Ultrastable Iron Oxide Nanoparticle Colloidal Suspensions Using

Oct 16, 2009 - We have found catechol-derivative anchor groups which possess irreversible binding affinity to iron oxide and thus can optimally disper...
13 downloads 15 Views 2MB Size
Download PDF
NANO LETTERS

Ultrastable Iron Oxide Nanoparticle Colloidal Suspensions Using Dispersants with Catechol-Derived Anchor Groups

2009 Vol. 9, No. 12 4042-4048

Esther Amstad, Torben Gillich, Idalia Bilecka, Marcus Textor, and Erik Reimhult* Laboratory of Surface Science and Technology, ETH Zurich, Switzerland, Laboratory for Multifunctional Materials, ETH Zurich, Switzerland Received July 10, 2009; Revised Manuscript Received September 21, 2009

ABSTRACT We have found catechol-derivative anchor groups which possess irreversible binding affinity to iron oxide and thus can optimally disperse superparamagnetic nanoparticles under physiologic conditions. This not only leads to ultrastable iron oxide nanoparticles but also allows close control over the hydrodynamic diameter and interfacial chemistry. The latter is a crucial breakthrough to assemble functionalized magnetic nanoparticles, e.g., as targeted magnetic resonance contrast agents.

The number of potential applications of superparamagnetic iron oxide nanoparticles (SPIONs) especially, but not exclusively, in the biomedical field is rapidly increasing with emerging technologies to tune and control their bulk and, even more importantly, surface properties. Due to the high biocompatibility of iron oxide nanoparticles1 and their superparamagnetic behavior, they are often used as, e.g., magnetic resonance (MR) contrast agents to achieve cellular or even molecular resolution2 and for cell tracking and separation purposes.3 However, the quality and performance of such iron oxide nanoparticles crucially depend on nanoparticle size, stability, dispersant layer thickness, and control over their interfacial chemistry.4,5 Iron oxide nanoparticles are typically sterically stabilized with dispersants if used for biomedical applications.6 Dispersants, such as dextran, which lack a well-defined high affinity anchor group lead to broad particle size distributions with average hydrodynamic diameters many times larger than the iron oxide core diameter.6 Needless to say the resulting interfacial chemistry is poorly defined and thus prevents efficient incorporation and controlled presentation of ligands.6 Steric stability can also be provided by low molecular weight dispersants strongly bound to the iron oxide nanoparticle surface at high packing densities. Low molecular weight dispersants which consist of one high affinity anchor group covalently linked to a hydrophilic polymer can be grafted to iron oxide nanoparticle surfaces in a well-defined way * Corresponding author: Erik Reimhult, Wolfgang-Pauli-Strasse 10, CH8093 Zurich, Switzerland; tel, +41 (0)44 633 7547; fax, +41 (0)44 633 1027; e-mail, [email protected]. 10.1021/nl902212q CCC: $40.75 Published on Web 10/16/2009

 2009 American Chemical Society

without the need for further in situ chemistry.7-11 Dopamine, a derivative of the amino acid DOPA which is abundantly present in the mussel adhesive protein Mytilus edulis,12 is the most considered high affinity binding group to stabilize iron oxide nanoparticles in water10 and physiologic buffers.8,9,13 Nevertheless, the suitability of dopamine as a high affinity anchor for iron oxide nanoparticle stabilization is still debated mainly because dopamine degradation has been claimed upon adsorption on iron oxide nanoparticles.13,14 Furthermore, aliquotted dopamine has been shown to be prone to oxidation.15 Despite the versatility of this method to disperse SPIONs and the tough requirements regarding affinity and stability of the anchor group, no systematic study on the influence of the anchor groups on nanoparticle stability has been reported yet. In particular, particle stability at 37 °C or higher, important for in vivo applications such as imaging and stimulated drug release, has not been addressed for iron oxide nanoparticles individually stabilized with low molecular weight dispersants. To this end, we have investigated eight different catechol derivatives selected to shed light on the design principles for optimized anchors for dispersants to stabilize iron oxide nanoparticles. Poly(ethylene glycol) (PEG) with a molecular weight of 5 kDa (PEG(5)) was coupled to each of the different anchor groups to yield dispersants of suitable size for SPION stabilization.7 We have found that despite the numerous applications of dopamine and DOPA as high affinity anchor group for iron oxide nanoparticle stabilization there are catechol-based anchors which, if coupled to PEG, result in greatly enhanced

Figure 1. Stabilization of iron oxide nanoparticles. (a) Chemical structures of investigated dispersants with a degree of polymerization of n ≈ 114 for PEG(5). (b) High-resolution transmission electron microscopy of iron oxide nanoparticles air-dried on a carbon supported Cu grid. The hexagonal pattern of the inverse spinel structure of Fe3O4 oriented in the (010) direction is clearly visible, proving the high crystallinity of these nanoparticles. Dispersants with different anchor groups have been grafted to as-synthesized iron oxide nanoparticles. (c) Stabilization of iron oxide cores with PEG(5)-dopamine which adsorbed reversibly (top) and the irreversibly adsorbing PEG(5)-nitroDOPA (bottom). (d) While PEG(5)-dopamine stabilized nanoparticles agglomerated and thus visibly precipitated (top), PEG(5)-nitroDOPA coated nanoparticles were stable even after being kept in HEPES containing 150 mM NaCl for 20 h at 90 °C (bottom). The initial nanoparticle concentration was 1 mg of iron oxide/mL; thus, the color difference between PEG(5)-dopamine and PEG(5)-nitroDOPA stabilized nanoparticle dispersions is a result of the precipitation of PEG(5)-dopamine stabilized nanoparticles.

iron oxide nanoparticle stability. Furthermore, it was shown that conventional dopamine-based dispersants fail to stabilize iron oxide nanoparticles at elevated temperature and desorb upon dilution of the iron oxide nanoparticle solution under physiologic conditions in strong contrast to some novel dispersants investigated here. This failure of dopamine-based dispersants to stabilize iron oxide nanoparticles under relevant conditions severely limits their applicability in the biomedical field where stability at body temperature in diluted suspensions and physiological ion concentrations is key. Thus, there are obvious advantages of replacing dopamine by the newly designed anchor groups demonstrated here, which yield optimal nanoparticle dispersion stability. Nano Lett., Vol. 9, No. 12, 2009

Iron Oxide Nanoparticle Stability. The nine different PEG-based dispersants investigated are shown in Figure 1a. Among these anchor groups, in addition to dopamine, also L-DOPA and hydroxydopamine have been reported to lead to good iron oxide nanoparticle stability.7,11 Furthermore, PEG(5)-COOH was investigated because dopamine and DOPA, and other investigated catechol derivatives chemically differ only in one carboxy group, motivating us to investigate the role of carboxylates in iron oxide binding of these anchors. Additionally, the carboxy group is thought to be the anchor of oleic acid that has frequently been used as a precursor during iron oxide nanoparticle synthesis16-18 and to sterically stabilize iron oxide nanoparticles in organic 4043

Figure 2. DLS measurements of individually stabilized iron oxide nanoparticles. Volume weighted diameters of iron oxide nanoparticles stabilized with PEG(5)-nitroDOPA (9), PEG(5)-nitrodopamine (0), PEG(5)-DOPA (2), PEG(5)-dopamine (4), PEG(5)mimosine (b), PEG(5)-hydroxypyridine (O) and PEG(5)hydroxydopamine (3) dispersed in HEPES containing 150 mM NaCl at room temperature: (a) as stabilized, (b) as a function of the number of times particles have been filtered with the aim of probing the binding reversibility of the dispersants, and (c) as a function of temperature. (d) The particle stability at elevated temperature depended on the amount of dispersant added as shown for PEG(5)-nitroDOPA stabilized with 1 mg of dispersant/mg of particle (], open diamond), 2 mg of dispersant/mg of particle (open diamond with crossed lines), 4 mg of dispersant/mg of particle (open box with crossed lines), 6 mg of dispersant/mg of particle ([), and 8 mg of dispersant/mg of particle (9). Standard deviations were calculated based on results measured on three to six identical, independent samples.

solvents.13,19 Single domain magnetite (Fe3O4) cores prepared by the microwave-assisted nonaqueous sol-gel route20 had a diameter of 6 ( 1 nm and were stabilized with the different dispersants through a simple grafting to method (Figure 1 and Figure S1 in the Supporting Information). As-synthesized iron oxide nanoparticle cores were added to a N,N-dimethylformamide (DMF) based dispersant solution. Dispersants were adsorbed at 50 °C for 24 h before excessive dispersants were removed and DMF was exchanged for Millipore water through 24 h dialysis against Millipore water. Stabilized iron oxide nanoparticles were freeze-dried and redispersed in the desired media at appropriate concentrations. PEG(5)-hydroxypyrone and PEG(5)-COOH stabilized iron oxide nanoparticles did not pass filters with a cutoff of 200 nm, which demonstrated that they instantaneously formed agglomerates larger than 200 nm in diameter. Consequently, no reliable dynamic light scattering (DLS) could be performed on them. The average volume-weighted hydrodynamic diameter of iron oxide nanoparticles individually stabilized with PEG(5)nitroDOPA, PEG(5)-nitrodopamine, PEG(5)-mimosine, PEG(5)-DOPA, PEG(5)-dopamine, PEG(5)-hydroxydopamine, and PEG(5)-hydroxypyridine respectively dispersed in 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) containing 150 mM NaCl (pH ) 7.4) at room temperature was 23 ( 2 nm (Figure 2a). The stabilized iron oxide 4044

Figure 3. Overview TEM micrographs of iron oxide nanoparticles. Iron oxide nanoparticles stabilized with (a) PEG(5)-nitroDOPA, (b) PEG(5)-nitrodopamine, (c) PEG(5)-mimosine, and (d) PEG(5)DOPA air-dried on a 10 nm thick carbon film supported Cu grid.

nanoparticles could be freeze-dried, stored as a powder for at least 3 months, or alternatively redispersed in Millipore water and stored for at least 4 months without noticeable change in particle stability and hydrodynamic diameter after redispersing and diluting them in HEPES containing 150 mM NaCl. A clear separation of the cores and the absence of aggregates of iron oxide nanoparticles stabilized with PEG(5)-nitroDOPA, PEG(5)-nitrodopamine, PEG(5)-mimosine, and PEG(5)-DOPA supports the DLS result of individually stabilized iron oxide nanoparticles (Figure 3). The binding reversibility of these dispersants was tested by successive filtrations and redispersions of the resulting iron oxide nanoparticle filter cakes in HEPES containing 150 mM NaCl.21 While dispersants which lead to low (dopamine, hydroxydopamine, hydroxypyridine, hydroxypyrone, and carboxyl) and intermediate (mimosine and DOPA) colloidal stability agglomerated within zero to three and five to seven filtrations, respectively, nanoparticles stabilized with nitroDOPA and nitrodopamine dispersants remained stable for more than nine filtrations (Figure 2b and Figure S2 in the Supporting Information). This indicated essentially irreversible binding of nitroDOPA and nitrodopamine to iron oxide nanoparticles under physiologic conditions, in stark contrast to other investigated anchor groups. The desorption rate of intermediate affinity dispersants, however, was significantly lower compared to that of low affinity dispersants, which indicated that although the binding was reversible the affinity of these intermediate dispersants toward iron oxide is much higher and their off rate lower. To further investigate the relative differences in the affinity of the different dispersants, individually stabilized iron oxide nanoparticles dispersed in HEPES containing 150 mM NaCl were subjected to repeated heating to set temperatures up to Nano Lett., Vol. 9, No. 12, 2009

Table 1. Dispersant Packing Density on Nanoparticles and Flat Surfacesa dispersant

dispersant/nm2 dispersant/nm2 (nanoparticle) (nanoparticle) dispersant/nm2 (flat) EG/nm2 EG/nm2 colloidal stability measured with TGA measured with XPS measured with XPS (nanoparticle) (flat)

PEG(5)-nitroDOPA high 2.8 ( 0.3 2.5 ( 0.2 0.31 ( 0.01 296 ( 27 36 ( 2 PEG(5)-nitrodopamine high 2.6 ( 0.2 2.5 ( 0.4 0.26 ( 0.01 290 ( 40 30 ( 1 PEG(5)-mimosine intermediate 2.5 ( 0.9 2.2 ( 0.3 0.32 ( 0.01 247 ( 37 36 ( 1 PEG(5)-DOPA intermediate 2.4 ( 0.9 1.8 ( 0.4 0.26 ( 0.00 210 ( 50 30 ( 0 PEG(5)-dopamine low 2.3 ( 0.5 2.2 ( 0.2 0.22 ( 0.01 254 ( 25 25 ( 1 PEG(5)-hydroxydopamine low 1.7 ( 0.6 2.2 ( 0.2 0.15 ( 0.00 254 ( 26 17 ( 0 PEG(5)-hydroxypyridine low 1.8 ( 0.6 1.7 ( 0.1 0.22 ( 0.01 197 ( 16 25 ( 1 PEG(5)-hydroxypyrone low 1.8 ( 0.2 2.1 ( 0.1 0.11 ( 0.00 244 ( 13 13 ( 0 PEG(5)-COOH low 1.3 ( 1.1 1.4 ( 0.3 0.10 ( 0.00 162 ( 37 11 ( 0 a The ethylene glycol (EG) density was calculated based on the dispersant packing density measured by XPS for nanoparticles and flat surfaces, respectively.

90 °C. Increasing the temperature probes both the density of the PEG shell as it collapses at high temperature causing the particles to reversibly (high density of PEG) or irreversibly (low density of PEG) aggregate as well as the strength of the anchor binding, since a temperature increase will speed up the exchange of reversibly binding dispersants. The result was that while iron oxide nanoparticles stabilized with PEG(5)-nitroDOPA and PEG(5)-nitrodopamine could be repeatedly heated up to 90 °C without noticeable agglomeration or changes in the derived count rates as measured by DLS, iron oxide nanoparticles stabilized with low affinity dispersants started to irreversibly agglomerate already below body temperature. Nanoparticles stabilized with dispersants which lead to intermediate colloidal stability agglomerated between 60 and 80 °C, as indicated by an increase in the hydrodynamic diameter (Figure 2c) and an increase followed by a decrease in the derived count rates due to particle agglomeration and sedimentation, respectively (see Figure S3 in Supporting Information). However, even for irreversibly binding dispersants, a concentration of at least 6 mg of dispersants/mg of nanoparticle was needed to completely prevent nanoparticle aggregation up to 90 °C (Figure 2d). Dispersant Packing Density. The dispersant packing density on iron oxide nanoparticles was determined by thermogravimetric analysis (TGA). Related to the packing density, the atomic CPEG:Fe ratio was investigated by X-ray photoelectron spectroscopy (XPS) (Table 1 and Figure 4). The obtained percentage weight losses measured between 200 and 400 °C by TGA (see Table S1 in Supporting Information) were converted into dispersant packing densities assuming spherical iron oxide nanoparticles with the core diameter distribution shown in Figure S1 in Supporting Information and an Fe3O4 density of 5.18 g/cm3 22 (Table 1). A clear trend toward higher particle stability with higher dispersant packing density was observed (cf. Table 1 and Figure 2). Furthermore, if particles were heated to 90 °C, before excessive dispersants were removed by filtration, the packing densities of PEG(5)-nitroDOPA and PEG(5)-mimosine stabilized nanoparticles remained in the range of packing densities of as-stabilized nanoparticles whereas that of PEG(5)-dopamine stabilized nanoparticles decreased to 26% of the initial packing density. This correlated well with the filtration experiments shown in Figure 2b where the binding reversibility of dopamine on iron oxide surfaces was markedly higher than those for mimosine and nitroDOPA. Nano Lett., Vol. 9, No. 12, 2009

Figure 4. TGA and DSC investigations of the optimal dispersant: iron oxide nanoparticle ratio. (a) DSC graphs of PEG(5)-nitroDOPA stabilized iron oxide nanoparticles where heat changes normalized to the mass of stabilized iron oxide nanoparticles are shown as function of temperature. Particles were stabilized with 1 mg of dispersant/mg of particle (]), 2 mg of dispersant/mg of particle (diamond with crossed lines), 4 mg of dispersant/mg of particle (box with crossed lines), 6 mg of dispersant/mg of particle ([), 8 mg of dispersant/mg of particle (9), respectively, and uncoated iron oxide nanoparticles (f). (b) The heat dissipated during temperatureinduced decomposition of the organic part of iron oxide nanoparticles stabilized with different amounts of dispersants was measured by DSC. The dissipated heat was normalized to the iron oxide core weight.

The exothermic reaction around 255 °C seen in differential scanning calorimetry (DSC) measurements for uncoated nanoparticles coincided with a mass loss of 20 wt % in TGA and corresponds to decomposition of weakly physisorbed impurities on the iron oxide nanoparticle surface (Figure 4a). A similar exothermic peak was measured for iron oxide nanoparticles stabilized with only 1 mg of dispersant/mg of particle. If particles were coated with 2 mg of dispersant/ mg of particle, two exothermic reactions around 255 and 380 °C were observed. The first reaction was assigned to the decomposition of remaining impurities. The second exothermic reaction, which coincided with the main mass loss consisting of equal amounts of CO2 and H2O as was measured in TGA coupled to a mass spectrometer (MS) (see Figures S4 and S5 in Supporting Information), was assigned to the decomposition of dispersants. For particles stabilized with at least 4 mg of dispersant/mg of particle, hardly any exothermic peak could be seen at 250 °C whereas the exothermic decomposition of PEG around 410 °C was evident. These results indicate that nitroDOPA could replace physisorbed impurities if particles were stabilized with a sufficient amount of dispersants. Dispersants adsorbed on flat Fe3O4 surfaces were analyzed by XPS. The atomic ratio of CPEG to Caliphatic, derived from 4045

Figure 5. XPS studies of dispersants adsorbed on flat 2D and nanoparticle surfaces. (a) The atom percentage of C of PEG (C-O, white), aliphatic C (C-C, C-H, dashed), and OH-conjugated carbons (C-OH, black) of dispersants adsorbed on flat Fe3O4 surfaces was determined from XPS analysis. Statistics were taken for three to six independent, identical samples. (b) The atomic ratio of CPEG:Fe of dispersants adsorbed on flat Fe3O4 surfaces correlates well with that of dispersants adsorbed on iron oxide nanoparticles indicating that dispersant binding affinity is independent of surface curvature whereas the dispersant packing density is considerably higher on nanoparticles as compared to flat surfaces.

the deconvoluted XPS C 1s spectra, in combination with the atomic ratio of CPEG:Fe can be taken as a measure of the binding affinity of the dispersants through the ability to replace aliphatic carbon contaminations. The atomic ratios of CPEG:Caliphatic and CPEG:Fe increased with increasing affinity of the dispersants (Figure 5 and Figure S6 in Supporting Information) and support DSC results where nitroDOPA was shown to replace weakly physisorbed carbon contaminations while low affinity anchor groups failed to do so. Moreover, dispersants yielding high colloidal stability resulted in higher dispersant packing densities compared to dispersants yielding low colloidal stability (Table 1). Furthermore, a good linear correlation of the XPS atomic CPEG:Fe ratios was obtained between catechol-based dispersants adsorbed on flat surfaces and on nanoparticles, respectively (Figure 5b). This indicates that the binding affinity of anchoring groups is independent of surface curvature, in contrast to the packing density which is considerably higher on nanoparticles than on flat surfaces (Table 1). The DSC, DLS, TGA, and XPS results all suggest that nitroDOPA and nitrodopamine were superior in terms of providing a high dispersant density on the nanoparticles, which directly correlated with irreversible binding and perfect particle stability under physiologic conditions. The hydrodynamic diameter of iron oxide nanoparticles stabilized with PEG(5)-nitroDOPA, PEG(5)-nitrodopamine, PEG(5)-mimosine, PEG(5)-DOPA, and PEG(5)-hydroxypyridine dispersed in HEPES at 40 °C was considerably lower (23 nm), and the particle size distribution was narrower than that for previously reported PEG(5)-silane stabilized iron oxide nanoparticles at room temperature (82 nm 23 and 44 nm,24 respectively) and PEG(3)-dopamine (nm) and PEG(6)dopamine stabilized iron oxide nanoparticles dispersed in phosphate buffered saline (PBS) at 37 °C (≈90 nm)9 despite similar iron oxide core diameters. An average dispersant shell thickness of 8.5 nm was calculated given an iron oxide core diameter of 6 nm and the hydrodynamic diameter of 23 nm. Superior particle stability and fewer agglomerates of PEG(5)nitroDOPA and PEG(5)-nitrodopamine compared to PEG(5)4046

dopamine stabilized nanoparticles were also directly shown by our experiments. The lower stability of iron oxide nanoparticles coated with PEG(5)-dopamine compared to, e.g., PEG(5)-nitrocatechol stabilized nanoparticles might partially be because dopamine is prone to oxidation especially in the presence of iron if kept aliquotted at room temperature,14,25,26 in contrast to, e.g., nitrocatechols.27 For dispersants leading to low or intermediate colloidal stability a solution temperature slightly below the corresponding PEG cloud point28 induced irreversible nanoparticle agglomeration. However, PEG(5)-nitroDOPA and PEG(5)nitrodopamine stabilized iron oxide nanoparticles aggregated reversibly by a temperature increase above the PEG cloud point29 and remained dispersed after repeated heating to 90 °C and cooling. A plausible explanation for temperatureinduced irreversible agglomeration of nanoparticles stabilized with poor or intermediate dispersants is that these dispersants are reversibly bound and show faster dissociation at elevated temperatures, while for high affinity anchors the bond formed with the substrate is so strong that for practical purposes it is irreversible. The high binding affinity of some of the dispersants correlated strongly with the unsurpassed dispersant packing density shown here. The maximum PEG(5)-catechol packing density previously reported was 0.3 PEG(5)/nm2 obtained for triple anchor PEG(5)-(DOPA)3 adsorbed on flat TiO2 surfaces30 and is comparable to that measured here for PEG(5)-nitroDOPA and PEG(5)-mimosine adsorbed on flat iron oxide surfaces (Table 1). A similar packing density of 0.4 molecules/nm2 has been reported for PEG-succinimidyl carbonate adsorbed on flat silane activated quartz, independent of the PEG molecular weight.31 Furthermore, the packing density of PEG-silanes adsorbed on Fe3O4 nanoparticles has been reported to be between 0.3 and 0.8 molecules/nm2 for PEG molecular weights between 0.1 and 0.35 kDa.32 Moreover, the maximum PEG(5) packing density measured on nanoparticles was 1.3 dispersants/nm2 for PEG(5)-hydroxydopamine7 adsorbed on iron oxide nanoparticles which is two times lower than the packing density measured here. The extremely high dispersant packing density obtained here appears, at first glance, to be too high for a grafting to immobilization process that requires diffusion of large polymers through a brush of already adsorbed polymers to reach completion. However, account must be taken of the high curvature of the nanoparticles. This is in agreement with the thiol packing density which has been reported to be up to 2.2 times higher for poly(ethylene oxide)thiols (PEO-SH) and 23 times for polystyrenethiols (PS-SH) if adsorbed on Au nanoparticles compared to 2D Au surfaces.33 The conical volume that can be occupied by each immobilized PEG chain will significantly reduce steric interchain repulsion, leading to a different chain density profile and thus different dispersant packing densities on highly curved surfaces of nanoparticles. Similar considerations were recently shown for the dependence of the DNA-SH packing density on the gold nanoparticle diameter.34 However, good correlation of the XPS CPEG:Fe atomic ratios between dispersants adsorbed Nano Lett., Vol. 9, No. 12, 2009

on flat 2D surfaces and on 3D nanoparticles (Figure 5b) still allows for screening of dispersants on 2D surfaces where more surface sensitive and quantitative characterization techniques are at hand. The obtained results can then be translated into the performance of these dispersants to stabilize nanoparticles of the same composition. The striking difference in binding affinities of these different anchor groups might be related to different pKa values. While nitroDOPA, nitrodopamine, and mimosine, the three anchor groups which lead to highest particle stability, all have one of their pKa values around 6.5, close to the point of zero charge (PZC) of iron oxide nanoparticles (6.7),7 DOPA, dopamine, and hydroxydopamine have pKa > 9 (Figure S8 in Supporting Information). Nitrocatechols have been reported to form more stable complexes if adsorbed on TiO2 surfaces than catechols.35 Furthermore, the binding affinity of catechols and catechol derivatives toward TiO2 has been shown to scale with their Brønstead acidity,35 which is influenced by the carboxy group and might be one of the reasons for the higher affinity of DOPA compared to dopamine toward iron oxide. Furthermore, the aromatic nitrogen of mimosine and hydroxypyridine enhances the Brønstead acidity of these anchors compared to DOPA and dopamine, respectively, which might result in higher packing densities of dispersants containing pyridine-based anchor groups compared to their catechol-based counterparts. Further characterization of how catechol derivatives bind to iron oxide is currently being performed in order to better understand the remarkable difference in binding affinities observed not only when varying the number of hydroxyl groups on the ring, but also upon coupling electronegative substituents directly to the catechol ring. It can also be noted that a mere increase in the number of hydroxyl groups does not significantly increase the affinity of the dispersant, suggesting that binding occurs with the ring oriented orthogonal to the substrate, sterically allowing only two hydroxyl groups to form simultaneous bonds with the iron oxide surface. Irreversible binding of low molecular weight dispersants which withstand physiologic conditions allow for controlled and efficient surface functionalization of individual oxide nanoparticles, e.g., with antibodies for targeting purposes or additional labels such as fluorophores without the risk of ligand dissociation caused by weak dispersant-nanoparticle interactions. Thus, so-functionalized nanoparticles are promising candidates for in vivo applications. Furthermore, the hydrodynamic diameter of individually stabilized nanoparticles solely depends on the iron oxide core diameter and the dispersant layer thickness and can therefore easily be controlled by adjusting the dispersant molecular weight and iron oxide core diameter. This is of great relevance for passive targeting of, e.g., tumor tissue, by the enhanced permeation and retention (EPR) effect as well as, in concert with a more stable presentation of bioactive ligands, for the active and specific targeting of cellular and tissue components. Furthermore, these particles can be freeze-dried and redispersed at the desired concentration and physiologic Nano Lett., Vol. 9, No. 12, 2009

conditions, which facilitates particle handling considerably. Finally, because catechols are known to have high affinity also toward a broad range of oxides,36 such dispersants are likely to be applicable to individually stabilize oxide nanoparticles in general through the easy and cost-effective grafting-to method. Thus, these catechol-based dispersants might pave the way for a variety of new applications of oxide nanoparticles where irreversible dispersant binding, good nanoparticle stability under harsh conditions, long shelf life, tailoring of the hydrodynamic diameter and interfacial chemistry, and easy particle handling are stringent requirements. In summary, individually stabilized sub-30-nm iron oxide nanoparticles have been developed where the dispersant packing density was twice as high as the currently highest reported PEG packing density. The particle stability was shown to be strongly correlated with the high dispersant packing density and the corresponding reported iron oxide nanoparticle stability vastly exceeded that of previously reported similar particles, especially at medically relevant conditions of elevated temperature and ionic strength. Irreversible binding of high affinity anchor groups such as nitroDOPA and nitrodopamine allowed diluting and heating iron oxide nanoparticle suspensions without nanoparticle agglomeration. These high affinity anchor groups will therefore allow for quantitative tailoring of molar ratios of differently functionalized dispersants at the nanoparticle surface, resulting in targeted and multifunctional SPIONs for in vivo use. This work is another successful demonstration of how materials and concepts inspired by Nature can be translated into development and engineering of advanced, highly multifunctional materials and devices. Acknowledgment. The authors thank Dr. Stefan Zu¨rcher for valuable discussions, suggestions, and careful proof reading. We also thank Professor Markus Niederberger for providing the infrastructure for the iron oxide nanoparticle synthesis and Dr. Lucio Isa for help with TEM image processing and evaluation. Furthermore, we are indebted to NETZSCH-Geraetebau GmbH which kindly performed some TGA-MS experiments for us. Professor L. Gauckler (ETH Zurich), EMEZ (ETH Zurich), and FIRST (ETH Zurich) are acknowledged for providing TGA, electron microscopy, and clean room facilities, respectively. COST Action No. D43 “Multifunctional Superparamagnetic Iron Oxide Nanoparticles for Targeted Magnetic Resonance Imaging” and ETH Zurich are acknowledged for their financial support. Supporting Information Available: Supporting figures describe the iron oxide core size distribution, the normalized count rates of DLS results shown in the main text of this article, temperature-dependent MS results of during TGA decomposed dispersant fragments, TGA results of PEG(5)nitroDOPA stabilized iron oxide nanoparticles, detailed C 1s XPS spectra of dispersants adsorbed on flat iron oxide surfaces, the influence of the oxygen partial pressure during reactive magnetron sputtering of thin iron oxide films on the iron oxide stochiometry, and pKa measurements of different anchors, MS, microelement analysis, 1H and 13C NMR 4047

spectra of dispersants, and experimental details. This material is available free of charge via the Internet at http:// pubs.acs.org. References (1) Weissleder, R.; Stark, D. D.; Engelstad, B. L.; Bacon, B. R.; Compton, C. C.; White, D. L.; Jacobs, P.; Lewis, J. Am. J. Roentgenol. 1989, 152 (1), 167–173. (2) Lee, J. H.; Huh, Y. M.; Jun, Y.; Seo, J.; Jang, J.; Song, H. T.; Kim, S.; Cho, E. J.; Yoon, H. G.; Suh, J. S.; Cheon, J. Nat. Med. 2007, 13 (1), 95–99. (3) Pittet, M. J.; Swirski, F. K.; Reynolds, F.; Josephson, L.; Weissleder, R. Nat. Protocols 2006, 1, 73–79. (4) Jun, Y. W.; Lee, J. H.; Cheon, J. Angew. Chem., Int. Ed. 2008, 47 (28), 5122–5135. (5) Lu, A. H.; Salabas, E. L.; Schuth, F. Angew. Chem., Int. Ed. 2007, 46 (8), 1222–1244. (6) Sun, C.; Lee, J. S. H.; Zhang, M. Q. AdV. Drug DeliVery ReV. 2008, 60 (11), 1252–1265. (7) Amstad, E.; Zurcher, S.; Mashaghi, A.; Wong, J. Y.; Textor, M.; Reimhult, E. Small 2009, 5 (11), 1334–1342. (8) Xu, C. J.; Xu, K. M.; Gu, H. W.; Zheng, R. K.; Liu, H.; Zhang, X. X.; Guo, Z. H.; Xu, B. J. Am. Chem. Soc. 2004, 126 (32), 9938–9939. (9) Xie, J.; Xu, C.; Kohler, N.; Hou, Y.; Sun, S. AdV. Mater. 2007, 19 (20), 3163-+. (10) Gu, H. W.; Yang, Z. M.; Gao, J. H.; Chang, C. K.; Xu, B. J. Am. Chem. Soc. 2005, 127 (1), 34–35. (11) Gao, J. H.; Liang, G. L.; Cheung, J. S.; Pan, Y.; Kuang, Y.; Zhao, F.; Zhang, B.; Zhang, X. X.; Wu, E. X.; Xu, B. J. Am. Chem. Soc. 2008, 130 (35), 11828–11833. (12) Waite, J. H.; Tanzer, M. L. Science 1981, 212 (4498), 1038–1040. (13) Xie, J.; Xu, C. J.; Xu, Z. C.; Hou, Y. L.; Young, K. L.; Wang, S. X.; Pourmond, N.; Sun, S. H. Chem. Mater. 2006, 18 (23), 5401–5403. (14) Shultz, M. D.; Reveles, J. U.; Khanna, S. N.; Carpenter, E. E. J. Am. Chem. Soc. 2007, 129 (9), 2482–2487. (15) Zurcher, S.; Wackerlin, D.; Bethuel, Y.; Malisova, B.; Textor, M.; Tosatti, S.; Gademann, K. J. Am. Chem. Soc. 2006, 128 (4), 1064– 1065. (16) Sun, S. H.; Zeng, H.; Robinson, D. B.; Raoux, S.; Rice, P. M.; Wang, S. X.; Li, G. X. J. Am. Chem. Soc. 2004, 126 (1), 273–279. (17) Willis, A. L.; Turro, N. J.; O’Brien, S. Chem. Mater. 2005, 17 (24), 5970–5975. (18) Bronstein, L. M.; Huang, X. L.; Retrum, J.; Schmucker, A.; Pink, M.; Stein, B. D.; Dragnea, B. Chem. Mater. 2007, 19 (15), 3624– 3632.

4048

(19) Jain, T. K.; Moralles, M. A.; Sahoo, S. K.; Lesllie-Pellecky, D. L.; Labhasetwar, V. Mol. Pharm. 2005, 2 (3), 194–205. (20) Bilecka, I.; Djerdj, I.; Niederberger, M. Chem. Commun. 2008, (7), 886–888. (21) While free dispersants were efficiently removed with centrifuge filters, leading to nanoparticle agglomeration, nanoparticles appeared to be stable for more than 30 days if dialyzed against Millipore water. This apparent discrepancy is due to hindered diffusion of free dispersants during dialysis despite dialysis tubing which had a cutoff of 12-14 kDa, more than twice the molecular weight of dispersants investigated here. Therefore, the former method is more suitable if binding reversibility is investigated. (22) Linde, D. R.; Frederikse, H. P. R. Handbook of Chemistry and Physics, 75th ed.; CRC: Boca Ratan, FL, 1995. (23) Herve, K.; Douziech-Eyrolles, L.; Munnier, E.; Cohen-Jonathan, S.; Souce, M.; Marchais, H.; Limelette, P.; Warmont, F.; Saboungi, M. L.; Dubois, P.; Chourpa, I. Nanotechnology 2008, 465608. 7 pp. . (24) Zhang, Y.; Sun, C.; Kohler, N.; Zhang, M. Q. Biomed. MicrodeVices 2004, 6 (1), 33–40. (25) Emerson, J. P.; Kovaleva, E. G.; Farquhar, E. R.; Lipscomb, J. D.; Que, L. Proc. Natl. Acad. Sci. U.S.A. 2008, 105 (21), 7347–7352. (26) Jang, H. G.; Cox, D. D.; Que, L. J. Am. Chem. Soc. 1991, 113 (24), 9200–9204. (27) Palumbo, A.; Napolitano, A.; Barone, P.; d’Ischia, M. Chem. Res. Toxicol. 1999, 12, 1213–1222. (28) Florin, E.; Kjellander, R.; Eriksson, J. C. J. Chem. Soc., Faraday Trans. 1 1984, 80, 2889–2910. (29) Pang, P.; Englezos, P. Nord. Pulp Pap. Res. J. 2000, 15 (5), 387– 394. (30) Dalsin, J. L.; Lin, L. J.; Tosatti, S.; Voros, J.; Textor, M.; Messersmith, P. B. Langmuir 2005, 21 (2), 640–646. (31) Burns, N. L.; Vanalstine, J. M.; Harris, J. M. Langmuir 1995, 11 (7), 2768–2776. (32) Butterworth, M. D.; Illum, L.; Davis, S. S. Colloids Surf., A 2001, 179 (1), 93–102. (33) Corbierre, M. K.; Cameron, N. S.; Lennox, R. B. Langmuir 2004, 20 (7), 2867–2873. (34) Kira, A.; Kim, H.; Yasuda, K. Langmuir 2009, 25, 1285. (35) Araujo, P. Z.; Morando, P. J.; Blesa, M. A. Langmuir 2005, 21 (8), 3470–3474. (36) Lee, H.; Dellatore, S. M.; Miller, W. M.; Messersmith, P. B. Science 2007, 318 (5849), 426–430.

NL902212Q

Nano Lett., Vol. 9, No. 12, 2009