Article pubs.acs.org/Langmuir
In Situ Characterization of Binary Mixed Polymer Brush-Grafted Silica Nanoparticles in Aqueous and Organic Solvents by Cryo-Electron Tomography Tara L. Fox,† Saide Tang,‡ Jonathan M. Horton,§ Heather A. Holdaway,† Bin Zhao,*,§ Lei Zhu,*,‡ and Phoebe L. Stewart*,† †
Department of Pharmacology and Cleveland Center for Membrane and Structural Biology, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, Ohio 44106-4965, United States ‡ Department of Macromolecular Science and Engineering, Case Western Reserve University, 2100 Adelbert Road, Cleveland, Ohio 44106-7202, United States § Department of Chemistry, University of Tennessee, 1420 Circle Drive, Knoxville, Tennessee 37996, United States S Supporting Information *
ABSTRACT: We present an in situ cryo-electron microscopy (cryoEM) study of mixed poly(acrylic acid) (PAA)/polystyrene (PS) brush-grafted 67 nm silica nanoparticles in organic and aqueous solvents. These organic−inorganic nanoparticles are predicted to be environmentally responsive and adopt distinct brush layer morphologies in different solvent environments. Although the self-assembled morphology of mixed PAA/PS brush-grafted particles has been studied previously in a dried state, no direct visualization of microphase separation was achieved in the solvent environment. CryoEM allows the sample to be imaged in situ, that is, in a frozen solvated state, at the resolution of a transmission electron microscope. Cryoelectron tomograms (cryoET) were generated for mixed PAA/PS brush-grafted nanoparticles in both N,N-dimethylformamide (DMF, a nonselective good solvent) and water (a selective solvent for PAA). Different nanostructures for the mixed brushes were observed in these two solvents. Overall, the brush layer is more compact in water, with a thickness of 18 nm, as compared with an extended layer of 27 nm in DMF. In DMF, mixed PAA/PS brushes are observed to form laterally separated microdomains with a ripple wavelength of 13.8 nm. Because of its lower grafting density than that of PAA, PS domains form more or less cylindrical or truncated cone-shaped domains in the PAA matrix. In water, PAA chains are found to form a more complete shell around the nanoparticle to maximize their interaction with water, whereas PS chains collapse into the core of surface-tethered micelles near the silica core. The cryoET results presented here confirm the predicted environmentally responsive nature of PAA/PS mixed brush-grafted nanoparticles. This experimental approach may be useful for the design of future mixed brush-grafted nanoparticles for nano- and biotechnology applications.
■
INTRODUCTION Mixed polymer brush-grafted particles represent a novel class of environmentally responsive materials with a wide range of potential nano- and biotechnology applications.1,2 These particles are composed of two distinct, immiscible polymers covalently linked at one end to the surface of a core particle. The use of immiscible polymers with random or alternate linking sites to the core particle leads to phase separation, which tends to decrease unfavorable contacts between the two components. Because the polymers are grafted by one end to the core particle, macroscopic level separation cannot occur, and the result is microphase separation with features on the nanometer scale.3 The environmental responsiveness of these particles arises from the ability of the two grafted polymers to reorganize in reaction to environmental variations, such as a solvent change.4−6 After exposure to a new environment, the polymers spontaneously rearrange to achieve the lowest free© 2015 American Chemical Society
energy states. In this manner, mixed polymer brush-grafted particles can adopt different morphologies and present altered surface properties depending on the environmental conditions. In addition to environmental responsiveness, mixed brushgrafted nanoparticles can also self-assemble into various patchy particles with organized nanodomains on the surface, as revealed by recent computer simulations.7−11 Motivated by these computer simulations, we have carried out a series of morphological studies on the self-assembly of poly(tert-butyl acrylate)/polystyrene (PtBA/PS) mixed brushes grafted on silica nanoparticles.12−17 Because of the technical difficulty in grafting binary polymer chains from small sized silica particles (e.g., < 20 nm), the silica core size has been limited to 60−200 Received: May 11, 2015 Revised: July 7, 2015 Published: July 14, 2015 8680
DOI: 10.1021/acs.langmuir.5b01739 Langmuir 2015, 31, 8680−8688
Article
Langmuir nm.13,14 Under the melt condition or cast from chloroform (a neutrally good solvent for both PtBA and PS), the mixed brushes underwent lateral phase separation, producing a wormlike bicontinuous morphology.16 When treated with n-octane, a selective solvent for PtBA, the hairy particles showed a morphology of surface-tethered micelles with the core composed of collapsed PS chains.17 By removing the t-butyl groups in PtBA, amphiphilic mixed poly(acrylic acid) (PAA)/ PS brush-grafted silica particles were obtained.17 These amphiphilic particles could be uniformly dispersed in both methanol (a selective solvent for PAA) and chloroform (a selective solvent for PS). By adding water to the dispersion of the mixed brush particles in N,N-dimethylformamide (DMF), the morphology of mixed PAA/PS brushes on silica particles changed from a bicontinuous ripple structure to a surfacetethered micellar structure with isolated PS domains;17 however, the morphological characterization was carried out for mixed brush particles cast and dried on transmission electron microscopy (TEM) grids. A true in situ morphological characterization for mixed brushes in various solvents was not achieved due to technical limitations. Cryo-electron microscopy (cryoEM) is a technique originally developed for biological samples that enables imaging of specimens in a frozen, near-native environment with the resolution of TEM.18 The main advantage of cryoEM over conventional TEM is the elimination of damage or artifacts caused by dehydration and adsorption of the sample to a support film. Therefore, in situ structures of specimens can be directly visualized by cryoEM. Typically, cryoEM specimens are suspended in a thin layer of frozen buffer stretched across a hole in a lacy carbon grid or in a prefabricated grid with holes of defined size and spacing.19 The development of 3D image reconstruction methods has enabled the structure determination of icosahedral viruses, other symmetrical biological assemblies such as GroEL, and asymmetric particles, such as the ribosome, by averaging of thousands of 2D cryoEM particle images.20 A complementary technique, cryo-electron tomography (cryoET), enables 3D evaluation of samples that have a pleomorphous structure.20,21 This approach involves recording a tilt series of images covering a wide angular range and computing a 3D density map, called a tomogram. If regular macromolecular complexes within the tomogram can be identified, these can be computationally averaged to obtain higher resolution structural information with a technique known as subtomogram averaging.22 The inherently heterogeneous structure of mixed polymer brush-grafted particles precludes the use of averaging 2D cryoEM particle images to generate a 3D structure because each particle is distinct from the others. In addition, the structural heterogeneity of the polymers on the surface of the particles precludes the use of subtomogram averaging; however, cryoET without additional subtomogram averaging is feasible for mixed polymer brush-grafted particles. One challenging aspect of cryoET is the collection of a complete tilt series without causing significant electron beam damage to the frozen specimen. Although conventional electron tomography has been applied to characterize nanostructures in polymers,23−26 cryoEM has just started to be utilized to directly visualize selfassembled polymers in solution.27 To the best of our knowledge, no research has been reported for direct in situ visualization of mixed polymer brush-grafted inorganic nanoparticles.
In this work, the in situ structure of mixed poly(acrylic acid) (PAA)/polystyrene (PS) brush-grafted 67 nm silica nanoparticles was investigated by cryoET. The particles were dispersed in two types of solvents, DMF and water. DMF is a nonselective good solvent for both PAA and PS brushes, and water is a selective solvent for PAA. The 3D structures of mixed PAA/PS brush-grafted silica nanoparticles were reconstructed by tomography software. Distinct microphase separation patterns are observed for mixed brushes in DMF and the aqueous solution, and the environmentally responsive behavior of mixed PAA/PS brush-grafted silica nanoparticles is understood.
■
EXPERIMENTAL SECTION
Materials. Mixed PtBA/PS brush-grafted 67 nm silica nanoparticles were prepared as previously described.13 In brief, bare silica nanoparticles were obtained from Nissan Chemical. The average diameter of these nanoparticles was 67 nm, based on particle analysis by TEM. Mixed PtBA/PS brushes were synthesized by sequential atomic transfer radical polymerization (ATRP) of tert-butyl acrylate and nitroxide-mediated radical polymerization (NMRP) of styrene from Y-initiator-functionalized silica nanoparticles. The numberaverage molecular weight (Mn) and the dispersity index (DI) of PtBA were 22.2 kDa and 1.19, respectively. The Mn and the DI of PS were 18.7 kDa and 1.18, respectively. The grafting densities of PtBA and PS were 0.54 and 0.28 chains/nm2, respectively. Mixed PAA/PS brush-grafted 67 nm silica nanoparticles were prepared from the mixed PtBA/PS nanoparticles by removing tertbutyl groups in PtBA using trifluoroacetic acid (TFA). Typically, 250 μL of TFA was added to 3.5 mL of dry dichloromethane containing mixed PtBA/PS brush-grafted silica nanoparticle (0.3 mg/mL). After the reaction mixture was stirred at ambient temperature for 3 days, excess TFA was removed using a rotary evaporator. A control experiment using a free PtBA homopolymer showed that the reaction was complete after 3 days because the peak located at 1.42 ppm in the proton nuclear magnetic resonance (1H NMR) spectrum, which is attributed to the three methyl groups in PtBA, disappeared. The Mn and degree of polymerization (DP) of prepared PAA were 12.5 kDa and 173, respectively. There were no cross-linking or degradation side reactions, as previously reported.14 The DP of the PS chains thus remained at 180. The mixed PAA/PS brush-grafted silica nanoparticles can be well-dispersed in DMF, which is a good solvent for both PAA and PS. To prepare PAA/PS nanoparticles in an aqueous solution, we slowly added 2 mL of water to 0.5 mL of nanoparticle dispersion in DMF under magnetic stirring. DMF was removed by placing the mixture solution in a 3 mL of Slide-A-Lyzer dialysis cassette, followed by thorough dialysis in deionized water for 2 days. (The water was replaced every 8 h.) The concentration of the mixed brush nanoparticles in DMF and water was adjusted to ∼0.05 and 0.1 mg/ mL, respectively. CryoEM Grid Preparation. For mixed PAA/PS brush nanoparticles in either DMF or aqueous solution, a 20.0 μL aliquot of the nanoparticle solution was stained by the addition of 0.5 μL of filtered 1% uranyl acetate aqueous solution. This solution was mixed via pipetting, microcentrifuging, and sonicating and then allowed to equilibrate on ice for at least 30 min before preparation of the cryoEM grids. Bovine serum albumin nanogold tracer solution with 10 nm nanogold fiducial markers (Electron Microscopy Sciences) was resuspended in distilled water and stored at 4 °C before use. Shortly before preparation of the cryoEM grids, the resuspended nanogold solution was briefly microcentrifuged (3×) to remove aggregates. Quantifoil R2/2 holey carbon grids with 200 mesh copper (Electron Microscopy Sciences) were glow-discharged using an EMITech K100X Glow Discharge unit. Grids were glow-discharged for 30 s on each side. Some cryoEM grids containing mixed brush particles in aqueous solution were prepared using an in-house manual plunger, and the rest of the grids containing the sample in both DMF and H2O solvents were prepared using an FEI Mark IV Vitrobot. All grids were 8681
DOI: 10.1021/acs.langmuir.5b01739 Langmuir 2015, 31, 8680−8688
Article
Langmuir
extracted nanoparticles in DMF and aqueous solution, we extracted the silica cores separately, and their average radii were determined. The expected volumes for spheres with these average radii were calculated. The actual volume of the density in the cores at various isosurface levels was measured with the UCSF Chimera Measure Volume and Area tool. The isosurface level selected for display of the extracted particles was chosen so that 60% of the expected volume within the core was displayed. As the cryoET results showed, the silica cores were not uniformly dense. All Figures were made with Adobe Photoshop, Microsoft Excel, and UCSF Chimera. Quantitative Measurements. To determine the radial density distribution of the mixed brush layer for the extracted nanoparticles, we set the isosurface level to display 90% of the expected volume within the core. Spherical shells of density 5 nm thick within the mixed brush layer were extracted by erasing density inside and outside of nested spheres using the UCSF Chimera Volume Eraser tool. The volume of density within each spherical shell was measured. The total expected volume per spherical shell was calculated and used to convert the radial density volume measurements to percentages. To measure the ripple wavelength within the mixed brush layer for nanoparticles in DMF, we created density slabs of 5 nm thickness with the UCSF Chimera Clipping tool. Images of these density slabs were imported into Adobe Photoshop, and the Measurement tool was used to measure the spacing between the strongest PAA density regions at the half-height of the brush layer. Ripple wavelength measurements were made on four density slabs from three different extracted nanoparticles in DMF. The Photoshop Measurement tool was also used to measure dimensions of the nanoparticles, including the thickness of the crossover (i.e., mixing) layer and of the overall polymer shell.
prepared in a temperature and humidity-controlled (20% relative humidity) environment. For manual plunging, a 3.0 μL aliquot of the stained PAA/PS mixed brush particles was applied to a grid and allowed to adhere for approximately 25−30 s before the addition of a 1.0 μL of nanogold aliquot. Whatman #1 filter paper was used to blot the grid for ∼3 to 4 s. Grids were immediately plunged into liquid ethane as a cryogen and promptly transferred to liquid nitrogen. For samples plunged on the Vitrobot, a 3.0 μL aliquot of the stained PAA/PS mixed brush particles was applied to a grid and allowed to adhere for ∼30 s before the addition of a 1.0 μL nanogold aliquot. Grids were blotted three times using Whatman #41 filter paper for the DMF samples and standard Vitrobot filter paper (Ted Pella) for the aqueous samples. Whatman #41 filter paper was used for wicking away DMF because it is a fast filter paper, and DMF wicked away much more slowly than water. The faster wicking property facilitated the formation of thin frozen layers of DMF on the grids suitable for cryoEM imaging. The temperature and humidity within the Vitrobot plunging chamber were maintained at 3 °C and ∼85% relative humidity for DMF samples and 3 °C and 100% relative humidity for water samples. DMF samples were plunged into liquid nitrogen, and aqueous samples were plunged into liquid ethane. All of the cryoEM grids were kept under liquid nitrogen until insertion into a cryoelectron microscope for imaging. CryoEM Imaging. Cryo-electron micrographs were collected on an FEI Tecnai TF20 (200 kV, field-emission gun) cryo-electron microscope with a Tietz TVIPS TemCam-F416 (4096 × 4096 pixels) using a Gatan 626 cryo-holder. A nominal magnification of 50 000× (actual magnification 71 595×) was used for the sample in DMF and 80 000× (actual magnification 114 313×) for the samples in water. Microscope defocus settings were in the range of −2 to −3 μm. Cryo-Electron Tomography. Tilt series of the PAA/PS mixed brush particles in both solvents were collected on a JEOL JEM2200FS (200 kV, field-emission gun, and in-column energy filter, Tietz TVIPS TemCam-F416) cryo-electron microscope using Serial EM tomography software28 and a Gatan 914 cryo-holder. The selected tilt series of the PAA/PS mixed brush particles in DMF was collected with an angle range of −64 to +64° in 2° increments. The selected tilt series of the PAA/PS mixed brush particles in aqueous solution was collected with an angle range of −66 to +62° in 2° increments. All tilt series images collected on the JEOL microscope were at a nominal magnification of 50 000× with a defocus of −5 μm and an exposure time of 0.1 s. The electron dose per tilt series image collected on the JEOL microscope, as measured by Serial EM, was 0.6 to 0.7 e/Å2·s. Tomograms were generated from the tilt series using the eTomo program in the IMOD software package.28 Nanogold fiducial markers were used for alignment of the tilt series images. Aligned images were binned by a factor of 4 to produce a pixel size of 0.92 nm. Contrast transfer function (CTF) correction was applied before generation of the tomogram with IMOD. UCSF Chimera software was used to filter the tomograms using a median 5 × 5 × 5 voxel filter for two iterations for the particles in DMF and four iterations for the particles in water.29 The number of iterations was selected based on the feature size in the tomograms. Additional tilt series of the PAA/PS mixed brush particles in aqueous solution were collected on the FEI Tecnai TF20 cryo-electron microscope using the TVIPS EM-Menu 4.0 tomography software and a Gatan 626 cryo-holder. Images in the selected tilt series were collected with an angle range of −64 to +74° in 2° increments and at a nominal magnification of 80 000× with a defocus of −2 μm and an exposure time of 0.1 s. Aligned images were binned by a factor of 4 to produce a pixel size of 0.55 nm. Processing of the tilt series to generate a tomogram was performed similarly to that described for the tilt series collected on the JEOL microscope. UCSF Chimera software was used to filter the tomograms using a median 9 × 9 × 9 voxel filter for four iterations.29 A larger voxel volume was used for filtering of the TF20 tilt series to account for the finer pixel size in the TF20 tilt series images compared with the JEOL tilt series images. The density for individual particles was extracted and displayed using the UCSF Chimera tools Subregion Selection, Volume Eraser, and Hide Dust.29 To normalize the density isosurface values for
■
RESULTS AND DISCUSSION CryoEM Images of Mixed PAA/PS Brush-Grafted Nanoparticles in DMF. A representative cryo-electron micrograph of the mixed PAA/PS brush-grafted silica nanoparticles in DMF is shown in Figure 1A,B. As observed previously for similar nanoparticles, the silica cores have a size distribution between 50 and 90 nm.13 In general, cryo-electron
Figure 1. (A) Cryo-electron micrograph of mixed PAA/PS brushgrafted silica nanoparticles in DMF. (B) One enlarged nanoparticle in panel A. (C) Cryo-electron micrograph of mixed PAA/PS brushgrafted silica nanoparticles in water. (D) One enlarged nanoparticle in panel C. Both micrographs were collected on an FEI TF20 cryoelectron microscope. The PAA chains were stained by uranyl acetate to bear a darker contrast. The scale bars represent 100 nm. 8682
DOI: 10.1021/acs.langmuir.5b01739 Langmuir 2015, 31, 8680−8688
Article
Langmuir
Figure 2. (A) Tilt image (+38.2°) of mixed PAA/PS brush-grafted silica nanoparticles in DMF, stained with uranyl acetate, collected on the JEOL 2200FS microscope. The inset shows one enlarged particle (indicated by the blue circle). (B) Middle z slice of the same area from the tomogram. (C) 5 nm thick density slab from the middle of the particle circled in red in panels A and B from the 3D tomogram. The blue density shows the PAA domains and the red density indicates the silica core. The scale bars represent 100 nm in panels A and B and 50 nm in panel C.
intermediate between the calculated end-to-end distance of an isolated chain, , and the length of an extended polymer chain of either PAA or PS. The PAA chains have an average degree of polymerization of 173 and a calculated in DMF of 8.8 nm.32 The PS chains have a similar degree of polymerization of 180 and a calculated in DMF of 9.0 nm.32 Assuming a repeat length of 0.254 nm for both polymers, the extended chain lengths would be 43 nm for PAA and 45 nm for PS; however, the fully extended chain lengths would be extremely unfavorable entropically. The measured thickness of ∼27 nm in cryo-electron micrographs for the PAA/PS mixed brush layer implies that in DMF the densely grafted PAA and PS brushes on the silica nanoparticles adopt close to extended chain conformations. CryoEM Images of Mixed PAA/PS Brush-Grafted Nanoparticles in Water. A representative cryo-electron micrograph of the mixed PAA/PS brush-grafted silica nanoparticles in water is shown in Figure 1C,D. Again, uranyl acetate was added to selectively stain the PAA chains prior to plungefreezing the sample on a cryoEM grid. Water is a selective solvent, which is good for PAA but poor for PS. As a result, the morphology of microphase-separated PAA/PS mixed brushes in water should look different from that in DMF (a nonselective good solvent). According to theoretical prediction3 and a previous experimental study,17 mixed brushes in a selective solvent adopt vertical phase separation, in which the solvophilic chains stretch out to form the top layer, and the solvophobic chains collapse into surface tethered micelles in the bottom layer. Comparing the cryo-micrographs in Figure 1C,D with those in Figure 1A,B, the morphology of the mixed brush particles in water appears similar to that in DMF (Figure 1A,B). Isolated bright PS nanodomains are seen on top of the silica particles in water as in DMF. In addition, some weak and irregular striations of lighter and darker gray are observed in the mixed brush halo around the edges of the nanoparticles in water. Between the silica core and the mixed brush layer, there is again a bright layer for the nanoparticles in water, which can be attributed to the crossover zone for mixed PAA/PS brushes. The only obvious difference lies in the quantitative measurement, where the mixed brush layer in water is only 18 ± 3 nm thick, more compact than the mixed brush layer in DMF (i.e., 27 nm). The thinner brush layer in water indicates that the PS chains collapse in the mixed brush layer, thus reducing the overall brush layer thickness. Obviously, conventional 2D TEM and 2D cryoEM are not powerful enough to unambiguously distinguish the 3D morphology of surface micelles from the
microscopy requires the use of low electron doses, typically 5− 20 electrons per Å2 for samples frozen in aqueous buffers.19 At higher electron doses, the nanoparticles become radiation damaged, and the frozen (amorphous) water layer begins to melt or devitrify. Frozen DMF is even more electron-beamsensitive than frozen water. Therefore, the electron dose used to collect the cryo-electron micrographs of the nanoparticles in DMF was limited to a lower level. In Figure 1A, the edge of the hole in the carbon support film is evident in the upper right corner. A few fiducial 10 nm nanogold particles are visible as electron dense spheres. Uranyl acetate was added to selectively stain the PAA chains, while the nanoparticles were at room temperature, prior to plunge-freezing the sample on a cryoEM grid. Because DMF is a nonselective good solvent for PAA and PS,17 both polymers are expected to maximize their contact with this solvent. In the cryo-electron micrographs (Figure 1A,B), the mixed brush layer is observed to form a halo around the electron dense silica cores with subtle striations of alternating lighter and darker gray within the halo. These striations correspond to microphase-separated domains of PS (lighter gray) and PAA (stained and darker gray). Because of the fact that these are projection images, certain isolated bright PS domains appear on top of the silica particle area. Considering both the top and edge views of the grafted mixed brushes, PS and PAA mixed brushes form a lateral microphase separation pattern, namely, a ripple structure. In between the dark silica core and the mixed brush halo, there is a thin bright layer, about 6 nm in thickness, for every nanoparticle. This layer was not clearly seen in our previous reports when the samples were stained by RuO4, which is a much stronger oxidative staining agent.30 In this work, the staining agent is uranyl acetate, which stains the PAA chains via electrostatic interactions, rather than oxidation. We consider that in addition to the initiator layer there must exist a finite crossover (i.e., intermixed) zone for mixed PAA/PS brushes to microphase separate laterally between the microphase-separated brush layer and the silica core.31 Because of the existence of hydrophobic PS brushes in the crossover zone, uranyl acetate may not be able to penetrate into this layer. As a result, the crossover zone will appear bright in cryoEM images. We speculate that the thin bright layer we observe in cryoEM images represents a finite crossover zone for densely grafted, laterally microphase-separated mixed brushes. The measured total thickness of the mixed brush layer in the frozen DMF in Figure 1A,B is ∼27 ± 3 nm. This value is 8683
DOI: 10.1021/acs.langmuir.5b01739 Langmuir 2015, 31, 8680−8688
Article
Langmuir
When this mixed brush particle is displayed as a solid object, it appears roughly spherical (Figure 3). The pattern and spacing
ripple structure, as expected for mixed brushes in a selective solvent and a nonselective good solvent, respectively. In the following, we use cryo-ET to characterize these different morphologies. 3D Morphology of Mixed PAA/PS Brush-Grafted Nanoparticles in DMF. Collection of a tilt series of the mixed brush particles in DMF proved to be challenging because of the electron beam sensitivity of frozen DMF. Initial attempts at collecting tilt series were performed on the FEI TF20 cryoelectron microscope using the TVIPS EM-Tools 4.0 tomography software package; however, only tilt series with limited angular ranges (e.g., a 40° range) were collected before there was visible electron damage to the sample. Successful collection of complete tilt series (−64 to +64°) was performed on a JEOL JEM2200FS cryo-electron microscope with the Serial EM tomography software package. The in-column energy filter of the JEOL microscope improved the image contrast in the lowdose tilt frame images and allowed a lower electron dose to be used while maintaining image tracking. Also, use of the Serial EM software package enabled the collection of a complete tilt series with fewer tracking and focusing micrographs than needed for the TVIPS software, resulting in a lower overall electron dose to the sample. A 3D tomogram was generated from the tilt series of the mixed brush particles in DMF with the IMOD software package.33 The aligned tilt series images are shown in Movie S1, and all of the z-slices from the tomogram are shown in Movie S2 in the Supporting Information. A tilt image from the tilt series and a representative z-slice from the middle of the tomogram are shown in Figure 2A,B. This z-slice cuts through the silica cores of about 16 nanoparticles. The cores of most mixed brush particles are spherical, although a few are more oval shaped or formed by two or more merged spherical cores. The mixed brush layer extends uniformly around all of the cores with alternating lighter (PS) and darker (uranyl acetate stained PAA) bands. The striated pattern of the brush layer is clearer in the tomogram z-slice (Figure 2B) than in the tilt image (Figure 2A). The particles within the tomogram volume are at various heights within the frozen DMF layer. The height variation can be seen more clearly in Movie S2 in the Supporting Information showing all of the tomogram z-slices. Most of the particles in Figure 2B are sliced through their silica cores, although a few are sliced more peripherally through their brush layers. All of the particles that are sliced through their silica cores show a clear “gap” of ∼6 nm between the core and the striated brush layer. Again, this unstained bright layer can be attributed to the crossover zone next to the tethering substrate.31 This mixing layer is required for the two immiscible polymer brushes to separate laterally and segregate into the rippled structure.31 The single mixed brush particle circled in red in Figure 2A,B was extracted from the tomogram, and its 3D structure was evaluated separately. A 5 nm thick density slab from the middle of the reconstructed nanoparticle is shown in Figure 2C. Slabs of density from the extracted nanoparticle show the ripple pattern of laterally separated PAA domains in blue color. Although the ripple pattern is not uniform throughout the mixed brush layer, there is a commonly repeated spacing for the PAA density. Some empty holes are seen in the silica core, which were also observed in previous 3D reconstructions of silica nanoparticles.15 We attribute this to certain artifacts caused by density fluctuations in the silica cores.
Figure 3. One PAA/PS brush-grafted nanoparticle extracted from a tomogram of particles frozen in DMF. Spherically cropped views from the top and bottom of the particle are shown to reveal the morphology of the PAA domains within the polymer shell. The silica core is displayed in red and the PAA domains are displayed in blue. The scale bar represents 50 nm.
of the PAA domains within the brush shell are revealed in spherically cropped views of the extracted particle (Figure 3) and are fairly uniform over the surface of the nanoparticle. The spherical appearance of the nanoparticles is in contrast with the overall bell shape observed for similar nanoparticles imaged by electron tomography for mixed brush particles cast from a neutrally good solvent, chloroform,15 where chloroform evaporation is thought to pull the mixed brushes down toward the carbon substrate, resulting in less polymer density on top of the nanoparticles than on the bottom. In the case of cryoET, the mixed brush particles are not sitting on a solid substrate but rather are suspended in a frozen layer of DMF (or water) within holes in the carbon support film of the grid. In addition, preparation of the cryoEM grids does not involve sample drying. Instead, the grids are prepared by flash freezing a droplet of the nanoparticles in solvent on the cryoEM grid, and the sample grid is kept frozen at liquid nitrogen temperature during image collection in a TEM. Therefore, the representation of the mixed brush-grafted particles obtained by cryoET corresponds to the in situ state of the mixed brush particles in solution. The PAA density spacing at the half-height of the mixed brush layer was measured in multiple density slabs. A histogram of the measurements is generated in Figure 4, and the average spacing was found to be 13.8 ± 3.9 nm. Theoretical calculations of Marko and Witten indicate that the ripple wavelength should be 1.97 times the polymer chain root-mean-square end-to-end
Figure 4. Distribution of the ripple wavelength features in cryoET density slabs of the PAA/PS mixed brush-grafted silica nanoparticles in DMF. 8684
DOI: 10.1021/acs.langmuir.5b01739 Langmuir 2015, 31, 8680−8688
Article
Langmuir
Figure 5. (A) Zero-degree tilt image of mixed PAA/PS brush-grafted silica nanoparticles in water from a tilt series collected on the FEI TF20 microscope. The inset shows a particle from a different tilt angle image (+6.4°). (B) Middle z-slice from the tomogram. (C) Central 5 nm thick density slab of the red circled nanoparticle in panels A and B. (D) Zero-degree tilt image of PAA/PS nanoparticles in water from a tilt series collected on the JEOL 2200FS microscope. The inset is of a particle from a different tilt angle image (+43.8°). (E) Middle z-slice from the tomogram. (F) Central 5 nm thick density slab of the red circled nanoparticle in panels D and E. In panels C and F the silica cores are displayed in gold and PAA domains in green. The scale bars represent 50 nm.
distance ().34,35 For PAA chains, the prediction would be 17.3 nm (1.97 × 8.8 nm), and for PS chains the prediction would be a ripple wavelength of 17.7 nm (1.97 × 9.0 nm); however, this theoretical prediction did not take into account grafting density and substrate curvature effects. Computer simulations have shown a strong dependence of the ripple wavelength on the relative grafting densities of two polymers in the brush layer.9,10 TEM measurements have shown a dependence of the ripple wavelength on grafting density: the ripple wavelength is observed to increase with decreasing overall grafting density.12 The cryoET ripple wavelength measurement of 13.8 nm is slightly higher than the 12.5 nm of the phase-separated domains on PtBA/PS grafted silica nanoparticles with the same mixed brush lengths and grafting densities. The slightly higher value can be attributed to the swelling of the brush layer in the DMF solvent. Although we have not examined mixed PAA/PS brushgrafted nanoparticles with different molecular weights of PAA and PS by cryoET, we expect that the ripple wavelength would increase with increasing molecular weights for symmetric mixed brushes as predicted by Zhulina and Balazs.31 The effect of composition of mixed brushes would be more complicated because biasing the composition (by changing either molecular weight disparity or grafting density) will change the morphology from the ripple structure to cylindrical or spherical structures. For small particles (core size
