Rabi and Ramsey Control

Rabi and Ramsey protocols are the basic control and diagnostic tools for a two-level quantum system. Rabi control drives population between two states. Ramsey control converts phase accumulation during free evolution into a measurable population difference. Together they calibrate drive strength, detuning, pulse phase, coherence time, and low-frequency noise.

This page treats them as open-system protocols. The closed-system Rabi formula belongs to Rabi Oscillations: First Encounter. The driven dissipative two-level equations belong to Optical Bloch Equations. Here the focus is how Rabi and Ramsey experiments are used for control when relaxation, dephasing, detuning, and calibration errors are present.

For the AMO connection from transition matrix elements and field intensity to pulse areas, measured population traces, chevrons, readout likelihoods, and model diagnostics, see Rabi Oscillations. The complementary AMO treatment of the separated-pulse propagator, finite-pulse fringes, detuning estimators, and atomic-clock discriminator is Ramsey Interferometry.

Two Primitives

In a rotating frame, a resonant two-level drive with phase $\phi$ is often written

$$H_{\mathrm{drive}}(t) = \frac{\hbar\Omega(t)}{2} \left( \cos\phi\,X+\sin\phi\,Y \right).$$

If relaxation and dephasing are negligible during the pulse, the pulse implements the rotation

$$U_\phi(\theta) = \exp \left[ - \frac{i\theta}{2} \left( \cos\phi\,X+\sin\phi\,Y \right) \right],$$

where the pulse area is

$$\theta = \int dt\,\Omega(t).$$

A Rabi experiment varies $\theta$ or the pulse duration and measures population oscillations. A Ramsey experiment uses two approximately $\pi/2$ pulses separated by a free-evolution time $T$, and measures the phase accumulated in between. For a compact notation reference for common pulse patterns, see Pulse Sequences.

The important distinction is:

Rabi control measures how well a drive rotates the state;
Ramsey control measures how phase evolves between drives.

Rabi Control

On resonance, a rectangular pulse with constant Rabi frequency $\Omega$ has pulse area

$$\theta=\Omega t.$$

The usual pulse times are

$$t_\pi = \frac{\pi}{\Omega}, \qquad t_{\pi/2} = \frac{\pi}{2\Omega}.$$

In the ideal closed two-level model, a $\pi$ pulse swaps ground and excited populations, while a $\pi/2$ pulse prepares an equal-population superposition. With detuning $\Delta$, the effective rotation axis tilts away from the equator, the oscillation frequency becomes $\sqrt{\Omega^2+\Delta^2}$, and the maximum transfer is reduced. Those closed-system formulas are derived in Rabi Oscillations: First Encounter.

In an open system, Rabi oscillations are damped by relaxation, pure dephasing, leakage, drive noise, inhomogeneity, and measurement imperfections. A minimal Markovian model is the optical-Bloch master equation

$$\dot\rho = - \frac{i}{\hbar} [H_{\mathrm{rot}},\rho] + \Gamma\,\mathcal D[\sigma_-]\rho + \frac{\gamma_\phi}{2} \mathcal D[Z]\rho.$$

Here $\Gamma=1/T_1$ is the population decay rate in the simple zero-temperature model, and

$$\frac{1}{T_2} = \frac{1}{2T_1} + \gamma_\phi$$

is the transverse Markovian coherence decay rate. If $\Omega$ is not large compared with the relevant decay rates, the oscillations may be overdamped or strongly distorted.

Ramsey Control

A Ramsey sequence has the form

π/2 pulse → free evolution for T → π/2 pulse → population measurement

During the free evolution, a detuning $\Delta$ produces relative phase

$$\varphi = \Delta T.$$

With a standard phase convention, an initial ground state gives an excited-state probability

$$P_e(T) = \frac12 \left[ 1+\cos(\Delta T+\phi_0) \right],$$

where $\phi_0$ includes the relative phase of the two control pulses and any fixed offset in the measurement convention. Changing the second pulse phase scans the fringe without changing $T$.

In the presence of decoherence, the fringe contrast decays:

$$P_e(T) = \frac12 \left[ 1+ C(T)\cos(\Delta T+\phi_0) \right],$$

with $0\le C(T)\le1$. In the simplest Markovian pure-dephasing model,

$$C(T)=e^{-T/T_2}.$$

In many experiments the observed Ramsey contrast instead defines $T_2^*$, because slow detuning drift and shot-to-shot inhomogeneity are included in the observed envelope. The distinction between $T_2$, $T_2^*$, and echo coherence times is canonical in Decoherence Timescales.

What Ramsey Measures

Ramsey control is sensitive to phase accumulation. That makes it a powerful diagnostic for:

• detuning calibration;
• clock and magnetometer signals;
• low-frequency dephasing noise;
• slow drift of fields, laser frequencies, or qubit splittings;
• ensemble inhomogeneity;
• pulse phase errors;
• free-evolution coherence time.

For a qubit with stochastic detuning noise $\xi(t)$,

$$H_{\mathrm{noise}}(t) = \frac{\hbar}{2} \xi(t)Z,$$

the Ramsey coherence factor is

$$W(T) = \left\langle \exp \left[ - i\int_0^T dt\,\xi(t) \right] \right\rangle.$$

If $\xi$ is quasi-static during each shot and Gaussian distributed between shots with variance $\sigma^2$, then

$$W(T) = e^{-\sigma^2T^2/2}.$$

This Gaussian envelope is not a Markovian exponential. It is one reason Ramsey $T_2^*$ can be much shorter than echo or dynamically decoupled coherence times.

Echo and Decoupling Connection

Ramsey free evolution has modulation function

$$y_{\mathrm{Ramsey}}(t)=1.$$

A spin echo inserts a $\pi$ pulse halfway through the free evolution, reversing the sign of the accumulated phase from slow longitudinal noise. In filter-function language, Ramsey and echo sample different parts of the same noise spectrum. Slow noise that strongly affects Ramsey can be suppressed by echo and by longer Dynamical Decoupling sequences.

The control lesson is practical: a quoted coherence time is not complete unless the pulse sequence is specified. A Ramsey time, an echo time, a CPMG time, and a relaxation time diagnose different physics.

Rabi versus Ramsey Diagnostics

Diagnostic	Vary	Main observable	Sensitive to
Rabi oscillation	pulse duration or amplitude	population oscillation	drive amplitude, detuning, damping, pulse area
Ramsey fringe	free time or second-pulse phase	fringe phase and contrast	detuning, low-frequency dephasing, phase errors
Echo	free time with refocusing pulse	refocused contrast	slower dephasing, pulse errors, nonstatic noise
Driven steady state	continuous drive strength	saturation or fluorescence	relaxation, dephasing, drive power broadening

Rabi and Ramsey are often used sequentially. A Rabi experiment calibrates the $\pi/2$ pulse area. A Ramsey experiment then calibrates detuning and measures coherence. An echo or decoupling experiment separates quasi-static dephasing from faster noise. None of these replaces a direct $T_1$ relaxation measurement.

Calibration Workflow

A typical two-level control workflow is:

1. Identify the transition and choose a rotating-frame convention.
2. Measure Rabi oscillations versus pulse length to calibrate $\Omega$ and pulse area.
3. Use a $\pi$ pulse to check population inversion and leakage.
4. Use a $\pi/2$ pulse pair to measure Ramsey fringes versus detuning or delay.
5. Fit the Ramsey frequency to correct detuning.
6. Fit the Ramsey contrast only after choosing an envelope model.
7. Add echo or dynamical-decoupling measurements to separate slow drift from faster dephasing.
8. Check the same controls under the measurement protocol actually used for the experiment.

The last step matters because readout backaction, state-preparation errors, finite pulse width, and drive leakage can change the apparent contrast.

Open-System Limits

Rabi and Ramsey control assume the intended two-level subspace remains meaningful. Watch for:

• leakage to nearby levels under strong drive;
• breakdown of the rotating-wave approximation;
• pulse bandwidth broad enough to excite unwanted transitions;
• amplitude noise that converts into pulse-area errors;
• phase noise that appears as detuning noise;
• relaxation during the pulse;
• measurement errors that masquerade as reduced contrast;
• non-Markovian or non-Gaussian noise envelopes fitted by an exponential without justification.

The Approximation Checklist is the place to record these assumptions when a model is used for quantitative inference.

Common Mistakes

• Calling every population oscillation a clean Rabi oscillation without checking leakage or detuning.
• Comparing Rabi frequencies without checking whether $\Omega$ or $\Omega/2$ appears in the Hamiltonian.
• Treating a $\pi/2$ pulse as phase-independent; its axis depends on the drive phase.
• Reporting “the coherence time” without saying Ramsey, echo, decoupled, or relaxation.
• Fitting a Gaussian quasi-static Ramsey envelope with a Markovian exponential and interpreting the result as $T_2$.
• Assuming a longer Rabi decay time means a longer Ramsey coherence time; the protocols sample different noise.
• Forgetting that finite pulse duration contributes to the total experimental time and to dephasing.

Exercises

Pulse areas

A resonant rectangular pulse has Rabi frequency $\Omega/2\pi=5\,\mathrm{MHz}$ in the convention $H=\hbar\Omega X/2$. Find $t_\pi$ and $t_{\pi/2}$.

Solution

The angular Rabi frequency is

$$\Omega = 2\pi(5\,\mathrm{MHz}).$$

Thus

$$t_\pi = \frac{\pi}{\Omega} = \frac{1}{2(5\,\mathrm{MHz})} = 100\,\mathrm{ns}.$$

The half pulse is

$$t_{\pi/2} = 50\,\mathrm{ns}.$$

These times depend on the convention for $\Omega$, so the Hamiltonian definition must always be checked.

Ramsey fringe phase

Use the Bloch-sphere convention in which the first $\pi/2$ pulse rotates the ground-state Bloch vector from $-z$ to $+y$. Free evolution with detuning $\Delta$ rotates around $z$ for time $T$. A second identical $\pi/2$ pulse maps the accumulated phase to population. Show that $P_e=(1+\cos\Delta T)/2$.

Solution

After the first pulse the Bloch vector is

$$\mathbf r_1=(0,1,0).$$

Free evolution around $z$ by angle $\Delta T$ gives

$$\mathbf r_2 = (-\sin\Delta T,\cos\Delta T,0).$$

The second $x$-axis $\pi/2$ pulse maps the $y$ component into the $z$ component, so

$$z_{\mathrm{final}} = \cos\Delta T.$$

Since $P_e=(1+z_{\mathrm{final}})/2$ in the convention $Z=\lvert e\rangle\langle e\rvert-\lvert g\rangle\langle g\rvert$,

$$P_e = \frac12 \left( 1+\cos\Delta T \right).$$

Changing pulse phases or the sign convention for detuning shifts the phase or flips the cosine, but the fringe contrast and frequency are the invariant information.

Quasi-static dephasing

A Ramsey experiment has quasi-static Gaussian detuning noise with standard deviation $\sigma$. The envelope is $C(T)=e^{-\sigma^2T^2/2}$. Find the time $T_*$ where $C(T_*)=e^{-1}$.

Solution

Set

$$\frac{\sigma^2T_*^2}{2}=1.$$

Therefore

$$T_* = \frac{\sqrt2}{\sigma}.$$

This is a Ramsey inhomogeneous-dephasing time, often reported as a version of $T_2^*$. It is not the same as the Markovian $T_2$ unless the noise model and protocol justify that identification.

Cross-Links

• Ramsey Interferometry for Quantum Estimation
• Driven Open Systems
• Control, Readout, and Calibration
• Pulse-Level Control
• Calibration Loops
• Pulse Sequences
• Rabi Oscillations: First Encounter
• Rabi Oscillations
• Ramsey Interferometry
• Optical Bloch Equations
• Bloch Sphere: Wave-Mechanics Perspective
• Rotating-Wave Approximation
• Decoherence Timescales
• Pure Dephasing Master Equation
• Pure Dephasing Model
• Noise Spectra
• One-Over-F Noise
• Dynamical Decoupling
• Decoherence Timescale Estimation
• Approximation Checklist

References

• I. I. Rabi, “Space quantization in a gyrating magnetic field,” Physical Review 51, 652–654 (1937).
• N. F. Ramsey, “A molecular beam resonance method with separated oscillating fields,” Physical Review 78, 695–699 (1950).
• E. L. Hahn, “Spin echoes,” Physical Review 80, 580–594 (1950).
• C. P. Slichter, Principles of Magnetic Resonance, 3rd ed., Springer (1990).
• C. Cohen-Tannoudji, J. Dupont-Roc, and G. Grynberg, Atom–Photon Interactions, Wiley (1992).
• L. Allen and J. H. Eberly, Optical Resonance and Two-Level Atoms, Dover (1987).
• H.-P. Breuer and F. Petruccione, The Theory of Open Quantum Systems, Oxford University Press (2002).