Effect of Surface Chemistry on the Stability of Gold Nanostructures

Nanoscale Synthesis and Characterization Laboratory, Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, California 94550. ‡ Instit...
0 downloads 0 Views 5MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Effect of Surface Chemistry on the Stability of Gold Nanostructures Juergen Biener,*,† Arne Wittstock,*,‡ Monika M. Biener,† Tobias Nowitzki,‡ Alex V. Hamza,† and Marcus Baeumer‡ †

Nanoscale Synthesis and Characterization Laboratory, Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, California 94550, and ‡Institute of Applied and Physical Chemistry, Universit€ at Bremen, Leobener Strasse, 28359 Bremen, Germany Received May 14, 2010. Revised Manuscript Received July 2, 2010

Understanding the role of surface chemistry in the stability of nanostructured noble-metal materials is important for many technological applications but experimentally difficult to access and thus little understood. To develop a fundamental understanding of the effect of surface chemistry on both the formation and stabilization of self-organized gold nanostructures, we performed a series of controlled-environment annealing experiments on nanoporous gold (np-Au) and ion-bombarded Au(111) single-crystal surfaces. The annealing experiments on np-Au in ambient ozone were carried out to study the effect of adsorbed oxygen under dynamic conditions, whereas the ion-bombarded Au single-crystal surfaces were used as a model system to obtain atomic-scale information. Our results show that adsorbed oxygen stabilizes nanoscale gold structures at low temperatures whereas oxygen-induced mobilization of Au surface atoms seems to accelerate the coarsening under dynamic equilibrium conditions at higher temperatures.

1. Introduction Nanoporous Au (np-Au) has recently attracted considerable interest, fueled by its potential use in catalysis,1-6 sensorics3,7-12 and actuator applications.13-15 For example, it has been shown that np-Au is a remarkable catalyst for low-temperature CO oxidation5,6 as well as the selective oxidation of alcohols.1,2 Another example is the recent discovery that surface-chemistryinduced changes in the surface stress can drive a macroscopic sample contraction/expansion of up to 0.5% that can be used for surface-chemistry-driven sensor and actuator technologies.13 The remarkable properties of np-Au are a direct consequence of its *Corresponding authors. E-mail: [email protected], [email protected]. (1) Wittstock, A.; Zielasek, V.; Biener, J.; Friend, C. M.; Baumer, M. Science 2010, 327, 319–322. (2) Dongqing, H.; Tingting, X.; Jixin, S.; Xiaohong, X.; Yi, D. ChemCatChem 2010, 2, 383–386. (3) Ding, Y.; Chen, M. W. MRS Bull. 2009, 34, 569–576. (4) Jia, J.; Cao, L.; Wang, Z. Langmuir 2008, 24, 5932–5936. (5) Xu, C. X.; Su, J. X.; Xu, X. H.; Liu, P. P.; Zhao, H. J.; Tian, F.; Ding, Y. J. Am. Chem. Soc. 2007, 129, 42–43. (6) Zielasek, V.; Jurgens, B.; Schulz, C.; Biener, J.; Biener, M. M.; Hamza, A. V.; B€aumer, M. Angew. Chem., Int. Ed. 2006, 45, 8241–8244. (7) Lang, X. Y.; Guo, H.; Chen, L. Y.; Kudo, A.; Yu, J. S.; Zhang, W.; Inoue, A.; Chen, M. W. J. Phys. Chem. C 2010, 114, 2600–2603. (8) Lang, X. Y.; Guan, P. F.; Zhang, L.; Fujita, T.; Chen, M. W. Appl. Phys. Lett. 2010, 96, 073701. (9) Lang, X. Y.; Chen, L. Y.; Guan, P. F.; Fujita, T.; Chen, M. W. Appl. Phys. Lett. 2009, 94, 213109. (10) Biener, J.; Nyce, G. W.; Hodge, A. M.; Biener, M. M.; Hamza, A. V.; Maier, S. A. Adv. Mater. 2008, 20, 1211–1217. (11) Dixon, M. C.; Daniel, T. A.; Hieda, M.; Smilgies, D. M.; Chan, M. H. W.; Allara, D. L. Langmuir 2007, 23, 2414–2422. (12) Kucheyev, S. O.; Hayes, J. R.; Biener, J.; Huser, T.; Talley, C. E.; Hamza, A. V. Appl. Phys. Lett. 2006, 89, 053102. (13) Biener, J.; Wittstock, A.; Zepeda-Ruiz, L. A.; Biener, M. M.; Zielasek, V.; Kramer, D.; Viswanath, R. N.; Weissmuller, J.; B€aumer, M.; Hamza, A. V. Nat. Mater. 2009, 8, 47–51. (14) Kramer, D.; Viswanath, R. N.; Weissmueller, J. Nano Lett. 2004, 4, 793–796. (15) Weissm€uller, J.; Viswanath, R. N.; Kramer, D.; Zimmer, P.; Wuerschum, R.; Gleiter, H. Science 2003, 300, 312–315. (16) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Nature 2001, 410, 450–453. (17) Erlebacher, J.; Sieradzki, K. Scr. Mater. 2003, 49, 991–996. (18) Roesner, H.; Parida, S.; Kramer, D.; Volkert, C. A.; Weissmueller, J. Adv. Eng. Mater. 2007, 9, 535–541.

13736 DOI: 10.1021/la1019422

characteristic spongelike morphology of interconnecting nanometer-sized ligaments (Figure 1a)16-19 that also makes it an intrinsically unstable material that tends to coarsen by curvaturedriven diffusion. For example, the length scale of both ligaments and pores of np-Au can be increased by more than 2.5 orders of magnitude by simply annealing the material.20 A related phenomenon is the size-dependent thermal stability of gold nanoparticles that is demonstrated by a size-dependent melting temperature21 and the tendency of Au nanoparticles to lose their catalytic activity at higher temperatures because of sintering (i.e., particle growth).22 Understanding what controls the characteristic length scale of noble-metal nanostructures is thus very important for many of the applications mentioned above. Nanoporous Au can be easily prepared by dealloying of Ag-Au alloys, and much progress has been made in understanding the physical mechanism of porosity formation during dealloying16,17 as well as the effects of various dealloying parameters such as electrolyte composition,23 potential,24 time,25 and temperature26 on the resulting length scale. However, the effect of surface chemistry on the stability of the material has attracted little attention, even though one would expect that both the surface diffusion and stability of under-coordinated Au atoms are strongly affected by the presence of adsorbed species during preparation and application. Good examples demonstrating the effect of surface chemistry on the coarsening kinetics are the observations that the presence of halides, specifically iodine, during dealloying can drastically increase the resulting porosity length scale23 and that the presence of chlorine can affect the (19) Fujita, T.; Qian, L. H.; Inoke, K.; Erlebacher, J.; Chen, M. W. Appl. Phys. Lett. 2008, 92, 251902. (20) Li, R.; Sieradzki, K. Phys. Rev. Lett. 1992, 68, 1168–1171. (21) Shim, J. H.; Lee, B. J.; Cho, Y. W. Surf. Sci. 2002, 512, 262–268. (22) Huang, J. H.; Akita, T.; Faye, J.; Fujitani, T.; Takei, T.; Haruta, M. Angew. Chem., Int. Ed. 2009, 48, 7862–7866. (23) Dursun, A.; Pugh, D. V.; Corcoran, S. G. J. Electrochem. Soc. 2003, 150, B355–B360. (24) Cattarin, S.; Kramer, D.; Lui, A.; Musiani, M. M. J. Phys. Chem. C 2007, 111, 12643–12649. (25) Ding, Y.; Kim, Y. J.; Erlebacher, J. Adv. Mater. 2004, 16, 1897–1900. (26) Qian, L. H.; Chen, M. W. Appl. Phys. Lett. 2007, 91, 083105.

Published on Web 07/29/2010

Langmuir 2010, 26(17), 13736–13740

Biener et al.

Letter

near the oxidation threshold of Au.24,29 Because the dissociation efficiency of molecular oxygen on Au is low, we used ozone (O3) and oxygen ions as more reactive sources of oxygen. Annealing experiments on np-Au in an ozone atmosphere allowed us to study the effect of adsorbed oxygen under dynamic conditions where both the adorption and desorption of oxygen are fast. Experiments on ion-bombarded Au single-crystal surfaces, however, were carried out in an ultrahigh vacuum environment and thus allowed us to obtain atomic-scale information by scanning tunneling microscopy (STM). Our experiments reveal that adsorbed oxygen stabilizes the nanoscale morphology of both np-Au and ion-bombarded Au(111) surfaces at low temperatures. At higher temperatures and dynamic equilibrium conditions, however, oxygen seems to accelerate coarsening, indicating an oxygen-induced mobilization of Au surface atoms.

2. Materials and Methods

Figure 1. Effects of temperature and annealing atmosphere on the coarsening of np-Au. Compared to annealing in an inert helium atmosphere, ozone (7 vol % O3 in O2) exposure has a stabilizing effect below 500 K but accelerates the coarsening of the nanoporous structure of np-Au at higher temperatures as demonstrated by the XSEM micrographs shown in the top panel. Below 500 K, ozone exposure leads to a stable oxygen adsorbate layer, which seems to reduce the surface diffusivity of Au. Above 500 K, rapid oxygen desorption leads to a dynamic adsorption/desorption equilibrium which, as a net effect, seems to mobilize Au surface atoms. The blue bar in the lower panel represents the oxygen desorption peak temperature as observed in single-crystal experiments.32

stability of initial structures formed during the dealloying of Cu3Au alloys.27 Surface chemistry can also be expected to be an important parameter in “electrochemical annealing”, which describes the general finding that the rates of many coarsening phenomena have been found to increase as the potential is raised toward more positive values.28 In the present study, we systematically investigate the effect of adsorbed oxygen on both the formation and stabilization of selforganized nanoscale Au structures by using a combination of controlled-environment annealing experiments on nanoporous gold and ion-bombarded Au(111) single-crystal surfaces. Oxygen was chosen for its role in Au-catalyzed oxidation reactions1-3,5,6 and in the surface-chemistry-driven actuator application.13 Furthermore, the presence of adsorbed oxygen seems to reduce the mobility of Au surface atoms as indicated by the experimental observation that smaller-length-scale structures result from dealloying (27) Renner, F. U.; Stierle, A.; Dosch, H.; Kolb, D. M.; Zegenhagen, J. Electrochem. Commun. 2007, 9, 1639–1642. (28) Giesen, M.; Beltramo, G.; Dieluweit, S.; M€uller, J.; Ibach, H.; Schmickler, W. Surf. Sci. 2005, 595, 127–137.

Langmuir 2010, 26(17), 13736–13740

Disk-shaped np-Au samples (diameter ∼5 mm, thickness ∼200 μm) were prepared by dealloying Ag0.7Au0.3 alloy samples in concentrated nitric acid (48 h, ∼65 wt % HNO3, Fluka Chemical Corp.). The Ag concentration in the resulting np-Au samples was below 1 atom % as determined by energy-dispersive X-ray spectroscopy. Annealing experiments with np-Au were carried out in a tube furnace (450-650 K, 3 h) under either an atmosphere of ultrapure He (5.0, Linde AG) or an ozone-oxygen mixture (∼7 vol % O3 in 4.7 O2, Linde AG) at a flow rate of 50 sccm (ozone generator type 802 N, ozone analyzer type 964, BMT Messtechnik Berlin). The coarsening of np-Au during annealing was assessed by cross-sectional scanning electron microscopy (XSEM). At least three samples were investigated for each temperature and gas environment, and several SEM images were taken from each of these samples from different areas of freshly prepared fracture surfaces (created after the annealing procedure) close (∼10 μm) to the outer surface. The average ligament diameter for each sample was then determined from ligament diameter distributions that were obtained by geometrical evaluation (i.e., measuring the diameter of randomly selected ligaments at their center). These average values are those displayed in Figure 1c. Ion-bombardment experiments on Au(111) single crystals were performed in an ultrahigh vacuum system (base pressure of ∼1  10-10 mbar) equipped with commercial instrumentation for scanning tunneling microscopy (STM, Omicron) and X-ray photoelectron spectroscopy (XPS). The Au(111) single crystal was cleaned by cycles of Arþ sputtering, and details of the experimental procedures can be found in an earlier publication.30 STM images were collected at room temperature, and Z-channel (topography) and I-channel (constant height) images were obtained simultaneously. Etched Pt0.8Ir0.2 tips from Molecular Imaging were used. The ion-induced structures studied in the present work were produced by removing roughly four monolayers (1 ML = 1.4  1015 cm-2) from well-annealed Au(111) surfaces using either 500 eV argon or oxygen ions, which is sufficient to reach a steadystate (in terms of a saturation of the step density31) surface morphology. The sputter rate for both argon and oxygen ions was determined by an STM-based method for yield determination.31 The oxygen coverage was determined by XPS.

3. Results and Discussions The results from the annealing experiments on np-Au are summarized in Figure 1. The as-prepared sample shows an average (29) Newman, R. C.; Corcoran, S. G.; Erlebacher, J.; Aziz, M. J.; Sieradzki, K. MRS Bull. 1999, 24, 24–28. (30) Biener, J.; Biener, M. M.; Nowitzki, T.; Hamza, A. V.; Friend, C. M.; Zielasek, V.; B€aumer, M. ChemPhysChem 2006, 7, 1906–1908. (31) Michely, T.; Comsa, G. Nucl. Instrum. Methods Phys. Res., Sect. B 1993, 82, 207–219. (32) Kim, J.; Samano, E.; Koel, B. E. Surf. Sci. 2006, 600, 4622–4632.

DOI: 10.1021/la1019422

13737

Letter

Biener et al.

ligament size of about 30 nm. After annealing at 450 K in an inert He atmosphere, the ligament size increased to approximately 90 nm, whereas the ligament size of the ozone-annealed sample did not change. This indicates that adsorbed oxygen (Oad) from ozone decomposition can stabilize the high-surface-area morphology of np-Au. When the sample is annealed at 650K, the stabilizing effect of ozone is no longer observed and the ligament size increases to approximately 400 nm in both the inert He atmosphere and the reactive O3 environment. Interestingly, the ligaments of the O3annealed sample are now slightly larger (∼450 nm) than those of the He-annealed sample (∼400 nm). The main difference between the two annealing temperatures discussed above is the stability of chemisorbed oxygen. As shown previously, np-Au is an active catalyst for ozone decomposition, and room-temperature ozone exposure leads to a saturation coverage of approximately one monolayer (1 ML = 1.4  1015 cm-2) of chemisorbed atomic oxygen (Oad) without further oxidizing the material,13 in good agreement with Au single-crystal experiments.32,33 Assuming simple first-order desorption kinetics, the mean residence time τ of an adsorbed species is generally given by τ ¼ ν - 1 expðEd =RTÞ where ν is the pre-exponential factor, Ed is the desorption activation energy, R is the universal gas constant, and T is the temperature in Kelvin. Using the values obtained from Au single-crystal experiments (ν = 1013 s-1, Ed = 34 kcal/mol)32 reveals that the mean residence time of Oad on Au at 450 K is on the order 103 s but is drastically reduced to only a few milliseconds at 650 K. The flux of impinging O3 can be estimated from kinetic gas theory and is on the order of 107 ML/s for an O3 partial pressure of ∼0.07 atm. Annealing experiments performed at 450 K thus result in a static oxygen coverage of approximately one monolayer because the rate of desorption is much smaller than the rate of adsorption (as long as the dissociation efficiency of O3 at 450 K is larger than 10-10). At 650 K, however, both the adsorption and desorption of oxygen are fast processes, and the resulting dynamic Oad coverage is lower than the saturation coverage but is not zero. The actual value, however, is difficult to determine because it depends on several parameters such as the sticking and decomposition probabilities of O3 at the annealing temperature. Another complication arises from diffusion limitations that can lead to the formation of ligament size gradients along the sample cross section. Details will be discussed in a separate publication. Nevertheless, one can conclude that the key difference between the 450 and 650 K experiment is the short residence time of adsorbed oxygen at 650 K. To assign the effect to adsorbed oxygen, we carried out experiments with a model system under ultrahigh vacuum (UHV) conditions where the ambient conditions can be precisely controlled and the surface chemistry can be investigated at the atomic level. To prepare a surface structure that resembles the morphology of np-Au, we ion bombarded Au(111) single-crystal surfaces with 500 eV argon or oxygen ions as described previously.34 The self-organized structures evolving during the dealloying of Ag-Au alloy samples and the ion bombardment of Au(111) single crystals are very similar: a 3D bicontinuous nanoporous structure in the case of dealloying (Figure 2a) and a 2D pit-andmound morphology in the case of ion sputtering (Figure 2b). This resemblance is not coincidental but reflects the similarities in the underlying microscopic processes (Figure 2c): dealloying16,17 as (33) Saliba, N.; Parker, D. H.; Koel, B. E. Surf. Sci. 1998, 410, 270–282. (34) Biener, M. M.; Biener, J.; Friend, C. M. Surf. Sci. 2005, 590, L259–L265.

13738 DOI: 10.1021/la1019422

Figure 2. Scanning electron micrograph of nanoporous gold (a) and STM image of an argon ion-bombarded Au(111) surface (b). Note the similarity in the patterns. (c) Schematic representation of the mechanism of pattern formation during dealloying (left) and ion bombardment (right). In both cases, the length scale of the resulting pattern is controlled by the diffusion and nucleation kinetics of adatoms and vacancies. Note that surface diffusion in an electrolyte is usually much faster than in a vacuum environment, which explains the different length scales obtained by dealloying (a) and sputtering (b).

well as sputtering35,36 produces adatoms and vacancies by removing surface atoms; the surface diffusion and nucleation of both vacancies and adatoms then lead to a pit-and-mound surface morphology (Figure 2b). In the case of dealloying, this pit-andmound structure develops into a 3D nanoporous structure as nucleated Au clusters locally passivate the surface against further corrosion.16 However, the steady-state morphology of ion-sputtered Au surfaces remains 2D. In any case, the feature size is controlled by the diffusion and nucleation kinetics of adatoms and vacancies. This suggests that ion-bombarded Au(111) surfaces can be used to study the effect of surfactants on pore formation during dealloying and on the stability of the resulting structure in a well-controlled environment compatible with standard surface science techniques. In the present work, we used argon and oxygen ions to study the effect of inert and reactive ions, respectively. Figure 3 compares typical STM images obtained from Au(111) surfaces after bombardment with 500 eV Arþ (a) and O2þ (b): both surfaces exhibit a similar pit-and-mound morphology, but the characteristic length scale observed after oxygen ion bombardment is smaller by roughly a factor of 2. Ion bombardment of Au(111) surfaces with 500 eV oxygen ions at 300 K leads not only to surface roughening but also to the chemisorption of oxygen with (35) Michely, T.; Comsa, G. Phys. Rev. B 1991, 44, 8411–8414. (36) Vishnyakov, V.; Donnelly, S. E.; Carter, G. Philos. Mag. B 1994, 70, 151–157.

Langmuir 2010, 26(17), 13736–13740

Biener et al.

Figure 3. STM images collected from Au(111) surfaces after bombardment with 500 eV argon (a, c, d) and oxygen (b, e) ions (fluence ∼3  1015 ions/cm2) at room temperature as well as after annealing at the indicated temperature. Note the smaller length scale as well as the higher thermal stability of the oxygen ioninduced surface structure. The scale bar always corresponds to 50 nm. The heating rate was 0.5 K/s, and the annealing time at the indicated temperature was always 300 s.

a saturation coverage of ∼0.3 ML (1 ML = 1.4  1015 cm-2)34 as determined by XPS and Auger electron spectroscopy (data not shown). However, argon was not retained in or on the Au surface using the current sputtering conditions. This suggests that adsorbed oxygen is responsible for the smaller length scale of the oxygen ion-induced surface morphology. Reducing the feature size requires an increase in the density of both vacancy and Au adatom clusters, which in turn points toward a reduced mobility of Au adatoms and vacancies in the presence of adsorbed oxygen.37,38 Indeed, adsorbed oxygen has been reported to reduce the surface mobility of Ag and Cu adatoms in metal deposition experiments.39 Although the observations discussed above can explain the fact that the presence of adsorbed oxygen leads to the formation of smaller features by suppressing surface diffusion, the question is open as to whether adsorbed oxygen also results in improved thermal stability. In an attempt to answer this question, we performed annealing experiments with both surfaces (Figure 3c-e). (37) Murty, M. V. R.; Curcic, T.; Judy, A.; Cooper, B. H.; Woll, A. R.; Brock, J. D.; Kycia, S.; Headrick, R. L. Phys. Rev. Lett. 1998, 80, 4713–4716. (38) Ratsch, C.; Venables, J. A. J. Vac. Sci. Technol., A 2003, 21, S96–S109. (39) Abermann, R.; Koch, R. Thin Solid Films 1986, 142, 65–76.

Langmuir 2010, 26(17), 13736–13740

Letter

These experiments reveal that adsorbed oxygen indeed stabilizes the ion-beam-induced nanometer-scale surface morphology against thermally induced coarsening. Specifically, we observed rapid annealing of the argon ion-induced surface damage at or above 400 K (Figure 3c,d). However, the even rougher pit-andmound surface morphology of the oxygen ion-bombarded Au surface remains stable up to at least 400 K (Figure 3e). This annealing temperature reduces the oxygen coverage, but only by about 40%.34 The stability of the oxygen ion-bombarded Au(111) surface against thermal coarsening is remarkable because the smaller length scale of the oxygen ion-induced roughness is expected to facilitate coarsening. The fact that even a low oxygen coverage (e0.2 ML) is sufficient to stabilize np-Au at 400 K suggests that oxygen decorates step edges and kink sites. In fact, as reported by Kim et al.,32 step sites do bind oxygen more tightly than terrace sites. Rapid coarsening is observed only during annealing at or above 450 K, the onset of oxygen desorption from Au surfaces.33 (Note that the situation is different for the ozone annealing experiments described above where the oxygen coverage is continuously replenished, thus stabilizing the structure of np-Au even at 450 K.) Coming back to Figure 1, an interesting observation can be made: at 650 K, ozone accelerates coarsening rather than slowing it down. We conjecture that this destabilization of np-Au in the presence of O3 at temperatures above 450 K is a consequence of a dynamic oxygen adsorption/desorption equilibrium. Recent STM studies on O3-exposed Au(111) surfaces revealed the formation of Au-O complexes, demonstrating that adsorbed oxygen atoms can “abstract” gold atoms from the surface sites.40 These Au-O complexes are stable up to ∼450 K where they start to decompose again, thus releasing the incorporated Au atoms. The accelerated coarsening kinetics of np-Au above 450 K could thus be explained by the dynamic equilibrium of Au-O complex formation and dissociation that effectively leads to the mobilization of Au surface atoms. Such an adsorbate-induced mobilization of Au surface atoms has also been observed for the case of adsorbed sulfur.41,42 The effect should be most pronounced in the 500-600 K temperature region, where the temperature is still low enough to guarantee a relatively high dynamic concentration of adsorbed oxygen. If, however, the oxygen coverage were the determining factor, then the outcome of the 650 K annealing experiments performed in He and ozone should be very similar because the oxygen coverage becomes very small at this temperature. Alternatively, one might attribute the accelerated coarsening of np-Au in the ozone annealing experiment at 650 K to the “cleaning” effect of ozone by removing residual carbon surface contamination.

4. Conclusions Our experiments demonstrate that surface chemistry is an important factor in controlling the stability of nanostructured noblemetal materials. Adsorbed oxygen, for example, stabilizes nanoscale gold structures at low temperatures whereas the oxygen-induced mobilization of Au surface atoms seems to accelerate the coarsening under dynamic equilibrium conditions at higher temperatures. Our finding also has profound implications for the potential application of gold-based catalysts that is currently hampered by the limited thermal stability of nanodispersed Au particles.43

(40) Min, B. K.; Alemozafar, A. R.; Biener, M. M.; Biener, J.; Friend, C. M. Top. Catal. 2005, 36, 77–90. (41) Biener, M. M.; Biener, J.; Friend, C. M. Langmuir 2005, 21, 1668–1671. (42) Biener, M. M.; Biener, J.; Friend, C. M. Surf. Sci. 2007, 601, 1659–1667. (43) Pattrick, G.; van der Lingen, E.; Corti, C. W.; Holliday, R. J.; Thompson, D. T. Top. Catal. 2004, 30-31, 273–279.

DOI: 10.1021/la1019422

13739

Letter

Acknowledgment. Work at LLNL was performed under the auspices of the U.S. DOE by LLNL under contract DE-AC5207NA27344. J.B. gratefully acknowledges financial support from the Hanse-Wissenschaftskolleg, Germany. We gratefully acknowledge

13740 DOI: 10.1021/la1019422

Biener et al.

the experimental support (SEM) of Ardalan Zargham and Torben Rohbeck (Prof. Falta and Prof. Hommel, Institute for Solid State Physics, University Bremen) and Petra Witte (Prof. Fischer, Crystallography, Geology Department, University Bremen).

Langmuir 2010, 26(17), 13736–13740