Physics note

EIT and dark states

Topic: Dark States and their consequences

Introduction

When coherent laser fields drive a multi-level atom, quantum interference between different excitation pathways produces remarkable effects: dark states, coherent population trapping, and electromagnetically induced transparency. These phenomena arise from the same quantum mechanical principles but manifest in different regimes, and they demonstrate how the light-matter interaction can be controlled with precise temporal and spatial shaping of the driving fields.

1. The Three-Level $\Lambda$-System and the Dark State

1.1 The Interaction Hamiltonian

We consider a three-level atom in the $\Lambda$-configuration, as introduced in the previous notes on master equations. Two ground states $| 1 \rangle$ and $| 2 \rangle$ (with energies $E_1$ and $E_2$, respectively) are coupled by electric dipole transitions to a common excited state $| 3 \rangle$ (energy $E_3$). The transition $| 1 \rangle \leftrightarrow | 2 \rangle$ is dipole-forbidden. Two classical, monochromatic laser fields drive the transitions:

In the rotating frame and under the rotating wave approximation (RWA), the effective Hamiltonian is (see previous notes for the full derivation): \begin{equation} \hat{\tilde{H}} = -\hbar \Delta_p | 3 \rangle\langle 3 | - \hbar(\Delta_p - \Delta_c) | 2 \rangle\langle 2 | + \frac{\hbar}{2} \left( \Omega_p | 3 \rangle\langle 1 | + \Omega_c | 3 \rangle\langle 2 | + \text{H.c.} \right), \label{eq:H_Lambda_rot} \end{equation} where $\Delta_p = \omega_p - \omega_{31}$ and $\Delta_c = \omega_c - \omega_{32}$ are the one-photon detunings, and \begin{equation} \delta = \Delta_p - \Delta_c = \omega_p - \omega_c - \omega_{21} \end{equation} is the two-photon (Raman) detuning, with $\hbar\omega_{21} = E_2 - E_1$.

1.2 The Condition for Two-Photon Resonance

We now focus on the case of two-photon resonance, $\delta = 0$, which means $\Delta_p = \Delta_c \equiv \Delta$. This is the condition under which the two-photon transition $| 1 \rangle \to | 3 \rangle \to | 2 \rangle$ is resonant even if the individual one-photon transitions are detuned. The Hamiltonian simplifies to \begin{equation} \hat{\tilde{H}}_{\delta=0} = -\hbar\Delta | 3 \rangle\langle 3 | + \frac{\hbar}{2} \left( \Omega_p | 3 \rangle\langle 1 | + \Omega_c | 3 \rangle\langle 2 | + \text{H.c.} \right). \label{eq:H_delta0} \end{equation}

1.3 The Dark State: Definition and Properties

We seek an eigenstate $| D \rangle$ of $\hat{\tilde{H}}_{\delta=0}$ that has zero eigenvalue and contains no component of the excited state $| 3 \rangle$. Such a state, if it exists, will be completely decoupled from the radiation fields: it cannot absorb photons (since it has no overlap with $| 3 \rangle$) and therefore cannot fluoresce. This is the dark state.

Writing $| D \rangle = \alpha | 1 \rangle + \beta | 2 \rangle + \gamma | 3 \rangle$ and imposing $\hat{\tilde{H}}_{\delta=0} | D \rangle = 0$, we obtain: \begin{align} \frac{\hbar}{2} \Omega_p^* \gamma &= 0, \label{eq:dark1} \\ \frac{\hbar}{2} \Omega_c^* \gamma &= 0, \label{eq:dark2} \\ \frac{\hbar}{2} \Omega_p \alpha + \frac{\hbar}{2} \Omega_c \beta - \hbar\Delta \gamma &= 0. \label{eq:dark3} \end{align} Equations (\ref{eq:dark1}) and (\ref{eq:dark2}) are satisfied if $\gamma = 0$ (no excited-state component). Equation (\ref{eq:dark3}) then requires \begin{equation} \Omega_p \alpha + \Omega_c \beta = 0. \end{equation} Together with the normalization condition $|\alpha|^2 + |\beta|^2 = 1$, this yields \begin{equation} \boxed{| D \rangle = \frac{\Omega_c | 1 \rangle - \Omega_p | 2 \rangle}{\sqrt{|\Omega_p|^2 + |\Omega_c|^2}}}. \label{eq:dark_state} \end{equation} This is the dark state. It is a coherent superposition of the two ground states, with relative amplitude and phase determined by the two Rabi frequencies. The dark state exists only under the condition of two-photon resonance ($\delta = 0$). For $\delta \neq 0$, the Hamiltonian has no zero-energy eigenstate orthogonal to $| 3 \rangle$.

The dark state is remarkable for a few reasons. It is an exact eigenstate of the interacting Hamiltonian with zero eigenvalue, so it is perfectly stationary and does not evolve in time. Its ground-state amplitudes are set by the two Rabi frequencies, and if those fields are varied slowly, the state follows them adiabatically. Most importantly, it has no overlap with the excited state, so the atom cannot scatter photons and the medium becomes transparent to the driving fields.

1.4 The Bright State

Orthogonal to the dark state, within the two-dimensional ground-state subspace, is the bright state: \begin{equation} | B \rangle = \frac{\Omega_p^* | 1 \rangle + \Omega_c^* | 2 \rangle}{\sqrt{|\Omega_p|^2 + |\Omega_c|^2}}. \label{eq:bright_state} \end{equation} The bright state couples maximally to the excited state. The interaction Hamiltonian in the $\{ | B \rangle, | D \rangle, | 3 \rangle \}$ basis takes the simple form \begin{equation} \hat{\tilde{H}} = \frac{\hbar}{2} \begin{pmatrix} 0 & 0 & \Omega_{\text{eff}} \\ 0 & 0 & 0 \\ \Omega_{\text{eff}}^* & 0 & -2\Delta \end{pmatrix}, \label{eq:H_bright_dark} \end{equation} where $\Omega_{\text{eff}} = \sqrt{|\Omega_p|^2 + |\Omega_c|^2}$ is the effective total Rabi frequency. The dark state is completely decoupled (entire row and column of zeros), while the bright state behaves like a two-level system coupled to the excited state with Rabi frequency $\Omega_{\text{eff}}$. The three-level $\Lambda$-system has thus been reduced to an effective two-level system (the bright-state–excited-state subsystem) plus a completely decoupled dark state. This decomposition is profoundly useful for understanding CPT, EIT, and STIRAP.

[Figure 1: (a) The three-level $\Lambda$-system with probe $\Omega_p$ and control $\Omega_c$ fields. (b) The dark state $| D \rangle$ and bright state $| B \rangle$ in the ground-state subspace. The bright state couples to $| 3 \rangle$ with effective Rabi frequency $\Omega_{\text{eff}}$; the dark state is completely decoupled. (c) The dressed-state picture in the $\{ | B \rangle, | 3 \rangle \}$ subspace, showing the Autler–Townes doublet and the location of the dark state at zero energy.]

2. Coherent Population Trapping (CPT)

2.1 The CPT Phenomenon

Coherent population trapping (CPT) is the phenomenon in which an atom, initially prepared in (or pumped into) the dark state $| D \rangle$, remains trapped there indefinitely. Because the dark state has no excited-state component, the atom does not scatter photons, even though both laser fields are present and resonant (or near-resonant). The population is "trapped" in the two ground states, and the excited-state population is identically zero in the steady state. CPT was first observed experimentally by Alzetta et al. in 1976 in sodium vapor and has since become a cornerstone of quantum optics.

2.2 Steady-State Solution of the Optical Bloch Equations for CPT

To see CPT emerge from the master equation, we consider the optical Bloch equations for the three-level $\Lambda$-system (derived in the previous notes). For simplicity, we assume equal decay rates from the excited state to the two ground states: $\Gamma_{31} = \Gamma_{32} = \Gamma/2$, so that the total spontaneous emission rate is $\Gamma$. We also neglect ground-state dephasing ($\gamma_{12} = 0$). The equations for the populations and coherences under the conditions $\Delta_p = \Delta_c = \Delta$ and $\delta = 0$ are: \begin{align} \dot{\rho}_{33} &= -\Gamma \rho_{33} - \frac{i}{2}(\Omega_p \rho_{13} - \Omega_p^* \rho_{31}) - \frac{i}{2}(\Omega_c \rho_{23} - \Omega_c^* \rho_{32}), \\ \dot{\rho}_{11} &= +\frac{\Gamma}{2} \rho_{33} + \frac{i}{2}(\Omega_p \rho_{13} - \Omega_p^* \rho_{31}), \\ \dot{\rho}_{22} &= +\frac{\Gamma}{2} \rho_{33} + \frac{i}{2}(\Omega_c \rho_{23} - \Omega_c^* \rho_{32}), \\ \dot{\rho}_{13} &= -\left(\frac{\Gamma}{2} + i\Delta\right) \rho_{13} + \frac{i}{2}\Omega_p^*(\rho_{33} - \rho_{11}) - \frac{i}{2}\Omega_c^* \rho_{12}, \\ \dot{\rho}_{23} &= -\left(\frac{\Gamma}{2} + i\Delta\right) \rho_{23} + \frac{i}{2}\Omega_c^*(\rho_{33} - \rho_{22}) - \frac{i}{2}\Omega_p^* \rho_{21}, \\ \dot{\rho}_{12} &= -i\delta \rho_{12} + \frac{i}{2}\Omega_p \rho_{32} - \frac{i}{2}\Omega_c^* \rho_{13} = \frac{i}{2}\Omega_p \rho_{32} - \frac{i}{2}\Omega_c^* \rho_{13} \quad (\text{since } \delta = 0). \end{align}

In the steady state, all time derivatives vanish. It is straightforward to verify that the following density matrix is an exact steady-state solution: \begin{equation} \hat{\rho}_{\text{CPT}} = \frac{1}{|\Omega_p|^2 + |\Omega_c|^2} \begin{pmatrix} |\Omega_c|^2 & -\Omega_p^* \Omega_c & 0 \\ -\Omega_p \Omega_c^* & |\Omega_p|^2 & 0 \\ 0 & 0 & 0 \end{pmatrix}. \label{eq:CPT_rho} \end{equation} This is precisely the pure-state density matrix $\hat{\rho}_{\text{CPT}} = | D \rangle\langle D |$, where $| D \rangle$ is the dark state (\ref{eq:dark_state}). The excited-state population is zero, and the atom is in a pure coherent superposition of the two ground states. This solution is independent of the detuning $\Delta$ (as long as $\delta = 0$) and of the absolute intensities of the fields (only their ratio matters). It exists even when the one-photon transitions are far off resonance ($|\Delta| \gg \Gamma$).

2.3 Experimental Signatures of CPT

The hallmark of CPT is a sharp dip in the fluorescence (or an increase in transmission) when the two-photon resonance condition $\delta = 0$ is satisfied. In a typical experiment, the control field frequency is fixed, and the probe field frequency is scanned. When $\omega_p - \omega_c = \omega_{21}$ (so $\delta = 0$), the atoms are pumped into the dark state, fluorescence drops dramatically, and the probe transmission increases. The width of the CPT resonance is determined by the ground-state decoherence rate, which can be extremely narrow (sub-kHz in buffer-gas cells, or even sub-Hz in cold atoms), making CPT a powerful tool for precision spectroscopy and atomic clocks.

[Figure 2: (a) The CPT resonance: fluorescence from a $\Lambda$-system as a function of the two-photon detuning $\delta$, showing a sharp dip at $\delta = 0$. (b) Energy-level diagram indicating the two-photon resonance condition. (c) Experimental CPT spectrum in $^{87}$Rb vapor, showing the narrow transparency window.]

3. Electromagnetically Induced Transparency (EIT)

3.1 From CPT to EIT: The Weak-Probe Limit

Electromagnetically induced transparency (EIT) is the phenomenon in which a medium that is normally opaque to a weak probe beam becomes transparent when a second, strong control beam is applied to a coupled transition. EIT is closely related to CPT but is distinct in an important way: in CPT, both fields are typically of comparable intensity, and the phenomenon is studied in the fluorescence spectrum. In EIT, one field (the control) is strong and the other (the probe) is weak, and the effect is observed in the probe's absorption and dispersion.

We consider the weak-probe limit: $|\Omega_p| \ll |\Omega_c|, \Gamma$. The control field is treated to all orders, while the probe is treated to first order in perturbation theory. The initial condition (zeroth order in $\Omega_p$) has all population in state $| 1 \rangle$: $\rho_{11}^{(0)} = 1$, $\rho_{22}^{(0)} = \rho_{33}^{(0)} = 0$, and all coherences vanish.

3.2 Linear Susceptibility of the EIT Medium

The probe field interacts with the atomic polarization on the $| 1 \rangle \leftrightarrow | 3 \rangle$ transition. The linear susceptibility $\chi_p(\omega_p)$ is proportional to the coherence $\rho_{13}$ induced by the probe, divided by the probe Rabi frequency: \begin{equation} \chi_p(\omega_p) = \frac{2 \mathcal{N} |\mathbf{d}_{31}|^2}{\varepsilon_0 \hbar \Omega_p} \rho_{13}, \end{equation} where $\mathcal{N}$ is the atomic density. Solving the optical Bloch equations to first order in $\Omega_p$ in the steady state yields the probe coherence: \begin{equation} \rho_{13}^{(1)} = \frac{i\Omega_p/2}{\dfrac{\Gamma}{2} + i\Delta_p + \dfrac{|\Omega_c|^2/4}{\gamma_{12} + i\delta}}. \label{eq:rho13_EIT} \end{equation} This is the central formula of EIT. The term in the denominator, \begin{equation} \frac{|\Omega_c|^2/4}{\gamma_{12} + i\delta}, \end{equation} represents the control-field-induced modification of the probe response. The linear susceptibility is therefore \begin{equation} \chi_p(\Delta_p, \delta) = \frac{i \mathcal{N} |\mathbf{d}_{31}|^2}{\varepsilon_0 \hbar} \, \frac{1}{\dfrac{\Gamma}{2} + i\Delta_p + \dfrac{|\Omega_c|^2/4}{\gamma_{12} + i\delta}}. \label{eq:chi_EIT} \end{equation}

3.3 The EIT Window: Absorption and Dispersion

The absorption coefficient for the probe is proportional to the imaginary part of the susceptibility: $\alpha_p \propto \operatorname{Im}[\chi_p]$. The refractive index is proportional to the real part: $n_p - 1 \propto \operatorname{Re}[\chi_p]$. Let us examine these on two-photon resonance ($\delta = 0$). In the ideal case of no ground-state dephasing ($\gamma_{12} = 0$), we obtain: \begin{equation} \chi_p(\Delta_p, \delta=0) = \frac{i \mathcal{N} |\mathbf{d}_{31}|^2}{\varepsilon_0 \hbar} \, \frac{1}{\dfrac{\Gamma}{2} + i\Delta_p + \dfrac{|\Omega_c|^2/4}{-i0}}. \end{equation} The term $|\Omega_c|^2/(4 \times 0)$ diverges, forcing the denominator to infinity and the susceptibility to zero. At the exact line center ($\Delta_p = 0$), the absorption vanishes completely: the medium becomes perfectly transparent. This is the EIT window.

For a more realistic analysis with finite $\gamma_{12}$, and setting $\Delta_p = \Delta_c \equiv \Delta$ (so $\delta = 0$), the susceptibility reduces to \begin{equation} \chi_p(\Delta) = \frac{i \mathcal{N} |\mathbf{d}_{31}|^2}{\varepsilon_0 \hbar} \, \frac{1}{\dfrac{\Gamma}{2} + i\Delta + \dfrac{|\Omega_c|^2/4}{\gamma_{12}}}. \end{equation} The absorption profile is a Lorentzian with a reduced peak value and a power-broadened width: \begin{equation} \alpha(\Delta) \propto \frac{\Gamma/2 + |\Omega_c|^2/(4\gamma_{12})}{\Delta^2 + \left(\Gamma/2 + |\Omega_c|^2/(4\gamma_{12})\right)^2}. \end{equation} The control field effectively broadens the transition, but the key point is that the peak absorption decreases with increasing $\Omega_c$. In the limit $\gamma_{12} \to 0$, the transparency becomes perfect at $\Delta = 0$.

The dispersion $\operatorname{Re}[\chi_p]$ is also dramatically modified. At $\Delta = 0$, the dispersion is zero. However, the slope $\mathrm{d}(\operatorname{Re}[\chi_p])/\mathrm{d}\Delta$ at $\Delta = 0$ is greatly enhanced in the EIT window compared to a regular two-level absorption line. This steep normal dispersion leads to a dramatic reduction of the group velocity of light pulses—the phenomenon of slow light—which has been observed with group velocities as low as a few meters per second.

[Figure 3: (a) The EIT window: probe absorption (Im$[\chi]$, blue) and dispersion (Re$[\chi]$, red) as functions of the probe detuning $\Delta_p$ for $\delta = 0$. The absorption vanishes at resonance, while the dispersion has a steep positive slope. (b) The Autler–Townes interpretation: the strong control field splits the excited state into two dressed states, and the probe absorption vanishes due to destructive interference between the two excitation pathways. (c) Experimental EIT spectrum showing the transparency window.]

3.4 Physical Interpretation: Dressed-State Picture

EIT can be elegantly understood in the dressed-state picture. The strong control field $\Omega_c$ couples the bare states $| 2 \rangle$ and $| 3 \rangle$, forming two dressed states: \begin{equation} | + \rangle = \sin\theta \, | 2 \rangle + \cos\theta \, | 3 \rangle, \qquad | - \rangle = \cos\theta \, | 2 \rangle - \sin\theta \, | 3 \rangle, \end{equation} with $\tan(2\theta) = |\Omega_c|/\Delta_c$. The probe field couples $| 1 \rangle$ to $| 3 \rangle$, and hence to both dressed states. There are thus two possible probe absorption pathways: $| 1 \rangle \to | + \rangle$ and $| 1 \rangle \to | - \rangle$. On two-photon resonance, these two pathways destructively interfere, resulting in zero net absorption. This is the quantum interference picture of EIT. The narrow EIT window is a direct consequence of the fact that the two excitation paths have equal and opposite Fano-type interference profiles.

3.5 The Interferometric Picture: "Which-Path" Information

An alternative and deeply physical interpretation of EIT invokes the concept of which-path information. The probe absorption amplitude is the sum of two indistinguishable quantum paths: direct excitation $| 1 \rangle \xrightarrow{\Omega_p} | 3 \rangle$, and indirect excitation $| 1 \rangle \xrightarrow{\Omega_p} | 3 \rangle \xrightarrow{\Omega_c^*} | 2 \rangle \xrightarrow{\Omega_c} | 3 \rangle$. When the control field is strong, the indirect path becomes equally probable as the direct path, but with a relative phase of $\pi$. The two paths interfere destructively, canceling the total absorption. In the language of measurement theory, the strong control field "which-path" information by correlating the atomic state with the photon number; on Raman resonance, this information is erased, and the interference is restored. EIT is thus a beautiful example of complementarity in quantum mechanics.

4. Applications of EIT and CPT

4.1 Slow Light and Stored Light

The steep normal dispersion at the EIT resonance leads to a dramatically reduced group velocity $v_g = \mathrm{d}\omega/\mathrm{d}k$ for a probe pulse. In 1999, Hau et al. used EIT in a Bose–Einstein condensate of sodium to slow light to $17$ m/s. Subsequently, the same group demonstrated the complete halting and storage of light pulses: a probe pulse was slowed, stopped, and then revived on demand by dynamically controlling the control field intensity. This "stored light" is the basis of quantum memories for photonic qubits—a critical component for long-distance quantum communication and quantum repeaters.

4.2 EIT-Enhanced Nonlinear Optics

While EIT eliminates linear absorption, it simultaneously enhances nonlinear optical processes. The steep dispersion increases the effective interaction time of light with the medium, and the transparency eliminates the usual trade-off between nonlinearity and absorption. This has enabled the demonstration of giant Kerr nonlinearities at the single-photon level, efficient four-wave mixing, and the generation of non-classical states of light such as squeezed vacuum and entangled photon pairs. EIT-based nonlinear optics is a promising avenue for quantum information processing with photons.

4.3 CPT Atomic Clocks and Magnetometers

The narrow CPT resonance can be used as a frequency reference for atomic clocks. In a CPT clock, a modulated laser field generates two phase-coherent frequency components that act as the probe and control. The CPT resonance is detected as a narrow transmission peak when the modulation frequency matches the ground-state hyperfine splitting. CPT clocks are being developed as compact, low-power alternatives to traditional microwave atomic clocks. Similarly, the sensitivity of the CPT resonance to magnetic fields (via the Zeeman shift of the ground-state sublevels) enables chip-scale atomic magnetometers with sensitivities rivaling SQUIDs.

4.4 STIRAP: Stimulated Raman Adiabatic Passage

STIRAP, briefly mentioned earlier, uses the dark state to transfer population between $| 1 \rangle$ and $| 2 \rangle$ with $100\%$ efficiency, without ever populating the excited state $| 3 \rangle$. The counter-intuitive pulse sequence (control before probe) adiabatically rotates the dark state from $| 1 \rangle$ to $| 2 \rangle$. Since the excited state is never populated, spontaneous emission losses are completely eliminated. STIRAP has found wide application in chemical reaction dynamics, atomic and molecular beam experiments, and quantum state preparation.

[Figure 4: (a) The slow-light experiment: a probe pulse is delayed by the EIT medium, while the control field is kept on. (b) Stored light: the control field is turned off while the probe pulse is inside the medium, mapping the photonic state onto a collective atomic spin wave; the control is turned back on to retrieve the pulse. (c) STIRAP pulse sequence: the control field ($\Omega_c$) is applied before the probe field ($\Omega_p$), adiabatically rotating the dark state from $| 1 \rangle$ to $| 2 \rangle$.]

5. EIT-Based Laser Cooling

5.1 Beyond Doppler and Sisyphus: The Quest for Sub-Recoil Cooling

Doppler cooling and polarization gradient (Sisyphus) cooling can achieve temperatures down to a few times the recoil limit $T_R = \hbar^2 k^2 / (2 m k_B)$. For the alkali atoms, $T_R$ is on the order of a few hundred nanokelvin. Reaching the recoil limit—or even surpassing it to achieve sub-recoil temperatures—requires cooling mechanisms that do not rely on the momentum diffusion inherent in photon scattering. The dark state, which by definition does not scatter photons, offers a path to such cooling: atoms with near-zero velocity are coherently trapped in a dark state and cease to interact with the cooling light, while atoms with non-zero velocity continue to scatter photons and undergo a random walk in momentum space until they too fall into the dark state. This is the principle of velocity-selective coherent population trapping (VSCPT).

5.2 Velocity-Selective Coherent Population Trapping (VSCPT)

In VSCPT, the two-photon Raman resonance condition depends on the atomic velocity via the Doppler effect. Consider a $\Lambda$-system with the two ground states $| 1 \rangle$ and $| 2 \rangle$ having a small energy splitting (e.g., different Zeeman sublevels of the same hyperfine ground state). The two laser beams propagate in opposite directions with wavevectors $\mathbf{k}_p$ and $\mathbf{k}_c$. The effective two-photon detuning for an atom moving with velocity $\mathbf{v}$ is \begin{equation} \delta_{\text{eff}}(\mathbf{v}) = \delta + (\mathbf{k}_p - \mathbf{k}_c) \cdot \mathbf{v}. \end{equation} By choosing $\mathbf{k}_p \approx \mathbf{k}_c$ (co-propagating beams) or $\mathbf{k}_p \approx -\mathbf{k}_c$ (counter-propagating beams), the velocity dependence can be tuned. For co-propagating beams, $\mathbf{k}_p - \mathbf{k}_c \approx 0$, and the two-photon resonance is essentially velocity-independent—this is the standard EIT/CPT configuration. For counter-propagating beams, the two-photon resonance condition is sharply velocity-selective: only atoms with a specific velocity component satisfy $\delta_{\text{eff}} = 0$.

In VSCPT, the counter-propagating configuration is used. Atoms with the "right" velocity (the velocity class that satisfies $\delta_{\text{eff}} = 0$) are pumped into the dark state and stop scattering photons. Atoms with other velocities continue to scatter, executing a random walk in momentum space due to photon recoil. Over time, this random walk allows atoms to diffuse into the narrow velocity class that is trapped in the dark state. The result is a cooling mechanism that can achieve temperatures below the recoil limit, limited only by the width of the dark-state velocity window, which is determined by the inverse of the interaction time (or the ground-state decoherence rate). VSCPT was first demonstrated by Aspect et al. in 1988 with metastable helium, achieving one-dimensional temperatures of $T \approx \hbar^2 k^2 / (2m) / 4 \approx T_R/4$.

5.3 Gray Molasses and EIT Cooling of Alkali Atoms

A particularly powerful and experimentally simple variant of EIT-based cooling is the gray molasses technique, which has revolutionized the all-optical production of quantum degenerate gases of lithium and potassium. In gray molasses, the cooling light is tuned near the D$_1$ line of the alkali atom, and the $\Lambda$-system is formed by the two ground-state hyperfine levels ($F=1$ and $F=2$ for $^{6}$Li, or the two lowest hyperfine states for $^{40}$K) and the excited state. The cooling and repump beams are both present, and a small magnetic field is applied to Zeeman-shift the sublevels.

The key to gray molasses is that the dark state is velocity-selective but also momentum-dependent in three dimensions. A specific configuration of laser polarizations and magnetic fields creates a situation where atoms with low kinetic energy are optically pumped into the dark state, where they are decoupled from the light and experience no further heating. Atoms with higher energy remain in the bright state and undergo Sisyphus-like cooling cycles until they find the dark state. The process efficiently cools the atomic cloud to temperatures well below the Doppler limit, and even below the recoil limit in some cases. The name "gray molasses" refers to the fact that the atoms are in a "gray" superposition state (the dark state), in contrast to the "bright" state of standard optical molasses.

For $^{40}$K and $^{6}$Li, D$_1$ gray molasses has been used to cool atomic clouds directly from a MOT to temperatures of a few microkelvin, achieving phase-space densities high enough for direct loading into an optical dipole trap and subsequent evaporative cooling to quantum degeneracy—all without the need for a magnetic trap or an intermediate sub-Doppler cooling stage. This "all-optical" route to BEC and degenerate Fermi gases has become standard practice in many ultracold atom laboratories.

5.4 EIT Ground-State Cooling of Trapped Ions

In trapped-ion quantum computing, the motional state of the ion must be cooled to near the quantum ground state for high-fidelity gate operations. EIT cooling (also known as dark-state cooling) uses a three-level $\Lambda$-system within the ion's internal energy structure. Two laser beams, both red-detuned from their respective transitions, create a dark state that is velocity-selective. An ion moving with a specific velocity is coherently trapped, while ions with other velocities are optically pumped and cooled. EIT cooling can achieve ground-state occupation probabilities exceeding $99\%$ and works over a broader range of trap frequencies than traditional sideband cooling. It has been demonstrated with $^{40}$Ca$^+$, $^{171}$Yb$^+$, and other ion species, and is becoming a standard tool in the trapped-ion quantum computing toolbox.

[Figure 5: (a) The VSCPT mechanism: counter-propagating beams create a velocity-selective dark state. Atoms with the correct velocity are trapped in the dark state and stop scattering; others continue to scatter and diffuse in momentum space until they reach the dark velocity. (b) D$_1$ gray molasses energy-level scheme for $^{40}$K: the $\Lambda$-system formed by the two ground hyperfine states and the D$_1$ excited state. (c) Temperature of a $^{40}$K cloud after gray molasses cooling as a function of cooling time, showing temperatures well below the Doppler limit.]

6. Theoretical Framework for EIT Cooling

6.1 The Cooling Force in the Weak-Probe Limit

The cooling force in EIT-based schemes can be derived from the radiation pressure on the atom. In the weak-probe limit, the probe beam is the primary source of momentum exchange. The average force on an atom is \begin{equation} \langle F \rangle = \hbar \mathbf{k}_p \, \Gamma \rho_{33}, \end{equation} where $\rho_{33}$ is the steady-state excited-state population, which depends on the atomic velocity through the velocity-dependent detunings and two-photon detuning. Near the two-photon resonance, the excited-state population is strongly suppressed due to CPT/EIT, leading to a dramatic reduction in the scattering rate. The velocity dependence of this suppression creates the cooling force.

The cooling rate and the ultimate temperature can be analyzed using a Fokker–Planck equation for the atomic momentum distribution, with the drift (cooling) and diffusion (heating) coefficients derived from the optical Bloch equations. The key result is that the equilibrium temperature scales as \begin{equation} k_B T_{\text{EIT}} \sim \frac{\hbar \Gamma}{2} \frac{\gamma_{12}}{\Gamma} \frac{\Gamma}{|\Omega_c|}, \end{equation} where $\gamma_{12}$ is the ground-state decoherence rate (which can be made very small) and $|\Omega_c|$ is the control Rabi frequency (which can be large). Thus, temperatures well below the Doppler limit are achievable.

6.2 Quantum Monte-Carlo Wavefunction Simulations

For quantitative predictions, the full three-dimensional, multi-level optical Bloch equations including the atomic motion are typically solved using quantum Monte-Carlo wavefunction (QMCWF) simulations. In this approach, the atomic wavefunction evolves under an effective non-Hermitian Hamiltonian (accounting for the coherent dynamics and the non-Hermitian decay terms), punctuated by random quantum jumps that simulate spontaneous emission events with their associated momentum recoil. Statistical averaging over many trajectories yields the momentum distribution and the cooling dynamics. QMCWF simulations have been essential for optimizing gray molasses parameters and predicting the performance of EIT cooling schemes.

References

  1. M. Fleischhauer, A. Imamoglu, and J. P. Marangos, "Electromagnetically induced transparency: Optics in coherent media," Reviews of Modern Physics 77, 633–673 (2005). The definitive review of EIT theory, experiments, and applications.
  2. E. Arimondo, "Coherent population trapping in laser spectroscopy," Progress in Optics 35, 257–354 (1996). A comprehensive review of CPT, its theoretical foundations, and its early experimental realizations.
  3. G. Alzetta, A. Gozzini, L. Moi, and G. Orriols, "An experimental method for the observation of r.f. transitions and laser beat resonances in oriented Na vapour," Il Nuovo Cimento B 36, 5–20 (1976). The original experimental observation of coherent population trapping.
  4. S. E. Harris, "Electromagnetically induced transparency," Physics Today 50(7), 36–42 (1997). An accessible overview of EIT and its physical interpretation.
  5. L. V. Hau, S. E. Harris, Z. Dutton, and C. H. Behroozi, "Light speed reduction to 17 metres per second in an ultracold atomic gas," Nature 397, 594–598 (1999). The landmark slow-light experiment.
  6. C. Liu, Z. Dutton, C. H. Behroozi, and L. V. Hau, "Observation of coherent optical information storage in an atomic medium using halted light pulses," Nature 409, 490–493 (2001). The first demonstration of stored light.
  7. A. Aspect, E. Arimondo, R. Kaiser, N. Vansteenkiste, and C. Cohen-Tannoudji, "Laser cooling below the one-photon recoil energy by velocity-selective coherent population trapping," Physical Review Letters 61, 826–829 (1988). The first demonstration of VSCPT sub-recoil cooling.
  8. D. Rio Fernandes, F. Sievers, N. Kretzschmar, S. Wu, C. Salomon, and F. Chevy, "Sub-Doppler laser cooling of fermionic $^{40}$K atoms in three-dimensional gray optical molasses," Europhysics Letters 100, 63001 (2012). Demonstration of D$_1$ gray molasses for $^{40}$K, a key technique for all-optical Fermi gas production.
  9. A. T. Grier, I. Ferrier-Barbut, B. S. Rem, M. Delehaye, L. Khaykovich, F. Chevy, and C. Salomon, "$\Lambda$-enhanced sub-Doppler cooling of lithium atoms in D$_1$ gray molasses," Physical Review A 87, 063411 (2013). Gray molasses cooling of $^{6}$Li and $^{7}$Li to sub-Doppler temperatures.
  10. G. Morigi, J. Eschner, and C. H. Keitel, "Ground state laser cooling using electromagnetically induced transparency," Physical Review Letters 85, 4458–4461 (2000). The theoretical proposal for EIT ground-state cooling of trapped ions.
  11. C. F. Roos, D. Leibfried, A. Mundt, F. Schmidt-Kaler, J. Eschner, and R. Blatt, "Experimental demonstration of ground state laser cooling with electromagnetically induced transparency," Physical Review Letters 85, 5547–5550 (2000). The first experimental realization of EIT cooling of a trapped ion.
  12. K. Bergmann, H. Theuer, and B. W. Shore, "Coherent population transfer among quantum states of atoms and molecules," Reviews of Modern Physics 70, 1003–1025 (1998). The comprehensive review of STIRAP.