18.2: Golden Rule Vibrational Relaxation Rates

Time Dependent Quantum Mechanics and Spectroscopy 2025e · LibreTexts

0%

Contents

    Text

    Size
    100%
    Theme
    Typeface
    Line length
    Spacing

    Read aloud

    How

    Speed: 1×
    Light up
    Highlight

    Symbols: Mulberry Symbols, © 2018–2026 Steve Lee, CC BY-SA 4.0 (licence, the set), the site's own copies.

    From LibreTexts, under CC BY-NC-SA 4.0, as the source states it; its equations are written in TeX, which the PDF shows typeset. This page at LibreTexts.

    18.2: Golden Rule Vibrational Relaxation Rates

    \( \let\bibref\eqref  %make-command_bibref-equivalent-to-eqref\)Let's treat the problem of a vibrational system \(H_S\) that relaxes through weak coupling \(V\) with a continuum of bath states \(H_B\) using perturbation theory. The eigenstates of \(H_S\) are \(|a\rangle\) and those of \(H_B\) are \(| \alpha \rangle\). Although our earlier perturbative treatment did not satisfy energy conservation, here we can take care of it by explicitly treating the bath states.

    \[\begin{gather}
    H=H_0+V \\[5pt]
    H_0=H_S+H_B \\[5pt]
    H_S=|a\rangle E_a\langle a|+| b\rangle E_b\langle b| \\[5pt]
    H_B=\sum_\alpha|\alpha\rangle E_\alpha\langle\alpha| \\
    H_0|a \alpha\rangle=\left(E_a+E_\alpha\right)|a \alpha\rangle
    \end{gather}\]

    Schematic diagram illustrating a quantum system and a bath, with states labeled and arrows indicating transitions.
    Figure \(\PageIndex{1}\): Illustration of the net energy conservation in the system and bath during vibrational relaxation of the system.

    We will describe transitions from an initial state \( |i\rangle=|a \alpha\rangle\) with energy \( E_a+E_\alpha\) to a final state \(|f\rangle=|b \beta\rangle\) with energy \(E_b+E_\beta\). Since we expect energy conservation to hold, this undoubtedly requires that a change in the system states will require an equal and opposite change of energy in the bath. Initially, we take \(p_a=1\) and \(p_b=0\). If the interaction potential is \(V\), Fermi's Golden Rule says the transition from \(|i\rangle\) to \(|f\rangle\) is given by

    \[\begin{aligned}
    k_{f i} & =\frac{2 \pi}{\hbar} \sum_{i, f} p_i \left|\langle i|V| f\rangle \vphantom{A^2} \right|^2 \delta\left(E_f-E_i\right) \\
    & =\frac{2 \pi}{\hbar} \sum_{a, \alpha, b, \beta} p_{a, \alpha} \left|\langle a \alpha|V| b \beta\rangle \vphantom{A^2} \right|^2 \goodbreak \delta \nobreak \left(\left(E_b+E_\beta\right)-\left(E_a+E_\alpha\right)\right) \\
    & =\frac{1}{\hbar^2} \int_{-\infty}^{+\infty} d t \! \! \sum_{\substack{a, \alpha, b, \beta}} p_{a, \alpha}\langle a \alpha|V| b \beta\rangle\langle b \beta|V| a \alpha\rangle \goodbreak \exp \left[\nobreak -\mathrm{i}\left(\left(E_b-E_a\right) \goodbreak +\left(E_\beta-E_\alpha\right)\right) t / \hbar \vphantom{A^2}\right]
    \end{aligned}\]

    Equation (18.2.8) is just a restatement of the time domain version of (18.2.6).

    \[k_{f i}=\frac{1}{\hbar^2} \int_{-\infty}^{+\infty} d t \, \langle V(t) V(0)\rangle \] \[V(t)=e^{\mathrm{i} H_0 t} V e^{-\mathrm{i} H_0 t}\]

    Now, the matrix element involves both evaluation in both the system and bath states, but if we write this in terms of a matrix element in the system coordinate \(V_{a b}=\langle a|V| b\rangle\):
    \[\langle a \alpha|V| b \beta\rangle=\left\langle\alpha\left|V_{a b}\right| \beta\right\rangle\]

    Then we can write the rate as
    \[\begin{aligned}
    k_{b a} &=\frac{1}{\hbar^2} \int_{-\infty}^{+\infty} d t  \, \sum_{\alpha, \beta} p_\alpha\left\langle\alpha\left|e^{+\mathrm{i} E_\alpha t} V_{a b} e^{-\mathrm{i} E_\beta t}\right| \beta\right\rangle\left\langle\beta\left|V_{b a}\right| \alpha\right\rangle e^{-\mathrm{i} \omega_{b a} t} \\
    &=\frac{1}{\hbar^2} \int_{-\infty}^{+\infty} d t\left\langle V_{a b}(t) V_{b a}(0)\right\rangle_B e^{-\mathrm{i} \omega_{a a} t}
    \end{aligned} \] \[ V_{a b}(t)=e^{\mathrm{i} H_B t} \ V_{a b} e^{-\mathrm{i} H_B t}\]

    Equation (18.2.12) says that the relaxation rate is determined by a correlation function

    \[C_{b a}(t)=\left\langle V_{a b}(t) V_{b a}(0)\right\rangle\]

    which describes the time-dependent changes to the coupling between \(b\) and \(a\). The time dependence of the interaction arises from the interaction with the bath; hence its time evolution under \(H_B\). The subscript \(\langle\cdots\rangle_B\) means an equilibrium thermal average over the bath states

    \[\langle\cdots\rangle_B=\sum_\alpha p_\alpha\langle\alpha|\cdots| \alpha\rangle\]

    Note also that eq. (18.2.12) is similar but not quite a Fourier transform. This expression says that the relaxation rate is given by the Fourier transform of the correlation function for the fluctuating coupling evaluated at the energy gap between the initial and final states. Alternatively, we could think of the rate in terms of a vibrational coupling spectral density, and the rate is given by its magnitude at the system energy gap \(\omega_{b a}\).
    \[k_{b a}=\frac{1}{\hbar^2} \tilde{C}_{b a}\left(\omega_{a b}\right)\]

    where the spectral representation \(\tilde{C}_{b a}(\omega_{a b})\) is defined as the Fourier transform of \(C_{b a}(t)\).

    Vibration coupled to a harmonic bath

    To evaluate these expressions, let's begin by considering the specific case of a system vibration coupled to a harmonic bath, which we will describe using a spectral density. Imagine that we prepare the system in an excited vibrational state, \(\left|1\right\rangle\), and we want to describe relaxation to the ground state \(\left|0\right\rangle\).\(\bibref{Cite1} %in-text-citation\)

    \[\begin{align}
    H_S &= \hbar \kern.05em \omega_0\left(\underset{^\sim}{P} \, ^2 + \underset{^\sim}{Q} \, ^2\right) \\
    H_B &= \sum_\alpha \hbar \, \omega_\alpha\left(\underset{^\sim}{p} \vphantom{l}_\alpha^2 + \underset{\phantom{.}^\sim}{q} \vphantom{l}_\alpha^2\right)=\sum_\alpha \hbar \, \omega_\alpha \left(a \vphantom{k}_\alpha^{\dagger} a\vphantom{g}_\alpha+\frac{1}{2}\right)
    \end{align}\]

    We will take the system-bath interaction to be linear in the bath coordinates:

    \[V=H_{S B}=\sum_\alpha c \vphantom{g}_\alpha \underset{^\sim}{Q} \underset{^\sim} {\alpha} \vphantom{g}_\alpha\]

    Here \(c_\alpha\) is a coupling constant which describes the strength of the interaction between the system and bath mode \(\alpha\). Note that this form suggests that the system vibration is a local mode interacting with a set of normal vibrations of the bath.

    Diagram showing a system labeled 'Q' with arrows pointing to four variables (qα, qβ, qγ, qδ) under a 'Bath' heading.

    Figure \(\PageIndex{2}\): Illustration of the system coordinate \(Q\) coupled bilinearly to a bath of harmonic modes, \(q_i\).

    For the case of single quantum relaxation from \(|a\rangle=|1\rangle\) to \(b=|0\rangle\), we can write the coupling matrix element as
    \[V_{b a}=\langle b|V| a\rangle=\sum_\alpha \xi_\alpha\left(a \vphantom{k}_\alpha^{\dagger}+a \vphantom{g}_\alpha\right)\]

    where \(\xi\) is a generalized coupling constant that depends explicitly on the the matrix element \(\langle b|Q| a\rangle\):
    \[\xi_\alpha=c_\alpha \frac{\sqrt{ \vphantom{N} m_{_Q} m_q \omega_0 \omega_\alpha}}{2 \hbar}\langle b|Q| a\rangle\]

    Evaluating eq. (18.2.14) is now much the same as problems we've had previously:

    \[\begin{aligned}
    \left\langle V_{a b}(t) V_{b a}(0)\right\rangle_B & =\left\langle e^{\mathrm{i} H_B t} \ V_{a b} \ e^{-\mathrm{i} H_B t} \ V_{b a}\right\rangle_B \\
    & =\sum_\alpha \xi \vphantom{A}_\alpha^2 \left[\left(\bar{n}_\alpha+1\right) e^{-\mathrm{i} \omega_\alpha t}+\bar{n}_\alpha e^{+\mathrm{i} \omega_\alpha t}\right]
    \end{aligned}\]

    here \(\bar{n}_\alpha=\left(e^{\beta \hbar \kern.05em \omega_\alpha}-1\right)^{-1}\) is the thermally averaged occupation number of the bath mode at \(\omega_\alpha\) In evaluating this we take advantage of relationships we have used before

    \[\begin{aligned}
    e^{i H_B t} a \vphantom{g}_\alpha e^{-\mathrm{i} H_B t} & =a\vphantom{g}_\alpha e^{-\mathrm{i} \omega_\alpha t} \\
    e^{\mathrm{i} H_B t} a \vphantom{k}_\alpha^{\dagger} e^{-\mathrm{i} H_B t} & =a \vphantom{k}_\alpha^{\dagger} e^{+i \omega_\alpha t}
    \end{aligned}
    \]\[
    \begin{aligned}
     \left\langle a\vphantom{g}_\alpha a \vphantom{k}_\alpha^{\dagger}\right\rangle & =\bar{n}\vphantom{g}_\alpha+1 \\
     \left\langle a \vphantom{k}_\alpha^{\dagger} a\vphantom{g}_\alpha\right\rangle &=\bar{n}\vphantom{g}_\alpha
    \end{aligned}\]

    So, now by Fourier transforming (18.2.22) we have the rate as

    \[k_{b a}=\frac{1}{\hbar^2} \sum_\alpha \xi \vphantom{A}_\alpha^2 \left[\vphantom{X^2}\right. \underbrace{\left(\bar{n}_\alpha+1\right) \delta\left(\omega_{b a}+\omega_\alpha\right)}_{\text{downward}} + \underbrace{\bar{n}_\alpha \delta\left(\omega_{b a}-\omega_\alpha\right)}_{\text{upward}} \left.\vphantom{X^2}\right]\]

    Diagram showing "downward" and "upward" transitions in quantum states with arrows indicating movement between states |a>, |b>, |c>, and |β>.
    Figure \(\PageIndex{3}\): Diagram of the net rate of energy relaxation as a sum of contributions from the upward and downward transition rates.

    This expression describes two relaxation processes which depend on temperature. The first "downward" process obeys \(-\omega_{b a} = \omega_{a}\) and becomes the dominant term as \(T \to 0 \, \text{K}\). This implies that \(E_a>E_b\) and that a loss of energy in the system is balanced by an equal rise in energy of the bath. That is \(|\beta\rangle=|\alpha+1\rangle\). The second "upward" term is only allowed for elevated temperatures in which thermal energy remains in the bath states. It describes relaxation of the system by transfer to a higher energy state \(E_b>E_a\), with a concerted decrease of the energy of the bath \((|\beta\rangle=|\alpha-1\rangle)\). Naturally, this process vanishes if there is no thermal energy in the bath.\(\bibref{Cite2} %in-text-citation\)

    To more accurately model the relaxation due to a continuum of modes, we can replace the explicit sum over bath states with an integral over a density of bath states \(W\)
    \[k_{b a}=\frac{1}{\hbar^2} \int d \omega_\alpha \, W(\omega_\alpha) \ \xi\vphantom{A}_\alpha^2(\omega_\alpha) \left[\left(\bar{n}(\omega_\alpha)+1\right) \delta(\omega_{b a}+\omega_\alpha) \ + \ \bar{n}(\omega_\alpha) \ \delta(\omega_{b a}-\omega_\alpha) \vphantom{X^2} \right]\]

    We can also define a spectral density, which is the vibrational coupling-weighted density of states:

    \[\rho_{b a}(\omega_\alpha) \equiv W(\omega_\alpha) \, \xi \vphantom{A}_\alpha^2(\omega_\alpha)\]

    Then the relaxation rate is:

    \[\begin{aligned}
    k_{b a} & =\frac{1}{\hbar^2} \int d \omega_\alpha \, W(\omega_\alpha) \xi_\alpha^2(\omega_\alpha) \left[\left(\bar{n}(\omega_\alpha\right)+1 ) \delta\left(\omega_{b a}+\omega_\alpha\right)+\bar{n}(\omega_\alpha) \ \delta\left(\omega_{b a}-\omega_\alpha\right)\right] \\
    & =\frac{1}{\hbar^2}\left[\left(\bar{n}(\omega_{b a})+1\right) \rho_{b a}(\omega_{a b})+\bar{n}(\omega_{b a}) \ \rho_{b a}(-\omega_{a b}) \vphantom{X^2}\right]
    \end{aligned}\]

    We see that the Fourier transform of the fluctuating coupling correlation function in eq. (18.2.22) is equivalent to the coupling-weighted density of states, \(\rho_{ba}\), which we evaluate at \(\omega_{ba}\) or \(-\omega_{ba}\) depending on whether we are looking at upward or downward transitions. Note that \(\bar{n}\) still refers to the occupation number for the bath, although it is evaluated at the energy splitting between the initial and final system states, \(\bar{n}(\omega_{ba})\). Equation (18.2.27) is a full quantum expression, and obeys detailed balance between the upward and downward rates of transition between two states:
    \[k_{b a}=\exp \left(-\beta \hbar \kern.05em \omega_{a b}\right) k_{a b}\]

    From our description of the two level system in a harmonic bath, we see that high frequency relaxation (\(k_BT\ll\hbar\omega_0\)) only proceeds with energy from the system going into a mode of the bath at the same frequency, but at lower frequencies (\(k_B T \approx \hbar \kern.05em \omega_0\)) that energy can flow both from the system into the bath and also from the bath back into the system. When the vibration has energies that are thermally populated in the bath, we return to the classical picture of a vibration in a fluctuating environment, in which the environment can both increase and dissipate the energy of the vibration.

    Note that, in a cascaded relaxation scheme like that shown in figure (18.1.1), as the excess energy of the system approaches \(k_B T\), the fraction of transitions that raise the system energy increases. Also note, the bi-linear coupling in eq. (18.2.19)—used again in our treatment of quantum fluctuations—can be associated with both changes in the bath's energy (relaxation) and shifts in its frequency (dephasing).


    References

    \[\tag
    {1}
    \label
    {Cite1}\]
    Note that we are using an equilibrium property, the coupling correlation function, to describe a nonequilibrium process, the relaxation of an excited state. Underlying the validity of the expressions are the principles of linear response. In practice this also implies a time scale separation between the equilibration of the bath and the relaxation of the system state. The bath correlation function should work fine if it has rapidly equilibrated, even though the system may not have. An instance where this would work well is electronic spectroscopy, where relaxation and thermalization in the excited state occurs on picosecond time scales, whereas the electronic population relaxation is on nanosecond time scales.
    \[\tag
    {2}
    \label
    {Cite2}\]
    There is an exact analogy between this problem and the interaction of matter with a quantum radiation field. The interaction potential is instead a quantum vector potential and the bath is the photon field of different electromagnetic modes. Equation (18.2.25) describes has two terms that describe emission and absorption processes. The leading term describes the possibility of spontaneous emission, where a material system can relax in the absence of light by emitting a photon at the same frequency.

    This page titled 18.2: Golden Rule Vibrational Relaxation Rates was last modified on Fri, 05 Jun 2026 02:03:57 GMT and is shared under a CC BY-NC-SA 4.0 license and was authored, remixed, and/or curated by Andrei Tokmakoff via source content that was edited to the style and standards of the LibreTexts platform.