Time-Dependent Quantum Mechanics and Spectroscopy Andrei Tokmakoff 417351 800107 5 Contents TIME‐DEPENDENT QUANTUM MECHANICS 1. INTRODUCTION 1. Time‐Evolution with a Time‐Independent Hamiltonian 2. Exponential Operators 3. Two‐Level System THE TIME‐DEPENDENT HAMILTONIAN 2. TIME‐EVOLUTION OPERATOR 1. Time‐Evolution Operator 2. IntegratinG the TDSE Directly 3. Transitions Induced by Time‐Dependent Potential 4. Resonant Driving of a Two‐Level System 5. Schrödinger and HeisenberG Representations 6. Interaction Picture 7. Time‐Dependent Perturbation Theory 8. Fermi’s Golden Rule 3. IRREVERSIBLE RELAXATION 4. THE DENSITY MATRIX 5. ADIABATIC APPROXIMATION 1. Born–Oppenheimer Approximation 2. Nonadiabatic Effects 3. Diabatic and Adiabatic States 4. Adiabatic and Nonadiabatic dynamics 5. Landau–Zener Transition Probability 6. INTERACTION OF LIGHT AND MATTER 1. Electric Dipole Hamiltonian 2. Supplement: Review of Free Electromagnetic Field 3. Absorption Cross Section CONCEPTS AND TOOLS FOR CONDENSED PHASE DYNAMICS 7. MIXED STATES AND THE DENSITY MATRIX 8. IRREVERSIBLE AND RANDOM PROCESSES 1. Concepts and Definitions 2. Thermal Equilibrium 3. Fluctuations 9. TIME‐CORRELATION FUNCTIONS 1. Definitions, Properties, and Examples 2. Correlation Function from a Discrete Trajectory 3. Quantum Time‐Correlation Functions 4. Transition Rates from Correlation Functions 10. LINEAR RESPONSE THEORY 1. Classical Linear Response Theory 2. Quantum Linear Response Functions 3. The Response Function and Energy Absorption 4. Relaxation of a Prepared State SPECTROSCOPY 11. TIME‐DOMAIN DESCRIPTION OF SPECTROSCOPY 1. A Classical Description of Spectroscopy 2. Time‐Correlation Function Description of Absorption Lineshape 3. Different Types of Spectroscopy Emerge from the Dipole Operator 4. Ensemble Averaging and Line‐Broadening 12. COUPLING OF ELECTRONIC AND NUCLEAR MOTION 1. The Displaced Harmonic Oscillator Model 2. Coupling to a Harmonic Bath 3. Semiclassical Approximation to the Dipole Correlation Function 13. FLUCTUATIONS IN SPECTROSCOPY 1. Fluctuations and Randomness: Some Definitions 2. Line‐Broadening and Spectral Diffusion 3. Gaussian‐Stochastic Model for Spectral Diffusion 4. The Energy Gap Hamiltonian 5. Correspondence of Harmonic Bath and Stochastic Equations of Motion APPLICATIONS 14. ENERGY AND CHARGE TRANSFER 1. Electronic Interactions 2. Förster Resonance EnerGy Transfer 3. Excitons in Molecular Aggregates) 4. Multiple Excitations and Second Quantization 5. Marcus Theory for Electron Transfer 15. QUANTUM RELAXATION PROCESSES 1. Vibrational Relaxation 2. A Density Matrix Description of Quantum Relaxation 1. INTRODUCTION 1.1. Time Evolution with a Time-Independent Hamiltonian The time evolution of the state of a quantum system is described by the time-dependent Schrödinger equation (TDSE): irtHrtrt ,,, ˆ (1.1) t Hˆ is the Hamiltonian operator which describes all interactions between particles and fields, and determines the state of the system in time and space. Hˆ is the sum of the kinetic and potential energy. For one particle under the influence of a potential 2 HVrtˆˆˆ 2 , (1.2) 2m The state of the system is expressed through the wavefunction rt, . The wavefunction is complex and cannot be observed itself, but through it we obtain the probability density 2 Prt , which characterizes the spatial probability distribution for the particles described by Hˆ at time t. Also, it is used to calculate the expectation value of an operator Aˆ Atˆˆ * r,, t A r tdr t A ˆ t (1.3) Physical observables must be real, and therefore will correspond to the expectation values of Hermitian operators AAˆˆ † . Our first exposure to time-dependence in quantum mechanics is often for the specific case in which the Hamiltonian Hˆ is assumed to be independent of time: HHrˆˆ . We then assume a solution with a form in which the spatial and temporal variables in the wavefunction are separable: rt, rTt (1.4) 1 Hrˆ r iTt (1.5) Tt t r Here the left-hand side is a function only of time, and the right-hand side is a function of space only ( r , or rather position and momentum). Equation (1.5) can only be satisfied if both sides are equal to the same constant, E . Taking the right-hand side we have Hrˆ r EHˆ rrEr (1.6) r Andrei Tokmakoff, 12/1/2014 1-2 This is the Time-Independent Schrödinger Equation (TISE), an eigenvalue equation, for which r are the eigenstates and E are the eigenvalues. Here we note that HHEˆˆ , so Hˆ is the operator corresponding to E and drawing on classical mechanics we associate Hˆ with the expectation value of the energy of the system. Now taking the left-hand side of (1.5) and integrating: 1 TiE iE Tt 0 (1.7) Tt t t Tt exp iEt / (1.8) So, in the case of a bound potential we will have a discrete set of eigenfunctions n r with corresponding energy eigenvalues En from the TISE, and there are a set of corresponding solutions to the TDSE. nnrt,exp/ r iEt n (1.9) Since the only time-dependence is a phase factor, the probability density for an eigenstate is 2 independent of time: Pt n = constant. Therefore, the eigenstates r do not change with time and are called stationary states. However, more generally, a system may exist as a linear combination of eigenstates: iEn t/ rt,, cnn rt cnne r (1.10) nn 2 where c are complex amplitudes, with c 1. For such a case, the probability density will n n n oscillate with time. As an example, consider two eigenstates iE12 t//iE t rt, 1211 c e c 22 e (1.11) For this state the probability density oscillates in time as 22 Pt12 22 **it21 **it21 cccc11 2 2 1 21 2e cce2121 (1.12) 22 122cos() 1221 t where nn E / . We refer to this state of the system that gives rise to this time-dependent oscillation in probability density as a coherent superposition state, or coherence. More generally, the oscillation term in eq. (1.12) may also include a time-independent phase factor that arises from the complex expansion coefficients. 1-3 As an example, consider the superposition of the ground and first excited states of the quantum harmonic oscillator. The basis wavefunctions, 0 ()x and 1()x , and their stationary probability densities Pxxii ( ) | i ( ) are If we create a superposition of these states with eqn. (1.11), the time-dependent probability density oscillates, with x(t) bearing similarity to the classical motion. (Here c0 = 0.5 and c1 = 0.87.) P(x,t) x 1-4 Readings 1. Cohen-Tannoudji, C.; Diu, B.; Lalöe, F., Quantum Mechanics. Wiley-Interscience: Paris, 1977; p. 405. 2. Nitzan, A., Chemical Dynamics in Condensed Phases. Oxford University Press: New York, 2006; Ch. 1. 3. Schatz, G. C.; Ratner, M. A., Quantum Mechanics in Chemistry. Dover Publications: Mineola, NY, 2002; Ch. 2. 1-5 1.2. Exponential Operators Throughout our work, we will make use of exponential operators of the form Teˆ iAˆ We will see that these exponential operators act on a wavefunction to move it in time and space. ˆ Of particular interest to us is the time-propagator or time-evolution operator, Ueˆ iHt/ , which propagates the wavefunction in time. Note the operator Tˆ is a function of an operator, f ()Aˆ . A function of an operator is defined through its expansion in a Taylor series, for instance n ˆ ()iAˆˆˆ AA Teˆ iA 1 iAˆ (1.13) n0 n!2 Since we use them so frequently, let’s review the properties of exponential operators that can be established with eq. (1.13). If the operator Aˆ is Hermitian, then Teˆ iAˆ is unitary, i.e., TTˆˆ†1 . Thus the Hermitian conjugate of Tˆ reverses the action of Tˆ . For the time-propagator Uˆ , Uˆ † is often referred to as the time-reversal operator. The eigenstates of the operator Aˆ also are also eigenstates of f ()Aˆ , and eigenvalues are functions of the eigenvalues of Aˆ . Namely, if you know the eigenvalues and eigenvectors of Aˆ , ˆ i.e., Annn a , you can show by expanding the function that ˆ fA nnn fa (1.14) Our most common application of this property will be to exponential operators involving the ˆ Hamiltonian. Given the eigenstates n , then HEnnn implies ˆ iHt/ iEn t eenn (1.15) Just as Ueˆ iHtˆ / is the time-evolution operator, which displaces the wavefunctionin ˆ ˆ ipx x/ time, Dex is the spatial displacement operator that moves along the x coordinate. If we ˆ ˆ ipx / define Dex () , then the action of is to displace the wavefunction by an amount . ˆ x Dxx () (1.16) ˆ Also, applying Dx () to a position operator shifts the operator by ˆˆ† DxDxxˆˆ x (1.17) ˆ Thus eipx x is an eigenvector of xˆ with eigenvalue x + instead of x. The operator ˆ ipˆ ˆ ipx ˆ y Dex is a displacement operator for x position coordinates. Similarly, Dey ˆ ˆ generates displacements in y and Dz in z. Similar to the time-propagator U , the displacement 1-6 operatorDˆ must be unitary, since the action of Dˆˆ†D must leave the system unchanged. That is if ˆ ˆ † D shifts the system to x from x0 , then D shifts the system from x back to x0 . We know intuitively that linear displacements commute. For example, if we wish to shift a particle in two dimensions, x and y, the order of displacement does not matter. We end up at the same position. These displacement operators commute, as expected from [px,py] = 0. ibp first move along x by a, xy eey iapx xy 22 11 then along y by B. iapx ibpy ee xy11 reverse order Similar to the displacement operator, we can define rotation operators that depend on the ˆ iL x angular momentum operators, Lx, Ly, and Lz.