Stimulus-Responsive Nanoporous System Based on a Redox-Active

Article pubs.acs.org/Langmuir

Stimulus-Responsive Nanoporous System Based on a Redox-Active Molecular Self-Assembled Monolayer Annik Abel,† Yanlin Wu,‡ and Julien Bachmann*,†,‡ †

Interdisciplinary Nanoscience Center Hamburg, University of Hamburg, Sedanstraße 19, D-20146 Hamburg, Germany Chair of Thin Film Materials Chemistry, Friedrich-Alexander University of Erlangen-Nürnberg, Egerlandstraße 1, D-91058 Erlangen, Germany

‡

S Supporting Information *

ABSTRACT: Ordered arrays of electrically conducting cylindrical nanotubes are created by atomic layer deposition of a thin titanium dioxide layer onto the pore walls of an anodic alumina matrix. All geometric parameters (pore length and diameter and TiO2 layer thickness) are defined and tunable experimentally. The titanium dioxide surface is subsequently functionalized with ferrocenylacetic acid. The chemisorbed ferrocene moieties are oxidized chemically and electrochemically. Monitoring the redox chemistry by UV− visible absorption spectroscopy allows for the quantification of the total density of redox-active units grafted to the surface, as well as the fraction of them oxidized at a given applied potential. The capillary properties of the surface can be adjusted by the applied potential, as quantified by contact angle measurements.

■

INTRODUCTION Many chemical, biological, and physical phenomena are caused by the exchange of energy, materials, or charge carriers at or across interfaces. Prominent examples include heterogeneous catalysis,1 as well as energy conversion in photovoltaics and photosynthesis.2 More specialized applications are based on a functional surface, in particular drug delivery,3 medical diagnostics, microfluidics,4 and electrophoresis. The enhanced reactivity of solid surfaces results from the interruption of the crystal lattice, which implies the lack of ligands in the last layer of ions or atoms. The properties of the surface can, to some extent, be controlled by adjusting the preparative method (growth or deposition under various kinetic controls) or experimental conditions such as the pH value of an electrolyte. This tunability of surfaces stands in contrast to the relatively static characteristics of bulk solids. Additionally, surface functionalization by molecular units opens a much broader range of possibilities; in particular, it allows for the design of dynamically responsive functionalities. For example, surface-grafted molecules can be switched by changing the pH value,5 applying a magnetic field,6 or by photochemical reactions.7 To date, most such fundamental studies have been performed on planar substrates. Here, we will use nanostructured surfaces to increase the surface-to-volume ratio of the samples.8 This is bound to render changes occurring at the grafted molecular monolayer more influential to the macroscopically observed properties of the bulk. The concrete objectives of the present study are: (1) preparation of nanoporous samples functionalized with a stimulus-responsive © XXXX American Chemical Society

self-assembled molecular monolayer, (2) characterization and quantification, with simple laboratory methods, of the selfassembled monolayer (SAM) switching between two states upon an external stimulus, and (3) demonstration that a macroscopic property of the sample can be tuned using the external stimulus. To achieve these goals, we prepare well-ordered nanostructures by controlled electrochemical oxidation (“anodization”) of aluminum metal in acidic electrolytes and subsequently confer the pores of the anodic aluminum oxide (AAO) membrane with electrical conductivity by coating them with a thin TiO2 layer by atomic layer deposition (ALD). This TiO2coated template is finally functionalized with redox-active molecular groups, ferrocenyl-acetic acid (FAA). We monitor the behavior of this redox-active molecular SAM upon application of an electrode potential by UV−visible absorption spectroscopy. We further demonstrate the changes in magnetism, capillarity, and color of the samples upon application of an electrical potential.

■

RESULTS AND DISCUSSION An overview of our preparative procedure is presented in Scheme 1. Nanoporous anodic alumina membranes produced by two-step electrochemical anodization of pure aluminum under a high voltage in an acidic electrolyte provide a uniform Received: June 7, 2017 Revised: July 22, 2017

A

DOI: 10.1021/acs.langmuir.7b01918 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir Scheme 1. Sample Preparation: (a) Two-Step Anodization of Al, (b) Wet Chemical Removal of the Al Substrate (Cu2+) and the Al2O3 Barrier Layer, Pore Widening (H+), (c) ALD of TiO2, (d) RF Sputter-Coating of the Transparent Conducting Oxide, and (e) Surface Functionalization with Molecular SAM

Figure 2. Spectroscopic ellipsometry of a Si wafer before (light green) and after (dark green) TiO2 deposition by ALD. Dotted lines are fitting curves, light red before ALD and dark red after ALD.

TiO2 ALD deposition are chosen so as to ensure homogeneous coating of deep pores: specifically, we used the parameters published previously for similar geometries.13 The samples must be annealed subsequently, a step which not only crystallizes the titania layer to its anatase phase (Figure S2) but also serves to generate oxygen vacancies which provide sufficient electrical conductivity.14 The transparent electrical contact is then defined by RF sputter-coating an indium tin oxide (ITO) layer onto one side of the sample (step (d) of Scheme 1). The surface functionalization with ferrocene moieties (step (e) of Scheme 1) follows a procedure used in the preparation of dye-sensitized solar cells.15,16 The solid samples are exposed to an ethanolic solution of ferrocenylacetic acid (3.5 × 10−4 M) for 4 days. They were then copiously rinsed with fresh ethanol and dried. The functionalization with FAA is accompanied by the appearance of a homogeneous brown color on the initially colorless samples (Figure 3c). The intensity of the color varies from light yellow-beige to dark brown depending on the membrane thickness. Treatment of the membrane with an excess of nitrosonium tetrafluoroborate in acetonitrile and dichloromethane (1:1) causes a further color change to green (Figure 3c), as expected for the oxidation of ferrocene derivatives to the corresponding ferrocenium cation.17 The choice of nitrosonium as the oxidant, a colorless species which upon reduction generates a gaseous byproduct, allows for the reaction to be characterized by UV−visible absorption spectroscopy (Figure 3b). In ethanolic solution, ferrocenylacetic acid absorbs broadly in the region 400−500 nm, whereas a peak shifted to 630 nm appears upon oxidation by NO+ (Figures 3a and S3). A similar behavior is also observed with the functionalized porous samples, although a significant background is due to the absorption of titania, which trails into the visible region (Figure 3b). This demonstrates that the structure and the redox capability of the ferrocene units are maintained after grafting onto the titania surface. In the next step, we investigate the suitability of our nanotubular titania samples (independently of molecular functionalization) to function as electrodes. To this goal, we compare such a nanostructured sample with a commercial glassy carbon electrode and with a flat TiO2 sample (prepared by sputtering 20 nm of Cr and 20 nm Au onto a piece of silicon wafer and then by depositing 12 nm TiO2 by ALD and

Figure 1. Scanning electron micrograph (in top view) of an alumina template prepared by two-step anodization in oxalic acid after under 40 V (step (a) of Scheme 1). The diameter of the pores is about 45 nm, length 20 μm, and interpore distance 100 nm.

alumina matrix in which cylindrical nanopores are arranged in a hexagonal, close-packed array (step (a) of Scheme 1, Figure 1). A subsequent isotropic widening of the pores in phosphoric acid allows for a fine-tuning of the pore diameter (step (b) of Scheme 1). In this paper, we will consider pores of 60 nm diameter arranged with a 105 nm period obtained by anodization in oxalic acid or of 200 nm diameter and 450 nm pitch in phosphoric acid. In both cases, the pore length is set to values between 10 and 100 μm by the anodization duration. The large aspect ratio of these pores (approximately 50−2000) prevents physical thin film coating techniques to be used for generating a continuous, electrically conducting layer on their walls. Instead, ALD, a thin film coating method controlled by well-defined surface chemistry and not by mass transport from the gas phase, is uniquely suited to the conformal coating of such porous structures.9−12 Thus, ALD is used to deposit a transparent, electrically conducting layer of TiO2 of approximately 10 nm thickness on the inner pore walls (step (c) of Scheme 1). This thickness obtained with 400−450 cycles of the ALD reaction between titanium isopropoxide and water at 120 °C is determined by spectroscopic ellipsometry and X-ray reflectivity (XRR) measured on a reference silicon wafer (Figures 2 and S1). The instrumental conditions of the B

DOI: 10.1021/acs.langmuir.7b01918 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

To quantify the absolute amount of FAA oxidized at any given applied potential, we apply chronocoulometry in a redox-inert electrolytic solution consisting of aqueous buffer only. Figure 5

Figure 3. (a) Absorption spectra of ferrocenylacetic acid in solution (c = 3.5 × 10−4 M in EtOH, light green) and after treatment with a stoichiometric oxidant NOBF4 (dark green). (b) Diffuse absorption spectra of AAO with 19 nm TiO2, after functionalization with ferrocenylacetic acid and after oxidation with NOBF4 (from light blue to dark blue). In both cases, the peak of the reduced, neutral state is between 400 and 500 nm, whereas a red shift to approximately 630 nm is caused by oxidation. (c) Photographs of TiO2/AAO samples after various treatments.

Figure 5. Chronocoulometry measurements on two distinct geometries of TiO2 surfaces functionalized with a SAM of ferrocenylacetic acid: a planar one (open circles) and a nanostructured sample (full disks). In each system, the green datapoints correspond to the functionalized sample, whereas the blue ones are recorded on a blank TiO2 surface (unfunctionalized). Red datapoints are obtained as the difference between the two and represent the charge associated with the oxidation of ferrocenylacetic acid.

annealing). If galvanic currents are to be observed, a pulsed method such as differential pulse voltammetry (DPV) must be used to eliminate most of the capacitive current associated with the large surface area of the samples. Figure 4 displays DPV

presents chronocoulometry data recorded on two distinct sample geometries, planar (open circles) and nanoporous (full disks). Beyond the redox potential of +150 mV (vs Ag/AgCl, determined in Figure 4), significant current flows in both FAAfunctionalized systems (green datapoints). The larger microscopic surface area of the nanoporous system is obvious at the first sight in the larger charges measured at any applied potential beyond 150 mV. The effect of irrelevant currents (associated with capacitive charging or any parasitic Faradaic process) can be eliminated by using reference samples bereft of the FAA functionalization (blue)the difference values, presented in red color in Figure 5, represent charges due to the oxidation of the surface-bound ferrocene moieties. The observation that undesired currents represent only a small fraction of the total in the nanostructured samples, as opposed to the planar case, substantiates the usefulness of our nanoporous strategy. It also supports the idea of a homogeneous adsorption of FAA on the whole TiO2 surface area. The curve shape q(E) evident in Figure 5 is very different from the sigmoidal profile obtained in diffusion-limited steadystate experiments of dissolved redox-active compounds and also expected based purely on the Nernst equation. The difference is due to the condensed state of our redox-active unit: in a compact surface monolayer, each redox center experiences strong Coulomb influence from already oxidized neighbors. This in effect generates a continuum of individual redox potentials beyond the value valid for the dissolved compound. A similar situation is encountered with the dissociation of polyacids: it has been recognized and modeled analytically.18 At this point, the compact nature of the SAM has not been demonstrated yet. In this system, however, the combination of redox activity with associated optical changes and a transparent electrode/matrix system allows one to resort to UV−visible

Figure 4. Differential pulse voltammograms recorded on one electrolytic system with three different working electrodes: planar TiO2 layer with an area 0.152 cm2, glassy carbon electrode with 0.0701 cm2, and TiO2/AAO sample with macroscopically determined area Amacro = 0.152 cm2 (from light green to dark green) in an aqueous ferrocenylacetic acid solution in K2HPO4 electrolyte at pH 7. The oxidation peak is at about 150 mV vs Ag/AgCl.

curves recorded on the same electrolyte (aqueous, buffered solution of FAA) with the three types of electrodes. Our nanostructured TiO2 (dark green curve) yields a DPV peak slightly broader than the standard glassy carbon electrode (forest green) but qualitatively similar. The peak maximum is also observed in the same potential range, 150 mV (vs Ag/ AgCl). The TiO2/AAO sample also delivers a better signal than the planar TiO2 system (light green), owing to the larger specific surface area. Let us now turn to the characterization of nanostructured TiO2/AAO samples functionalized with a monolayer of FAA. C

DOI: 10.1021/acs.langmuir.7b01918 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

molar amount derived from the integrated current flown at +700 mV (8 μC) corresponds to 8 × 10−11 mol oxidized. Thus, only 0.2% of the redox-active units of the SAM are switched at this quite large overpotential (owing to the prominent electrostatic repulsions in this confined system). However, the presence of such minute amounts of oxidized ferrocenium centers is substantiated by electron paramagnetic resonance (EPR) spectroscopy, Figure 7. The blue curve, recorded on an

absorption spectroscopy to quantify the areal density of FAA molecules grafted on the surface. The absorption spectra of both oxidation states (FAA and FAA+) in solution are known and can be expressed as molar extinction coefficients ε, in the historic units M−1 cm−1 or the SI units m2 mol−1. Their difference spectrum Δε, presented in green color in Figure 6,

Figure 6. UV−visible difference absorption spectra used for the calculation of areal molecular density. Green, difference spectrum obtained in the FAA/FAA+ system in solution expressed in molar extinction coefficient. Blue, difference spectrum obtained of a nanoporous and FAA-functionalized sample compared before and after oxidation with excess NOBF4. Sample geometry: pore length 53 μm, pore diameter 45 nm (roughness factor: 800).

Figure 7. EPR spectra (30 K, 9.288 GHz) of FAA+ in solution (red) and as a SAM on a nanoporous sample (blue), compared to the background signal for reference (black).

electrochemically oxidized sample, displays a clear, albeit weak, signal at g = 4.17. This feature fits the spectrum recorded on a dilute frozen solution of oxidized ferrocenylacetic acid (red curve), for which the classic ferrocenium features are obtained at g∥ = 4.17 and g⊥ = 1.98. It is also consistent with values measured on ferrocenium adsorbed in MCM-A zeolite (g∥ = 4.19, g⊥ = 2.07).19 Not only the sample’s color and magnetic behavior (Figures 3 and 7, respectively) are affected by electrochemical oxidation. Another macroscopically accessible physical property of the system can be tuned at the turn of the applied voltage knob: its capillary (hydrophilic/hydrophobic) behavior. Figure S4 demonstrates this feature in a comparison of water contact angles on dry nanoporous samples. While the TiO2-coated sample without SAM is hydrophilic (Figure S4a, φ = 55°), its functionalization with the highly apolar ferrocene units renders it much more hydrophobic (Figure S4b, φ = 81°). Chemical oxidation with NOBF4 regenerates the hydrophilic character (Figure S4c, φ = 49°) due to the generation of charged centers, as does electrochemical oxidation (Figure S4d, φ = 33°).

exhibits a characteristic shape with a maximum near 640 nm and minimum near 730 nm. The difference ΔΔε = 15 m2 mol−1 between them is a property of the FAA/FAA+ system, independent of their state of aggregation (Figure 3 demonstrates that the absorption spectra are not affected by adsorption). Thus, it can be used to quantify porous, surfacefunctionalized samples as well. The blue curve of Figure 6 presents the difference spectrum obtained form the comparison of an FAA-grafted nanoporous sample before and after oxidation with NOBF4. From the value ΔΔA = 0.04 determined at 640 nm/730 nm, a simple division by ΔΔε yields an effective molar areal density of 2.7 mmol m−2. Taking into account the pore geometry (Figure 6 caption), one realizes that the microscopy surface area is 800 times larger than the macroscopic sample area (“roughness factor”). With this, the real molar areal density scales to 3.4 μmol m−2 or roughly to 2 × 1018 molecules per m2. In other words, each molecule occupies 50 Å2 on the surface. Considering that ferrocene has a lateral size on the order of 6 Å, our extremely simple UV−visible spectroscopy assessment confirms that a compact SAM is present on our nanostructured sample. We note that techniques such as scanning tunneling microscopy are typically used to observe the formation of SAMs directly, but are limited to planar samples compatible with ultra-high vacuum conditions, and may not yield data representative of the full sample. With this piece of information on the areal density of redoxactive groups, a quantitative interpretation of the data in Figure 5 is possible. Given the sample area (macroscopic 1.5 × 10−5 m2 and microscopic 0.012 m2) and the density value 3.4 μmol m−2, the total molar amount of ferrocene units interrogated in the experiment is 4.13 × 10−8 mol. Meanwhile, the absolute

■

EXPERIMENTAL SECTION

Nanoporous Al2O3 Matrix Preparation. Water was purified in a Millipore Direct-Q-system. All chemicals were ordered from Strem, Sigma-Aldrich or Alfa Aesar and used as received. Aluminum plates of 0.5 mm thickness (99.999% from Goodfellow) were electropolished in a 1:3 (v/v) mixture of 70% HClO4 and EtOH under 20 V for 4 min. A first anodization was performed in 0.3 M oxalic acid at 8 °C under 40 V for 15 h. This incompletely ordered first layer of alumina was dissolved in chromic acid (1.8 g CrO3 and 4.15 mL 85% H3PO4 to 100 mL water) at 45 °C overnight. Subsequently, the second anodization was performed in the same conditions as the first. Its duration determines the length of the pores and was set to values ranging from 3 to 22 h.20 The underlying layer of metallic aluminum was removed in an aqueous copper chloride solution (13.6 g CuCl2·2H2O in 400 mL D

DOI: 10.1021/acs.langmuir.7b01918 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir water + 100 mL 37% HCl). Finally, the barrier layer of alumina closing the pore extremities was opened by dilute phosphoric acid. During this step, the initial pore diameter of 40 nm was simultaneously increased to 60 nm. A thin layer of gold (approximately 60 nm) was then sputtered onto one side of the samples (Cressington sputter coater 108 auto, 3 min, 30 mA). More gold was then deposited galvanically from a commercial AURUNA electrolyte from Umicore until a metallic Au film of approximately 3 μm thickness was obtained. ALD Coating. The inner walls of the pores of the membranes were coated with a thin layer of TiO2 by ALD. ALD was performed in a home-made hot-wall reactor from titanium tetrakis(isopropoxide) at 120 °C. Ti(OiPr)4 was held at 90 °C and H2O at 40 °C. The pulse, exposure, and purge durations were 2, 50, and 60 s for Ti(OiPr)4 and 0.5, 50, and 60 s for water. The samples were then annealed at 400 °C for 4 h in the oven under N2-atmosphere for better crystallinity. The thickness and the crystallinity of the TiO2 deposition were determined on reference samples, obtained by performing ALD in the same conditions on pieces of silicon wafers (from Si-Mat, with 200 nm of thermal oxide), and characterized by spectroscopic ellipsometry, XRR, and X-ray diffraction (XRD). Spectroscopic ellipsometry data were collected from 400 to 1000 nm under a 70° incidence angle with an instrument model EL X-02 P Spec from DRE-Dr. Riss Ellipsometerbau GmbH. Fits were performed using the database of material files provided with the instrument. XRD measurements were performed on an X’Pert-diffractometer from Philips in the Bragg−Brentano geometry using the Cu Kα radiation (potential 45 kV, current 40 mA). XRR data were collected with a PANalytical X’Pert Pro instrument with triple axis monochromators and point detectors. UV−Visible Absorption Spectroscopy. UV−visible absorption spectroscopy was performed with a DH-2000-L light source, an HR4000 spectrometer, and an ISP-50-8-R integrating sphere from Ocean Optics. For solid samples, the diffuse transmittance T was measured, and because the reflection losses were minute, we converted the data to absorption A as A = −log T. Solutions were measured using a quartz cuvette with 1 cm optical path length. Electrochemistry. Electrochemistry was performed in a phosphate buffer solution at pH 7 with a Gamry Reference 600 or G300 instrument as the potentiostat and in a three-electrode configuration using a Pt counter-electrode and Ag/AgCl/KCl(sat.) as the potential reference (0.199 V vs NHE). Spectroelectrochemistry. The samples were prepared with a 300 nm thick layer of ITO deposited by HF sputtering instead of gold. They were then contacted on the one side using silver glue and held in the electrolytic solution in a quartz cuvette. Scanning Electron Microscopy. Scanning electron microscopy images were taken on a Sigma or Evo MA10 instrument from Zeiss at 5 kV (without sample pre-treatment) or at 20 kV (with preliminary sputter-coating of the samples with approximately 3 nm Au). Electron Paramagnetic Resonance. Ferrocene was oxidized by concentrated sulfuric acid, and the FAA solution was oxidized by NOBF4 for EPR measurements. EPR data were collected on a homemade instrument based on a Hewlett-Packard 85021A microwave generator, Hewlett-Packard 85021A bridge, Varian V-4531 multipurpose cavity, Ithaco Dynatrac 393 AFC-lock-in amplifier, EG&G 5210 signal lock-in, and Hewlett-Packard 34401A signal analyzer. Characterization of Capillary Behavior. Water was dropped onto the samples. The contact angles were determined from the photos taken by a digital camera.

the fraction of oxidized molecules. Thereby, the capillary properties can be adjusted at the turn of a knob (using the applied electrical potential). One can consider this functionalized nanoporous membrane as a valve which opens and closes access of aqueous solutions to the pores upon an electrical stimulus, in a manner somewhat reminiscent of some ion channels of biological membranes. Potential applications can be foreseen in microfluidic devices, where actuators must provide function in the absence of mechanically moving parts. To this goal, subsequent steps will need to increase the magnitude of the response and its immediacy. This could be achieved by manipulation of the pores’ geometric parameters.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.7b01918. TiO2 ALD growth rate of 0.045 (±0.005) nm per cycle, X-ray diffraction of a Si substrate after different treatments, titration of FAA with the stoichiometric oxidant NOBF4, and capillary response of nanoporous systems (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Julien Bachmann: 0000-0001-6480-6212 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was funded at the University of Hamburg by the excellence cluster “Nano-spintronics” of the Freie und Hansestadt Hamburg and at FAU by the excellence cluster “Engineering of Advanced Materials” of the Deutsche Forschungsgemeinschaf t. We thank Sascha Bohse for XRR, Dr. Hauke Heller for XRD measurements, as well as Dr. Detlef Görlitz and Stephan Martens for EPR measurements.

■

REFERENCES

(1) Ertl, G. Reactions at Surfaces: From Atoms to Complexity (Nobel Lecture). Angew. Chem., Int. Ed. 2008, 47, 3524−3535. (2) Mrksich, M. A surface chemistry approach to studying cell adhesion. Chem. Soc. Rev. 2000, 29, 267−273. (3) Gupta, A. K.; Gupta, M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials 2005, 26, 3995−4021. (4) Hau, W. L. W.; Trau, D. W.; Sucher, N. J.; Wong, M.; Zohar, Y. Surface-chemistry technology for microfluidics. J. Micromech. Microeng. 2003, 13, 272−278. (5) Lahann, J.; Mitragotri, S.; Tran, T.-N.; Kaido, H.; Sundaram, J.; Choi, I. S.; Hoffer, S.; Somorjai, G. A.; Langer, R. A reversibly switching surface. Science 2003, 299, 371−374. (6) Katz, E.; Sheeney-Haj-Ichia, L.; Basnar, B.; Felner, I.; Willner, I. Magnetoswitchable Controlled Hydrophilicity/Hydrophobicity of Electrode Surfaces Using Alkyl-Chain-Functionalized Magnetic Particles: Application for Switchable Electrochemistry. Langmuir 2004, 20, 9714−9719. (7) Muraoka, M.; Gillett, S. L.; Bell, T. W. Reversible Photoinsertion of Ferrocene into a Hydrophobic Semiconductor Surface: A Chemionic Switch. Angew. Chem., Int. Ed. 2002, 41, 3653−3656.

■

CONCLUSIONS The present work demonstrates the functionalization of an electrically conductive substrate of high specific surface area with a SAM of a redox-active molecule based on ferrocene. The system is amenable to facile quantitative characterization with bulk analytical methods such as UV−visible absorption spectroscopy. The redox-active units within the SAM can be oxidized by chemical means or electrochemically. The latter method provides the advantageous capability of finely tuning E

DOI: 10.1021/acs.langmuir.7b01918 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (8) Shi, W.; Shen, Y.; Ge, D.; Xue, M.; Cao, H.; Huang, S.; Wang, J.; Zhang, G.; Zhang, F. Functionalized anodic aluminum oxide (AAO) membranes for affinity protein separation. J. Membr. Sci. 2008, 325, 801−808. (9) Elam, J. W.; Routkevitch, D.; Mardilovich, P. P.; George, S. M. Conformal Coating on Ultrahigh-Aspect-Ratio Nanopores of Anodic Alumina by Atomic Layer Deposition. Chem. Mater. 2003, 15, 3507− 3517. (10) Bachmann, J.; Zierold, R.; Chong, Y. T.; Hauert, R.; Sturm, C.; Schmidt-Grund, R.; Rheinländer, B.; Grundmann, M.; Gösele, U.; Nielsch, K. A Practical, Self-Catalytic, Atomic Layer Deposition of Silicon Dioxide. Angew. Chem., Int. Ed. 2008, 47, 6177−6179. (11) Bachmann, J. Atomic layer deposition, a unique method for the preparation of energy conversion devices. Beilstein J. Nanotechnol. 2014, 5, 245−248. (12) Nielsch, K.; Bachmann, J.; Daub, M.; Jing, J.; Knez, M.; Gösele, U.; Barth, S.; Mathur, S.; Escrig, J.; Altbir, D. Ferromagnetic Nanostructures by Atomic Layer Deposition: From Thin Films Towards Core-Shell Nanotubes. ECS Trans. 2007, 11, 139−148. (13) Wu, Y.; Assaud, L.; Kryschi, C.; Capon, B.; Detavernier, C.; Santinacci, L.; Bachmann, J. Antimony sulfide as a light absorber in highly ordered, coaxial nanocylindrical arrays: Preparation and integration into a photovoltaic device. J. Mater. Chem. A 2015, 3, 5971−5981. (14) Mathews, N. R.; Morales, E. R.; Cortés-Jacome, M. A.; Antonio, J. A. T. TiO2 thin filmsInfluence of annealing temperature on structural, optical and photocatalytic properties. Sol. Energy 2009, 83, 1499−1508. (15) Grätzel, M. Dye-sensitized solar cells. J. Photochem. Photobiol., C 2003, 4, 145−153. (16) Chen, C.-Y.; Wu, S.-J.; Wu, C.-G.; Chen, J.-G.; Ho, K.-C. A Ruthenium Complex with Superhigh Light-Harvesting Capacity for Dye-Sensitized Solar Cells. Angew. Chem. 2006, 118, 5954−5957. (17) Connelly, N. G.; Geiger, W. E. Chemical Redox Agents for Organometallic Chemistry. Chem. Rev. 1996, 96, 877−910. (18) Neto, A. A.; Filho, E. D.; Fossey, M. A.; Neto, J. R. Polyacids Self-Dissociation Model. J. Phys. Chem. B 1997, 101, 9833−9837. (19) Toda, Y.; Ishimaru, S.; Ikeda, R.; Mitani, T.; Kitao, S.; Seto, M. Oxidation of ferrocene molecules adsorbed in MCM-41. J. Phys. Chem. Solids 2004, 65, 471−473. (20) Gemmer, J.; Hinrichsen, Y.; Abel, A.; Bachmann, J. Systematic catalytic current enhancement for the oxidation of water at nanostructured iron(III) oxide electrodes. J. Catal. 2012, 290, 220− 224.

F

DOI: 10.1021/acs.langmuir.7b01918 Langmuir XXXX, XXX, XXX−XXX