MIT Department of Chemistry 5.74, Spring 2004: Introductory Quantum Mechanics II� Instructor: Prof. Robert Field 17 – 1 5.74 RWF Lecture #17 Normal ↔ Local Modes: 6-Parameter Models Reading: Chapter 9.4.12.5, The Spectra and Dynamics of Diatomic Molecules, H. Lefebvre-Brion and R. Field, 2nd Ed., Academic Press, 2004. Last time: ω1, τ, ω2 (measure populations) experiment ω2 ω ( abcd ) ← ( AB) ←1 g two polyads. populations in (1234) depend on τ. E res,k to devise optimal plucks for more complex situations E res (choice of plucks and probes) * multiple resonances * more than 2 levels in polyad could use f k = Overtone Spectroscopy nRH single resonance nRH + 1RH double resonance dynamics in frequency domain Today: Classical Mechanics: 2 1 : 1 coupled local harmonic oscillators QM: Morse oscillator 2 Anharmonically Coupled Local Morse Oscillators H eff Local . Antagonism. Local vs. Normal. Whenever you have two identical subsystems, energy will flow rapidly between them unless something special makes them dynamically different: * anharmonicity * interaction with surroundings spontaneous symmetry - breaking Next time: H eff Normal . 5.74 RWF Lecture #17 17 – 2 Two coupled identical harmonic oscillators: Classical Mechanics (R = Right, L = Left ) H = T ( PR , PL ) + V (QR , QL ) PR 1 T = ( PR , PL )G 2 PL geometry and masses [ ( ) ] ) ] 1 Grr PR2 + PL2 + 2Grr ′ PR PL 2 QR 1 V = (QR QL )F 2 QL = force constants [ ( 1 = Frr QR2 + QL2 + 2Frr ′QR QL 2 1 1 2 1 2 1 2 H = Grr PR + Frr QR + Grr PL + Frr QL2 2 2 2 2 ( ) ( ) H L0 H R0 +Grr ′ PR PL + Frr ′QR QL kinetic coupling potential (anharmonic) coupling 5.74 RWF Lecture #17 17 – 3 3 φ 1 2 Frr = k Frr ′ = kRL 1 1 1 1 m1 + m3 1 + = + = = m1 m3 m2 m3 m1m3 µ 1 (projection of velocity of ③ for cosφ Grr ′ = ① — ③ stretch onto ③ — ② m3 direction) kinetic coupling gets small for large m or φ = π/2 Grr = Each harmonic oscillator has a natural frequency, ω0: 1 1 k 1/ 2 ω 0 = F G = [ ] 2 π c µ 2 πc rr rr 1/ 2 and the coupling is via 1 : 1 kinetic energy and potential energy coupling terms. Uncouple by going to symmetric and anti-symmetric normal modes. Qs = 2–1/2[QR + QL] Qa = 2–1/2[QR – QL] Ps = 2–1/2[PR + PL] Pa = 2–1/2[PR – PL] plug this into H and do the algebra 1 1 1 H = + Grr ′ Ps2 + ( k + k RL ) Qs2 2 2 µ 1 1 1 + − Grr ′ Pa2 + ( k − k RL ) Qa2 2 2 µ no coupling term! 5.74 RWF Lecture #17 17 – 4 1/ 2 1 1 + G k k + ωs = rr ′ ( RL ) 2πc µ 1/ 2 1 1 − k − k G ωa = rr ′ ( RL ) 2πc µ simplify to ωs = ω0 + β + λ ωa = ω0 + β –λ kRL Grr ′ β = 2 ( ) 2 2πc ω 0 (algebra, not power series) Grr′ can have either sign. It is usually negative because φ > π/2. ω0 β λ= 1 − [ kRL k + µGrr ′ ] 2 ω0 Can have either sign. Positive if right bond gets stiffer when left bond is stretched. sign of λ determined by whether potential or kinetic coupling is larger (or by the signs of kRL and Grr′). 5.74 RWF Lecture #17 17 – 5 Morse Oscillator The Morse oscillator has a physically appropriate and mathematically convenient form. It turns out to give a vastly more convenient representation of an anharmonic vibration than V ( r) = 1 1 1 f rr x 2 + f rrr x 3 + f rrrr x 4 2 6 24 treated by perturbation theory. VMorse ( r) = De [1− e − ar ] (V (0) = 0, V (∞) = De ) 2 r = R − Re Power series expansion of VMorse ( r) = If we use ( ) ( ) ( ) 1 1 1 14a 4 De r 4 . 2a 2 De r 2 − 6a 3 De r 3 + 24 6 2 frr = 2a2De frrr = –6a3De frrrr = 14a4De in the framework of nondengenerate perturbation theory, we get much better results than we expect or deserve. Why? Because the energy levels of a Morse oscillator have a very simple form: E Morse (v ) hc = E 0Morse hc + ω m (v + 1 / 2) + x m (v + 1 / 2) 2 and an exact solution for the energy levels gives E 0Morse = 0 1 2a 2 De ωm = 2πc µ xm = − 1/ 2 µ= m1m2 m1 + m2 a2h 4πcµ ( ) we get the exact same relationship between (De,a) and E 0Morse ,ω m , x m by perturbation theory (with a twist) 5.74 RWF Lecture #17 17 – 6 1 1 2 P f rr r 2 + 2 2µ quartic term treated 1 (1) 4 to 1st order only! Ev = f vr v 24 rrrr 2 1 v − 1r 3 v − v + 1r 3 v v − 3r 3 v − v + 3r 3 v ( 2) f rrr + Ev = 6 ωm 1 3 H ( 0) = This works better than we could ever have hoped, and therefore we should never look a gift horse in the mouth. We always use Morse rather than an arbitrary power series representation of V(r). Sometimes we n even use a power series ∑ an [1− exp(−ar)] . n Armed with this simplification, consider two anharmonically coupled local stretch oscillators. WHY? What promotes or inhibits energy flow between two identical subsystems? * * * * ubiquitous Local and Normal Mode Pictures are opposite limiting cases Heff contains antagonistic terms that preserve and destroy limiting behavior eff the roles are reversed for H eff Local and H Normal See Section 9.4.12.3 of HLB-RWF Extremely complicated algebra HLocal defined identically to H Local, but with diagonal anharmonicity. Convert to dimensionless P, Q, H and then to a,a†. exploit the convenient V(Q) ↔ E(v) properties of VMorse. van Vleck transformation to account for the effect of out of polyad coupling terms from [Grr′PRPL + kRLQRQL] BUT NOT from VMorse. 5. Simplest possible fit model — relationships (constraints) between fit parameters imposed by the identical Morse oscillator model. 6. Next time — transformation from HLocal to HNormal. 1. 2. 3. 4. 5.74 RWF Lecture #17 17 – 7 ( ) () hR0 1. H Local hR1 1 2 1 2 = PR + Q R + V anh (Q R ) 2µ 2µ 1 2 1 2 + PL + Q L + V anh (Q L ) 2µ 2µ +Grr ′PR PL + kRL Q R Q L () 1 H RL 1 V anh (Q) = VMorse (Q) − kQ 2 2 This enables us to use Harmonic-Oscillators for basis set but Morse simplification for the separate local oscillators. We are going to expand Vanh(Q) and keep only the Q3 and Q4 terms and treat them, respectively, by secondorder and first-order perturbation theory, as we did for the simple Morse oscillator. ( ) ( ) () () () 1 H Local = hR0 + hL0 + hR1 + hL1 + H RL ( ) H0 5.74 RWF Lecture #17 2,3. 17 – 8 ˆ , Pˆ , H ˆ → a , a † , a ,a † Q, P, H → Q R R L L ˆ Qi = α i−1/ 2Q i P = hα 1/ 2Pˆ H i Local i i ˆ Local = h(2πcω M )H αi = 2πcωi µ i h note inconsistency between ωM in H and ωi in α 1 ki µ i ]1/ 2 ωi = [ 2πc ˆ = 2 −1/ 2 a + a † etc. Q R R R ( ) Pˆ R = 2 −1/ 2 i(a †R − a R ) etc. H Local = h(2πcω M )[ v R v L v R v L ]{(vR + 1 / 2) + (vL + 1 / 2) + F [(vR + 1 / 2)2 + (vL + 1 / 2)2 ]} 2 D + C + v R ± 1,v L m 1 v R v L v R + 1 / 2 ± 1 / 2)(v L + 1 / 2 m 1 / 2)] ( [ 2 1/ 2 D − C + v R ± 1,v L ± 1 v R v L v R + 1 / 2 ± 1 / 2)(v L + 1 / 2 ± 1 / 2)] ( [ 2 F=− 2 −1/ 2 ( ha) 4π(µDe ) 1/ 2 C = Grr ′µ k k D = RL = RL 2 kM 2De a dimensionless (a, De from Morse) dimensionless First 2 lines of HLocal are polyad, third line is out of polyad. 5.74 RWF Lecture #17 17 – 9 4. (D − C)2 F 2 2 eff ˆ H Local = v R v L v R v L (v R + v L + 1) 1 − + (v R + v L + 1) + (v R − v L ) 2 8 1/ 2 D + C + v R ± 1, v L m 1 v R v L v r + 1 / 2 ± 1 / 2)(v L + 1 / 2 m 1 / 2)] ( [ 2 [ ] F 2 v R − v L ) tries to preserve local mode limit. The D+C ( 2 coupling term tries to destroy the local mode limit. 2 Polyad P = vR + vL ( ) ( ) Overall width of polyad: E( P0 / 2,P / 2) − E(O0 ,P ) = − F 2 P 2 ( F < 0) (0,P) and (P,0) are at low energy extreme because of anharmonicity: ω(v + 1 / 2) − x (v + 1 / 2) . 2 Off-diagonal matrix elements are smallest between () H(10,P )(1,P − 1) = (0, P ) ~ (1, P − 1) and ( P, 0) ~ ( P − 11 ,) D + C 1/ 2 P 2 Off diagonal matrix elements are largest between ( P / 2, P / 2) ~ ( P / 2 − 1, P / 2 + 1) 1/ 2 D + C ( ) ( ) [ ] H( P / 2,P / 2)( P / 2 − 1,P / 2 +1) = P /2 P /2 +1 2 (1) 1/ 2 larger by a factor of [( P / 4 ) + 1 / 2] . 5.74 RWF Lecture #17 5. 17 – 10 General (minimal fit model) H eff Local hc = v R v L v R v L {ω R (v R + 1 / 2) + ω L (v L + 1 / 2) } + x R (v R + 1 / 2) + x L (v L + 1 / 2) + x RL (v R + 1 / 2)(v L + 1 / 2) 2 + v R ± 1,v L m 1 v R v L 2 {(H hc )[(v R + 1 / 2 ± 1 / 2)(v L + 1 / 2 m 1 / 2)] 1/ 2 RL But, in the two identical 1 : 1 coupled Morse local oscillator picture (D − C)2 ω R = ω L = ω M 1 − = ω′ 8 xR = xL = xM a2h =− 4πcµ x RL = 0 D + C H RL hc = ω M 2 }