Surface Charge Control through the Reversible Adsorption of a

Article pubs.acs.org/JPCB

Surface Charge Control through the Reversible Adsorption of a Biomimetic Polymer on Silica Particles Emily D. E. Hyde,† Roberto Moreno-Atanasio,† Paul A. Millner,‡ and Frances Neville*,§ †

School of Engineering and §School of Environmental & Life Sciences, The University of Newcastle, Callaghan, New South Wales 2308, Australia ‡ Faculty of Biological Sciences, The University of Leeds, Leeds, LS2 9JT, U.K. S Supporting Information *

ABSTRACT: The control of the physicochemical properties of silica particles is of paramount importance to achieve full functionality in specific applications. A novel facile method of silica particle synthesis, requiring only two reactants, was developed. Control of the surface charge of these newly synthesized silica particles was achieved via the rapid electrostatic adsorption and acidic desorption of the branched, biomimetic polymer, polyethylenimine (PEI). Successful adsorption/desorption of PEI was supported by ATR-FTIR spectra, an adsorption isotherm, and ζ-potential curves. PEI adsorption above a threshold PEI concentration was determined to categorically change the topography of the silica particles’ ζ-potential curve. The results from our study convey a rapid, reversible, and reliable method of silica particle surface charge control. This may be of particular use in tailoring surface interactions of silica or silica-coated particles for applications in drug delivery, biomedical technologies, catalysis, and coatings.

■

a nucleation site for silica formation.22,23 This modified Stöber method has been shown to produce spherical particles with a narrow size distribution.7 However, the expense of the polyamine catalyst may make this biomimetic modified Stöber process difficult to viably scale up. Modification of the silica surface has been extensively studied in the literature, especially in relation to increasing colloidal silica suspension stability.24−28 The relative inertness of the silica surface means that surface modifications are made either by conditioning the surface for covalent attachment or by electrostatic adsorption at the appropriate pH.4 The most common form of silica surface modification uses a covalent siloxane linkage mechanism which provides a robust method to adhere organic materials to the silica surface.29−31 However, modification via covalent bonding is an irreversible process and could be detrimental in some applications due to the reduction in system flexibility. Electrostatic adsorption is a reversible process allowing the modifying agent to be attached and stripped from the surface by changes in the solution pH or salt concentration. PEI is a biomimetic cationic polymer and has been used successfully to modify silica4,24,32−36 and other particle surfaces.37−40 However, a direct and comprehensive study of the surface charge shifts caused by PEI modification of silica nanoparticles has yet

INTRODUCTION Silica nanoparticles have attracted considerable research attention due to their large number of current and potential applications in areas such as drug delivery,1−4 biomedical technologies,5−8 catalysis,9,10 and optimization of coatings.11,12 Stö ber et al.13 developed a robust method of silica production that allowed the synthesis of spherical, monodisperse, micron-size silica particles. Stöber’s technique13,14 involves the hydrolysis and condensation polymerization of a tetraalkoxysilane (generally tetraethyl orthosilicate (TEOS)) in a water/ethanol medium using ammonia as a basic catalyst to promote condensation and polymerization.15 In 2001, Dingsøyr and Christy16 were able to successfully replace the ammonia catalyst with sodium hydroxide (NaOH). However, in both cases, the harsh chemicals and organic solvent used may become a safety and environmental concern for the commercial scaling up of the Stöber process. The issues associated with the Stöber method have prompted a number of recent studies7,17−22 to modify the technique, taking inspiration from biological silication processes. These studies demonstrated that polybasic peptide mimics, such as polyethylenimines (PEIs), are efficient basic catalysts due to the multiple amine groups they contain.22 The modified Stöber method developed by Neville et al.7 requires the prereaction acidic hydrolysis of the less hazardous silane precursor, trimethoxymethylsilane (TMOMS). The polymerization of the hydrolyzed silane is then performed at room temperature and physiological pH buffered conditions using a conglomeration of the biomimetic polymer, PEI, and phosphate ions as © 2014 American Chemical Society

Received: October 4, 2014 Revised: November 23, 2014 Published: December 26, 2014 1726

DOI: 10.1021/jp5100439 J. Phys. Chem. B 2015, 119, 1726−1735

Article

The Journal of Physical Chemistry B

achievable with PEI adsorption. This high degree of surface charge control will allow the surface chemistry of modified particles to be precisely tailored for specific applications.

to be published. Neither has there been a study characterizing the PEI adsorbed to the silica surface via infrared spectroscopy. Lindquist and Stratton35 investigated the stabilizing effect of PEI adsorption on Ludox AM colloidal silica particles. The amount of PEI remaining in the supernatant was determined spectroscopically in order to calculate the amount of polymer adsorbed on the silica colloidal particles.35 Mészáros et al.36 investigated the kinetics of PEI adsorption on a silica wafer surface using reflectometry to determine the amount adsorbed. These authors36 also briefly considered the reversal of surface charge associated with PEI adsorption and the effect of different electrolyte concentrations using a zeta (ζ)-potential curve. During these studies the PEI concentrate remained as the medium with no attempt to remove the supernatant from the modified surface, so the stability of the coating could not be deduced from this study.36 There was also no attempt to expand the study of surface charge to include the ability to desorb PEI molecules nor investigate the dependency on PEI concentration.36 Electrostatic adsorption of PEI has been successfully employed to neutralize the negative surface charge of the silica surface in silica lined capillaries to improve the performance of capillary electrophoresis analysis of proteins.33,34,41 Erim et al.33 found the electrostatically adsorbed PEI coating to be stable for a pH range of 3−11. Silica or silica-coated nanoparticles have enjoyed widespread success as drug delivery agents mainly due to their low toxicity and potential biodegradability.4,42,43 One major obstacle for effective drug delivery via inorganic nanoparticle carriers is the ability for the carriers to pass through cell barriers. Thus, the flexibility of the silica surface chemistry allowing the drug delivery systems to be modified with different functionalities depending on their target cell or tissue is another highly favorable characteristic of silica particles.4,32,38,39,44 PEI functionalization of nanoparticles provides a promising solution by introducing positively charged amine groups shown to protect the carrier drug against enzymatic digestion and assist the transportation of the drug carrier into the cell.45−48 In fact, it has been proposed that with careful surface charge control silica particles may serve the dual purpose of drug and nucleic acid delivery.49 Due to the negative surface charge of nucleic acid, effective delivery of the acid necessitates the positive modification of the silica surface charge, thus allowing it bind to DNA and siRNA.4 Xia et al.4 studied the effectiveness of PEI (with molecular weights between 0.6 and 25 kDa) coated (via electrostatic adsorption) mesoporous silica particles as a drug delivery agent. PEI-coated particles were found to possess a high binding effectiveness and a high rate of cellular uptake making them promising candidates for an enhanced drug delivery system.4 In our study, the polyamine polypeptide mimic used by Neville et al.7 was replaced by NaOH. Hence our silica synthesis required only two chemicals, prehydrolyzed TMOMS and the NaOH catalyst, to be reacted at room temperature within an aqueous solution. By reducing the number of chemicals required, the novel method of silica synthesis detailed in our study provides a cheaper, simplified alternative to other silica particle production methods. Infrared and UV−visible spectroscopies and ζ-potential measurements were used to characterize the silica particles with adsorbed and desorbed PEI. A ζ-potential study on the effect of PEI concentration during adsorption on particles demonstrated the high level of surface charge control

■

METHODS AND MATERIALS Materials. Trimethoxymethylsilane, hydrochloric acid (HCl), sodium hydroxide, Na2 HPO 4 ·2H 2 O, NaH 2 PO 4 , branched 25 kDa polyethylenimine, and sodium nitrate (NaNO3) were purchased from Sigma-Aldrich and used without further purification except where noted. Ultrapurified water (18.2 mΩ) was used to make all solutions and for washing of particles and glassware. Solid NaOH and NaNO3 were used to make 1 M solutions and were diluted where needed. A 290 mM phosphate buffer was prepared by adding a combination of 290 mM Na2HPO4·2H2O and NaH2PO4 dissolved in water until pH 7.4. NaOH−TMOMS Silica Particle Synthesis. The silane precursor, TMOMS, was hydrolyzed by incubation in 1 mM HCl for 15 min. The concentration of the working solution of TMOMS was 0.1 M. The particles were synthesized in 1 mL aqueous batches by thoroughly mixing 0.2 M (final concentration) hydrolyzed TMOMS with 25 mM (final concentration) NaOH and reacting the mixture for 30 min at room temperature. The reaction was stopped by centrifugation of the particles at 16500g for 1 min and removal of the supernatant. The synthesized NaOH−TMOMS particles were then washed twice with water. The average yield of NaOH−TMOMS particles synthesized using this method was calculated using a sample set of no less than five 1 mL reaction tubes. The yield was defined as the mass of silica particles per milliliter of reaction. The percentage yield was calculated as the weight percent of hydrolyzed TMOMS consumed in the reaction. Particles were synthesized in tubes of known mass and desiccated for more than 3 days before mass measurements were taken. PEI−TMOMS Silica Particle Synthesis. For the purpose of comparison, PEI−TMOMS particles were made using the method outlined by Neville et al.7 An increased final concentration of 0.2 M hydrolyzed TMOMS was used during the synthesis (as opposed to the 0.1 M hydrolyzed TMOMS used in previous work,7,23), corresponding to the hydrolyzed TMOMS concentration used for the production of NaOH− TMOMS particles in this study. PEI Electrostatic Adsorption and Desorption. The NaOH−TMOMS particles undergoing modification and a stock solution of 1 mM PEI were prepared and stored at 3 °C. Generally, PEI was adsorbed by resuspending one 1 mL tube of NaOH−TMOMS particles in 145 mM pH 7.4 phosphate buffer (PB) and different concentrations of PEI (0−0.500 mM). The suspended particles were incubated at room temperature in the PB/PEI solution for 1 h with mixing at 15 min intervals to retain the particles in suspension. The PEI remaining in the supernatant after the required reaction time was removed after centrifugation for 30 s at 16500g, and the modified particles were washed twice with water to remove any remaining unbound PEI. For PEI desorption the particles were resuspended in 2 mM HCl and incubated for 1 h followed by washing twice with water. Unmodified and Modified NaOH−TMOMS Particle Characterization. Scanning Electron Spectroscopy (SEM). Particles were prepared for SEM by drying drops of the particle suspensions directly onto aluminum stubs. A Zeiss Sigma field 1727

DOI: 10.1021/jp5100439 J. Phys. Chem. B 2015, 119, 1726−1735

Article

The Journal of Physical Chemistry B

were measured intermittently throughout the experiment to account for any drifting. A standard curve of PEI (25 kDa) concentrations within the linear range between 1 × 10−3 and 1 × 10−5 mM was created by diluting the stock PEI to the required amount and adding PB and CuSO4. After the addition of all reactants the mixture was shaken and allowed an incubation time of 1 min before the absorbance measurement was made. The supernatant from each PEI concentration incubation and subsequent washes was saved and diluted until the absorbance at 275 nm was within the linear range dictated by the standard curve. For each sample the intrinsic amount of phosphate buffer was calculated taking into account the dilution and if required additional phosphate buffer was added to the cell to bring the total concentration of PB up to the constant 4.83 mM used during the standard measurements. The adsorption measurements were carried out in triplicate and the results shown are the average of the measurements, with the standard deviation given as the error bars. The Beer−Lambert law was applied by using the equation of the calibration curve to determine the PEI concentration in the supernatants after the adsorption measurements had been carried out. Assuming all the PEI not adsorbed to the particles was found in the supernatant and wash fractions, the PEI adsorbed could be calculated by subtracting the measured amount from the initial concentration used during incubation. The adsorption isotherm was fitted using the Boltzmann sigmoidal growth equation (Origin lab software fitting module).

emission gun scanning electron microscope (FEGSEM) was used for analysis of the particles, and no sputter coating was required. The SEM images were analyzed for particle size using the Image Tool (UTHSCA) software. Dynamic Light Scattering (DLS). DLS measurements were made by using a Malvern Zetasizer Nano ZS instrument. Intensity fluctuations measured by the scattering of a laser beam directed through the sample allows the Brownian motion of the particles to be detected. The Stokes−Einstein equation is then used to relate the measured Brownian motion to particle size. The intensity average (hydrodynamic diameter) size distributions were obtained. The approximate particle concentration used was 0.5 g/L. For each particle type the mean intensity average was calculated from the average of nine DLS measurements made for each one of the six separate sample tubes. The error of each data point was calculated as the standard deviation of all DLS measurements of that particular sample type. Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR-FTIR). ATR-FTIR measurements were performed using a Spectrum Two (PerkinElmer) instrument. FTIR allows for the identification of specific molecular groups, by detecting the wavelengths of infrared radiation absorbed by molecular bond vibrations. In ATR-FTIR the IR beam is internally reflected by a crystal; the IR absorbed by the sample can be detected by the evanescent wave which extends beyond the crystal and into the sample which is placed in contact with the crystal. An air background was taken before measurements on the dried particles were taken. For the unmodified, adsorbed, and desorbed types of particles, six samples of each were prepared, and before measurements, all particles of the same treatment were combined and desiccated for 3 days before being further dried in an oven at 110 °C for 24 h. The error of the measurement was taken as the precision limit of the instrument used as given by the manufacturer, which was 4 cm−1. ζ-Potential Measurements. ζ-Potential measurements were made with Malvern ZetaSizer Nano ZS instrument. Within the software, the Henry equation was used to convert the electrophoretic mobility to the ζ-potential. Particles were suspended in 10 mM NaNO3 base electrolyte with pH adjustments made using 10 mM HCl or NaOH to maintain constant the ionic strength. The results for each data point are the average of at 30 DLS sample measurements made of three separate sample tubes. The error for each data point was taken as whichever gave the largest value between the ζ-deviation given by the ZetaSizer and the standard deviation of all measurements of a particular data set. PEI Adsorption Isotherm. The amount of PEI in the supernatant was detected spectroscopically40 via a method based on the work of Ungaro et al.50 and Wen et al.51 The selective chelation of PEI with copper(II) ion creates a complex with a maximum absorbance peak at 275 nm.50 Quartz cells were used to avoid cell absorbance at the wavelength of interest, and a Shimadzu UV-1800 spectrophotometer was used to make the absorbance measurements. In all experiments the cells contained a final concentration of 2 × 10−4 mM copper(II) sulfate (CuSO4) as the source of copper ion and 4.83 mM phosphate buffer (pH 7.4) to take into account the PB remaining in the supernatant after incubation. Reagent blanks (containing the working concentration of CuSO4 and phosphate buffer) were used for zeroing the instrument and

■

RESULTS AND DISCUSSION NaOH−TMOMS Silica Particle Synthesis and Characterization. Silica particles were synthesized using a modified version of the Neville et al.7 method. In both methods the silane precursor, in this case TMOMS, is treated to prereaction acid hydrolysis to increase the rate of particle formation. A basic catalyst is then added to the prehydrolyzed mixture to initiate the nucleation and growth of silica particles. In the new method proposed here, NaOH is added to the reaction as the catalyst for the silica particle formation instead of the PEI/phosphate buffer catalyst which is employed in the original PEI−TMOMS method.7 The advantages of the NaOH catalyst are the reduced number of reactants and the reduced cost of the NaOH species compared with the polymer synthetic catalyst PEI. In both cases a set reaction time of 30 min was employed, which according to Seyfaee et al.21 allowed PEI−TMOMS particles to grow to maturity. The NaOH−TMOMS particles synthesized were characterized by yield determination (as defined under Methods and Materials), DLS sizing, and SEM imaging. In Table 1 the average NaOH−TMOMS particle yield and size are tabulated and compared with the PEI/PB−TMOMS particles synthesized using the Neville et al.7 method. The mass yields of NaOH−TMOMS and PEI−TMOMS dried particles were obtained, as shown in Table 1. The low standard deviation of the yield measurements suggests the method produces a consistent yield of particles each time. A slightly higher yield was measured for the PEI−TMOMS particles. However, the slight advantage in particle yield via the Neville et al.7 method is offset by the expense of the PEI catalyst in contrast to the NaOH catalyst employed by our method. This increase in yield was also reflected in the percentage of reactants used in the reaction. In both cases approximately 50 1728

DOI: 10.1021/jp5100439 J. Phys. Chem. B 2015, 119, 1726−1735

Article

The Journal of Physical Chemistry B Table 1. Comparsion of the Physical Propeties of TMOMS Silica Particles Made with PEI/PB and NaOH physical property mass yield (mg of silica particles formed per mL of reaction) percentage yield (wt % of hydrolyzed TMOMS consumed in reaction) average intensity particle size (via DLS) (nm) average particle size (via SEM) (nm)

PEI/PB− TMOMS particles

NaOH− TMOMS particles

10.0 ± 0.3

9.0 ± 0.4

52

47

990 ± 30

950 ± 50

795 ± 100

685 ± 50

wt % TMOMS remained unreacted after the 30 min reaction time. Neville and Seyfaee23 observed that PEI−TMOMS particle diameter will plateau within approximately 15 min of reaction time. Hence the unreacted TMOMS is likely the result of a thermodynamic limitation as opposed to a kinetics problem caused by a reaction time restraint. The sizes of the particles were determined using a wet method and a dry method, namely DLS and SEM imaging, respectively. The NaOH−TMOMS particle hydrodynamic diameter distribution determined via DLS is plotted in Figure 1 with the average size given in Table 1.

Figure 2. Scanning electron microscopy image of desiccator dried NaOH−TMOMS particles. The scale bar is 1 μm.

Silica particles were observed by SEM imaging (Figure 2) to be highly spherical. As expected52 the DLS sizing technique produced average particle sizes greater than those obtained via SEM image analysis, due to differences in the wet and dry methods, respectively. A slight roughness was observed on the outside of the particles (Figure 2). Seyfaee et al.21 attributes this roughness to the mechanism of particle formation via the aggregation of nanometer sized subparticles, a concept first proposed to explain Stöber method silica formation by Bogush and Zukoski.53 Hence, it is likely that a similar mechanism of particle growth is followed by the novel NaOH−TMOMS silica particles. The silica particles synthesized via the proposed NaOH− TMOMS method have been characterized based on particle size, size distribution, and yield. Comparison with PEI− TMOMS particles synthesized using the Neville et al.7 method has shown that our novel method of silica particle fabrication produced particles with similar characteristics. These similar properties are summarized in Table 1 and include comparable particle diameters (795 ± 100 nm (SEM) PEI−TMOMS particles and 685 ± 50 nm (SEM) NaOH−TMOMS particles) and mass yields (10.0 ± 0.3 g mL−1 PEI−TMOMS particles and 9.0 ± 0.4 g mL−1). However, the NaOH−TMOMS particles possess the advantages of requiring fewer reactants and possessing a simplified reaction method. The similarities in properties discussed in this work are likely to extend to the mechanism of formation. However, the determination of the similarity and differences of the growth mechanism may form the body of further research exploring this novel NaOH− TMOMS synthesis method. Surface Modification of NaOH−TMOMS Particles. Adsorption−Desorption of PEI. Modification of the particle surface was conducted through the electrostatic adsorption of PEI at pH 7.4 and its desorption at pH 2.6. The surface modification with PEI was characterized using ATR-FTIR spectroscopy, UV−visible spectroscopy (for the adsorption isotherm), and ζ-potential measurements. The ATR-FTIR spectra of the unmodified, PEI adsorbed, and PEI desorbed dried particles are presented in Figure 3. Note that the data are offset for clarity. In the region of wavenumbers below 1400 cm−1 (Figure 3a) the ATR-FTIR spectra of modified and unmodified particles were observed to be approximately equivalent. The 1400−500 cm−1 region mostly includes peaks related to the vibrational

Figure 1. NaOH−TMOMS hydrodynamic particle size distribution against percent intensity, as measured using dynamic light scattering (DLS).

The mean hydrodynamic diameter of NaOH−TMOMS particles, determined using DLS, was found to be 950 ± 50 nm (Table 1). In contrast, the PEI/PB−TMOMS particles reacted for the same period had a slightly higher average intensity diameter of about 990 ± 30 nm (Table 1), which is in good agreement with Seyfaee et al.21 The similar standard deviation of the NaOH−TMOMS and PEI/PB−TMOMS particles suggests a similarity in particle size distribution between the two methods. In Seyfaee et al.21 TMOMS concentration and reaction time were observed to be the main contributing factors determining PEI−TMOMS particle size. It was also observed that the catalyst (PEI/PB) concentration played only a small role in determining the equilibrium PEI−TMOMS final particle size, but significantly influenced the time taken for the particles to reach equilibrium. A scanning electron microscopy image of the NaOH− TMOMS particles produced using our new method is presented in Figure 2. The average size was determined via the Image Tool software (UTHSCA) and is given in Table 1. 1729

DOI: 10.1021/jp5100439 J. Phys. Chem. B 2015, 119, 1726−1735

Article

The Journal of Physical Chemistry B

vibration in this region of the spectra, had a signal at wavenumber 3402 ± 4 cm−1 (Figure 4). In contrast, the peak for the PEI adsorbed particles increased to 3455 ± 4 cm−1 due to the addition of the PEI amine groups, meaning both NH and OH stretching were observed. Hence the ATR-FTIR spectra indicated that the adsorption of PEI was successful. Acidic desorption of the PEI was also investigated using ATR-FTIR. The 3424 ± 4 cm−1 ATR-FTIR peak detected for the desorbed particles decreases in wavenumber from the adsorbed particle signal. This suggests that the PEI was desorbed when the signal was compared to the unmodified particles, which was also further confirmed using ζ-potential measurements. To further characterize the modified particles, ζ-potential measurements were made. In Figure 5a, the curves of ζ-

Figure 3. (a) Full ATR-FTIR spectra and (b) ATR-FTIR spectra between wavelengths 3900 and 3100 cm−1 of NaOH−TMOMS unmodified particles, PEI adsorbed NaOH−TMOMS particles, and PEI desorbed NaOH−TMOMS particles. The data are offset for clarity.

energies of the silica.23,54−56 Therefore, this similarity between the data was expected. The ATR-FTIR technique penetrates several micrometers (0.5−5 μm) into the sample;57 hence the signal related to the silica center of the particles is much stronger than the signal of the particle surface where the PEI coating can be detected. However, a significant shift in wavenumber was detected for the broad peak signal at approximately 3450 cm−1corresponding to the O−H and N−H stretching bonds54 which allowed distinction between modified and unmodified particles. These peak wavenumbers for the modified, PEI adsorbed, and PEI desorbed particles are plotted in Figure 4. The excitation of OH and NH bond stretching occurs at 3650−3200 and 3500−3200 cm −1 , respectively.54 The unmodified particles, which contain only OH stretching

Figure 5. (a) Experimentally determined ζ-potential curves and (b) calculated surface charge density, σζ (mC m−2), of unmodified, PEI adsorbed, and PEI desorbed NaOH−TMOMS particles suspended within a 10 mM NaNO3 base electrolyte with pH controlled by 10 mM HCl or NaOH for pH between 2 and 12. The experimental data (a) also show the ζ-potential curve of NaOH−TMOMS particles with PEI readsorbed after one cycle of adsorption−desorption.

potential versus pH for the unmodified, PEI adsorbed, PEI desorbed, and readsorbed particles are plotted. The ζ-potential data were also transformed into surface charge density using the method outlined by Delgado et al.58 The expression used to transform the ζ-potential, ψζ, to electrokinetic surface charge density, σζ, is given in eq 1,59 and the transformed data are given in the plot in Figure 5b. Figure 4. Signal peaks in the ATR-FTIR broad peak at approximately 3450 cm−1 for unmodified, PEI adsorbed, and PEI desorbed particles.

σζ = 1730

ψζε0εR (1 + κa) a

(1) DOI: 10.1021/jp5100439 J. Phys. Chem. B 2015, 119, 1726−1735

Article

The Journal of Physical Chemistry B where ε0 is the vacuum permittivity, εR is the relative permittivity of water, κ is the Debye constant, and a is the particle radius. The ζ-potential measurements of the unmodified NaOH−TMOMS silica particles (Figure 5a) correlate well with the ζ-potential measurements of a silica wafer measured via streaming potential.36 The silica surface comprises of two functional groups: siloxane groups (SiOSi) or silanol groups (SiOH).26 The silanol groups take on different forms dependent on the distance between similar groups, i.e., isolated silanols for far separations and vicinal silanols when the groups are close enough to form hydrogen bonds with their neighbors.26 Work by Ong et al.60 found one form of silanol site (isolated) to have a pKa of 4.5 and a prevalence of 19%, while the other type occupied 81% of silanol sites (vicinal) and had a pKa of 8.5. In agreement with these pKa values, Figure 5 shows that the unmodified NaOH−TMOMS particles became negatively charged at pH ∼4 and the ζ-potential dropped significantly at pH ∼9 due to the loss of protons from isolated and vicinal silanol groups, respectively. Published research has indicated the point of zero charge (PZC) of PEI is at pH 10.8.35,37 In addition, Wang et al.61 determined the ζ-potential of a PEI microsphere as a function of pH. The ζ-potential of the PEI microsphere61was found to correlate well with the ζ-potential of PEI adsorbed particles shown in Figure 5a and our PZC also agrees well with the value provided in the literature by Lindquist and Stratton,35 suggesting that the successful adsorption of PEI was achieved. Furthermore, the 0.200 mM concentration of PEI used during adsorption was enough to completely mask the silica silanol groups. The PEI adsorbed particles had a positive ζ-potential at pH greater than the PZC due to amine groups on the PEI gaining a proton. At pH less than 4 there is a decrease in the ζ-potential of the PEI adsorbed particles which may be linked to a decrease in PEI adsorption on Ludox colloidal silica particles when the pH during incubation was decreased as observed by Lindquist and Stratton.35 As the pH approaches the pKa1 (4.5) of the silica surface,60 the negative surface charge of silica decreases toward neutral and the electrostatic attraction between the PEI and the silica surface diminishes. The decrease in pH was also observed to increase the size of the PEI molecule as a consequence of the higher charge density and thus increased steric repulsion between adjacent branches of the PEI molecule.35 Lindquist and Stratton35 proposed that it is the combination of these two factors that leads to the decrease in amount of PEI adsorbed at pH below 4. The mechanism of desorption takes advantage of the neutralization of the silica surface at pH less than ∼4. The ζpotential of particles with PEI desorbed (Figure 5a) corresponds to the unmodified NaOH−TMOMS particles, showing how incubation in the acidic environment successfully desorbs the PEI surface coating. The PEI desorption suggested by the ATR-FTIR results (Figures 3 and 4) was corroborated by the ζ-potential results. However, the ζ-potential is most representative of the adsorption process since the ζ-potential is measured at the particle surface and the ATR-FTIR data are taken from throughout the whole particle sample. The ζ-potential curve of the readsorbed PEI particles, where adsorption was conducted after one cycle of PEI adsorption and desorption, correlated with the original PEI adsorption curve (Figure 5a). This result suggests that the process of

charge control via PEI adsorption−desorption is fully reversible. The electrokinetic surface charge density of the particles at various solution pH values was calculated from the ζ-potential data using eq 1 and is given in Figure 5b. The unmodified and desorbed particles were expected to possess a similar surface charge density as regular silica surfaces. Lindquist and Stratton35 investigated the surface charge density of Ludox silica nanoparticles using titration. Their results were used to verify our study. Adjusting to negate the contribution of aluminum ions found in Ludox nanoparticles, the authors determined the surface charge density of their silica particles decreased from approximately 0 to −25 mC m−2 with increasing pH from 3 to 11.35 In Figure 5b, the surface charge density of the unmodified and desorbed particles was determined as approximately 0 mC m−2 at pH below 4. At pH exceeding 4, upon the ionization of silanol surface groups, the silica surface gained a negative charge density reaching a maximum negative charge density of −7 to −10 mC m−2 at pH 11. Both charge density curves are largely similar, despite differences in the method of determination and the type of silica particle used, thus supporting the hypothesis that the silanol groups determine the surface charge in both the unmodified and desorbed particle systems. The surface charge density of the PEI modified particles is strongly positive, which is also expected from the ζ-potential results. In 2002, Mészáros et al.36 determined the surface charge density of silica wafer surfaces modified via PEI adsorption by both titration and transformation of the ζ-potential values. A maximum positive surface charge density occurred at pH 6 with surface charge densities of 10 and 20 mC m−2 for the titration and ζ-potential transformation measurement techniques, respectively.36 The PEI adsorbed particles from our study possessed a similar surface charge density maximum of approximately 10 mC m−2 at pH 6−8 (Figure 5b). The similarity in surface charge density between the literature and our study further supports the claim of successful PEI adsorption. Effect of PEI Concentration. To study the effect of PEI concentration used during adsorption, the ζ-potential curves were measured for PEI concentrations between 0.010 and 0.500 mM. For clarity, the full set of concentrations used is not shown in Figure 6. However, the full set of data is available in Figure S1 of the Supporting Information. In Figure 6, PEI concentrations of 0.125 mM and larger during the incubation stage were observed to have significantly positive ζ-potentials for pH values above the PEI PZC, as expected. PEI concentrations between 0.200 and 0.500 mM PEI (Supporting Information, Figure S1) were observed to have similar ζ-potential curves, indicating a PEI adsorption limit had been reached. In addition, the particles adsorbed with PEI concentrations from 0.125 to 0.500 mM had a reduced ζpotential when the pH was below 4, once again as expected due to the partial desorption of PEI from the silica surface at pH below the PZC of silica.35 PEI concentrations of 0.110 mM and lower gave negative/ neutral ζ-potentials for all pHs tested (Figure 6). However, even for the experiments with the smallest concentrations of PEI there was a clear increase in the values of ζ-potential compared to the unmodified particles between pH 5 and 10 (Figure 6). This suggests that, even for the smallest concentrations used, PEI was adsorbed and masked some of 1731

DOI: 10.1021/jp5100439 J. Phys. Chem. B 2015, 119, 1726−1735

Article

The Journal of Physical Chemistry B

concentration greater than the critical amount are assumed to take on the form of a PEI ζ-potential curve.61 For the cases of very low concentration (