Nanoparticle Synthesis and Assembly from Atomistic Simulation Studies

Transcription

1 Nanoparticle Synthesis and Assembly from Atomistic Simulation Studies By Takumi Hawa School of Aerospace & Mechanical Engineering Oklahoma Supercomputing Symposium 2008 October 7 Supported by NSF, ARL, & NIST

2 Why Nanoparticle Some 75% of chemical manufacturing processes involve fine particles at some point. food tires & toners pharmaceuticals personal care & cosmetics Proper manufacturing processes: improve cost improve quality minimize waste waste provide safety Design & handling of these fine particles makes the difference between success & failure

3 Particle Synthesis Liquid Phase Easier to control size Production rate is low Vapor Phase High production rate Inexpensive! Difficult to control size Challenge: Control size in vapor phase!

4 Vapor phase synthesis of nanoparticles chemical precursor vapor + carrier gas Nucleation of monomer via chemical or physical gas to particle conversion Rapid particle growth via coagulation Particle number decreases Early stage of veryfine aggregate formation

5 Collision & Sintering Time t sintering t collision Zachariah et al (1996) t collision Zachariah et al (1996)

6 Vapor phase synthesis of nanoparticles sintering faster than collisions chemical precursor vapor + carrier gas collisions faster than sintering Nucleation of monomer via chemical or physical gas to particle conversion Rapid particle growth via coagulation Particle number decreases Early stage of veryfine aggregate formation Large spheres form Large aggregates form Particle growth and morphology are determined by the competition of collisions and sintering

7 Hydrogen Passivation Surface Reactivity? Property? Sintering? H t collision & t sintering

8 Contents Size of nanoparticles Shape of nanoparticles Assembly of nanoparticles.

9 PARTICLE SIZE ASSEMBLY CONTROL

10 Collision of H coated particles (liquid & solid) (cross-section view) 6 nm particles at 600 & 1500 K

11

12 Contact Surface Area Analysis Droplet interaction is proportional to the contact surface area. a0 KEapp Johnson et al. (1971) suggested πa 2 ~ N atoms 3 2 kt = 1 2 mv 2 KE app = 1 3 mv 2 k Phys. Rev. B 69, (2004)

13 Reactivity of the Coated Particles Critical approach energy for reaction 3 2 kt = 1 2 mv 2 KE app = 1 3 mv 2 k bounce reaction size, harder to react T, easier to react No thermal reaction - t col Phys. Rev. B 69, (2004)

14 Mathematical Model Assumption 1) Viscous fluid 2) Maintain a spherical shape (made by Frenkel (1945)) Continuity ΔS = 4πa 2 [ Δθ( sinθ) + O(Δθ 2 )] S: surface area Energy balance 16 3 πa3 ηγ 2 + 4πa 2 βξ(θ)γ = 4πa 2 σγ Energy dissipation due to viscosity, η Work done by relocation of surface atoms Work done by surface tension, σ σ = surface tension η = viscosity γ = velocity gradient Phys. Rev. B 71, (2005)

15 Effective passivation surface area ξ = Passivated surface Effective contact area = πaf c π af c + a 2 sin 2 θ ( ) fc ξ=1 (initially) Most of the energy is consumed by relocation of the surface atoms ξ=0 towards the end Phys. Rev. B 71, (2005)

16 Mathematical Model Solving the energy equation gives t = 2 3 ηd π 2 0 sinθ σ βξ( θ) dθ β=0 (bare) β>0 (coated) t Frenkel = 2ηd 3σ by Frenkel (1945) d = diameter = 2a Phys. Rev. B 71, (2005)

17 Sintering (Bare vs. Coated) (cross-section view) 6 nm droplets at 1500 K KEapp = 110,000 K

18

19 Dynamics of Sintering t = 2 3 ηd π 2 0 sinθ σ βξ( θ) dθ η = 5.9 centipose σ = 0.83 J/m 2 a 0 = 2.35 Å 6 nm particles at 1500K (σ-β)/σ=3.25*10^(-6) Work done by relocation of surface atoms dominates the initial process Mb Mc σ(coated)/σ(bare)=0.54 Mb/Mc =0.48 After the initial process, the surface tension dominates the sintering process Phys. Rev. B 71, (2005)

20 40 particle chain aggregate T = 1500 K 2.5 nm primary particles

21

22 Sintering Time for a Chain Aggregate t = η 1 d σ 2 A = 2 3 Atan B = 2 ln [ ] ln 1+ 2 π V C = ln ( ) 1 3 L 0 Atan π V 3 [ ] ln 1 2 π V 1 3 A + B + C ( ) L 0, and L π V, 2 3 L 0 Independent of primary particle diameter. Universal relationship that only depends on chain length. Phys. Rev. B 76, (2007)

23 Sintering Time for a Chain Aggregate t = η 1 d σ 2 A = 2 3 Atan B = 2 ln η = 5.9 centipose σ = 0.83 J/m 2 3 [ ] ln 1+ 2 π V C = ln a 0 = 2.35 Atan π V 3 Å [ ] ln 1 2 π V 1 3 A + B + C ( ) ( ) 1 3 L 0 L 0, and L π V, 2 3 L 0 Excellent agreement with MD. Phys. Rev. B 76, (2007)

24 Sintering Time for a Chain Aggregate t d = 2η t = t Frenkel d 3σ N 1 d t Frenkel = 2ηd 3σ by Frenkel (1945) η = 5.9 centipose σ = 0.83 J/m 2 a 0 = 2.35 ( ) 0.68 ( ) 0.68 Å Depends on the number of particle connections in a chain. J. Aerosol. Sci. 38, 793 (2007)

25 Fractal Aggregate Sintering Fractal Dimension, Df m R D f Df = 1 Df = 1.9 Df = 3 J. Aerosol. Sci. 38, 793 (2007)

26 Fractal Aggregate (Df = 1.9) 66 particles T = 1500 K 2.5 nm primary particles

27

28 Sintering time for fractal aggregate Fractal Dimension, Df m R D f Df = 1: wire Df = 1.9: aerosol aggregates Df = 3: compact Monotonic increase w/ Np. t decrease w/ Df. t d = 2η ( ) 0.68D f 3σ N 1 t Frenkel = 2ηd 3σ J. Aerosol. Sci. 38, 793 (2007)

29 Sintering time for fractal aggregate Fractal Dimension, Df m R D f Df = 1: wire Df = 1.9: aerosol aggregates Df = 3: compact Monotonic increase w/ Np. t decrease w/ Df. t d = 2η ( ) 0.68D f 3σ N 1 t Frenkel = 2ηd 3σ J. Aerosol. Sci. 38, 793 (2007)

30 Sintering time for fractal aggregate Fractal Dimension, Df m R D f Df = 1: wire Df = 1.9: aerosol aggregates Df = 3: compact Monotonic increase w/ Np. t decrease w/ Df. t d = 2η ( ) 0.68D f 3σ N 1 t Frenkel = 2ηd 3σ J. Aerosol. Sci. 38, 793 (2007)

31 PARTICLE SHAPE ASSEMBLY CONTROL

32 Vapor phase synthesis of nanoparticles sintering faster than collisions chemical precursor vapor + carrier gas Nucleation of monomer via chemical or physical gas to particle conversion Rapid particle growth via coagulation Particle number decreases Early stage of veryfine aggregate formation Large spheres form Nanoparticles are described as being sphere

33 Plasma Synthesis TEM images of cubic particles Experiments by Kortshagen et al at U of Minnesota

34 Stability of Nanoparticles Liquid Minimum surface area Solid Crystal structure Surface structure?

35 Stability of Nanoparticles Bare surface to volume ratio

36 Transition from cube to truncated octahedron 2980 Si atoms 4 x 4 x 4 nm cube

37

38 Stability of Nanoparticles Assume: particle surfaces are covered by H Bare surface to volume ratio Coated Additional H energies surface to volume ratio PE / Si atoms

39 Etching of Spherical Particles Sasaki et al, Vacuum 51, 537 (1998) Sasaki et al, Jap. J. Appl. Phys. 37, 402 (1998) > > (100) (110) (111) Experiment: R( 100) R( 111) 1.5 MD simulation: R( 100) R( 111) 1.37

40 PARTICLE ASSEMBLY

41 building blocks Nanoparticle Based Devices Assembly is a biggest challenge in Nanotechnology based device development. Microelectronic, optoelectronic devices Sensors Ag Ag Ag Au Au Ag Ag Ag P N P Need to control the location of particles in deposition process

42 Electrostatic Directed Assembly stamp (+) insulated substrate Charge patterns are unstable Stamp is easily damaged Non-insulated surface? We want to have: Stable charge patterns Stable structure Adjustable charge strength Available and reliable technology Use P-N junction

43 P-N Patterned Substrate Apply reverse bias voltage N-type substrate Silicon doped n-type GaAs substrate P-type stripes P-type Contact pad 1 µm Zinc doped p-type stripes and contacts are patterned by the photolithography plus ion implantation. The Spacing between p-type stripes are 30µm in width P-type 1 µm Monodisperse particles P-type /cm & flow rate = 1 lpm N-type N-type 30 µm N-type nozzle: 2mm in diameter & 1 cm above the substrate

44 Particles Particles Deposition Deposition on on PN PN - + N N P N P N P N N N N P P N P N

45 P-N Model & Simulations Experiment: nozzle P-N substrate Simulation:

46 P-N Model & Simulations We can summarize the factors involved In the deposition process: 1. External force, F ext : Electrostatic (F e ), van der Waals (F vdw ) and image forces (F i ) 2. Convective flow (only in x-direction) 3. Diffusion force (Brownian motion): random, non-directional force C c : slip correction factor 4. Drag force, F D : to resist the momentum change v: particle velocity v g Ag P-N Substrate From Langevin Equation we can derive the particle trajectory: (in terms of velocity) F D F ext Brownian motion ß=1/(particle relaxation time) Affected by electrostatic force Affected by diffusion force

47 Effect of Electric Field 1 Coverage Selectivity Simulation 30nm Experiment A Coverage selectivity increases as the reverse bias voltage increases Control location of particle deposition Reverse Bias (-Volts) Control coverage selectivity by voltage Coverage Selectivity

48 Size of Nanoparticles. Hydrogen passivation surface prevented particle growth. Slowed sintering process & developed viscous flow model. Shape of Nanoparticles. Hydrogen stabilizes Si crystals to be a cube. Assembly of nanoparticles. Used P-N junction. Developed a Dynamics Model & Simulations. Mechanics of nanoparticles. Sensitivity of morphology. Summary

49 Summary Acknowledgement Prof. M. R. Zachariah (U of Maryland & NIST) Dr. D. Tsai (Cabot Microelectronics) Prof. J. Zhu (U of Western Ontario, Canada)