1. Yoshiyuki Miyamoto ’ s talk Please standby. Yoshiyuki Miyamoto ’ s talk Please standby.

2. © NEC Corporation 2005 2 “Computational Challenges and Tools for Nanotubes" Sunday 26 June 2005, 13:00 -??:?? Studenternas Hus, Götabergsgatan 17, S-41134 Göteborg, Sweden Simulation of Excited State Dynamics in Nanotubes with Use of the Earth Simulator Acknowledgements: Osamu Sugino, ISSP, U-Tokyo Noboru Jinbo (RIST→TOSHIBA) Hisashi Nakamura (RIST) Savas Berber (Michigan State Univ.→Tsukuba Univ.) Mina Yoon (Michigan State Univ.→ORNL) Angel Rubio (Univ. Páis Páis Vasco, Spain) David Tománek (Michigan State Univ.) Fundamental and Environmental Res. Labs. Yoshiyuki Miyamoto

3. © NEC Corporation 2005 3 How to simulate excited state dynamics? (An approximated way) (An approximated way) |g> Reaction coordinate Potential |e> Constraint DFT Sugino & Miyamoto PRB 59, 59, 2579 (1999) ; ibid, B 66, 66, 89901(E) (2002). Time-dependent method can follow with no need of occupation assignment at every time-step.

4. © NEC Corporation 2005 4 CPU 0: ρ(t+Δt)=Σ|Ψ i (t+Δt)| 2 H KS (t+Δt)=H KS {ρ(t+Δt)} CPU 1: Ψ 1 (t+Δt)=exp{-i/ħΔtH KS (t)} Ψ 1 (t) CPU 2: Ψ 2 (t+Δt)=exp{-i/ħΔtH KS (t)} Ψ 2 (t) CPU n: Ψ n (t+Δt)=exp{-i/ħΔtH KS (t)} Ψ n (t) |Ψ 1 (t+Δt)| 2 |Ψ 2 (t+Δt)| 2 |Ψ n (t+Δt)| 2 H KS (t+Δt) MPI_Reduce & MPI_Bcast Parallel computing with respect to each wave function

5. © NEC Corporation 2005 5 Textbook for ‘TDDFT’ from Springer will soon appear (I hope)! Edited by E. U. K Gross, et al. - Contents - Principle (Runge & Gross) Perturbation theory Current density functional Non-adiabatic xc potential Real-time for finite systems (Rubio, et al.) Real-time for extended systems (Sugino & Miyamoto) etc..

6. © NEC Corporation 2005 6 1.Elimination of impurities from nanotubes by optical surgery. 2.Hot-carrier relaxation in nanotubes 3.Some notes for fluorescence from (n,0) tubes 4.Summary

7. © NEC Corporation 2005 7 Electronic excitation as a tool of cleaning!

8. 8 O Potential Profile (Oxidized)-(O-free) (3,3) defected (3,3) defected tube with O NEXAFS A. Kuznetsova, et al., J. Am. Chem. Soc. 123, 10699 (2001)

9. © NEC Corporation 2005 9 O CNT V.B. CNT C.B. O-related electronic levels O 2p O 2s

10. © NEC Corporation 2005 10 30 fs 60 fs 126 fs H2H2 Auger decay after O 1s to 2p excitation (~520 eV) CNT V.B. O 1s O 2s O 2p C.B. CNT V.B. O 1s O 2s O 2p C.B. Nanotube can heal its wound! Atomic scale surgery of O-extraction from NT. A gaping wound Combination of H2 H2 introduction and electronic excitation: Post fabrication processing [More details, Miyamoto, et al., PRB70, 233408 (2004).]

11. © NEC Corporation 2005 11 Electronic excitation and subsequent decay

12. © NEC Corporation 2005 12 Nanotube devices: FET, non-linear optical devices Switching Frequency – lifetime of excited carriers Ichida, et al., Physica B 323, 237 (2002) Hertel and Moos, PRL 84, 5002 (2000) e-e comes before e-ph

13. © NEC Corporation 2005 13 hole electron Hot carrier decay in (3,3) nanotube under assumed lattice temperature

14. © NEC Corporation 2005 14 electron-phonon is dominant Energy Energy transfer between electrons and ions electron-electron is dominant /96 atom

15. © NEC Corporation 2005 15 Some notes for fluorescence from (n,0) tubes

16. © NEC Corporation 2005 16 Identificati Identification of semiconductor nanotubes by fluorescence Miyauchi et al., CPL 387, 198, (2004) Excitation

17. © NEC Corporation 2005 17 Many possible decay paths into optically forbidden e-h pairs Electronic structure of (7,0): LDA HO-LU pair is optically inactive! Time-constant of the possible decays should be checked by TDDFT-MD simulations! Question! Can we see light-emission from thinner (n,0) tubes? Can we see light-emission from thinner (n,0) tubes?Question! π-band T.B. model cannot reproduce this band. (LDA:0.94eV) (LDA: 2.31eV) (LDA:2.33eV) (LDA:2.98eV) (LDA:2.94eV)

18. © NEC Corporation 2005 18 1.TDDFT-MD simulation has become able beyond 500 fs. 2.We start from constraint DFT with Hellmann-Feynman force (Ehrenfest approach) 3.Application to nanotubes photo-induced defect dynamics decay dynamics of hot carriers examination of luminescence from (n,0) tubes (future works)

19. © NEC Corporation 2005 19 t = 0: Promote the electronic occupations to mimic the excited states. Then perform the static SCF calculation. t > 0: Solve ψ n (t+Δt)=exp{- i ΔtH(t)} ψ n (t). No Observation of the nonradiative decay! lifetime, decay path Yes Do MD. Hellmann-Feynman theorem works How to simulate excited state dynamics? |g> Reaction coordinate Potential |e> Is the matrix of H Kohn-Sham diagonal? 1.No need of level assignment for a hole and an excited electron except at the beginning. 2. Automatic monitoring of the nonradiative decay (lifetime, decay path) without experiences.

20. © NEC Corporation 2005 20 |g> Reaction coordinate Potential |e> Ground state theory! Solving the eigenvalue problem Many software packages (VASP, CASTEP, GAUSSIAN, etc.) Excited state dynamics Solving the time-dependent problem First-Principles Simulation tool for Electron-Ion Dynamics TM Sugino & Miyamoto PRB 59, 59, 2579 (1999) ; ibid, B 66, 66, 89901(E) (2002). What code is used? Computational conditions (on the basis of standard band structure calculation 1.TDLDA (with adiabatic xc potential) 2.Troullier-Martins type separable pseudopotentials 3.Plane wave Cutoff energy: 40 Ry, 60 Ry 4.Time step: 0.08 a.u. (=0.0019 fs), 0.05 a.u.(=0.0012 fs) Computational conditions (on the basis of standard band structure calculation 1.TDLDA (with adiabatic xc potential) 2.Troullier-Martins type separable pseudopotentials 3.Plane wave Cutoff energy: 40 Ry, 60 Ry 4.Time step: 0.08 a.u. (=0.0019 fs), 0.05 a.u.(=0.0012 fs)