Publisher Correction: Interactions of the Calcite {10.4} Surface with Organic Compounds: Structure and Behaviour at Mineral - Organic Interfaces.

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has been fixed in the paper.

Interactions of the Calcite {10.4} Surface with Organic Compounds: Structure and Behaviour at Mineral - Organic Interfaces.

The structure and the strength of organic compound adsorption on mineral surfaces are of interest for a number of industrial and environmental applications, oil recovery, CO2 storage and contamination remediation. Biomineralised calcite plays an essential role in the function of many organisms that control crystal growth with organic macromolecules. Carbonate rocks, composed almost exclusively of calcite, host drinking water aquifers and oil reservoirs. In this study, we examined the ordering behaviour of several organic compounds and the thickness of the adsorbed layers formed on calcite {10.4} surfaces. We used X-ray reflectivity (XRR) to study calcite {10.4} surfaces that were prepared in three alcohols: methanol, isopropanol and pentanol and one carboxylic acid: octanoic acid. All molecules adsorbed in self-assembled layers, where thickness depended on the density and the length of the molecule. For methanol and isopropanol, molecular dynamic simulations (MD) provided complementary information, which allowed us to develop a surface model. Branching in isopropanol induced slightly less ordering because of the additional degree of freedom. Pentanol and octanoic acid adsorbed as single monolayers. The results of this work indicate that adhered organic compounds from the surrounding environment can affect the surface behaviour, depending on properties of the organic compound.

Adsorption of ethanol and water on calcite: dependence on surface geometry and effect on surface behavior.

Molecular dynamics (MD) simulations were used to explore adsorption on calcite, from a 1:1 mixture of ethanol and water, on planar {10.4} and stepped, i.e. vicinal, surfaces. Varying the surface geometry resulted in different adsorption patterns, which would directly influence the ability of ethanol to control calcite crystal growth, dissolution, and adsorption/desorption of other ions and molecules. Ethanol forms a well-ordered adsorbed layer on planar faces and on larger terraces, such as between steps and defects, providing little chance for water, with its weaker attachment, to displace it. However, on surfaces with steps, adsorption affinity depends on the length of the terraces between the steps. Long terraces allow ethanol to form a well-ordered, hydrophobic layer, but when step density is high, ethanol adsorption is less ordered, allowing water to associate at and near the steps and even displacing pre-existing ethanol. Water adsorbed at steps forms mass transport pathways between the bulk solution and the solid surface. Our simulations confirm the growth inhibiting properties of ethanol, also explaining how certain crystal faces are more stabilized because of their surface geometry. The -O(H) functional group on ethanol forms tight bonds with calcite; the nonpolar, -CH3 ends, which point away from the surface, create a hydrophobic layer that changes surface charge, thus wettability, and partly protects calcite from precipitation and dissolution. These tricks could easily be adopted by biomineralizing organisms, allowing them to turn on and off crystal growth. They undoubtedly also play a role in the wetting properties of mineral surfaces in commercial CaCO3 manufacture, oil production, and contamination remediation.

Predicting the pKa and stability of organic acids and bases at an oil-water interface.

We have used density functional theory and the implicit solvent model, COSMO-RS, to investigate how the acidity constant, pKa, of organic acids and bases adsorbed at the organic compound-aqueous solution interface changes, compared to its value in the aqueous phase. The pKa determine the surface charge density of the molecules that accumulate at the fluid-fluid interface. We have estimated the pKa by comparing the stability of the protonated and unprotonated forms of a series of molecules in the bulk aqueous solution and at an interface where parts of each molecule reside in the hydrophobic phase and the rest remains in the hydrophilic phase. We found that the pKa for acids is shifted by ∼1 pH unit to higher values compared to the bulk water pKa, whereas they are shifted to lower values by a similar amount for bases. Because this pKa shift is similar in magnitude for each of the molecules studied, we propose that the pKa for molecules at a water-organic compound interface can easily be predicted by adding a small shift to the aqueous pKa. This shift is general and correlates with the functional group. We also found that the relative composition of molecules at the fluid-fluid interface is not the same as in the bulk. For example, species such as carboxylic acids are enriched at the interface, where they can dominate surface properties, even when they are a modest component in the bulk fluid. For high surface concentrations of carboxylic acid groups at an interface, such as a self-assembled monolayer, we have demonstrated that the pKa depends on the degree of deprotonation through direct hydrogen bonding between protonated and deprotonated acidic headgroups.

Energies of the adsorption of functional groups to calcium carbonate polymorphs: the importance of -OH and -COOH groups.

The adsorption behavior of calcium carbonate is an important factor in many processes in nature, industry, and biological systems. We determined and compared the adsorption energies for a series of small molecules of different sizes and polarities (i.e., water, several alcohols, and acetic acid) on three synthetic CaCO3 polymorphs (calcite, aragonite, and vaterite). We measured isosteric heats of adsorption from vapor adsorption isotherms for 273 < T < 293 K, and we used XRD and SEM to confirm that samples did not change phase during the experiments. Density functional calculations and molecular dynamics simulations complemented the experimental results and aided interpretation. Alcohols with molecular mass greater than that of methanol bind more strongly to the calcium carbonate polymorphs than water and acetic acid. The adsorption energies for the alcohols are typical of chemisorption and indicate alcohol displacement of water from calcium carbonate surfaces. This explains why organisms favor biomolecules that contain alcohol functional groups (-OH) to control which polymorph they use, the crystal face and orientation, and the particle shape and size in biomineralization processes. This new insight is also very useful in understanding organic molecule adsorption mechanisms in soils, sediments, and rocks, which is important for predicting the behavior of mineral-fluid interactions when the challenge is to remediate contaminated groundwater aquifers or to produce oil and gas from reservoirs.

Geometry, reduction potential, and reorganization energy of the binuclear Cu(A) site, studied by density functional theory.

The dimeric Cu(A) site found in cytochrome c oxidase and nitrous oxide reductase has been studied with the density functional B3LYP method. We have optimized the structure of the realistic (Im)(S(CH(3))(2))Cu(SCH(3))(2)Cu(Im)(CH(3)CONHCH(3)) model in the fully reduced, mixed-valence, and fully oxidized states. The optimized structures are very similar to crystal structures of the protein, which shows that the protein does not strain the site significantly. Instead, inorganic model complexes of the protein site are strained by the macrocyclic connections between the ligand models. For the mixed-valence (Cu(I)+Cu(II)) state, two distinct equilibrium structures were found, one with a short Cu-Cu distance, 248 pm, similar to the protein structure, and one with a longer distance, 310 pm, similar to what is found in inorganic models. In the first state, the unpaired electron is delocalized over both copper ions, whereas in the latter, it is more localized to one of the ions. The two states are nearly degenerate. The potential energy surfaces for the Cu-Cu, Cu-S(Met), and Cu-O interactions are extremely flat. In fact, all three distances can be varied between 230 and 310 pm at an expense in energy of less than 8 kJ/mol, which explains the large variation observed in crystal structures for these interactions. Inclusion of solvation effects does not change this significantly. Therefore, we can conclude that a variation in these distances can change the reduction potential of the Cu(A) site by at most 100 mV. The model complex has a reorganization energy of 43 kJ/mol, 20 kJ/mol lower than for a monomeric blue-copper site. This lowering is caused by the delocalization of the unpaired electron in the mixed-valence state.

Inner-sphere reorganization energy of iron-sulfur clusters studied with theoretical methods.

Models of several types of iron-sulfur clusters (e.g., Fe(4)S(4)(SCH(3))(4)(2-/3-/4-)) have been studied with the density functional B3LYP method and medium-sized basis sets. In a vacuum, the inner-sphere reorganization energies are 40, 76, 40, 62, 43, and 42 kJ/mol for the rubredoxin, [2Fe-2S] ferredoxin, Rieske, [4Fe-4S] ferredoxin, high-potential iron protein, and desulfoferrodoxin models, respectively. The first two types of clusters were also studied in the protein, where the reorganization energy was approximately halved. This change is caused by the numerous NH.S(Cys) hydrogen bonds to the negatively charged iron-sulfur cluster, giving rise to a polar local environment. The reorganization energy of the iron-sulfur clusters is low because the iron ions retain the same geometry and coordination number in both oxidation states. Cysteine ligands give approximately the same reorganization energy as imidazole, but they have the advantage of stabilizing a lower coordination number and giving more covalent bonds and therefore more effective electron-transfer paths.

On the role of strain in blue copper proteins.

Theoretical investigations of the structure and function of the blue copper proteins are described. We have studied the optimum vacuum geometry of oxidised and reduced copper sites, the relative stability of trigonal and tetragonal Cu(II) structures, the relation between the structure and electronic spectra, the reorganisation energy, and reduction potentials. Our calculations give no support to the suggestion that strain plays a significant role in the function of these proteins; on the contrary, our results show that the structures encountered in the proteins are close to their optimal vacuum geometries (within 7 kJ/mol). We stress the importance of defining what is meant by strain and of quantifying strain energies or forces in order to make strain hypotheses testable.

The influence of axial ligands on the reduction potential of blue copper proteins.

The reduction potentials of blue copper sites vary between 180 and about 1000 mV. It has been suggested that the reason for this variation is that the proteins constrain the distance between the copper ion and its axial ligands to different values. We have tested this suggestion by performing density functional B3LYP calculations on realistic models of the blue copper proteins, including solvent effects by the polarizable continuum method. Constraining the Cu-S(Met) bond length to values between 245 and 310 pm (the range encountered in crystal structures) change the reduction potential by less than 70 mV. Similarly, we have studied five typical blue copper proteins spanning the whole range of reduction potentials: stellacyanin, plastocyanin, azurin, rusticyanin, and ceruloplasmin. These studies included the methionine (or glutamine) ligand as well as the back-bone carbonyl oxygen group that is a ligand in azurin and is found at larger distances in the other proteins. The active-site models of these proteins show a variation in the reduction potential of about 140 mV, i.e., only a minor part of the range observed experimentally (800 mV). Consequently, we can conclude that the axial ligands have a small influence on the reduction potentials of the blue copper proteins. Instead, the large variation in the reduction potentials seems to arise mainly from variations in the solvent accessibility of the copper site and in the orientation of protein dipoles around the copper site.

Quantum chemical calculations of the reorganization energy of blue-copper proteins.

The inner-sphere reorganization energy for several copper complexes related to the active site in blue-copper protein has been calculated with the density functional B3LYP method. The best model of the blue-copper proteins, Cu(Im)2(SCH3)(S(CH3)2)(0/+), has a self-exchange inner-sphere reorganization energy of 62 kJ/mol, which is at least 120 kJ/mol lower than for Cu(H2O)4(+/2+). This lowering of the reorganization energy is caused by the soft ligands in the blue-copper site, especially the cysteine thiolate and the methionine thioether groups. Soft ligands both make the potential surfaces of the complexes flatter and give rise to oxidized structures that are quite close to a tetrahedron (rather than tetragonal). Approximately half of the reorganization energy originates from changes in the copper-ligand bond lengths and half of this contribution comes from the Cu-S(Cys) bond. A tetragonal site, which is present in the rhombic type 1 blue-copper proteins, has a slightly higher (16 kJ/mol) inner-sphere reorganization energy than a trigonal site, present in the axial type 1 copper proteins. A site with the methionine ligand replaced by an amide group, as in stellacyanin, has an even higher reorganization energy, about 90 kJ/mol.

The cupric geometry of blue copper proteins is not strained.

The geometry of several realistic models of the metal coordination sphere in the blue copper proteins has been optimised using high-level quantum chemical methods. The results show that the optimal vacuum structure of the Cu(II) models is virtually identical to the crystal structure of oxidised blue copper proteins. For the reduced forms, the optimised structure seems to be more tetrahedral than the one found in the proteins, but the energy difference between the two geometries is less than 5 kJ/mol, i.e. within the error limits of the method. Thus, the results raise strong doubts against hypotheses (entatic state and the induced-rack theory) suggesting that blue copper proteins force the oxidised metal coordination sphere into a structure similar to that preferred by Cu(I) in order to minimise the reorganisation energy of the electron transfer reaction. Instead, a small reorganisation energy seems to be reached by an appropriate choice of metal ligands. In particular, the cysteine thiolate ligand appears to be crucial, changing the preferred geometry of the oxidised complexes from square-planar to a more trigonal geometry.