pubs.acs.org/Langmuir © 2009 American Chemical Society
Silsesquioxane/Polyamine Nanoparticle-Templated Formation of Star- Or Raspberry-Like Silica Nanoparticles )
:: :: Lucy Kind,† Felix A. Plamper,‡,¥ Ronald Gobel,^ Alexandre Mantion,†,§ Axel H. E. Muller,‡ ,†,^,# ,† Uwe Pieles, Andreas Taubert,* and Wolfgang Meier* Department of Chemistry, University of Basel, CH-4056 Basel, ‡Macromolecular Chemistry II, University of :: :: Bayreuth, D-95440 Bayreuth, §Bundesantstalt fur Materialprufung and -forschung (BAM), D-12489 Berlin, :: Fachhochschule Nordwestschweiz, Hochschule fur Life Science, CH-4132 Muttenz, ^Institute of Chemistry, University of Potsdam, D-14476 Golm, and #Max-Planck-Institute of Colloids and Interfaces, D-14476 Golm. ¥ Current address: Laboratory of Polymer Chemistry, A.I. Virtasen Aukio 1, 00014 University of Helsinki, Finland.
)
†
Received January 19, 2009. Revised Manuscript Received February 19, 2009 Silica is an important mineral in biology and technology, and many protocols have been developed for the synthesis of complex silica architectures. The current report shows that silsesquioxane nanoparticles carrying polymer arms on their surface are efficient templates for the fabrication of silica particles with a star- or raspberry-like morphology. The shape of the resulting particles depends on the chemistry of the polymer arms. With poly(N,N-dimethylaminoethyl methacrylate) (PDMAEMA) arms, spherical particles with a less electron dense core form. With poly{[2-(methacryloyloxy)ethyl] trimethylammonium iodide} (PMETAI), star- or raspberry-like particles form. Electron microscopy, electron tomography, and small-angle X-ray scattering show that the resulting silica particles have a complex structure, where a silsequioxane nanoparticle carrying the polymer arms is in the center. Next is a region that is polymer-rich. The outermost region of the particle is a silica layer, where the outer parts of the polymer arms are embedded. Time-resolved ζ-potential and pH measurements, dynamic light scattering, and electron microscopy reveal that silica formation proceeds differently if PDMAEMA is exchanged for PMETAI.
Introduction Silica nanoparticles have gained increasing attention because they have favorable properties like low density (compared to other inorganics), good biocompatibility, and inertness to many substances. Moreover, they are easy to synthesize and functionalize. As a result, functional silica nanoparticles have been studied for applications in biomedicine and bioimaging,1,2 sensing and diagnostics,3,4 or in chemistry and engineering.5-7 Silica/organic hybrid materials add a further dimension to the flexibility and variability of silica materials. Silica/polymer hybrid materials combine the advantages of silica with the advantages of polymers such as pH and temperature responsiveness or different solubilities.8,9 For example, silica/polymer core shell particles are interesting for applications in biomedicine.
*Corresponding authors. A.T., Institute of Chemistry, University of Potsdam, Karl-Liebknecht-Str. 24-25, D-14476 Golm, Germany. Tel. 0049 (0)331 977 5773,
[email protected]. W.M., Department of Chemistry, University of Basel, Klingelbergstr. 80, CH-4056 Basel, Switzerland. Tel. 0041 (0)61 267 3802,
[email protected]. (1) Jovanovic, A. V.; Flint, J. A.; Varshney, M.; Morey, T. E.; Dennis, D. M.; Duran, R. S. Biomacromolecules 2006, 7, 945–949. (2) Ow, H.; Larson, D. R.; Srivastava, M.; Baird, B. A.; Webb, W. W.; Wiesner, U. Nano Lett. 2005, 5, 113–117. (3) Burns, A.; Ow, H.; Wiesner, U. Chem. Soc. Rev. 2006, 35, 1028–1042. (4) Wang, Q.; Liu, Y.; Lin, C.; Yan, H. Nanotechnology 2007, 18, 405604 (7 pp). (5) Suzuki, T. M.; Nakamura, T.; Akimoto, Y.; Kazuhisa, Y. Chem. Lett. 2008, 37, 124–125. (6) Mizutani, T.; Arai, K.; Miyamoto, M.; Kimura, Y. Prog. Org. Coating 2006, 55, 276–283. (7) Spahn, P.; Ruhl, T. Polym. Mater. Sci. Eng. 2006, 95, 65–66. (8) Zhang, Y.; Luo, S.; Liu, S. Macromolecules 2005, 38, 9813–9820. (9) Wu, C.; Szymanski, C.; McNeill, J. Langmuir 2006, 22, 2956–2960. (10) Nann, T.; Mulvaney, P. Angew. Chem., Int. Ed. 2004, 43, 5393–5396. (11) Darbandi, M.; Thomann, R.; Nann, T. Chem. Mater. 2005, 17, 5720–5725. (12) Ahn, B. Y.; Seok, S. I.; Baek, I. C.; Hong, S.-I. Chem. Comm. 2006 189–190.
Langmuir 2009, 25(12), 7109–7115
There are different methods to obtain polymer/silica :: hybrid particles. In contrast to the Stober10 or microemulsion 11,12 the template method is promising, because much approach, less surfactant or organic solvent is needed. Indeed, amine-based polymer templates, which act as in situ catalysts for silica hydrolysis and condensation, lead to silica shells with polymer cores.13 Among others, polymers containing aliphatic amines,14 pyridine, imidazoles,15 or polymers with a dendritic structure16-18 have been used as templates. For example, Khanal et al.19 templated silica with poly(styrene-block-2-vinyl pyridine-blockethylene oxide) (PS-PVP-PEO) triblock copolymer micelles. At low pH, the PVP is protonated and causes a selective deposition of silica precursor. The resulting particles have a diameter of 30 nm, and the PEO block prevents further aggregation. The main issue of this approach is that the synthesis and modification of block copolymers, in particular, multiblock copolymers, is often tedious and only yields small amounts of polymer. This prevents the fabrication of large amounts of uniform materials, which in turn prevents large-scale uses and sometimes even detailed studies. Moreover, if the polymer/silica hybrid material is to be used for, e.g., drug transport, the encapsulation of a small molecule will also need to be controlled by the multiblock copolymer. Depending on the nature of the encapsulant, this will require an additional chemical modification (13) Bergna, H. E.; Roberts, W. O. Colloidal Silica; CRC Press: Boca Raton, FL, 2006; Chapter 33, pp 397-424. (14) Yuan, J.-J.; Mykhaylyk, O. O.; Ryan, A. J.; Armes, S. P. J. Am. Cem. Soc. 2007, 129, 1717–1723. (15) Percy, M. J.; Barthet, C.; Lobb, J. C.; Khan, M. A.; Lascelles, S. F.; Vamvakaki, M.; Armes, S. P. Langmuir 2000, 16, 6913–6920. (16) Neofotistou, E.; Demadis, D. Colloids Surf., A 2004, 242, 213–216. (17) Knecht, M. R.; Wright, D. W. Langmuir 2004, 20, 4728–4732. (18) Knecht, M. R.; Sewell, L. S.; Wright, D. W. Langmuir 2005, 21, 2058–2061. (19) Khanal, A.; Inoue, Y.; Yada, M.; Nakashima, K. J. Am. Chem. Soc. 2007, 129, 1534–1535.
Published on Web 04/03/2009
DOI: 10.1021/la900229n
7109
Article
Kind et al.
Scheme 1. Sketch of the template (the silsesquioxane with 24 polymer arms) used in the current study and the mineralization reaction. Magnified view of the square region shows the structure of poly(N,N-dimethylaminoethyl) methacrylate (PDMAEMA) and poly{[2-(methacryloyloxy)ethyl] trimethylammonium iodide} (PMETAI).
of the multiblock copolymer. As a result, polymers that (i) are easy to manufacture in large quantities, (ii) support encapsulation of some cargo, and (iii) catalyze silicification are thus interesting candidates for the synthesis of (responsive and biocompatible) polymer/cargo/silica nanoparticles. The current report shows that, for the synthesis of raspberrylike polymer/silica hybrid particles, only a rather simple, starshaped polyelectrolyte is necessary. The polyelectrolyte poly(N, N-dimethylaminoethyl) methacrylate (PDMAEMA) or its quarternized equivalent poly{[2-(methacryloyloxy)ethyl] trimethylammonium iodide} (PMETAI) is attached to a silsequioxane core,20 yielding a 24-armed template, which, by its basic nature, catalyzes its own silicification. Depending on the chemistry of the arms (PDMAEMA or PMETAI), spherical particles with a lowdensity core or raspberry-like particles, which are rare in the literature,15,21,22 form (Scheme 1). The main advantage of our silsesquioxane/polymer template is that it is preorganized and, because all connections within the template are covalent, is not prone to decomposition or internal rearrangements.
Experimental Section Materials. The synthesis and characterization of the templates has been reported before.20 Here, we use a star-shaped poly(N,N-dimethylaminoethyl) methacrylate (PDMAEMA) and poly{[2-(methacryloyloxy)ethyl] trimethylammonium iodide} (PMETAI). Each polymer star has an arm number fn = 24 (number average; PDI in arm number distribution ≈ 1.4) and a number-average degree of polymerization per arm Pn(arm) = 240 (PDI of arms ≈ 1.3). The resulting average composition of the stars is (PDMAEMA240)24 and (PMETAI240)24. Tetraethyl orthosilicate (TEOS) (puriss., g99.0%, GC) was purchased from Fluka. Ultrapure water (resistivity of 18 MΩ/cm) from a Millipore water purification system was used in all experiments. For pH adjustments, 0.01 M HCl and 0.01 M NaOH were used. Nanoparticle Synthesis. One milligram of the respective star-shaped polymer was dissolved in 1 mL of water at pH 5.6. Then 5, 10, or 20 μL (22.4, 44.8, or 89.6 mM) of TEOS was added. The reaction mixture was stirred at room temperature for (20) Plamper, F. A.; Schmalz, A.; Penott-Chang, E.; Drechsler, M.; Jusufi, A.; :: Ballauff, M.; Muller, A. H. E. Macromolecules 2007, 40, 5689–5697. (21) Qiao, X.; Chen, Min; Zhou, Juan; Wu, L. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 1028–1037. (22) Chen, M.; Zhou, S.; You, B.; Wu, L. Macromolecules 2005, 38, 6411–6417.
7110 DOI: 10.1021/la900229n
170 h (7 days). The resulting transparent suspensions were directly used for further analysis. pH Measurements. pH vs the reaction time was determined with a Mettler Toledo MP 220 pH meter with an InLab 423 electrode. The pH meter was calibrated with buffer standard solutions of pH 4, 7, and 10. ζ-Potential and Size Measurements. ζ-potentials and hydrodynamic diameters were measured on a Malvern Instruments Zetasizer Nano ZS with a 4 mW He-Ne laser (633 nm) and a detection angle of 173. Size distributions were determined by fitting the experimental correlation curve to a multiexponential using the CONTIN algorithm. ζ-potentials were calculated via the Smoluchowski equation. Atomic Force Microscopy (AFM). Tapping-mode AFM was done on a Molecular Imaging PicoLE system with Super Sharp Silicon SFM-Sensors (SSS-NCHR-10, Nanosensor) with a tip radius of 2 nm, spring constant of 10-130 N/m, and resonance frequency of 204-497 kHz (values given by the manufacturer). All measurements were performed at very soft tapping conditions to minimize structural changes of the previously deposited star-shaped polymer. Additionally, imaging was done at low scan speeds (0.1-1 Hz) to minimize lateral forces exerted by the SFM tip on the sample. Samples were prepared from 0.01 wt % solutions in ultrapure water on mica. Scanning Electron Microscopy (SEM). Particle suspensions were diluted 100-fold with ultrapure water and deposited on a mica substrate. The dry samples were coated with platinum (4 nm) in a BalTec MED 020. Samples were imaged on a Hitachi S-4800 with a field emission source at 5 kV. Transmission Electron Microscopy (TEM). Particle suspensions were diluted 100-fold with ultrapure water. Five microliter aliquots of the suspension were deposited on a carboncoated 300 mesh copper grid. After 1 min, the remaining liquid was removed and samples were imaged with a Philips Morgani 268D with a tungsten source operated at 80 kV. Cryogenic Transmission Electron Microscopy (cryoTEM). Particle suspensions were diluted 100-fold with ultrapure water. Five microliter aliquots of the suspension were deposited on a lacey carbon-coated 300 mesh copper grid. Most of the drop was removed with blotting paper. Subsequently, the samples were shock-frozen in liquid ethane and held at approximately 90 K by a temperature-controlled freezing unit. CryoTEM experiments were done using a Zeiss 912 Omega Microscope operated at 120 kV. 3D Transmission Electron Microscopy (3D-TEM). Samples were prepared as above. The tilt series for 3D analysis were acquired in bright-field mode on a Tecnai F20 operated at Langmuir 2009, 25(12), 7109–7115
Kind et al.
Article
Figure 1. AFM images of PDMAEMA (top) and PMETAI (bottom) nanoparticles with the respective height profiles (right) at selected locations.
200 kV with a 0.23 nm point-to-point resolution and a high-tilt sample holder. Several series of images were recorded on a 2048 pixels 2048 pixels cooled CDD array detector in different locations of the grid containing the structures of interest. Samples were tilted from -65 to 65, with an image recorded every 2 between -40 and 40 and 1 elsewhere. During the acquisition, TEM parameters like defocusing, horizontal specimen shift, and specimen tilt were controlled automatically. Calibrated 5 nm diameter gold nanoparticles were added on the grid before acquisition to simplify data analysis. The Imod software23 was used for data processing, including alignment of all projections, corrections related to the geometry of the acquisition, and calculation of the volume from the series of projections using a back-projection technique,24 which leads to a 3D volume image of the sample. ImageJ25,26 was used for further data filtering and data reduction. Slicer27 was used to surface-render the tomograms. Small Angle X-ray Scattering (SAXS). SAXS curves were recorded at room temperature with a Nonius rotating anode instrument (4 kW, Cu KR radiation) with pinhole collimation and a MARCCD detector (pixel size: 79). The distance between sample and detector was 74 cm, covering a range of the scattering vector q = 2/λ sin θ = 0.04-0.7 nm-1 (θ = scattering angle, λ = 0.154 nm). 2D diffraction patterns were transformed into a 1D radial average of the scattering intensity. For measurements, samples were placed in quartz Mark tubes (Hilgenberg) with 1 mm in diameter and 10 μm wall thickness. Tubes were sealed (23) Kremer, J. R.; Mastronarde, D. N.; McIntosh, J. R. J. Struct. Biol. 1996, 116, 71–76. (24) Weyland, M. Top. Catal. 2002, 21, 175–183. (25) Rasband, W. S. ImageJ; National Institutes of Health: Bethesda, MD; http://rsb.info.nih.gov/ij/; 1997-2004. (26) Messaoudi, C.; Boudier, T.; Sorzano, C. O. S.; Sergio, M. BMC Bioinformatics 2007, 8, 288–297. (27) Slicer v 2; Massachusetts Institute of Technology Artificial Intelligence Laboratory and the Surgical Planning Laboratory at Brigham and Women’s Hospital, Boston, MA; www.slicer.org.
Langmuir 2009, 25(12), 7109–7115
before measurement. Concentrations of all samples were 2 mg/mL in water.
Results Structure of the Particles. Figure 1 shows atomic force microscopy (AFM) images of the two polymer stars (PDMAEMA and PMETAI, Scheme 1). The silsesquioxane cores of the stars have a height and diameter of 2 to 2.5 nm. The purpose of the silsesquioxane core is simply to provide the central anchor for the polymer arms, and indeed, all cores are surrounded by polymer arms. The average diameter of the PDMAEMA polymer stars (core plus polymer chains) is about 70 nm. This is in good agreement with values reported for PDMAEMA,20 where the root of the z-average of the mean square radius of gyration ÆR2gæ0.5 z is 29 nm, as determined by static light scattering (SLS) in acetone. For the PMETAI polymer star (∼80 nm), no reference value is available, but the PMETAI stars were prepared from the PDMAEMA polymer stars. Therefore, a similar diameter can be expected. Overall, AFM shows that the silsesquioxane/polymer stars are uniform and have a reasonably narrow size distribution. This is consistent with an earlier solution study on these stars.20 Figure 2 shows representative scanning (SEM) and transmission electron microscopy (TEM) images of the silica particles obtained after mineralization of the stars shown in Figure 1 with TEOS (Scheme 1). Both SEM and TEM show that relatively uniform nanoparticles form. The average diameter of the silica particles is ca. 25 and 50 nm for the PDMAEMA and the PMETAI stars, respectively. TEM also shows that the particles grown with the PDMAEMA stars have a bright dot in the center. The bright area is assigned to regions with less electron density. A less electron dense region is most likely due to the silsesquioxane core and some polymer. The diameter of the bright spot (2 to 3 nm) is DOI: 10.1021/la900229n
7111
Article
Kind et al.
Figure 2. (a) TEM image of silica core-shell nanoparticles based on PDMAEMA. (b) Cryo-TEM, (c) TEM, and (d,e) SEM images of silica core-shell nanoparticles based on PMETAI. All scale bars are 50 nm. Arrows point to less electron dense particle interior. Silsesquioxane concentration was 1 mg/mL, TEOS concentration was 44.8 mM and reaction time 3 days.
consistent with a silsesquioxane core of 2 to 2.5 nm, measured by AFM, in a polymer corona. Given the rather small diameter of the bright center, the polymer corona is most likely not entirely located in the bright dot, but partly encapsulated in the surrounding silica sphere obtained through the TEOS condensation reaction. The particles grown in the presence of PMETAI show a less common morphology in that they consist of a central, spherical particle carrying short arms. Depending on the sample preparation, different morphologies are observed. In cryo-TEM, that is, without drying after the mineralization reaction, the particles exhibit long arms. In conventional SEM and TEM, where samples are dried before microscopy, the particles exhibit bulges rather than real arms. As a result, we suspect that the arms on the particle surface retain some flexibility even after mineralization, which allow the shrinking or backfolding of the arms upon drying, but this hypothesis will need to be tested in the future. Figure 3 shows further data that were obtained on the PMETAI-based particles using 3D-TEM tomography. 7112 DOI: 10.1021/la900229n
Figure 3. (a) Representative TEM image from a tilt series of silica core-shell nanoparticles with PMETAI cores. The brighter inner parts (arrows) show regions of less electronic density. The dark dots (circles) are 5 nm gold particles used to match the single images for 3D reconstruction. The sample is aggregated due to drying on the TEM grid. (b,c) Surface renderings of the sample shown in (a). Panel (b) shows the silica particle surface. Panel (c) is the same tomogram as panel (b), but contrast was adjusted to better show the less electron dense core. Gold dots have been removed from the reconstruction in panels (b,c). Labels (L, R, A, P, I, S) are guides for the eye to simplify navigation in 3D. Panel (a) has been rotated such that all panels (a-c) have the same orientation. Silsesquioxane concentration was 1 mg/mL, TEOS concentration was 44.8 mM and reaction time 3 days.
3D-TEM tomography provides evidence that all particles have a less electron-dense core, although sometimes it is not clearly visible in conventional TEM images. The diameter of the central, low-electron-density region is 5 to 7 nm, which is slightly Langmuir 2009, 25(12), 7109–7115
Kind et al.
Figure 4. SAXS data of particles grown with the PDMAEMA and the PMETAI core particles.
larger than the diameters obtained from conventional TEM. The diameter of the silsesquioxane core particles is 2 to 3 nm.20 As a result, electron tomography shows that there is a gap of ca. 1 to 2 nm, which is presumably filled by the polymer, before the outer silica layer begins. Figure 4 shows small-angle X-ray scattering (SAXS) patterns of the samples shown above. Both samples show a q-4 decay, which is consistent with a two-phase system, such as the particles observed in electron microscopy suspended in a liquid. Patterns of samples grown with PDMEAMA show an additional q-0.5 and samples grown with PMETAI an additional q-1 decay. The latter can be assigned to the presence of rod-like features. Closer inspection of the q-0.5 decay suggests that it is in fact a q-1 decay with a small plateau. However, as the signal is weak and broad, we cannot unambiguously distinguish between the two. Moreover, as TEM shows that the particles contain cores with a lower electron density, interpretation of the SAXS patterns is not straightforward. The complex composition of the samples could lead to an overlap from hollow spheres and other features, and it is therefore not possible to extract more information from the SAXS data. Overall, AFM, SEM, TEM, and SAXS show that the silsesquioxane cores are efficient templates for the synthesis of silica nanoparticles containing a central region with a lower electron density. Depending on the chemistry of the polymer surface and the sample treatment, spheres or star-like particles form. In order to better understand the formation of these different species, timeresolved pH, ζ-potential, and TEM experiments were performed. Formation Mechanism of the Particles. Figure 5 shows the ζ-potential of the neat stars (Figure 1) in aqueous solution. The ζ-potential of the PMETAI stars is roughly pH-independent. In contrast, the ζ-potential of the PDMAEMA polymer stars decreases with increasing pH. This behavior can be assigned to the different states of charging in the two polymer stars: while PDMAEMA can take up or release protons via its amine groups, the quaternary ammonium groups in PMETAI cannot. Figure 6 shows time-resolved pH and ζ-potential data for the mineralization of PDMAEMA and PMETAI stars with three TEOS concentrations. With PDMAEMA, the pH of the initial polymer star solution is 8.7 and decreases to ca. 7.7 as the reaction proceeds. The ζ-potential initially increases to ca. +30 mV and subsequently decreases again. The decrease is more pronounced with higher TEOS concentration. The corresponding data for the PMETAI stars show that the pH rapidly decreases between 20 and 30 h from ca. 7.2 to 3-4. The decrease is followed by a very slow increase up to 170 h of reaction. The corresponding ζ-potential increases from an Langmuir 2009, 25(12), 7109–7115
Article
Figure 5. ζ-potential vs solution pH for both polymer stars (polymer concentration of 1 mg/mL). The horizontal line indicates the 0 mV potential.
initial value of +28 mV to +40 to +50 mV, followed by a concentration-dependent decrease. The strongest decrease to ca. +18 mV is observed for the highest TEOS concentration. Time-resolved TEM experiments show that the PDMAEMA stars lead to spherical particles with a brighter central part, which do not change anymore as the reaction proceeds. Silicification of the PMETAI stars initially also leads to spherical particles. After ca. 120 h, however, we observe the formation of the rod-like features shown in Figure 2. This suggests that, here, a different growth mechanism takes place (see discussion below). Figure 7 shows results from time-resolved dynamic light scattering investigations of particle growth. The bubble plots represent the hydrodynamic diameters Dh of samples grown in the presence of PDMAEMA and PMETAI stars vs the reaction time. The bubble size is proportional to the intensity-weighted fraction of particles with the respective Dh. There are multiple bubbles for most x-axis (time) values. This is due to the fact that in most cases more than one Dh population was observed, which is in agreement with TEM data. For example, the vertical line in panel (a) highlights a case at 55 h of reaction time where two populations with different Dh values are observed. Figure 7 indicates that there are significant differences between the two polymer stars. The particle growth with PDMAEMA stars as templates can be divided into three stages. During the first growth period up to 20 h, the average particle sizes are on the order of 50 to 100 nm. After a reaction time of 20 h, Dh increases, and the most dominant Dh populations are between 500 and 1000 nm. After ca. 55 h, the solution becomes turbid and a large Dh increase is observed. Here, the largest Dh populations are between 3000 and 5500 nm. Later on, sedimentation of the sample occurs, indicating that the large aggregates are colloidally unstable. In contrast, with PMETAI stars, Dh increases from ca. 20 nm to approximately 200-300 nm at 55 h. Thereafter, bimodal distributions with Dh1 around 100 nm and Dh2 between 250 and 380 nm are observed. Typically, Dh1 is the dominant population, although the populations are quite similar in size.
Discussion Structure of the Hybrid Particles after Silicification. AFM (Figure 1) shows that the silsesquioxane cores carrying the PDMAEMA or PMETAI chains are uniform and the cores have a diameter of 2 to 3 nm, consistent with literature.20,21 In a sense, the approach is similar to earlier reports, but as the experimental conditions (TEOS vs Na-silicate, concentrations, salt, or buffer solutions vs pure water, etc.) are different, it is difficult to quantitatively compare our work with these earlier DOI: 10.1021/la900229n
7113
Article
Kind et al.
Figure 7. Hydrodynamic diameters Dh of samples grown in the presence of PDMAEMA and PMETAI stars vs reaction time. Bubble diameters are proportional to the percentage of particles (weighted by intensity) with the respective Dh values. Vertical line in the top panel highlights the transition mentioned in the text. The data shown in the figure are for TEOS concentrations of 44.8 mM. Figure 6. pH and ζ-potential vs reaction time for PDMAEMA and PMETAI cores at different TEOS concentrations. (a) pH vs reaction time, (b) ζ-potential vs reaction time, (c,d) TEM images of samples isolated after different reaction times for both polymer stars. All scale bars are 50 nm. Samples shown in the TEM images were obtained with a silsesquioxane concentration of 1 mg/mL and a TEOS concentration of 44.8 mM.
reports.16-18 Conventional TEM and cryo-TEM (Figure 2) indicate that, with PDMAEMA, spherical particles with a diameter of ca. 25 nm and a bright center form. Cryo-TEM (Figure 2) further shows that with PMETAI star-like particles with a spherical core and ca. 10 nm long arms form. Moreover, SEM and TEM (Figure 2) suggest that the number of arms or bulges in the PMETAI samples correlates with the number of polymer chains on the silsesquioxane cores. This is however difficult to quantify, as we do not see the entire particle (surface) in the electron microscope, but only one side, and the number of arms has to be inferred from this half-view. Conventional TEM and SEM (Figure 2) show that upon drying the arms collapse onto the spherical core and result in a raspberry-like particle morphology. Similar to the particles grown with PDMAEMA, they have a less electron dense (bright) center. Electron tomography (Figure 3) confirms the presence of a less electron dense core in every particle after silicification. SAXS (Figure 4) confirms the structural differences between the two polymer chemistries. In the case of PDMAEMA, SAXS 7114 DOI: 10.1021/la900229n
detects a q-4 and q-0.5 decay. The latter could also be a q-1 decay preceded by a small plateau, but these two cases cannot be identified unambiguously. In the case of PMETAI, SAXS detects a q-4 and q-1 decay. The q-4 and q-1 decays can be assigned to a two-phase system such as particles in a suspension (q-4) and rod-like features (q-1). The reason for the q-0.5 decay is unclear at the moment, but the fact that it could also be a q-1 behavior suggests that the samples grown with PDMEMA contain protrusions and, possibly, that the surfaces of the particles are not perfectly flat. In summary, electron microscopy and SAXS show that the chemistry of the polymer chains on the silsequioxane cores has a strong influence on the structure of the particles formed after silicification. The different morphologies obtained with the two polymers can be explained as follows. Tertiary or quaternary amine or ammonium groups can act as an “inner” catalyst in the condensation of silica monomers.11,12,28,29 In the current study, the initial solution containing the PDMAEMA star polymers is basic (pH 8.7) showing that the PDMAEMA (pKa = 5.89)30 chains are only weakly protonated. As a consequence, the PDMAEMA chains are less extended but rather collapsed (28) Kang, S. M.; Lee, K.-B.; Kim, D. J.; Choi, I. S. Nanotechnology 2006, 17, 4719–4725. (29) Behrens, P.; Jahns, M.; Menzel, H. Handbook of Biomineralization WILEY-VCH Verlag GmbH & Co KgaA: Weinheim. 2007; Chapter 1, pp 3-18. :: (30) Plamper, F.A.; Ruppel, M.; Schmalz, A.; Borisov, O.; Ballauff, M.; Muller, A.H.E. Macromolecules 2007, 40, 8361–8366.
Langmuir 2009, 25(12), 7109–7115
Kind et al.
around the silsesquioxane core. In this case, spherical particles form upon silicification. The lighter region observed in the TEM is assigned to a less electron dense region caused by the silsequioxane core and parts of the (collapsed) polymer chains. The diameter of the core of ca. 5 to 7 nm suggests that parts of the polymer chains are buried in the outer silica layer, leading to a [silsequioxane/polymer/SiO2 + polymer] structure of the particles. The reason for the lower density of the core lies in the fact that we do not directly observe the silsesquioxane particle (which should have the same density as the other silica), but rather a “composite” central region, where the polymer arms also contribute to the overall electron density. This yields a central region with a density between that of the polymer and the silsesquioxane. With PMETAI, the striking difference is the presence of the ca. 10 nm rods. The reason for the formation of these rods can be found in the high charge density of the PMETAI chains, which is (unlike PDMAEMA) pH-independent. Due to the high number of charges on each chain, the polymer chains are not collapsed onto the silsequioxane cores, but extend into the solution, even at high pH. A fully extended chain is ca. 60 nm long. As a result, a silsequioxane with fully extended PMETAI chains should have a diameter of ca. 120 nm. However, fully extended chains are physically unrealistic. The rods observed in TEM and SAXS, which are only ca. 10 nm long, can therefore be explained with partly extended chains being mineralized. Due to the chargecharge repulsion, in ideal cases, each chain is mineralized individually. Overall, the silicified particles have a [silsequioxane/polymer/SiO2 + (partly extended) polymer] structure. Particle Growth Process. The pH of the PDMAEMA star solution is 8.7, and the pH of the PMETAI star solution is 7.2 at the beginning of the mineralization. ζ-potential measurements (Figure 5) show that the PDMAEMA stars only have a low potential of ca. +15 mV at mineralization pH (8.7). In contrast, the PMETAI stars have a higher potential of ca. +28 mV at mineralization pH (7.2). This finding supports the above hypothesis of collapsed PDMAEMA and extended PMETAI chains. The increase of the ζ-potential could be due to the fact that, with the PDMAEMA arms, the pH of the reaction mixture roughly reaches the pKa value of PDMAEMA, which could result in a more pronounced protonation of the amino groups of the PDMEMA. In contrast, the PMETAI is insensitive to protonation. Possibly, there is an additional effect of the iodide counterions, but the details of the initial increase are not clear at the moment. During mineralization of the PDMAEMA stars, the pH drops from 8.7 to ca. 7.7. This is reasonably close to the pKa value (5.89) of PDMAEMA. In contrast to the pH, the ζ-potential shows a strong dependence of the TEOS concentration vs reaction time. We explain this observation with the fact that the solution pH is buffered at ca. pH 7.7 by the PDMAEMA, which leads to deprotonation of surface silanol groups on the silica particles and hence a negative ζ-potential. Indeed, Pham et al. report a ζpotential of -20 mV for silica particles at pH 6.67,31 which is reasonably close to our measurements. The more TEOS is present in the solution, the more silanol groups are present and the more negative the overall potential becomes. During mineralization of the PMETAI star solution, the pH rapidly drops from ca. 7.2 to 3-4 between ca. 20 and 30 h of reaction time. Silica has an isoeletric point at around pH 2.6 to 3.31 The pure polymer solution has a ζ-potential of ca. +28 mV (Figure 5). Upon mineralization, the ζ-potential increases to (31) Pham, K.; Fullston, D.; Sagoe-Crentsil J. Colloid Interface Sci. 2007, 315, 123–127.
Langmuir 2009, 25(12), 7109–7115
Article
+40 to +50 mV and then decreases again. The decrease is concentration-dependent, with the strongest decrease being that of the reaction mixture with the highest TEOS concentration. We explain this behavior with an exchange of iodide ions against 13 silicic acid, SiO44 , or related species (active silica). As a result, iodide is released into the solution and the silicic acid species rapidly react to form SiO2. The resulting particles carry a number of positive charges, which are more and more obscured by the growing SiO2, and thus, the lowest positive potentials are observed for the highest TEOS concentrations. Time-resolved TEM experiments (Figure 6) show that the decrease in the pH and ζ-potential is not necessarily correlated with the particle shape. The main result of these experiments is that the rod-like features in the PMETAI samples develop relatively late in the process, suggesting that the PMETAI chains extend throughout the entire spherical shell and are able to template the rods later on, whereas the PDMAEMA is collapsed and may not extend throughout the whole silica shell. The hypothesis of PMETAI chains sticking out of the silica shells even late in the reaction is further supported by the fact that the particles grown with PMETAI show a much lower aggregation tendency than the samples grown with PDMAEMA. In the latter case, we observe aggregate sizes up to a few 1000 nm in diameter, and in the former case, the largest aggregates have a diameter of around 400 nm (Figure 7). From TEM, we also conclude that the features observed in DLS are indeed aggregates and not large single silica particles. Finally, the idea of a single polymer star template per particle is further supported by the fact that AFM shows star diameters of dPDMAEMA of ca. 80 nm and dPMETAI of ca. 100 nm. This is reasonably close to the sizes of the mineralized particles with dPDMAEMA of ca. 25 nm and dPMETAI of ca. 50 nm (TEM, Figure 2).
Conclusion The current paper presents a facile route to nanosized silica hybrid particles from star-shaped polyelectrolytes as templates at ambient conditions. Depending on the polymer chemistry, either spherical particles with a low electron density core or particles with ca. 10 nm arms form, where the number of arms is roughly related to the number of polymer chains on the template. All polymers are encapsulated in a silica shell formed by the condensation of TEOS. As the polymers catalyze TEOS hydrolysis, there is, similar to other polyamine-controlled silicification reactions, no need for additional catalysts such as ammonium hydroxide solution. ζ-potential and pH measurements suggest that the silica hybrid particles are responsive to external stimuli such as pH, which could be interesting for future applications. Acknowledgment. We thank G. Morson (U. of Basel) for the cryo-TEM images, D. Mathys (U. of Basel) for SEM measurements, Dr. O. Ersen (Institut de Genetique et de Biologie Moleculaire et Cellulaire de Strasbourg, CNRS - U. Louis Pasteur, Strasbourg) for 3D-TEM experiments and helpful discussion, the Institut de Genetique et de Biologie Moleculaire et Cellulaire de Strasbourg (IGBMC) for access to the TECNAI microscope, and Dr. V. Malinova (U. of Basel) and Dr. J Weber (MPI-KG) for helpful discussion. A.M. thanks the AdolfMartens-Fonds e.V. for an Adolf Martens Fellowship. This work was part of the EUROCORES Program SONS and financially supported by the NCCR Nanosciences and the Swiss National Science Foundation. DOI: 10.1021/la900229n
7115
