Rheological Behavior of Acid-Swellable Cationic Copolymer Latexes

Oct 15, 2009 - James McParlane , Damien Dupin , Jennifer M. Saunders , Sarah Lally , Steven P. Armes , Brian R. Saunders. Soft Matter 2012 8, 6239 ...
2 downloads 0 Views 3MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Rheological Behavior of Acid-Swellable Cationic Copolymer Latexes Beng H. Tan Institute of Materials Research and Engineering (IMRE), A*STAR (Agency for Science, Technology and Research), 3 Research Link, Singapore 117602

Kam C. Tam* Department of Chemical Engineering, University of Waterloo, 200 University Avenue West, Waterloo, Ontario N2L 3G1, Canada

Damien Dupin* and Steven P. Armes Department of Chemistry, Dainton Building, The University of Sheffield, Brook Hill, Sheffield, South Yorkshire S3 7HF, U.K. Received July 28, 2009. Revised Manuscript Received September 28, 2009 2-Vinylpyridine (2VP) was copolymerized with four different cross-linker densities ranging from 0.05 to 0.31 wt % divinylbenzene (DVB) via aqueous emulsion polymerization to produce a series of submicrometer-sized, lightly crosslinked P2VP latexes. Protonation of the 2VP residues leads to a latex-to-microgel transition due to interchain electrostatic repulsion, as confirmed by dynamic light scattering. The DVB content of these pH-responsive copolymer particles strongly affects their rheological behavior. The particle size and viscosity of the swollen cationic microgels exhibit a maximum at ∼0.11 wt % DVB. Static light scattering results confirm this density as the minimum amount of DVB required to ensure that all P2VP chains are cross-linked (i.e. that there is no soluble fraction), thus allowing optimal swelling of the microgels. Viscosity studies shows that the solution viscosity of a P2VP microgel at low pH follows two models, depending on its concentration. For volume fractions below 0.30, the P2VP microgels behave as hard spheres, as predicted by the Batchelor equation. For more concentrated P2VP microgels (volume fractions above 0.30), the rheological behavior can be predicted using the Krieger-Dougherty model for strong particle-particle interactions; thus, this semiempirical approach provides a useful description of the aqueous solution behavior of microgel.

Introduction In recent years, numerous studies have provided new insights into the rheological behavior of pH-responsive colloids, such as alkali-swellable latexes. The aqueous solution viscosity and shearthinning behavior of such systems increase dramatically, leading to various commercial applications.1-4 The pronounced increase in aqueous solution viscosity and shear-thinning behavior is attributed to the increased solubility and electrostatic repulsion between polymer chains in their swollen microgel form.4 Previous studies have focused mainly on particles that may exhibit substantial dissolution behavior, in addition to particle swelling, upon neutralization.5,6 The resulting complex mixture of dissolved linear polymer chains and compositionally heterogeneous swollen particles does not readily lend itself to an improved understanding of the dispersion structure and its influence on rheology. A major contribution regarding the correlation of swelling behavior and structure of near-monodisperse, cross-linked, *Corresponding authors. E-mail: [email protected] (K.C.T.), [email protected] (D.D.). (1) Quadrat, O.; Mrkvickova, L.; Walterova, Z.; Stern, P.; Bradna, P.; Snuparek, J. Prog. Org. Coat. 2003, 46, 1–7. (2) Snuparek, J.; Quadrat, O.; Horsky, J. Prog. Org. Coat. 2005, 54, 99–103. (3) Saunders, B. R.; Laajam, N.; Daly, E.; Teow, S.; Hu, X.; Stepto, R. Adv. Colloid Interface Sci. 2009, 147-148, 251–262. (4) Tan, B. H.; Tam, K. C. Adv. Colloid Interface Sci. 2008, 136, 25–44. (5) Verbrugge, C. J. J. Appl. Polym. Sci. 1970, 14, 911–928. (6) Muroi, S.; Hosoi, K.; Ishikawa, T. J. Appl. Polym. Sci. 1967, 11, 1963–1978. (7) Wolfe, M. S. Prog. Org. Coat. 1992, 20, 487–500.

2736 DOI: 10.1021/la9027699

nonaqueous microgels with their bulk rheology is credited to Wolfe.7 He demonstrated that the solution viscosity is reduced with increasing cross-linker density; the observed changes, which are dependent on the microgel concentration, cross-linker density, and solvent quality can all be correlated using the microgel effective volume fraction. In addition, Rodriguez et al. studied changes in the internal polymer segment density and particle swelling on addition of alkali to cross-linked methacrylic acid-ethyl acrylate (MAA-EA) microgels using static and dynamic light scattering.8 Upon ionization of the MAA residues, a reduction in the ratio of the radius of gyration (Rg) to the hydrodynamic radius (Rh) from the known value of 0.778 for homogeneous spheres, along with the curvature in Guinier plots of the angular-dependent light scattering intensity, indicated a nonuniform polymer segment density distribution within these microgels. Stieger et al. investigated the internal structure of thermoresponsive poly(N-isopropylacrylamide) (PNIPAM) particles in dilute solution using small-angle neutron scattering (SANS).9 In their swollen microgel state, the polymer segment density distribution was found to be inhomogeneous. In contrast, when the particles were collapsed at temperatures above the LCST, the segment density distribution can be described by a boxlike profile, which is characteristic of homogeneous hard spheres. (8) Rodriguez, B. E.; Wolfe, M. S.; Fryd, M. Macromolecules 1994, 27, 6642– 6647. (9) Stieger, M.; Richtering, W.; Pedersen, J. S.; Lindner, P. J. Chem. Phys. 2004, 120, 6197.

Published on Web 10/15/2009

Langmuir 2010, 26(4), 2736–2744

Tan et al.

Article

After exploring the internal structure of microgel particles in dilute solution, the Richtering group also investigated interparticle interactions in more concentrated dispersions.10 Microgels with differing degrees of cross-linking and particle diameters behaved as hard spheres up to effective volume fractions, φeff, of 0.35. At higher φeff values, strong deviation from hard-sphere behavior was observed. Interpenetration of the less cross-linked outer regions of the microgels and significant particle compression were observed under these conditions. The structure of PNIPAM colloidal crystals has been extensively studied by several groups.11-18 Assuming that the particle form factor P(q) was independent of concentration and that the φeff of dilute dispersions is sufficiently small to neglect interparticle interactions, SANS analysis indicated the formation of a face-centered-cubic (fcc) lattice.12 Debord et al. directly observed compression of soft PNIPAM-based microgel particles when close-packed in a crystal using laser scanning confocal microscopy.13-15 Such swollen microgel particles are deformable at higher concentrations and exhibit soft-sphere behavior. Lyon’s group has also shown that, at high microgel concentration, the maximum microgel volume fraction is similar to that for hard spheres (0.74). The degree of particle compression, or assembly “overpacking”, was defined as r/σ, where r is the center-to-center distance at a given microgel concentration and σ is the microgel diameter in dilute solution.15,16 Another significant contribution to the rheological behavior of stimuli-responsive microgels can be attributed to the pioneering work of Ballauff and Richtering.11,19-21 They reported that the effective volume fraction, φeff, of microgels can be determined using Batchelor’s equation22 η0 =ηs ¼ 1 þ 2:5φeff þ 5:9φeff 2

ð1Þ

in the dilute regime, where φeff is substituted with the term kc. Here η0 is the viscosity of the suspension and ηs the viscosity of the medium (or solvent), c is the mass concentration, and k is the specific volume, which is a constant and also the only adjustable parameter. Recently, Tan et al. confirmed that values of k determined in the dilute regime could not account for the prevailing physics of soft colloidal particles at moderate to high concentrations. Instead, they suggested that k should decrease with increasing concentration until it approached the hard-sphere limit. They proposed a semiempirical approach where a variable specific volume, k, was introduced to convert c into φeff.4,23,24 The values (10) Stieger, M.; Pedersen, J. S.; Lindner, P.; Richtering, W. Langmuir 2004, 20, 7283–7292. (11) Senff, H.; Richtering, W. J. Chem. Phys. 1999, 111, 1705. (12) Hellweg, T.; Dewhurst, C. D.; Br€uckner, E.; Kratz, K.; Eimer, W. Colloid Polym. Sci. 2000, 278, 972. (13) Debord, J. D. J. Phys. Chem. B 2000, 104, 6327. (14) Debord, J. D.; Eustis, S.; Debord, S. B.; Lofye, M. T.; Lyon, L. A. Adv. Mater. 2002, 14, 658. (15) Debord, S. B.; Lyon, L. A. J. Phys. Chem. B 2003, 107, 2927. (16) Wu, J.; Huang, G.; Hu, Z. Macromolecules 2003, 36, 440. (17) St. John, A. N.; Breedveld, V.; Lyon, A. L. J. Phys. Chem. B 2007, 111, 7796–7801. (18) Lyon, A. L.; Debord, J. D.; Debord, S. B.; Jones, C. D.; McGrath, J. G.; Serpe, M. J. J. Phys. Chem. B 2004, 108, 19099. (19) Senff, H.; Richtering, W.; Norhausen, C.; Weiss, A.; Ballauff, M. Langmuir 1999, 15, 102–106. (20) Senff, H.; Richtering, W. J. Colloid Polym. Sci. 2000, 278, 830–840. (21) Ballauff, M. Macromol. Chem. Phys. 2003, 204, 220–234. (22) Batchelor, G. K. J. Fluid Mech. 1977, 83, 97–117. (23) Tan, B. H.; Tam, K. C.; Lam, Y. C.; Tan, C. B. J. Rheol. 2004, 48, 915–926. (24) Tan, B. H.; Tam, K. C.; Lam, Y. C.; Tan, C. B. Polymer 2004, 45, 5515– 5523.

Langmuir 2010, 26(4), 2736–2744

of k were determined from a mathematical fit to the data using the modified Krieger-Dougherty equation η0 =ηs ¼ ð1 - kc=φm Þ -½ηφm

ð2Þ

where the intrinsic viscosity, [η], is 2.5 and the maximum volume fraction, φm, is 0.63 for hard spheres.25 The relationship between the specific volume, k, and concentration of particles, c, can be written as23,24 k - kmin =k0 - kmin ¼ ½1 þ ðc=c0 Þ2  -m

ð3Þ

where kmin describes the limiting condition when the soft particles are compressed to the hard-sphere equivalent volume at high concentration. The constant c0 denotes the critical concentration at which the concentration of free counterions in the solution is large enough to induce osmotic deswelling of the soft particles, resulting in a smaller k. The reduction in k with particle concentration is described by the parameter m, which can be obtained from the slope of (k - kmin)/(k0 - kmin) versus c curve on a log-log plot. The φeff for microgels at various particle concentrations can be determined from eq 4 by substituting eq 3 into φeff = kc; hence φeff ¼ kmin þ ðk0 - kmin Þ½1 þ ðc=c0 Þ2  -m c

ð4Þ

This expression corrects for changes in the volume fraction of soft particles compared to that of the equivalent hard spheres. Tan et al. observed excellent agreement with the K-D model, η0/ηs = (1 - φ/φm)-[η]φm, for moderate to high volume fractions: a master curve was obtained when the viscosities of methacrylic acid-coethyl acrylate (MAA-EA) microgels were plotted using the modified φeff determined from eq 4.23,24 This strongly suggests that this semiempirical approach is appropriate for understanding the rheological behavior of soft, deformable particles. Moreover, eq 2 can be used to predict the relative viscosities of concentrated aqueous microgel solutions by taking account of the corrected volume fraction (eq 4) due to particle compression. There are very few reported studies on the rheological behavior of acid-swellable cross-linked latexes. Such systems acquire microgel character at low pH due to protonation of basic groups while maintaining a robust internal microstructure at all solution pH. Recently, Dupin et al. reported the preparation of acidswellable cross-linked latexes comprising 2-vinylpyridine (2VP) copolymerized with divinylbenzene (DVB) cross-linker via aqueous emulsion polymerization.26 These poly(2-vinylpyridine) (P2VP) particles exhibit a latex-to-microgel transition at around pH 4.1 when the solution pH is lowered on addition of acid. Stopped-flow techniques confirmed that the kinetics of swelling of such P2VP particles occurred within time scales of tens of milliseconds, whereas the kinetics of deswelling were significantly slower due to retarded salt excretion.27,28 Adsorption of these P2VP particles at the air/water interface was also studied: P2VP latex-stabilized foams could be obtained at either neutral or alkaline pH, but rapid foam destabilization occurred below pH 4 due to the hydrophilic character of the highly swollen cationic P2VP microgels.29 (25) Krieger, I. M.; Dougherty, T. J. Trans. Soc. Rheol. 1959, 3, 137–152. (26) Dupin, D.; Fujii, S.; Armes, S. P.; Reeve, P.; Baxter, S. M. Langmuir 2006, 22, 3381–3387. (27) Dupin, D.; Rosselgong, J.; Armes, S. P.; Routh, A. F. Langmuir 2007, 23, 4035–4041. (28) Yin, J.; Dupin, D.; Li, J.; Armes, S. P.; Liu, S. Langmuir 2008, 24, 9334– 9340. (29) Dupin, D.; Howse, J.; Armes, S. P.; Randall, D. J. Mater. Chem. 2008, 18, 545–552.

DOI: 10.1021/la9027699

2737

Article

Tan et al.

Figure 1. Schematic representation of the synthesis of poly(2-vinylpyridine) [P2VP] latex via aqueous emulsion polymerization at 10% solids with varying amounts of DVB cross-linker. The milky latex dispersion obtained at neutral pH becomes a transparent free-standing cationic gel at only 1.0 wt % solids by lowering the solution pH to 3, as shown in the digital photograph. Table 1. Hydrodynamic Radius, Rh, and Weight-Average Molecular Weight, M w, for the Four P2VP Latexes/Microgels Prepared with Different DVB Cross-Linker Contents at pH 5 and pH 3 entry no.

DVB content (wt %)

M w at M w at Rh at Rh at numberpH 5 pH 3 pH 5 pH 3 averagea radius (nm) (nm)b (nm)b (g mol-1)c (g mol-1)c

1 2 3 4

0.05 320 330 1000 780 480 0.11 350 360 1600 810 750 0.21 300 310 1390 828 800 0.31 300 320 1150 813 763 a Estimated by scanning electron microscopy (over 200 particles). b Measured by dynamic light scattering at 25 °C. c Measured by static light scattering at 25 °C.

The present study investigates the microstructure and the behavior of acid-swellable cross-linked P2VP microgels. Four P2VP latexes were synthesized with DVB contents ranging from 0.05 to 0.31 wt % via aqueous emulsion polymerization, as already reported by Dupin et al.26 These particles were characterized by electron microscopy and their swelling behavior was studied by both dynamic and static light scattering. The effects of the microstructure of these cationic microgels on the bulk rheological properties at conditions that are relevant to industrial applications have not been elucidated. In the present study, the rheological properties of these cationic microgels were assessed at both high and low volume fractions. The scope and limitations of the semiempirical approach proposed by Tan et al. were examined for the present which is an excellent model for soft particles since the particle size (and hence the volume fraction) can be readily controlled by adjusting the solution pH and the degree of cross-linking.

Figure 2. Representative scanning electron micrographs obtained for the four P2VP latexes prepared with different DVB contents: (a) 0.05, (b) 0.11, (c) 0.21, and (d) 0.31 wt %.

Experimental Section Materials. 2-Vinylpyridine (97%, 2VP; Aldrich) and divinylbenzene (80 mol % 1,4 divinyl content, DVB; Fluka, UK) were treated with basic alumina in order to remove inhibitor. Aliquat 336 (Aldrich, UK) and R,R0 -azodiisobutyramidine dihydrochloride (97%, AIBA; Aldrich, UK) were used as received. Monomethoxy-capped poly(ethylene glycol) methacrylate (PEGMA) macromonomer (Mn = 2000; Mw/Mn = 1.10) was supplied by Cognis Performance Chemicals (Hythe, UK) as a 50 wt % aqueous solution. Doubly distilled deionized water was used in all the polymerizations. PEGMA-Stabilized Poly(2-vimylpyridine) Syntheses via Emulsion Polymerization. The Aliquat 336 surfactant (1.0 g) and the PEGMA stabilizer (2.0 g) were dissolved in deionized water (82.0 g) in a 250 mL three-necked round-bottomed flask. A comonomer mixture of 2VP (10.0-9.97 g) and DVB (0-30 mg) was then added, causing the solution pH to increase to pH ∼8.3. 2738 DOI: 10.1021/la9027699

Figure 3. Variation of the hydrodynamic radius measured by dynamic light scattering from pH 10 to 2 of P2VP latexes prepared with different DVB contents: ([) 0.05, (9) 0.11, (2) 0.21, and (b) 0.31 wt % (see entries 1-4 in Table 1). Note that the error bars for hydrodynamic radii above pH 4.5 are omitted to avoid overcrowding the data points (measurement errors estimated from three readings were below 5%). Langmuir 2010, 26(4), 2736–2744

Tan et al.

Article

Figure 4. Relaxation time distributions obtained from Gendist for P2VP latexes prepared with different DVB contents: (a) 0.05, (b) 0.11, (c) 0.21, and (d) 0.31 wt % (see entries 1-4 in Table 1). Curves are plotted with increasing pH; i.e., the lowest curve represents pH 2, and the highest curve represents pH 10. The flask was sealed with a rubber septum, and the aqueous solution was degassed at ambient temperature using five vacuum/ nitrogen cycles. The degassed solution was stirred at 250 rpm using a magnetic stirrer and heated at 60 °C with the aid of an oil bath, and then the initiator solution (0.10 g of AIBA dissolved in 5.0 g of water) was added after 20 min. The copolymerizing solution turned milky-white within 10 min, and stirring was continued for 24 h at 60 °C. Purification. The latex particles were centrifuged at 8000 rpm for 40 min, followed by careful decantation of the supernatant, replacement with fresh water, and redispersion of the sedimented particles with the aid of an ultrasonic bath. This protocol was used to remove residual 2VP monomer, excess Aliquat 336 surfactant, and any nongrafted PEGMA stabilizer. Purification was continued until the serum surface tension was close to that of pure water (71 ( 1 mN m-1). Characterization. All microgel solutions were prepared using 10 mM NaCl as background electrolyte, and the solution pH was adjusted using either 1 M NaOH or 1 M HCl. Scanning electron microscopy (SEM) images were obtained using a field emission Inspect F instrument operating at a voltage of 20 kV. Dried samples were mounted on adhesive carbon disks placed on aluminum stubs and then sputter-coated with gold so as to minimize sample charging. Light scattering measurements were conducted using a Brookhaven BI-200SM goniometer equipped with a BI9000AT digital correlator. The inverse Laplace transform of REPES supplied with the GENDIST software package was used to analyze the time correlation functions, and the probability of reject was set to 0.5. DLS measures the intensity fluctuations with time and correlates these fluctuations to the properties of the scattering objects. Using the expression Γ = Dq2, the translational diffusion coefficient, D, can be determined. Γ is the decay rate, which is the inverse of the relaxation time, τ; q is the scattering vector, which is defined as q = (4πn sin(θ/2)/λ) (where n is the refractive index of the solution, θ is the scattering angle, and λ is the wavelength of the incident laser light in a vacuum). The apparent hydrodynamic Langmuir 2010, 26(4), 2736–2744

radius (Rh) can be determined using the Stokes-Einstein relationship30,31 Rh ¼

kT 6πηDapp

ð5Þ

where kB, η, and T are the Boltzmann constant, solution viscosity, and the absolute temperature, respectively. Static light scattering (SLS) data were collected at 25 °C to determine the weight-average molecular weight (Mw), the z-average radius of gyration (Rg), and the second virial coefficient (A2) of the aqueous microgel (or latex) particles from Zimm plots according to the relation32,33 " # Kc 1 q 2 Rg 2 þ 2A2 c ¼ 1þ ð6Þ 3 ΔRθ Mw where K is the optical constant, which depends on the refractive index increment of the microgel solution (K = 4π2n2(dn/dc)2/ NAλ4). Here, c is the microgel (or latex) concentration, n is the refractive index of the solvent, θ is the scattering angle, λ is the wavelength of laser light, ΔRθ is the excess Rayleigh ratio [ΔRθ = Rθ(solution)- Rθ (solvent)], dn/dc is the refractive index increment of the microgel (or latex), and NA is Avogadro’s constant. The scattering angles ranged from 60° to 135° at 15° intervals while the copolymer concentration ranged from 0.001 to 0.01 g/L. A plot of (Kc/Rθ) vs [sin2(θ/2) þ kc] (where k is a convenient arbitrary constant) can be used to determine the molecular parameters. By extrapolating the data to zero angle and concentration, Rg and A2 can be obtained from the two slopes,  ipanek, P. Data analysis in dynamic light scattering. In Dynamic Light (30) St Scattering - The Method and Some Applications; Brown, D., Ed.; Clarendon Press: Oxford, 1993; pp 177-241. (31) Chu, B. Laser Light Scattering: Basic Principles and Practice, 2nd ed.; Academic Press: New York, 1991. (32) Zimm, B. H. J. Chem. Phys. 1948, 16, 1093–1099. (33) Zimm, B. H. J. Chem. Phys. 1948, 16, 1099–1116.

DOI: 10.1021/la9027699

2739

Article

Tan et al.

respectively. This double extrapolation to zero angle and concentration also yields a common intercept, which is simply the inverse of the M w. Dilute solution viscosities of the P2VP microgels were determined using a Cannon Ubbeldohe capillary viscometer at a temperature of 25 ( 0.02 °C maintained with a thermostatic bath. The efflux time was kept relatively long (preferably more than 100 s) to minimize the need to apply kinetic corrections to the observed data. The efflux time measured using the capillary viscometer is related to the reduced and inherent viscosity by assuming that the density of the aqueous microgel solution is similar to the solvent. The relevant formulas for the reduced and inherent viscosities are as follows: ηred ¼

ηsp η - ηs t - ts ¼ ¼ c ηs c ts c

ð7Þ

where ηred is the reduced viscosity, ηinh is the inherent viscosity, ηsp is the specific viscosity defined as (η - ηs)/ηs, c is the solution concentration, η is the polymer solution viscosity, η0 is the solvent viscosity, ηr is the relative viscosity defined as (η/ηs), t is the efflux time of the polymer solution, and ts is the efflux time of the solvent. When the solution concentration approaches zero, both the reduced and inherent viscosities approach the intrinsic viscosity [η], which is given by ½η ¼ lim

cf0

ηsp ln ηr ¼ lim cf0 c c

ð8Þ

For suspensions, the intrinsic viscosity is the dilute limit of the viscosity increment per unit particle volume fraction divided by the solvent viscosity. It indicates the ability of the microgel particles to increase the solution viscosity in the absence of any intermolecular interactions. The intrinsic viscosity was obtained by extrapolation of the reduced and inherent viscosity data to zero polymer concentration using the Huggins and Kraemer equations. The Huggins equation is represented as34 ηsp =c ¼ ½η þ Kh ½η2 c

ð9Þ

where Kh is the Huggins coefficient, which is a constant for a series of polymers of different molecular weights in a given solvent and temperature. The Kraemer equation is35

Figure 5. Typical Zimm plots of the P2VP latex prepared with 0.11 wt % DVB content (entry 2 in Table 1) at (a) pH 5 (in its neutral latex form) and (b) pH 2 (in its cationic microgel form).

ð10Þ

viscosity data presented in this paper were collected under equilibrium viscosity conditions.

ðln ηr =cÞ ¼ ½η þ Kk ½η2 c

where Kk is the Kraemer coefficient and, theoretically, Kh - Kk = 0.5. The Huggins constant, Kh is a parameter that is related to the particle-solvent interaction. In a θ solvent, Kh is close to 0.5, and the medium provides an exact compensation for the excluded volume effect. In a “good” solvent, i.e., one that shows a zero or negative heat of mixing with the polymer, the chains are loosely extended, and the intrinsic viscosity is high. If the Huggins constant, Kh, is less than 0.5, this represents a strong particlesolvent interaction. In a “poor” solvent, i.e., one that shows a positive heat of mixing, the polymer chains interact more strongly with themselves than with surrounding solvent molecules. The polymer chains adopt a more compact configuration that results in a lower intrinsic viscosity. For “poor” solvents Kh is greater than 0.5, which indicates relatively weak particle-solvent interactions.36 Rheological studies of semidilute and concentrated solutions were collected using a Carri-Med CSL500 controlled-stress rheometer and a Contraves LS40 controlled rate rheometer. Cone and plate (40 mm, 2°) with a cup and bob geometry were employed to measure high and low viscosities. All experiments were conducted at a temperature of 25 ( 0.1 °C. The steady-shear (34) Huggins, M. L. J. Am. Chem. Soc. 1942, 64, 2716–2718. (35) Kraemer, E. O. Ind. Eng. Chem. 1938, 30, 1200–1203. (36) Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953; Chapter 13.

2740 DOI: 10.1021/la9027699

Results and Discussion Four lightly cross-linked P2VP latexes were prepared via oneshot batch emulsion copolymerization of 2-vinylpyridine with divinylbenzene using PEGMA macromonomer as a reactive steric stabilizer, a DVB cross-linker, a cationic azo initiator, and Aliquat 336 as a cosurfactant, as described by Dupin et al.26 A schematic representation of this latex synthesis is shown in Figure 1. These P2VP latexes all formed transparent free-standing gels on lowering the solution pH to 3, even at an initial latex concentration of only 1.0 wt % solids, except when the P2VP latex was prepared with the lowest amount of DVB cross-linker (entry 1 in Table 1) which gives a highly viscous microgel dispersion as shown in Figure S1 (see Supporting Information). Table 1 summarizes the hydrodynamic radii, Rh, obtained by DLS for the four P2VP latexes at pH 5 and the corresponding P2VP microgels at pH 3. The SEM images shown in Figure 2 confirm spherical, near-monodisperse morphologies for these P2VP latexes when dried at pH 6. The mean radii estimated from these images are generally lower than those obtained from DLS studies (see Table 1). There are several reasons for this apparent discrepancy. First, electron microscopy reports number-average diameters, whereas DLS reports intensity-average diameters. Thus particle size distributions with finite polydispersities are Langmuir 2010, 26(4), 2736–2744

Tan et al.

Figure 6. Static light scattering data illustrating the effect of varying the DVB content on (a) the weight-average molecular weight Mw and (b) the second virial coefficient A2 for four P2VP latexes (entries 1-4 in Table 1) at (b) pH 5 and (O) pH 2.

always oversized by DLS. Second, DLS is also sensitive to the solvated PEGMA chains surrounding each P2VP particle, whereas this steric stabilizer layer contributes negligible thickness under the high-vacuum conditions required for SEM. The four pH-responsive P2VP latexes/microgels were characterized in dilute aqueous solution using DLS from pH 2 to pH 10 in the presence of 10 mM NaCl. Above pH 5, these particles behaved like hard spheres with hydrodynamic radii, Rh, of 310-360 nm, as shown in Table 1 and Figure 3, where the systematic variation of the DVB cross-linker content appears to have a negligible effect on the Rh of these particles. In each case, a critical latex-to-microgel swelling transition was observed between pH 4.0 and 4.5. As anticipated, microgels with lower levels of DVB cross-linker exhibited stronger swelling behavior: the maximum Rh for entry 2 in Table 1 is 1600 nm compared to 1150 nm for entry 4 in Table 1. This swelling is due to the enhanced osmotic pressure exerted by counterions trapped inside the P2VP network and electrostatic repulsion between the highly cationic, protonated chains. It is noteworthy that the maximum particle size of entry 1 is lower than that of entries 2-4 in Table 1; this was not expected since entry 1 has the lowest cross-linker density. Figure 4 shows the relaxation time distributions obtained from Gendist for the four P2VP microgels, i.e., entry 1 (Figure 4a), entry 2 (Figure 4b), entry 3 (Figure 4c), and entry 4 (Figure 4d). Each microgel exhibits a unimodal distribution at both low and Langmuir 2010, 26(4), 2736–2744

Article

Figure 7. (a) Reduced viscosity (ηsp/c) and inherent viscosity (ln ηr/c) data as a function of microgel concentration, c, obtained from Ubbelohde viscometry measurements for fully protonated P2VP microgel cross-linked with 0.11 wt % DVB (entry 2 in Table 1) in 10 mM NaCl at pH 2. (b) Relationship between the intrinsic viscosity, [η] (open symbols), and the Huggins constant, Kh (filled symbols), as a function of DVB content for P2VP microgels at pH 2.

high pH except for entry 1, where a bimodal distribution is observed in the low pH region (corresponding to Rh values of around 50 and 1000 nm). In order to better understand the bimodal distribution associated with entry 1 (Figure 4a), we conducted static light scattering experiments to determine the weight-average molecular weight, Mw, and second virial coefficient, A2, for these four P2VP latexes/microgels prepared with differing levels of DVB cross-linker at pH 5 and 2, respectively. The Zimm plots obtained for entry 2 at pH 5 and pH 2 are shown in parts a and b of Figure 5, respectively. Figure 6a shows clearly that, at pH 5, the latexes possess molecular weights of ∼8.0  108 g mol-1, regardless of their DVB content. This is because the emulsion copolymerization process employed produces near-monodisperse, crosslinked particles of similar dimensions that contain comparable numbers of insoluble monomer units. The second virial coefficient, A2, is around -1.0  10-5 cm3 mol g-2 for all four P2VP latexes as depicted in Figure 6b. A negative A2 value suggests that the particle-solvent interaction is poor, which is consistent with the insoluble latex form of the P2VP particles at pH 5. However, when the P2VP chains become extensively protonated below pH 4, positive A2 values of ∼(2.2-8.8)  10-4 cm3 mol g-2 were obtained, which is consistent with the formation of highly solvated P2VP microgels in acidic solution. Moreover, the Mw of the resulting cationic swollen microgel containing 0.05 wt % DVB (entry 1) decreased to ∼5.0  108 g mol-1. The reduction in DOI: 10.1021/la9027699

2741

Article

Tan et al.

Figure 8. Viscosity versus shear rate of aqueous P2VP microgels at concentrations of (a) 0.1 wt % and (b) 3.0 wt % and prepared with varying DVB cross-linker contents: (]) 0.05, (0) 0.11, (4) 0.21, and (O) 0.31 wt % (see entries 1-4 in Table 1). (c) shows the relative low shear rate viscosities of P2VP microgels as a function of DVB contents at five microgel concentrations: (O) 0.1, (0) 0.5, (4) 1.0, (]) 2.0, and (b) 3.0 wt %.

Mw and the concomitant large increase in A2 for this microgel suggests that, in addition to particle swelling, there is substantial expulsion of soluble linear (i.e., non-cross-linked) P2VP chains from the microgel particles, which enhances the overall polymer-solvent interaction. This is consistent with the bimodal distribution observed for this microgel in Figure 4a at low pH and indicates incomplete cross-linking at this lowest DVB content. As the level of DVB cross-linker was increased, the molecular weight of the fully protonated cationic microgel increased up to an asymptotic value of ∼8.0  108 g mol-1, which is essentially the same as that determined for the four P2VP latexes at pH 5. Thus, on the basis of these empirical data, 0.11 wt % appears to be the optimum DVB content for P2VP microgels to yield optimally cross-linked swollen microgels. The reduction in A2 at higher DVB contents indicates that the extent of solvation of the protonated cationic P2VP chains is progressively reduced at greater degrees of cross-linking. This is similar to results reported by Tan et al. for alkali-swellable microgels.37 In addition, we also synthesized a linear P2VP latex in order to assess the mean degree of polymerization of the P2VP chains in the absence of the DVB cross-linker. Unfortunately, SLS analysis of this linear latex was not undertaken. However, GPC analysis (37) Tan, B. H.; Tam, K. C.; Lam, Y. C.; Tan, C. B. Adv. Colloid Interface Sci. 2005, 113, 111–120.

2742 DOI: 10.1021/la9027699

(THF eluent, refractive index detector, six near-monodisperse P2VP calibration standards) indicated an Mw of ∼229 000 and an Mw/Mn of 2.67. Thus, assuming that the PEGMA content is negligible and that the 2VP and DVB comonomers are both fully reacted during the emulsion copolymerization, we calculated a weight-average mean degree of polymerization of ∼2150. Thus, for the 0.055 wt % DVB formulation (entry 1 in Table 1), we estimated one DVB unit per ∼2475 2VP units, which corresponds to just under the minimum amount required to fully cross-link all the P2VP chains. This observation is consistent with the soluble fraction of linear chains indicated for this sample by the static light scattering data (see Figure 5). In contrast, using 0.11 wt % DVB cross-linker (entry 2 in Table 1) corresponds to ∼1240 2VP units per chain, which is sufficient to ensure complete crosslinking (and hence there is no soluble fraction). Aqueous electrophoresis measurements were conducted from pH 2 to 10 for the four P2VP latex/microgels, as depicted in Figure S2 (see Supporting Information). On lowering the pH, the zeta potential increased from þ5 mV at pH 10 up to þ45 mV at pH 2 due to protonation of the 2VP residues. Above pH 5, the zeta potential was quite close to zero due to deprotonation of the 2VP residues and the shielding effect of the nonionic PEGMA stabilizer. Viscometry measurements were collected using a Ubbelohde viscometer for microgels with varying DVB contents in the dilute solution regime at pH 2. The data were analyzed using the Langmuir 2010, 26(4), 2736–2744

Tan et al.

Article

Huggins and Kraemer equations (eqs 9 and 10, respectively) so as to calculate the intrinsic viscosity [η] and particle-solvent interaction parameter, Kh. An example data set is shown in Figure 7a, which shows the reduced viscosity (ηsp/c) and inherent viscosity (ln ηr/c) as a function of microgel concentration, c, for fully protonated P2VP microgel (entry 2 in 10 mM NaCl at pH 2). The solid lines represent the extrapolation of the reduced and inherent viscosities to zero P2VP concentration to obtain [η] and Kh from the Huggins and Kraemer equations, respectively. Figure 7b shows the relationship between [η] and Kh as a function of DVB content at pH 2. The intrinsic viscosity decreases with increasing DVB content due to a reduction in the swollen coil dimensions, as expected. However, microgels prepared using 0.05 wt % DVB display a much lower intrinsic viscosity, since they possess a smaller particle size due to the loss of linear P2VP chains upon acid-induced swelling. From the definition of the Huggins constant, a smaller polymer-solvent interaction Kh implies that the polymer-solvent interaction is enhanced, as described above. As expected, Figure 7b shows that Kh is reduced on lowering the amount of DVB cross-linker due to greater microgel swelling (and also dissolution of linear P2VP chains if the DVB content is too low) which enhances the polymer-solvent interaction. This result is also in good agreement with the second virial coefficient, A2, determined from the static light scattering data (see Figure 6b). Figure 8 displays the flow properties of aqueous P2VP microgels cross-linked with different amounts of DVB at concentrations of 0.10 wt % (Figure 8a) and 3.0 wt % (Figure 8b). At the lower concentration, this dispersion displays Newtonian behavior, with the low shear rate viscosity, η0, being determined by extrapolation to zero shear. The aqueous microgel viscosity increased progressively and became shear-thinning at higher concentration. This shear rate dependence on viscosity occurs when the shear rate is sufficiently high to perturb the interparticle interaction forces. The value of the low shear rate viscosity plateau, η0, was determined from a regression fit of the viscosity versus shear rate data to the Cross model given in eq 11: η 1 ¼ η0 1 þ ðK γÞm

ð11Þ

which is known to provide a reliable description of the viscosity of a colloidal suspension. The model fits to eq 11 are represented by the solid lines in Figure 8b. The low shear rate relative viscosity η0/ηs for the four P2VP microgels with varying DVB contents obtained at five microgel concentrations is plotted in Figure 8c. For a given microgel concentration, the relative viscosity is reduced at higher DVB content because the more heavily crosslinked microgel is less swollen and hence exhibits a lower effective volume fraction. Note that microgels cross-linked with 0.05 wt % DVB display a much lower viscosity as they possess a smaller particle size due to the loss of linear P2VP chains upon swelling, which is consistent with the data presented above. A maximum relative viscosity was observed when the P2VP latex was prepared using 0.11 wt % DVB cross-linker. This indicates that the optimum amount of DVB required to fully cross-link the P2VP latex is located between 0.055 and 0.11 wt %. Adopting the approach proposed by Ballauff and co-workers,19 the effective volume fractions of these pH-responsive P2VP microgels were determined using Batchelor’s equation. Microgel cross-linked with 0.05 wt % DVB (entry 1) has been excluded from the following analysis as it is a complex mixture of dissolved linear P2VP chains and compositionally heterogeneous swollen microgel particles that does not readily lend itself to an improved Langmuir 2010, 26(4), 2736–2744

Figure 9. Relative viscosity of P2VP microgels measured at pH 2 as a function of (a) mass concentration, c, and (b) the effective volume fraction φeff (as calculated from (a)) for microgels prepared with varying DVB cross-linker contents: (0) 0.11, (4) 0.21, and (O) 0.31 wt % (see entries 2-4 in Table 1).

understanding of the effect of polymer microstructure on the solution rheology. From the intrinsic viscosity measurements, a constant specific volume, k, was determined using eq 1 for each microgel (entries 2-4) at pH 2. The relative viscosities, η0/ηs, are shown in Figure 9a as a function of microgel concentration, c, for different DVB contents. The solid lines represent fits according to eq 1 where φeff = kc. In Figure 9b, the relative viscosity of each P2VP microgel was plotted against its effective volume fraction, φeff, calculated from the constant k determined from Figure 9a. In the dilute solution regime, all the relative viscosity data obeyed Batchelor’s equation for hard spheres. This suggests that k is independent of microgel concentration in the dilute regime (φeff < 0.3). The semiempirical theory developed by Tam’s group for predicting microgel viscosities assumes that the specific volume, k, is inversely proportional to the microgel volume fraction.23,24 Values of k are determined by fitting the data to the modified Krieger-Dougherty equation (eq 2). Figure 10a shows the variation in specific volume, k, as a function of microgel concentration, c, for the three P2VP microgels (entries 2-4) at pH 2. The solid lines in Figure 10a denote the mathematical fits according to eq 3.23,24 Microgels that are highly swollen at low pH (i.e., prepared with low DVB contents) exhibit a strong dependence of k on c; (k/k0) may decrease from 1.0 to 0.15 for such samples. The reduction in the (k/k0) term is explained by microgel deswelling due to the balance of osmotic pressure DOI: 10.1021/la9027699

2743

Article

Tan et al.

suggests that this semiempirical approach provides a useful description of the aqueous solution behavior of these microgels, and this form of eq 2 can be used to predict the solution viscosities.

Conclusions

Figure 10. (a) Specific volume, k, measured at pH 2 as a function of microgel concentration, c, for three P2VP microgels prepared using varying amounts of DVB cross-linker: (9) 0.11, (2) 0.21, and (b) 0.31 wt % (see entries 2-4 in Table 1). The solid lines denote the mathematical fits determined according to eq 3. (b) Relative viscosity as a function of the effective volume fraction φeff as calculated from eq 4 for P2VP microgels prepared with varying DVB cross-linker contents: (9) 0.11, (2) 0.21, and (b) 0.31 wt % at pH 2. The solid line denotes the Krieger-Dougherty model, η0/ηs = (1 - φ/φm)-[η]φm.

between the bulk and the interior of highly swollen cationic particles. The reduction in microgel dimensions in the concentrated regime is in good agreement with the particle compression effects observed by Lyon and co-workers for PNIPAM colloidal crystals.17,18 The φeff of these P2VP microgels at various concentrations can be determined from eq 4 by substituting eq 3 into φeff =kc, where this expression corrects for changes in volume fraction of soft particles to produce that of the equivalent hard sphere. Excellent agreement with the K-D model [η0/ηs = (1 - φ/φm)-[η]φm] was observed for moderate to high volume fractions: a master curve was obtained when the viscosity data were plotted using the modified φeff determined from eq 4 (see Figure 10b). This strongly

2744 DOI: 10.1021/la9027699

The rheological properties of acid-swellable P2VP particles were investigated as a function of pH and DVB cross-linker content. The hydrodynamic radius, Rh, increases dramatically as the solution pH is reduced below the pKa of the P2VP chains, indicating substantial particle swelling due to the enhanced osmotic pressure inside the P2VP microgel particles and electrostatic repulsion between the highly cationic, protonated P2VP chains. However, the extent of microgel swelling generally decreases with increasing cross-linker density, as expected. The exception was when a P2VP latex was prepared using 0.05 wt % DVB cross-linker, which proved insufficient to cross-link all the P2VP chains, as judged by relaxation time distributions and static light scattering studies. A pronounced increase in solution viscosity and shear-thinning behavior were observed at low pH, which is attributed to the swollen cationic P2VP microgels occupying larger effective volume fractions, resulting in reduced interparticle separation and hence a stronger particle interaction potential. The optimum DVB cross-linker density for P2VP latex/ microgel was found to be between 0.05 and 0.11 wt % DVB, since the latter formulation produced microgels with the largest size and highest solution viscosity upon swelling at low pH, as observed by dynamic light scattering and rheological studies, respectively. Beyond 0.11 wt % DVB, the P2VP particles were fully cross-linked, and their microgel dimensions at pH 2 decreased with increasing DVB content. Rheological studies of aqueous P2VP microgels at pH 2 confirm that the solution viscosity is reduced at higher DVB contents. However, two distinct behaviors were observed depending on the P2VP microgel concentration. In the dilute solution regime at pH 2, 0.1 wt % P2VP aqueous microgels exhibit Newtonian behavior, whereas shear-thinning behavior was observed for a 3.0 wt % microgel solution at the same pH. For volume fractions below 0.30, the intrinsic viscosities of these P2VP microgels can be fitted to Batchelor’s equation, which indicates that the particles behave as hard spheres with negligible interparticle interactions. Moreover, for P2VP microgel volume fractions greater than 0.30, the intrinsic viscosity can be predicted by the Krieger-Dougherty model, which implies strong interparticle interactions. Acknowledgment. S.P.A. is the recipient of a five-year Royal Society-Wolfson Research Merit Award. We thank The University of Sheffield for postdoctoral funds to support D.D. B.H.T. acknowledges financial support provided by Nanyang Technological University, Singapore. We also thank Ms. Lim Hui Ying, who assisted with some of the measurements. Supporting Information Available: Digital images of the P2VP microgel dispersions at pH 3 and zeta potential vs pH of the P2VP particles. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(4), 2736–2744