ARTICLE pubs.acs.org/Langmuir
Molecular Ordering of Ethanol at the Calcite Surface I. S. Pasarín,*,† M. Yang,† N. Bovet,† M. Glyvradal,‡ M. M. Nielsen,§ J. Bohr,^ R. Feidenhans’l,‡ and S. L. S. Stipp† †
Nano-Science Center, Department of Chemistry, University of Copenhagen, Universitetsparken 5, 2100 Copenhagen, Denmark ‡ Niels Bohr Institute, University of Copenhagen, Blegdamsvej 17, 2100 Copenhagen, Denmark § Materials Research Division, Risø National Laboratory for Sustainable Energy, Technical University of Denmark (DTU), Frederiksborgvej 399, Post Office Box 49, DK-4000 Roskilde, Denmark ^ Fuel Cell and Solid State Chemistry Division, Risø National Laboratory for Sustainable Energy, Technical University of Denmark (DTU), Frederiksborgvej 399, Post Office Box 49, DK-4000 Roskilde, Denmark ABSTRACT: To produce biominerals, such as shells, bones, and teeth, living beings create organic compounds that control the growth of the solid phase. Investigating the atomic scale behavior of individual functional groups at the mineralfluid interface provides fundamental information that is useful for constructing accurate predictive models for natural systems. Previous investigations of the activity of coccolith-associated polysaccharides (CAP) on calcite, using atomic force microscopy (AFM) [Henriksen, K.; Young, J. R.; Bown, P. R.; Stipp, S. L. S. Palentology 2004, 43 (Part 3), 725743] and molecular dynamics (MD) modeling [Yang, M.; Stipp, S. L. S.; Harding, J. H. Cryst. Growth Des. 2008, 8 (11), 40664074], have suggested that OH functional groups control polysaccharide attachment. The purpose of this work was to characterize, using X-ray reflectivity (XR) combined with molecular dynamics (MD) simulations, the structuring on calcite of a layer of the simplest carbon chain molecule that contains an OH group, ethanol (CH3CH2OH). We found evidence that EtOH forms a highly ordered structure at the calcite surface, where the first layer molecules bond with calcite. The ethanol molecules stand up perpendicularly at the interface or nearly so. As a consequence, the fatty, CH3 ends form a new surface, about 6 Å from the termination of the bulk calcite, and beyond that, there is a thin gap where ethanol density is low. Following is a more disordered layer that is two to three ethanol molecules thick, about 14 Å, where density more resembles that of bulk liquid ethanol. The good agreement between theory and experiment gives confidence that a theoretical approach can offer information about behavior in more complex systems.
’ INTRODUCTION Organisms produce biominerals by creating organic compounds that control the growth of a solid phase, such as for shells, bones, and teeth. Some of these organic molecules are extremely large and complex; therefore, investigating how the individual functional groups interact with mineral surfaces to control growth processes will provide a base for studying more complex systems. Defining the properties of mineralfluid interfaces at the fundamental level is a necessary first step for describing natural systems with models; therefore, we will be able to predict long-distance or long-time outcomes, and it is also the key for discovering the mysteries in atomic scale systems. For example, calciteorganic compound interactions play a role in reducing atmospheric CO2 through carbon mineralization, a global-scale geochemical process, and defining the atomic-scale mechanisms that control biomineralization will help make dreams of producing biomimetic compounds come closer to reality. Calcite, CaCO3, is a particularly interesting mineral. Through the activity of complex organic molecules, it forms the shells of oysters, deposits scale on our teeth and tea pots, as well as r 2011 American Chemical Society
choking water pipes, and it is the main component of chalk, limestone, and marble. Chalk can be composed of 95% or more biogenic calcite. Vast beds of chalk form the aquifers that supply water to southern England, northern France, southern Sweden, and most of Denmark, and it forms oil and gas reservoirs in the North Sea, Texas, and many other locations worldwide. Chalk is composed of the fossil remains of some species of algae called the coccolithophorids. These planktonic algae, many species of which still thrive in modern oceans, cover their one cell with an interlocking set of coccoliths, tiny discs made of 2060 individual calcite crystals that are less than 1 μm in length. Proteins, which are used by molluscs to control shell growth,3,4 have not been implicated in coccolith production. For algae, each species engineers its own intricate coccolith design through the activity of complex polysaccharides.5 Received: June 9, 2011 Revised: October 11, 2011 Published: November 07, 2011 2545
dx.doi.org/10.1021/la2021758 | Langmuir 2012, 28, 2545–2550
Langmuir Previous investigations of the activity of coccolith-associated polysaccharides (CAP) on calcite using atomic force microscopy (AFM)1 and molecular dynamics (MD) modeling2 have established that the OH groups on these complex sugars play an important role in their attachment to calcite. The simplest OH compound is water, HOH; therefore, understanding its bonding at calcite surfaces provided some insight. Several studies68 have established that water is highly ordered in the first layers that are in contact with calcite. Simulations and reflectivity give the same qualitative result (i.e., ordered layers), but they show a 15% discrepancy in the position of the first water layer. Improving upon this, some recent simulations by Raiteri et al.9 achieved more accurate results, with a surface water layer separation 0.3 Å smaller than the experimental results. The simplest organic molecule that has a fatty, CH3 end, and a polar, OH end, is ethanol, CH3CH2OH or EtOH. Cooke et al.10 and Sand et al.11,12 have recently studied the behavior of ethanolwater solutions on calcite using AFM and MD. Also recently, the importance of the position of the OH group of hydroxyproline in collagen has been discussed in the context of biomineralization.13 It makes sense to extend the ethanol calcite work using X-ray reflectivity (XR) because of the unparalleled capacity of this technique for characterizing layered, flat structures down to the Ångstrom scale. Calcite is a rhombohedral mineral with three symmetrically identical, perfect cleavage directions. With careful cleavage, surfaces can be produced that are atomically flat over several micrometers; therefore, surface roughness is far below the level needed for resolving monolayers of atoms or low-molecular-weight molecules, such as water and ethanol. These physical properties make calcite very well suited for XR studies of adsorbed layers.8 Previous investigations of the calcite surface with XR have established without a doubt that the first layer of water in contact with calcite is structured by the atomic arrangement of the solid surface.14 Chiarello et al.15 showed that the water layer adsorbed on calcite from the humidity in air was 19 Å, and Bohr et al.8 demonstrated that the water layer thickness, which they found to be about 15 Å, is constant, regardless of the atmospheric humidity, indicating that water thickness depends upon the properties of the solid surface and not the partial pressure of water vapor. Bohr and colleagues also confirmed structuring in the first water layer in contact with calcite and showed that there is a gap between this first layer and the rest of the water, just as one sees empty space between the layers of atoms in a crystal. Chattopadhyay et al. also reported the presence of such a gap on water interacting with very hydrophobic surfaces.16 Thus far, no XR studies of organic compounds with calcite have been reported, but some recent work using grazing incidence X-ray diffraction (GIXRD) analyzed the adsorption of glycine and water on the surface of calcite or other usual biominerals, such as fluorapatite. They confirmed the presence of ordering in the organic liquid interacting with the mineral.17,18 The purpose of this work was to characterize the structure of an ethanol layer in contact with calcite using XR and to test the conceptual model resulting from the data using MD simulations. The coupling of experimental and theoretical approaches gives an excellent means to check the consistency of the experimental interpretations. It is also the start in the development of a predictive tool to screen complicated systems quickly, allowing one to focus on promising systems for further experimental study.
ARTICLE
’ EXPERIMENTAL SECTION Experimental Study. We used single crystals of optical-quality Iceland spar calcite (from Chihuahua, Mexico, purchased from Ward’s Scientific, Rochester, NY). XR requires flat surfaces that are at least 5 mm on a side; therefore, we cleaved broad thin slices using the method described by Stipp and Hochella.6 Briefly, we scratched along the cleavage direction with a scalpel until a fracture propagated through the crystal and the sample fell away. The samples were about 10 mm long, about 8 mm wide, and about 1.5 mm thick. Such surfaces are atomically flat over several micrometers. The steps between terraces, however, do produce a shadow for the incident or reflected beam; therefore, a correction for the loss of intensity must be made during data analysis. The samples were cleaved under ethanol to avoid adsorption of water vapor from the air, which affects the morphology and structure of the calcite surface.11 We used clean tools and avoided touching the surface of the sample. The single crystal was mounted in the sample holder with clean, stainless-steel tweezers that only touched the sides of the crystal at two edges and not the analyzed surface. The samples were kept immersed under liquid ethanol in covered vessels until used to minimize water adsorption from air humidity and the accumulation of adventitious carbon, the hydrocarbon contamination that comes from the air, or solutions in contact with the sample, to which calcite is particularly attractive.6 The ethanol had been distilled to remove the impurities normally present,19 for example, hydrocarbons from storage in plastic bottles and zinc from the process to remove water from commercially produced ethanol. After storage for 6 h in distilled ethanol, we mounted each sample in the analytical chamber, which is a kapton cylinder supported on an aluminum frame. The chamber is airtight, allowing for control of the atmosphere. To minimize contamination, beam damage, and the loss of X-ray intensity by air scattering, the chamber was filled with helium at 1 atm pressure. We used a flow-through system, where dry helium bubbled through distilled ethanol and then into the chamber to maintain ethanol saturation. XR has become a standard method for studying the molecular structure on interfaces.14,20,21 The reflected intensity of an X-ray beam striking the sample at a low angle provides information about the electron density of the interacting layers. The incident angle ranges from below the critical angle of the material to where the reflected intensity fades to the background or the first Bragg diffraction peak appears. The intensity of the reflected beam depends upon the structure of the surface. For cleaved calcite, a scaling parameter was required to account for the X-ray beam shadow cast by cleavage steps. These steps generally follow a zigzag pattern; therefore, rotating or changing the sample alignment does not solve the problem. The scaling parameter varied in magnitude depending upon the sample used, namely, how many and how large the cleavage steps were, and their orientation, i.e., the degree to which the steps faced or screened the X-ray beam. Measured intensity also depends upon the extent of internal reflections in the various layers of solid, liquid, and gas that form the interface, and these are characterized by layer thickness (Z), mass density (F), and the roughness of the interfaces between the layers (σ).15,2224 We performed the experiments at beamline BW2 in the synchrotron radiation facility at DESY, in Hamburg, Germany.2527 X-ray energy was 10 keV (λ = 1.24 Å), and the slits that defined the beam size had an aperture of 2 mm for the horizontal direction and 0.2 mm for the vertical. For the data analysis, we used the implementation for MatLab R2009b by Whiting et al.28 of the Parratt method, as outlined by Zhou and Chen.29 A five-box model provided the best fit. Two boxes corresponded to the calcite substrate and the helium atmosphere. The other three modeled the ethanol. To evaluate the goodness of fit, the software implements a R value calculated from the residuals.30 2546
dx.doi.org/10.1021/la2021758 |Langmuir 2012, 28, 2545–2550
Langmuir
Figure 1. Example of XR data (black dots) and the best fit (red line). The fit provides the parameters to build a density profile of the calciteethanol interface (Figure 2), and Table 1 summarizes the physical properties derived from the fit.
Theoretical Study. In parallel with the experiments, we performed classical MD simulations of the calciteethanol interface to gain a better understanding of the system details. First, we built the calcite {1014} surface with Materials Studio 4.231 and the ethanol molecule with AMBER 9.0.32 Given the arrangement of atoms in a calcite crystal, we built a surface with an area of 32.38 24.95 Å2 instead of a more conventional squared area and covered it with 150 ethanol molecules. We input these elements into the simulation software, DL_POLY 2.20,33 with a set of potentials designed for use at bioinorganic interfaces34 and assigned periodic boundary conditions. First, we built a box of ethanol. The box had a size of 32 24.5 25 Å3 to match the dimensions of the calcite surface, with its height, 25 Å, being adjustable depending upon the number of molecules, in this case, 150 ethanol molecules. A Nose-Hoover NVT thermostat with a 0.1 ps relaxation time was applied to maintain a constant temperature. The temperature of the ethanol box was increased stepwise from 0 to 300 K with a step size of 50 K, and a MD simulation of 250 ps was carried out for each temperature. Second, we put the final configuration of the ethanol box after the previous MD simulations to the calcite surface that we built, and the initial distance between the solvent box and the surface was set at 2.0 Å. Finally, we applied the same stepwise method to increase the temperature of this system from 0 to 300 K, and the final configuration was used as the starting configuration for a long-time MD simulation. If the simulation of the combined system had started directly from 300 K, it may have unrealistic and strong interactions between ethanol molecules and the calcite surface, which would cause an abrupt temperature increase that would destroy the system. Actually, the calculations ran for a long enough time to eliminate the influence of the initial specific configuration. The total time for the MD simulation was 8 ns, with a time step of 1 fs. For data analysis, we used the molecular trajectories from the last 4 ns; the first 4 ns allowed the system to equilibrate.
’ RESULTS Observations. Figure 1 shows an example of the data from one XR experiment (black dots). The red line represents the best fit that could be achieved using a five-box model, where each box or layer is represented by three parameters, thickness (Z), mass density (F), and roughness (σ). The parameters derived from the model optimization to fit the data are listed in Table 1, and a density profile is presented as Figure 2. From the best fit, the total thickness of the ethanol layer adsorbed on freshly cleaved calcite is 20.3 Å. Immediately bonded at the termination of the bulk crystal is a structured
ARTICLE
Figure 2. Comparison between the density profile derived from the XR data (red) and that obtained from a statistical average of 4000 MD simulation snapshots (green). The values from the experimental profile (red) are presented in Table 1. The simulated profile (green) was derived from MD calculations (Figure 5). The agreement is good, matching within uncertainty the densities and location of the lowdensity layer, the gap (EtOH-2), which lies between the highly ordered first ethanol layer (EtOH-1) and the rest of the ethanol, which has more bulk-like properties (EtOH-3).
Table 1. Summary of the Physical Parameters for the Density Profile in Figure 2a Z (Å)
F (g cm3)
σ
2.710
0.7 ( 0.1
EtOH-1
5.4 ( 0.3
0.8 ( 0.1
0.5 ( 0.1
EtOH-2
0.9 ( 0.2
0.4 ( 0.1
0.2 ( 0.1
EtOH-3
14.0 ( 0.5
0.8 ( 0.1
1.9 ( 0.1
calcite
He
0.010
For each layer, Z represents thickness, F represents mass density, and σ represents roughness of the interface between the layer and the next layer; the smaller the roughness, the sharper the interface. The first layer represents the bulk substrate, calcite, and the fifth layer represents helium, the atmosphere in contact with the sample. EtOH-2 has parameters that fit a layer with very low density and represents the space between ethanol structured on the calcite template, EtOH-1, and the rest of the ethanol, EtOH-3. This is clear in the cross-section (Figure 5). The density of the calcite and the He layers are most precise because we used known values for bulk material; therefore, we fixed them as calibration parameters. It is possible that the calcite surface restructuring causes the calcite layer in contact with ethanol to be slightly different from bulk calcite, but this difference is likely to be small. The densities derived from the model for the two ethanol layers (EtOH-1 and EtOH-3) are the same, within uncertainty, as the density of bulk ethanol, 0.789 g cm3.35 a
layer of ethanol. This first layer, which we call EtOH-1, is 5.4 Å thick. This thickness compares well to the length of an ethanol molecule, which is about 6 Å,33 suggesting that the ethanol attaches as a brush protuding out from the surface. Ethanol, with its fatty CH3 end and its polar OH end, is forced by chemical affinity to orient with the OH end toward the ionic calcite surface. Thus, we can interpret that EtOH adsorbs to calcite through hydrogen bonding to an O from a CO3. The fatty ends of the molecule are oriented away from the surface and create a hydrophobic layer. The XR data fitting describes this layer with a density of about 0.8 g cm3. This is, within the uncertainty, in agreement with the density of liquid ethanol (0.789 g cm3 35). 2547
dx.doi.org/10.1021/la2021758 |Langmuir 2012, 28, 2545–2550
Langmuir
ARTICLE
Figure 3. Zoom of the data (black dots) for high values of q provides an opportunity to examine the sensitivity of the fit to the conceptual model used. If the first adsorbed molecular layer of EtOH is highly ordered and the next layer is somewhat ordered (red line), a plane must exist where atomic density is low, just as it exists between the atomic planes for a crystalline substance (Figure 5). If EtOH was not highly ordered at the surface, then the data would follow the blue line and the density profile would be close to flat (Figure 4). The lack of a gap between the first and second adsorbed layers would reflect a lack of ordering.
Figure 4. Density profile for a model where ethanol is not highly ordered; therefore, there is no gap between the first and subsequent molecular layers, such as the case represented by the blue curve in Figure 3. The physical parameters for this model are listed in Table 3.
Table 3. Summary of the Physical Parameters for a Model Where There Is No Structuring in the First or Subsequent Ethanol Layers, Thus No Gap (EtOH-20 )a
Table 2. Summary of the Physical Values from the Density Profile Resulting from the MD Simulation (Figure 2)a Z (Å) calcite
EtOH-20
2.71
0.7
EtOH-30 He
4.5
0.84
0.4
2.0
0.37
0.4
EtOH-3
EtOH-10
σ
EtOH-2
F (g cm3)
σ
2.710
1.2 ( 0.1
7.1 ( 0.3
0.6 ( 0.1
0.1 ( 0.1
13.1 ( 0.5
0.8 ( 0.1
1.7 ( 0.1
calcite
F (g cm3)
EtOH-1
Z (Å)
a
0.010
The density profile of Figure 4 represents this model.
0.79
a
The parameters compare very well to those for the XR results (Table 1), with Z representing layer thickness, F representing mass density, and σ representing the roughness of the interface between the layer and the next layer. The system reaches equilibrium after the initial 4 ns simulation, and the fluctuation in the density profile at different time intervals, which is the only kind of error that can be evaluated by statistical analysis, is insignificant, thus a lack of uncertainties associated with the data.
Once the first molecular layer of EtOH is established, the next is forced to organize on the hydrophobic terminations, through either its polar or fatty end. The fatty end is preferable, but there is less overall drive than for hydrogen bonding to the ionic surface. Thus, the second and subsequent layers become less and less ordered. For crystals, layers of atoms are highly ordered and the steric and charge constraints produce space between the atomic planes. For the first and second molecular layers of ethanol, which show better ordering than the rest, one could expect there to be space between the fatty ends of the two layers, just as there is space between the layers of atoms in a crystal. The fitting model predicts a layer (EtOH-2) that is 0.9 Å thick, where density is 0.4 g cm3, significantly less than that of the first structured ethanol layer (EtOH-1). It manifests itself as a gap in the density versus the distance profile (Figure 2). The disruption in ordering in the second ethanol layer allows for more freedom for the orientation of subsequent molecular layers; therefore, after several ethanol equivalent lengths, one would expect the ordering to degenerate into random orientation. The fitting model predicts that this less structured layer (EtOH-3) is 14.0 Å thick and has a density of 0.8 g cm3, which is consistent with the density of liquid ethanol.
The ethanol layers were established in a helium atmosphere at room temperature and atmospheric pressure, where ethanol vapor was at saturation. The total thickness of the ethanol layer was the same for all of the experiments. We did not change the partial pressure of the ethanol in the chamber; therefore, we cannot state that it is constant for other ethanol partial pressures, as could be stated for the thickness of a water layer for various atmospheric humidities.8 The thickness of the gap (EtOH-2) between the first structured layer (EtOH-1) and the subsequent, more bulk-like layer of ethanol (EtOH-3) lies close to the limit of experimental resolution defined by the data quality, signal-to-noise ratio, and the measured q range (Table 1). To verify that it was a physical feature and not a fitting artifact, we tested several models, some of which assumed that ethanol was not ordered in the first and second layers; thus, there would be no gap. Figure 3 compares the fit from the best of those tests (blue line) to the original data and the fit from the model that includes the gap (Figure 1) for the part of the curve that represents the first adsorbed layer, which is the last bump in both figures. Table 3 lists the parameters resulting from the fit without a gap, and its corresponding density profile is presented as Figure 4. The fit without ethanol ordering is significantly worse. The R factor, indicating the goodness of fit, was 49% worse than that for the original model. Also, the parameters describing the first structured ethanol layer suggest it to be 7.1 Å thick with a density of 0.6 g cm3. Although one could imagine a layer where the ethanol molecules were not oriented normal to the surface and they might organize themselves with more space between them than in bulk ethanol, 2548
dx.doi.org/10.1021/la2021758 |Langmuir 2012, 28, 2545–2550
Langmuir
Figure 5. Snapshot time step from the MD simulations and density profile derived from them. The bulk calcite crystal is composed of atomic layers of Ca (green) that are octahedrally coordinated with O (red) from six CO3. The trigonal planar carbonate is composed of three O surrounding each C (blue). Ethanol, CH3CH2OH, orients with the OH bonded through H (gray) to the O of CO3. The first layer of ethanol is highly ordered, and the second is somewhat ordered, which shows more clearly in the density profile. There is a space between the two ordered layers, as there is between the planes of atoms in the bulk of a crystal, such as calcite.
the concepts of chemical behavior for this model are not as consistent with ethanol properties as the first model. Simulations. Figure 5 shows a snapshot of a time step in one of the MD simulations together with its density profile. The bulk calcite crystal is composed of atomic layers of Ca (green) that are octahedrally coordinated with O (red) from six CO3. The trigonal planar carbonate is composed of three O surrounding each C (blue). In the bulk calcite structure, the planar CO3 ions are tipped relative to the atomic planes but the surface atomic arrangement is restructured, partly because it is the termination of the solid and partly because of bonding with EtOH. Ethanol, CH3CH2OH, orients with its OH bonded through H (gray) to the O of surface CO3. The first ethanol layer is highly ordered by its interaction with calcite. The cross-section in Figure 5 shows clear ordering in the first layer, and the second layer is somewhat ordered. There is a gap between these first and second layers, as we see between the atomic planes in the bulk calcite crystal below. The ethanol molecules are ordered normal to the calcite surface, with OH interacting with the surface and fatty ends of the first and second layers in close proximity. All of these features derived from molecular modeling are consistent with the interpretation of the XR results. The green line in Figure 2 represents the density profile derived from the statistical average of 4000 snapshots of the MD simulations. The simulated density profile stops at 15 Å because ethanol does not show any structure beyond that point, which means the influence by the calcite surface is negligible, and, also, because the simulations were made in vacuum; therefore, beyond 15 Å, they are not comparable to the ethanolhelium interface present in the reflectivity experiments. Nevertheless, in the range were they are comparable, the simulation agrees well with the best fit model of the experimental data (red line). The gap in the simulated density profile appears closer to the calcite surface and slightly thicker. These differences as well as the minor variations between simulated and measured densities are within the uncertainties of the experimental and theoretical approaches (Tables 1 and 2).
’ CONCLUSION AND IMPLICATION The experimental observations and the theoretical simulations reveal a consistent picture of the structure for a calciteethanol
ARTICLE
interface where water is absent. Both show that ethanol forms a highly ordered layer directly adsorbed to the surface by hydrogen bonding from EtOH to O of CO3 on the calcite surface, where the fatty ends of the ethanol point perpendicularly away from the termination of the bulk solid. The structured ethanol layer is 5.4 Å thick, about the length of an ethanol molecule, and this first layer is separated from a less ordered, more liquid-like layer that is about 14.0 Å, by a low-density layer, producing a gap of 0.9 Å. The total ethanol thickness in contact with ethanol-saturated He at room temperature and pressure is 20.3 Å. We did not change the partial pressure of ethanol; therefore, we cannot confirm that this thickness is independent of the concentration of ethanol vapor in the gas. However, the constant thickness of water on calcite, regardless of relative humidity, could suggest that the ethanol layer would also have constant thickness independent of the ethanol vapor pressure, given that the mineralfluid attachment is in both cases through hydrogen bonds and ethanol binds to calcite more tightly than water. Nevertheless, this is just a hypothesis that should be tested by further experiments. There is good agreement between theory and experiment, especially taking into account that the calcite used was not a perfect, pure, synthetic crystal but one of natural origin that contained a significant number of step edges, imperfections, and impurities. The small differences between experiments and modeling results, for which we used atomically perfect surfaces, give confidence that a theoretical approach could provide useful information about more complex systems where experimentation would be difficult. Knowing how the simplest organic chain molecule attaches to calcite, creating a protective layer as shown by Cooke et al.10 and Sand et al.,12 provides a fundamental knowledge base, which helps with the interpretation of the behavior of more complex systems and contributes to improve the understanding of, for example, how polysaccharides control the elegant design of coccoliths, bringing us closer to producing biomimetic materials. Another use of better understanding into the controls on the interaction between organic compounds and calcite is for developing non-surfactant surface-tension modifiers for a range of applications. The binding of various functional groups on calcite offers insight into what compounds might be feasible for developing more effective and environmentally friendly methods to enhance oil recovery (EOR) from depleted chalk reservoirs, for example.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT We thank Finn Engstrøm, Henning Osholm Sørensen, Klaus Bechgaard, and Keld West for challenging questions and discussion, the NanoGeoScience group members for their cheerful willingness to help, and the staff from the DESY beamline and Karina Sand for discussion. I. S. Pasarín thanks Christian C. Rein for his help and insight during preliminary experiments and is grateful for the fellowship from the European Commission Marie Curie Early Stage Training Network called MIR (EC-MC-ESTMineral Interface Reactivity). Research funding was provided by Maersk Oil and Gas A/S and the Danish National Advanced Technology Foundation through the Nano-Chalk Venture. Support for MD simulations came from MIB, an EPSRC Frame Grant. 2549
dx.doi.org/10.1021/la2021758 |Langmuir 2012, 28, 2545–2550
Langmuir
ARTICLE
’ REFERENCES (1) Henriksen, K.; Young, J. R.; Bown, P. R.; Stipp, S. L. S. Palentology 2004, 43 (Part 3), 725–743. (2) Yang, M.; Stipp, S. L. S.; Harding, J. H. Cryst. Growth Des. 2008, 8 (11), 4066–4074. (3) Addadi, L.; Yellin, Z.; Weissbuch, I.; Lahav, M.; Leiserowitz, L. Mol. Cryst. Liq. Cryst. 1983, 96, 1–17. (4) Addadi, L.; Joester, D.; Nudelman, F.; Weiner, S. Chem.—Eur. J. 2006, 12, 981–987. (5) Marsh, M. E. Protoplasma 1994, 177, 108–122. (6) Stipp, S. L.; Hochella, M. F., Jr. Geochim. Cosmochim. Acta 1991, 55, 1723–1736. (7) Stipp, S. L. S. Geochim. Cosmochim. Acta 1996, 63, 1–8. (8) Bohr, J.; Stipp, S. L. S.; Morris, P. M.; Wogelius, R. A. Geochim. Cosmochim. Acta 2010, 74, 5985–5999. (9) Raiteri, P.; Gale, J. D.; Quigley, D.; Rodger, P. M. J. Phys. Chem. C 2010, 5997–6010. (10) Cooke, D. J.; Gray, R. J.; Sand, K. K.; Stipp, S. L. S.; Elliot, J. Langmuir 2010, 26, 14520–14529. (11) Sand, K. K.; Stipp, S. L. S.; Hassenkam, T.; Yang, M.; Cooke, D.; Makovicky, E. Mineral. Mag. 2008, 72, 353–357. (12) Sand, K. K.; Yang, M.; Makovicky, E.; Cooke, D. J.; Bechgaard, K.; Stipp, S. L. S. Langmuir 2010, 26, 15239–15247. (13) Bohr, J.; Olsen, K. Theor. Chem. Acc. 2010, DOI: 10.1007/ s00214-010-0761-3. (14) Geissb€uhler, P.; Fenter, P.; DiMasi, E.; Srajer, G.; Sorensen, L. B.; Sturchio, N. C. Surf. Sci. 2004, 573, 191–203. (15) Chiarello, R. P.; Wogelius, R. A.; Sturchio, N. C. Geochim. Cosmochim. Acta 1993, 57, 4103–4110. (16) Chattopadhyay, S.; Uysal, A.; Stripe, B.; Ha, Y.; Marks, T. J.; Karapetrova, E. A.; Dutta, P. Phys. Rev. Lett. 2010, 105. (17) Magdans, U.; Torrelles, X.; Angermund, K.; Gies, H.; Rius, J. Langmuir 2007, 23, 4999–5004. (18) Pareek, A.; Torrelles, X.; Angermund, K.; Rius, J.; Magdans, U.; Gies, H. Langmuir 2009, 25, 1453–1458. (19) Lund, H.; Bjerrum, J. Ber. Dtsch. Chem. Ges. 1931, 64 (1), 210–213. (20) Cowley, R. A.; Ryan, T. W. J. Phys. D: Appl. Phys. 1987, 20, 61–68. (21) Fenter, P.; Sturchio, N. C. Prog. Surf. Sci. 2005, 77, 171–258. (22) Parratt, L. G. Phys. Rev. 1954, 95, 359–369. (23) Weber, W.; Lengeler, B. Phys. Rev. B 1992, 46, 7953–7956. (24) Wogelius, R. A.; Farquhar, M. L.; Fraser, D. G.; Tang, C. C. Structural evolution of the mineral surface during dissolution probed with synchrotron X-ray techniques. In Growth, Dissolution and Pattern Formation in Geosystems; Jamtveit, B., Meakin, P.; Kluwer Academic Publishers: Norwell, MA, 1999; pp 269289. (25) Drube, W.; Schulte-Schrepping, H.; Schmidt, H.-G.; Treusch, R.; Materlik, G. Rev. Sci. Inst. 1995, 66, 168. (26) Schulte-Schrepping, H.; Heuer, J.; Hukelmann, B. J. Synchrotron Radiat. 1998, 5, 682. (27) Drube, W.; Grehk, T. M.; Materlik, R. T. G. J. Electron Spectrosc. Relat. Phenom. 1998, 683, 88–91. (28) Whiting, G. L.; Snaith, H. J.; Khodabakhsh, S.; Andrasen, J. W.; Breiby, D. W.; Nielsen, M. M.; Greenham, N. C.; Friend, R. H.; Huck, W. T. S. Nano Lett. 2006, 6, 573–578. (29) Zhou, X. L.; Chen, S. H. Phys. Rep. 1995, 257, 223–348. (30) Farquhar, M. L.; Wogelius, R. A.; Charnock, J. M.; Wincott, P.; Tang, C. C.; Newville, M.; Eng, P. J.; Trainor, T. P. Mineral. Mag. 2003, 67, 1205–1219. (31) Accelrys Software, Inc. MS Materials Visualizer, Release 4.2; Accelrys Software, Inc.: San Diego, CA, 2006. (32) Case, D. A.; et al. AMBER 9; University of California: San Francisco, CA, 2006. (33) Smith, W.; Forester, T. R. J. Mol. Graphics 1996, 14 (3), 136–141. (34) Freeman, C. L.; Harding, J. H.; Cooke, D. J.; Elliott, J.; Lardge, J. S.; Duffy, D. M. J. Phys. Chem. C 2007, 111 (32), 11943–11951. (35) Lide, D. R. Handbook of Chemistry and Physics; CRC Press (Taylor and Francis Group): Boca Raton, FL, 19981999. 2550
dx.doi.org/10.1021/la2021758 |Langmuir 2012, 28, 2545–2550
