Quantum Optics XI Ramsey Spectroscopy
fengxiaot Lv4

Ramsey spectroscopy is one of the most powerful techniques for precision frequency measurements in quantum optics. It is the basic operating principle behind atomic clocks. It rests on a simple yet profound idea: coherent control of a two-level system using two separated pulses.

Ramsey Spectroscopy

Rotating Frame Hamiltonian

Consider a two-level system interacting with a monochromatic field

H=12ℏω0σz+ℏΩ0cos⁡(ωt+ϕ0)σxH = \frac{1}{2} \hbar \omega_0 \sigma_z +\hbar\Omega_0 \cos(\omega t+\phi_0) \sigma_x

H=12ℏω0σz+12ℏΩ0[ei(ωt+ϕ0)+e−i(ωt+ϕ0)](σ†+σ)H=\frac{1}{2}\hbar\omega_0 \sigma_z +\frac{1}{2} \hbar\Omega_0 \left[\mathrm{e}^{\mathrm{i}(\omega t+\phi_0)} +\mathrm{e}^{-\mathrm{i}(\omega t+\phi_0)}\right] (\sigma^\dagger +\sigma)

where σ†\sigma^\dagger is the raising operator ∣↑⟩⟨↓∣\ket{\uparrow}\bra{\downarrow} and σ\sigma is the lowering operator ∣↓⟩⟨↑∣\ket{\downarrow}\bra{\uparrow}. Notice that the counter-rotating term is far-off-resonant, so it will not affect the state too much, but it will lead to a stark shift of the energy level

ω0→ω0+Ω022(ω0+ω)\omega_0 \to \omega_0 + \frac{\Omega_0^2}{2(\omega_0+\omega)}

Typically atomic frequency ω0\omega_0 should be around 2π×100 THz2\pi \times 100 \, \mathrm{THz}, while the Rabi frequency Ω0\Omega_0 is only 2π×100 KHz2\pi \times 100 \, \mathrm{KHz}, so the shift caused by the counter-rotating term should be very very very small. It is kept only for completeness.

Hence we can remove this counter-rotating term and write

H=12ℏ[ω0+Ω022(ω0+ω)]σz+12ℏΩ0[σei(ωt+ϕ0)+σ†e−i(ωt+ϕ0)]H=\frac{1}{2}\hbar \left[\omega_0 + \frac{\Omega_0^2}{2(\omega_0+\omega)}\right] \sigma_z +\frac{1}{2} \hbar\Omega_0 \left[\sigma\mathrm{e}^{\mathrm{i}(\omega t+\phi_0)}+\sigma^\dagger \mathrm{e}^{-\mathrm{i}(\omega t+\phi_0)}\right]

Defining U=ei2ωtσzU = \mathrm{e}^{\frac{\mathrm{i}}{2}\omega t\sigma_z}, and switching to the rotating frame

H~=UHU†+iℏU˙U†=−12ℏΔσz+12ℏΩ0(σ†e−iϕ0+σeiϕ0)\tilde{H} = U H U^\dagger + \mathrm{i}\hbar \dot{U} U^\dagger = -\frac{1}{2}\hbar\Delta \sigma_z + \frac{1}{2} \hbar \Omega_0 (\sigma^\dagger\mathrm{e}^{-\mathrm{i}\phi_0} + \sigma\mathrm{e}^{\mathrm{i}\phi_0})

where Δ=ω−[ω0+Ω022(ω0+ω)]\Delta = \omega - \left[\omega_0 + \frac{\Omega_0^2}{2(\omega_0+\omega)}\right], The evolution operator in the rotating frame will be

U~=exp⁡(−iℏH~t)=[cos⁡Ωt2+iΔΩsin⁡Ωt2−isin⁡Ωt2Ω0Ωe−iϕ0−isin⁡Ωt2Ω0Ωeiϕ0cos⁡Ωt2−iΔΩsin⁡Ωt2]\tilde{U} = \exp\left( - \frac{\mathrm{i}}{\hbar} \tilde{H} t\right) = \begin{bmatrix} \cos\frac{\Omega t}{2} + \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega t}{2} & -\mathrm{i} \sin\frac{\Omega t}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{-\mathrm{i}\phi_0} \\ -\mathrm{i} \sin\frac{\Omega t}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{\mathrm{i}\phi_0} & \cos\frac{\Omega t}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega t}{2} \end{bmatrix}

The state of the spins

∣ψ~(t)⟩=U~∣ψ~(0)⟩=U~∣↓⟩\ket{\tilde{\psi}(t)} = \tilde{U} \ket{\tilde{\psi}(0)} = \tilde{U} \ket{\downarrow}

and switching back to the lab frame (Schrödinger picture)

∣ψ(t)⟩=e−i2ωtσz∣ψ~(t)⟩=e−i2ωtσzU~∣↓⟩=[−isin⁡Ωt2Ω0Ωe−iϕ0e−i2ωtei2ωt(cos⁡Ωt2−iΔΩsin⁡Ωt2)]\ket{\psi(t)} = \mathrm{e}^{-\frac{\mathrm{i}}{2}\omega t\sigma_z} \ket{\tilde{\psi}(t)} = \mathrm{e}^{-\frac{\mathrm{i}}{2}\omega t\sigma_z} \tilde{U} \ket{\downarrow} = \begin{bmatrix} -\mathrm{i} \sin\frac{\Omega t}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{-\mathrm{i}\phi_0} \mathrm{e}^{-\frac{\mathrm{i}}{2}\omega t} \\ \mathrm{e}^{\frac{\mathrm{i}}{2}\omega t} \left( \cos\frac{\Omega t}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega t}{2} \right) \end{bmatrix}

Experiment

Assume the atoms start in the ground state. At t=0t=0, we apply the first pulse. We choose the field amplitude Ω0\Omega_0 such that the field drives a π/2\pi/2 pulse, i.e., Ω0τ=π/2\Omega_0 \tau = \pi/2.

Then, we let the atom undergo free evolution precession about the z-axis for a time TT.

Finally, we apply the second π/2\pi/2 pulse, and its effect depends on the state of the atom after the precession
stage.

First pulse

At time τ=π2Ω0\tau = \frac{\pi}{2\Omega_0}, the probability to find the spin-up state is

∣⟨↑∣ψ(τ)⟩∣2=Ω02Ω02+Δ2sin⁡2Ω02+Δ2τ2=Ω02Ω2sin⁡2Ωτ2|\langle \uparrow | \psi(\tau)\rangle|^2= \frac{\Omega_0^2}{\Omega_0^2+\Delta^2} \sin^2\frac{\sqrt{\Omega_0^2+\Delta^2} \tau}{2} = \frac{\Omega_0^2}{\Omega^2} \sin^2\frac{\Omega \tau}{2}

Free precession

The spins process freely for time TT, the corresponding Hamiltonian in the rotating frame is

H=12ℏω0σz→H~=12ℏ(ω0−ω)σz=−12ℏΔσzH = \frac{1}{2} \hbar \omega_0 \sigma_z \to \tilde{H}= \frac{1}{2}\hbar (\omega_0-\omega)\sigma_z = -\frac{1}{2}\hbar\Delta \sigma_z

thus the evolution operator is

U~=exp⁡(−iℏH~T)=exp⁡(i2ΔTσz)=[ei2ΔT00e−i2ΔT]\tilde{U} = \exp\left( - \frac{\mathrm{i}}{\hbar} \tilde{H} T \right) = \exp\left(\frac{\mathrm{i}}{2} \Delta T \sigma_z\right) = \begin{bmatrix} \mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T} & 0\\ 0 & \mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T} \end{bmatrix}

The state of the spins after the procession in the rotating frame is

∣ψ~(τ+T)⟩=[ei2ΔT00e−i2ΔT][−isin⁡Ωt2Ω0Ωe−iϕ0cos⁡Ωt2−iΔΩsin⁡Ωt2]=[−isin⁡Ωτ2Ω0Ωe−iϕ0ei2ΔTe−i2ΔT(cos⁡Ωτ2−iΔΩsin⁡Ωτ2)]\ket{\tilde{\psi}(\tau+T)} = \begin{bmatrix} \mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T} & 0\\ 0 & \mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T} \end{bmatrix} \begin{bmatrix} -\mathrm{i} \sin\frac{\Omega t}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{-\mathrm{i}\phi_0} \\ \cos\frac{\Omega t}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega t}{2} \end{bmatrix} = \begin{bmatrix} -\mathrm{i} \sin\frac{\Omega \tau}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{-\mathrm{i}\phi_0} \mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T} \\ \mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T} \left( \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \right) \end{bmatrix}

Second pulse

The second pulse has the same duration τ\tau but may have an adjustable phase φ\varphi relative to the first pulse. Repeat what we’ve done above, but substituting ϕ0\phi_0 with ϕ0+φ\phi_0 +\varphi, we get

∣ψ~(2τ+T)⟩=U~∣ψ~(τ+T)⟩\ket{\tilde{\psi}(2\tau +T)} = \tilde{U} \ket{\tilde{\psi}(\tau+T)}

∣ψ~(2τ+T)⟩=[cos⁡Ωτ2+iΔΩsin⁡Ωτ2−isin⁡Ωτ2Ω0Ωe−i(ϕ0+φ)−isin⁡Ωτ2Ω0Ωei(ϕ0+φ)cos⁡Ωτ2−iΔΩsin⁡Ωτ2][−isin⁡Ωτ2Ω0Ωe−iϕ0ei2ΔTe−i2ΔT(cos⁡Ωτ2−iΔΩsin⁡Ωτ2)]\ket{\tilde{\psi}(2\tau +T)} = \begin{bmatrix} \cos\frac{\Omega \tau}{2} + \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} & -\mathrm{i} \sin\frac{\Omega \tau}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{-\mathrm{i}(\phi_0+\varphi)} \\ -\mathrm{i} \sin\frac{\Omega \tau}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{\mathrm{i}(\phi_0+\varphi)} & \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \end{bmatrix} \begin{bmatrix} -\mathrm{i} \sin\frac{\Omega \tau}{2} \frac{\Omega_0}{\Omega} \mathrm{e}^{-\mathrm{i}\phi_0} \mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T} \\ \mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T} \left( \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \right) \end{bmatrix}

after some horrible calculations, we finally obtain

∣ψ~(2τ+T)⟩=[−isin⁡Ωτ2Ω0Ωe−iϕ0[ei2ΔT(cos⁡Ωτ2+iΔΩsin⁡Ωτ2)+e−iφe−i2ΔT(cos⁡Ωτ2−iΔΩsin⁡Ωτ2)]−sin⁡2Ωτ2(Ω0Ω)2eiφei2ΔT+e−i2ΔT(cos⁡Ωτ2−iΔΩsin⁡Ωτ2)2]\ket{\tilde{\psi}(2\tau + T)} = \begin{bmatrix} -\mathrm{i} \sin\frac{\Omega \tau}{2} \frac{\Omega_0}{\Omega}\mathrm{e}^{-\mathrm{i}\phi_0} \left[\mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T}\left(\cos\frac{\Omega \tau}{2} + \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2}\right) + \mathrm{e}^{-\mathrm{i}\varphi}\mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T}\left( \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \right)\right] \\ - \sin^2 \frac{\Omega \tau}{2} \left(\frac{\Omega_0}{\Omega}\right)^2 \mathrm{e}^{\mathrm{i}\varphi} \mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T} + \mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T} \left( \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \right)^2 \end{bmatrix}

Switching back to the lab frame just add an global dynamical phase to each component

∣ψ(2τ+T)⟩=[−isin⁡Ωτ2Ω0Ωe−iϕ0e−i2ω(2τ+T)[ei2ΔT(cos⁡Ωτ2+iΔΩsin⁡Ωτ2)+e−iφe−i2ΔT(cos⁡Ωτ2−iΔΩsin⁡Ωτ2)]−sin⁡2Ωτ2(Ω0Ω)2eiφei2ΔTe−i2ω(2τ+T)+e−i2ΔTei2ω(2τ+T)(cos⁡Ωτ2−iΔΩsin⁡Ωτ2)2]\ket{\psi(2\tau + T)} = \begin{bmatrix} -\mathrm{i} \sin\frac{\Omega \tau}{2} \frac{\Omega_0}{\Omega}\mathrm{e}^{-\mathrm{i}\phi_0} \mathrm{e}^{-\frac{\mathrm{i}}{2} \omega (2\tau+T)}\left[\mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T}\left(\cos\frac{\Omega \tau}{2} + \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2}\right) + \mathrm{e}^{-\mathrm{i}\varphi}\mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T}\left( \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \right)\right] \\ - \sin^2 \frac{\Omega \tau}{2} \left(\frac{\Omega_0}{\Omega}\right)^2 \mathrm{e}^{\mathrm{i}\varphi} \mathrm{e}^{\frac{\mathrm{i}}{2}\Delta T}\mathrm{e}^{-\frac{\mathrm{i}}{2} \omega (2\tau+T)} + \mathrm{e}^{-\frac{\mathrm{i}}{2}\Delta T} \mathrm{e}^{\frac{\mathrm{i}}{2} \omega (2\tau+T)}\left( \cos\frac{\Omega \tau}{2} - \mathrm{i} \frac{\Delta}{\Omega} \sin\frac{\Omega \tau}{2} \right)^2 \end{bmatrix}

It does not depend on ϕ0\phi_0. Only the phase difference φ\varphi of two pulse matters. ϕ0\phi_0 is just a global phase.

The probability is

P↑=∣⟨↑∣ψ(2τ+T)⟩∣2=4Ω02Ω2sin⁡2Ωτ2∣cos⁡Ωτ2cos⁡(ΔT2−φ2)−ΔΩsin⁡Ωτ2sin⁡(ΔT2−φ2)∣2P_\uparrow = |\langle \uparrow | \psi(2\tau+T)\rangle|^2= \frac{4\Omega_0^2}{\Omega^2} \sin^2\frac{\Omega \tau}{2} \left| \cos \frac{\Omega \tau}{2} \cos\left(\frac{\Delta T}{2} - \frac{\varphi}{2}\right) -\frac{\Delta}{\Omega}\sin \frac{\Omega \tau}{2} \sin\left(\frac{\Delta T}{2} - \frac{\varphi}{2}\right) \right|^2

Again, substitute τ=π2Ω0\tau = \frac{\pi}{2\Omega_0}

P↑=4Ω02Ω2sin⁡2(ΩΩ0π4)∣cos⁡(ΩΩ0π4)cos⁡(ΔT2−φ2)−ΔΩsin⁡(ΩΩ0π4)sin⁡(ΔT2−φ2)∣2P_\uparrow = \frac{4\Omega_0^2}{\Omega^2} \sin^2 \left(\frac{\Omega}{\Omega_0}\frac{\pi}{4}\right) \left| \cos \left(\frac{\Omega}{\Omega_0}\frac{\pi}{4}\right) \cos\left(\frac{\Delta T}{2} - \frac{\varphi}{2}\right) -\frac{\Delta}{\Omega}\sin \left(\frac{\Omega}{\Omega_0}\frac{\pi}{4}\right) \sin\left(\frac{\Delta T}{2} - \frac{\varphi}{2}\right) \right|^2

Near resonance, Δ≪Ω\Delta \ll \Omega, the equation becomes

P↑=Ω02Ω2sin⁡2(ΩΩ0π4)cos⁡2(ΔT2−φ2)P_\uparrow = \frac{\Omega_0^2}{\Omega^2} \sin^2 \left(\frac{\Omega}{\Omega_0}\frac{\pi}{4}\right)\cos^2\left(\frac{\Delta T}{2} - \frac{\varphi}{2}\right)

It yields a fringe pattern in detuning Δ\Delta.

As shown in the figure, the laser frequency ω\omega is scanned, the population oscillates periodically. The fringe period is

ωfringe=2πT\omega_\text{fringe} = \frac{2\pi}{T}

while the overall envelope of the curve follows the Rabi lineshape

P↑envelope=Ω02Ω2sin⁡2(ΩΩ0π4)=Ω02Ω02+Δ2sin⁡2(Ω02+Δ2Ω0π4)P_\uparrow^\text{envelope}=\frac{\Omega_0^2}{\Omega^2} \sin^2 \left(\frac{\Omega}{\Omega_0}\frac{\pi}{4}\right)=\frac{\Omega_0^2}{\Omega_0^2+\Delta^2} \sin^2 \left(\frac{\sqrt{\Omega_0^2+\Delta^2}}{\Omega_0}\frac{\pi}{4}\right)