. 0.76 k eq. () 0.78 0.6 0.6 0.8 0.
Intrinsic viscosity 98 0. η /c 0. 0. η /c = 9c + 0.8 R = 0.99 k = 0.688 0. 0 0. 0. 0.6 0.8.0 η /c.0 0 0. 0. 0.6 0.8 Fig. Dependence of the reduced viscosity as a function of concentration for solutions of PVA (samples -, Table ) in ater at 0 C. The insert shos an enlargement of the plot obtained for sample. Upard changes in the slope of η / c as a function of concentration ere reported for polyelectrolytes, due to the electrostatic repulsion beteen the charged groups of the macromolecules strongly increases upon dilution. 7 Moreover, experimental results have shoed that the dependences of η / c versus polymer concentration reach a maximum, hich in very dilute solutions is folloed by a rapid decrease of the reduced viscosity.,7 We depicted a such maximum for sample (clearer evidenced in the insert from Figure ), hich can be explained as being due to the fact that in this point the polymer reach probably its final state of expansion in a finite domain of concentration. The reduced viscosity further decreases ith the dilution. For sample the maximum is less evidenced and for higher molecular eight samples probably this maximum can be reached for loer polymer concentrations and it is very difficult to find this experimentally. Recently, - a ne alternative method as proposed for the determination of the intrinsic viscosity of polyelectrolytes in absence of salts, according to the folloing equation: c ln η = [ η ] + Bc [ η] [ η] + Bc[ η] () here B represents a system ecific constant and [ η] is the characteristic ecific hydrodynamic volume. We applied this method to neutral solutions, i.e., polyacrylonitrile in dimethylformamide, 6 and the results are close to those determined by applying the uggins equation to the experimental data obtained in the dilute regime. Figure gives the concentration dependence of lnη for PVA samples in ater at 0 C. By modeling the viscometric dependences in terms of lnη vs. concentration, the values of the intrinsic
98 Irina-Elena Lămătic et al. viscosity ([η] ) ere evaluated, as ell as the values of parameters B and [ η] (Table ). It can be observed that the values of the intrinsic viscosity determined by applying the equation () to the experimental data for hich. < η <.9 are closed to those determined from the initial slope (eq. ) (Table ). In order to discuss the hydrodynamic interactions, Wolf et al. deduced the folloing interation for the uggins constant: k = B () hich only holds true for the range of pair interactions beteen the solute.,8 The particular situation of B = 0 consequently translates to k = 0. (eak polymer-solvent interactions). 0. ln(η ) 0. 0. 0 0. 0.. ln(η ) 0.8 0. 0 0. 0. 0.6 0.8 η Fig. Plot of ln( ) as a function of concentration for PVA samples in ater at 0 C. The insert shos an enlargement of the plot for very lo PVA concentrations. The values of k calculated ith the equation () are given in Table. The k value is close to 0. for sample, shoing unfavorable polymersolvent interactions for a sample 80% hydrolyzed, hereas for samples, having high degree of hydrolysis (98 99%) k presents values close to 0., suggesting favorable polymer-solvent interactions. Also, these values can be considered more realistic that those determined according to the equation of uggins, because the last dependence becomes non-linear at very lo polymer concentrations and the extrapolation procedure becomes inapplicable.
Intrinsic viscosity 98 Mark-oink ationship The dependence of the intrinsic viscosity on the molecular eight (double logarithmic plot) for PVA in ater at 0 C is shon in Figure and leads to the equation: [ ].8x0 M 0.6 η = () 0. log[η] -0. -0.8 log M 6 Fig. [η] - M double logarithmic plot for PVA samples in ater at 0 C. For syndiotactic-rich PVA samples ith molecular eight in the range. 0 g/mol. 0 g/mol, literature 9 gives the folloing dependence: [ ] 0.7x0 M 0.6 η = () The differences beteen our data and literature data can be due to the sample characteristics or the method used for the determination of [η]. Aqueous solutions of syndiotactic-rich PVA samples are very unstable; i.e., belo 80 C they become turbid, can form easily aggregates or gels. PVA is knon to dissolve more or less easily in ater, due to the degree of hydrolysis of the poly(vinylacetate) precursor polymer. At high degree of hydrolysis (> 98%), PVA is highly crystalline and forms strong interchain hydrogen bonds, affecting its solubility in ater. 0 Thus, a solubilization temperature higher than 80 C is required to dissolve PVA completely. Upon cooling at 0 C, PVA remains in solution, but its storage time at this temperature does not exceed days to avoid the formation of aggregates. The effect of the association is influenced by the flo-induced crystallization of PVA solutions and by the thermal history in the preparation of the solutions. 9, 0 Such solution does not form fibrous precipitate in capillary viscometer if the flo is made at sufficiently lo shear rate. In our measurements, for a given sample, changes of the solution viscosity in time ere found negligible and all samples remained homogeneous during the experimental investigations. Rheological measurements carried out at 0 C did not evidence floinduced crystallization of PVA solutions. EXPERIMENTAL PART Four polyvinyl alcohol samples (denoted,, and, Table ) ere purchased from Aldrich and a PVA sample (sample, Table ) as provided from Loba Chemie. omogeneous PVA solutions ere prepared by heating the PVA and ater mixture at 80 C, aiting days for reaching the thermodynamic equilibrium. The viscometric measurements ere carried out at 0 C ith an Ubbelohde suended-level viscometer. The flo time for the solvent as 6.7 s and the kinetic energy corrections ere found to be negligible. The flo volume of the viscometer as greater then ml, making drainage errors unimportant. In order to avoid any aging effects of PVA solutions, flo time measurements have been made on fresh solutions in less than 8 h. All the solutions and solvents ere filtered through 0. µm pore size membranes before being introduced in the viscometer. The solutions ere prepared starting from an initial stock solution successively diluted in the viscometer. Each run time as repeated five times to give flo time measurements ith an accuracy of s.
986 Irina-Elena Lămătic et al. CONCLUSIONS In aqueous solutions, there are to possible intermolecular interactions among PVA chains or beteen PVA and ater hich can determine the association or aggregation: -bonding beteen hydroxyl groups on PVA chains and -bonding beteen the O groups of PVA and ater molecules. The competition beteen these types of interactions ill influence the properties of PVA solutions. The analysis of the concentration dependence of ln( η ), according to a ne method proposed by Wolf, instead of the traditional uggins plots or similar evaluations, provides a general facility to obtain intrinsic viscosity. This method as verified for aqueous solutions of polyelectrolytes in the absence/presence of salt and e demonstrated its validity for solutions of non-ionic polymers in organic solvents. The intrinsic viscosity values can be determined more accurately by modeling the viscometric data in the region of very lo concentrations according to this ne procedure instead of the uggins one. The viscosity molecular eight ationship as determined for PVA in ater at 0 C, shoing some differences as compared ith literature data. Thus, a Mark-ouink exponent of 0.6 as obtained in the present study, shoing less favorable PVA-ater interactions. Previously, 9 a value of 0.6 as reported for this parameter. These differences can be due to the both sample characteristics and the method used for the determination of the intrinsic viscosity. Acknoledgement: The research activity as realized as part of an exploratory research project (contract number 6/009). The authors ould like to thank the Roumanian National University Research Council for the financial support. Also, the authors are grateful to Prof. Bernhard A. Wolf from Johannes Gutenberg Universität Mainz for the useful discussions and comments and to CS I Dr. Constantin Ciobanu from Petru Poni Institute of Macromolecular Chemistry, Iasi (Roumania) for kindly providing the sample of polyvinyl alcohol. REFERENCES. M. Bercea, S. Morariu, S. Ioan, C. Ioan and B.C. Simionescu, Eur. Polym. J., 999,, 09-0.. C. Cincu, L.M. Butac and M. Iliescu, Eur. Polym. J., 000, 6, 7-77.. Y. Pan, W. Fu, F. Xue, J. Gu and R. Cheng, Eur. Polym. J., 00, 8, 9-6.. M. Xu, X. Yan, R. Cheng, and X. Yu, Polym. Int., 00, 0, 8-... Yang, P. Zhu, C. Peng, S. Mas, Q. Zhu and C. Fan, Eur. Polym. J., 00, 7, 99-9. 6. P. Zhu,. Yang, C. Peng, and X. Zhang, Macromol. Chem. Phys., 00, 0, 80-8. 7.. Yang,. Li, P. Zhu, Y. Yang, Q. Zhu and C. Fan, Polym. Testing, 00,, 897-90. 8.. Yang, Y. Yan,, P. Zhu,. Li, Q. Zhu and C. Fan, Eur. Polym. J., 00,, 9-0. 9. R.G. Larson, J. Rheol., 00, 9, -70. 0. J. Cai, R. Cheng and S. Bo, Polymer, 00, 6, 07-06.. C.E. Brunchi, M. Bercea and S. Morariu, igh Perform. Polym., 008, 0, - and references included therein.. B.A. Wolf, Macromol. Rapid Commun., 007, 8, 6-70.. J. Eckelt, A. Knopf and B.A. Wolf, Macromolecules, 008,, 9-98.. M.V. Badiger, N.R. Gupta, J. Eckelt and B.A. Wolf, Macromol. Chem. Phys., 008, 09, 087-09.. F. Samadi, B.A. Wolf, Y. Guo, A. Zhang and A. Dieter Schluter, Macromolecules, 008,, 87-880. 6. C.E. Brunchi, M. Bercea and S. Morariu, e-polymers, 009, no. 06, -0. 7. L. Ghimici and E.S. Dragan, Tranort properties of soluble polyelectrolytes, in Focus on Ionic Polymers, E.S. Dragan Ed., 9-9, 00. 8. L. Ghimici, M. Nechifor and B.A. Wolf, J. Phys. Chem. B, 009,, 800-80. 9. K. Yakamura, K. irata, S. Tamura and S. Matsuzaa, J. Polym. Sci., Polym. Phys. Ed., 98,, 70-7. 0. C. Damas, T. Leprince, T..V. Ngo and R. Coudert, Colloid Polym. Sci., 008, 86, 999-007.