ClayPoly Project

1
Clay-Poly Project

• Quarterly Meeting
• UCL 30th March 2004

2
Contents

• Objective 5 - Systematic Monte Carlo calculations
• Objective 5 - Investigate systems using
  large-scale molecular dynamics
• Objective 1 - Investigating reactivity in
  clay-polymer systems
• Objective 15 - Using electronic structure
  methods to investigate clay-polymer rectivity
• Objective 5 - Develop informatics approaches

3
Monte Carlo Simulation Results
4
Monte Carlo SimulationsCalculated Loadings
Number of Monomers
5
Monte Carlo SimulationsCalculated Loadings
Organic Content
6
Monte Carlo Simulations - Summary

• GCMC Gives a loading of circa 30 organic content
  for all systems.
• The strongest interaction energy from CGMC
  results is usually one of the linear
  conformations.
• The CMC results shows good agreement with the
  experiment data for the PEO backbone monomers -
  so far.
• CMC data suggests that where PPO backbone
  monomers are used, some of the monomer is outside
  the gallery in the TGA measurement.

7
Objective 1Investigating Reactivity in
Clay-Polymer SystemsEffect of monomer backbone
lengthIncreased reactivity systems
8
Predicting Reactivity

• Now we can simulate large systems we can look at
  simulating reactivity as a function of chain
  length. The size of molecule means we must use
  classical methods.
• The close contacts for -OH H-O group condensation
  reactions can be examined.
• The close-contacts between CC groups can also be
  investigated.
• It is hard to examine orientation between the
  reacting groups, which is an important aspect for
  reactivity.

9
Predicting Reactivity

• We need some experimental evidence to base this
  on.
• The test of all knowledge is experiment.
  Experiment is the sole judge of scientific
  truth.
• R. P. Feynman

10
Objective 1 5Simulating electronic
structure/reactivity relationships
11
Ab initio Calculations

• Polymerisation of acrylates

H
12
Methods

• Gaussian PW91, 6-311G
• Gas phase
• Model for the molecule acid CH2CHCOOH
• Cambridge Serial Total Energy Package (CASTEP)
• Density Functional Theory RPBE
• supercell approximation (3D periodic system)
• plane waves and pseudopotentials
• Geometry minimisations (0 K)
• Cut-off for the energy 300 eV

13
Results with Gaussian
DH? 3.0 kcal/mol DH 51.0 kcal/mol
14
CASTEP in progress...

• Model acrylate/clay

DEint. 22.8 kcal/mol

• Reaction

DH 36.0 kcal/mol
15
Objective 5Develop informatics
approachesEstablish materials- molecular
properties links
16
Informatics

• A simulation/materials database structure was
  devised.
• This is being implemented at Durham.
• Kathryn will present this tomorrow.

17
Summary

• Objective 1
• Now have the capability to investigate the effect
  of monomer backbone length on the reactivity.
• We can investigate those systems which show
  greater reactivity by either electronic/classical
  methods.
• We do need to discover some method to
  experimentally determine the reactivity.

18
Summary

• Objective 5
• We have performed systematic Monte Carlo
  simulations.
• A new code, Towhee, will allow us to extend these
  studies to very large systems.
• Large-scale MD calculations are under way.
• Ab initio methods have been used, and will be
  investigated further later this year, to study
  clay-monomer interactions.
• The database is under continuous development.

19
Clay-Poly Project

• Quarterly Meeting
• UCL 31st March 2004

20
Contents

• Predicting the monomer loading in Mmt
• Populating the Clay-Poly database
• Large-scale simulations
• Forcefield parameterisation
• Future work

21
Predicting the Monomer Loading A comparison
between Na-Mmt and K-Mmt PEODiAc 259Mw
22
Predicting the Monomer Loading

• Calculating the uptake energetics
• Used the method of Wang et al for the hydration
  of hydrotalcite.
• Uptake energy defined by
• Where
• N Number of monomers

23
Loading Energetics of PEO-DiAc259K-Mmt and Na-Mmt
Energy (kCal/mol)
Monomers
24
Summary of Loading Energetics

• It is clear why K-Mmt might act as a swelling
  inhibitor. The K cation is weakly hydrated
  resulting in a smaller d-spacing and more layer -
  K interaction.
• This results in a very high initial barrier to
  swelling.
• It is not clear why K-Mmt PEODiAc should
  exfoliate, and not the Na-Mmt system.

25
Populating the Database
26
Database Data - Subset 1
27
Database Data - Subset 2
28
Database Simulations Underway
29
Large-Scale Simulations
30
Large-Scale Simulations

• We have run systems with circa 15K atoms.
• However, large-scale is usually gt100K atoms.
• There has been a problem with generating
  large-scale (gt100K atom) models for simulations.
• 3 Possible routes have been explored
• i) Use DL-POLY builder
• ii) Use LAMMPS replicator
• iii) Use latest commercial Materials
  Studio

31
Large-Scale Simulations

• The DL-POLY builder relies on a Cerius model -
  the same problem as LAMMPS.
• The LAMMPS replicator is only for biomolecules -
  no repeat unitcell.
• Materials Studio does not do what the
  manufacturers say it does!
• SOLUTION - Matt Harvey has written a unit-cell
  replicator. This uses .car and .mdf format
  allowing PDLMD to still be used to process the
  data.

32
Large-Scale Simulations

• Currently we have running
• 20,000 atom simulation at CCS
• 150,000 atom simulation at HPCx
• Using a simpler model - hydrotalcite

33
Large-Scale Simulations - Summary

• We can now build and run large clay model
  systems.
• There is a small problem with one part of the
  process for assigning the forcefield types to
  atoms - hope to get this sorted soon.
• The latest version of LAMMPS has a converter to
  allow the large simulations to be visualised in
  many free visualiser software.

34
Forcefield Parameterisation
35
Forcefield Parameterisation

• Currently, the Teppen forcefield lacks parameters
  for the glycidyl ether group and the conjugated
  carboxylate group.
• However, other forcefields, e.g. Dreiding, use
  more general parameters. These cann be applied
  within the Teppen FF.
• Crystal structures will be downloaded from the
  CCDC and minimised with both force-fields.
• If no appreciable difference is noted we will use
  these more general parameters

36
Future Work

• Complete all MC/MD simulations using Discover to
  populate database.
• Trial large-scale Monte Carlo simulations.
• Calculate volume fractions of monomers.
• Electronic structure calculations to understand
  the clay reactivity.
• Complete database.