PaperlayerView exact revision on arXiv

Source-linked AI summary

The SLH framework for modeling quantum input-output networks

Joshua Combes, Joseph Kerckhoff, Mohan Sarovar

arXiv:1611.00375v2quant-ph

TL;DR

Quantum input-output networks require models that capture localized quantum systems, propagating fields, and their interconnections. This review explains the SLH framework, using operator triples and algebraic composition to model arbitrary networks, and surveys its applications, extensions, and limitations. The framework provides modular, control-oriented modeling for coherent networks, while its scope remains constrained by approximations such as Markovian dynamics and by practical technology limitations.

  • Problem

    Quantum technologies need precise models for networks of quantum systems connected by propagating fields, including arbitrary interconnections and coherent feedback.

  • Method

    The review explains how SLH triples represent localized components and how algebraic composition, quantum stochastic equations, and input-output theory model interconnected networks.

  • Results

    The review synthesizes SLH applications and extensions, including coherent feedback, light-matter interactions, squeezed-state sources, distributed Kerr media, and delayed propagation.

  • Takeaways & Limitations

    SLH offers a modular framework for control-theoretic analysis and design of quantum input-output networks, while also supporting models of materials and light-matter interactions.

  • Takeaways & Limitations

    The framework relies in part on Markov and short-delay approximations, while general long-delay and complex feedback networks remain incompletely modeled.

Abstract

from arXiv · show

Many emerging quantum technologies demand precise engineering and control over networks consisting of quantum mechanical degrees of freedom connected by propagating electromagnetic fields, or quantum input-output networks. Here we review recent progress in theory and experiment related to such quantum input-output networks, with a focus on the SLH framework, a powerful modeling framework for networked quantum systems that is naturally endowed with properties such as modularity and hierarchy. We begin by explaining the physical approximations required to represent any individual node of a network, eg. atoms in cavity or a mechanical oscillator, and its coupling to quantum fields by an operator triple $(S,L,H)$. Then we explain how these nodes can be composed into a network with arbitrary connectivity, including coherent feedback channels, using algebraic rules, and how to derive the dynamics of network components and output fields. The second part of the review discusses several extensions to the basic SLH framework that expand its modeling capabilities, and the prospects for modeling integrated implementations of quantum input-output networks. In addition to summarizing major results and recent literature, we discuss the potential applications and limitations of the SLH framework and quantum input-output networks, with the intention of providing context to a reader unfamiliar with the field.

I. INTRODUCTION

The review presents SLH as a modular framework for modeling and analyzing quantum input-output networks, from individual component descriptions through interconnected systems and feedback. It surveys the framework’s physical foundations, applications, extensions, and practical limitations.

  • Framework and workflow: Connecting component modules into networks generates more complex behavior, including coherent feedback and feedforward interactions.The framework treats components as black boxes with prespecified input-output behavior and supports control-theoretic analysis.
  • Framework and workflow: SLH represents localized quantum components as triples (S, L, H) that capture their interactions with incident and scattered fields.These triples support systematic derivation of equations of motion for field modes and internal degrees of freedom.
  • Framework and workflow: The review develops SLH from quantum input-output theory, physical approximations, cascading, quantum stochastic equations, and algebraic network composition.It emphasizes physical motivation and intuition while covering the workflow from component specification to network analysis.
  • Applications and advantages: Coherent feedback can outperform measurement-based feedback because it processes both non-commuting output quadratures without the same additional noise.The cited results include lower cost for a coherent controller in a linear-quadratic-Gaussian control problem and squeezing dynamics unavailable to measurement-based control in a two-quadrature setting.
  • Applications and advantages: SLH has been applied to light-matter interactions, quantum nondemolition detection, distributed Kerr media, and models of finite-bandwidth squeezed states.These applications extend beyond basic network construction to modeling fundamental properties of materials and light-matter interactions.
  • Limitations and outlook: The framework’s validity and adoption remain bounded by modeling assumptions, immature quantum technologies, and unresolved treatment of long propagation delays.The Markov approximation assumes a memoryless bath, while some delay treatments are valid only when γL ≪ν; practical adoption also depends on costs and convenience.

B. Cascaded systems

Cascading routes one system’s output unidirectionally into another, allowing the composite network to be represented by effective dynamics while preserving directional information flow.

  • B. Cascaded systems: Cascading connects one cavity’s output to another cavity’s input through a unidirectional probe field without back scattering.A circulator can enforce the required propagation direction.
  • B. Cascaded systems: Cascaded coupling breaks time-reversal symmetry by establishing a clear direction of information flow between components.This distinguishes cascading from simple Hamiltonian coupling.
  • B. Cascaded systems: Taking the inter-cavity propagation time to zero eliminates the intervening field degrees of freedom and yields effective operators Heff and Leff for the composite system.The approximation is appropriate when propagation is much faster than the relevant 1/γ_i dynamics, but can fail at GHz rates over centimeter-scale distances.
  • B. Cascaded systems: The effective composite model extends beyond two cavities to more complex interconnected networks, including feedback connections.The framework generalizes the first-principles treatment of cascaded systems.
  • B. Cascaded systems: The first cavity drives the second cavity’s evolution, while the second does not drive the first, demonstrating directional information flow.The output field contains information from both cavities.

IV. QUANTUM STOCHASTIC DIFFERENTIAL EQUATIONS

Quantum stochastic differential equations provide the calculus for integrating inherently stochastic quantum dynamics, with distinct Stratonovich and Itō forms and conversion rules between them.

  • IV. QUANTUM STOCHASTIC DIFFERENTIAL EQUATIONS: Quantum stochastic dynamics require specialized integration because broadband input fields have singular commutation relations and nonzero second-order increment products.Time-integrated field quantities are introduced to handle these singularities.
  • IV. QUANTUM STOCHASTIC DIFFERENTIAL EQUATIONS: Quantum noise increments are noncommuting analogues of classical Wiener-process increments.Their products can be proportional to dt under vacuum expectation.
  • IV. QUANTUM STOCHASTIC DIFFERENTIAL EQUATIONS: Itō and Stratonovich integrals differ for quantum stochastic increments because those increments remain irregular in the continuum limit.The two forms are equivalent only when increments are regular, as in standard calculus.
  • IV. QUANTUM STOCHASTIC DIFFERENTIAL EQUATIONS: Physical QSDEs typically arise in Stratonovich form, whereas Itō form is easier for analytical and numerical calculations.A straightforward conversion procedure relates the two representations.
  • IV. QUANTUM STOCHASTIC DIFFERENTIAL EQUATIONS: The review subsequently works exclusively with QSDEs in Itō form, using vacuum Itō tables to derive localized-system equations of motion.The single-sided resonator example recovers the corresponding Itō equation of motion.

V. GENERAL QUANTUM INPUT-OUTPUT NETWORKS AND THE SLH FRAMEWORK

The SLH framework represents localized quantum components with operator triples and combines them into network models whose dynamics are fixed by their system-field interactions.

  • V. GENERAL QUANTUM INPUT-OUTPUT NETWORKS AND THE SLH FRAMEWORK: The SLH framework was developed to describe quantum input-output networks of localized components interacting through itinerant quantum bosonic fields.It aims to support modular modeling of large assemblies of quantum coherent systems.
  • V. GENERAL QUANTUM INPUT-OUTPUT NETWORKS AND THE SLH FRAMEWORK: Under weak-coupling and Markov approximations, a component’s system-field dynamics are represented by a unitary propagator parameterized by (S, L, H).The coupling is assumed weak and linear, with Markovian field interactions.
  • V. GENERAL QUANTUM INPUT-OUTPUT NETWORKS AND THE SLH FRAMEWORK: L couples localized system operators to external fields, H is the system Hamiltonian, and S describes scattering of input fields.For multiple modes, S and L become operator-valued matrices and vectors subject to unitarity constraints.
  • V. GENERAL QUANTUM INPUT-OUTPUT NETWORKS AND THE SLH FRAMEWORK: Specifying an SLH triple together with the initial component and field states completely determines the network’s properties.The coupling prescription fixes the time evolution of all network components.
  • V. GENERAL QUANTUM INPUT-OUTPUT NETWORKS AND THE SLH FRAMEWORK: Phase shifters, beamsplitters, and cavities can each be represented by SLH triples, with scattering constraints applying to components such as beamsplitters.Components without internal degrees of freedom have no system Hamiltonian or coupling operators.

B. SLH composition rules

SLH composition rules construct arbitrary quantum input-output networks algebraically by grouping components, routing channels, and eliminating internal feedback connections.

  • B. SLH composition rules: The SLH composition rules provide algebraic prescriptions for combining component triples according to their field interconnections.They are the central mechanism for composing network models without repeating first-principles derivations.
  • B. SLH composition rules: The composition rules assume Markovian interactions, dispersionless linear propagation with negligible delay, and vacuum input fields.These assumptions define the physical scope of the basic rules.
  • B. SLH composition rules: The series product cascades matched outputs into inputs, while concatenation groups components with independent input-output fields and direct coupling adds a Hamiltonian interaction.The series product is generally order-dependent because linking fields are directional.
  • B. SLH composition rules: Feedback reduction connects an output port to an input port, removes the internal link, and produces a reduced SLH triple.The reduced system depends on which ports are connected.
  • B. SLH composition rules: Multiple-port interconnections require careful index tracking and may use permutations, padding, and block-contiguous elimination of internal nodes.Permutation components model rerouting, while padding accommodates unequal port counts.
  • B. SLH composition rules: Arbitrary networks are built by concatenating all components first and then applying feedback reduction to implement their connections.The series product is a special case of feedback reduction.

C. Network Heisenberg equation of motion and network input-output relations

The SLH framework derives network output fields and localized-system dynamics from a network’s SLH triple using quantum stochastic calculus. These relations also clarify how cascading and field connections shape component dynamics and outputs.

  • The network SLH triple determines output-field increments through scattering and coupling operators in the Heisenberg picture.The general relation is expressed for multiple input and output fields.
  • For a single port, the input-output relation reduces to the familiar form when S = I and L = √γa.This recovers the Itō input-output relation for a cavity.
  • Zero-delay cascading makes the first component’s output immediately become the second component’s input, so downstream dynamics depend on upstream fields and operators.The cascade relation is obtained by identifying dBin(2, t) with dBout(1, t).
  • The Heisenberg equation of motion for arbitrary network operators follows from the propagator and quantum stochastic calculus.The framework provides general multiple-port and single-port forms for these equations.
  • In the two-cavity example, cascading causes the second cavity to be driven by the first cavity and both input fields, while an output contains superpositions of both cavity modes.The network is assembled from cavity, beamsplitter, and padding components using series and concatenation products.

D. Master equation description

Tracing over vacuum input-output fields converts the stochastic network description into a deterministic master equation for the localized degrees of freedom. If some outputs are monitored instead, conditioned stochastic master equations describe the localized dynamics.

  • Tracing over propagating fields yields a statistical description of localized-system dynamics by averaging over field degrees of freedom.The construction assumes vacuum input fields and uses the factorized initial state.
  • The resulting master equation is deterministic because the stochastic field quantities have been averaged over.
  • When selected output fields are monitored, the localized systems can instead be described by stochastic master, quantum trajectory, or quantum filtering equations.
  • Under the vacuum-input assumption, the scattering matrix S does not appear in the localized-system master equation.For non-vacuum inputs, the master equation can depend on S.

VI. LINEAR QUANTUM NETWORKS

Linear quantum networks model harmonic modes with quadratic Hamiltonians and linear field couplings, optionally retaining linearity under Gaussian measurements. Their dynamics and input-output behavior can be represented using techniques adapted from classical linear systems theory.

  • The framework restricts attention to linear quantum networks because they are experimentally accessible and extensively studied, particularly in the optical regime.
  • Linear quantum networks contain harmonic localized modes with quadratic Hamiltonians and linear couplings to propagating fields.Gaussian measurements such as homodyne and heterodyne detection can preserve a linear description.
  • Classical linear-systems techniques have been transferred to quantum linear systems, including methods relevant to control.

A. Passive linear quantum networks

Passive and active linear quantum networks admit SLH and ABCD descriptions of their internal dynamics and input-output behavior, but quantum constraints restrict the allowable ABCD matrices. Coherent feedback provides an example in which an active optical network changes squeezing behavior.

  • A. Passive linear quantum networks: Passive linear QIONs use scalar scattering, couplings linear in annihilation operators, and quadratic Hamiltonians that conserve total photon number.
  • A. Passive linear quantum networks: For linear QIONs, specifying the ABCD matrices is equivalent to specifying the network with an SLH triple.The ABCD matrices are determined by the coupling and Hamiltonian parameters.
  • A. Passive linear quantum networks: Quantum ABCD matrices are not freely specifiable because unitarity and the uncertainty principle impose dependencies among A, B, C, and D.
  • A. Passive linear quantum networks: The transfer function matrix is sufficient to specify a passive QION’s input-output behavior in the Laplace domain.
  • A. Passive linear quantum networks: Active linear QIONs include components such as squeezers and amplifiers, requiring doubled-up state vectors and matrices while retaining linear equations of motion.
  • A. Passive linear quantum networks: In coherent-feedback squeezing enhancement, deamplification of one input quadrature increases as η → 0 while the conjugate quadrature is amplified by the same amount.The quadrature phase space becomes increasingly squeezed while preserving total area.

C. Survey of results regarding linear quantum networks

Linear quantum networks inherit central analysis concepts from classical linear systems, while coherent-feedback design must additionally satisfy quantum physical-realizability constraints. The review surveys controller design, robustness, synthesis, and extensions of the SLH framework for practical network modeling.

  • Survey of results regarding linear quantum networks: Hurwitz stability, controllability, and observability extend to linear quantum networks through eigenvalue and matrix-rank characterizations.Hurwitz stability is determined by the eigenvalues of the A matrix, while controllability and observability use analogous rank conditions.
  • Survey of results regarding linear quantum networks: Coherent-feedback control designs controllers from network outputs, but quantum realizability determines whether the specified controller can be physically implemented.Unlike classical control, quantum controller specifications must correspond to hardware realizable with standard optical components.
  • Survey of results regarding linear quantum networks: Linear quantum systems satisfying the realizability conditions preserve the canonical commutation relations of the underlying system degrees of freedom.This preservation supplies the fundamental physical-realizability requirement for the linear model.
  • Survey of results regarding linear quantum networks: Optimal and robust coherent-controller design remains comparatively underdeveloped, particularly when model uncertainty or performance guarantees must be addressed.Quantum LQG design is complicated by realizability constraints, and the resulting optimal controller need not guarantee stability or robustness.
  • Survey of results regarding linear quantum networks: Controller synthesis addresses how realizable coherent controllers can be constructed from basic optical components.This extends controller design from specifying a physically valid system to realizing it as an optical network.
  • Survey of results regarding linear quantum networks: SLH extensions preserve modular network structure while representing non-vacuum inputs, experimental imperfections, and more complex component behavior.A common strategy is to approximate complex behavior through interactions among freely propagating fields and standard SLH building blocks.

1. Coherent states

The review describes source-system and cascaded-network constructions for coherent, squeezed, Fock, and related non-vacuum field states. These models translate field-state preparation into SLH components and derive corresponding system dynamics.

  • Coherent states: Continuous-mode coherent states describe pulsed laser light through a temporal profile α(t), with single-mode coherent states and vacuum as special cases.The wave-packet profile is square-normalized, and the mean photon number is determined by |α|2 and the profile integral.
  • Coherent states: A coherent-state source can be cascaded into a target system, allowing vacuum-driven SLH dynamics to reproduce coherent driving.The source coherently displaces the field, and the resulting construction supports derivation of the coherent-state master equation.
  • Coherent states: The coherent-state master equation contains terms involving the scattering operator S when the system is driven by a non-vacuum field.The review presents equivalent Lindblad-form and expanded expressions for the dynamics.
  • Finite-bandwidth squeezed states: Finite-bandwidth squeezed light can be modeled with a cavity-based source, whose output has quadrature squeezing determined by the source parameters.The model explicitly represents a degenerate optical parametric oscillator and identifies the squeezed quadrature.
  • Fock states: Fock states contain temporal superpositions that induce temporal correlations, so their interacting systems are non-Markovian but can be represented using a larger Markovian source model.Source constructions include two-level-atom models for single-photon and vacuum superpositions and generalizations to Fock states.

4. Cat states

The review presents cat-state and broader non-vacuum-input methods, including basis decompositions, generalized state equations, and Gaussian-state extensions. These approaches broaden SLH modeling while introducing representational trade-offs.

  • 4. Cat states: Cat states are continuous-mode superpositions of coherent states with distinct temporal amplitudes and normalized complex weights.Source constructions use either a qudit with time-dependent couplings or a multimode cavity.
  • Alternatives to source models: When source-model construction is difficult, arbitrary input fields can be decomposed into workable bases such as Fock, N-photon, selected multiphoton, or cat states.The field may be truncated in the chosen basis before deriving the dynamics.
  • Fock-state input calculations: For Fock-state inputs, generalized state matrices obey coupled master equations that yield system observables and output quantities such as mean photon flux.The method has been extended to multiple input-output modes and spectrally entangled states.
  • General Gaussian input states: Gaussian states encompass experimentally important coherent, squeezed, and thermal fields and are characterized by first and second moments.Their covariance parameters obey a quantum-validity constraint, with M=0 describing thermal states and nonzero M indicating squeezing.
  • General Gaussian input states: The SLH composition rules can model arbitrary Gaussian inputs, but the resulting network triple must be interpreted through corresponding Stratonovich dynamical equations.This trade-off applies when modular intrinsic component descriptions, general composition rules, and direct non-vacuum Gaussian inputs are all required.

C. Emission and propagation losses

The SLH framework models emission and propagation losses with fictitious modes or beam splitters, while representing circulator imperfections and bidirectional propagation through expanded network models. Finite propagation distances can be incorporated as phase shifts, but the approximation requires negligible delays.

  • Emission and propagation losses: Losses are modeled by introducing fictitious modes or ports with vacuum inputs and tracing over the unmonitored modes.Distributed propagation losses can be represented by fictitious beam splitters, effectively adding fictitious output ports.
  • Circulator non-idealities: Real circulator imperfections include loss, imperfect isolation, impedance mismatch, backreflection, and finite bandwidth.Finite-bandwidth models have been developed for three-port, four-port, and more general circulators.
  • Circulator non-idealities: Circulator scattering is represented by a unitary S matrix whose transmission, reflection, and isolation coefficients obey normalization and orthogonality constraints.For a symmetric imperfect circulator, the coefficients satisfy |t|^2 + |r|^2 + |b|^2 = 1 and rt* + tb* + br* = 0.
  • Circulator non-idealities: |t| ≫ |r|, |b| is desirable for a circulator with strong transmission and small reflection and isolation errors.
  • Bidirectional propagation: Bidirectional waveguides require separate right- and left-propagating modes, because ordinary SLH cascades naturally describe co-propagating fields.Correct counter-propagation modeling requires cascading components for each mode and then concatenating the modes.
  • Finite-length propagation: A finite distance L between cascaded components is represented by a phase shift, which can generate nontrivial coupling between components and different effective Hamiltonians for counter- and co-propagation.The phase-shift treatment is valid when γL ≪ ν; time-delayed propagation requires other methods outside this regime.

F. Model reduction by adiabatic elimination of fast degrees of freedom

Adiabatic elimination reduces SLH models by removing fast degrees of freedom under a timescale-separation limit. The reduced model acts on the slow subspace, and elimination commutes with network composition under stated conditions.

  • Reduction procedure: Adiabatic elimination approximates a fast-timescale SLH node by a slow model acting on a projected subspace H0.The original operators depend on a parameter k that scales fast rates; the reduced triple removes this dependence in the k → ∞ limit.
  • Reduction conditions: The procedure requires pre-elimination operators with appropriate fast, slow, and coupling components and additional preconditions for the limiting propagator to converge.Operators depending on k^2 generate fast dynamics, k-dependent operators couple timescales, and k-independent operators describe slow dynamics.
  • Reduction procedure: The projection P0 selects slow dynamics, while P1 = I − P0 projects onto fast dynamics that are eliminated.Choosing P0 requires physical insight into the network dynamics and must satisfy structural conditions.
  • Applications and limitations: Adiabatic elimination can reduce QION model complexity and simulation costs, but subtleties arise when scaling coherent-state input amplitudes.Special procedures were developed to address mathematical complications in that setting.
  • Composition and reduction: Adiabatic elimination and SLH network composition produce the same final SLH parameters whether elimination occurs before or after composition.This commutativity result was established first in a special setting and later in full generality.
  • Example: cavity QED: In a cavity-QED example, the large-κ, large-g limit restricts the internal system to the joint ground state with no effective Hamiltonian or coupling dynamics.The guided input reflects with an additional π phase shift, while the fictitious spontaneous-emission mode acquires no phase shift.

G. Modeling distributed transformations

The SLH framework can model spatially distributed field transformations by approximating a continuous medium as many cascaded discrete components and taking a continuum limit. A gradient echo memory illustrates this construction, while the approach remains relatively unexplored.

  • Distributed field transformations can be approximated by cascading many discrete components that implement infinitesimal transformations, then taking their continuum limit.
  • Figure 12 maps the construction from an experimental schematic, to a discrete SLH cascade of N cavities, to a continuum SLH model spanning x = 0 to x = L.
  • A gradient echo memory is modeled by dividing its spatially distributed atomic ensemble into thin slices whose outputs feed the next slice.In the weak excitation limit, each slice can be approximated as a coherent exchange with a bosonic mode, formally resembling a cavity.
  • In the continuum limit, the k’th cavity is located at x(k) = k × ∆x with ∆x = L/N, and N →∞ converts the discrete spacing into dx.
  • The continuum model determines output fields through the input-output relation, but solving the output requires finding the local modes a(x, t).
  • Modeling material properties through continuum limits of SLH networks remains relatively unexplored and has significant potential.

H. SLH and scattering theory

The SLH framework connects network input-output dynamics with scattering theory, allowing frequency-domain scattering matrices to be derived from SLH descriptions. This connection extends scattering calculations from individual components to arbitrary SLH networks, but standard SLH assumptions constrain dispersive propagation.

  • The S-matrix is a unitary matrix connecting asymptotic input and output field states, providing the central scattering-theory link to SLH models.
  • SLH input-output relations can be used to calculate scattering-matrix elements by relating frequency-domain input and output fields to the scattering interaction.
  • For a single photon, the output remains a wavepacket whose profile is deformed by the scattering interaction through the matrix Sω,ν.
  • The framework derives a single-photon scattering expression for a two-level atom and has been extended to multiphoton, multimode, and coherent-state scattering cases.
  • Scattering calculations can in principle be performed for arbitrary networks of SLH components, including networks with finite spatial distances between scattering elements.
  • Standard SLH assumes dispersionless propagation with negligible travel time, and dispersive propagation can invalidate the Markov approximation.This limitation is especially relevant to integrated implementations such as silicon photonic waveguides.

J. Time delayed field propagation

The review describes three approaches for modeling finite propagation delays beyond the zero-delay SLH approximation, while emphasizing that none is yet complete. These approaches introduce fictitious components, cascade past outputs, or explicitly discretize fields for tensor-network simulation.

  • J. Time delayed field propagation: The standard SLH framework assumes negligible propagation delay, an approximation that becomes questionable in physically large networks and feedback configurations.The assumption is especially problematic when propagation is long relative to component dynamics or when fields propagate in both directions.
  • Approach 1: introduce fictitious SLH components: Approach 1 models time delays by introducing additional fictitious SLH components, treating delay as a special case of linear dispersion.This strategy is motivated by earlier methods for approximately modeling dispersive propagation.
  • Approach 3: explicit representation of the in-loop fields: Approach 3 explicitly discretizes input, output, and in-loop fields and simulates the resulting representation with tensor-network methods.Explicit field modeling permits calculation of subsystem, input/output-field, and in-loop-field observables, including arbitrary propagation delays.
  • Approach 2: cascades from the past: Approach 2 represents delayed feedback by cascading systems driven by outputs from past systems; simulating through k delay intervals requires k cascades.The approach defines a propagator for the entire delayed system and can yield reduced system states after tracing out auxiliary degrees of freedom.
  • Approach 3: explicit representation of the in-loop fields: Explicit field representations become infeasible as networks grow because their degrees of freedom become intractable despite concise matrix-product-state descriptions.This trade-off contrasts with SLH’s philosophy of eliminating intermediary fields from the model.
  • Integrated implementations: For integrated platforms, SLH models capture many superconducting microwave systems but face limitations from dispersion, scattering, heating, loss, and non-Markovian effects in photonics and general microwave networks.Nonlinear scattering in silicon photonics usually cannot be represented within standard SLH because it generally does not couple to a Markovian reservoir.
  • IX. OUTLOOK: The review identifies systematic relaxation of the zero-delay approximation, through successive corrections toward fully delayed models, as an important direction for QION modeling.Such work would help connect canonical SLH models with models that fully represent propagation delay.

Appendix A: SLH representation of some basic components

The appendix collects SLH triples and scattering descriptions for commonly encountered quantum network components. Examples span passive optical elements, cavities, nonlinear and optomechanical systems, atoms, circulators, and related devices.

  • Appendix A: SLH representation of some basic components: The appendix lists SLH triples for commonly encountered network components, providing basic building blocks for QION models.The catalog includes both passive components and dynamical quantum systems.
  • Passive and cavity components: A phase shifter, beam splitter, coherent drive, one-sided cavity, Kerr cavity, and Fabry–Perot cavity represent basic optical network elements.Beam-splitter scattering matrices must satisfy the unitarity condition S†S = I.
  • Nonlinear and optomechanical systems: Crossed cavities, degenerate OPOs, two-mode squeezing systems, and optomechanical cavities extend the catalog to nonlinear and hybrid light–matter components.The optomechanical model includes photon decay, optical and mechanical detunings, electromagnetic–mechanical coupling, and thermal phonon-bath parameters.
  • Atomic and atom–cavity systems: Two-level atoms coupled to waveguides, atoms in harmonic traps, and atom–cavity systems provide SLH descriptions of emitters and open quantum systems.The atom–cavity examples include both Rabi and rotating-wave Jaynes–Cummings models.
  • Circulators: Three-port and four-port circulators are described by scattering matrices whose coefficients obey unitarity constraints and quantify transmission, reflection, and isolation errors.For a symmetric imperfect three-port circulator, desirable operation has |t| much greater than |r| and |b|.
Loading 1611.00375v2…