A virtual research organization enabled by eMinerals minigrid: An integrated study of the transport and immobilization of arsenic species in the environment.

A virtual research organization enabled by eMinerals minigrid:

An integrated study of the transport andimmobilization of arsenic species in the

environment

Zhimei Du

September 2006

A NERC’s eScience testbed project

Environment from the Molecular Level

History

A NERC’s eScience testbed project

Environment from the Molecular Level

The aim: Incorporate grid technologies with computer simulations to tackle complex environmental problems.

Phase 1: Establish eMinerals minigrid and a functional virtual organisation.

Phase 2: Fully explore the established infrastructure to

perform simulations of environmental processes.

Components of eMinerals minigrid

TeamPeople: Scientists, Code developers, escientists

Institutions:

University of Bath, Birkbeck College London, UCL,

The Royal Institution, Cambridge,

the CCLRC Daresbury, Reading

www.eminerals.orgwww.eminerals.org

A NERC’s eScience testbed project

Environment from the Molecular Level

Problem

• A pressing environmental issue: the contamination of groundwater sources by arsenic.

• Rise massive human health problems.

• The scale: millions people worldwide at risk. The scale of this environmental disaster has never been seen before.

• It has become a worldwide catastrophe.

A NERC’s eScience testbed project

Environment from the Molecular Level

Global arsenic occurrence

A NERC’s eScience testbed project

Environment from the Molecular Level

Arsenic occurrence in Asia

A NERC’s eScience testbed project

Environment from the Molecular Level

Solution

• Possible way: selective adsorption.• Promising adsorbents: iron-bearing minerals.

• A computational approach: A comprehensive study of the capabilities of different iron-bearing minerals.

A NERC’s eScience testbed project

Environment from the Molecular Level

Challenge Needs• Simulations at different levels.• Various methodologies • Many iron-bearing minerals. • People• Techniques• Infrastructures• Workflows, high-throughput, • Data management, computing resources.

A NERC’s eScience testbed project

Environment from the Molecular Level

A real challenge !!!!

Grid techniques

Communication tools

AG meeting: acting as a valuable management tool.

for members contributing their ideas to collaborative papers.

Wiki: exchange ideas, deposit news, edit collaborative papers.

disadvantage: not support instant communication.

Instant message useful for members of the project developing new tools.

A NERC’s eScience testbed project

Environment from the Molecular Level

Grid environment support for workflow

Job submission: using my_condor_submit (MCS) --- a meta-scheduling job submission tool , where Condor’s DAGman functionality and storage resource brokers (SRB) are used.

Workflow in three steps:

download input from SRB MCS decides where to run job upload output to SRB.

Our practice has shown: the SRB is of prime importance for data management in such collaboration.

A NERC’s eScience testbed project

Environment from the Molecular Level

Benefits of grid techniques for scientific studies

Example: Quantum mechanical studies of the structures of Goethite, Pyrite and Wüstite

Problem: The electronic and magnetic structures of many iron-bearing minerals

are not very well represented by traditional density functional theory. Minerals: Goethite, Pyrite and Wüstite.

Task: Compare GGA and hybrid-functional calculations with experimental

data to decide the best way to describe these minerals. Magnetic structures: Ferric iron (3+): AFM, FM Ferrous iron (2+): AFM, FM, NM

A NERC’s eScience testbed project

Environment from the Molecular Level

Maria Alfredsson

Example (continue):

Calculations needed: • 5-10 different hybrid-functionals for each mineral• 10 to 20 runs needed for each mineral.

• These are compute intensive calculations !!!!

• All calculations are independent of each other.

• Performed on UCL Condor-pool (> 1000 processors) using the MCS job submission tool.

• Calculations are completed within a couple of months

Prior to this eScience technology, this type of study might have taken a year or longer

A NERC’s eScience testbed project

Environment from the Molecular Level

Science outcomes

Bulk calculationsBulk calculationsSurface stabilitiesSurface stabilitiesHydration processesHydration processes

Computational methods used:Computational methods used: Quantum mechanical calculations(e.g.DFT)

Interatomic Potential Methods

• Static lattice energy minimisation

• Molecular dynamics simulations

A NERC’s eScience testbed project

Environment from the Molecular Level

• Pyrite plays an important role in the transport of arsenic.

• Experiment: Arsenic substitutes for sulphur, forming AsS di-anion

groups rather than As2 groups Arsenic substitutes for iron. • Using first-principles calculations How As incorporated ??

Where it sits in the lattice??At Fe or S sites??

Incorporation mechanism of arsenic in pyrite (FeS2)

Marc Blanchard

A NERC’s eScience testbed project

Environment from the Molecular Level

DFT (CASTEP code):• The AsS configuration is the most energetically favourable when pyrite

precipitates or is stable.• Incorporation of arsenic as a cation is energetically unfavourable in pure pyrite.

As in Fe site As in S site

AsS As2 As instead of S2

As in Fe site As in S site

AsS As2 As instead of S2

A NERC’s eScience testbed project

Environment from the Molecular Level

Zhimei Du

Structures and stabilities of iron (hydr)oxide mineral surfaces

• Iron (hydr)oxide minerals: promising adsorbents to immobilise

active arsenic and other toxic species in groundwater.

• A large number of simulations required to examine the surface structures and stabilities of these minerals.

{100}

{011}{101}

{100}

{011}{101}{111}

{001}

{010}{101}

{111}

{110}

{001}

{012}

{101}

{100}

(a)

Calculated bulk structures and the dry (bottom left) and hydrated (bottom right) thermodynamic morphologies of (a) Hematite, (b) Goethite, (c) pure iron hydroxide.

(b) (c)

A NERC’s eScience testbed project

Environment from the Molecular Level

Calculated surface energies for three iron (hydr)oxide minerals, including both dry and hydroxylated surfaces

• Surface energies of Fe(OH)2, are lower compared to those of Fe2O3 and FeOOH due to the open layered structure of Fe(OH)2.

• In general, the hydroxylated surfaces are more stable than corresponding dry surfaces.

Fe2O3

Surface 001 012a 012b 012c 100 101 110 111a 111b Dry surface

energy (Jm-2) 1.78 1.87 2.75 2.35 1.99 2.34 2.02 2.21 2.07

Hydrated surface energy (Jm-2)

0.90 0.38 0.28 0.38 0.27 0.04 0.20 0.33 0.28

FeOOH Surface 010 100 110 001 011 101 111

Dry surface energy(Jm-2)

1.92 1.68 1.26 0.67 1.18 1.72 1.33

Hydrated surface energy (Jm-2)

0.51 1.17 0.68 0.52 0.34 1.32 1.12

Fe(OH)2

Surface 001 010 011 101 110 111 Dry surface

energy (Jm-2) 0.04 0.38 0.35 0.35 0.64 0.60

A NERC’s eScience testbed project

Environment from the Molecular Level

ARNALD

Arnaud Marmier

Molecular dynamics simulations of aqueous solution/goethite interfaces

• Immobilisation processes concern the adsorption from solution, but the exact structure of aqueous solutions in contact with surfaces is not yet completely elucidated.

Reasons:

• The distribution and local concentration of the various

species is difficult to observe experimentally.• Expensive ab initio calculations are unable to cope with

the amount of water required.

A NERC’s eScience testbed project

Environment from the Molecular Level

• In pure water, layering water structure formed near the surface.

• Layering structures appear near the surface for both sodium and chloride ions.

• There is a clear build up of negative charge near the surface.

Reason: the adsorption of chloride ions.

A NERC’s eScience testbed project

Environment from the Molecular Level

Conclusions

Our experience showed that

• With the support of grid technologies it is very promising to solve complex scientific problems.

• We can achieve our goals in a much quicker, more comprehensive and detailed way.

A NERC’s eScience testbed project

Environment from the Molecular Level

Acknowledgement

The work was funded by NERC via grants NER/T/S/2001/00855, NE/C515698/1 and NE/C515704/1.

A NERC’s eScience testbed project

Environment from the Molecular Level