Environment from the Molecular Level

A NERC eScience testbed project

Science areas in the eMinerals project

1 Adsorption of atoms and molecules

Dissolved metal ions and organic molecules sorb onto mineral surfaces by a variety of mechanisms (e.g., inner-sphere complexes, surface precipitates, ion exchange). These sorption reactions play a fundamental role in the transport of heavy metals and organic pollutants (e.g. PCB’s, dioxins, DDT’s) in soil and groundwater. However, until we understand the mechanisms sorption of metal cation and organic molecules down to the molecular level we cannot develop models of the fate and transport of contaminants in the environment. Several groups in the UK Earth Science community are using X-ray absorption spectroscopy (based on synchrotron radiation) to probe the nature of metal complexes at mineral surfaces. However, spectroscopic data only gives us the local molecular environment (< 5 Å) about a sorbed cation. The energetics of sorption depends on the structure of the electrical double layer at the mineral-water interface. This is a mesoscale problem which requires that we know the structure and dynamics of the mineral water interface over distances of 10–1000 nm. No spectroscopic technique can give us this information. We propose, therefore, to develop the capability for very large (106–107 atoms) simulations of the atomistic structure and dynamics of the mineral water interface over 10 ns. From such computational experiments we could investigate the structure of the electrical double layer and the electrostatic potential at the mineral-water interface for geochemically important systems. The structure and energetics of the mineral-water interface must be know for thermodynamic models of contaminant transport. In practice, simple continuum models (e.g., constant capacitance model) are used by environmental geochemists even though such models have no strong physical foundation. An atomistic picture of the mineral-water interface would be a major advance from these continuum models and may allow more reliable estimates of sorption equilibria under natural conditions.

1. Collins CR, Ragnarsdottir KV, Sherman DM. Geochim Cosmochim Acta, 63, 2989 (1999).
2. Randall SR, Sherman DM, Ragnarsdottir KV, Collins CR. Geochim Cosmochim Acta, 63, 2971 (1999).
3. Collins CR, Sherman DM, Ragnarsdottir KV. J Colloid Interface Science. 219, 345 (1999).
4. MS Craig, MT Dove. Surface Science (submitted).

2 Crystal growth and dissolution

Many minerals, e.g. calcite and baryte, grow from a range of surface defects such as steps and spiral dislocations [1,2]. In order to model growth from such defects under sedimentary conditions (aqueous environment), solvent effects must be included explicitly in the calculations [3]. In addition, we need to take into account the presence of foreign ions in the solution. Impurity ions will compete with the pure material during the crystal growth process [4] and their presence affects the growth rate of the various crystal faces to differing extents, hence altering the morphology of the growing crystal [5].

The presence of organic material has a twofold effect: Firstly, in many biological situations, minerals such as calcium carbonates or hydroxy-apatites in mammalian bones, grow on an organic matrix which directs the growing crystal to a specific morphology [6,7]. Hence, the choice of organic template material affects the final shape and/or phase of the bio-mineral [8]. In experimental studies of bio-mineral nucleation and growth, natural conditions are often mimicked by a careful choice of solution, containing a host of different ions resembling blood plasma, but it is difficult to evaluate separately the effects of the various processes involved. However, a combination of several computational techniques is ideally suited to identify the crucial steps in the mineral growth process. The second role of organic material is as a growth inhibitor [9–11]. Scale is a major industrial problem (for example, in oil and water transportation pipes and boilers). Water softening agents, traditionally used in large quantities to de-scale industrial installations, pose a serious threat to the environment. Hence, there is a need to search for alternative scale inhibitors, which prevent the formation of scale rather than dissolve the crystal once formed. In addition to organic matter, the presence of cations has been shown to have a growth inhibiting effect. Magnesium ions, for example, have been shown to inhibit both calcite and baryte growth [12] through incorporation into the growing crystal [13].

Crystal dissolution and weathering is often found to occur by initial leaching of ions from rock surfaces and cracks, at varying pH and temperatures [14]. The challenge is to model these dissolution processes under realistic circumstances, for example using MD simulations to include temperature in the calculations, while new models need to be developed to accurately model pH in the fluid surrounding the rock faces, which has been shown to affect the rate of dissolution of, for example, olivine minerals [15]. In addition, the presence of solvated ions in the fluid may influence the dissolution rate and the type of ions leaching out first [16].

1. A.J. Gratz, P.E. Hillner, P.K. Hansma, Geochim. Cosmochim. Acta, 57, 491 (1993).
2. P.E. Hillner, S. Manne, P.K. Hansma, Faraday Discuss., 95, 191 (1993).
3. N.H. de Leeuw, S.C. Parker, J.H. Harding, Phys. Rev. B, 60, 13792 (1999).
4. M. Deleuze, S.L. Brantley, Geochim. Cosmochim. Acta, 61, 1475 (1997).
5. S.C. Parker, E.T. Kelsey, P.M. Oliver, J.O. Titiloye, Faraday Discuss., 94, 75 (1993).
6. G. Falini, S. Albeck, S. Weiner, L. Addadi, Science, 271, 67 (1996).
7. D. Walsh, S. Mann, Nature, 377, 320 (1995).
8. S. Mann, B.R. Heywood, S. Rajam, J.D. Birchall, Nature, 334, 692 (1988).
9. D.S. Cicerone, A.E. Regazzoni, M.A. Blesa, J. Colloid Interface Sci., 154, 423 (1992).
10. M.A. Nygren, D.H. Gay, C.R.A. Catlow, M.P. Wilson, A.L. Rohl, J. Chem. Soc. Faraday Trans., 94, 3685 (1998).
11. W.J. Benton, I.R. Collins, I.M. Grimsey, G.M. Parkinson, S.A. Rodger, Faraday Discuss., 95, 281 (1995).
12. R.G. Compton, C.A. Brown, J. Colloid Interface Sci., 165, 445 (1994).
13. N.H. de Leeuw, S.C. Parker, Phys. Chem. Chem. Phys., 3, 3217 (2001).
14. E. Wieland, B. Wehrli, W. Stumm, Geochim. Cosmochim. Acta, 52, 1969 (1988).
15. S.E. Ghazy, M.A. Kabil, A.M. Abeida, N.M. El-Metwally, Sep. Sci. Technol., 31, 829 (1996).
16. U. Becker, P. Risthaus, D. Bosbach, A. Putnis, Mol. Simul., in press (2001).

3 Iron sulphides and the problem of acid mine waters

The Fe-S minerals are widespread in occurrence and are particularly common in sulphide ore deposits. At temperatures above 350° C, pyrite (FeS2) and pyrrhotite (Fe1-xS) are the only stable forms. Below this temperature, the phase relations are far more complex, with pyrite and pyrrhotite phases ranging in composition from FeS to Fe7S8 [1–3]. The pyrrhotite phases pose a major problem when associated with waste generated by mining activity. Bacterially assisted breakdown of pyrrhotites and other Fe sulphides leads to the generation of very acid waters which accelerates dissolution of minerals containing Pb, As, Hg, Cu and other toxic metals. The stability of the different pyrrhotite structures is related to the distribution and ordering of vacancies. Other Fe-sulphides, generated by the action of sulphate reducing bacteria, have been shown to scavenge toxic elements form the environment. Furthermore, this ability appears to be related to the magnetic properties [4]. Thus there is a need to understand the relationship between structure, vacancy distribution and magnetic properties. The study of individual vacancies and magnetic spin ordering is not currently possible by experimental/analytical means, and requires advanced QM modelling techniques.

The simulation of defects in ionic materials has had considerable success using classical atomistic methods and QM methods using HF and DFT. However, in the case of strongly covalent materials such as Fe-sulphides, the classical approach is not appropriate. Periodic DFT methods give a much more accurate description of the system under consideration but, due to their periodic nature, will always include a contribution from defect-defect interactions. Furthermore, the treatment of charged defects is difficult because of the need to introduce a neutralising background charge. Recently, hybrid embedded cluster schemes have been developed to allow the defect and surrounding lattice to be treated at a high level of theory [5]. This QM cluster is then embedded an array of lattice points which are treated classically. The advantage of this scheme is that it allows neutral and charged defects to be treated at a QM level while avoiding all defect-defect interactions.

If the technique is to be made applicable to the Fe-sulphide minerals, then the description of the embedding region must be replaced by a more accurate description of the system. This can be achieved through the integration of embedded cluster methods with the SIESTA approach. This will allow large systems to be studied at a high level of theory but with a manageable computational cost.

1. D.J. Vaughan and A.R. Lennie (1991) Sci. Progress Edinburgh 75, 371.
2. A.R. Lennie and D.J. Vaughan (1996) Mineral Spectroscopy: Geochemical Society Special Publication No. 5, 117–131. Eds. M.D. Dyar, C. McCammon and M.W. Schafer.
3. JWA Kondoro (1999) J. Alloys & Compounds 289, 36.
4. J.H.P. Watson, B.A. Cressy, A.P. Roberts, D.C. Ellwood, J.M. Charnock and A.K. Soper. J. Magnetism & Magnetic Materials 214, 13.
5. P.V. Sushko, A.K. Shluger, and C.R.A. Catlow (2000) Surf. Sci. 450, 153

4 Radiation damage, leeching and encapsulation of radioactive waste

The problem of long-term storage of high-level radioactive waste materials is one of the most pressing environmental problems. It is likely that nuclear power will begin to take a more important role in energy strategies, and the problem of long-term storage of high-level radioactive waste will become even more urgent. Any encapsulation must be physically and chemically stable within a geochemical environment over many generations, and must not be adversely affected by continued radiation damage. Among the types of materials proposed as matrices for encapsulation are glasses, crystalline ceramics, clays and cements. The potential use of crystalline ceramics has arisen from the observation that several minerals (e.g. zircon, ZrSiO4, and titanite, CaTiSiO5) have contained natural radioactive cations (such as uranium and hafnium) for geological timescales without significant leaching [1,2]. Encapsulation in crystalline ceramics may replace encapsulation in glasses for solid storage of radioactive waste because leach rates in crystalline matrices are likely to be significantly lower than in glasses. Experimental studies are focusing on understanding the formation and nature of radiation damage on an atomic scale, and how the damage changes the mobility of radioactive cations [3,4]. This programme of work is now being developed through funding by BNFL.

Realistic computer simulation work is of paramount importance in this work [5,6]. Simulations can give information for time scales that are inaccessible to experiment (i.e. over times shorter than 10–10 s), and length scales that range from interatomic distance to 100 nm. However, the simulations that have been performed to date (in Cambridge [5,6], France and the USA) have been limited in scope because the energy of the recoiling nucleus following a radioactive decay event has been an order of magnitude smaller than actual recoil energies. This has meant that the sample sizes need not be as large as would otherwise be required, and the size of damaged regions of the crystal structure following a single radioactive decay event is not overly large for detailed inspection. However, there are certain critical features of damaged materials that are not captured by these low-energy simulations, most notably the experimental observation that there is a 15–20% reduction in density in highly damaged materials. Such discrepancies should either vanish or become explicable when the recoil energies sizes of simulation samples can be scaled up.

We now need to develop the ability to perform simulation studies of radiation damage in crystalline materials with realistic energies and realistically large length scales (i.e. dealing with millions of atoms rather than thousands). Higher energies imply the need for finer sampling of the time scale, leading to an unfavourable scaling with current algorithms. The main tool for these simulations will be classical molecular dynamics simulation with empirical interatomic potentials. However, these simulations are probing length and energy scales of these potentials that are not included in the databases from which they are obtained. Moreover, both NMR experiments and the low-energy computer simulations provide evidence for significant rebonding. For example, in zircon, the isolated SiO4 tetrahedra in the crystalline form are polymerised into chains of tetrahedra in the damaged regions. It is necessary to compare the classical simulations with more realistic models for rebonding energies, which necessarily requires the use of QM models. These need to be constructed in a way that can handle very large samples.

1. Weber WJ et al. J Mat Res. 13, 1434 (1998).
2. Ewing RC, Can Min, 39, 697 (2001).
3. Rios S, Salje EKH, Zhang M, Ewing RC J Phys: Cond Matter, 12, 2401 (2000).
4. Capitani GC, Leroux H, Doukhan JC, Rios S, Zhang M, Salje EKH. Phys Chem Min, 27, 545 (2000).
5. K Trachenko, MT Dove, E Salje. J Appl Phys 87, 7702 (2000).
6. KO Trachenko, MT Dove, EKH Salje. J Phys: Cond Matter 13, 1947 (2001).

Page is being maintained by Arnaud Marmier