Applications and Experimental Platforms

Experimental platforms differ in hardware, accessible observables, and dominant noise sources, but they are compared most cleanly through the same questions:

1. Which degrees of freedom define the system?
2. Which degrees of freedom are treated as an environment?
3. Which environmental outputs are actually measured?
4. Which coherent and incoherent rates control the experiment?
5. On what time and energy scales is the effective model trustworthy?

This chapter is an application map. It does not make measurement and open-system theory the canonical home for atomic, optical, condensed-matter, chemical, or sensing hardware. Instead, it shows how the common language of channels, master equations, noise spectra, measurement records, and decoherence is instantiated in each platform.

A Shared Experimental Template

A useful starting point is a partition

$$H_{mathrm{tot}} = H_S+H_E+H_{SE},$$

with interaction

$$H_{SE} = \sum_\alpha S_\alpha\otimes B_\alpha.$$

The partition is a modeling choice. A resonator mode may be part of the system in a cavity-QED calculation, part of an engineered reservoir after adiabatic elimination, or part of a detector chain in a readout model. None of those descriptions is intrinsically preferred; each must be justified by the observables and scales of interest.

After suitable approximations, an unconditional Markovian model often takes the form

$$\dot\rho = -\frac{i}{\hbar}[H,\rho] + \sum_\mu \mathcal D[L_\mu]\rho,$$

where

$$\mathcal D[L]\rho = L\rho L^\dagger - \frac{1}{2} \{L^\dagger L,\rho\}.$$

This equation alone does not specify an experiment. One must also say which channels are observed. For example, a decay operator can be decomposed schematically as

$$L \longrightarrow \left\{ L_{\mathrm{obs}}, L_{\mathrm{unobs}} \right\},$$

with the observed part producing clicks or a continuous current and the unobserved part contributing only to reduced-state evolution. The same unconditional generator can admit several measurement schemes and therefore several trajectory descriptions.

Reading Path

Read this page	Use it for
Quantum Optics	Photon counting, quadrature detection, field damping, correlations, and input–output language.
Cavity QED	Emitter–cavity coupling, cavity loss, cooperativity, Purcell physics, and monitored output fields.
Circuit QED	Dispersive readout, resonator loss, measurement-induced dephasing, Purcell decay, and superconducting-qubit trajectories.
Trapped Ions	Fluorescence readout, motional heating, sideband cooling, dephasing, and quantum jumps.
Neutral Atoms	Imaging, atom loss, photon scattering, dephasing, and Rydberg-state decay.
Spin Qubits	Relaxation, dephasing, charge and hyperfine noise, dynamical decoupling, leakage, and readout.
NV Centers and Solid-State Defects	Optical pumping and readout, magnetic noise, relaxometry, and defect-spin sensing.
Optomechanics	Radiation-pressure coupling, mechanical damping, imprecision, backaction, and sideband cooling.
Mesoscopic Transport	Leads, quantum dots, tunneling rates, electric current, shot noise, and counting statistics.
Molecular and Chemical Environments	Vibronic baths, spectral densities, energy transfer, Redfield and Förster regimes, and nonadiabatic caveats.
Quantum Sensing	Decoherence-limited estimation, backaction, squeezing, noise spectroscopy, and relaxometry.

A reader interested in one device can start with its application page. A reader comparing devices should first fix the system boundary and then compare coherent rates, dissipative rates, measured outputs, and approximation regimes.

For mesoscopic devices, Quantum Dots supplies the complementary confinement, charging-energy, and Coulomb-diamond ledger before the open-system rate and noise description.

Cross-Platform Comparison

Platform	Retained degrees of freedom	Common environment or noise	Typical record
quantum optics	selected field modes, emitters	vacuum and thermal field continua, loss	photon counts, homodyne or heterodyne current
cavity QED	emitter plus one or a few cavity modes	mirror loss, free-space emission, dephasing	transmitted or reflected field, fluorescence
circuit QED	superconducting qubit and resonator modes	dielectric loss, flux and charge noise, transmission lines	amplified microwave quadrature voltage
trapped ions	internal states and selected motional modes	electric-field noise, laser noise, spontaneous emission	state-dependent fluorescence counts
neutral atoms	internal, motional, and sometimes Rydberg states	photon scattering, technical fields, collisions, loss	fluorescence or absorption image, atom-resolved counts
spin qubits	one or several spin and charge states	nuclei, phonons, charge and magnetic noise	charge-sensor current or spin-to-charge outcome
solid-state defects	defect spin and selected nuclear spins	spin bath, phonons, optical cycling imperfections	fluorescence counts or microwave response
optomechanics	optical or microwave mode plus mechanical mode	cavity loss, mechanical thermal bath	output-field quadrature or photon count
mesoscopic transport	dot, island, or conductor states	fermionic leads, phonons, electromagnetic circuit	transferred charge, current, waiting times
molecular systems	electronic states and selected vibrations	solvent and intramolecular vibrational continua	spectra, populations, reaction products
quantum sensing	sensor state plus signal coupling	the signal itself and nuisance noise	population, phase, fluorescence, or continuous current

The table is only a first pass. In every row, degrees of freedom can move across the system–environment boundary. For example, a strongly coupled vibration should often be retained explicitly rather than hidden inside a smooth spectral density.

Rates, Spectra, and Records

Three distinct layers should be kept separate.

Dynamical parameters

Hamiltonian parameters include detunings, tunneling amplitudes, drive strengths, and coherent couplings such as $g$. Dissipative parameters include cavity linewidths, spontaneous-emission rates, heating rates, and relaxation times. They enter the state equation.

Environmental characterization

Noise spectra and bath correlation functions describe fluctuations before they are reduced to a few rates. For a stationary operator $B(t)$, a conventionally ordered spectrum may be written

$$S_{BB}(\omega) = \int_{-\infty}^{\infty} dt\, e^{i\omega t} \langle B(t)B(0)\rangle.$$

The ordering, normalization, and one-sided or two-sided convention must be stated. Positive- and negative-frequency noise need not be equal in a quantum environment.

Detector data

A laboratory record has gain, offset, finite bandwidth, inefficiency, latency, dark counts, dead time, digitization, and classical technical noise. An ideal Poisson or Wiener increment is a model of calibrated data, not a literal description of the entire acquisition chain.

Confusing these layers causes common errors: fitting a detector bandwidth as if it were a physical decoherence rate, treating every Lindblad operator as a monitored event, or assigning a single noise spectrum without saying how it is ordered.

Platform Guides

Quantum optics

Quantum optics supplies the cleanest operational examples of open-system theory. A selected mode may be damped by a continuum, an emitter may radiate into monitored and unmonitored modes, and the outgoing field can be measured by direct detection, homodyne detection, or heterodyne detection.

The key distinction is between an intracavity or source operator and the measured output field. Input–output theory connects them, while a detector model turns the field into a classical record. Photon correlations such as $g^{(2)}(\tau)$ probe statistics that are not contained in a mean intensity alone.

Cavity QED

Cavity QED retains an emitter and a discrete resonator mode. The Jaynes–Cummings coupling $g$ competes with cavity decay $\kappa$ and emitter decay $\gamma$. A frequently used convention defines the single-emitter cooperativity as

$$C = \frac{4g^2}{\kappa\gamma}.$$

Factors of two vary with linewidth conventions, so a numerical claim about strong coupling or cooperativity should state whether quoted rates refer to amplitude decay, energy decay, half-width, or full width. In the dispersive regime the cavity can act as a probe; near resonance it can exchange excitations coherently with the emitter.

Circuit QED

Circuit QED implements analogous light–matter physics with superconducting circuits and microwave resonators. In a dispersive model, the resonator frequency depends on the qubit state through a shift $\chi$. The resulting output phase or amplitude carries information about $\sigma_z$.

The same photons that carry information also dephase the qubit. Measurement rate, measurement-induced dephasing, amplifier efficiency, cavity ring-up, Purcell decay, higher transmon levels, and leakage therefore belong to one consistent readout model. Treating readout as an instantaneous projective measurement hides these experimentally important scales.

Trapped ions

For trapped ions, the retained system commonly includes internal electronic states and selected harmonic motional modes. State-dependent fluorescence maps an internal state to a photon-count distribution, while laser cooling and motional heating are open-system processes acting on the oscillator.

The resolved-sideband condition compares the trap frequency with optical linewidths and drive scales. Motional heating is often summarized by $d\langle n\rangle/dt$, but a physical model should identify the electric-field-noise spectrum sampled at the motional frequency. Fluorescence thresholds are detector decisions built on overlapping count distributions, not fundamental projectors.

Neutral atoms

Neutral-atom experiments retain internal states, center-of-mass motion, and, when relevant, Rydberg levels. Photon scattering from trapping and imaging light produces both information and recoil. Atom loss is different from dephasing: it changes the accessible particle-number sector and may require an enlarged Hilbert space or a trace-decreasing conditional description.

Rydberg blockade models introduce coherent interactions and decay, but their validity depends on level structure, laser detuning, Doppler effects, blackbody-induced transitions, and technical fluctuations. Imaging can be destructive or nondestructive only relative to a stated observable and timescale.

Spin qubits

Spin-qubit models often begin with a two-level spin subspace, then add orbital or charge states when control and readout require them. Energy relaxation is characterized by $T_1$, coherence decay by $T_2$, and pure dephasing by $T_\phi$. For a simple Markovian qubit with independent relaxation and pure dephasing,

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

This relation is not universal. Quasistatic noise, nonexponential decay, pulse-sequence filtering, leakage, and correlated noise can invalidate a single-rate description. Charge-sensor or spin-to-charge readout also requires a model of tunneling, finite bandwidth, and assignment error.

NV centers and other solid-state defects

Defect spins combine optical initialization and fluorescence readout with microwave spin control. Their environments may include nearby nuclear spins, paramagnetic impurities, strain, electric fields, phonons, and optical-cycle dynamics. Different pulse sequences sample different spectral bands.

In sensing, environmental coupling is deliberately retained as signal while other couplings are treated as noise. Ramsey spectroscopy, dynamical decoupling, and $T_1$ relaxometry therefore do not measure one universal noise strength; each has a different filter function or transition-frequency selectivity.

Optomechanics

Optomechanics couples a resonator mode to a mechanical displacement. A standard interaction is

$$H_{\mathrm{int}} = -\hbar g_0 a^\dagger a (b+b^\dagger),$$

where $g_0$ is the single-photon coupling. Under a strong coherent drive, linearization produces an enhanced coupling whose form depends on detuning.

The output field carries displacement information, while radiation-pressure fluctuations exert backaction. Mechanical damping and thermal occupation determine heating; cavity linewidth and mechanical frequency determine sideband resolution. Statements about a standard quantum limit must specify the estimator, bandwidth, loss, and balance between imprecision and backaction.

Mesoscopic transport

Mesoscopic transport treats electronic reservoirs as particle sources and sinks characterized by chemical potentials, temperatures, and spectral densities. A minimal sequential-tunneling model for an empty or occupied dot has rates $\Gamma_L$ and $\Gamma_R$. The measured current is an ensemble average, while individual electron transfers define a counting record.

Rate equations are appropriate only when coherence between relevant states is negligible on the transport timescale. Cotunneling, strong coupling, Kondo correlations, superconducting leads, and non-Markovian circuits require richer descriptions. Full counting statistics carries information beyond the mean current, including shot noise and higher cumulants.

Molecular and chemical environments

Molecular open-system models must choose which electronic states, nuclear coordinates, and solvent modes are explicit. Weakly coupled residual modes can be summarized by a spectral density; strongly coupled or resonant modes often need to be retained.

Redfield theory, Förster transfer, polaron transformations, and hierarchical equations describe different parameter regimes. None is a universal upgrade of the others. The comparison depends on electronic coupling, reorganization energy, bath correlation time, temperature, vibronic structure, and the observable of interest. Positivity failures in perturbative reduced equations should be diagnosed rather than silently clipped away.

Quantum sensing

Quantum sensing turns a coupling to an external parameter into a change in outcome statistics. The sensor is open because signal acquisition, control, readout, and decoherence all involve couplings outside the ideal two-level model.

Sensitivity is not determined by coherence time alone. It depends on signal response, contrast, preparation and readout fidelity, cycle time, estimator, and total averaging time. Entanglement or squeezing can improve an ideal statistical scaling while loss, correlated noise, or imperfect readout removes the advantage. A trustworthy claim states the resource count and the benchmark being beaten.

Choosing the Right Description

The following hierarchy is practical rather than absolute.

Description	Use it when	Essential check
enlarged unitary model	a few environmental modes are coherent and resolvable	convergence with environment truncation
quantum channel	only input and output states at specified times matter	complete positivity and trace preservation
time-local master equation	a reduced state is needed continuously	approximation regime and positivity
quantum trajectory	a time-resolved record is available or simulated	ensemble average recovers the master equation
classical stochastic Hamiltonian	noise acts as an externally prescribed classical process	comparison with quantum detailed balance and backaction
memory-kernel or enlarged-state model	bath correlations persist on system timescales	recovery of Markovian and weak-coupling limits
rate equation	coherences are dynamically irrelevant	separation between dephasing and population timescales

A model is not made more physical merely by adding states or parameters. The useful model is the least elaborate one that retains the measured observables and passes independent validation tests.

A Model-Selection Checklist

Before assigning a channel or master equation to a platform, record the following.

1. System boundary: List every retained state or mode and every eliminated one.
2. Frame and units: State rotating frames, whether frequencies are angular or cyclic, and whether $\hbar=1$.
3. Coherent scales: Give splittings, detunings, couplings, drives, and tunneling amplitudes.
4. Dissipative scales: Give relaxation, dephasing, loss, heating, and detector rates with conventions.
5. Bath evidence: Identify measured spectra, correlation times, temperatures, and stationarity assumptions.
6. Observed outputs: Specify which channels are monitored and the detector efficiency and bandwidth.
7. Approximation chain: Name the weak-coupling, rotating-wave, Markov, secular, linearization, or elimination steps used.
8. Initial correlations: State whether a product system–environment state is assumed.
9. Validation observable: Reserve at least one decay curve, spectrum, correlation function, or record statistic for testing.
10. Failure regime: Say what parameter change would invalidate the model.

The Approximation Checklist expands this audit for open-system derivations.

One Damped Mode, Three Physical Roles

Consider a bosonic mode with Hamiltonian $\hbar\omega a^\dagger a$ and loss rate $\kappa$. Formally, each of the following may contain the term

$$\kappa\mathcal D[a]\rho.$$

In an optical cavity, $a$ may be the field whose leaked photons are counted. In circuit QED, it may be a readout resonator whose amplified quadrature estimates a qubit state. In optomechanics, it may both probe mechanical displacement and exert radiation-pressure backaction.

The identical dissipator does not make the experiments equivalent. The coupling Hamiltonian, monitored fraction of the loss, input state, detector observable, calibration chain, and scientific question determine the operational meaning.

Canonical Boundaries

This chapter owns comparative maps of how measurement and open-system ideas appear in experimental platforms.

• Open Quantum Systems owns system–bath reductions and approximation logic.
• Markovian Master Equations owns GKSL generators and Lindblad dynamics.
• Continuous Measurement and Quantum Trajectories owns record-conditioned evolution.
• Quantum Noise, Dissipation, and Baths owns spectra, correlation functions, fluctuation–dissipation relations, and input–output theory.
• Quantum Control and Feedback owns control protocols, feedback, and reservoir engineering.
• Quantum Thermodynamics owns work, heat, entropy production, and thermodynamic resource accounting.
• Hardware-focused volumes own detailed device construction, spectroscopy, materials, and many-body phase structure.
• Quantum Information owns full treatments of metrology resources, error correction, tomography, and information-processing performance.

Application pages should cross-link to those canonical derivations rather than reproducing them.

Common Mistakes

• Calling a platform “open” without defining the retained system.
• Treating a fitted Lindblad rate as a microscopic explanation.
• Mixing angular-frequency and ordinary-frequency conventions in rate ratios.
• Identifying every collapse operator with a detector click.
• Ignoring the unmonitored fraction of a physical output channel.
• Equating loss, dephasing, leakage, and detector assignment error.
• Using $T_2=2T_1$ as a universal identity rather than a relaxation-limited bound in a simple model.
• Applying a Markovian generator when the environment contains a resolvable coherent mode.
• Reporting a noise spectrum without its ordering and normalization convention.
• Inferring a quantum advantage without stating resources, overhead, and the classical benchmark.
• Validating a model only on the data used to fit its parameters.
• Treating numerical positivity repair as evidence that an approximation is physically valid.

Exercises

Observed and unobserved cavity loss

A cavity has a monitored output port with rate $\kappa_m$ and internal loss with rate $\kappa_i$. Write collapse operators for the two channels, give the total unconditional dissipator, and identify the ideal monitored photon flux.

Solution

Choose

$$L_m=\sqrt{\kappa_m}\,a, \qquad L_i=\sqrt{\kappa_i}\,a.$$

The unconditional damping is

$$\mathcal D[L_m]\rho + \mathcal D[L_i]\rho = (\kappa_m+\kappa_i) \mathcal D[a]\rho.$$

For ideal direct detection of the monitored port, the conditional count rate is

$$\lambda_m = \operatorname{Tr} (L_m^\dagger L_m\rho_c) = \kappa_m \langle a^\dagger a\rangle_c.$$

Internal loss contributes to state damping but not to this record.

Cooperativity and conventions

Using the convention $C=4g^2/(\kappa\gamma)$, calculate $C$ for $g/(2\pi)=10\,\mathrm{MHz}$, $\kappa/(2\pi)=2\,\mathrm{MHz}$, and $\gamma/(2\pi)=1\,\mathrm{MHz}$. Why must the linewidth convention still be reported?

Solution

Because all three quantities use the same conversion between cyclic and angular frequency, the factors of $2\pi$ cancel:

$$C = \frac{4(10)^2}{(2)(1)} = 200.$$

The number is meaningful only with the stated definition. Some sources use field-amplitude decay rather than energy decay, or half-width rather than full width, which moves factors of two among $\kappa$, $\gamma$, and the definition of $C$.

Relaxation and pure dephasing

A qubit has $T_1=40\,\mu\mathrm{s}$ and a measured exponential $T_2=30\,\mu\mathrm{s}$. Assuming the simple independent Markovian model, find $T_\phi$ and check the relaxation-limited bound.

Solution

Use

$$\frac{1}{T_\phi} = \frac{1}{T_2} - \frac{1}{2T_1}.$$

Thus

$$\begin{aligned} \frac{1}{T_\phi} &= \frac{1}{30} - \frac{1}{80} \quad \mu\mathrm{s}^{-1}, \\ T_\phi &\simeq 48\,\mu\mathrm{s}. \end{aligned}$$

The simple model requires $T_2\le 2T_1$. Here $30\,\mu\mathrm{s}<80\,\mu\mathrm{s}$, so the bound is satisfied. Agreement with the bound does not establish that the decay is truly Markovian or single exponential.

Sequential-tunneling current

A dot can be empty or occupied. Electrons enter from the left at rate $\Gamma_L$ and leave to the right at rate $\Gamma_R$. Find the stationary occupation and mean current.

Solution

If $p_1$ is the occupied probability, then

$$\dot p_1 = \Gamma_L(1-p_1) - \Gamma_Rp_1.$$

At stationarity,

$$p_1^{\mathrm{ss}} = \frac{\Gamma_L} {\Gamma_L+\Gamma_R}.$$

Each rightward transition transfers charge $e$, so

$$I = e\Gamma_Rp_1^{\mathrm{ss}} = e \frac{\Gamma_L\Gamma_R} {\Gamma_L+\Gamma_R}.$$

This rate model omits coherent tunneling, reverse transport, cotunneling, and detector bandwidth.

Decoherence-limited Ramsey time

Suppose a Ramsey sensor has contrast $e^{-t/T_2}$, phase response proportional to $t$, negligible dead time, and independent repetitions during a fixed total duration. Show that the information rate is maximized at $t=T_2/2$ in this simplified model.

Solution

Near the most sensitive operating point, the Fisher information per shot scales as

$$\mathcal I_1(t) \propto t^2e^{-2t/T_2}.$$

The number of shots in fixed total time scales as $1/t$, so the total information rate scales as

$$\frac{\mathcal I_1(t)}{t} \propto te^{-2t/T_2}.$$

Differentiating its logarithm gives

$$\frac{d}{dt} \left( \ln t- \frac{2t}{T_2} \right) = \frac{1}{t} - \frac{2}{T_2}.$$

The stationary point is therefore $t=T_2/2$, and it is a maximum. Dead time, nonexponential contrast, adaptive protocols, and control overhead can shift the optimum.

Cross-Links

• Quantum Dots
• Quantum Optics
• Cavity QED
• Circuit QED
• Trapped Ions
• Neutral Atoms
• Spin Qubits
• NV Centers and Solid-State Defects
• Optomechanics
• Mesoscopic Transport
• Molecular and Chemical Environments
• Quantum Sensing
• Open Quantum Systems
• Markovian Master Equations
• Continuous Measurement and Quantum Trajectories
• Common Noise Spectra
• Approximation Checklist

References

• H.-P. Breuer and F. Petruccione, The Theory of Open Quantum Systems, Oxford University Press (2002).
• C. W. Gardiner and P. Zoller, Quantum Noise, 3rd ed., Springer (2004).
• A. A. Clerk, M. H. Devoret, S. M. Girvin, F. Marquardt, and R. J. Schoelkopf, “Introduction to quantum noise, measurement, and amplification,” Reviews of Modern Physics 82, 1155–1208 (2010).
• S. Haroche and J.-M. Raimond, Exploring the Quantum: Atoms, Cavities, and Photons, Oxford University Press (2006).
• A. Blais, A. L. Grimsmo, S. M. Girvin, and A. Wallraff, “Circuit quantum electrodynamics,” Reviews of Modern Physics 93, 025005 (2021).
• D. Leibfried, R. Blatt, C. Monroe, and D. Wineland, “Quantum dynamics of single trapped ions,” Reviews of Modern Physics 75, 281–324 (2003).
• M. Aspelmeyer, T. J. Kippenberg, and F. Marquardt, “Cavity optomechanics,” Reviews of Modern Physics 86, 1391–1452 (2014).
• R. Hanson, L. P. Kouwenhoven, J. R. Petta, S. Tarucha, and L. M. K. Vandersypen, “Spins in few-electron quantum dots,” Reviews of Modern Physics 79, 1217–1265 (2007).
• C. L. Degen, F. Reinhard, and P. Cappellaro, “Quantum sensing,” Reviews of Modern Physics 89, 035002 (2017).
• U. Weiss, Quantum Dissipative Systems, 4th ed., World Scientific (2012).