A computational chemical study of penetration and displacement of water films near mineral surfaces
© The Royal Society of Chemistry and the Division of Geochemistry of the American Chemical Society 2001
Received: 11 June 2001
Accepted: 07 August 2001
Published: 28 August 2001
A series of molecular dynamics simulations have been performed on organic–water mixtures near mineral surfaces. These simulations show that, in contrast to apolar compounds, small polar organic compounds such as phenols can penetrate through thin water films to adsorb on these mineral surfaces. Furthermore, additional simulations involving demixing of an organic–water mixture near a surfactant-covered mineral surface demonstrate that even low concentrations of adsorbed polar compounds can induce major changes in mineral surface wettability, allowing sorption of apolar molecules. This strongly supports a two-stage adsorption mechanism for organic solutes, involving initial migration of small polar organic molecules to the mineral surface followed by water film displacement due to co-adsorption of the more apolar organic compounds, thus converting an initial water-wet mineral system to an organic-covered surface. This has profound implications for studies of petroleum reservoir diagenesis and wettability changes.
Adsorption of organic solutes on mineral surfaces is an important process in many natural and engineered environments and is particularly important in controlling the distribution of organic compounds in the subsurface. Molecular transformations related to the chemisorption of organic matter can remove organic contaminants from soils and sediments[1, 2] and have also been associated with diagenetic alterations in biomarker compounds. Physisorption of organic molecules on mineral surfaces controls the wettability of permeable rocks and thus capillary pressure terms in multiphase fluid flow and has been proposed as a factor controlling the composition of migrating petroleum. [5–7] Furthermore, oil wetting of mineral surfaces has been proposed to retard or stop diagenesis and cementation of oil reservoirs. In this paper we report on investigations using molecular dynamics calculations[9, 10] (MD) on a possible chain of events that could change an initially water-wet mineral surface to an oil (or more apolar)-wet surface. We hypothesise that this chain of events involves a penetration of the water film covering the mineral surface by small relatively hydrophilic polar organic molecules which physisorb to reduce mineral surface polarity, followed by an adsorption of apolar organic molecules on the thus preconditioned surface. To evaluate this two-stage adsorption mechanism a series of simulations have been performed studying the stability of water films adsorbed on mineral surfaces as a function of the polarity of an adjacent organic phase. For the organic phase a set of compounds ranging from cyclohexane (highly apolar), carbazole (apolar), phenol (polar) to acetic acid (highly polar) was used. For each organic phase two simulations were performed, one starting with a randomly distributed organic–water mixture and the other with completely demixed water and organic phases, with the water film wetting the mineral surfaces.
The inorganic component of the simulations constituted of a (1,-1,1) calcite and a (1,1,1) α-quartz surface. Both calcite and quartz are important contributors to the sediment mineral matrix and authigenic cements. To facilitate comparison of mineral surface effects on organic–water phase composition a completely hydroxylated α-quartz surface was used in the simulations, instead of replacing part of these hydroxyl groups with oxygen bridges (which is more realistic at neutral pHs). Thus, both the calcite and the α-quartz surface presented a homogeneous surface to the organic/water phases.
Finally, to assess the surfactant behaviour of small polar molecules, a set of simulations was performed in which the calcite surface was selectively and partially covered with phenol molecules. This system was covered with a water film and brought into contact with an apolar cyclohexane phase. Sorption of small polar molecules such as phenols to mineral surfaces has been shown to occur during both simulated and natural petroleum migration[11, 12] through initially water-wet rocks. By comparing the results from these simulations with those from the cyclohexane–water–mineral surface simulations the influence of pre-adsorption of small polar organic molecules on the final wettability of mineral surfaces to apolar components could be monitored.
EEM parameters as derived to reproduce peptide MKS charge distributions
To allow a proper description of Coulomb interactions between charges qi and qj on neighbouring atoms in the inorganic phases a shielding-term σij was incorporated, adjusting the Coulomb interaction for electron orbital overlap as described in eqn. (1), where qi and qj are the atomic charges (in atomic units), e is the elementary charge, 4πεo is the vacuum permittivity and rij is the interatomic distance (in Å).
Charge distributions and shielding parameters σi used for the mineral phases
Charge/atomic units a
For the α-quartz phase the parameters derived by Demiralp et al. were used, expanded with parameters to describe the surface hydroxyl groups. For the calcite phase force field parameters were determined using the Morse functional form in eqn. (2) as described by Demiralp et al. This Morse-potential was used to describe both the inter-and intramolecular van der Waals interactions as well as the covalent Si-O, C-O and O-H bonds in the mineral phases.
Morse potential parameters used for the mineral phases a
CH2 united atom b
The simulations were performed using the Delphi program. All simulations were performed at a simulated temperature of 298 K using a time-step of 2 fs. The system temperature was held constant using the algorithm described by Berendsen et al. using a temperature damping constant of 2000 fs. Bonds and bond angles were held stationary during the simulation using the approach described by Andersen. Apart from the surface hydroxyl groups, mineral phase atom positions were kept fixed during the simulations.
A cutoff radius of 8.5 Å was used in all simulations. The system energy was corrected for van der Waals interactions beyond this range. To ensure a stable simulation of the Coulomb interactions a 7th order taper function was applied as described by de Vos Burchart.
System compositions used in the MD simulations
# organic molecules
# water molecules
Results and discussion
Demixing simulations on water-wet mineral surfaces
Due to the small timescales used in the MD simulations the steady-state configurations obtained from the randomly orientated systems may reflect metastable intermediate configurations rather than the thermodynamic end points. To check this, a second set of simulations was performed in which each organic–water system was initialised with completely water-wet mineral surfaces.
Position-dependent diffusion coefficient analysis
The analysis presented in the previous section has primarily been qualitative, based on visual observations regarding the relative positions of the organic and water phases with respect to the mineral surfaces. One of the strong features of MD simulations is that by monitoring the development of the molecular position through time it gives the opportunity for a more quantitative analysis. To thus scrutinise the observations made in the previous section we have calculated the molecular diffusion constants for each molecule in the various organic–water–inorganic systems studies. To this end, atom positions were saved at 1000 iteration intervals (equivalent to a time interval of 1000 × 2 fs = 2 ps). From these trajectory files the molecular diffusion constant dm was calculated from the cartesian coordinates of the molecular centres of mass (xm, ym, zm) using eqn. (3).
Fig. 11 and 12 confirm the observations that, in contrast to the apolar cyclohexane molecules, small polar compounds such as phenol and acetic acid can migrate to the mineral surface, building up significant concentrations of adsorbed organic species. Fig. 11 also demonstrates that this observation can be made regardless of the starting conditions used in the MD simulations, as the randomly distributed phenol–water and acetic acid–water mixtures yield similar equilibrated concentration profiles compared to the water-wet starting configurations. Molecular diffusion of both phenol and acetic acid is significantly reduced near the mineral surfaces, indicating that these molecules indeed have a physical interaction with mineral surface sites.
Fig. 11 shows that raising the simulated temperature from 298 to 373 K affects the positions and diffusion constants of the phenol molecules in the equilibrated system, as an increased phenol concentration near the calcite surface is observed. Furthermore, the phenols associated to the calcite surface are separated from the other phenol molecules in the system by a water layer of almost 10 Å and phenol molecules non directly adsorbed on a mineral surface show a higher diffusion constant. Apart from these differences the results at 298 and 373 K are not dissimilar; at both temperatures the phenol molecules have access to both mineral surfaces and do not form a distinctly separated phase, distinguishing them in these aspects from the apolar cyclohexane molecules (Fig. 10).
Impact of surfactant phenol molecules on cyclohexane-water phase separation
Comparison between calcite and α-quartz surface
In general terms, the calcite and α-quartz surfaces show comparable behaviour with respect to accessibility to organic compounds. Apolar compounds like cyclohexane do not partition to either surface (Fig. 10) while both phenol and acetic acid do (Fig. 11, 12 and 14). On closer analysis we find that the diffusion constants of acetic acid near the calcite surface are significantly lower than those near the α-quartz surface. This effect, if present, is far less distinct for phenol. This indicates that the highly polar acetic acid molecules get tightly bound to the ionic calcite surface while they associate less strongly to the more covalent α-quartz. This could signify that highly polar mineral surfaces could become oil (or organic) wet in contrast to more apolar surfaces due to their high affinity for polar organic surfactants.
The results in this study have potentially profound implications for several processes central to petroleum geology. Clearly rapid sorption of polar surfactants such as phenols can transform the wetting properties of minerals, thus aiding the development of permanently hydrophobic surfaces following direct sorption of more apolar and/or higher molecular weight compounds from apolar phases. This has implications for the wettability of petroleum reservoir rocks, and thus fluid flow, and also for the preservation of water films necessary in the distribution of silica and other inorganic solutes involved in mineral diagenesis.
A series of MD simulations were performed to study the phase behaviour of organic–water mixtures in the presence of α-quartz and calcite surfaces. Upon changing the polarity of the organic phase by going from a charge-neutral cyclohexane via apolar carbazole and polar phenol to highly polar acetic acid distinct changes in phase behaviour are observed. Cyclohexane and carbazole form discrete organic phases removed from the mineral surfaces by a water film; phenol and acetic acid show a sufficient water solubility to penetrate through these films to compete with the water molecules for mineral surface adsorption sites. Potential biases in these observations, due to simulation conditions, were tested by changing the MD-starting organic–water configurations from completely random to completely phase-separated and by changing the system size and simulation temperature, and neither of these were found to have significant effects on the accessibility of the mineral surfaces for the organic compounds.
Further MD simulations with pre-adsorbed phenol compounds on a calcite surface demonstrate that these small polar molecules can have a profound impact on surface wettability, making the mineral surface accessible to even completely non-polar compounds. This provides strong support for a two-stage process incurring wettability changes in mineral systems, commencing with small polar species migrating to the initially water-wet mineral surface, thus changing its surface characteristics, after which the rest of the water phase gets replaced by more apolar compounds adsorbing on and around these polar surfactants.
This work was supported by a Royal Society Research Fellowship and by TMR grant No. ERBFMBICT971871 for ACTvD. We thank the reviewers for their helpful suggestions and remarks.
- Alexander M: Environ Sci Technol. 1995, 29: 2713-10.1021/es00011a003.View ArticleGoogle Scholar
- Kubicki JD, Itoh MJ, Schroeter LM, Apitz SE: Environ Sci Technol. 1997, 31: 1151-10.1021/es960663+.View ArticleGoogle Scholar
- Sieskind O, Joly G, Albrecht P: Geochim Cosmochim Acta. 1979, 43: 1675-10.1016/0016-7037(79)90186-8.View ArticleGoogle Scholar
- Anderson WG: J Pet Technol. 1986, 1125-Google Scholar
- Nagy B: Geochim Cosmochim Acta. 1960, 42: 77-95.Google Scholar
- Krooss BM, Brothers L, Engel MH: Petroleum migration: special publication of the Geological Society. Edited by: England WA, Fleet AJ. 1991, the Geological Society, London, 59: 149-Google Scholar
- Larter SR, Bowler BFJ, Li M, Chen M, Brincat D, Bennett B, Noke K, Donohoe P, Simmons D, Kohnen M, Allan J, Telnaes N, Horstad I: Nature. 1996, 383: 593-10.1038/383593a0.View ArticleGoogle Scholar
- Worden RH, Oxtoby NH, Smalley PC: Pet Geoscl. 1998, 4: 129-View ArticleGoogle Scholar
- Allen MP, Tildesley DJ: Computer Simulation of Liquids. 1987, Clarendon Press, OxfordGoogle Scholar
- Frenkel D, Smit B: Understanding molecular simulation. 1996, Academic Press, San DiegoGoogle Scholar
- Taylor P, Larter SR, Jones DM, Dale J, Horstad I: Geochim Cosmochim Acta. 1997, 61: 1899-10.1016/S0016-7037(97)00034-3.View ArticleGoogle Scholar
- Larter SR, Bowler B, Clarke E, Wilson C, Moffat B, Bennett B, Yardley G, Carruthers D: Geochem Trans. 2000, 9:Google Scholar
- Mortier WJ, Ghosh SK, Shankar S: J Am Chem Soc. 1986, 108: 4315-10.1021/ja00275a013.View ArticleGoogle Scholar
- Njo SL, Fan J, van de Graaf B: J Mol Catal A: Chem. 1998, 134: 79-10.1016/S1381-1169(98)00024-7.View ArticleGoogle Scholar
- Cornell WD, Cieplak P, Bayly C, Gould IR, Merz KM, Ferguson DM, Spellmeyer DC, Fox T, Caldwell JW, Kollman PA: J Am Chem Soc. 1995, 117: 5179-10.1021/ja00124a002.View ArticleGoogle Scholar
- Demiralp E, Çagin T, Goddard WA: Phys Rev Lett. 1999, 82: 1708-10.1103/PhysRevLett.82.1708.View ArticleGoogle Scholar
- Tobias DJ, Klein ML: J Phys Chem. 1996, 100: 6637-10.1021/jp951260j.View ArticleGoogle Scholar
- DeBolt SE, Kollman PA: J Am Chem Soc. 1995, 117: 5316-5340. 10.1021/ja00124a015.View ArticleGoogle Scholar
- Jorgensen WL, Chandrasekhar J, Madura JD: J Chem Phys. 1983, 79: 926-935. 10.1063/1.445869.View ArticleGoogle Scholar
- van de Graaf B, Baas JMA: J Comput Chem. 1984, 5: 314-10.1002/jcc.540050406.View ArticleGoogle Scholar
- Berendsen HJC, Postma JPM, van Gunsteren WF: J Chem Phys. 1984, 81: 3684-10.1063/1.448118.View ArticleGoogle Scholar
- Andersen HC: J Comput Phys. 1983, 52: 24-10.1016/0021-9991(83)90014-1.View ArticleGoogle Scholar
- de Vos Burchart E: 1992, PhD Thesis, Delft University of Technology, the NetherlandsGoogle Scholar
- Bjørlykke K, Ramm M, Saigal GC: Geol Rundsch. 1989, 78: 243-268. 10.1007/BF01988363.View ArticleGoogle Scholar