pubs.acs.org/Langmuir © 2010 American Chemical Society
On the Water Dispersibility of a 1:1 Stoichiometric Complex between a Cationic Nanogel and Linear Polyanion Ryo Doi and Etsuo Kokufuta* Graduate School of Life and Environmental Sciences and Institute of Applied Biochemistry, University of Tsukuba, 1-1-1 Ten-noudai, Tsukuba, Ibaraki 305-8572, Japan Received May 9, 2010. Revised Manuscript Received June 24, 2010 This work aimed to obtain information on the water dispersibility of a 1:1 stoichiometric polyelectrolyte nanogel complex (SPENC). We synthesized a cationic polyelectrolyte nanogel (CPENG) composed of a cross-linked copolymer of 1-vinylimidazole and N-isopropylacrylamide. SPENC was then prepared at 25 C from the mixing of equimolar amounts (based on fixed charges) of CPENG and potassium poly(vinyl alcohol) sulfate, which were dissolved in an aqueous solution without adding salt and at pH 3.0. We carefully observed at 25 C the reduction of the imidazole-based cationic charge in the CPENG component of SPENC as a function of pH. Dynamic and static light scattering techniques were employed in combination with electrophoretic light scattering experiments. The amount of cationic charge in the SPENC was estimated from the potentiometric titration data of CPENG. It was found that, during the charge reduction process, the complex underwent aggregation, followed by a phase separation. The aggregation started at about 25% of the charge reduction (i.e., at pH ≈ 4.9), and the phase separation took place when almost half of the charge was eliminated (at pH ≈ 5.5). However, the phase-separated complexes became water-soluble again when about 90% of the charge was eliminated (pH ≈ 6.6). By colloid titration, the dissociated free polyanions were not detected in the aqueous SPENC solution before the phase separation but were detected in the complex-redispersed solution. When the pH (9.0) of the redispersion was slowly decreased to the original level (pH 3.0) by the gradual addition of HCl so as to cause again the phase separation, an intraparticle complex was reformed, the physical quantities of which were close to those of the initial SPENC. These findings clearly indicate that the whole and a part (segment) of the complexed polyanions undergoes dissociation-association reactions on the surface of a SPENC particle, depending on the ionization state of the cationic gel component. As a result, these reactions seem to be a key factor for the water dispersibility of the SPENC.
Introduction Since the early work on polyelectrolyte complexes (PECs) by Fuoss and Sadek1 in 1941 and by Michaels and Miekka2 in 1961, an enormous number of studies have been published in this field.3 Most of the aqueous PEC systems reported so far can be divided into two types on the basis of macroscopic observations: watersoluble PECs and precipitated or phase-separated PECs. In some cases, the phase-separated types are distinct from the complex precipitates and further subdivided into stable turbid dispersions, flocculating systems, and coacervates upon the state of PECs in aqueous solutions. In the classification of PECs, it is also controversial whether the complex formation is stoichiometric or nonstoichiometric.3 *To whom correspondence should be addressed. E-mail: kokufuta@ sakura.cc.tsukuba.ac.jp. Fax: 81-298-53-4605. (1) Fuoss R. M.; Sadek H. Sterling Chemistry Laboratory, Yale University: New Haven, 1949; Number 25, Vol. 110, pp 552-554. (2) Michaels, A. S.; Miekka, R. G. J. Phys. Chem. 1961, 65, 1765. (3) For examples, see: (a) Bekturov, E. A.; Bimendina, L. A. Adv. Polym. Sci. 1981, 41, 99–147. (b) Tsuchida, E.; Abe, K. Adv. Polym. Sci. 1982, 45, 1–123. (c) Smid, J.; Fish, D. In Encyclopedia of Polymer Science and Engineering; Mark, H. F., Bikales, N. M., Overberger, C. G., Menges, G., Eds.; Wiley: New York, 1988; Vol. 11, pp 720-739. (d) Philipp, B.; Dautzenberg, H.; Linow, K.-J.; Koetz, J.; Dawydoff, W. Prog. Polym. Sci. 1989, 14, 91–172. (e) Kabanov, V. A. In Macromolecular Complexes in Chemistry and Biology; Dubin, P., Bock, J., Davies, R. M., Schulz, D. N., Thies, C., Eds.; Springler-Verlag: Berlin, 1994; pp 151-173. (f) Kokufuta, E. In Macromolecular Complexes in Chemistry and Biology; Dubin, P., Bock, J., Davies, R. M., Schulz, D. N., Thies, C., Eds.; Springler-Verlag: Berlin, 1994; pp 301-325. (g) Bell, C. L.; Peppas, N. A. Adv. Polym. Sci. 1995, 122, 125–175. (h) Kabanov, A. V.; Kabanov, V. A. Adv. Drug Delivery Rev. 1998, 30, 49–60. (i) Wandrey, C.; Grigorescu, G.; Humkeler, D. Prog. Colloid Polym. Sci. 2002, 119, 84–91. (j) Kabanov, A. V. In Multilayer Thin Films; Decher, G., Schlenoff, J. B., Eds.; Wiley-VCH: Weinheim, Germany, 2003; pp 47-86. (k) Koetz, J.; Kosmella, S. Polyelectrolytes and Nanoparticles; Springer: Berlin, Heidelberg, 2007; pp 36-46.
Langmuir 2010, 26(16), 13579–13589
Polyelectrolytes of opposite charge in aqueous media associate with each other to form primary PECs due to a strong electrostatic interaction. This process results in the release or diffusion of counterions that are condensed around the polyion, thus causing an increase in entropy in the system. Although other intermolecular forces such as hydrogen bonding and hydrophobic interactions would play an important role in the formation of the primary complex and/or its further aggregation or structuralization, the stoichiometry of PEC formation is generally discussed in terms of the electrostatic association between the fixed charges of opposite sign. In other words, in principle, one-to-one stoichiometric formation of PECs means that the reaction follows a 1:1 “ion-molar” stoichiometry (Aþ þ B- f AB). In many cases, however, the complexes from the mixing of anionic and cationic polyelectrolytes at an equal ion (or unit) molar ratio are referred as to “1:1 stoichiometric” or simply “stoichiometric” PECs. Such complexes usually fall into the precipitated or phase-separated types. Nonstoichiometric PECs are well-known to be soluble or dispersible in aqueous media in the presence of low molecularweight salts. There are two types of nonstoichiometric PECs in empirical classification: (i) molecular complexes consisting of a long host polyion and shorter sequentially attached guest polyions of the opposite charge;4 and (ii) large nonequilibrium aggregates of PECs formed with an excess of either anionic or cationic components.5 Both PEC systems, where the polyion charges exceeded the stoichiometric amount for the complexation, would provide water-solubility or dispersibility characteristics. The excess (4) Kabanov, V. A.; Zezin, A. B. Makromol. Chem., Suppl. 1984, 6, 259. (5) (a) Dautzenberg, H. Macromolecules 1997, 30, 7810. (b) Karibyants, N.; Dautzenberg, H. Langmuir 1998, 14, 4427.
Published on Web 07/14/2010
DOI: 10.1021/la101852b
13579
Article
charge of aggregated PEC particles is on their surfaces, while in the molecular complexes this excess charge exists as a “void” between the sequentially complexed sections of the host chain (see Scheme A in ref 4). It has been believed that in the latter PEC system there is an association-dissociation equilibrium between the host and guest polyions. In this context, it may be said that the guest polyion is able to “walk” along the host polyions (see page 38 of ref 3k). PECs of the water-insoluble types have been used as industrial materials,3k for example, membranes, coatings, binders, and flocculants. On the other hand, it is expected that the improvement of the water solubility or dispersibility of PECs allows the potential application in such biotechnological and pharmaceutical fields as microencapsulation or immobilization of biological objects and carriers for gene therapy or drug delivery.3e For this reason, most studies have focused on the synthesis and properties of water-soluble PECs of not only the nonstoichiometric but also of the stoichiometric types. Since the stoichiometric PECs are usually water-insoluble, molecular improvement of anionic or cationic polyelectrolyte components has been studied in an attempt to form stoichiometric water-soluble PECs.6-17 Many studies have used polyelectrolytes bearing nonionic and watersoluble blocks6-12 in their backbones or side chains,13-16 except that a few cases17 have employed hydrophilic copolypeptides. Dynamic light scattering (DLS) was used in almost all of the studies6,8-12,14-17 to measure the hydrodynamic radius (Rh) of a PEC particle. Also used in several studies were static light scattering (SLS)6b,9-13 and/or electrophoretic light scattering (ELS).6a,9,10,15,16 These measurements were mainly performed as a function of the polyanion/polycation ion-molar ratio in the presence or the absence of low-molecular-weight salts. It should be noted that to estimate the weight-average molecular weight (M w) of a PEC from SLS the measurements of the concentration and the refractive index increment are needed, although these values in the aforementioned studies6b,9-13 were obtained from calculations (e.g., the concentration of PEC under an assumption of a 1:1 stoichiometric complexation). By limiting the focus to the stoichiometric and water-soluble or dispersible PECs as quoted above,6-17 we can find that the complex particle has a core-shell or corona structure in which the outer layer (shell or corona) contributes to the water solubility or dispersibility. This was the case when polyelectrolyte-neutral diblock copolymers were complexed with macroions of opposite surface charge, that is, charged dendrimers8 (4-5 nm in diameter, as calculated by a computer) and citrate-coated inorganic colloids (Rh =5-18 nm).12 The complexation of star-shaped polyanions with linear polycations was found to yield water-soluble PECs (6) (a) Harada, A.; Kataoka, K. Macromolecules 1995, 28, 5294. (b) Harada, A.; Kataoka, K. Science 1999, 283, 65. (7) Kabanov, A. V.; Bronich, T. K.; Kabanov, V. A.; Yu, K.; Eisenberg, A. Macromolecules 1996, 29, 6797. (8) (a) Stapert, H. R.; Nishiyama, N.; Jiang, D.-L.; Aida, T.; Kataoka, K. Langmuir 2000, 16, 8182. (b) Zhang, G.-D.; Nishiyama, N.; Harada, A.; Jiang, D.-L.; Aida, T.; Kataoka, K. Macromolecules 2003, 36, 1304. (9) Gohy, J. F.; Varshney, S. K.; Antoun, S.; Jerome, R. Macromolecules 2000, 33, 9298. (10) Zintchenko, A.; Dautzenberg, H.; Tauer, K.; Khrenov, V. Langmuir 2002, 18, 1386. (11) Holappa, S.; Andersson, T.; Kantonen, L.; Plattner, P.; Tenhu, H. Polymer 2003, 44, 7907. (12) Berret, J. F. Macromolecules 2007, 40, 4260. (13) Dautzenberg, H. Macromol. Chem. Phys. 2000, 201, 1765. (14) Sotiropoulou, M.; Cincu, C.; Bokias, G.; Staikos, G. Polymer 2004, 45, 1563. (15) Matralis, A.; Sotiropoulou, M.; Bokias, G.; Staikos, G. Macromol. Chem. Phys. 2006, 207, 1018. (16) Shovsky, A.; Varga, I.; Makuska, R.; Claesson, P. M. Langmuir 2009, 25, 6113. (17) Vasilevskaya, V. V.; Leclercq, L.; Boustta, M.; Vert, M.; Khokhlov, A. R. Macromolecules 2007, 40, 5934.
13580 DOI: 10.1021/la101852b
Doi and Kokufuta
with a core-shell or corona structure.18 Although the complexes obtained were a mixture of small particles (a major fraction) and large particles (a minor fraction), the latter was reported to be an aggregate with “base-molar stoichiometry” close to 1:1 (i.e., a 1:1 molar ratio of charged basic group to charged plus uncharged acidic group; see ref 18a). From these different types of stoichiometric water-soluble PECs, a simple question naturally arises as to what macromolecular structure(s) can be considered for the interior core in which many of the ion-pairs between the fixed charges could be concentrated. (Is it like a “scrambled egg” model proposed by Michaels? See Figure 2 in ref 19.) The answer to this question would provide another approach to understand the water solubility or dispersibility of the stoichiometric PECs in which the charge neutrality condition is satisfied. A clue to answer the above question would be the use of polyelectrolyte architectures which are unable to form the scrambledegg-like structure of an interior core, but are able to provide a water-soluble PEC based on a one-to-one stoichiometry. Polyelectrolyte nanogels (PENGs), that is, lightly cross-linked polyelectrolyte gel particles with diameters in the range of tens to hundreds of nanometers, are regarded as a good candidate for meeting these two requirements. This approach is derived from our previous studies on the complex formation20 of linear polyions, as well as on the geometric characteristic21 of PENG particles, the outlines of which are as follows. Our PENGs with Rh = 100-160 nm have an obscure surface covering with dangling chain hairs and have ionic groups bound to both the gel network and dangling chain;20 thus, they are distinguishable from the macroions as reported in refs 8 and 12. These conclusions have been obtained from a combination of DLS and SLS,20 that is, the measurements of Rh from DLS and radius gyration (mean-square radius; Rg) as well as M w from SLS. These light scattering techniques were also usable in the studies20 of the PEC formation, even in the absence of low-molecularweight salts (i.e., salt-free system). The use of the salt-free system has an experimental advantage because one may assume a 1:1 stoichiometric charge neutralization in the complexation without consideration of the screening effect of counterions from added salts. Then, the stoichiometry may be discussed on the “formal” charges (in ion moles per mole of a polyelectrolyte chain or a PENG particle) which can be determined by experiments. As a result, it has been demonstrated that a cationic PENG (abbreviated as CPENG) forms a water-soluble “intraparticle” complex at the molar ratio (rm) of the fixed anion to the fixed cation of e1.0, according to the following reaction:20 Gmþ þ aPn - f Gmþ ðPn - Þa
ð1Þ
where Gmþ denotes a CPENG particle with þm charges, Pn- is a polyanion chain with -n charges, and a = (m/n)rm (note that þm and -n are measurable quantities by a combination of SLS and potentiometric titration). Gmþ(Pn-)a at rm = 1, abbreviated hereafter as stoichiometric polyelectrolyte nanogel complex (SPENC), is obviously different from water-soluble stoichiometric PECs6-17 of the core-shell type, because the binding of a moles of Pn- ions on the surface of a (18) (a) Pergushov, D. V.; Babin, I. A.; Plamper, F. A.; Alexander, B.; Zezin, A. B.; Muller, A. H. E. Langmuir 2008, 24, 6414. (b) Pergushov, D. V.; Babin, I. A.; Plamper, F. A.; Schmalz, H.; M€uller, A. H. E.; Zezin, A. B. Doklady Phys. Chem. 2009, 425, 343. (19) Michaels, A. S. Ind. Eng. Chem. 1965, 57, 32. (20) (a) Ogawa, K.; Sato, S.; Kokufuta, E. Langmuir 2005, 21, 4830. (b) Ogawa, K.; Sato, S.; Kokufuta, E. Langmuir 2007, 23, 2097. (21) Kokufuta, E.; Ogawa, K.; Doi, R.; Farinato, R. S. J. Phys. Chem. B 2007, 111, 8634.
Langmuir 2010, 26(16), 13579–13589
Doi and Kokufuta
Article
Table 1. Properties of CPENG, KPVS, and SPENC in a Salt-Free Aqueous Solution at pH 3 and at 25 C sample
formal charge (mol/mol)
M w (g/mol)
Rh (nm)
Rg (nm)
Rg/Rh
U (104 cm2/(V s))
1.17 107 170 257 1.512 4.13 CPENG 2.04 104 4.19 105 -b -b -b -4.32c KPVS 2.58 103a d 7 SPENC 1.55 10 85 117 1.376 2.38 a Determined in an aqueous 0.2 M KCl and at 25 C. b Note that in the absence of salts such as KCl the data cannot be obtained by DLS and SLS due to polyelectrolyte effects; e.g., see ref 22. c Determined at a relatively high concentration (5 g/L). d Expected value is zero.
Gmþ particle could lead to formation of an external hydrophobic shell. Then, one would have a question about the water solubility of SPENC: that is, it is essential to know a key factor that prevents the complex particles from aggregating. Taking into account the nonstoichiometric PECs, we may assume that an associationdissociation equilibrium between Gmþ and Pn- would become the factor. However, direct experimental confirmation of this assumption is very difficult, even though it has been suggested that a polyelectrolyte exchange reaction takes place in a SPENC/ CPENG or SPENC/polycation mixture.20 The reason is that the whole and a part (segment) of the Pn- chain can undergo the dissociation and association reactions on the surface of a SPENC particle. As an idea to overcome this experimental difficulty, we attempt here to study the dissociation of Pn- from the SPENC particle (i.e., Gmþ(Pn-)a at rm=1) by the reduction of þm charges. The SPENC was prepared with a new type of CPENG, which is composed of a cross-linked copolymer of 1-vinylimidazole (VI) and N-isopropylacrylamide (NIPA) (note that the SPENC used in ref 20 was prepared from a positively charged polyampholytic nanogel). The complex formation was performed with a strong polyanion, potassium poly(vinyl alcohol) sulfate (KPVS). In the present SPENC system, we may change the value of þm per complex upon the addition of a base (e.g., KOH) when the complex undergoes an association-dissociation reaction as
continuous addition of the calculated volume (1.74 mL containing a total of 2.175 10-6 moles of the sulfate anion) of the KPVS solution (pH 3.0) to the CPENG dispersion (50 mL; pH 3.0) containing the imidazole cations, the amount of which is equal to that of the sulfate anion. From Table 1 and Figure 1, it is found that the size (Rh and Rg) of CPENG decreases in the complexation with KPVS. However, the electrophoretic mobility (U in cm2/ (V s)) of SPENC did not become zero. Then, one might have a question about precision in our SPENC preparation, that is, whether the complexation was conducted with the “exact” 1:1 stoichiometric amounts of CPENG and KPVS. To remove this concern, we have carried out the measurements of DLS and ELS with the CPENG dispersion to which a 10% excess of the stoichiometric amount of KPVS was added (see Figure 1a and 1c). The addition of this excess causes an increase of Rh but little change in the mobility; Rh = 162 nm and U = 2.12 10-4 cm2/ (V s) for the complex at rm = 1.1. Before discussing the DLS and ELS results of the nanogel complexes of CPENG with the stoichiometric or excess amount of KPVS, we attempt to calculate how many KPVS chains are required for the formation of one SPENC particle. This can be provided by the ratio of formal charges; that is, 7.91 chains per particle, which may be approximated as 8 chains in almost all the cases, due to the existence of a molecular weight distribution in synthetic polymer systems. Therefore, we may conclude that the M w (1.55 107 g/mol) of SPENC agrees well with the calculated value (1.51 107 g/mol) under the assumption that the binding of 8 KPVS chains to 1 CPENG particle results in a 1:1 stoichiometric complex (i.e., SPENC). Nevertheless, we note that the molar mass of our SPENC was estimated from the SLS data (see Figure 1b) on the basis of the Rayleigh-Gans approximation: ! KCx 1 16π2 2 2 θ þ 2A2 Cx ¼ 1 þ 2 Rg sin 2 Rθ Mw 3λ
ð3Þ
K ¼ 4π2 =λ40 NA n0 2 ðdn=dcÞx 2
ð4Þ
and The reasons are as follows: (i) the ionization degree (R) of CPENG-bound imidazole groups depends on pH, but (ii) the charge (-n) of KPVS is independent of pH over a wide pH range. In this work, we have investigated the dissociation of KPVS ions as a function of pH (therefore R) by a combination of various experimental methods: DLS, SLS, ELS, potentiometric titration, and colloid titration. Also investigated was whether the dissociated KPVS ions reassociate again to form the SPENC. The study of this sort could be helpful in understanding the nature of water solubility not only of SPENCs but also of different types of stoichiometric and nonstoichiometric PECs, through which we can provide a significant base for the molecular design of water-soluble PEC architectures.
Here, Cx denotes the complex concentration (in g/mL), A2 is the second virial coefficient, λ and λ0 are wavelengths of the light in the medium and in a vacuum, respectively, NA is Avogadro’s number, n0 is the refractive index of the medium, and (dn/dc)x is the change in the refractive index of the complex with Cx. The value of (dn/dc)x was not measurable due to a strong light scattering characteristic of the SPENC particle and is thus calculated from dn 1 dn β dn ¼ þ ð5Þ dc x 1 þ β dc g 1 þ β dc p
Results and Discussion Preparation and Properties of SPENC. Table 1 shows the properties of our SPENC, together with those of CPENG and KPVS. The preparation of SPENC was performed by slow and
where (dn/dc)g (0.1831 mL/g) and (dn/dc)p (0.1145 mL/g) represent the changes in the refractive index of CPENG and KPVS, respectively, and β is the weight ratio of KPVS to CPENG. For this reason, we have noted in the Introduction that the M w of
Langmuir 2010, 26(16), 13579–13589
DOI: 10.1021/la101852b
13581
Article
Doi and Kokufuta
“ladder” or “scrambled egg” model proposed by Michaels.19 Such a complex should be a big particle and has a tendency to aggregate with each other. In our PEC formation system, however, all the sulfate ions (-OS(dO)2O-) may not form a “tightly linked ion-pair” with the protonated imidazole cations (tNHþ) bound to a CPENG particle, because most of them are within the cross-linked polymer network. Then, one is allowed to assume that (i) a slight excess charge of the complexed KPVS anion is screened by potassium ions (as the counterion) so as to maintain the “quasi-overall” charge neutrality of a SPENC particle even if it has 8 chains of KPVS, but (ii) several of the KPVS chains dissociate from the complex under an applied electric field in ELS. Also, (iii) upon the addition of an excess (10%) of KPVS, a more stabilized complex is formed, consisting of two or more CPENG particles in an aggregation state. These assumptions allow one to explain the results in Figure 1, but may not be supported solely by a study on the formation process of SPENC with DLS and SLS as well as ELS. Therefore, we have paid attention to the dissociation of the bound KPVS ions from the SPENC particle when its positive charges are gradually eliminated by the pH-induced deprotonation (see eq 2). Effect of pH on the SPENC Dispersion. First, we have followed the pH dependence of R for the imidazole groups which are bound to CPENG. Potentiometric titration was then performed, and R was estimated from
CHþ - COH R ¼ Rn þ CM
ð6Þ
where Rn is the degree of neutralization of the protonated imidazole group (as an acid) with KOH as a titrant; CHþ and COH- denote the molar concentrations of Hþ and OH- ions, respectively; and CM is the CPENG concentration (in mol/L) based on the imidazole group. Figure 2 shows the pH versus R curve of an aqueous CPENG dispersion in the absence of lowmolecular-weight salts. Also shown in Figure 2 is the titration data given by the Henderson-Hasselbalch (H-H) equation: 1-R pH ¼ pK þ n log R
Figure 1. Results of characterization for SPENC by DLS (a), SLS (b), and ELS (c). In (a) and (c), the results of CPENG, as well as of a complex (at rm = 1.1) of the CPENG with 10% excess of KPVS, are also shown for reference. R The function f(Rh) was normalized by the scattered intensity (I ∼ ¥ 0 f(Rh) dRh) of the SPENC dispersion with the concentration of 0.031 g/L. The A2 value of SPENC obtained from a Zimm plot was 6.60 10-5 cm3 mol/g2.
PECs6b,9-13 (involving our SPENCs20) has not yet been determined by SLS without any assumptions and approximations. Next, let us reconsider the nature of PEC formation reactions in which a strong electrostatic association between the fixed charges of opposite sign plays a key role. For this, it would be reasonable to consider that when the number of charges of either the anionic or cationic component is slightly different from that of another component, their association continues until the electroneutrality condition is fulfilled, resulting in a PEC like the 13582 DOI: 10.1021/la101852b
ð7Þ
Here, pK is the dissociation constant obtained at R = 0.5 and n is unity for monomeric acids or bases. For polyacids or polybases, however, pK is an apparent dissociation constant and n is larger than unity due to the effect of polyion charges (e.g., see ref 23). In our case, the plot of pH versus log{(1 - R)/R} shows a linear line in the range of R between 0.01 and 0.95; thus, we can obtain pK = 5.48 and n = 1.16 in the H-H equation. By using the results in Figure 2, we have examined the reduction of the imidazole charge in the SPENC by DLS. The pH (= 3.0) of the complex dispersion was increased to a desired value by addition of a very slight volume of an aqueous KOH solution (0.001-1 mol/L). After that, the dispersion was equilibrated with gentle stirring for 1 day, followed by standing for 3 days. To avoid the effect of carbon dioxide on the solution pH, these procedures were performed under a nitrogen atmosphere. Figure 3 shows how the scattering intensity (Is) varies with pH or R for the SPENC or CPENG dispersion, together with the results of visual observation of the pH-induced precipitation and redispersion (22) Schmitz, K. S., Ed., Macro-ion Characterization: From Dilute Solutions to Complex Fluids; American Chemical Society: Washington, DC, 1994; Chapters 25 and 26. (23) Morawetz, H. Macromolecules in Solutions; John Wiley Interscience Publishers: New York, London, Sydney, 1965; pp 348-356.
Langmuir 2010, 26(16), 13579–13589
Doi and Kokufuta
Figure 2. Results of potentiometric titration of CPENG (as the HCl salt) with KOH (0.1 mol/L) in pure water: (a) relation of pH and the degree (R) of ionization (or protonation); (b) fitting of the data in (a) by the Henderson-Hasselbalch equation. For more details, see the text.
of the complex particles. R was scaled on the upper abscissa of Figure 3a, so as to satisfy eq 7 with pK = 5.48 and n = 1.16; thus, the range of the scale becomes narrow around R = 0.5. It is found that the decrease of R with increasing pH leads to a sigmoidal increase in the Is of CPENG (see open circles), but to a peculiar change in the Is of SPENC around R ≈ 0.48 (pH ≈ 5.51) due to a phase separation, and around R ≈ 0.11 (pH ≈ 6.56) due to the redispersion (the pH or R leading to the phase separation is hereafter abbreviated as pHp or Rp, and similarly the pH or R leading to redispersion of the phase-separated complex is written as pHr or Rr). The pictures in Figure 3b and c clearly show that the complex is phase-separated, followed by precipitated, and dispersed again during the reduction of the positive charge in the nanogel component. The redispersed complex particle exhibits the Is value close to that of the CPENG particle under conditions where most of the charge is removed via the deprotonation of imidazole cations (see the plot with x at R < 0.05). Generally Is increases with increasing the size as well as the concentration of scattering particles. In the case of polyelectrolyte gel particles, however, the network collapse leads to an increase of Is; for example, the complexation of CPENG with KPVS causes an increase of Is accompanied by a decrease of Rh (see Figure 1a). Therefore, we must consider this fact in comparison with the dependencies of Is and Rh on R. The data in the Supporting Information (Figure S1) show how the Rh of CPENG or SPENC varies with the pH-dependent R change. In the case of CPENG, the Rh decreases sigmoidally with decreasing R, contrary to the Langmuir 2010, 26(16), 13579–13589
Article
change of Is with R, indicating the gel collapse due to the deprotonation of the network-bound imidazole cations. At R > 0.75, the R dependencies of Rh and Is for SPENC are similar to those for CPENG. This means a possibility that several imidazole cations in the gel component of the complex undergo deprotonation so as to reduce the positive charge which interacts with the sulfate ion of KPVS, as shown in eq 2. In other words, there is a possibility of the whole or partial dissociation of a KPVS ion complexed with CPENG. In the range of R between 0.75 and Rp (0.48), in which CPENG shows monotonous changes in Rh and Is, SPENC displays a rapid increase in Is as well as in Rh (see Supporting Information Figure S1). This is an indication that the complex particles aggregate with each other before the phase separation at Rp. At R e Rr (0.11), both Is and Rh of SPENC decrease rapidly and then become close to those of CPENG, indicating that the aggregated complex particles dissociate from each other when most of the positive charges of the CPENG component are eliminated. These results also suggest the dissociation of the KPVS ions from the complex surface during the charge reduction of the cationic gel component. Detection of the Dissociated KPVS Ions by Colloid Titration. In this measurement, it is necessary to separate the dissociated KPVS ions from the complex dispersion. We thus have attempted to search for a good filter which allows the passage of the polyions only but not the complex particles. Through several preliminary experiments, a PALL Acrodisc syringe filter with 0.1 μm Supor membrane (Pall Life Sciences, NY) was found to satisfy the present demand. As can be seen from Figure 4, there is a good agreement between the colloid titration data of a standard KPVS solution (4.2 10-5 mol/L based on the sulfate ion; 50 mL) and its filtrate, both of which were titrated with a polycation, poly(diallyldimethylammonium chloride) (PDDA; 2.5 10-3 mol/L based on the ammonium ion). In addition, the DLS measurements of the filtrates, which were obtained from the CPENG dispersions over the range of pH 3-9, showed that there are no scattering particles in each sample. As a result, it has become apparent that the present technique allows us to detect the KPVS ion when it could dissociate from the complex. The colloid titration with PDDA was performed with the filtrates (50 mL) obtained from the SPENC dispersions at various pHs covering all of the R range (0-1.0). As shown in Figure 4, no KPVS ions were detected in the range of pH < pHp (5.51), that is, R > Rp (0.48), and also in the phase separated solution from which the precipitated complex had been removed. However, we can find the free KPVS ions in the filtered samples obtained from the complex redispersions. As shown in Table 2, the amount of the dissociated KPVS increases with increasing pH, and at pH = 8.2 (R ≈ 0) the dissociation becomes almost half of the stoichiometrically complexed amount. With a further increase in pH to 9, 80% of the complexed KPVS was dissociated. This indicates that 20% of the tightly linked ion-pairs between the sulfate anion and the imidazole cation remain unseparated from the surface of the complex particle, even at pH 9 at which the CPENG-bound imidazole cation is completely deprotonated. From eq 2, one might argue whether the separation of the ionpair is due to (i) the pH-induced charge elimination (i.e., deprotonation of a charged imidazole) or (ii) the effect of the resulting KCl during the increase in pH on addition of KOH (i.e., salt effect). The original complex dispersion was adjusted to pH 3.0, so that the raising of the pH to 9.0 with KOH yields about 10-3 mol/L of KCl. On the basis of this estimation, we prepared a SPENC dispersion (pH 3.0) containing 5 10-3 mol/L (5 times excess) of KCl and subjected it to the filtration. Also prepared as a control sample was a KPVS solution under the same conditions of pH DOI: 10.1021/la101852b
13583
Article
Doi and Kokufuta
Figure 3. Changes in the dispersion state of SPENC and CPENG as a function of pH or R: (a) Light scattering intensities (Is) of SPENC and CPENG; (b) white precipitate (see the bottom of the bottle) resulting from the SPENC dispersion when the pH was adjusted to 5.8 to lead phase separation, followed by allowing the dispersion to stand for 3 days; (c) redispersion of the precipitated SPENC with stirring at pH 6.6 for 3 h. The results of CPENG in panel (a) are shown in two magnifications (see open circles and circles with cross). Also in panel (a), pHp or Rp shows the value at which the phase separation takes place, whereas at pHr or Rr the redispersion occurs. Note that we can see a laser beamline in photo (c) due to the Tyndall effect. Table 2. Amount of Dissociated KPVS Ions as a Function of pH and r dissociated KPVS ions pH
R
(%)
(chains/particle)
1.00 nda 0.94 nda 0.68 nda b 0.26 nda 0.07 18 0.00 47 0.00 80 a Denotes no detection. b Phase separation occurred.
3.0 4.1 5.1 6.0b 6.8 8.2 9.0
Figure 4. Results of colloid titration with PDDA (0.0025 mol/L) of the following samples (50 mL each): standard KPVS solution (O) at pH 6.0 and its filtrate (b), and filtrates of the complex dispersions at pH 3.0 (4), 5.1 (2), 6.0 (), 6.8 (9), 8.2 (0), and 9.0 (þ). The control titration performed with the standard solution at the same pH as that of each filtrated sample showed no difference from the result at pH 6.0 (see curve indicated by open circles). Note that the results (4,2,) at pH e 6.0 are seen on the bottom of the x axis. 13584 DOI: 10.1021/la101852b
1.42 (≈1) 3.72 (≈4) 6.33 (≈6)
and KCl concentration. The colloid titration of these samples showed no dissociation of the KPVS from the SPENC. As a result, it has become apparent that a part of the complexed KPVS ions dissociate entirely (but not segmentally) from the SPENC particles upon the addition of KOH as shown in eq 2. Detection of the Dissociated KPVS Ions by ELS. Our initial question of the ELS behavior of SPENC was why its mobility was a positive but not a neutral value. Then, we assumed the dissociation of the complexed KPVS ion during the ELS measurement. Unfortunately, this assumption was not confirmed by the colloid titration. However, as can be seen from ELS diagram (a) in Figure 5, the complex showed a negative mobility (-2.3 10-4 cm2/(V s)) at pH 8.2 (R ≈ 0) at which almost half of Langmuir 2010, 26(16), 13579–13589
Doi and Kokufuta
Figure 5. Electrophoretic diagrams of the following samples: (a) SPENC at pH 8.2; (b) SPENC at pH 11; (c) KPVS (5 g/L) at pH 11; (d) NIPA nanogel (0.025 g/L) at pH 11.0; and (e) NIPA nanogel and KPVS mixture (pH 11.0) having the same polymer concentrations as those of the SPENC dispersion.
the complexed KPVS ions were dissociated (see Table 2). We may understand the observed negative mobility at pH 8.2 by considering that the polyanions which remain persistently bound to the complex become dominant in ELS. Nevertheless, the dissociated KPVS ions did not appear in ELS diagram (a). To resolve this contradiction, we consider that the KPVS ion dissociated at pH 8.2, whose concentration is 3.21 10-3 g/L from the data in Table 2, is detectable in the colloid titration but not in the ELS measurement. To confirm this consideration, we have attempted to perform ELS at pH 11 at which a considerable amount (more than 85%) of the dissociation of KPVS was “expected” from the colloid titration. We did not show the titration data at pH 11 in Table 2 because a control titration at this pH did not agree well with those at the other pH levels (3.0-9.0), mainly due to an effect of KCl resulting from the rise in the solution pH with KOH. In the ELS measurement, however, such a salt effect was not dominant; therefore, ELS diagram (b) of the complex solution at pH 11 showed a small peak at the peak position of diagram (c) which was obtained with an authentic KPVS solution of a high concentration (5 g/L). The accuracy of the ELS data at pH 11 is also illustrated by the following: (i) the main peak of diagram (b), which is assignable to a complex particle with the remaining bound KPVS, is shifted toward the zero mobility line as compared with diagram (a), and (ii) the intensity (equivalent to the peak area) of the main peak of diagram (b) is stronger than that of diagram (a), due to the network collapse which seems to be caused by the aforementioned salt effect. To make more clear the detection of a slight amount of KPVS by ELS in the presence of nanogel particles, we have investigated the ELS behavior of aqueous nonpolyelectrolyte nanogel dispersions with and without the KPVS ions. ELS diagrams (d) and (e) on the upper side of Figure 5 are the results at pH 11 for the gel particles (2.42 10-2 g/L) of NIPA homopolymer24 in the absence and the presence of KPVS (6.83 10-3 g/L), respectively. The concentrations (in g/L) of the polymers (NIPA gel and KPVS) for the nonpolyelectrolyte nanogel dispersions are the same as those for the SPENC dispersion; thus, these samples may be used as a control for comparison with the complex dispersion (24) Suzuki, H.; Kokufuta, E. Colloids Surf., A 1999, 147, 233.
Langmuir 2010, 26(16), 13579–13589
Article
at pH 11. It is found that the ELS intensity of the NIPA gel dispersion without KPVS is very much lower than that of the complex. This is due to a difference in the polymer densities of both gel particles, rather than in their particle sizes (Rh = 57 nm for NIPA nanogel and 78 nm for SPENC) at pH 11. The reason is that the NIPA nanogel is in a swollen state but the complex is in a fully collapsed state. We should note that the reason for a slight negative mobility of the NIPA gel is due to a few polymer-bound sulfate ions which come from an initiator (ammonium persulfate). Also noted is that the shoulders on both sides of the peak were artificially generated, because a single and sharp peak was observed at a high concentration (e.g., 1 g/L) of NIPA nanogel dispersion. Now, let us look at the KPVS-containing dispersion of NIPA gel particles, that is, ELS diagram (e). Then we can barely observe “vague protuberances” at the positions of NIPA nanogel and of KPVS. This means that in the ELS experiments it is very difficult to detect amounts less than half of the complexed KPVS, as was seen in diagram (a) for the complex dispersion at pH 8.2. As a result, it is still undeniable that the observation of positive mobilities for the SPENC is related to the dissociation of the polyanions during the ELS measurement. Mobility Changes as a Function of pH or r. It has become apparent that in ELS the complex behaves as a cationic particle at pH 3 but as an anionic particle at pHs 8 and 11; thus, we have studied the mobility change over a wide pH range. As shown in Figure 6, the mobility of the complex is positive in the range of pH < pHp or R > Rp but negative in the range of pH > pHr or R < Rr. First, let us look at the ELS results in the pH range below pHp. It is found that the positive mobility of SPENC almost remains unchanged at R > 0.95 (pH < 4) and then gradually decreases upon raising the pH so as to decrease R. An overlap can be seen in the results between SPENC and CPENG when each observed mobility of the latter is shifted by a constant value of 1.80 10-4 cm2/(V s) to the negative direction (see cross symbols). The difference (1.80 10-4 cm2/(V s)) remains constant up to pHp. This means that the mobility change of SPENC with pH is dominated by the pH dependence of R for a “part” of the CPENG component of the complex. Such a cation-rich part in the complex should be free from the binding of the sulfate ion of KPVS, so that we must return to the issue of why the SPENC prepared at pH 3 has the positive mobility. The cation-rich part in the complex seems like a segment in a cationic polyelectrolyte when we recall a free draining model (FDM)25 to explain the electrophoretic behavior of polyelectrolytes. Indeed, our previous study26 has demonstrated that the pH dependence of the mobility for a polyampholyte (M w ≈ 1.92 105 g/mol) having both the imidazole (1.16 mmol/g) and the carboxyl (0.31 mmol/g) groups agrees well with that of an ampholytic nanogel (M w ≈ 5.24 107 g/mol) whose imidazole and carboxyl group contents are very close to those of the ampholytic polymer; in more detail, the mobilities measured in 0.05 M KCl at pH 3.0 were 2.20 10-4 cm2/(V s) for the polymer and 2.25 10-4 cm2/(V s) for the nanogel, as well as the isoelectric points were pH = 6.50 for both. The cationic nanogel used here contains the imidazole groups (1.74 mmol/g), which is 1.5 times those of the ampholytic polymer or nanogel reported in ref 26. In addition, the ELS of the present CPENG was performed in the salt-free system, so that the mobility at pH 3.0 should be larger than that of the positively charged ampholytic nanogel. For the SPENC, its (25) (a) Hermans, J. J.; Fujita, H. Proc. K. Ned. Akad. Wet., Ser. B: Phys. Sci. 1955, 58, 182. (b) Overbeek, J. T. G.; Stigter, D. Recl. Trav. Chim. Pays-Bas 1956, 75, 543. (c) Hermans, J. J. J. Polym. Sci. 1955, 18, 529. (26) Ogawa, K.; Nakayama., A.; Kokufuta., E. J. Phys. Chem. B 2003, 107, 8223.
DOI: 10.1021/la101852b
13585
Article
Doi and Kokufuta
Figure 6. Changes of electrophoretic mobilities (U) of SPENC (closed circles) and CPENG (open circles) as a function of pH or R. A
difference of 1.80 10-4 cm2/(V s) from the observed mobility of CPENG at each pH or R is represented by the cross symbol, to indicate that there is a good overlap between the mobility curves of SPENC and CPENG in the range of pH < pHr or R > Rr.
composition is chemically equivalent to an ampholytic nanogel having equimolar amounts (1.74 mmol/g) of the imidazole and the sulfate groups. Such an imidazole-based ampholyte would have zero mobility at R = 1.0 because there is no protonation of the sulfate ions over a wide pH range. This is supported by the fact that KPVS forms a 1:1 stoichiometric complex with PDDA even at pH 3.0 as mentioned in the colloid titration. Taking this into account, we may suggest that the overall dissociation of one or two KPVS ion(s) from the SPENC particle causes its positive mobility, the value of which is expected to be of the same order of magnitude as that of CPENG, as well as of the positively charged ampholytic polymer and nanogel as mentioned above. Nevertheless, in the solution, the SPENC particle having the surface with the KPVS ions bound both loosely and tightly seems to undergo a segmental dissociation-association reaction; therefore, the complex does not separate into the cationic particle and the polyanion. This is the reason that we did not detect the polyanion by the colloid titration at pH e pHr. In our view, such a phenomenon is similar to the fact that a macromolecular coil with charges, which behaves in sedimentation or diffusion as impermeable, should behave in electrophoresis as if it were free draining. Electrophoretic behaviors of ionic nanogels have theoretically been explained in terms of the following two models other than FDM: charged surface model (CSM) established by Pelton et al.27 in collaboration with Rowell, and charged hairy surface layer model (CHSLM) initially proposed by Ohshima et al.28 Let us consider the issue of the ELS behavior in the range of pH < pHr using CSM and CHSLM. Then, we see that the results cannot be described by the CSM in which the mobility (as the absolute value) is inversely proportional to Rh, because of a decrease of Rh with pH in the range of R > 0.8, that is, pH < 4.8 (see Supporting Information Figure S1). Also, it is seen that the surface of SPENC on which the KPVS ions are condensed is the reverse of the CHSLM. Even when such polyanions undergo a segmental dissociation so as to form a charged hairy surface layer, the mobility based on CHSLM should be negative. The theory based (27) Pelton, R. H.; Pelton, H. M.; Morphesis, A.; Rowell, R. L. Langmuir 1989, 5, 816. (28) Ohshima, H.; Makino, K.; Kato, T.; Fujimoto, K.; Kondo, T.; Kawaguchi, H. J. Colloid Interface Sci. 1993, 159, 512.
13586 DOI: 10.1021/la101852b
on FDM, in which the mobility is given on a segment composed of one or more monomer units having a fixed charge, allows us to understand why the mobility of SPENC varies depending on R. Finally, we look at the ELS results in the range of pH > pHr or R < Rr. The mobility change in this range can be explained in terms of the pH-induced deprotonation of the CPENG and of the dissociation of the whole (but not a segment) of the KPVS ion. Recall that the negative mobility of the complex is related to the KPVS anion remaining bound to the CPENG component. Such an anion-rich complex loses the positive charge with increasing pH as indicated by the mobility curve of CPENG, thereby causing an increase of the negative mobility. However, the amount of the dissociation of the complexed KPVS increases with increasing pH (see Table 2). As a result, the negative mobility increases around pH 6.6 (R ≈ 0.1) and becomes almost constant in the range of pH 7-8 (i.e., R ≈ 0.05-0). Upon the further increase of pH after the attainment of R = 0, the negative mobility of the complex tends to come closer to zero due to the dissociation of KPVS via the separation of the tightly linked ion-pair between the sulfate anion and the imidazole cation. Reassociation of KPVS Ions by the pH Change from 9.0 to 3.0. The SPENC dispersion whose pH had changed from 3.0 to 9.0 becomes a mixture of the dissociated KPVS ions (80 mol %) and anionic nanogel particles with the remaining undissociated KPVS ion. Thus, it is interesting and important to investigate the process of “reformation” of SPENC by the decrease in pH from 9.0 to 3.0. There are two ways to vary the pH leve: either a rapid or a gradual addition of HCl. The former (“method I”) with a pHjump provides the dispersion without causing the phase separation. In the latter (“method II”), we slowly lowered the pH of the redispersion up to the appearance of a slight turbidity around pHr (6.56) or Rr (0.11). After that, the pH was fixed at 5.8 for 1 h in the phase separation range (see a light-yellow zone in Figures 3a and 6, as well as in Supporting Information Figure S1), followed by lowering the pH to 3.0 with gentle stirring. Figure 7 shows the results of DLS, SLS, and ELS for the reformed complex dispersions obtained by methods I and II. The SLS measurements were performed at one concentration because the A2 value of SPENC was negligibly small (see Figure 1b). The complex prepared in method I cannot return to the original state, as indicated by a strong scattering intensity and a large Rh (107 nm) in DLS, as well Langmuir 2010, 26(16), 13579–13589
Doi and Kokufuta
Article
It should be discussed that the Rg (167 nm) of the intraparticle complex that resulted in method II is larger than that (117 nm) of the original SPENC. Recall that the SPENC particle redispersed at pH 9.0 has ca. 20% of the remaining undissociated polyanions. In addition, the dissociated KPVS ions bind again to the cationrich complex particle during the change of pH from 9.0 to 3.0, leading to an increase in the concentration of KCl (i.e., from 0.001 to 0.002 mol/L). Therefore, it is conceivable that the reformed intraparticle complex is not geometrically equivalent to the original SPENC. Nevertheless, the results obtained in method II would allow us to rewrite eq 1 as Gmþ þ aPn- / Gmþ(Pn-)a, meaning that Gmþ(Pn-)a undergoes the association and dissociation reactions in water media without added salts, when mþ is varied with the solution pH. Another important result obtained from the ELS is that the mobility (2.20 10-4 cm2/(V s)) of the complex aggregate formed in method I is slightly but significantly smaller than that (2.38 10-4 cm2/(V s)) of the original SPENC. This is similar to the case of the complex at rm = 1.1. However, note that the latter with U = 2.12 10-4 cm2/(V s) were the aggregated CPENG particles with a 10% excess of KPVS. The complex aggregate in method I resulted from a “mixture” containing 1:1 stoichiometric amounts of CPENG and KPVS; nevertheless, its mobility is smaller than that of the SPENC. This means that the number of the tightly linked ion-pair between the sulfate anion and the imidazole cation is smaller in the former (the original SPENC) than in the latter (the reformed complex in method I). As a result, it seems that the dissociation of KPVS ions from the complex reformed in method I becomes difficult under an applied electric field in ELS.
General Discussion and Conclusions
Figure 7. Results of DLS (a), SLS (b), and ELS (c) for the complexes formed from the SPENC redispersion (at pH 9.0) by reduction of pH to 3.0. The reduction of pH was performed by two ways: method I in which the pH was promptly varied so as not to cause the phase separation; and method II in which the phase separation takes place. In each panel, the results of SPENC in Figure 1 are shown for reference. For more details, see the text.
as by both large Rg (213 nm) and M w (1.70 108 g/mol) in SLS. In method II, however, the complex reformed provides the data of Rh (99 nm), M w (1.58 107 g/mol), and mobility (2.25 10-4 cm2/(V s)), the values of which are close to those of the original SPENC (see Table 1). In particular, it is of importance to note that an intraparticle complex forms in method II because its molar mass is less than twice that of CPENG, whereas in method I the resulting complex is an aggregate composed of about 11 particles of SPENC. Langmuir 2010, 26(16), 13579–13589
There have been two types of water-soluble or dispersible PECs: nonstoichiometric PECs and 1:1 stoichiometric PECs having a core-shell or corona structure. On the basis of the previous20 and the present experiments, we have demonstrated that the SPENCs under investigation fall into the category of a watersoluble and 1:1 stoichiometric PEC. Thus, it is of importance to discuss the water dispersible nature of SPENCs, in comparison with those of the above two types of PECs. As was mentioned in the Introduction, the water solubility or dispersibility of nonstoichiometric PECs is due to an excess charge of either anionic or cationic component. In the case of 1:1 stoichiometric PECs of the core-shell (or corona) type, the outer shell or corona layer of the complex particles contributes to their water solubility. The structural difference between the coreshell type PEC and the SPENC is the location of the distributed anion-cation pairs. The former has the ion pairs condensed within the internal core, whereas in the latter case the ion pairs are clustered on the surface. In general, colloidal stability depends upon the balance of attraction which causes aggregation, and steric or electrostatic forces that oppose aggregation. The SPENC is a kind of cationic colloid on which surface a 1:1 stoichiometric amount of polyanions is bound to form a cluster of ion-pairs. Such a particle is generally unfavorable to colloid stability in water. However, this is not the case in the SPENC particles prepared in the previous20 and the present studies. Then, we can pay attention to an association-dissociation equilibrium between the host and guest polyions in nonstoichiometric molecular PEC systems. Indeed, a polyelectrolyte exchange reaction has been observed in the previous SPENC-water system.20 There seems to be two types of dissociation reactions in the PEC system: One is based on the whole of a polyion chain, and the other is on a part (segment) of the polyion. The former can be suggested (or evidenced) by the observation of polyelectrolyte DOI: 10.1021/la101852b
13587
Article
exchange reactions.20 For the latter, however, there are large difficulties in experimental observations. Therefore, to discuss this issue, we have to rely on carefully collected experimental data. In this study, we carried out the monitoring of the reduction of the imidazole-based cationic charge in the nanogel component (CPENG) of SPENC as a function of pH. It has been found that, during the charge reduction process, the complex undergoes aggregation, followed by a phase separation. The aggregation starts at about 25% of the charge reduction (R ≈ 0.75), and the phase separation takes place when almost half of the charge is eliminated (R = Rp ≈ 0.48). However, at R ≈ 0.11 (Rr), the redispersion of the phase-separated complex was observed. It should be noted that the dissociated free polyanions are not detected in the aqueous SPENC solution before the phase separation but are detected in the redispersed solution. Also noted is that when the pH (9.0) of the redispersed solution was slowly decreased to the original level (pH 3.0) by the gradual addition of HCl so as to cause the phase separation, we obtained an intraparticle complex whose physical quantities are close to those of the initial SPENC. A detailed examination of these findings clearly indicates that the dissociation-association reactions of both the chain’s whole and segment types take place on the surface of a SPENC particle, depending on the ionization state of the cationic gel component. This aspect is schematically illustrated as
Doi and Kokufuta
rm < 0.2. In the range of rm from 0.2 to 1.0, however, the holding of KCl within the complex is about 20% and independent of rm. This would provide a key for understanding why the association-dissociation reactions take place in the SPENC system. The report is under preparation and will be submitted soon.
Experimental Section Materials. All chemicals were obtained from commercial sources: NIPA from Kojin Chemical Co. (Tokyo, Japan); N,N0 methylenebisacrylamide (Bis, cross-linker), ammonium persulfate (APS; initiator), and sodium dodecylbenzene sulfonate (NaDBS, surfactant) from Wako Pure Chemical Co. (Osaka, Japan); and 1-vinylimidazole (VI) from Tokyo Chemical Industry Co. (Tokyo, Japan). NIPA was recrystallized from a 65:35 mixture of hexane and benzene. VI was twice distilled under reduced pressure (9 mm Hg at 45 C) immediately before use. All pregel solutions were prepared with distilled water passed through a Milli-Q water purification system. We used in this study two polyelectrolyte samples: KPVS for the complex formation and PDDA for colloid titration. Both polymers (lyophilized powder) were of analytical grade for colloidal titration and obtained from Wako Pure Chemical Co. Several physical quantities are as follows: M w (measured in 0.2 M KCl and at 25 C by SLS), 4.19 105 g/mol for KPVS, and 1.42 105 g/mol for PDDA; charge density, 6.16 mmol/g (based on the sulfate group) for KPVS and 6.18 mmol/g (based on the quaternary ammonium group) for PDDA; Rh (in 0.2 M KCl and at 25 C), 16 nm for KPVS and 14 nm for PDDA; and Rg (under the same conditions as Rh measurements), 33 nm for KPVS and 37 nm for PDDA. A nonpolyelectrolyte nanogel consisting of Bis-cross-linked NIPA used as a reference sample in the ELS measurements was a stock which was prepared in our previous study.24 The Rh value was 54 nm, as measured in pure water at 25 C by DLS. Synthesis of Cationic Polyelectrolyte Nanogel (CPENG) and Its Complexation with KPVS. We adopted for the synth-
In some more detail we believe that the aggregation, followed by the phase separation, takes place via the same mechanism that amphoteric macromolecules in water media associate with each other around an isoelectric point via electrostatic interactions between their fixed charges of opposite sign. The number and length of the segments that undergo dissociation-association on the surface of a SPENC particle become the dominant factor for the aggregation. In conclusion, the present study illustrates that the water dispersibility of a 1:1 stoichiometric complex between a cationic nanogel and linear polyanion can be discussed in terms of the association-dissociation reactions between both of the polyelectrolyte components. Such reactions have long been considered as the key factor for the water solubility of nonstoichiometric PECs; however, we can suggest that PECs from polyelectrolyte nanogels (cross-linked polyelectrolytes) become water-soluble or dispersible, independent of whether the mixing ratio of anionic to cationic charges is stoichiometric (R = 1) or nonstoichiometric (R > Rp and R < Rr), and also even in the “absence” of low-molecularweight salts. This would provide a significant base for the molecular design of water-soluble PEC architectures. Another work focused on the formation mechanism of the present SPENC has been finished during the reviewing process of this manuscript. It has become apparent that the complexation given by eqs 1 and 2 brings about a complete release of KCl at 13588 DOI: 10.1021/la101852b
esis of CPENG an aqueous redox polymerization initiated by APS in the presence of NaDBS (surfactant) and of Bis (crosslinker) at a temperature above the lower critical solution temperature (32 C) of NIPA polymers. The polymerization was carried out using 300 mL of aqueous NaDBS solution (10 mM) containing NIPA (56 mM), VI (24 mM), and Bis (3.9 mM). The O2-free monomer solution was placed in the usual separable flask (500 mL) equipped with a cooler and a magnetic stirrer. To remove oxygen well, nitrogen gas was continuously supplied above the surface of the solution for 1 h before the reaction. The polymerization reaction was initiated by adding 1 mL of aqueous O2-free solution of APS (7.5 w/w %), allowed to continue several hours (2-3 h), and terminated by blowing oxygen through the reactor. After that, residual monomers and NaDBS were removed from the resulting reaction mixture by a dialysis method using a Spectra/Por CLC500 tube with a molecular weight cutoff of 100 000. Further purification was carried out by passing the dialyzed solution through a mixed bed of anion and cation exchange resins. The purified nanogel dispersion was then lyophilized for 3 days. The complexation was carried out at pH 3.0 and at 25 C in the absence of low-molecular-weight salts (i.e., salt-free system). The KPVS solution (1.25 mmol/L based on the fixed charges) was added into the dispersion (50 mL) with a fixed concentration (0.025 g/L) of positively charged CPENG particles, using an automatic titrator (model ART-3, Hirama Co., Tokyo, Japan). The total volume of KPVS addition (1.74 mL for the stoichiometric complex, SPENC) was controlled within a precision of (0.001 mL. To avoid pH change during the complexation, the KPVS solution was previously adjusted to pH 3.0 with HCl. The dispersion of two kinds of the complexes obtained at rm = 1.0 and 1.1 was then adjusted to a desired pH with aqueous KOH or HCl solutions (1 mM to 1 M). Because of the use of the different acid or Langmuir 2010, 26(16), 13579–13589
Doi and Kokufuta base concentrations, the decrease in the complex concentration caused by the pH adjustment was less than 2% at the maximum. All the complex solutions before as well as after the pH adjustment were allowed to stand for more than 12 h. Light Scattering Measurements. DLS was carried out at a scattering angle (θ) of 90 using an Otsuka DLS 7000 apparatus (Osaka, Japan) equipped with an argon ion laser (75 mW; NEC model GLG-3112) as the light source. We analyzed the autocorrelation functions with the CONTIN program. The same apparatus as used in DLS was employed for SLS measurements. The calibration was made by using pure (>99.5%) toluene, and the optical alignment was ensured by less than (2% over the range of 30 e θ e 140. We determined the Rayleigh ratio (Rθ) as the average of five different measurements for the same sample. Then, Zimm plots were obtained with a SLS data fitting program provided by Otsuka Co. Ltd., Osaka, Japan. The change in the refractive index with concentration, which has been written as (dn/dc)g for nanogel and as (dn/dc)p for KPVS, was measured at 25 C using an electrophotometric differential refractometer (Otsuka model DRM-1021). The measurements were carried out using vertically polarized light from an iodine arc with a spectrum filter (wavelength = 488 nm). The instrument constant was obtained by using a standard KCl solution (0.01481 g/mL) with a known dn/dc value of 0.1344. ELS was made at a fixed scattering angle of 20 with an Otsuka ELS-6000 apparatus (Osaka, Japan). The electric field was applied at a constant current of 10 mA. The temperature of the thermostatted chamber was maintained at 25 C. The mobilities obtained in this work were repeatable to within less than 5%. Potentiometric Titration. A CPENG dispersion (2 g/L; 50 mL) was titrated with 0.1 N KOH at 25 C under nitrogen. Several modifications were made to a TOA model AUT-501 titration apparatus to allow the following operations to be made
Langmuir 2010, 26(16), 13579–13589
Article with the aid of a computer: (i) An accurate volume (0.05 mL) of titrant can be automatically added when a pH remains constant for 30 min within the range of (0.01 unit (measuring limit), the condition of which was determined from careful preliminary experiments. (ii) Error due to drift of electrode potential with time is avoided by using the same pair of pH electrodes; one for continuously measuring the solution pH, and the other for confirming the measured values by periodic calibration (at least at half-day intervals) using two standard solutions (phthalate buffer for pH 4.01, citrate/phosphate buffer for pH 6.86). (iii) All the data can be recorded by the computer for later analysis. Colloid Titration. The sample solutions (50 mL) containing different amounts of KPVS and at different pH levels were titrated with the PDDA titrant (2.5 10-3 mol/L) whose solution conditions were exactly adjusted to those of the sample. For example, for the sample obtained by the filtration of a complex dispersion at pH 8.2, the control sample at pH 8.2 was prepared from the KPVS solution (4.2 10-5 mol/L, 50 mL) at pH 3.0 by the pH adjustment with a very slight volume (less than 0.1 mL) of a KOH solution and then titrated with the PDDA titrant whose pH was also adjusted to 8.2. The titration was performed at room temperature using a Hirama automatic titrator (model ART-3). The colloid titration curve was then given by the change of absorbance at 460 nm with the titrant volume.
Acknowledgment. This work was supported in part by Grantsin-Aid for Scientific Research to E.K. from the Japan Society for the Promotion Science (No. 20550183). Supporting Information Available: Figure S1 showing changes of hydrodynamic radii of SPENC and CPENG as a function of pH or R. This material is available free of charge via the Internet at http://pubs.acs.org/.
DOI: 10.1021/la101852b
13589
