360 likes | 371 Views
Learn about the crucial role of electron-driven chemistry in various applications, including plasma technology and radiation damage. Explore the dissociative attachment of water molecules and the quantum dynamics of polyatomic collisions.
E N D
Dissociative Attachment to H2O C. William McCurdy, Lawrence Berkeley National Lab and The University of California Collaborators:D. Haxton (UCB), Z. Zhang, T. Rescigno (LBNL), and H-D. Meyer (University of Heidelberg)
Electron-Driven Chemistry Plays a Crucial Role in a Multitude of Applications High Intensity Plasma Arc Lamp (OSRAM-Sylvania) Plasma Flat Panel Display (Fujitsu) Plasma vapor deposition and plasma etching – no chemistry without electrons. LBNL-AMO
Electron-Driven Chemistry Associated with Ionizing Radiation Secondary electron cascades in mixed radioactive/ chemical waste drive much of the chemistry that determines how those materials age, change, and interact with the natural environment. Cascades of secondary electrons from ionizing radiation Low energy electrons with energies significantly below the ionization energies of DNA molecules can initiate single and double strand-breaks by attaching to components of DNA molecules or the water around them and driving bond dissociation. Most energy deposited in cells by ionizing radiation is channeled into secondary electrons between 1eV and 20eV (Research group of L. Sanche) LBNL-AMO
Secondary electrons play a pivotal role in radiation damage • Dissociative attachment and resonant processes occur primarily for electron energies below 20 eV • The energy distribution of secondary electrons emphasizes those processes Figure from Thom Orlando, Ga Tech LBNL-AMO
A primer on the electron-driven chemistry of water • The ground state configuration of water is • 1a122a121b223a121b12 1A1 • Electron-driven chemistry through both dissociative excitation and dissociative attachment Electron impact dissociation Dissociative Attachment LBNL-AMO
Simple molecular orbital picture of the dissociation processes • Dissociative excitation proceeds by excitation of low-lying dissociative electronic states, e.g, those observed in EELS spectra: • 1a122a121b223a121b114a111,3 B1 • 1a122a121b223a111b124a11 1,3A1 • Dissociative attachment of H2O proceeds through Feshbach resonances • 1a122a121b223a121b114a122B1 (~ 6.5 eV) • 1a122a121b223a111b124a12 2A1 (~ 9 eV) • 1a122a121b213a121b124a12 2B2 (~ 12 eV) LBNL-AMO
H- production is primarily through the 2B1 Resonance C. E. Melton, Journal of Chemical Physics, 57, pp.4218-25, (1972). LBNL-AMO
DA products are vibrationally (and rotationally) excited • Up to v = 7 is observed for OH stretch in the H- production channel • Rotational excitation: J=7 most populated for ground state OH, j=4 or 5 for v =4 OH stretch D. S. Belic, M. Landau and R. I. Hall, Journal of Physics B 14, pp.175-90 (1981) Vibrational excitation cross sections H- energy distribution for various electron energies LBNL-AMO
The problem: From first principles, solve the scattering problem including the nuclear dynamics, predict the cross sections and show how they display the polyatomic dynamics of the collision. • Breaking up the problem into two parts: • Electron scattering for fixed nuclei: Calculate the position and lifetime of the Feshbach resonances • Nuclear dynamics during the resonant collision: Calculate the quantum molecular dynamics leading to dissociative attachment LBNL-AMO
O r R H H A Complete ab initio Treatment of Polyatomic Dissociative Attachment: 2B1 Resonance • Electron scattering: Calculate the energy and width of the resonance for fixed nuclei • Complex Kohn calculation of fixed-nuclei electron scattering cross sections (7 state close coupling)– produce • CI calculations -- produce • Fitting of complete resonance potential surface including the electronically bound (product) regions • Nuclear dynamics in the local complex potential model on the anion surface • Multiconfiguration Time-Dependent Hartree (MCTDH) • Flux correlation function calculation of DA cross sections LBNL-AMO
A D B C Complex Kohn Variational Method Variational Functional for the T-Matrix (scattering amplitude) Trial wave function for the N+1 electron system target continuum exchange Correlation and Polarization Continuum functions are further expanded in combined basis of Gaussians and continuum functions LBNL-AMO
ER and G are determined from the eigenphase sums in the Complex Kohn calculation near the resonance E.g. at equilibrium geometry G = 0.005819eV LBNL-AMO
Real part of resonance energy, ER(r,R,g), nearly parallels the “parent” 3B1 state When the resonance becomes bound, the CI calculation of ER(r,R,g) must dissociate properly Dominant configuration is 1a122a121b223a121b114a12 A robust basis set is necessary: au-cc-pvTZ (augmented, correlation consistent, polarization and valence triple zeta) CI (898,075 configurations): Singles and doubles excitation from the CAS reference space (1b2, 3a1, 1b1, 4a1, 5a1, 2b2)7 4a1 is the “resonance” orbital, 5a1 and 2b2are important for correct dissociation and correlation Configuration Interaction calculation for real part of the 2B1 (2A’’) resonance surface LBNL-AMO
r1 O r2 H H Entire 2B1 (2A’’) potential surface is fit with combination analytic fit and 3-D spline. OH +H- Break up OH +H- LBNL-AMO
r1 O r2 H H Q = 00 O-+ H2 150 350 OH +H- 700 104.50 1250 OH +H- 1500 1800 q LBNL-AMO
O H H A view of the O- channel H2 +O- LBNL-AMO
O r R g H H As a function of angle in Jacobi coordinates LBNL-AMO
Width of 2B1 (2A’’) resonance J. D.Gorfinkiel, L. A. Morgan, and J.Tennyson, J. Phys. B 35 543 (2002). Function of r1 with r2 and q fixed at equilibrium: (at equilibrium geometry G = 0.005819eV) LBNL-AMO
Width of 2B1 (2A’’) resonance q = 104.5 r2 = 1.81 C2v symmetry LBNL-AMO
O r R H H A Complete ab initio Treatment of Polyatomic Dissociative Attachment: 2B1 Resonance • Electron scattering: Calculate the energy and width of the resonance for fixed nuclei • Complex Kohn calculation of fixed-nuclei electron scattering cross sections – produce • CI calculations -- produce • Fitting of complete resonance potential surface including the electronically bound (product) regions • Nuclear dynamics in the local complex potential model on the anion surface • Multiconfiguration Time-Dependent Hartree (MCTDH) • Flux correlation function calculation of DA cross sections LBNL-AMO
Local Complex Potential Model W(R) = ER(r,R,g)-i G(r,R,g) Propagate on complex resonance potential surface Solution, z(R), is Fourier Transform Cross section from the asymptotic behavior of z(R) LBNL-AMO
r R g H H Hamiltonian for Nuclear Motion O • For J = 0 the wave function is a function of only internal coordinates • For arbitrary J: LBNL-AMO
Solving the multidimensional Time-Dependent Schrödinger Equation • Multi-configuration time-dependent Hartree (MCTDH) method in collaboration with Prof. H. D. Meyer, University of Heidelberg • Can be converged to the exact solution -- efficient “adaptive” method. • Start with a time-independent orthonormal product basis (DVR) • The MCTDH wave function is a time-dependent linear combination of configurations (products of single-particle functions) • The time-dependent single-particle functions are represented as • Derive equations of motion for both the coefficients and the single-particle functions LBNL-AMO
Local complex potential wave packet starting from ground state H2O LBNL-AMO
H- from single excitation of bend LBNL-AMO
Cross Sections for OH vibrational states compared with experiment Total 0 1 2 3 4 5 6 7 D. S. Belic, M. Landau and R. I. Hall, Journal of Physics B 14, pp.175-90 (1981) LBNL-AMO
Cross Sections compared with experiment – experiment scaled to match peak of total cross section 0 1 2 3 4 5 6 7 LBNL-AMO
Rotational State Distributions for OH + H- OHv = 0 Integrated over incident electron energy OHv = 3 LBNL-AMO
O-+H2 channel Experimental peak at about 6.5 eV is LBNL-AMO
Experimental Isotope Effect D/H H- from H2O D- from D2O LBNL-AMO
Theoretical Isotope Effect D/H D- H- LBNL-AMO
It is now possible to treat dissociative attachment to a triatomic in full dimensionality from first principles. Dynamics of the 2B1 (2A’’) resonance leads almost exclusively to H - +OH An ab initio treatment reproduces the cross sections for producing OH in excited vibrational states to within a factor of 2 Interesting energy dependence is predicted for cross sections from H2O excited in symmetric stretch The dynamics of the other two resonances (2A1 and2B2) remains unknown. This calculation does not account for all of the observed O – production. Is the surface not accurate enough or is there production of O - from Renner-Teller or other nonadiabatic coupling between resonances? Isotope effect still fails to agree with experiment – Is the shape of the width “surface” the cause? Conclusions and Open Questions LBNL-AMO