Quantifying Surface Coverage of Colloidal Silica by a Cationic

Quantifying Surface Coverage of Colloidal Silica by a Cationic Peptide Using a .... Upon increasing the total KWK concentration (C0 from eq 1) in the ...
0 downloads 0 Views 95KB Size
Download PDF
1852

Langmuir 2006, 22, 1852-1857

Quantifying Surface Coverage of Colloidal Silica by a Cationic Peptide Using a Combined Centrifugation/Time-Resolved Fluorescence Anisotropy Approach Dina Tleugabulova and John D. Brennan* Department of Chemistry, McMaster UniVersity, Hamilton, Ontario L8S 4M1, Canada ReceiVed October 4, 2005. In Final Form: NoVember 28, 2005 Recent experimental studies have shown that time-resolved fluorescence anisotropy (TRFA) is a promising methodology for in situ characterization of the surface modification of aqueous silica nanocolloids. Here we provide a more fundamental insight into the principle of this approach and discuss how the adsorption parameters for a cationic peptide, Lys-Trp-Lys (denoted using the standard shortform KWK), onto Ludox nanoparticles (NPs) are linked to the rotational dynamics of rhodamine 6G (R6G) dispersed in the KWK/Ludox mixture. First, the adsorption isotherm of KWK on hydrophilic controlled pore glass (CPG-3000) was obtained using the traditional centrifugation method, which provides the total molar amount of KWK per unit surface area of the silica. Assuming that both CPG and Ludox particles possess identical surface properties when suspended in the same aqueous buffer, both materials should also have identical adsorption properties. Thus, the adsorbed amount of KWK per unit area at a given total KWK concentration, as determined by the centrifugation method, can be plotted against the fractions of R6G anisotropy decay components at the same KWK concentration to relate the anisotropy components to the absolute surface coverage. Using this approach, it was determined that the concentration of KWK at which the CPG surface was saturated corresponded to the condition g ) 0 in the R6G decay, where g is the fraction of the nondecaying anisotropy component. This condition means that there is no R6G bound to the fraction of Ludox NPs with a radius R > 2.5 nm at maximum KWK coverage, consistent with the adsorbed peptide forming a continuous layer on the Ludox surface. Hence, the g value obtained from TRFA analysis can be used to assess the absolute surface coverage of monolayer coatings on colloidal nanoparticles.

Introduction (TRFA)1,2

Time-resolved fluorescence anisotropy has been recently extended to in situ monitoring of growth and surface modification of colloidal silica systems.3-7 For TRFA studies, silica nanoparticles (NPs) are dispersed in water and labeled ionically with a micromolar amount of the cationic fluorescent dye rhodamine 6G (R6G).3 The R6G molecule forms multiple bonds with the oppositely charged surface of the silica NP.4 Such tight binding locks any local motion of the adsorbed dye and the decay of R6G anisotropy actually reflects the Brownian tumbling of the whole silica NP.5 The R6G decay is collected using time-correlated single photon counting instrumentation and then fitted to the model of rigid spheres.4 The fit generates numerical values for the fast correlation time φ1, attributed to rotation of free, silica-unbound R6G, the slow correlation time φ2, attributed to rotation of small silica NPs of radius R < 2.5 nm and their respective fractions (f1 and f2), in addition to the fraction g of the nondecaying component φ3 . φ2 attributed to * To whom correspondence should be addressed. Tel: (905) 525-9140 (ext. 27033). Fax: (905) 527-9950. E-mail: [email protected]. Internet: http://www.chemistry.mcmaster.ca/faculty/brennan. (1) Steiner, R. F. Topics in Fluorescence Spectroscopy; Plenum: New York, 1991; Vol. 2, p 1. (2) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Klu¨wer Academic Plenum Publishers: New York, 1999. (3) (a) Tleugabulova, D.; Duft, A. M.; Brook, M. A.; Brennan, J. D. Langmuir 2004, 20, 101. (b) Tleugabulova, D.; Zhang, Z.; Chen, Y.; Brook, M. A.; Brennan, J. D. Langmuir 2004, 20, 848. (c) Sui, J.; Tleugabulova, D.; Brennan, J. D. Langmuir 2005, 21, 4996. (4) Tleugabulova, D.; Sui, J.; Ayers, P.; Brennan, J. D. J. Phys. Chem. B 2005, 109, 7850. (5) Geddes, C. D. J. Fluoresc. 2002, 12, 343. (6) Tleugabulova, D.; Duft, A. M.; Zhang, Z.; Chen, Y.; Brook, M. A.; Brennan, J. D. Langmuir 2004, 20, 5924. (7) Tleugabulova, D.; Zhang, Z.; Brennan, J. D. J. Phys. Chem. B 2003, 107, 10127.

slow tumbling of larger silica NPs of radius R > 2.5 nm, which are the main fraction of Ludox particles. According to the model of rigid spheres,4 the redistribution of R6G anisotropy between φ1, φ2 and the nondecaying component reflects the real-time equilibrium between free R6G in solution and that bound ionically to silica NPs. This methodology has allowed insight into silica particle growth (NP metrology approach),5,6 sol-to-gel evolution,7 microviscosity within aqueous silica monoliths,7,8 dynamics of entrapped biomolecules,9 as well as the monitoring of silica surface modification3 in diluted sols of Ludox AM-30 (average radius, 6 nm; 30 wt % SiO2; pH ) 8.9).10,11 Ludox has been used as a model of colloidal silica in numerous studies.3,5,12-14 The adsorption of nonfluorescent watersoluble polymers,3a organosilane precursors,3b or cationic peptides3c onto Ludox NPs displaces the adsorbed R6G from the silica surface to solution and leads to a decrease in g and an increase in f1. Although the dependence of f1 and g on the modifier concentration in the bulk Ludox formally generates the relative degree of silica surface modification,3c a clear understanding how the rotational properties of R6G are linked to the surface coverage of silica by the nonfluorescent modifier is still missing. To bridge this gap, the TRFA principle was reconciled here with the Langmuir theory of adsorption.15 For this purpose, we compared our recent TRFA data for R6G/Ludox modified with (8) Ferrer, M. L.; del Monte, F.; Levy, D. J. Sol-Gel Sci. Technol. 2003, 26, 353. (9) Sui, X.; Cruz-Aguado, J. A.; Chen, Y.; Zhang, Z.; Brook, M. A.; Brennan, J. D. Chem. Mater. 2005, 17, 1174. (10) DuPont, Product information. Ludox colloidal silica. (11) van der Meeren, P.; Saveyn, H.; Bogale Kassa, S.; Doyen, W.; Leysen, R. Phys. Chem. Chem. Phys. 2004, 6, 1408. (12) Milosavljevic, B. H.; Meisel, D. J. Phys. Chem. B 2004, 108, 1827. (13) Vertegel, A, A.; Siegel, R. W.; Dordick, J. S. Langmuir 2004, 20, 6800. (14) Lundqvist, M.; Sethson, I.; Jonsson, B. H. Langmuir 2004, 20, 10639. (15) Langmuir, I. Chem. ReV. 1933, 13, 147.

10.1021/la0526941 CCC: $33.50 © 2006 American Chemical Society Published on Web 01/05/2006

Surface CoVerage of Colloidal Silica

the cationic peptide Lys-Trp-Lys (KWK)3c to the actual adsorption isotherm of KWK on macroporous silica CPG-3000 particles of radius R ) 17-35 µm as determined by the centrifugation method, which allowed us to calculate directly the surface coverage of CPG (KWK molecules/nm2). The need to compare Ludox to CPG arises because neither TRFA nor centrifugation studies are amenable to both systems. The classical centrifugation approach is not amenable to studies involving Ludox, because an unknown fraction of these particles do not sediment during centrifugation. Thus, it is not possible to properly assess binding to the silica surface using this system. On the other hand, it is not possible to perform TRFA studies for R6G when bound to CPG, since CPG is composed of large particles which sediment from water. Ludox is widely recognized as an excellent model of colloidal silica.3,5,11,12,16 Both CPG and Ludox particles possess identical surface properties, given that the adsorption sites (SiO-) are of the same nature and their density distribution is independent of the particle size or type of amorphous silica.17 In addition, the maximum density of ionic binders at saturation is also independent of particle size.12 A comparative analysis between KWK/CPG and KWK/Ludox systems, monitored by the centrifugation method and TRFA, respectively, emphasizes the similarity and distinctive features of the information available from the two techniques. The data also show that it should be possible to assess the absolute surface coverage of nanoparticle surfaces using TRFA, which may have important ramifications in the development of nanoscale materials or devices. Experimental Section Chemicals. Lys-Trp-Lys acetate salt (KWK), rhodamine 6G, and melittin from honey bee venom (MW 2846.46; g 97% pure by HPLC) were obtained from Sigma (St. Louis, MO). Ludox AM-30, 30 wt % SiO2 in water (d ) 1.210, average particle radius of 6 nm, specific surface area of 220 m2/g)10 was purchased from Aldrich (Milwaukee, WI). Controlled pore silica CPG-3000 (particle size, 35-70 µm; mean pore diameter, 297 nm; specific surface area of 8.4 m2/g) was purchased from CPG Inc. (Lincoln Park, NJ). All water was distilled and deionized using a Milli-Q Synthesis A10 water purification system. All other reagents were of analytical grade and were used as received. Procedures. All samples were prepared by diluting the commercial Ludox in 5 mM Tris-HCl, pH 7.4 containing 0-5 mM KWK, as described previously.3c The samples contained a fixed amount of silica (0.90 wt % SiO2 in 2.0 mL buffer; calculated surface area, 3.993 m2). After equilibration for at least 5 min at 295 K, 1 µM R6G was added to the KWK/Ludox mixture. R6G binds electrostatically to Ludox NPs, which are negatively charged over the entire pH range18 and have both deprotonated and protonated silanol groups.16,19 The same procedure was used to prepare melittin/Ludox samples containing 0-140 µM melittin. The upper concentration limit of the modifier corresponded to the critical flocculation concentration of Ludox. The KWK structure was built in extended form using the Insight/ Biopolymer molecular modeling package (INSIGHT-II; Biosym Technologies). To reduce atomic overlaps and to relax torsional and dihedral constraints, the peptide model was energy minimized using the Insight/Discover molecular modeling package. The minimization consisted of 100 iterations with a conjugate gradient method in the gas phase, neglecting electrostatic interactions. The minimization did not significantly alter the extended structure of KWK and gave dimensions of 17 Å × 11 Å × 7 Å. Since the adsorption process involves both the -NH3+ and R-NH3+ groups of the peptide, the (16) Iler, R. K. The Colloidal Chemistry of Silica; American Chemical Society: Washington, DC, 1994. (17) Zhuravlev, L. T. Langmuir 1987, 3, 316. (18) Kosmulski, M. Colloid Surf. A 2003, 222, 113. (19) van Blaaderen, A.; Vrij, A. The Colloid Chemistry of Silica; American Chemical Society: Washington, DC, 1994; p 99.

Langmuir, Vol. 22, No. 4, 2006 1853

Figure 1. Adsorption isotherm for KWK on CPG-3000 in 5 mM Tris-HCl, pH 7.4. projection of the extended KWK molecule onto the silica surface corresponds to a surface area AKWK ≈ 9 × 12 ) 108 Å2 ) 1.08 nm2. KWK Adsorption on CPG-3000. The adsorption isotherm of KWK on controlled pore silica CPG-3000 was obtained by means of a standard centrifugation method to measure the surface concentration of KWK. A small amount (∼ 0.01 g) of the CPG powder was placed in a preweighed 2 mL Eppendorf tube, and the exact CPG weight (( 0.00001 g) was recorded. 1.5 mL-aliquots of KWK solutions (0-3.5 mM) in 5 mM Tris-HCl, pH 7.4 buffer were added to the Eppendorf tubes containing CPG. Preliminary experiments showed that, within 5 min after mixing of KWK and CPG, adsorption reached a steady state. The samples were agitated for 20-30 min at room temperature and centrifuged for 15 min at 8000 rpm to remove CPG. The absorption intensity of the supernatant at 280 nm was measured with a Cary 400 UV-visible spectrophotometer and the value was extrapolated on the KWK calibration curve to determine the peptide concentration in the supernatant, Ceq (M). The adsorbed amount per unit surface area, Γ(mol/m2), was calculated from the difference between the added KWK concentration (C0, M) and that measured in the supernatant after removal of CPG by centrifugation (Ceq), using mass balance considerations and the specific surface area of CPG-3000 Γ)

(C0 - Ceq)V mSCPG

(1)

where V ) 1.5 × 10-3 L is the sample volume, m is the exact amount of CPG in each sample in grams, and SCPG ) 8.4 m2/g is the specific surface area of CPG. The error in the calculated surface concentration was estimated using a Gaussian error propagation method.20 The experimental Γ values were plotted as a function of the equilibrium concentration of KWK in solution (Figure 1) to obtain a simple Langmuir adsorption isotherm15 Γ ) Γmax

KCeq 1 + KCeq

(2)

where Γmax (mol/m2) is the maximum surface coverage of CPG3000 by KWK and K (M-1) is the adsorption equilibrium constant at 295 K in 5 mM Tris-HCl, pH 7.4 buffer. The values of Γmax and K were determined by fitting the experimental data to eq 2 (solid line in Figure 1) using nonlinear regression routines available in the Sigma Plot 2000 program. TRFA. Time-resolved fluorescence intensity and anisotropy decay data for R6G (λex ) 495 nm; λem ) 551 nm) were acquired in the time-domain with an IBH 5000U time-correlated single photon counting fluorimeter using procedures that are described in detail elsewhere.4 The anisotropy decay was fit to a two-component hindered (20) Barrante, J. R. In: Applied Mathematics for Physical Chemistry; Prentice Hall: New York, 1974; p 173.

1854 Langmuir, Vol. 22, No. 4, 2006

TleugabuloVa and Brennan

rotor model according to the following equation:21 r(t) ) f1r0 exp(-t/φ1) + f2r0(-t/φ2) + gr0

(3)

where r0 is the limiting anisotropy at t ) 0, φ1 (φ1 < φ2) reflects isotropic rotation of free R6G in solution, φ2 reflects slower reorientation of R6G bound to silica particles ( 2.5 nm in radius) that rotate more slowly than can be measured with the R6G probe (φ > 60 ns) and thus is assessed as “nonmotion”.22 It should be noted that the value of gr0 is equivalent to the residual anisotropy, r∞, which is the anisotropy at t f ∞. The fluorescence lifetime of R6G is identical in the free and adsorbed states7 and does not depend on the adsorption of KWK to the silica surface,3c and thus, the fraction of fluorescence is equivalent to the fraction of probe in each state. Fits were considered acceptable if the reduced chi-squared (χR2) was close to 1.0 and the residuals showed a random pattern.

Results and Discussion The TRFA principle for the KWK/Ludox analysis relies upon a dramatic difference in the rotational mobility between the free and Ludox-bound R6G dye. Upon increasing the total KWK concentration (C0 from eq 1) in the Ludox sol,3c the fraction of Ludox-bound R6G decreases that produces a change in the shape of R6G decay. However, the decay itself possesses no information regarding KWK (Ceq and Γ in eq 2). To gain more insight into the link between the R6G anisotropy and the surface coverage of Ludox by KWK, we need to compare the R6G decay to the adsorption isotherm of KWK in the Ludox sol. The adsorption isotherms of molecules at liquid/solid interfaces (eq 2) are usually measured by means of the centrifugation method, which provides the surface saturation limit, the surface area occupied by one adsorbed molecule at saturation and the strength of binding.15 As noted earlier, Ludox NPs are too small in size to be removed from the equilibrated KWK/Ludox mixture by centrifugation, and thus, the traditional centrifugation method cannot provide the adsorption isotherm of KWK on Ludox NPs. However, the centrifugation method is applicable if Ludox is substituted by larger silica particles, such as CPG-3000. We assume that the adsorption isotherm of KWK onto Ludox NPs is identical to that obtained using micro-size CPG particles. Ludox and CPG are both amorphous (synthesized by precipitation of silica) and thus expose the same hydrophilic OH-terminated surface to the added peptide. On both adsorbents, the deprotonated silanol groups form a lattice of noninteracting, identical active sites. The density of these sites is independent of the particle size,12 type of silica,17 or synthesis procedure used to obtain it.23 On the Ludox surface, the active sites are totally accessible to KWK and solvent molecules due to the large surface area-tovolume ratio and the absence of any internal surface. Similarly, the accessibility of adsorption sites on CPG is provided by large and uniform pores (mean pore diameter, ∼300 nm). All of these aspects suggest that the adsorption behavior of KWK in both CPG and Ludox suspensions should be similar. Adsorption Isotherm of KWK on CPG-3000. Figure 1 shows the adsorption isotherm for KWK on CPG particles. The plateau is reached at Ceq ∼ 1 mM KWK. The fit of the experimental data to eq 2 gave r2 ) 0.927 (solid line in Figure 1), a maximum (21) Geddes, C. D.; Karolin, J.; Birch, D. J. S. J. Phys. Chem. B 2002, 106, 3835. (22) Tleugabulova, D.; Sui, J.; Ayers, P.; Brennan, J. D. J. Phys. Chem. B 2005, 109, 7850. (23) Heston, W. M.; Iler, R. K.; Sears, G. W. J. Phys. Chem. 1960, 64, 147.

Figure 2. Molecular model of KWK. The fully extended form has molecular dimensions of ca. 9 Å × 12 Å.

surface density of Γmax ) 1.6 × 10-6 mol/m2 and an equilibrium constant of K ) 1.5 × 104 M-1 (295 K, 5 mM Tris-HCl, pH 7.4). The area occupied by one KWK molecule on the CPG surface calculated from the measured adsorption density is 1.1 nm2 (1.6 × 10-6 mol/m2 × 6.023 × 1023 molec/mol ) 9.63 × 1017 molec/ m2 ) 0.96 molec/nm2, or 1.1 nm2/molec). The area estimated from molecular modeling of the fully extended KWK structure is 1.08 nm2 (see Figure 2). Hence, the saturation coverage of CPG corresponds to a tightly packed peptide monolayer. Apparently, such tight packing does not occur if the peptide sequence is extended to 14 or 75 lysine units, as shown by the adsorption study of poly-L-lysine on negatively charged polystyrene particles in aqueous solution.24 The surface area that each polylysine molecule occupies is nearly 1 order of magnitude larger than the size of the molecule in its extended form. The authors24 suggested that the low adsorption density is likely a result of Coulombic repulsion between the positive charges on the amino acid units of polylysine, which caused the peptide to extend away from the silica surface. The high adsorption density of KWK on the CPG surface at saturation coverage suggests that there is no intermolecular Coulombic repulsion between adsorbed KWK molecules, and thus, the NP surface can be efficiently coated by the tripeptide monolayer. The measured adsorption constant of 1.5 × 104 M-1 for KWK on CPG is characteristic for molecules that adsorb strongly to silica surfaces.25 For example, adsorption of a polyamidoamine dendrimer onto the surface of fused silica through two primary amino groups exhibits an equilibrium constant that is the same order of magnitude, K ) 5 ( 1 × 104 M-1, as that of KWK.25 Adsorption of KWK onto CPG is expected to occur through electrostatic interactions between protonated -NH3+ and R-NH3+ groups of the peptide and deprotonated silanols on the silica surface.26,27,28,29 In addition, the flexibility of KWK side chains allows them to rearrange for greater contact with the silica surface. Modeling of the lysine residue near a silica surface indicates that the -NH3+ and R-NH3+ groups are situated at distances of ∼1.5 and 3 Å, respectively, from the silica surface. At such a close (24) Eckenrode, H. M.; Dai, H.-L. Langmuir 2004, 20, 9202. (25) McCain, K. S.; Schluesche, P.; Harris, J. M. Anal. Chem. 2004, 76, 930. (26) Gambino, G. L.; Lombardo, G. M.; Grassi, A.; Marletta, G. J. Phys. Chem. B 2004, 108, 2600. (27) Adalsteinsson, H.; Maulitz, A. H.; Bruice T. C. J. Am. Chem. Soc. 1996, 118, 7689. (28) Buemi, G. J. Mol. Struct. (THEOCHEM) 2000, 499, 21. (29) Flodstrom, K.; Wennerstrom, H.; Alfredsson, V. Langmuir 2004, 20, 680.

Surface CoVerage of Colloidal Silica

proximity to the silica surface, the KWK molecule is totally immersed in the first hydration layer (thickness ∼1 nm based on the hydrated radius of 7 nm for Ludox calculated from TRFA data)5 adjacent to the silica surface. Hence, the adsorbed KWK peptide will be surrounded by the electrical double layer, which is ∼4.3 nm in thickness under the ionic strength conditions used in our work. This provides additional stabilization of the complex formed by three ionic bonds and multiple hydrogen bonds between the primary amino groups of KWK and the silanols on the silica surface. This adsorption model is supported by tryptophan anisotropy measurements3c that indicate almost complete restriction in the backbone mobility of Ludox-bound KWK. R6G Adsorption vs KWK Adsorption. Adsorption of R6G at water/silica interfaces has been the subject of several independent studies, which elucidated different aspects of the R6G adsorption process, such as the thermodynamics,30,31,32 molecular geometry, and dynamics4 of the R6G:silica complex. In a global context, all of these studies provide a concise picture of the strong R6G-silica interaction. At low surface coverage (Γ < < Γmax), R6G adsorbs as a monomer and the adsorption isotherm is Langmuirian (eq 2).30-32 The equilibrium constant is K ) 4.1 × 105 M-1 (298 K, 5 mM KCl, pH 7),32 as shown by single-microparticle injection and microadsorption techniques. The high K value indicates that the adsorbed R6G would not be easily displaced from the silica surface. Previous studies by our group3,4,6,7 and Geddes and Birch5 indicate that R6G interacts with silica via electrostatic interactions. A similar situation also holds for KWK.3c Studies by Geddes have shown that R6G does not bind to silica in the presence of high levels of methanol,5 whereas our group has shown that the R6G:silica interaction is disrupted at high ionic strength (above 10 mM),7 ruling out hydrogen bonding as a significant bonding interaction and highlighting the dominance of ionic interactions in the R6G: silica system. The adsorbed dye adopts a flat orientation, lying with the xanthene plane nearly parallel to the silica surface, as determined by second harmonic generation interferometry/linear dichroism measurements30,33,34,35 and X-ray diffraction/elemental analysis/ polarized UV-visible spectroscopy analysis.36,37,38 The strong binding of R6G to silica surfaces and the flat orientation of adsorbed R6G molecules are in excellent agreement with the absence of wobbling motion for R6G on hydrated silica surfaces, as determined by TRFA analysis and molecular modeling.4 Because of the flat orientation, the surface area occupied by one R6G molecule can be calculated from its molecular dimensions (1.6 nm × 0.8 nm)39 or from the hydrodynamic radius of the R6G sphere, RR6G. The RR6G value can be calculated from the rotational correlation time of the R6G sphere in water,40 which is measured by TRFA, or from the translational diffusion coefficient provided by fluorescence correlation spectroscopy (FCS).31 The three independent calculations give an R6G surface (30) Simpson, G. J., Rowlen, K. L. Anal. Chem. 2000, 72, 3407. (31) Leng, X., Starchev, K., Buffle, J. Langmuir 2002, 18, 7602. (32) Sekine, T., Nakatani, K. Langmuir 2002, 18, 694. (33) Kikteva, T.; Star, D.; Zhao, Z.; Baisley, T. L.; Leach, G. W. J. Phys. Chem. B 1999, 103, 1124. (34) Gruzdkov, Y. A., Parmon, V. N. J. Chem. Soc., Faraday Trans. 1993, 89, 4017. (35) Ishibashi, K.; Sato, O.; Baba, R.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J. Colloid Interface Sci. 2001, 233, 361. (36) Sasai, R.; Fujita, T.; Iyi, N.; Itoh, H.; Takagi, K. Langmuir 2002, 18, 6578. (37) Bujda´k, J., Iyi, N., Kaneko, Y., Czı´merova´, A., Sasai, R. Phys. Chem. Chem. Phys. 2003, 5, 4680. (38) Vieira Ferreira, L. F.; Lemos, M. J.; Reis, M. J.; Botelho do Rego, A. M. Langmuir 2000, 16, 5673. (39) Mubarekyan, E.; Santore, M. Langmuir 1998, 14, 1597. (40) Tleugabulova, D.; Zhang, Z.; Brennan, J. D. J. Phys. Chem. B 2003, 107, 10127.

Langmuir, Vol. 22, No. 4, 2006 1855

Figure 3. Equilibrium KWK concentration as a function of total KWK concentration in Ludox (0.90 wt % SiO2).

area of 1.28 nm2 (molecular dimensions), 1.96 nm2 (TRFA), and 1.86 nm2 (FCS). The first value is close to the cross-sectional area (1.25 nm2) of the most stable molecular conformation of R6G obtained from computer simulations in the gaseous phase,41 whereas the other two values correspond to the surface area of a water-solvated R6G sphere. Since the present study is carried out in aqueous solution, an average surface area of 1.9 nm2 is chosen to calculate the saturation surface coverage of Ludox (radius, 6 nm; 0.90 wt % SiO2) by R6G molecules. On the basis of a surface area of 1.9 nm2/molecule, we calculate a theoretical monolayer saturation value for Ludox NPs at a R6G surface density of 5.3 × 1017 molecules/m2 or Γmax ≈ 8.7 × 10-7 mol/m2. Under the conditions of the TRFA experiment used in the present work, ΓR6G ) (0.95 × 10-6 M × 0.0020 L)/3.993 m2 ) 4.8 × 10-10 mol/m2, given that 95% of the added R6G (1 µM) is bound to Ludox at 0.90 wt % SiO2, pH 7.4.3c This means that the dye occupies only 0.055% of the available Ludox surface, and thus, the use of R6G as an adsorption sensor probe does not modify the surface properties of Ludox or interfere with KWK adsorption. Assuming that the surface of one Ludox NP of average radius R ) 6 nm is A ) 4πR2 ) 4.52 × 10-16 m2, the number of Ludox NPs (omitting polydispersity) in our samples is roughly 3.993/4.52 × 10-16 ) 8.8 × 1015 particles. Hence, we estimate that no more than one R6G molecule is adsorbed per every eight Ludox particles in the absence of KWK. This is important for TRFA analysis since the binding of several R6G molecules per one silica NP may lead to FRET-based depolarization, which would cause a loss of orientational information. R6G Decay in KWK/Ludox vs Saturation Coverage in KWK/CPG. The centrifugation method gives the adsorbed amount of KWK as a function of the equilibrium peptide concentration, Ceq, whereas the TRFA decay is monitored as a function of the total KWK concentration, C0. In both measurements, C0 is the common known parameter regarding KWK. Clearly, at any point in the adsorption isotherm, the value of Ceq is given by C0 - Cads, where Cads is the concentration of KWK adsorbed. Beyond ∼0.2 mM KWK, Cads becomes constant and thus C0 and Ceq become proportional (Ceq ) C0 - constant, see Figure 3). Since the surface chemistries of Ludox and CPG are similar, one should expect a similar degree of adsorption per unit surface area on both surfaces. The initial concentration of KWK used in the centrifugation and TRFA experiments are different because Ludox has a much larger surface area than CPG. Thus, to allow direct comparisons between these systems, it is necessary to normalize the initial concentration to the surface area of the silica system under study. Using C0, we can approximate the total amount of KWK added (41) Churaev, N. V.; Sergeeva, I. P.; Sobolev, V. D.; Derjaguin, B. V. J. Colloid Interface Sci. 1981, 84, 451.

1856 Langmuir, Vol. 22, No. 4, 2006

TleugabuloVa and Brennan

Figure 4. Fractions f1 (open circles), f2 (triangles), and g (closed circles) of R6G anisotropy in KWK/Ludox and the surface amount of KWK on CPG-3000 silica (diamonds) as a function of the total moles of KWK per unit surface area. R6G, 1 µM; Ludox, 0.90 wt % SiO2. Dotted line indicates the flocculation limit for KWK in Ludox.

to Ludox or CPG per surface unit area, N (mol/m2)

N ) C0V/mS ) (Γ - CeqV/mS)

(4)

where m is the amount of silica in grams, S the specific surface area of the adsorbent, and V the sample volume. In the Ludox sol, m ) 1.815 × 10-2 g, S ) 220 m2/g,10 and V ) 2.0 mL, N ) 5.01 × 10-4 × C0 mol/m2. In the CPG suspension, V ) 1.5 mL, S ) 8.4 m2/g, and m varies from sample to sample. The normalized N parameter was used to plot the Γ values for KWK/CPG and the f1, f2, and g fractions of R6G anisotropy in KWK/Ludox on the same graph (Figure 4). In the absence of KWK (N ) 0), ∼95% of the emission intensity of R6G corresponds to the Ludox-bound form (f2 + g ) 0.95).3c With increasing values of N, the g value decreases to almost zero, whereas f1 and f2 increase. When f1, f2, and g reach their respective plateau values, the KWK/Ludox system approaches the flocculation limit (dotted line in Figure 4). The flocculation of Ludox at high values of N reflects the ability of KWK to efficiently screen the Ludox surface charge. As a result, the double-layer repulsion existing between plain Ludox NPs is reduced. As discussed above, the adsorbed KWK molecules likely form a thin adsorption monolayer less than 1 nm in thickness, which is not able to keep the point of closest approach outside of the range of the attractive vander Waals forces, which is typically 2.5 nm.4,10 For R6G binding to occur, the Ludox NP should expose a free adsorption site (SiO-) and at least a 2 nm2 area of free surface surrounding it. Thus, adsorption to larger particles should occur preferentially over smaller particles with insufficient surface area to allow multipoint binding of the KWK molecule. Hence, the condition g ≈ 0 indicates that the Ludox surface charge is efficiently screened by KWK molecules. In previous studies, the density of adsorption sites on Ludox for ionic binding was estimated to be ∼1.2 groups per nm2.11 Based on this, the condition g ≈ 0 also means that the Ludox surface is evenly covered by KWK and there is no available silica surface for R6G binding. The decrease in the g value leads to increases in both f1 and f2. The f1 value is related to the nonadsorbed R6G that is free in solution. As KWK is added and R6G is displaced, it is expected that there will be a higher amount of probe available to redistribute between the free state and the surface of small silica particles. Thus, the increase in f1, which follows the trend observed for the decrease in g, is fully consistent with the higher level of free probe in solution upon blockage of the surface of large particles by KWK. By definition, f2 corresponds to the fraction of R6G-labeled primary Ludox NPs of radius R < 2.5 nm. In general, the primary silica NPs of radius R ) 1-2 nm are formed by approximately 20 tetrahedral units of monosilicic acid, Si(OH)4, which are joined together through siloxane bonds.5 The NP interior can be regarded as dense SiO2, whereas the NP surface exposes more than 50% of the Si atoms. Hence, there should be a few surface-exposed silanol groups available for KWK adsorption. As g deceases and f1 increases, one should expect a higher driving force for adsorption of R6G to small particles and, thus, an increase in f2. This will be balanced by binding of KWK to small particles, which will partially block the surface. The f2 fraction does not reach zero but instead increases upon increasing the amount of KWK until a plateau value is reached. This means that R6G is able to bind to the primary Ludox NPs independent of the presence of KWK. The coadsorption of R6G and KWK on the same NP would occur, if the adsorbed KWK does not form a continuous monolayer, which might be related to the surface curvature effect, and the lower surface area of the smaller NPs.13 The continuous layers are likely formed on larger Ludox NPs (g ≈ 0), which cause R6G molecules to redistribute to the uncovered areas on smaller Ludox NPs, thus explaining the gradual increase in f2 as the Ludox modification with KWK proceeds (Figure 4). Recent studies13,14 have shown that the NP size and curvature have dramatic effects on the secondary structure of adsorbed proteins as well as on the stoichiometry of protein-silica binding. The fact that even a small peptide, such as KWK, exhibits differential binding to small versus larger silica NPs is relevant in this respect. Based on the TRFA data in KWK/Ludox mixtures, KWK is not able to form monolayers on silica particles of radius 0) in the R6G decay, which indicate the formation of continuous adsorption layers on Ludox NPs of radius R > 2.5 nm and submonolayers on NPs of radius R < 2.5 nm. This kind of information is not available from the centrifugation method, which provides the maximum adsorbed amount but does not prove the continuity or efficiency of molecular coatings in polydisperse systems. As shown here, this can be successfully addressed by the TRFA method. For small peptides, the concentration of the solute at which g reaches zero corresponds to the monolayer coating of large silica NPs. Based on the trends in g values as a function of the solute concentration, it is possible to assess whether continuous monolayers can be formed and what fraction of the NPs can be covered under the conditions employed, at least up to the flocculation limit. This kind of characterization will be useful in the analysis of biocompatible surfaces for medical applications. The present work follows a trend in recently published adsorption studies13,24,26,30 that tries to reconcile the information available from sophisticated spectroscopic measurements with the classical Langmuir theory. The fact that the TRFA data does follow the Langmuir adsorption model provides a useful basis for interpreting such data when applied to different model systems. Future work will focus on extending the TRFA model to more complex systems, including larger solutes (proteins, polypeptides) and sol-gel derived silica monoliths, to better assess the nature of biomolecule:surface interactions in such systems. Acknowledgment. The authors thank the Natural Sciences and Engineering Research Council of Canada, MDS-Sciex, the Canadian Foundation for Innovation and the Ontario Innovation Trust for financial support of this work. The authors are also grateful to Dr. K. Nakatani from the University of Tsukuba for providing the adsorption parameters for R6G, Ms. J. Sui for performing the adsorption studies with melittin, and Dr. W. Czardybon from McMaster University for molecular modeling studies. J.D.B. holds the Canada Research Chair in Bioanalytical Chemistry. LA0526941