Influence of Electronegative Substituents on the Binding Affinity of

Dec 15, 2010 - 2009 , 9 , 4042−4048) that compared different catechol .... ACS Applied Materials & Interfaces 2015 7 (44), 24656-24662 ... Iron Oxid...
1 downloads 0 Views 1MB Size
Download PDF
J. Phys. Chem. C 2011, 115, 683–691

683

Influence of Electronegative Substituents on the Binding Affinity of Catechol-Derived Anchors to Fe3O4 Nanoparticles Esther Amstad,† Andreas U. Gehring,‡ Håkon Fischer,‡ Venkatamaran V. Nagaiyanallur,† Georg Ha¨hner,§ Marcus Textor,† and Erik Reimhult*,†,| Laboratory of Surface Science and Technology and Geophysics Institute, ETH Zurich, Switzerland, EaStChem, School of Chemistry, UniVersity of St. Andrews, U.K., and Department of Nanobiotechnology, UniVersity of Natural Resources and Life Sciences (BOKU), Vienna, Austria ReceiVed: NoVember 16, 2010

Successful applications of nanoparticles are often limited by insufficient nanoparticle stability due to low binding affinity of dispersants. However, excellent Fe3O4 nanoparticle stability was reported in a recent study (Nano Lett. 2009, 9, 4042-4048) that compared different catechol derivative-anchored low molecular weight dispersants. Here, we investigate mechanistic binding aspects of five different anchors from this study that showed radically different efficiencies as dispersant anchors, namely nitroDOPA, nitrodopamine, DOPA, dopamine, and mimosine, using electron paramagnetic resonance, Fourier transform infrared spectroscopy, and UV-vis spectroscopy. We demonstrate enhanced electron delocalization for nitrocatechols binding to Fe2+ compared to unsubstituted catechols if they are adsorbed on Fe3O4 surfaces. However a too high affinity of mimosine to Fe3+ was shown to lead to gradual dissolution of Fe3O4 nanoparticles through complexation followed by dissociation of the complex. Thus, the binding affinity of anchors should be optimized rather than maximized to achieve nanoparticle stability. Introduction Controlled stabilization and functionalization of oxide nanoparticles mainly, but not exclusively, for biomedical applications have received increasing attention in recent years.1-3 Prominent examples are iron oxide nanoparticles for hyperthermia4 and as magnetic resonance (MR) contrast agents.5-8 Many of these applications require dispersants to be irreversibly bound with a controlled orientation and conformation to the nanoparticles, which is especially important when targeting of specific cells, tissue, or organ under physiological salt concentrations and elevated temperatures is desired.1,8 Only then can good nanoparticle stability, sufficient circulation in vivo and a controlled presentation of (bio)functional targeting groups on the particle surface be achieved. Almost all dispersants used today are either electrostatically adsorbed on nanoparticles (e.g., dextran)9,10 or attached through reversibly binding anchors such as oleic acid or dopamine.11 Steric stabilization of oxide nanoparticles with low molecular weight dispersants that are adsorbed on the nanoparticle surface through a straightforward and cost-effective “grafting to” approach is attractive. It allows coadsorption of differently functionalized dispersants resulting in multifunctional nanoparticles. Furthermore, the thickness and end-functionality of the thin stabilizing coating can be closely controlled without the need for sophisticated in situ chemical reactions. The “graftingto” approach relies on high affinity anchors that can create the bond to the nanoparticle surface under suitable conditions. Naturally, anchors that meet the stringent requirements for * To whom correspondence should be addressed. Address: WolfgangPauli-Strasse 10, CH-8093 Zurich, Switzerland. Telephone: +41 (0)44 633 7547. Fax: +41 (0)44 633 1027. E-mail: [email protected]. † Laboratory of Surface Science and Technology, ETH Zurich. ‡ Geophysics Institute, ETH Zurich. § University of St. Andrews. | University of Natural Resources and Life Sciences (BOKU).

coupling dispersants to nanoparticles by “grafting to” can perform the same function to graft nonfouling polymer brushes to planar substrates and also be used as anchors for initiators to graft polymers from a substrate surface. Catechol derivatives are often proposed as suitable dispersant anchors to various oxides including TiO2 and iron oxides. Despite the close chemical similarity of different catechol derivatives, their affinities to TiO212 and Fe3O411 have been shown to differ considerably. Because binding affinities of these anchors to Fe3O4 directly correlate with polymer brush density and Fe3O4 nanoparticle stability,11 these differences are not only of scientific interest but also crucial for many industrial and medical applications. In particular, it has been demonstrated that anchors can both lead to too weak reversible binding and to too strong binding depending on their affinity to the cations of the investigated oxide.11 Whereas too low binding affinity leads to nanoparticle aggregation, too strong binding of anchors results in nanoparticle dissolution. Thus, in the quest to optimize the binding strength to oxides in general and iron oxide in particular it is of highest importance to understand the binding mechanisms in detail. The structure13-16 and electronic interactions17,18 of different catechols complexed with Fe3+, primarily dopamine and LDOPA, a post-translationally modified amino acid abundantly present in the mussel adhesive protein Mytulis edulis,19 have been studied in detail. The crystallographic and electronic structure of peptides,20 proteins21 and models for catechol dioxygenases16,22-24 complexed with iron ions have been reported. Furthermore, iron-catalyzed catechol degradation has been thoroughly studied in view of its biological relevance.25-31 However, to the best of our knowledge, no such studies on the adsorption of catechols and catechol derivatives on iron oxides have been reported. Here we present detailed insights into the electronic structures and coordination to iron ions of five different catechol derived anchors upon adsorption to Fe3O4

10.1021/jp1109306  2011 American Chemical Society Published on Web 12/15/2010

684

J. Phys. Chem. C, Vol. 115, No. 3, 2011

Amstad et al.

Figure 1. While catechols bind weakly and reversibly to iron oxide nanoparticles, the binding affinity of mimosine is high enough to remove Fe3+ ions through complexation which gradually dissolves iron oxide nanoparticles. Nitrocatechols have an intermediate affinity to iron oxide nanoparticles and strongly adsorb on Fe3O4 nanoparticles without dissolving them. R ) H for dopamine and nitrodopamine and R ) COOH for DOPA and nitroDOPA respectively.

surfaces (Figure 1), which explain the unusually high affinity of nitrocatechols to magnetite (Fe3O4) nanoparticles. A deeper understanding of how these anchors bind may make it possible to theoretically predict binding affinities of different anchors toward desired oxides and provide design rules for selecting dispersants with optimal binding affinities without the need for laborious experimental screening. Experimental Methods Materials. NitroDOPA and nitrodopamine, poly(ethylene glycol)-nitroDOPA where the poly(ethylene glycol) had a molecular weight of 5 kDa (PEG(5)-nitroDOPA) and PEG(5)dopamine were synthesized as described previously.32 DOPA (purity ) 99%) was purchased from Acros, 3-hydroxytryamine-

hydrochloride (purity > 98.5%), FeCl3 (97%), tris(hydroxymethyl)aminomethane, KBr, NaCl from Fluka, mimosine (purity ) 98%), Fe(II)acetate (99.995%, batch 517933, Lot 03901JJ), N,N-dimethylformamide (DMF) (puriss) and FeCl2 (98%) from Sigma, and ethanol (analytical grade) from Scharlau. Complexation. For FTIR and UV-vis measurements, 0.137 µmol anchors were added to 0.137 µmol FeCl2 and FeCl3, respectively, which was dissolved in 1 mL Millipore water (R ) 18.2 Ω, TOC < 6 ppb), 1 mL 10 mM tris(hydroxymethyl)aminomethane containing 150 mM NaCl (Tris), EtOH and DMF respectively. For EPR investigations 40 µmol/mL nitroDOPA was complexed with Fe3+ and Fe2+ at a molar ratio of nitroDOPA:iron ion ) 3:1 and 1:1, respectively, in Millipore water. Unless stated otherwise, these solutions were

Binding Affinity of Anchors to Fe3O4 Nanoparticles left at RT for 1 h before they were analyzed with UV-vis spectroscopy or freeze-dried for FTIR and EPR investigation (freeze-dryer ALPHA 1-2/LDplus, Kuhner LabEquip, Switzerland). Iron Oxide Nanoparticles. Fe3O4 nanoparticles were synthesized as described earlier using a microwave-assisted nonaqueous sol-gel route.32 Briefly, 173 mg Fe(ac)2 was dissolved in 5 mL benzylalcohol and heated for 3 min to 180 °C in the microwave (Discover S-class, CEM, NC, USA).33 Particles were washed once with 10 mL ethanol before they were redispersed in fresh 10 mL ethanol. The particles were coated within 4-8 h after synthesis. To obtain surface oxidized nanoparticles, Fe3O4 nanoparticles were kept in EtOH in air at 4 °C for 4 months before they were stabilized. Oxidation of the aged nanoparticles was confirmed with X-ray photoelectron spectroscopy (XPS) measurements (Supporting Information Figure S1). For FTIR samples, 200 µg of particles were added to 1 mL DMF or EtOH containing 3.85 µmol of the respective anchor. Pure anchors, rather than PEG-anchors, were chosen for FTIR to minimize the interference from the C-O and C-C stretching bands of the PEG moieties with vibrations from anchors. For EPR studies, 100 µg particles, dispersed in 1 mL EtOH, were stabilized with 1.54 µmol PEG(5)-nitroDOPA, PEG(5)-dopamine, or PEG(5)mimosine. Because unstabilized nanoparticles are less susceptible to agglomeration in organic solvents, compared to aqueous solutions, they were stabilized in EtOH and DMF, respectively. Furthermore, highest PEG grafting densities can be achieved if PEG is adsorbed in a collapsed state.34 High packing densities have been shown to directly translate into high nanoparticle stability.11 Therefore, most of the nanoparticles were stabilized in EtOH in which PEG has poor solubility. Anchors were adsorbed for 24 h at 50 °C while they were constantly mechanically mixed at 500 rpm (Thermomixer comfort, VaudauxEppendorf, Switzerland). Excessive PEG(5)-dispersants were removed by 24 h dialysis against Millipore water using dialysis membranes with a cutoff of 25 kDa (Spectra/Por dialysis membrane, spectrum laboratories, Netherlands), while iron oxide nanoparticles coated with anchors only and analyzed by FTIR were purified by washing them 10 times with Millipore water through centrifugation for 10 min at 13 400 rpm. Purified nanoparticles were freeze-dried (freeze-dryer ALPHA 1-2/ LDplus, Kuhner LabEquip, Switzerland). Flat Iron Oxide Surface Preparation. Thin ( 200 °C. (c) PEG(5)-nitroDOPA stabilized Fe3O4 (solid line) and to Fe2O3 surface oxidized (dashed line) nanoparticles.

Fe3O4 nanoparticles (∆B ) 92 ( 6 mT). The lower ∆B indicates less magnetic interparticle interactions between stabilized as compared to unmodified Fe3O4 nanoparticles. Moreover, the stabilized nanoparticles revealed two additional signals at Br ) 352 mT (resulting in g ) 2.0) and Br ) 180 mT (g ) 3.9). The sharp signal at g ) 2.0 is characteristic for free electrons. Such signals were only found if dispersants were adsorbed on Fe3O4 nanoparticles (Supporting Information Figure S4). The low field resonance at Br ) 180 mT (g ) 3.9) can be assigned to rhombohedrally distorted, magnetically decoupled Fe3+.39,40 Despite that the two signals occur simultaneously, their relative intensity, deduced from their peak heights, were different. The relative intensity ratio of free electrons (g ) 2.0) to Fe3+ (g ) 3.9) was highest for PEG(5)-nitroDOPA stabilized nanoparticles. This ratio decreased by 14 ( 5% for PEG(5)-mimosine and 32 ( 1% for PEG(5)-dopamine compared to PEG(5)-nitroDOPA stabilized Fe3O4 nanoparticles. Therefore, electron delocalization is suggested to be highest for PEG(5)-nitroDOPA stabilized nanoparticles. The occurrence of a g ) 3.9 signal associated with magnetite nanoparticles indicates that Fe3+ is magnetically decoupled from the bulk material.41 Such a configuration is most likely to occur at the surface and caused by the binding of anchors. To verify this interpretation, dispersants were decomposed by subjecting stabilized iron oxide nanoparticles to T > 200 °C. The removal of the dispersants from the nanoparticle surfaces should result in EPR spectra similar to those of uncoated Fe3O4 nanoparticles. Indeed, after subjecting PEG(5)-nitroDOPA and uncoated Fe3O4 nanoparticles to T > 200 °C for 10 min, the signal arising from superparamagnetic nanoparticles exhibited a similar, broad line width for all samples. The broadening for the coated nanoparticles is due to an increase of magnetostatic interactions and clearly shows that the thermal treatment decomposed the coating which was further supported by FTIR measurements (Supporting Information Figure S5). Furthermore, X-ray diffraction (XRD) spectra showed that the inverse spinel structure of the iron oxide cores was retained after exposing the cores to elevated temperatures (Supporting Information Figure S6). Upon heating both signals at g ) 2.0 and g ) 3.9 disappeared (Figure 2c). The correlated appearance and loss of the EPR signals of free electrons and magnetically decoupled Fe3+ ions upon, respectively, adsorption and removal of dispersants on Fe3O4 nanoparticles demonstrates that these two signals were related to each other. Hence adsorbed nitroDOPA was preferentially bound to Fe3+ rather than to Fe2+, which is not directly detectable with EPR.42 We have also shown that nitroDOPA/Fe3+ complexes lead to a signal at g ) 4.25. Thus, the g ) 3.9 signal cannot be assigned to nitroDOPA/Fe3+ complexes. The magnetically decoupled Fe3+ signal with g )

Figure 3. (a) Suggested binding mechanism of nitrocatechols which initially bind to Fe2+ ions. (b) Iron-catalyzed catechol degradation. Fe3+ complexed catechols are oxidized to semiquinones while Fe3+ is reduced to Fe2+ before semiquinones are further degraded through reactions with O2.26 (c) Mimosine complexed with Fe3+. R1 ) H for nitrodopamine and dopamine and COOH for nitroDOPA and DOPA whereas R2 were amines for pure feet and PEG(5) for dispersants used to stabilize nanoparticles.

3.9 is thus due to strong interactions of the nitroDOPA with Fe3+ ions at the nanoparticle surface. Upon adsorption of nitroDOPA on Fe3O4 nanoparticles, nitroDOPA can interact with Fe2+ or Fe3+. If nitroDOPA initially binds to Fe3+ ions, the simultaneous appearance of free electrons and a magnetically decoupled Fe3+ signal cannot be explained, since the charge transfer to create the radical is unaccounted for. Alternatively, nitroDOPA can bind to Fe2+ which leads to an increased electron density in the π* system of nitroDOPA and a pronounced electron depletion at the surface bound iron ion (see suggested reaction in Figure 3a). For the latter case, the simultaneous appearance of a magnetically decoupled Fe3+ signal and free electrons would be expected. In the literature, a reduction of nitrobenzenes43 as well as facilitated oxidation of Fe2+ to Fe3+ in the presence of chelates that strongly bind to Fe3+ have already been reported.44,45 PEG(5)-nitroDOPA stabilized oxidized nanoparticles revealed a significantly weaker signal at g ) 2.0 compared to PEG(5)-

Binding Affinity of Anchors to Fe3O4 Nanoparticles

J. Phys. Chem. C, Vol. 115, No. 3, 2011 687

Figure 4. FTIR spectra of (A) pure anchors were compared to those of anchors complexed with (B) Fe2+ and (C) Fe3+ ions in Millipore water at a molar ratio of anchor:iron ion ) 1:1. Complexes were kept for 24 h at 50 °C before they were freeze-dried. Furthermore, anchors were adsorbed on (D) Fe3O4 nanoparticles. (E) Uncoated Fe3O4 nanoparticles and (F) benzoquinones are shown as references. The investigated anchors were (a) nitroDOPA, (b) nitrodopamine, (c) DOPA, (d) dopamine, and (e) mimosine. The most apparent vibrations measured on anchors adsorbed on Fe3O4 nanoparticles were the C-C ring out of plane vibration (*), the C-O stretch vibrations (b) and for nitrocatechols asymmetric (9) and symmetric (2) NO2 vibrations.

nitroDOPA stabilized Fe3O4 nanoparticles, supporting the interpretation that the change in EPR spectra is a result of a strong electron delocalization and charge transfer between nitroDOPA and Fe2+ ions. A clear decoupled Fe3+ signal at g ) 3.9 could also not be detected for stabilized surface-oxidized nanoparticles (Figure 2c). This provides clear evidence for weaker interaction of nitroDOPA with surface oxidized (Fe2O3) compared to as-synthesized Fe3O4 nanoparticles. Further evidence that nitroDOPA initially binds to Fe2+ if adsorbed on Fe3O4 nanoparticles is given by EPR studies of nitroDOPA complexed with Fe2+ and Fe3+ ions, respectively (Supporting Information Figure S7). The intensity ratios of the peak heights resulting from delocalized electrons to Fe3+ were 6 ( 1 and 8 ( 1 for molar ratios of nitroDOPA/Fe2+ equal to 1:1 and 3:1. These were considerably higher than those of nitroDOPA/Fe3+ complexes (2 ( 0 and 3 ( 1 for molar ratios of nitroDOPA/Fe3+ equal to 1:1 and 3:1, respectively) despite that the signal/noise ratio of the Fe3+ signal was comparable for these complexes. The Fe3+ signal of nitroDOPA/Fe2+ complexes at g ) 4.2 further indicates that Fe2+, which is not directly observable with EPR,42 is oxidized to EPR active Fe3+ if complexed with nitroDOPA. More importantly, it shows that interactions of nitroDOPA with Fe2+ lead to stronger electron delocalization compared to nitroDOPA/Fe3+ complexes. The enhanced electron delocalization of nitroDOPA/Fe2+ complexes thus likely results from a strong electron delocalization and charge transfer between nitroDOPA and Fe2+, which is in

support of the similar observations and conclusions described above for the surface bound ions. FTIR Studies. To further investigate structural changes that occur upon adsorption, FTIR studies were performed on nitroDOPA, nitrodopamine, DOPA, dopamine, and mimosine adsorbed on Fe3O4 nanoparticles. Additionally, FTIR spectra of anchors adsorbed on Fe3O4 nanoparticles were compared to spectra of anchors complexed with Fe2+ and Fe3+ ions. FTIR spectra revealed two pronounced peaks upon adsorption of anchors to Fe3O4 nanoparticles, one at around 1280 cm-1 assigned to in-plane CO stretching vibrations and one at 1496 cm-1 assigned to tangential normal C-C vibration modes of the aromatic ring46 (Figure 4a,b and Supporting Information Table S2). Changes in the C-C ring vibration of nitrocatechols upon complexation with iron ions or adsorption on Fe3O4 nanoparticles were small (Figure 4a,b and Supporting Information Table S2). However, shifts of the C-O stretch vibrations from 1290 to 1280 cm-1 for nitroDOPA and from 1294 to 1280 cm-1 for nitrodopamine upon adsorption on Fe3O4 nanoparticles compared to reference spectra are close to that of nitrocatechols complexed with Fe2+ (1285 cm-1for nitroDOPA and 1278 cm-1 for nitrodopamine) in contrast to nitrocatechol/Fe3+ complexes (1287 cm-1 for nitroDOPA and 1290 cm-1 for nitrodopamine). Moreover, the C-O ring vibrations of Fe2+-complexed nitrocatechols were gradually shifted toward lower wavenumbers with time (Figure 5 and Supporting Information Table S2). The

688

J. Phys. Chem. C, Vol. 115, No. 3, 2011

Amstad et al.

Figure 5. FTIR spectra of (A) pure anchors, anchors complexed with (B) Fe2+ and (C) Fe3+ kept for 1 h at RT, (D) Fe2+ and (E) Fe3+ kept for 24 h at 50 °C and (F) anchors adsorbed on Fe3O4 nanoparticles for (a) nitroDOPA, (b) nitrodopamine, (c) DOPA, (d) dopamine, and (e) mimosine. The molar ratio of the anchor-iron complexes was always 1:1. Identical to Figure 4, the most apparent vibrations were assigned to the C-C ring out of plane vibration (*), the C-O stretch vibrations (b) and for nitrocatechols asymmetric (9) and symmetric (2) NO2 vibrations.

good agreement of the C-O ring vibrations of Fe2+-complexed nitrocatechols with that of nitrocatechols adsorbed on Fe3O4 nanoparticles and the time dependent shift toward even closer agreement further supports the conclusions from the EPR results of initial binding of nitroDOPA to Fe2+ followed by strong electron delocalization and charge transfer between nitroDOPA and Fe2+. The FTIR spectra also display marked shifts to higher wavenumbers for the symmetric and asymmetric NO2 vibrations upon adsorption of nitrocatechols on Fe3O4 nanoparticles (e.g., 1317 and 1551 cm-1 for symmetric and asymmetric NO2 vibrations of nitroDOPA) compared to reference spectra (1331 and 1535 cm-1)47 (Figure 4a,b). This indicates that the electron transfer to nitroDOPA resulted in an increased electron density on the nitro group. Increased electron density in the lowest unoccupied molecular orbital (LUMO) upon deprotonation has already been calculated,47 which agrees with the observed increased electron density in the LUMO upon binding of nitroDOPA to Fe3O4 nanoparticles. Differences in FTIR spectra of catechols complexed with Fe2+ and Fe3+ and that of catechols adsorbed on Fe3O4 nanoparticles were more pronounced relative to those of nitrocatechols (Figure 4c,d and Supporting Information Table S2). Whereas the C-C ring vibrations of nitrocatechols were only slightly shifted toward higher wavenumbers upon adsorption to Fe3O4 nanoparticles, marked shifts toward the position of benzoquinones were recorded for (unsubstituted) catechols if adsorbed on Fe3O4 nanoparticles (Figure 4c,d). Furthermore, the C-C ring vibrations of DOPA complexed with iron ions were significantly less

apparent after complexes were kept at 50 °C for 24 h compared to those measured on DOPA that was complexed with iron ions for 1 h at RT (Figure 5c). This agrees with a gradual degradation of DOPA. The similarity of FTIR spectra of especially dopamine adsorbed on Fe3O4 nanoparticles with the reference spectra of benzoquinone is striking. Therefore, the peak at 1484 cm-1 measured for catechols adsorbed on Fe3O4 nanoparticles can likely be assigned to a C-O stretching vibration of semiquinones.48-50 The absence of an electronegative NO2 group on catechols compared to nitrocatechols renders catechols more prone to oxidation.17,51,52 Iron-catalyzed catechol degradation that results in semiquinones, quinones, and eventually in carboxy containing species has been thoroughly described in literature.23,25-27,44 It can explain the weak C-C ring and C-O stretch vibrations of catechols compared to nitrocatechols seen in FTIR. Furthermore, carboxy groups, which can be the result of iron-catalyzed catechol degradation, have been shown to poorly bind to Fe3O4 nanoparticles.11 This might explain the considerably worse stability of PEG(5)-catechol compared to PEG(5)-nitrocatechol stabilized Fe3O4 nanoparticles.11 In contrast to catechol/iron and nitrocatechol/iron complexes, FTIR spectra of mimosine/iron complexes did not change significantly with complexation time and temperature (Figure 5 and Supporting Information Table S2). This indicates a fast reaction between mimosine and iron which resulted in stable complexes and might be due to the low pKa values of mimosine53 (Figure 3c). Similar to nitrocatechols, the location of the C-C ring vibration did not change significantly upon

Binding Affinity of Anchors to Fe3O4 Nanoparticles

J. Phys. Chem. C, Vol. 115, No. 3, 2011 689

TABLE 1: Locations of HOMOfLUMO Transition UV-vis Peaks of Uncomplexed and with Fe2+ and Fe3+ Ions Complexed Nitrocatechols Dissolved at Different pHs and in Organic Solvents, Respectively, Where the Molar Ratio of Nitrocatechols/Iron ) 1:1 pH ) 5 pH ) 7.4 pH ) 12 EtOH nitroDOPA nitroDOPA/Fe2+ nitroDOPA/Fe3+ nitrodopamine nitrodopamine/Fe2+ nitrodopamine/Fe3+

352 353 391 351 353 390

422 414 406 422 418 405

499

355 382 372 356 353 345

501

DMF 364/441 369 366 446 421 363

TABLE 2: Peak Locations of Electron Transition UV-vis Peaks of Uncomplexed and with Fe2+ and Fe3+ Ions Complexed Anchors Dissolved at Different pHs and in Solvents, Respectively, Where the Molar Ratio of Anchors/ Iron Ions is 1:1 pH ) 5

pH ) 7.4

EtOH

2+

nitroDOPA/Fe nitroDOPA/Fe3+ nitrodopamine/Fe2+ nitrodopamine/Fe3+ DOPA/Fe2+ DOPA/Fe3+ dopamine/Fe2+ dopamine/Fe3+ mimosine/Fe2+ mimosine/Fe3+

644 644 402/742 405/744 396/463 513

broad broad broad 574 573 575 561 395/452 448

667 739? 356/739

DMF 406 420/s 520/s 637

356

359 602/471 364

363

361

adsorption of mimosine on Fe3O4 nanoparticles (Figure 4e). However, the C-O ring vibration of mimosine and nitrocatechols complexed with Fe2+ was shifted toward lower wavenumbers compared to the respective reference spectra. Interestingly, in contrast to nitrocatechols, the C-O ring vibration of mimosine adsorbed on Fe3O4 nanoparticles was closer to that of mimosine complexed with Fe3+ compared to mimosine/Fe2+ complexes. This indicates that mimosine binds directly to Fe3+ and is unlikely to bind to Fe2+ and undergo a redox reaction as observed for nitrocatechols. UV-vis Spectroscopy. To further elucidate the role of adsorption conditions on the binding kinetics of these anchors to Fe3O4 nanoparticles, UV-vis spectroscopy measurements on complexes of anchors with free iron ions were performed (Supporting Information Figures S8-S10 and Tables 1 and 2). Primarily electron transitions between the highest occupied orbital (HOMO) and the lowest unoccupied orbital (LUMO) were investigated because these electron transitions give insights into anchor-iron interactions. Reference spectra were compared to UV-vis spectra of anchors complexed with Fe2+ and Fe3+ ions at different pHs and in organic solvents respectively. The pH of Millipore water based complex solutions varied between 3 and 5. Tris has been reported not to interfere with iron ions37 and was chosen to buffer the solutions to pH ) 7.4. Furthermore, anchors were complexed with iron in organic solvents, namely EtOH and DMF, because iron oxide nanoparticles were stabilized in these solvents. The prominent peak located between 350 and 500 nm was assigned to the HOMOfLUMO transition resulting from a charge transfer of the HOMO localized on the aromatic ring of nitrocatechols to the LUMO mainly localized on the nitro group. The wavelength of the HOMOfLUMO transition peak maxima increased from 350 at pH 5 to 420 at pH 7.4 and to 500 nm at pH ≈ 12. Considering the pKa values of nitrocatechols (pKa1 ≈ 6.5, pKa2 ≈ 10),11,54 these peaks were assigned to the fully protonated, once deprotonated, and twice deprotonated state of

nitrocatechols, respectively. A comparison of the pH-dependent reference spectra of nitrocatechols aliquotted in aqueous solutions to reference spectra of nitrocatechols dissolved in EtOH and DMF reveals that nitrocatechols were fully protonated if dissolved in EtOH whereas a significant fraction of nitrocatechols was once deprotonated if aliquotted in DMF. While the peak locations of nitroDOPA and nitrodopamine aliquotted in DMF were identical, relative peak intensities differed markedly indicating that a higher amount of nitrodopamine was partially deprotonated compared to nitroDOPA (Supporting Information Figure S8). One hour after complexation, changes in the HOMOfLUMO transition peak of Millipore water-based nitrocatechol/Fe2+ complex solutions compared to the reference (free) nitrocatechol spectra were negligible (Table 1 and Supporting Information Figure S8 and S9). This was in contrast to nitrocatechols complexed with Fe3+ where nitrocatechol/Fe3+ interactions were seen already 1 h after complexation (Supporting Information Figure S10). Well in agreement with FTIR results, a significant broadening of the HOMOfLUMO peak became apparent 24 h after nitrocatechols were complexed with Fe2+. No further change in the UV-vis spectra of nitrocatechol/Fe3+ complexes was measured if these complexes were kept in Millipore water at RT for 24 h compared to spectra taken 1 h after complexation (Supporting Information Figure S11). These results indicate that fully protonated nitrocatechols interacted faster with Fe3+ compared to Fe2+. However, fully protonated nitrocatechols started to interact with Fe2+ within 24 h. The different UV-vis and FTIR peak locations and shapes of nitrocatechol/Fe2+ and nitrocatechol/Fe3+ complexes indicate different interactions of nitrocatechols with Fe2+ and Fe3+. Thus, time dependent changes in the UV-vis spectra of nitrocatechol/Fe2+ complexes cannot be assigned to a nitrocatechol-independent oxidation of Fe2+ to Fe3+ but should be ascribed to a slow reaction between nitrocatechols and Fe2+ which is absent if nitrocatechols are complexed with Fe3+. HOMOfLUMO transition peaks of nitrocatechol/iron complexes were considerably broader and shifted toward higher wavelengths already 1 h after complexation if once deprotonated nitrocatechols were complexed with Fe2+ compared to fully protonated nitrocatechol/Fe2+ complexes (Table 2, Supporting Information Figure S9). This indicates that once deprotonated nitrocatechols interacted faster especially with Fe2+ ions compared to the fully protonated nitrocatechols. Hence, adsorbing nitrocatechols in a deprotonated state accelerates their binding to Fe3O4 surfaces and increases the adlayer formation rate which might be beneficial for industrial processes. Nitrocatechol adsorption on Fe3O4 nanoparticles in a fully protonated form from EtOH and in a partially once deprotonated form from DMF resulted in equal EPR and FTIR spectra. Therefore, adsorption of nitrocatechols on Fe3O4 nanoparticles in a fully protonated form slows their adsorption down but does not prevent or alter it. Deprotonation of nitrocatechols upon adsorption to iron oxide nanoparticles is well in agreement with literature where protons from alcohols of catechols have been reported to dissociate upon adsorption on TiO2.55-58 As is shown here, adsorption of catechol derivatives is facilitated if these anchors are at least partially deprotonated already prior to adsorption. The NO2 group lowers pKa1 values of nitrocatechols to 6.749 compared to that of catechols (pKa > 8.5).59 Therefore, nitrocatechols can more readily bind to iron ions and adsorb on Fe3O4 surfaces in a pH range between 6.5 and 9 compared to catechols.

690

J. Phys. Chem. C, Vol. 115, No. 3, 2011

No differences between reference spectra of catechols were seen whether they were aliquotted in Millipore water or Tris buffer, indicating that they were protonated under these conditions. This was expected considering their pKa values (pKa > 9)59 (Supporting Information Figure S8). Furthermore, the absence of electron transfer peaks of Millipore water based catechol/Fe2+ complex solutions points to negligible interactions of catechols with Fe2+ ions if dispersed in Millipore water irrespective of complexation time, well in agreement with what has been reported.60 This is in contrast to catechol/Fe3+ complexes that showed electron transfer peaks above 500 nm. The considerably higher pKa1 value of catechols compared to nitrocatechols markedly decreases the affinity of catechols to Fe2+ and prevents catechol/Fe2+ interactions if complexed in Millipore water irrespective of complexation time and in contrast to what was observed for nitrocatechols. While UV-vis spectra of Tris and Millipore water-based mimosine/Fe3+ complexes were independent of the complexation time and temperature, the electron transfer peak of mimosine/Fe2+ complexes shifted toward that of the respective mimosine/Fe3+ complexes with increasing complexation time (Supporting Information Figure S9). This might indicate oxidation of Fe2+ to Fe3+ and hints to strong mimosine/Fe3+ interactions but a low affinity of mimosine to Fe2+ which is in stark contrast to nitrocatechols. Whereas nitrocatechols have a high affinity to iron61 and lead to good Fe3O4 nanoparticle stability,11 catechols have been reported to undergo iron catalyzed degradation62 and thus lead to poor Fe3O4 nanoparticle stability.11 However, mimosine is known to have a high affinity toward Fe3+(see ref 63) but only leads to intermediate Fe3O4 nanoparticle stability if used as an anchor group for grafting low molecular weight dispersants to nanoparticles.11 To elucidate this apparent contradiction, UV-vis spectra of anchor/iron complexes were compared to spectra taken from supernatants of Fe3O4 nanoparticles stabilized with PEG(5)-anchors and dispersed in Millipore water. UV-vis spectra revealed a charge transfer peak in supernatants of Fe3O4 nanoparticles stabilized with PEG(5)-mimosine and dispersed in Millipore water identical to Millipore based mimosine/Fe3+ complexes (Supporting Information Figure S12). This close similarity suggests that the strong complexation of PEG(5)mimosine removed Fe3+ ions from the surface resulting in gradual Fe3O4 nanoparticle dissolution and likely is the reason for only intermediate nanoparticle stability if PEG(5)-mimosine was used as dispersant.11 No such PEG(5)-anchor/iron ion complexes were found in supernatants of PEG(5)-nitrocatechol or PEG(5)-catechol stabilized Fe3O4 nanoparticles (Supporting Information Figure S12). Further evidence for iron oxide dissolution upon adsorption of mimosine was given by near edge X-ray absorption fine structure spectroscopy (NEXAFS) recorded on flat Fe3O4 surfaces coated with mimosine and DOPA, respectively. Significant differences in the oxygen NEXAFS spectra of mimosine coated iron oxide surfaces compared to reference spectra are indicative of changes in the iron oxide stochiometry upon mimosine adsorption. Furthermore, mimosine adsorption led to increased electron density of surface iron atoms (Supporting Information Figure S13). None of these changes were detected on DOPA-coated iron oxide surfaces. Thus, the NEXAFS results support UV-vis findings where iron oxide dissolution was observed upon adsorption of mimosine but absent if catechols were adsorbed on iron oxide nanoparticles. Conclusions We have shown that the addition of electronegative groups such as a NO2 to the catechol ring greatly enhances electronic

Amstad et al. interactions between these anchors and both iron ions in solution and, importantly, in iron oxides. Consequently, catechol derivatives that are electronegatively substituted lead to much enhanced Fe3O4 nanoparticle stability if covalently linked to a spacer molecule such as PEG. NitroDOPA (and nitrocatechols in general) was shown to bind to Fe2+ leading to strong electron delocalization. This results in an enhanced electron density at nitrocatechol anchors and an electron depletion at the chelated, surface bound iron ion. The latter was seen as rhombohedrally distorted Fe3+ EPR signal at g ) 3.9. The significance of this strong electron delocalization on the binding affinity of said anchors and thus on the nanoparticle stability was illustrated by iron oxide nanoparticles oxidized to Fe2O3. On Fe2O3 surfaces, no Fe2+ is presented. For these nanoparticles, no rhombohedrally distorted Fe3+ signal at g ) 3.9 was measured with EPR and the corresponding nanoparticle stability of PEG(5)-nitroDOPA-coated Fe2O3 nanoparticles was poor. Furthermore, the addition of an electronegative group to the catechol ring not only lowered the pKa1, which facilitated their adsorption on Fe3O4 surfaces through partial deprotonation at physiologic pH, but also rendered nitrocatechols more oxidation resistant. Oxidation was observed for (unsubstituted) catechols adsorbed on Fe3O4 surfaces and led to significantly reduced binding affinity, which severely influenced stability of PEG(5)-catechol coated Fe3O4 nanoparticles. Important in the context of further optimization of dispersant anchors for nanoparticles, our findings demonstrate that there is an optimal binding affinity of anchors toward the metal ion of (oxide) nanoparticles. Whereas low binding affinities of anchors to iron oxide surfaces lead to reversible adsorption and thus poor nanoparticle stability, too high binding affinity resulted in gradual Fe3O4 nanoparticle dissolution as was exemplified for mimosine where mimosine/Fe3+ complexes were found in the supernatant of PEG(5)-mimosine stabilized Fe3O4 nanoparticles. Thus, the affinity should be high enough to ensure essentially irreversible adsorption in the salt, temperature, and pH range in which sterically stabilized nanoparticles will be applied, but below the limit that causes substrate dissolution through complex formation. Acknowledgment. The authors thank Torben Gillich for the synthesis of nitrocatechols, Professor Markus Niederberger and Idalia Bilecka (Department of Material Science, ETH Zurich) for their support for the iron oxide nanoparticle synthesis, FIRST (ETH Zurich) for providing clean room facility, EMEZ (ETH Zurich) for providing electron microscopy facility and the Swiss Light Source for synchrotron beam time. Furthermore, we are indebted to Dr. Jo¨rg Raabe and Benjamin Watts for their help with NEXAFS measurements at PolLux (PSI, Villigen, Switzerland). COST Action No. D43, EU-FP7-NMP-ASMENA and ETH Zurich are acknowledged for their financial support. Supporting Information Available: TEM images of unstabilized and PEG(5 kDa)-nitroDOPA stabilized iron oxide nanoparticles, XPS spectra of surface oxidized and assynthesized Fe3O4 nanoparticles, DLS of PEG(5)-nitroDOPA stabilized surface oxidized and as-synthesized iron oxide nanoparticles, EPR spectra of pure dispersants and nitroDOPA complexed with Fe2+ and Fe3+ ions at molar ratios of anchor/ iron ) 1:1 and 1:3. Furthermore, a comprehensive list of FTIR peaks is presented. UV-vis spectra of uncomplexed and with Fe2+ and Fe3+ ions complexed anchors in different solvents as well as time dependent UV-vis spectra of anchor/iron complexes aliquotted in Millipore water are given. UV-vis spectra of supernatants of Fe3O4 nanoparticles stabilized with PEG(5)-

Binding Affinity of Anchors to Fe3O4 Nanoparticles anchor dispersants are compared with the respective anchor/ iron complex spectrum. Finally, O K-edge and Fe L-edge NEXAFS spectra of flat Fe3O4 surfaces coated with DOPA and mimosine respectively are shown. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Lin, M. M.; Kim, D. K.; El Haj, A. J.; Dobson, J. IEEE Trans. Nanobioscience 2008, 7 (4), 298–305. (2) Namdeo, M.; Saxena, S.; Tankhiwale, R.; Bajpai, M.; Mohan, Y. M.; Bajpai, S. K. J. Nanosci. Nanotechnol. 2008, 8 (7), 3247–3271. (3) Latham, A. H.; Williams, M. E. Acc. Chem. Res. 2008, 41 (3), 411–420. (4) Sharma, R.; Chen, C. J. J. Nanopart. Res. 2009, 11 (3), 671–689. (5) Cheon, J.; Lee, J. H. Acc. Chem. Res. 2008, 41 (12), 1630–1640. (6) Waters, E. A.; Wickline, S. A. Basic Res. Cardiol. 2008, 103 (2), 114–121. (7) Sosnovik, D. E.; Nahrendorf, M.; Weissleder, R. Basic Res. Cardiol. 2008, 103 (2), 122–130. (8) McCarthy, J. R.; Weissleder, R. AdV. Drug DeliVery ReV. 2008, 60 (11), 1241–1251. (9) Lunov, O.; Syrovets, T.; Buchele, B.; Jiang, X. E.; Rocker, C.; Tron, K.; Nienhaus, G. U.; Walther, P.; Mailander, V.; Landfester, K.; Simmet, T. Biomaterials 2010, 31 (19), 5063–5071. (10) Gamarra, L. F.; Amaro, E.; Alves, S.; Soga, D.; Pontuschka, W. M.; Mamani, J. B.; Carneiro, S. M.; Brito, G. E. S.; Neto, A. M. F. J. Nanosci. Nanotechnol. 2010, 10 (7), 4145–4153. (11) Amstad, E.; Gillich, T.; Bilecka, I.; Textor, M.; Reimhult, E. Nano Lett. 2009, 9 (12), 4042–4048. (12) Zurcher, S.; Wackerlin, D.; Bethuel, Y.; Malisova, B.; Textor, M.; Tosatti, S.; Gademann, K. J. Am. Chem. Soc. 2006, 128 (4), 1064–1065. (13) Lynch, M. W.; Valentine, M.; Hendrickson, D. N. J. Am. Chem. Soc. 1982, 104 (25), 6982–6989. (14) Zirong, D.; Haltiwanger, R. C.; Bhattacharya, S.; Pierpont, C. G. Inorg. Chem. 1991, 30 (22), 4288–4290. (15) Attia, A. S.; Bhattacharya, S.; Pierpont, C. G. Inorg. Chem. 1995, 34 (17), 4427–4433. (16) Heistand, R. H.; Roe, A. L.; Que, L. Inorg. Chem. 1982, 21 (2), 676–681. (17) Girerd, J. J.; Boillot, M. L.; Blain, G.; Riviere, E. Inorg. Chim. Acta 2008, 361 (14-15), 4012–4016. (18) Grillo, V. A.; Hanson, G. R.; Wang, D. M.; Hambley, T. W.; Gahan, L. R.; Murray, K. S.; Moubaraki, B.; Hawkins, C. J. Inorg. Chem. 1996, 35 (12), 3568–3576. (19) Waite, J. H.; Tanzer, M. L. Science 1981, 212 (4498), 1038–1040. (20) Weisser, J. T.; Nilges, M. J.; Sever, M. J.; Wilker, J. J. Inorg. Chem. 2006, 45 (19), 7736–7747. (21) Sever, M. J.; Weisser, J. T.; Monahan, J.; Srinivasan, S.; Wilker, J. J. Angew. Chem., Int. Ed. 2004, 43 (4), 448–450. (22) Lauffer, R. B.; Heistand, R. H.; Que, L. Inorg. Chem. 1983, 22 (1), 50–55. (23) Bruijnincx, P. C. A.; Lutz, M.; Spek, A. L.; Hagen, W. R.; van Koten, G.; Gebbink, R. J. M. K. Inorg. Chem. 2007, 46 (20), 8391–8402. (24) Citadini, A. P. S.; Pinto, A. P. A.; Araujo, A. P. U.; Nascimento, O. R.; Costa, A. J. Biophys. J. 2005, 88 (5), 3502–3508. (25) Cox, D. D.; Benkovic, S. J.; Bloom, L. M.; Bradley, F. C.; Nelson, M. J.; Que, L.; Wallick, D. E. J. Am. Chem. Soc. 1988, 110 (7), 2026– 2032. (26) Jang, H. G.; Cox, D. D.; Que, L. J. Am. Chem. Soc. 1991, 113 (24), 9200–9204. (27) 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. (28) ElAyaan, U.; Herlinger, E.; Jameson, R. F.; Linert, W. J. Chem. Soc., Dalton Trans. 1997, (16), 2813–2818. (29) Linert, W.; Jameson, R. F.; Herlinger, E. Inorg. Chim. Acta 1991, 187 (2), 239–247.

J. Phys. Chem. C, Vol. 115, No. 3, 2011 691 (30) Kalyanaraman, B.; Felix, C. C.; Sealy, R. C. EnViron. Health Perspect. 1985, 64, 185–198. (31) Funabiki, T.; Fukui, A.; Hitomi, Y.; Higuchi, M.; Yamamoto, T.; Tanaka, T.; Tani, F.; Naruta, Y. J. Inorg. Biochem. 2002, 91 (1), 151–158. (32) Amstad, E.; Zurcher, S.; Mashaghi, A.; Wong, J. Y.; Textor, M.; Reimhult, E. Small 2009, 5 (11), 1334–1342. (33) Bilecka, I.; Djerdj, I.; Niederberger, M. Chem. Commun. 2008, (7), 886–888. (34) Malisova, B.; Tosatti, S.; Textor, M.; Gademann, K.; Zurcher, S. Langmuir 2010, 26 (6), 4018–4026. (35) Park, T. J.; Sambasivan, S.; Fischer, D. A.; Yoon, W. S.; Misewich, J. A.; Wong, S. S. J. Phys. Chem. C 2008, 112 (28), 10359–10369. (36) Charkoudian, L. K.; Franz, K. J. Inorg. Chem. 2006, 45 (9), 3657– 3664. (37) Andjelkovic, M.; Van Camp, J.; De Meulenaer, B.; Depaemelaere, G.; Socaciu, C.; Verloo, M.; Verhe, R. Food Chem. 2006, 98 (1), 23–31. (38) Sharma, V. K.; Waldner, F. J. Appl. Phys. 1977, 48 (10), 4298– 4302. (39) Barreto, W. J.; Barreto, S. R. G.; Moreira, I.; Kawano, Y. Quim. NoVa 2006, 29 (6), 1255–1258. (40) Castner, T.; Newell, G. S.; Holton, W. C.; Slichter, C. P. J. Chem. Phys. 1960, 32 (3), 668–673. (41) Gehring, A. U.; Karthein, R.; Reller, A. Naturwissenschaften 1990, 77 (4), 177–179. (42) Abragam, A.; Bleaney, B. Electron paramagnetic resonance of transition ions; Clarendon P. Oxford: 1970; pp xiv+911. (43) Schmitt, H.; Altenburger, R.; Jastorff, B.; Schuurmann, G. Chem. Res. Toxicol. 2000, 13 (6), 441–450. (44) Harris, D. C.; Aisen, P. Biochim. Biophys. Acta 1973, 329 (1), 156– 158. (45) Bates, G. W.; Workman, E. F.; Schlabac., Mr Biochem. Biophys. Res. Commun. 1973, 50 (1), 84–90. (46) Araujo, P. Z.; Morando, P. J.; Blesa, M. A. Langmuir 2005, 21 (8), 3470–3474. (47) Cornard, J. P.; Lapouge, C.; Merlin, J. C. Chem. Phys. 2007, 340 (1-3), 273–282. (48) Floriani, C.; Henzi, R.; Calderaz, F. J. Chem. Soc., Dalton Trans. 1972, (23), 2640–2642. (49) Sofen, S. R.; Cooper, S. R.; Raymond, K. N. Inorg. Chem. 1979, 18 (6), 1611–1616. (50) Suzuki, H.; Nagasaka, M.; Sugiura, M.; Noguchi, T. Biochemistry 2005, 44 (34), 11323–11328. (51) Higuchi, M.; Hitomi, Y.; Minami, H.; Tanaka, T.; Funabiki, T. Inorg. Chem. 2005, 44 (24), 8810–8821. (52) Kawabata, T.; Schepkin, V.; Haramaki, N.; Phadke, R. S.; Packer, L. Biochem. Pharmacol. 1996, 51 (11), 1569–1577. (53) Stunzi, H.; Harris, R. L. N.; Perrin, D. D.; Teitei, T. Aust. J. Chem. 1980, 33 (10), 2207–2220. (54) Gelb, R. I.; Laufer, D. A.; Schwartz, L. M.; Wairimu, K. J. Chem. Eng. Data 1989, 34 (1), 82–83. (55) Moser, J.; Punchihewa, S.; Infelta, P. P.; Gratzel, M. Langmuir 1991, 7 (12), 3012–3018. (56) Rego, L. G. C.; Batista, V. S. J. Am. Chem. Soc. 2003, 125 (26), 7989–7997. (57) Rajh, T.; Chen, L. X.; Lukas, K.; Liu, T.; Thurnauer, M. C.; Tiede, D. M. J. Phys. Chem. B 2002, 106 (41), 10543–10552. (58) Redfern, P. C.; Zapol, P.; Curtiss, L. A.; Rajh, T.; Thurnauer, M. C. J. Phys. Chem. B 2003, 107 (41), 11419–11427. (59) Ishimitsu, T.; Hirose, S.; Sakurai, H. Chem. Pharm. Bull. 1978, 26 (1), 74–78. (60) Jewett, S. L.; Eggling, S.; Geller, L. J. Inorg. Biochem. 1997, 66 (3), 165–173. (61) Hakkinen, P. Finn. Chem. Lett. 1984, (3), 59–62. (62) Shultz, M. D.; Reveles, J. U.; Khanna, S. N.; Carpenter, E. E. J. Am. Chem. Soc. 2007, 129 (9), 2482–2487. (63) Ellis, B. L.; Duhme, A. K.; Hider, R. C.; Hossain, M. B.; Rizvi, S.; vanderHelm, D. J. Med. Chem. 1996, 39 (19), 3659–3670.

JP1109306