On an Intraparticle Complex of Cationic Nanogel with a Stoichiometric

On an Intraparticle Complex of Cationic Nanogel with a. Stoichiometric Amount of Bound Polyanions. Kazuyoshi Ogawa, Seigo Sato, and Etsuo Kokufuta*...
0 downloads 0 Views 241KB Size
Langmuir 2007, 23, 2095-2102

2095

On an Intraparticle Complex of Cationic Nanogel with a Stoichiometric Amount of Bound Polyanions Kazuyoshi Ogawa, Seigo Sato, and Etsuo Kokufuta* Graduate School of Life and EnVironmental Sciences and Institute of Applied Biochemistry, UniVersity of Tsukuba, Tsukuba, Ibaraki 305-8572, Japan ReceiVed June 20, 2006. In Final Form: September 26, 2006 A polyelectrolyte nanogel (PENG) particle consisting of lightly cross-linked terpolymer chains of N-isopropylacrylamide, acrylic acid, and 1-vinylimidazole has positive charges in an aqueous medium at pH 3 due to protonation of the imidazole groups, and thereby forms a polyelectrolyte complex with the linear polyanion, potassium poly(vinyl alcohol) sulfate (KPVS). It has been demonstrated that the hydrodynamic radius (Rh), by dynamic light scattering (DLS), and the radius of gyration (Rg), by static light scattering (SLS), of the complex particles are smallest at ∼1:1 mixing ratio (rm) of anions to cations, in the absence of simple salts such as KCl (Langmuir 2005, 21, 4830). Here, we aimed to study the nature of the complex formed at rm ) 1 and examined the complex formation process by electrophoretic light scattering (ELS). It was found that the mobility of the cationic PENG with a stoichiometric amount of bound KPVS anions (i.e., the complex formed at rm ) 1) is positive but not zero at 25 °C. This was also the case when the complex was examined by ELS at 45 °C, where DLS and SLS show a temperature-driven collapse of the complex. We thus assumed that (a) electroneutrality is maintained in the complex particle with the aid of counterions, but (b) the complex is highly polarizable, and hence (c) during ELS the KPVS anions would dissociate in part from the complex. This hypothesis was supported by the following results: (i) Mixing complexed and uncomplexed PENG particles at different ratios brings about an increase in Rh and a decrease in the light scattering intensity of the complex at the same time, suggesting a polyelectrolyte exchange reaction. (ii) The same phenomenon is seen when poly(diallyldimethylammonium chloride) (PDDA as a polysalt) is added to the complex dispersion, meaning that the PDDA takes out the KPVS from the complex to form a stable PDDA-KPVS complex. (iii) Upon addition of KCl, the complex undergoes little change in Rh (62-67 nm) at a salt concentration (Cs) e 0.02 M, aggregates with each other at Cs from 0.03 to 0.2 M, and becomes water-soluble at Cs > 0.2 M. (iv) The Rh (78 nm) of the soluble complex at Cs from 0.3 to 0.5 M is larger than that at Cs < 0.02 M, suggesting dissociation of the KPVS ions. (v) Complexation between KPVS and PDDA as mentioned in (ii) is facilitated in the presence of 0.01 M KCl.

Introduction

Gm+ + aPn- f Gm+‚(Pn-)a

In contrast to hard colloids such as proteins and charged polymer latexes, polyelectrolyte nanogel (PENG) particles undergo a large change in size depending on their ionization state,1-6 as observed in polyelectrolyte bulk gels.7 Our previous study8 dealt with polyelectrolyte complex formation of a positively charged PENG with a linear polyanion in aqueous solutions containing different concentrations (Cs) of KCl, as a function of the mixing ratio (rm) based on moles of anions to cations. It has become apparent that there is a critical mixing ratio (cmr) at which both size and molar mass of the resulting particles abruptly increase, indicating the formation of an “interparticle” complex. The cmr value at Cs e 0.01 M () mol/L) was observed to be at rm around unity. At rm < cmr, the measured molar mass of the complex agreed with that of a PENG particle to which the calculated amount of the polyanion is bound, meaning that the formation of an “intraparticle” complex (Gm+‚(Pn-)a) follows the stoichiometry of the reaction,

where Gm+ denotes a PENG particle with +m charges, Pn- is a polyanion chain with -n charges, and a ) (m/n)rm. During the process of intraparticle complex formation, both hydrodynamic radius (Rh), by dynamic light scattering (DLS), and mean square radius of gyration (Rg), by static light scattering (SLS), were found to decrease with an increase of rm and become minimum at rm ) 1, the point at which a stoichiometric amount (a mol) of Pn- ions was added to the dispersion of Gm+ particles. This means the gel collapse at that value rm is due to complexation before the formation of an interparticle complex, which is an aggregate of intraparticle complexes. Another important feature of our complexation system8 was that a polyelectrolyte exchange reaction occurs when uncomplexed cationic PENG (Gm+) is mixed with the complex (Gm+‚(Pn-)a) prepared at rm ) 1 (i.e., a ) m/n). We do not expect tight ion pairing between all the anionic and cationic charges in the complex. Rather, we have to consider the polyanions in the complex as “loosely” bound because Pn- has a very small and uniform spacing between anionic charges, while cationic charges of Gm+ would be distributed all over the particle, which consists of lightly cross-linked polymer chains and has an obscure (or a diffuse) surface due to dangling chains (see refs 4 and 8). Within the intraparticle complex formed at rm ) 1.0, bulk electroneutrality should be maintained by counterions, but there must be local regions of excess positive and negative charges. Therefore, it seems that the complex should be highly polarizable. Here we aimed to study in detail the nature of polyanionbound PENG complex formed at rm ) 1, in particular, in terms

* To whom correspondence should be addressed. Fax: 81-298-53-4605. E-mail: [email protected]. (1) Pelton, R. AdV. Colloid Interface Sci. 2000, 85, 1. (2) Suzuki, H.; Wang, B.; Yoshida, R.; Kokufuta, E. Langmuir 1999, 15, 4283. (3) Ogawa, K.; Wang, B.; Kokufuta, E. Langmuir 2001, 17, 4704. (4) Ogawa, K.; Nakayama, A.; Kokufuta, E. Langmuir 2003, 19, 3178. (5) Ogawa, K.; Nakayama, A.; Kokufuta, E. J. Phys. Chem. B 2003, 107, 8223. (6) Miyake, M.; Ogawa, K.; Kokufuta, E. Langmuir 2006, 22, 7335. (7) Kokufuta, E. Langmuir 2005, 21, 10004. (8) Ogawa, K.; Sato, S.; Kokufuta, E. Langmuir 2005, 21, 4830.

10.1021/la061767t CCC: $37.00 © 2007 American Chemical Society Published on Web 01/12/2007

(1)

2096 Langmuir, Vol. 23, No. 4, 2007

of the removal of the polyanions via (i) the exchange reaction with different amounts of the uncomplexed PENG, (ii) the complexation with polycations, and (iii) the addition of low molecular weight salt. The experimental advantage of our system is that it allows an accurate monitoring of items (i)-(iii) through changes of both scattering intensity and size (Rh and/or Rg) using light scattering techniques. We expect to obtain information about the behavior of polyions that were bound to an oppositely charged PENG particle, the surface of which is covered with ionic chains. We also expected from this study to obtain a basic knowledge relevant to similar changes that have been found in many biological systems; for example, see ref 9. This is the reason we used a polyampholytic terpolymer nanogel under pH conditions where the basic group is fully protonated. A study to look at the stability of the same nanogel complex as a function of pH has already been started. Experimental Section Materials. We used a polyampholytic terpolymer nanogel abbreviated as G(1/1) as described in ref 4. This PENG sample is composed of 14 mol % acrylic acid (AAc) and 12 mol % 1-vinylimidazole (VI) in an N-isopropylacrylamide (NIPA) network cross-linked by N,N′-methylenebisacrylamide. The important physical quantities are weight-average molar mass (M h w), 1.49 × 107 g/mol; charge densities, 1.36 mmol/g for anions due to carboxyl groups and 1.12 mmol/g for cations due to imidazole groups; and isoelectric point (pI), 5.3 (in 0.01 M KCl solution). In addition, Rh and Rg are 87 and 81 nm, respectively, as determined in 0.1 M KCl at pH ) pI and at 25 °C. Potassium poly(vinyl alcohol) sulfate (KPVS) and poly(diallyldimethylammonium chloride) (PDDA) were used as strong polyanion and polycation, respectively. Both polymers (lyophilized powder) were the same sample as used in our previous studies.2,7,8 Several physical quantities are as follows: M h w (in g/mol), 4.19 × 105 for KPVS and 1.42 × 105 for PDDA; charge density, 6.16 mmol/g (due to sulfate groups) for KPVS and 6.18 mmol/g (due to quaternary ammonium groups) 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. Preparation of Intraparticle KPVS-PENG Complexes. The complexation was carried out at pH 3 and at 25 °C in the absence of KCl (Cs ∼ 0). The KPVS solution (0.202 g/L; equivalent to 1.25 × 10-3 mol/L based on the ionizable groups) was added into the dispersion (50 mL) with a fixed concentration (0.025 g/L) of positively charged PENG particles, using a Hirama automatic titrator (model ART-3). To avoid pH change during the titration, the KPVS solution was previously adjusted to pH 3 with HCl. The mixing ratio (rm) was then given as rm ) (CKPVSDKPVS/CPENGDPENG)(VKPVS/VPENG), where C is the concentration (in g/ mL), D is the charge density (in mol/g), and V is the volume (in mL) of the species indicated by the subscripts. DLS. The measurements were carried out at a scattering angle (θ) of 90° using an Otsuka DLS 7000 apparatus (Osaka, Japan) equipped with a He-Ne laser as the light source. We analyzed the autocorrelation functions with the CONTIN program. SLS. 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 characterized by having a normalized intensity variation less than (2% over the range of 30° e θ e140°. We determined the Rayleigh ratio (Rθ) as the average of five different measurements for the same sample and at the same θ. The molar mass (M h x) and Rg of the KPVS-PENG complexes obtained at different rm were estimated as reported in ref 8, using the following relations: (9) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell, 4th ed.; Garland Science, Inc.: New York, 2002.

Ogawa et al.

(

)

KCx 1 16π2 2 2 θ ) 1+ Rg sin + 2A2Cx Rθ 2 M h 3λ2

(2a)

K ) (4π2/λo4NA)no2(dn/dc)x2

(2b)

x

(dndc) ) 1 +1 β(dndc) x

+ PENG

β dn 1 + β dc

Cx ) (CPENGVPENG + CKPVSVKPVS)

( )

KPVS

1 (VPENG + VKPVS)

(2c) (2d)

Here, λ in eq 2a and λo in eq 2b denote the wavelength of the light in the medium and in a vacuum, respectively. In addition, A2 is the second virial coefficient, NA is Avogadro’s number, no is the refractive index of the medium, and (dn/dc)x is the change in the refractive index of the complex with its concentration (Cx), which is given by eq 2d. As mentioned in ref 8, accurate determination of (dn/dc)x was very difficult due to a strong scattering of light from the complex particles; thus, we calculated (dn/dc)x using eq 2c in which β (∼(CKPVSVKPVS/CPENGVPENG)) denotes the mass ratio of bound KPVS to a PENG particle. Note that in our analytical method the Rayleigh-Gans approximation is tacitly assumed; however, this assumption is permissible when Rg is smaller than 258 nm for the He-Ne laser (see ref 6). ELS. The measurements were 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 thermostated chamber was maintained at either 25 or 45 °C.

Results and Discussion Formation of KPVS-PENG Complex Particles Examined by DLS and ELS. At the beginning of the present study we combined ELS and DLS to see how both size and charge of the complex particle vary depending on rm. As can be seen from Figure 1, Rh decreases with increasing rm at rm < cmr, in particular, at rm between 0.5 and 1.0. This size change has been interpreted as a result of the neutralization of the cationic nanogel charges due to the binding of the polyanions.8 From the ELS experiments, a very gradual decrease in the mobility (U in cm2‚s-1‚V-1) was observed at rm < cmr; however, U does not become zero until rm > cmr. When considering that U is proportional to the net charge of PENG particles, one would raise a question about our interpretation for the Rh change from DLS. Therefore, we start from this point our discussion about the nature of the intraparticle KPVS-PENG complex. To make clear why U > 0 at rm ) 1 at which Rh becomes minimum, we studied first the Rh and U changes upon addition of KOH, instead of KPVS, into the PENG dispersion. The Rh and U were plotted against the molar ratio (R) of KOH to a total of the protonated imidazole cations (see Figure 2). Our PENG is a polyampholyte with both carboxyl and imidazole groups; thus, the R dependencies of Rh and U in Figure 2 should provide good control for the rm dependencies of Rh and U in Figure 1. In addition, to obtain a better understanding of the results in Figure 2, the pH dependence of the ionization degree (Ri) for the carboxyl and the imidazole groups bound to NIPA-based nanogels would be needed. For this purpose the pH titration curves are presented in Figure 3, using the data obtained by the titration of the VI-containing PENG with HCl and of the AAc-containing PENG with KOH.10 The results in Figure 3 show that the addition of KOH to raise solution pH causes deprotonation of both the (10) Potentiometric titrations were performed at 25 °C and at Cs ) 0 M using two PENG samples synthesized in our previous studies: (a) Ito, S.; Ogawa, K.; Suzuki, H.; Wang, B.; Yoshida, R.; Kokufuta, E. Langmuir 1999, 15, 4289 and (b) Suzuki, H.; Wang, B.; Yoshida, R.; Kokufuta, E. Langmuir 1999, 15, 4283.

Intraparticle Complex of Cationic Nanogel

Figure 1. Formation of KPVS-PENG complex studied by DLS and ELS in the KCl-free system as a function of rm. Open circles and circles with cross represent hydrodynamic radius (Rh); the former is the data reported in ref 8, and the latter is from the present DLS measurements which were performed to test reproducibility. Filled circles show electrophoretic mobility (U). At rm > cmr (gray shaded area) there is a lack of reproducibility of DLS data due to the aggregation of intraparticle complexes. At rm ) 1, however, we obtained Rh ) 67 ( 10 nm from three measurements (including one previous measurement) with three separately prepared samples. Conditions of the complex formation are described in detail in the Experimental Section.

Figure 2. Changes of Rh (open circles) and U (filled circles) upon addition of KOH into the PENG dispersion in the absence of KCl. R is equivalent to rm in Figure 1 when we consider KOH instead of Pn- in eq 1; thus, R ) an/m where an is the moles of KOH added. Square and triangle plots are the Rh data obtained at KCl concentrations (Cs) ) 0.01 and 0.1 M, respectively, because the PENG particles aggregate with each other at R ) 1 and at Cs ) 0 M (see ref 4).

imidazole cations and the carboxyl groups. The former eliminates positive charges, whereas the latter generates negative charges. In the case of polyampholyte nanogels, these changes take place at the same time, depicted below as

and

The ionization of a COOH group given by eq 4 is accompanied by the formation of an “ion pair” with an imidazolyl cation bound to the nanogel. This neutralizes the imidazole cationic

Langmuir, Vol. 23, No. 4, 2007 2097

Figure 3. Relations between pH and the ionization degree (Ri) for the carboxyl group (filled circles) and the imidazole group (open circles) obtained by potentiometric titrations10 of the NIPA-VI and NIPA-AAc copolymer nanogels in the absence of KCl. The expected pI value for the present ampholyte nanogel is indicated by arrow b in the figure. Arrows a and c show apparent pK values for the ≡NH + and -COOH groups, respectively; however, note that the proper expression of pK for PENG particles should be as a function of R as used for polyacids and polybases (see (b) in ref 10).

charges. Therefore, increasing R for R < 1.0 leads to a decrease in U according to eq 3 as well as to eq 4, and U becomes zero at R ) 1.0 corresponding to an isoelectric point (pI) (see Figure 2). The validity of these interpretations for the mobility change in Figure 2, which are based on the results in Figure 3, can be supported by the fact that the pI (∼5.3 from ELS at Cs ) 0.01 M; see Experimental Section) of the present G(1/1) nanogel is close to the pH (∼5.7; as the expected pI) obtained from the intersection of the pH vs Ri curves in Figure 3. As a result, we can see a good correlation between the dependencies of U and Rh on R in Figure 2; that is, the PENG shrinks at R < 1.0 because of the elimination of positive charges but swells at R > 1.0 because of the generation of negative charges. Note that the present G(1/1) PENG particles aggregate with each other at pH ) pI (i.e., R ∼ 1.0) in a salt-free system or under conditions of low Cs (see ref 4); therefore, we cannot measure the Rh at R ) 1.0 by DLS. Nevertheless, the interpolation of the R vs Rh curve in Figure 2 implies the existence of a minimum Rh at R ∼ 1.0. In addition, the Rh measured in the presence of KCl (see square and triangle symbols in Figure 2) became minimum at R ∼ 1.0 at which U ) 0. In conclusion, a complete neutralization of positive charges bound to the present PENG can be demonstrated by the observation of zero mobility as well as a minimum Rh. Taking the above into account, we are forced to assume that during ELS the KPVS anions would dissociate in part from the KPVS-PENG complex formed at rm ) 1.0 at which the stoichiometric amount of the polyanions binds to a cationic PENG particle so as to maintain electroneutrality. Note that this complex is equivalent to Gm+‚(Pn-)a in eq 1 where a ∼ m/n; thus, we refer to it as “stoichiometrically charge-neutral” complex (SCNC) unless otherwise noted. Moreover, the dissociated KPVS anions may not be detected by ELS because of their weak light-scattering intensity, in particular, in a salt-free system containing the nanogel particles with a very strong scattering intensity. Effect of Temperature on Size and Mobility of SCNC. In the next stage of ELS study, we paid attention to the fact that our PENG is composed of the NIPA chain and therefore undergoes a collapse in diameter upon heating due to hydrophobic interaction (e.g., see refs 1 and 4). As can be seen from Figure 4, this is the case for the SCNC; that is, both Rh and Rg decrease with increasing temperature from 25 to 55 °C. Therefore, it would be interesting

2098 Langmuir, Vol. 23, No. 4, 2007

Ogawa et al.

Figure 4. Thermal collapse of SCNC (Gm+‚(Pn-)a where a ∼ m/n) particles. SLS data showed that the molar mass (M h x ∼ 1.80 × 107 g/mol) of the complex is independent of temperature within experimental errors.

Figure 5. Results of ELS for the complex sample used in Figure 4 at 25 and 45 °C. The mobilities (U) were determined from each peak shown in the figure.

to see whether the binding between the KPVS and the PENG in the SCNC becomes tight or loose upon heating. We performed the ELS measurements at 25 and 45 °C. As shown in Figure 5, the U (3.12 × 10-4 cm2‚s-1‚V-1) at 45 °C becomes 1.7 times larger than that (1.85 × 10-4 cm2‚s-1‚V-1) at 25 °C. To interpret this difference in terms of ionic charges of the complex, we may take notice of two previous ELS studies of NIPA-based anionic nanoparticles.11,12 Both studies have demonstrated an increase in the U but a decrease in the Rh with a rise in temperature above the cloud point (∼33 °C) of poly(NIPA) in water. One is the study by Pelton et al.11 who used a NIPA nanogel particle with bound anions resulting from an initiator, the other is by Ohshima et al.12 using a nanoparticle composed of poly(styrene-co-ethylhexyl methacrylate) core whose surface is covered with a thin NIPA gel layer. In the former in which all of the charges were assumed to be concentrated on the particle surface, the results were interpreted as an increase in the surface charge density (N) due to the temperature-induced decrease in the Rh. In the latter, the authors attributed the increase of U to both an increase of N and a decrease of “softness parameter” 1/λf (λf is the square root of the ratio of the frictional coefficient of the surface layer to the viscosity of medium) caused by the collapse of the NIPA gel surface layer. Even when both studies are considered, in our results in Figures 4 and 5 there is (11) Pelton, R. H.; Pelton, H. M.; Morphesis, A.; Rowell, R. L. Langmuir 1989, 5, 816. (12) Makino, K.; Yamamoto, S.; Fujimoto, K.; Kawaguchi, H.; Ohshima, H. J. Colloid Interface Sci. 1994, 166, 251.

Figure 6. Polyelectrolyte exchange reaction in the KPVS-bound PENG system demonstrated by DLS measurements: Time courses of intensity (a) and Rh (b) are shown as a function of the PENG: SCNC ratios (i.e., m:an in eq 1): 1:1 (open squares, cited from Figure 4b in ref 8); 3:1 (filled circles); and 9:1 (open circles). Each of the broken lines in (a) and (b) are obtained from the DLS for the KPVS-PENG mixtures at different rm (see Figure 1), the value of which is expected to provide both intensity and Rh of the Gm+‚(Pn-)a complex resulting from the polyelectrolyte exchange reaction under a given PENG:SCNC ratio.

no reason for suggesting that the binding between the polyanion and the gel particle in the SCNC becomes tight or stable under conditions where the complex particle was fully collapsed. KPVS Exchange Reaction between KPVS-Bound and KPVS-Free PENG Particles. Our tentative conclusion from the previous sections may be summarized as follows: (i) Bulk electroneutrality in the SCNC is maintained by counterions, but (ii) the polyanions in the complex loosely bind and hence (iii) the ion pairs are labile. Such a complex appears to have the ability to undergo a polyelectrolyte exchange reaction with other polyelectrolytes or PENG particles. Indeed, we have briefly reported that the Pn- exchange takes place between Gm+ (Rh ∼ 158 nm) and Gm+‚(Pn-)a (a ∼ m/n) (Rh ∼ 67 ( 10 nm) in their 1:1 mixture.8 This process could be monitored as a function of time using DLS. It was found that a hydrodynamic radius distribution (f(Rh)) curve obtained with the CONTIN program is monomodal, and 12 h after mixing of components the distribution curve comes close to that for rm ) 0.5. To make more clear the KPVS exchange reaction in the KPVSbound PENG system, mixtures with different PENG:SCNC ratios (m:an ) 1:1, 3:1, and 9:1) were examined as a function of time. In our DLS measurements, at least 10 min is needed to obtain the autocorrelation function, from which Rh can be estimated by the CONTIN analysis. In contrast, the scattering intensity (Is ∼ ∫0∞f(Rh)dRh) can be recorded almost continuously after the mixing of the two sample dispersions. We thus show in Figure 6 the time courses of Rh and Is. The Is decreases rapidly at first, then slowly, and finally levels off to a certain value. In contrast,

Intraparticle Complex of Cationic Nanogel

the Rh increases with time and reaches a constant value. These clearly indicate that the SCNC undergoes a swelling change upon addition of the PENG with cationic charges. Each of the asymptotic Is values agrees well with the ones for (see broken line in Figure 6a) of the Gm+‚(Pn-)a complexes prepared at different rm: 0.1 (a ) 0.1m/n), 0.25 (a ) 0.25m/n), and 0.5 (a ) 0.5m/n). The values of Rh (∼158 nm) of the complex prepared at these rm almost remain unchanged (see Figure 1), and hence in Figure 6b the steady state Rh value is within 155-160 nm. Although there is a distribution of the data in Figure 6b, the results from previous8 and the present studies may indicate that the KPVS exchange reaction takes place between the PENG and the SCNC. Effects of Polysalt and Low Molecular Weight Salt on SCNC. From the results in the above sections, one would be interested in looking at aspects of the dissociation of the bound KPVS ions from the SCNC upon addition of other electrolytes. For this, we used PDDA as a polysalt and KCl as a simple salt. The former has been well-known to form a stable polyelectrolyte complex with KPVS in aqueous media,13 and the latter was found in our previous study8 to hinder the formation of KPVS-PENG complexes at concentrations >0.5 M. Figure 7 shows the results of DLS for mixtures of SCNC and PDDA. The measurements were performed 2 days after mixing, to obtain data at an equilibrium state in which the KPVS dissociation from the complex had been completed. The mixing ratio was based the molar ratio of PDDA-bound dN+) charges versus KPVS-bound -OS(dO)2O- charges in the SCNC. To distinguish this ratio from rm, we represent it as ma. As can be seen from Figures 7a and 7b, an increase in ma leads to both an increase of Rh and a decrease of the scattering intensity, indicating the swelling of the complex particle as mentioned in the above section. This is due to the dissociation of the KPVS ions from the SCNC to form a stable KPVS-PDDA complex, the formation of which can be supported by the following (i) a peak in the Rh range of 10-20 nm is reproducibly observed from the CONTIN analysis and (ii) a very diluted aqueous solution containing a KPVS-PDDA complex shows a peak in the same Rh range. CONTIN analysis of DLS data showed a bimodal profile for all of the mixtures examined; thus, one might argue against the application of SLS to such a mixture. As shown in Figure 8, however, the plots of KCx/R(θ) vs sin2(θ/2) give a good straight line at ma e 0.5 over the range of θ from 30° to 140°. At ma ) 1.0 there is a lack of linearity at θ < 70°, but we were able to estimate both M h x and Rg with a curve-fitting program from our SLS equipment. Then, it was found that all of the M h x values at different ma levels are between the molar masses14 of the KPVSfree PENG (rm ) 0 and ma ) 0) and of the SCNC (rm ) 1 and ma ) 0). Also found is the increase in Rg with increasing ma. This is very similar to the ma-dependent change of Rh. As a result, the swelling of the SCNC particle due to the dissociation of the KPVS ions upon addition of PDDA can also be supported by SLS. It is worth noting that in this dissociation process aggregation did not take place. Another important observation from Figures 7 and 8 is that at ma ) 1.0 both Rh and Rg of the SCNC do not return to those of the original PENG (i.e., Gm+). This means an incomplete (13) For example, see Kokufuta, E. et al. (a) Nihon Kagaku Kaishi (in Japanese) 1976, 1335. (b) Macromolecules 1981, 14, 1178. (c) J. Appl. Polym. Sci. 1981, 26, 2601. (e) Biotechnol. Bioeng. 1988, 31, 382. (f) Prog. Polym. Sci. 1992, 17, 647. (14) The M h w of Gm+ as HCl salt (i.e., KPVS-free PENG) was 1.49 × 107 g/mol as determined by SLS (see ref 4) and the M h x of Gm+‚(Pn-)a (a ∼ m/n) was 1.74 h x agrees well with the result × 107 g/mol as calculated by eq 1. This calculated M (1.80 × 107 g/mol) obtained by SLS (see the data at ma ) 0 and at rm ) 1 in Figure 8a).

Langmuir, Vol. 23, No. 4, 2007 2099

Figure 7. Results of DLS for mixtures of SCNC (Gm+‚(Pn-)a where a ∼ m/n) with PDDA as a polysalt. In parts (a) and (b) the mixing ratio (ma) is given by dividing the moles of PDDA-bound cations by all the moles of KPVS-bound anions (an) in the SCNC. (a) Hydrodynamic radius distributions f(Rh) normalized by the scattered intensity (Is ∼ ∫0∞ f(Rh) dRh) of SCNC: (1) rm ) 0 and ma ) 0 (i.e., uncomplexed PENG); (2) rm ) 1 and ma ) 0 (SCNC); (3) rm ) 1 and ma ) 0.1; (4) rm ) 1 and ma ) 0.5; and (5) rm ) 1 and ma ) 1. (b) Changes of Rh (open circles) and intensity (filled circles) with ma. All the measurements were performed 2 days after the mixing. Several samples were re-examined after a week and found that there is no difference in the results after 2 days and a week within experimental errors.

dissociation of the KPVS ions from the complex; in other words, a few parts of the KPVS ions are still bound to the SCNC particle even when an equimolar amount of PDDA was added. As a reason for this we may consider different states of the bound KPVS anions. The cationic charges fixed to the dangling chains on the surface of the PENG particle should form a stable complex with KPVS. The other charges within the gel particle may also attract the KPVS anions to the surface via the long-distance Coulomb interaction, but the binding seems to be considerably looser. Such loosely bound KPVS ions can be removed by forming a stable complex with the PDDA added, whereas the removal of stably bound KPVS ions would be difficult due to restrictions on chain configurational entropy for both PDDA and KPVS (i.e., “steric hindrance” effect). Taking the above into account, it would be interesting to see the effect of simple salts on the stability of the SCNC particle. Figure 9 shows the change of Rh with KCl concentration (Cs). The complex undergoes little change of Rh with increasing Cs at Cs < 0.02 M, but begins to aggregate with each other around Cs ∼ 0.03 M as demonstrated by an increase of Rh. This is not the “salting out” effect as observed in water solutions of poly(NIPA)15 because the complex becomes water-soluble upon further addition of KCl (i.e., Cs g 0.2 M). Therefore, this (15) Schild, H. G.; Tirrell, D. A. J. Phys. Chem. 1990, 94, 4352.

2100 Langmuir, Vol. 23, No. 4, 2007

Ogawa et al.

Figure 10. Normalized hydrodynamic radius distribution (f(Rh)) curves obtained by CONTIN analysis of DLS data for (1) uncomplexed PENG at Cs )0.01 M, (2) SCNC/PDDA mixture (ma ) 1) in the KCl-free system, and (3) SCNC/PDDA mixture (ma ) 1) in the presence of KCl (Cs ) 0.01 M). The normalization was made by the scattered intensity of SCNC (rm ) 1 and ma ) 0).

Figure 8. Results of SLS for the samples subjected to the DLS measurements in Figure 7. (a) Plots of KCx/R(θ) vs sin2(θ/2); the numbers in the parentheses are the same as those in Figure 7a. (b) Change of Rg with ma. The SLS data were analyzed as shown in the Experimental Section.

Figure 9. Effect of KCl on Rh of SCNC. The measurements were performed 2 days after the sample preparation and therefore were not able to be done in the gray shaded area of Cs due to the precipitation of aggregated complex particles.

phenomenon is the same as the “antipolyelectrolyte behavior” observed in aqueous polyampholyte systems (e.g., see ref 4). To compare the Rh (∼78 nm) of the SCNC at Cs from 0.3 to 0.5 M with that of the KPVS-free PENG at the same Cs ranges, we performed the DLS at Cs ) 0.5 M using the PENG sample and obtained Rh ∼ 77 nm. As a result, it seems that the SCNC in the KCl solution at Cs from 0.3 to 0.5 M is equivalent to the KPVSfree PENG and is more swollen than that at Cs e 0.02 M ranges in which the Rh ∼ 62 ∼ 67 nm. Ionic polymers and gels shrink with increasing Cs (e.g., see ref 7); thus, the results in Figure 9 strongly suggest that almost all of the KPVS ions dissociate from the complex in the presence of high concentrations of KCl.

We believe that the aggregation observed at Cs from 0.03 to 0.2 M is due to the electrostatic interaction of the KPVS anions dissociated partially, but not separated from the complex particle, with the other complex whose cationic charges resulted from partial dissociation of KPVS. To confirm the above assertion, we performed DLS for the PDDA/SCNC mixture at ma ) 1 in the absence and the presence of KCl (Cs ) 0.01 M). The size distribution curves were compared with that of the uncomplexed PENG in the presence of 0.01 M KCl (see Figure 10). The important results are summarized as follows: (i) The distribution curve at ma ) 1 and at Cs ) 0.01 M shows a clear bimodal profile having two peaks, see curve (3); (ii) the peak at Rh ∼ 17 nm is assignable to the KPVSPDDA complex; and (iii) another peak at Rh ∼116 nm is equal to that (Rh ∼ 116 nm) of the uncomplexed PENG in 0.01 M KCl solution (i.e., rm ) 0 and Cs ) 0.01 M). However, (iv) there is no good agreement between the distribution curves of the uncomplexed PENG at Cs ) 0.01 M and the PDDA/SCNC mixture at both ma ) 1 and Cs ) 0; compare curves (1) and (2). The uncomplexed PENG in 0.01 M KCl solution should provide the baseline data for the SCNC, from which all the KPVS ions completely dissociate. Therefore, results (i)-(iv) strongly suggest that KCl acts as an accelerator in the process in which the complexed KPVS dissociates to form a more stable complex with PDDA. General Discussion. In the present study the first question we had to answer was why the SCNC does not show zero mobility in ELS measurements. For this we assumed that the KPVS ion (Pn- in Gm+‚(Pn-)a) dissociates under an applied electric field. To directly verify this assumption through experiments, however, there are many difficulties. Therefore, we studied changes of Rh by DLS and/or Rg by SLS of the SCNC upon addition of cationic PENG, PDDA, and KCl. These light scattering experiments strongly suggested the dissociation of KPVS from the SCNC. This is shown schematically below. The exchange reaction of the SCNC with the cationic PENG gives rise to a nanoparticle as shown by (A). This process results in swelling of the SCNC. Upon addition of PDDA as a polycation, a stable KPVS-PDDA complex was formed, instead of nanoparticle (A). Therefore, it seems that both processes take place via the dissociation of the KPVS ion as shown in (B) followed by that in (C). (Note that the dissociated KPVS ion immediately forms the complex with the PDDA cation.) The KPVS dissociation was more clearly demonstrated by addition of a high concentration (Cs g 0.2 M) of KCl. However, moderate

Intraparticle Complex of Cationic Nanogel

concentrations (0.04 M < Cs < 0.1 M) of KCl bring about aggregation, perhaps due to a partial dissociation of the KPVS ion as shown in (B). It is described in Figures 4-24 in ref 9 that a nucleosome core particle around which the DNA is tightly wrapped dissociates into an octameric histone core and 146-nucleotide-pair DNA double helix with a high concentration of salt. The effect of KCl on the dissociation of the KPVS ion, as shown in (C), seems to be closely related to this biochemical change; thus, the present SCNC can serve as a simple model for understanding the complicated biological processes in which electrostatic interaction plays a key role. The second point to which we pay attention is the stoichiometric charge neutralization in macromolecular systems in which one polyelectrolyte has a nonlinear chain structure. Kokufuta16 has studied the polyelectrolyte complex formation of KPVS with branched poly(ethyleneimine) (BPEI; Mw ∼ 1.1 × 105 g/mol) by the method of “colloid titration” which was originated in 1952 by Terayama.17 It was reported that the -NH3+ and -NH2+- groups in BPEI form salt linkages or ion pairs with the sulfate groups in KPVS according to a 1:1 stoichiometry; however, this stoichiometry does not hold in the salt linkage formation of the tertiary ammonium ions (>NH+-) at the branching point of BPEI. Taking these results into account, we may suggest from the present study that the stoichiometric charge neutralization between PENG and KPVS observed here is due to both the loose binding of KPVS ions and the moderate collapse of the resulting SCNC. In other words, the complexation of BPEI with KPVS could result in highly collapsed and hydrophobic complex particles, which immediately aggregate with each other and precipitate before the participation of the tertiary amino groups in the complex formation. Another significant suggestion is that the present results may provide a key for understanding the nature of polyelectrolyte gel swelling. As mentioned in ref 7, there are two models for the swelling force that works in ionic gels; i.e., the Flory model and the Katchalsky model. Flory considered that a lightly crosslinked ionic gel closely resembles a Donnan membrane system, while such a gel in the Katchalsky’s treatment was considered to be an extension of the corresponding polyion. The shrinking of PENG by complexation with KPVS ions and the swelling of SCNC by the dissociation of KPVS ions can be understood in terms of electrostatic interactions between the fixed charges of polyelectrolytes. Note that counterions (circles with minus signs) within the gel phase in (C) act to maintain electrical neutrality, but do not raise the osmotic pressure.

Conclusions The formation of an intraparticle polyelectrolyte complex between cationic PENG and KPVS anions was examined by (16) Kokufuta, E. Macromolecules 1979, 12, 350. (17) Terayama, H. J. Polym. Sci.. 1952, 8, 243.

Langmuir, Vol. 23, No. 4, 2007 2101

ELS. It was found that the complex formed at rm ) 1 (i.e., cationic PENG with the stoichiometric amount of bound KPVS ions) shows a positive, but not zero, mobility at both 25 and 45 °C, at which point the complex is in a fully collapsed state. To understand this result, we assumed that (i) bulk electroneutrality is maintained in the complex with the aid of counterions, but (ii) the complex is highly polarizable, and hence (iii) during ELS the KPVS anions would dissociate in part from the complex. The following results have indicated that this assumption is reasonable: (i) DLS shows an increase in the Rh of the complex, accompanied by a decrease in the light scattering intensity, upon addition of the KPVS-free PENG. That is an indication of the polyelectrolyte exchange reaction. (ii) The same phenomenon can be observed by DLS and SLS when PDDA was added to the complex dispersion, indicating that the PDDA removes KPVS from the complex. (iii) Upon addition of KCl, the complex undergoes little change in Rh at Cs e 0.02 M, aggregates with each other at Cs from 0.03 to 0.2 M, and becomes water-soluble at Cs > 0.2 M. (iv) The Rh of the soluble complex at Cs from 0.3 to 0.5 M is larger than that at Cs < 0.02 M, suggesting the swelling of a complex particle due to the dissociation of the KPVS ions. (v) The complexation between KPVS and PDDA which results in the KPVS dissociation as mentioned in (ii) is facilitated in the presence of 0.01 M KCl. Acknowledgment. We thank Reviewer I for her/his help to make a better account of the present work, which was supported in part by Grant-in-Aids for Scientific Research to E.K. from the Japan Society for the Promotion of Science (No. 15350127 and No. 18655089). List of Abbreviations A2 ) second virial coefficient BPEI ) branched poly(ethyleneimine) cmr ) critical mixing ratio at which both size and molar mass of the resulting complex particles abruptly increase CKPVS ) KPVS concentration in g/mL CPENG ) PENG concentration in g/mL Cx ) complex concentration in g/mL Cs ) salt concentration in mol/L (dn/dc)KPVS ) change in the refractive index of KPVS with its concentration (dn/dc)PENG ) change in the refractive index of PENG with its concentration (dn/dc)x ) change in the refractive index of complex with its concentration DKPVS ) charge density in mol/g for KPVS DPENG ) charge density in mol/g for PENG DLS ) dynamic light scattering ELS ) electrophoretic light scattering f(Rh) ) hydrodynamic radius distribution Gm+ ) a PENG particle with +m charges Is ) scattering intensity KPVS ) potassium poly(vinyl alcohol) sulfate ma ) molar ratio of PDDA-bound dN+d ions to the KPVSbound -OS(dO)2O- ions which are in existence in a SCNC particle M h w ) weight-average molar mass M h x ) weight-average molar mass of the complex no ) refractive index of the medium N ) surface charge density NA ) Avogadro’s number pI ) isoelectric point PDDA ) poly(diallyldimethylammonium chloride) PENG ) polyelectrolyte nanogel Pn- ) a polyanion chain with -n charges rm ) molar ratio of KPVS-bound -OS(dO)2O- ions to PENGbound imidazole cations Rg ) mean square radius of gyration Rh ) hydrodynamic radius Rθ ) Rayleigh ratio

2102 Langmuir, Vol. 23, No. 4, 2007 SCNC ) stoichiometrically charge-neutral complex SLS ) static light scattering U ) electrophoretic mobility in cm2‚s-1‚V-1 VKPVS ) volume in mL for KPVS solution VPENG ) volume in mL for PENG dispersion R ) molar ratio of KOH to a total of PENG-bound imidazole cations Ri ) ionization degree for carboxyl or imidazole groups bound to NIPA-based nanogels

Ogawa et al. β ) mass ratio of bound KPVS to a PENG particle λ ) wavelength of laser light in medium λo ) wavelength of laser light in a vacuum λf ) square root of the ratio of the frictional coefficient of surface layer to the viscosity of the medium θ ) scattering angle LA061767T