Trapped Ions

Trapped ions are among the cleanest platforms for seeing open quantum systems at the single-particle level. Internal electronic states form long-lived qubits, the ion’s motion is a quantized harmonic oscillator, lasers drive coherent transitions, fluorescence produces detector records, and environmental electric-field noise heats motional modes.

This page is an application map. Ion Traps is the canonical AMO treatment of Paul-trap confinement, secular motion, micromotion, Coulomb crystals, cooling, and platform calibration. The purpose here is narrower: to show how the measurement and open-system language used throughout this volume appears in trapped-ion experiments.

Trapped-Ion Qubits owns the complementary processor-level account of encodings, gates, connectivity, shuttling, photonic interconnects, and scaling.

Basic Degrees of Freedom

For one ion in one motional mode, a minimal Hamiltonian is

$$\frac{H_0}{\hbar} = \frac{\omega_0}{2}\sigma_z + \nu a^\dagger a,$$

where $\omega_0$ is an internal transition frequency and $\nu$ is the trap frequency of the motional mode. The zero-point length is

$$x_0 = \sqrt{\frac{\hbar}{2m\nu}},$$

and the Lamb–Dicke parameter for laser wavevector $k$ is

$$\eta = kx_0 .$$

In the Lamb–Dicke regime,

$$\eta\sqrt{\langle n\rangle+1} \ll 1,$$

the ion’s optical phase varies only weakly over the motional wavepacket. This makes sideband-resolved control and cooling possible.

Laser Coupling and Sidebands

The same optical or Raman field can address an internal carrier, exchange an internal excitation with a phonon on a red sideband, or create both excitations on a blue sideband. Trapped-Ion Control owns the interaction-picture derivation, exact motional matrix elements, Lamb–Dicke expansion, and sideband Rabi rates.

For open-system analysis, the important distinction is what follows the coherent sideband pulse. A red-sideband pulse alone transfers excitation unitarily between internal and motional degrees of freedom. Cooling requires an irreversible reset, usually optical pumping with spontaneous emission, to export entropy. The same sideband coupling can instead entangle spin and motion or mediate a spin–spin gate when the reset channel is absent.

Fluorescence Detection

State detection often uses a cycling transition. One internal state is bright: it scatters many photons when illuminated. Another state is dark: it is far off resonance or shelved in a metastable level and scatters few photons.

If the detected photon count in an interval is $n$, a simple model is

$$P(n\mid b) = e^{-\mu_b} \frac{\mu_b^n}{n!}, \qquad P(n\mid d) = e^{-\mu_d} \frac{\mu_d^n}{n!},$$

where $\mu_b$ and $\mu_d$ are the mean detected counts for bright and dark states. Usually

$$\mu_b\gg\mu_d .$$

A threshold on $n$ implements a high-fidelity projective measurement in the internal-state basis. The measurement is strong because many photons are scattered, and each scattered photon can carry which-state information into the environment.

The detection error is not only photon shot noise. It also includes off-resonant pumping, decay out of the cycling transition, detector dark counts, collection inefficiency, finite detection time, and motional heating from recoil.

Measurement Backaction

Fluorescence detection is a vivid example of measurement backaction. The internal state becomes correlated with a macroscopic photon record, and unobserved scattered photons still decohere superpositions of bright and dark states.

For an idealized measurement with bright-state scattering rate $\Gamma_b$ and negligible dark scattering, a bright-dark coherence decays on a timescale set by the rate at which the environment can distinguish the two alternatives. A schematic dephasing model is

$$\dot\rho = \Gamma_\varphi \mathcal D[\sigma_z]\rho,$$

with $\Gamma_\varphi$ determined by the measurement strength and unobserved scattering. The exact coefficient is convention- and level-structure-dependent.

Backaction also affects motion. Photon recoil changes the motional state, and state-dependent optical forces can entangle internal and motional degrees of freedom. High-fidelity internal readout therefore does not automatically mean nondestructive motional readout.

Motional Heating

Ion motion is a harmonic oscillator exposed to electric-field noise, technical drive noise, background-gas collisions, and anharmonic trap imperfections. A common Markovian model for one mode is

$$\dot\rho = - \frac{i}{\hbar}[H,\rho] + \Gamma_\downarrow\mathcal D[a]\rho + \Gamma_\uparrow\mathcal D[a^\dagger]\rho .$$

The mean phonon number obeys

$$\frac{d\langle n\rangle}{dt} = - (\Gamma_\downarrow-\Gamma_\uparrow) \langle n\rangle + \Gamma_\uparrow .$$

If $\Gamma_\downarrow=0$, the mode heats at rate

$$\frac{d\langle n\rangle}{dt} = \Gamma_\uparrow .$$

Electric-field noise at the motional frequency gives a heating rate often written, up to spectral-density convention, as

$$\dot{\bar n} = \frac{e^2} {4m\hbar\nu} S_E(\nu).$$

Here $S_E(\nu)$ is the electric-field noise spectral density at the ion. This formula is why surface noise, trap-electrode distance, filtering, and cryogenic operation matter so much for trapped-ion open-system performance.

Laser Cooling as Engineered Dissipation

Laser cooling is reservoir engineering. The lasers and spontaneous emission create an effective dissipative channel that removes motional quanta.

In resolved-sideband cooling, one drives the red sideband:

$$|g,n\rangle \longrightarrow |e,n-1\rangle .$$

Spontaneous emission then returns the ion to $|g\rangle$ while, in the Lamb–Dicke regime, usually leaving $n-1$ approximately unchanged. Repeating the cycle removes phonons.

The effective motional dynamics can again be written as

$$\dot\rho_m = \Gamma_-\mathcal D[a]\rho_m + \Gamma_+\mathcal D[a^\dagger]\rho_m,$$

where $\Gamma_-$ is the cooling rate and $\Gamma_+$ is the residual heating rate. The steady-state mean occupation is

$$\bar n_{\mathrm{ss}} = \frac{\Gamma_+} {\Gamma_- - \Gamma_+}, \qquad \Gamma_-\gt\Gamma_+ .$$

Doppler cooling, electromagnetically induced transparency cooling, sympathetic cooling, and sideband cooling differ in how they engineer these effective rates and in which limits set the final temperature.

Quantum Jumps

Trapped ions provided some of the classic observations of quantum jumps. In a shelving experiment, the ion alternates between bright periods, where many photons are detected, and dark periods, where the ion is shelved in a long-lived state.

The observed fluorescence record is a time series of counts. A simple two-state hidden Markov picture has transitions

$$b \rightleftarrows d$$

with state-dependent photon count distributions. In a quantum trajectory description, detected photons update the conditional state, while the absence of photons also carries information.

This is not just a metaphor for collapse. The detector record is an experimental time series, and the conditional state is the state assigned given that record. Averaging over all records recovers the unconditional master equation.

See Quantum-Jump Trajectories.

Dephasing of Internal States

Internal-state coherence can be limited by:

• magnetic-field noise;
• laser phase noise;
• intensity noise and AC Stark shifts;
• spontaneous scattering;
• motional-state fluctuations;
• off-resonant coupling to nearby levels.

A simple phenomenological model is

$$\dot\rho = - \frac{i}{\hbar}[H,\rho] + \frac{\Gamma_\phi}{2} \mathcal D[\sigma_z]\rho .$$

The measured coherence time depends on the experiment. Ramsey measurements are sensitive to low-frequency detuning noise, while spin echo and dynamical decoupling filter the noise spectrum differently. A quoted $T_2$ therefore always has a sequence and noise-bandwidth context.

Ion Chains and Normal Modes

In an ion chain, motion decomposes into collective normal modes. A mode expansion has the schematic form

$$x_j = \sum_m b_{jm}x_{0m} (a_m+a_m^\dagger),$$

where $j$ labels ions and $m$ labels modes. Lasers couple internal states to these collective modes, and environmental noise can heat modes differently.

Open-system questions become more layered:

• which modes are cooled or heated;
• whether noise is common-mode or local;
• whether sympathetic cooling disturbs stored qubits;
• how spontaneous emission affects multiqubit entanglement;
• whether motional dephasing limits gates or sensing protocols.

The single-mode formulas remain useful, but they must be applied to the correct normal modes and couplings.

Common Mistakes

• Treating fluorescence detection as passive observation rather than a strong scattering process.
• Calling a state measurement nondestructive without checking optical pumping and recoil.
• Using a single heating rate without specifying the motional mode and trap frequency.
• Forgetting that $T_2$ depends on the pulse sequence and noise spectrum.
• Applying Lamb–Dicke sideband formulas when $\eta\sqrt{\langle n\rangle+1}$ is not small.
• Ignoring off-resonant carrier and sideband couplings during cooling or gates.
• Equating high internal-state readout fidelity with preservation of the motional state.
• Treating all ion-chain modes as if they couple identically to electric-field noise.

Exercises

Coherent Sideband or Dissipative Cooling?

An ideal red-sideband $\pi$ pulse maps

$$|g,n\rangle \longrightarrow |e,n-1\rangle.$$

Explain why repeating only this coherent pulse does not constitute cooling, and identify the step that exports entropy in resolved-sideband cooling.

Solution

The sideband pulse is unitary. Once it maps $|g,n\rangle$ to $|e,n-1\rangle$, another identical pulse drives the same two-state subspace back toward $|g,n\rangle$; no entropy has left the ion.

Cooling adds dissipative repumping,

$$|e,n-1\rangle \longrightarrow |g,n-1\rangle,$$

implemented through spontaneous emission or another engineered reset. The emitted field carries away entropy and makes the cycle directional. In the Lamb–Dicke regime, repumping usually preserves the reduced motional number, though recoil supplies a residual heating channel. Repeating the combined coherent transfer and dissipative reset accumulates population near $|g,0\rangle$.

Bright-Dark Threshold

Suppose bright and dark detection counts are Poisson distributed with means $\mu_b$ and $\mu_d$. If the rule is “bright when $n\ge n_{\mathrm{th}}$,” write the two assignment-error probabilities.

Solution

The bright state is misassigned as dark when $n\lt n_{\mathrm{th}}$:

$$P(d\mid b) = \sum_{n=0}^{n_{\mathrm{th}}-1} e^{-\mu_b} \frac{\mu_b^n}{n!}.$$

The dark state is misassigned as bright when $n\ge n_{\mathrm{th}}$:

$$P(b\mid d) = \sum_{n=n_{\mathrm{th}}}^{\infty} e^{-\mu_d} \frac{\mu_d^n}{n!}.$$

Real experiments add state preparation errors, off-resonant pumping, detector dark counts, and finite detection bandwidth.

Cooling Steady State

For

$$\frac{d\langle n\rangle}{dt} = - (\Gamma_- - \Gamma_+) \langle n\rangle + \Gamma_+,$$

find the steady-state occupation when $\Gamma_-\gt\Gamma_+$.

Solution

Set the time derivative to zero:

$$0 = - (\Gamma_- - \Gamma_+) \bar n_{\mathrm{ss}} + \Gamma_+.$$

Solving,

$$\bar n_{\mathrm{ss}} = \frac{\Gamma_+} {\Gamma_- - \Gamma_+}.$$

The condition $\Gamma_-\gt\Gamma_+$ is required for a finite cooling steady state.

Heating and Noise Spectrum

Using

$$\dot{\bar n} = \frac{e^2} {4m\hbar\nu} S_E(\nu),$$

explain how the heating rate scales with ion mass, trap frequency, and electric-field noise at fixed convention for $S_E$.

Solution

At fixed $S_E(\nu)$, the heating rate is inversely proportional to the ion mass $m$ and trap frequency $\nu$. Heavier ions and higher-frequency modes are less sensitive in this simple formula.

The dependence on $S_E(\nu)$ means that only noise near the motional frequency directly contributes to this Markovian heating rate. Changing electrode filtering, surface noise, or trap distance changes the spectral density and therefore the heating.

Cross-Links

• Quantum Harmonic Oscillator
• Rabi Oscillations
• Quantum Optics
• Measurement Backaction
• Quantum-Jump Trajectories
• Radiation Pressure
• Amplitude Damping Master Equation
• Pure Dephasing Master Equation
• Reservoir Engineering
• Quantum Thermometry
• AMO References

References

• D. J. Wineland and W. M. Itano, “Laser cooling of atoms,” Physical Review A 20, 1521-1540 (1979).
• W. Nagourney, J. Sandberg, and H. Dehmelt, “Shelved optical electron amplifier: Observation of quantum jumps,” Physical Review Letters 56, 2797-2799 (1986).
• T. Sauter, W. Neuhauser, R. Blatt, and P. E. Toschek, “Observation of quantum jumps,” Physical Review Letters 57, 1696-1698 (1986).
• J. C. Bergquist, R. G. Hulet, W. M. Itano, and D. J. Wineland, “Observation of quantum jumps in a single atom,” Physical Review Letters 57, 1699-1702 (1986).
• D. Leibfried, R. Blatt, C. Monroe, and D. Wineland, “Quantum dynamics of single trapped ions,” Reviews of Modern Physics 75, 281-324 (2003).
• H. Häffner, C. F. Roos, and R. Blatt, “Quantum computing with trapped ions,” Physics Reports 469, 155-203 (2008).
• R. Blatt and D. Wineland, “Entangled states of trapped atomic ions,” Nature 453, 1008-1015 (2008).
• M. Brownnutt, M. Kumph, P. Rabl, and R. Blatt, “Ion-trap measurements of electric-field noise near surfaces,” Reviews of Modern Physics 87, 1419-1482 (2015).