Case Study Report

1
Case Study Report

• By Manish Shrivastava
• 19th April, 2006.

• Formation of soot (Products of Incomplete
  combustion of hydrocarbons)
• Deterioration of a silver catalyst surface due to
  adsorption of O2,Ar,Kr

2
Combustion Sources Pollution

• Emitted from various sources directly into the
  atmosphere
• Gasoline and diesel vehicles
• Biomass burning
• Industrial sources
• Other forms of combustion

3
What is fine particulate matter?
Diesel Soot
Small Suspended Particles -- Span wide size
range -- Complex shapes -- Complex composition --
Multiphase -- Many sources
Soot
Flame
SEM Images Pittsburgh Air Quality Study Gary
Casuccio, R.J. Lee Group, Monroeville, PA
4
Why do we care Particles reduces visibility
Yosemite, CA
Pittsburgh, PA
EPA (2001) National Air Quality Status and Trends
5
Soot Global Warming
Ship Tracks over Pacific.
Biomass burning in the Amazon.
Source Kulmala, 1996
6
Paper 1

• Violi A., Voth G.A., Sarofim A.F., 2005. The
  relative roles of acetylene and aromatic
  precursors during soot particle inception.
  Proceedings of the Combustion Institute, 30,
  1343-1351.
• Objective Describe soot formation in benzene
  flames
• Reactions occur at wide range of time scales
• Pico or nanoseconds for intramolecular processes
• Milliseconds for formation of first soot
  precursors
• Moleular dynamics
• Resolving atomic vibrations Requires time-scales
  of femtoseconds
• Reaching µ-seconds Large Computational
  Resources
• Kinetic-Monte Carlo
• Allows to generate different states bypassing
  barriers between states
• Bypassing high energy barriers Difficult in MD
• Allows to directly sample states at widely
  varying time scales

7
Approach Combine KMC and MD
Gas
Atoms

• Identify gas phase species (for molecular growth)
  (CHEMKIN package)
• Local environment for each atom Gas species and
  temperature
• Check Enough room to accomodate the gas phase
  species next to particle
• Select a reaction Proabibility proportional to
  reaction rate
• KMC modifies the reaction site according to
  reaction transition
• Variable Time step Consistent with rate
  constants for the chosen reaction
• MD Energy minimization for the new configuration
• Reaction sites for new particle Counted and loop
  back to step 1

8
MD Simulation details

• Thermodynamic Ensemble NVT
• Temperature control Noose Hoover thermostat
• Potential AIREBO describes
• Covalent bonding interactions
• Torsional energies of single bond rotations
• Van-der Waals using 12-6 Lennard Jones

9
Modeling Configuration

• Soot particle seed ensemble of few to 400 C
  atoms
• Two flame heights 8 mm and 10 mm
• Corresponding Temperatures 1800 K and 2300 K
• Benzene high upto 8 mm
• Acetylene high beyond 8 mm
• Interestingly Curvature of soot formed depends
  on temperature
• 2100 K more curved particles Due to breaking of
  C-H bonds and cage closure reactions
• High curvature Lower space for gas phase species
• Chemical structure different at different flame
  heights

10
Results Modeling Cage Closure Reactions

• Structure 1 Kinetic Monte-Carlo (breaking C-H
  and adding acetylene)
• First 2ps C1-C2 brought together while all other
  atoms are allowed to relax(Steps of 5A-6A)
• After ring closure Full relaxation performed

11
Results KMC-MD simulation
12
Conclusions Paper 1

• KMC-MD used to predict soot formation in a simple
  fuel-rich benzene acetylene flame
• Could model upto milliseconds using this
  procedure
• Comments
• Soot is very complex
• Depends on combustion conditions and material
  being burnt
• For eg. Soot formed when driving a vehicle on an
  inclined hill may be very different from driving
  the same vehicle on level road.
• Understanding soot formation and properties in
  the atmosphere Challenge
• Neverthless, starting from soot structures we
  understand is a first step to understanding and
  modeling its formation

13
Paper 2

• Shibahara M., Katsuki M., 2006. Molecular
  dynamics study of effects of adsorbed molecules
  on reaction probability and energy transfer.
  Combustion and Flame, 144, 17-23.
• Process Oxidation of a silver surface with
  oxygen molecules
• Objective Model reduction in reaction
  probability and energy transfer by oxidation due
  to adsorbates oxygen, krypton, xenon
• Motivation To understand processes involved in
  nanoscale fabrication or nanoscale catalytic
  combustion.

14
Simulation Details

• 12-6 Lennard Jones as both intra and inter
  molecular potentials between all atoms and
  molecules
• Chemical reaction Complex potential between
  oxygen molecule and silver surface
  (London-Eyring-Polanyi-Sato) based on good
  agreement with exp. Reaction probability data
• Depends on projected positions of O2 atoms on
  Ag, O-O bond length, configuration of Ag atoms on
  surface

15
Parameters

• Doesnt mention NVT or NVE
• I guess it is NVT since surface temperature
  maintained at a fixed value.
• Surface adsorbate coverage 25
• Incident angle of O2 molecules 30?
• Rotational and Vibrational temperature 1053 K
  and 1132 K respectively
• Number of silver molecules used Not given
• I calculated this from the diagram as 100
• Monte Carlo used to analyze surface states and
  target collision positions of O2 atoms on silver
  surface

16
Potential Energy Surface
O-O Ag (molecular adsorption)
O-O
-Ag-Ag- O-O
-Ag-Ag-
Barrier for molecular adsorption
Barrier for dissociative adsorption
17
Results

• Reaction probability increases with increasing
  translational energy of O2 molecules
• Adhesion Strength on surface larger Well depth ?
  is larger
• ? lowest for O2, highest for Xe
• Higher adhesion Lower mobility of adsorbed
  molecules on surface
• Higher mobility of adsorbed molecules More
  energy of colliding O2 molecule absorbed
• O2 spends more time near the surface Higher
  reaction probability

18
Results

• For O2 and Ar Surface reaction Probability
  increases with surface temperature, levels out at
  higher O2 translational energies
• For Krypton Surface reaction probability
  decreases with increase in surface temperature
• For a clean silver surface with nothing adsorbed
  same effect as Kr seen
• So adsorbed Krypton molecules act as a clean
  silver surface and dont affect the reaction
  probability

19
Conclusions and Comments

• Nature of adsorbates affects the reaction
  probability on a silver surface
• Mobility of adsorbates on surface increases for
  Ar and O2 but no effect on Kr molecules
• Surface reaction between Ag and O2 could be
  modeled using MD
• The study suggests that oxidation in presence of
  argon would be more efficient than in presence of
  krypton
• Comments No equations describing calculation of
  reaction probability
• Not clear why surface reaction probability on a
  clean silver surface decreases with increasing
  temperature.

20

• THANKS for your ears!!!!!!