1 GODDARD - MSC/Caltech GODDARD - MSC/Caltech PASI-Caltech 1/5/04 PASI-Caltech 1/5/04 First principles Simulations of Nanoscale Materials Technology William A. Goddard III Charles and Mary Ferkel Professor of Chemistry, Materials Science, and Applied Physics Director, Materials and Process Simulation Center California Institute of Technology, Pasadena, California 91125 [http://www.wag.caltech.edu] NSF Pan-American Advanced Studies Institutes (PASI) Workshop January 5-16, 2004 Computational Nanotechnology and Molecular Engineering MSC/Caltech: Dr. Mario Blanco and Dr. Mamadou Diallo
79
Embed
NSF Pan-American Advanced Studies Institutes (PASI) Workshop January 5-16, 2004
NSF Pan-American Advanced Studies Institutes (PASI) Workshop January 5-16, 2004. Computational Nanotechnology and Molecular Engineering MSC/Caltech: Dr. Mario Blanco and Dr. Mamadou Diallo. First principles Simulations of Nanoscale Materials Technology William A. Goddard III - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Born El Centro, California (in Southern California desert east San Diego) Public schools of El Centro, Delano, Indio, Lodi, Firebaugh, MacFarland, Oildale (Bakersfield), Modesto, Yuma
108 PhD’s, 550 publications, Member National Academy of Science, Int. Acad. Quantum Molec. Sci.8 patents in protein structure prediction, new polymerization catalysts, semiconducting processing modelingCo-founded 5 companies (all still thriving)
What is Materials and Process Simulation Center? (MSC)
Senior Staff: All involved in nanotechnolgyWilliam A. Goddard III, DirectorBlanco, Director Process Simulations and Industrial TechnologyCagin, Director Mesoscale and Materials Science Technologyvan Duin, Director Force Field and Materials TechnologyVaidehi, Director of Biotechnology and PharmaMeulbroek, Director Software Integration and Databases Oxgaard, Director of Quantum and Catalysis TechnologyDiallo, Manager Molecular Environmental Technology Molinero, Manager of Complex Materials Simulations Willick, Manager of Computer Technology and Networks
Each student involved with bothdeveloping tools (theory and software) to solve impossible problems (1/2 effort)•using simulation (and experiment) to develop new materials and processes materials (1/2 effort)
Nearly every project has a strong coupling with experimentExperiment stimulates current theory with impossible problemsResults from simulation guide and interpret experimentsExperiment essential to validate theory and simulation
Ultimately, de novo simulation will be at the heart of chemistry, biology, materials science
Theory Essential to solve Grand Challenges in Technology
Materials Science: Nanoscale Technology• Opportunity: Tremendous potential for new functional materials
(artificial machines smaller than cells)• Problems: Synthesis, Characterization, Design• Need Multiscale Modeling: couple time/length scales from
electrons/atoms to manufacturingBiology: Protein Folding: Predict all structures of life• Opportunity Will soon have Genomes for All Life
Now: Over 700,000 genes but only ~10,000 protein structures • Problem: experiment can do only ~ 2000/year ($200 million)• Need: Prediction of Reliable Protein Structure and Function for
1,000,000 proteinsChemistry: Methane (CH4) Activation, Gas to Liquid• Opportunity Enormous reserves of CH4 for energy, chemicals,
and materials, mostly wasted• Problem: no efficient, selective, low temperature catalyst• Need: Predictive Mechanism to predict new catalystsFirst Principles Theory will lead developments of new technologies in 21st Century
• GM advanced propulsion: Fuel Cells (store H2, membrane, cathode)• Chevron Corporation: CH4 to CH3OH, Alkylation, Wax Inhibition• General Motors - Wear inhibition in Aluminum engines• Seiko-Epson: Dielectric Breakdown in nm oxide films, TED (B/Si)• Asahi Kasei: Ammoxidation Catalysis, polymer properties• Berlex Biopharma: Structures and Function of CCR1 and CCR5 (GPCRs)• Aventis Pharma: Structures and Function of GPCR’s
Stimulation toward solving impossible problemsCollaborations with Industry
Asahi Glass: Fluorinated Polymers and CeramicsAvery-Dennison: Nanocomposites for computer screens Adhesives, CatalysisBP Amoco: Heterogeneous Catalysis (alkanes to chemicals, EO)Dow Chemical: Microstructure copolymers, Catalysis polymerize polar olefinsExxon Corporation: Catalysis (Reforming to obtain High cetane diesel fuel)Hughes Satellites/Raytheon: Carbon Based MEMSHughes Research Labs: Hg Compounds for HgCdTe from MOMBEKellogg: Carbohydrates/sugars (corn flakes) Structures, water contentMMM: Surface Tension and structure of polymersNippon Steel: CO + H2 to CH3OH over metal catalystsOwens-Corning: Fiberglas (coupling of matrix to fiber)Saudi Aramco: Demulsifiers, AsphaltenesEach project (3 Years) supports full time postdoc and part of a senior scientist
Many experimental efforts to make nanoscale electronic devices based on molecules sandwiched between conducting surfaces (e.g., Jim Heath/UCLA-Caltech, Charley Lieber/Harvard, Jan Schön/Lucent, Phaedon Avouris/IBM)
This could be most useful. For example a future MEMS-scale device (say 20 microns in size) might have an onboard computer based on nanometer sized elements with built in sensors and logic to respond to local environment without the necessity of communicating to remote computer.
Thus to be useful the nanosized switches need not be as fast as current computer elements (GHz). They could be even as slow as KHz and still be useful.
Unfortunately little is known about the atomic-level structure and properties of these nanoelectronics systems, making difficult the design of improved devices.
Step 1: Use theory to predict the structures and mechanical properties of Nanoelectronics systems
Step 2: Use theory to predict electrical performance from 1st principles.
To predict electrical performance of experimental systems we must develop QM based methods to predict current/voltage (I/V) performance from first principles.
We have been working on this over the last 2 years (with support from an industrial partner). We can now predict I/V both for isolated molecules and for 2-dimensional slabs (still need to make the software faster for routine use).
We are now in the position to design new systems that can be synthesized and tested.
Napthyl and TTF nearly equally good donors Rotaxane ring binds to TTF > 300KRotaxane binds to napthyl > 250K Assume ring moves when apply external voltage which cause diode to switch.
The theory has already shown which site blue box sits on in the On/Off statesWe are now determining how the blue box moves: applied voltage or chemical reductant
Get fundamental understanding of how the device is switched.Currently the switching is done using chemical oxidants, reductants (experiments may take minutes) or by applying a voltage stepIt is not known what the limiting speed is or how it depends on the design
Failure of Theory? Bell Labs Molecular Transistor (Schön)
Drain
SiO2
SS SS SSSS
Source
Gate
PURESAM
SAM transistor: Nature (2001) Vol, 413, p. 713Ion/Ioff = 106
Single molecule transistor:Science (2001) Vol 294, p. 2138 Ion/Ioff = 450 at 1:5000 4K temperature
Goals in theoretical Study:• Validate the ability to predict field-effect modulation behavior • Determine the reason for the difference in behavior of these systems•Design new improved systems
Shift in Gate voltage can modify MOs near Fermi energy leading to a transistor effect but the maximum effect is only a factor or 20 not the 450 that Schoen reported. Also this value of 20 depends on the optimum placement of the energies of the MOs.
Our work was done in Sept. 2001 and not published since it seemed so inconsistent with experiment.
Since then the experimental results have been withdrawn because of possible fraud.
Thus the theory might be ok.
-6.5
-6
-5.5
-5
-4.5
-4
-3.5
-3
-2.5
-2
0 0.1 0.2 0.3 0.4 0.5
Gate field (V/A)
Energy (eV)
LUMO
HOMO
Under Drain voltage (1V)
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
0 0.05 0.1 0.15 0.2 0.25
Gate Field (V/A)
Drain current (uA)
Molecular orbitals shift under Gate voltage, can shift to fermi energy of electrodes
Predicted field-effect modulation under Gate voltage
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
0.00 0.20 0.40 0.60 0.80 1.00
Drain bias (V)
Drain Current (uA)
0
0.05
0.1
0.15
0.2
0.25
Experiment (Schoen)Field-effect modulation behavior Single molecule transistor
Manipulate Graphene, Graphite, and Single Walled Carbon Nanotubes
Motivation: Hongjai Dai, J. Am. Chem. Soc. 2001, 123, 3838-3839The experiment provided evidence of 1-PBSE absorption by imaging gold clusters presumably attached to the ester tails of the molecules absorbed on SWNT, but what is really happening?
1-Pyrene Butanoic Acid Succinimidyl Ester
Purpose: Understand the interaction of 1-PBSE with carbon nanotubes and with one another so that we can use non-covalent sidewall functionalization of carbon nanotubes to for molecular electronics and self assembly.
concept
The point: people are looking at various devices incorporation SWNTs. Non-covalent functionalization is one way of making compatible with these schemes.
More on MethodsStep Six: Put PBSE on (10,10) nanotube and compare the free energies of different packing configurations. Try different nanotube sizes and chiralities.
Validating Tools – Friction: Our tools need to produce the right energy barrier for translation of the absorbed molecules on the graphene surface. We should have the correct stresses for the shearing of graphite.
Validating Tools – Bonding forces: We should have the right shape for our molecules.
Validating Tools – Non-Bond forces: We should produce the correct inter-plane separation in graphite. Our QEq parameters should give the right charges.
The point: simulation can complement experiment by revealing things that experiment can't.
Surprises So FarFor isolated PBSE clusters on graphite, both the ester tail and the pyrene part of PBSE like to lay flat.
Van der Waals interaction is responsible for absorption onto graphite, graphene, and nanotube, but there is strong Coulombic interaction between absorbed PBSE’s.
On Graphene, Coulombic interactions between neighboring PBSE’s has led to tighter than expected optimum packing of PBSE. Whereas a single pyrene covered 50 graphene atoms, a single PBSE can cover 24 or 40 graphene atoms. Results on graphite and nanotubes are pending.
Ester tail
Pyrene
PBSE 4x3
PBSE 4x5
The point: simulation can complement experiment by revealing things that experiment can't.
SAM packing on Au(111) from MD simulation Coverage-dependence conformation (-stacking)
(1) Electronic structure of essential components from QM Role of the ring 1: provide low-lying LUMO’s Role of the ring 2: stabilize levels of nearby stations -orbitals dominant around HOMO-LUMO
(Aliphatic linkers and anchors negligible?)
(3) I-V Calculation from periodic QM and Green’s matrix ona. Fully extended (unfolded) formb. Fully folded formc. Fully folded form w/o linker/anchor
P-QEq Force Field ModelProper description of Electrostatics is critical
Pair-wise Nonbond Terms between all atomsShort range Pauli Repulsion plus Dispersion (EvdW)
Include Electrostatic interactions between all atoms (ECoulomb) •Describe Charges as distributed (Gaussians)
not point charges •Core charge is fixed to the mass of the atom
(total charge +4 for Ti) •Electron or shell charge is allowed to move
wrt the core (atomic polarizability)•This includes Shielding as charges overlap•Allow charge transfer (shell charges not fixed)•Self-consistent charge equilibration
Charge Equilibration (QEq) Require that the chemical potential (dE/dQi) be the same at every atomFix core charge, allow valence charge to transfer and to shift center
Electrostatic energy
Atomic self energy
Pairwise electrostatic interaction(Shielded Coulomb) ( ) l
kkir
rErflj
kiij QQQQrE
ij
ijklij
klij ⎟⎠
⎞⎜⎝
⎛ ⋅+
⋅
=ηη
ηη
,,int
Five universal parameters for each element describes charge distribution, polarization
Original Paper: Rappe and Goddard, J. Phys. Chem. 1991
QEq parameters are fitted to reproduce QM charges on molecules and polarizability of atoms
• containers- fill and empty upon change in pH, T permanent disposal• Sensors- detect change in local state• Energy source - nano fuel cell (H2/O2) • Templates- to organize parts into a regular array
Use Multiscale Theory to design and characterize systems
Nanodevice based on DNA B-Z motorFig. 13 a & 13 b : The nanomechanical device based on B-Z transition device shownin fig. 1. 13a : before operation 13b is after operation.
Double cross over DX in light blue. Each circle represents a DNA double helix.Dark blue circle has a fluorescent quencherB-Z device shown in red.After B-Z transition - quenchingseveral such motors can be mounted in a 2-D array that could cause circles in light blue to move
Before operation
After operation
Design and experimental synthesis and testing: Ned Seeman NYUTheory: Testing and modify: Goddard CaltechFunding: NSF-NIRT (Mike Roco
Performance is not adequate but considerable uncertainly regarding: The mechanism of Cathode catalysis, Transport of proton from anode to electrolyte to cathodeRole of polymer nanostructure and interfacesUse atomistic simulations to explain current system and then improve performance
Nafion: Polymer electrolyte Membrane in Fuel Cells
Nafion used in PEMFC (Polymer Electrolyte Membrane Fuel Cell) due to its high proton conductivity as well as electrochemical and mechanical stability.
Molecular level study of nanostructure and transport in hydrated Nafion using atomistic simulation
Most studies on this system have focused on
1. Structure of Nafion 2. Transport characteristics under the various conditions
(water content, temperature etc)
Motivation for studyNeed to run catalyst at 120C to avoid CO poisoning of catalyst, but nafion operates only <80CNeed to develop new generation of non-water membranes that operate at 120C and also do not freeze at -20CNeed first to understand nafion
We find that for low q, DR=0.1 has larger intensity than the DR=1.1.This implies that DR=0.1 (blocky Nafion) has more phase-segregation than DR=1.1.
Use calculated structure to predict the results of Small-angle scattering experiments.SAXS: electron density contrastSANS: deuteron density contrast
How measure the extent of phase-segregation?
Theory can predict many details about how structure and properties and predicts S(q) Experimental measurement of S(q) not give detailed structureCombining theory and experiment to conclude full detail about structure and properties
Computational Materials Design Facility (CMDF)Funded by DARPA
To integrate these disparate software into a Virtual Test FacilityWe are developing the CMDF (an integrated Materials Design Facility)• ICARUS visualization and GUI environment• Common-Generalized Materials Data Model• External-Plug in Materials Properties Simulation Tools as modules• Driven by Python as scripting Language• Archiving MP simulations in MP Database• Query/Search MP database• Generate and regenerate Materials Properties• Sockets for 3rd party MP application modules
Materials Properties Modeling software spans modeling scales from electrons, to atoms, to mesoatoms(QM, FF, MM/MD/MC, unit processes, analysis)
• Describe Chemistry (i.e., reactions) of molecules-Fit QM Bond dissociation curves for breaking every type of bond
(XnA-BYm), (XnA=BYm), (XnA≡BYm)-Fit angle bending and torsional potentials from QM-Fit QM Surfaces for Chemical reactions (uni- and bi-molecular)-Fit Ab initio charges and polarizabilities of molecules
•Pauli Principle: Fit to QM for all coordinations (2,4,6,8,12)•Metals: fcc, hcp, bcc, a15, sc, diamond•Defects (vacancies, dislocations, surfaces)•cover high pressure (to 50% compression or 500GPa)
•Generic: use same parameters for all systems (same O in O3, SiO2, H2CO, HbO2, BaTiO3)
Require that One FF reproduces all the ab-initio data (ReaxFF)
Three universal parameters for each element:1991 paper : use experimental IP, EA, Ri; ReaxFF get from fitting QM
ci
si
ci
oi
oi qRRJ &,,,?
Keeping: ∑ =i
i Qq
•Self-consistent Charge Equilibration (QEq)•Describe charges as distributed (Gaussian)•Thus charges on adjacent atoms shielded (interactions constant as R0) and include interactions over ALL atoms, even if bonded (no exclusions)•Allow charge transfer (QEq method)
short distance Pauli Repulsion + long range dispersion(pairwise Morse function)
•Valence Terms (EVal) based on Bond Order: dissociates smoothly• Bond distance Bond order Bond energy•Allow angle, torsion, and inversion terms where needed•Includes resonance (benzene, allyl)•describes forbidden (2s + 2s) and allowed (Diels-Alder) reactions •Atomic Valence Term (sum of Bond Orders gives valency)
•Pair-wise Nonbond Terms between all atoms•Short range Pauli Repulsion plus Dispersion (EvdW)
• Decomposition of High Energy (HE) Density Materials (HEDM)• MD simulations of shock decomposition and of cook-off• Reaction Kinetics from MD simulations• Ferroelectric oxides (BaTiO3) domain structure, • Pz/Ez Hysteresis Loop of BaTiO3 at 300K, 25GHz by MD• Catalysts: Pt Fuel cell anode, cathode, CH transformations• Ni catalyzed growth of bucky tubes• MoOx catalysts: catalyzed growth of bucky tubes• Si-SiO2 and Si-SiOxNy interfaces• Metal alloy phase transformations (crystal-amorphous)• Enzyme Proteolysis • Si-Al-Mg oxides: Zeolites, clays, mica, intercolation with
• Use QM to Extend and develop the detailed reaction mechanism for reactions of organics, organometallics– Unimolecular and bimolecular reactions
• Extend ReaxFF to fit all possible unimolecular and bimolecular reactions.
• Use ReaxFF in computational experiments to describe fundamental processes as function temperature and pressure including uni-,bi-,ter, -tetra-molecular reactions
• Include defects, finite grain size, binder, plasticizer in these simulations to determine models for their effect on microstructure and performance of materials
ReaxFF references:1: van Duin, A.C.T.; Dasgupta, S.; Lorant, F.; Goddard III, W.A. J. Phys. Chem. A 2001, 105, 9396.2: van Duin, A.C.T.; Strachan, A.; Stewman,S; Zhang, Q.; Goddard III, W.A., J. Phys. Chem. A 2003, 107, 3803.3: Strachan, A.; van Duin, A.C.T.; Chakraborty, D.; Dasgupta, S.;Goddard III, W.A. Phys. Rev. Letters. 2003, 91, 098301. 4. Zhang. Q.; Cagin, T.; van Duin, A.C.T.; Goddard III, W.A. Phys. Rev. B, subm