PaperlayerView exact revision on arXiv

Source-linked AI summary

Quantum Plasmonic Sensors

Changhyoup Lee, Benjamin Lawrie, Raphael Pooser, Kwang-Geol Lee, Carsten Rockstuhl, Mark Tame

arXiv:2012.01525v2quant-phcond-mat.mes-hallphysics.app-phphysics.optics

TL;DR

Plasmonic sensors are approaching the shot-noise limit, prompting interest in applying quantum metrology to them. This review organizes the quantum-plasmonic framework and key works; beyond-shot-noise performance has been demonstrated in practical optical sensors, though progress in squeezed interferometric plasmonic sensors remains limited.

  • Problem

    Plasmonic sensors are beginning to reach the sensitivity and precision limit set by quantum fluctuations of light, motivating the question of whether quantum-metrology techniques can improve their performance.

    Source: §I · p. 2

  • Method

    The review introduces the general framework and basic theory of quantum plasmonic sensing, surveys the research landscape, and examines key works and their operating principles.

    Source: §I · p. 2

  • Results

    Quantum-metrology concepts have been shown to achieve sensitivity and precision beyond the shot-noise limit in practical optical sensors.

    Source: §I · p. 2

  • Takeaways & Limitations

    Surface plasmon resonance sensing uses the dependence of the resonance condition on analyte permittivity to detect environmental changes.

    Source: B. Surface plasmon resonance sensing · p. 5

    Progress in squeezed interferometric plasmonic sensors has been limited, with increased interferometric loss identified as one reason.

    Source: §2 · p. 47

Abstract

from arXiv · show

The extraordinary sensitivity of plasmonic sensors is well known in the optics and photonics community. These sensors exploit simultaneously the enhancement and the localization of electromagnetic fields close to the interface between a metal and a dielectric. This enables, for example, the design of integrated biochemical sensors at scales far below the diffraction limit. Despite their practical realization and successful commercialization, the sensitivity and associated precision of plasmonic sensors are starting to reach their fundamental classical limit given by quantum fluctuations of light -- known as the shot-noise limit. To improve the sensing performance of these sensors beyond the classical limit, quantum resources are increasingly being employed. This area of research has become known as `quantum plasmonic sensing' and it has experienced substantial activity in recent years for applications in chemical and biological sensing. This review aims to cover both plasmonic and quantum techniques for sensing, and shows how they have been merged to enhance the performance of plasmonic sensors beyond traditional methods. We discuss the general framework developed for quantum plasmonic sensing in recent years, covering the basic theory behind the advancements made, and describe the important works that made these advancements. We also describe several key works in detail, highlighting their motivation, the working principles behind them, and their future impact. The intention of the review is to set a foundation for a burgeoning field of research that is currently being explored out of intellectual curiosity and for a wide range of practical applications in biochemistry, medicine, and pharmaceutical research.

B. Surface plasmon resonance sensing

This section establishes classical plasmonic sensing through surface-plasmon resonance, where analyte-dependent refractive-index or permittivity changes are detected through resonance shifts and reflected-light intensity or phase. It also covers how sensitivity and measurement noise shape detection limits, alongside localized plasmons as a nanoscale sensing route.

I. INTRODUCTION

  • The review places conventional plasmonic sensing near the shot-noise limit and introduces quantum plasmonic sensing, consolidating its framework, theory, key works, working principles, and future impact.It notes that SPR sensors are commercialized and used to study biomolecular interactions.Source: §I · p. 2

II. PLASMONIC SENSORS

  • Plasmonic sensors combine high sensitivity with sub-diffraction-limited confinement, enabling compact sensing of small analyte quantities using classical light in this section.Their sensing process uses light interacting with materials at the metal–dielectric interface.Source: §II · pp. 2–3

A. Surface plasmon polaritons

  • SPPs are interface-bound light–electron hybrid modes whose confinement makes them sensitive to dielectric properties; direct illumination cannot excite them, so prism coupling and other schemes are used.The Kretschmann prism setup is widely used and has enabled successful, including commercialized, plasmonic sensing applications.Source: A. Surface plasmon polaritons · pp. 3–5

B. Surface plasmon resonance sensing

  • SPR sensing exploits resonance’s dependence on analyte permittivity: a reflectance dip reveals SPP excitation, and refractive-index changes support kinetic or material-parameter estimation.SPR sensors can also use a surface receptor to bind a specific ligand, changing the refractive index above the metal.Source: B. Surface plasmon resonance sensing · pp. 5–6
  • In the Kretschmann configuration, prism illumination of a thin metal film produces an evanescent field that couples to the analyte–metal SPP at a specific resonance angle.The resulting attenuated total internal reflection can be identified by a drop in reflectance.Source: B. Surface plasmon resonance sensing · pp. 5–6, Figure 2 · p. 5

1. Angular interrogation

  • Angular interrogation tracks the reflectance-minimum resonance angle, whose sensitivity increases as wavelength decreases but becomes singular at short wavelengths; gold and BK7 glass can improve sensitivity in the stated conditions.Typical angular sensitivities are reported as 10-103 degrees/RIU; gold outperforms silver in the optical range, and BK7 is favored over SF14 for an analyte with na = 1.32.Source: §1 · p. 7

2. Spectral interrogation

  • Spectral interrogation tracks the resonance wavelength at fixed incidence angle; sensitivity rises with wavelength and |ϵ′m(ω)|, making silver more sensitive than gold, while BK7 is favored over SF14 for na = 1.32.The spectral sensitivity has a singularity in the long-wavelength limit, and typical values are reported in the range 103-105.Source: §2 · p. 7

3. Limit of detection

  • Quantum resources can lower classical plasmonic sensors’ minimum LOD of ∼10^-6–10^-7 RIU by reducing noise below shot noise, without changing setup-dependent sensitivity or specificity.The LOD can be reduced by increasing sensitivity and reducing noise; in quantum sensing, noise reduction at fixed integration time and intensity is the crucial improvement.Source: §3 · p. 7

C. Intensity vs phase sensing

  • Phase sensing can be a few orders of magnitude more sensitive than intensity sensing; one reported Kretschmann sensor achieved an LOD of 10^-8, two orders better than intensity sensing.Intensity modulation fixes the incident angle at the steepest off-resonance point and monitors reflected intensity; phase sensing detects phase changes across the resonance, often interferometrically.Source: C. Intensity vs phase sensing · p. 8

D. Localized surface plasmon resonance sensing

  • Localized surface plasmons in metallic nanoparticles produce near-field enhancement and resonance shifts when the surrounding permittivity changes, enabling refractive-index sensing; larger or nonspherical particles require higher-mode or alternative modeling.For small spheres, the dipolar resonance follows the Fröhlich condition; larger spheres may require Mie theory, ellipsoids Gans theory, and complex structures numerical electromagnetic methods.Source: D. Localized surface plasmon resonance… · pp. 8–9

1. Sensitivity and figure of merit

  • Single-particle LSP sensitivities are on the order of 102-103 nm/RIU, and their FOM balances sensitivity against linewidth; spherical-particle FOM is about an order below SPP sensors.For non-Lorentzian spectra with ill-defined linewidth, an alternative FOM∗ uses relative intensity change per refractive-index change; non-spherical particles or metamaterial unit cells can yield comparable FOM or FOM∗.Source: §1 · pp. 9–10

2. Metamaterials

Plasmonic metamaterials use collective responses from subwavelength metallic unit cells to offer sensing architectures with tunable optical properties. Reported work demonstrates high sensitivity and parallel sensing, while miniaturized platforms still face integration and noise challenges.

2. Metamaterials

  • Metamaterial sensors use periodic subwavelength metallic unit cells whose collective response forms a macroscopic material response, with designs in 2D metasurfaces or 3D arrays [148–150].Compared with ensemble LSP sensors, coupled nanostructures within or between unit cells can influence the bulk response, though the distinction is not always clear.Source: §2 · p. 10
  • 32,000 nm/RIU sensitivity and FOM 330 were reported for a gold-nanorod metamaterial biosensor, exceeding LSP and SPP sensor benchmarks, respectively.The sensitivity was about two orders of magnitude above LSP sensors; the FOM was one order above SPP sensors and two above LSP sensors, with corresponding LOD improvements.Source: §2 · p. 10
  • Different split-ring-resonator reflectance resonances enabled parallel sensing of biological interactions at different length scales, and the strategy was later used for dual sensing.The thin-film approach analyzes transmitted or reflected fields without prism coupling.Source: §2 · pp. 10–11
  • Further sensitivity gains are constrained by analyte adsorption–desorption noise and the fabrication of accurate, small unit-cell features.Source: §2 · p. 11

3. Miniaturization

  • Fiber and waveguide SPP sensors offer compactness, but integrating components such as microfluidic channels is challenging because of precise-alignment and assembly requirements.A polarization-maintaining fiber reached an LOD of 10^-6 RIU; integrated sensors realized so far have sensitivity and LOD at best comparable to traditional SPP sensors.Source: §3 · p. 11
  • Miniaturization efforts include low-cost LEDs, diffractive mirrors, micro-prisms, and fiber-based plasmonic designs.Progress toward integration is expected to benefit from advances in optoelectronics, integrated optics, surface functionalization, microfluidics, and nanofabrication; many proposed sensors have passed proof of principle and are being developed into products.Source: §3 · p. 11
  • Classical sensors are limited by noise, ultimately shot noise, while quantum techniques could potentially improve compact plasmonic sensors if their designs preserve the exploited quantum effects.Source: §3 · p. 11

4. Plasmon-enhanced fluorescence and Raman scattering

Plasmon-enhanced fluorescence and SERS use localized plasmonic fields to strengthen optical signals, although the plasmon does not itself transduce the analyte in these modalities. Their performance depends on coupling, emitter distance, and nanostructure design.

Plasmon-enhanced fluorescence and Raman scattering

  • PEF and SERS can improve signal-to-noise ratio by enhancing fluorescence or Raman signals, and their nanostructures can provide sub-diffraction-scale imaging resolution.In these modalities the plasmon enhances the signal rather than serving as the sensor transduction mechanism.Source: §4 · p. 11

Plasmon-enhanced fluorescence

  • 30-times fluorescence enhancement was reported for a single terrylene molecule coupled to a spherical gold nanoparticle, with κ ∼80, ηcoupled ∼0.3, and η0 ∼0.8.The enhancement combines excitation-rate enhancement with changes to emission dynamics; the figure depicts the single-molecule setup and lifetime comparison.Source: §4 · p. 12, Figure 7 · p. 12
  • At distances below ∼5 nm, higher-order LSP modes dissipate transferred energy through Ohmic losses and quench fluorescence; at approximately 10–30 nm, dipolar-mode coupling can enhance it.The dipolar mode radiates but also has nonradiative Ohmic losses; when the original fluorophore quantum efficiency is only a few percent, coupling to LSPs can increase it.Source: §4 · p. 12, Figure 6 · p. 12
  • Fluorescence enhancement combines plasmon-enhanced excitation with emission changes and, below saturation, is given by κηcoupled/η0.Excitation enhancement requires spectral matching between the fluorophore absorption and LSP resonance, with the molecule positioned within the LSP evanescent decay depth.Source: §4 · pp. 11–12
  • Small nanostructures below 20 nm are usually used to quench fluorescence, while particles above 100 nm can lose quantum efficiency and radiation efficiency as higher-order modes contribute.Smaller structures can provide better spatial resolution; rods, stars, dimers, and arrays can improve PEF, and sharp features or nanogaps can support higher field enhancement [67, 200–204].Source: §4 · pp. 12–13
  • Silver-nanoparticle-assisted PEF detected streptavidin and human IgE at 0.25 ng/mL, while gold-nanorod-array hotspots detected single-strand DNA at 10 pM.These examples remain in the weak-coupling regime; strong coupling can provide additional quantum-plasmonic sensing functionalities.Source: §4 · p. 13

Surface-enhanced Raman scattering

  • SERS enhancement combines stronger local fields with chemical enhancement from charge transfer, and the signal enhancement is commonly estimated to scale with the fourth power of the local electric field.Source: §4 · p. 13

E. Surface plasmon resonance imaging

SPR imaging parallelizes spatially resolved intensity measurements, while quantum sensing theory provides a framework for understanding estimation limits and possible noise reduction. The section distinguishes single- and multiparameter estimation and relates shot-noise scaling to practical sensing constraints.

E. Surface plasmon resonance imaging

  • SPR imaging monitors spatial intensity changes at a resonance inflection point, enabling parallel detection across sensor arrays and spatial mapping of refractive-index variations.Functionalizing array elements for different molecules supports scaling to multiple targets; imaging systems can also introduce more noise than single-pixel detectors, motivating quantum imaging schemes.Source: E. Surface plasmon resonance imaging · pp. 13–14

III. QUANTUM SENSORS

  • Coherent-state probes have Poissonian photon-number fluctuations that limit resolution, and squeezing is introduced as a way to reduce this shot noise.The section uses parameter-estimation theory to establish bounds for classical and quantum sensing.Source: §III · p. 14

1. Single-parameter estimation

  • Parameter estimation comprises state preparation, parameter encoding, measurement, and estimation; estimator bias and variance characterize accuracy and uncertainty, while greater Fisher information favors precision and the Cramér–Rao bound is asymptotically saturable by maximum-likelihood estimation.Quantum Fisher information depends on probe state and encoding and can be maximized with an optimal probe; the lowest bound over input states is called the ultimate quantum limit. When a Poisson-distributed random feature with mean N dominates uncertainty, the lower bound is the SNL and Δxest scales with N−1/2.Source: §1 · pp. 14–16

2. Multiparameter estimation

  • Multiparameter estimation bounds the covariance matrix using the Fisher information matrix and supports estimating linear combinations such as relative or average phase.The classical bound can be saturated asymptotically by maximum likelihood estimation, while a quantum Fisher information matrix gives a further quantum bound.Source: §2 · pp. 16–17

B. Shot-noise limited sensing

  • Intensity sensing estimates transmitted intensity, whereas phase sensing reads the outgoing phase; coherent probes give N^-1/2 uncertainty scaling, while quantum comparisons assume fixed measurement repetitions.Matching quantum scaling by increasing classical repetitions or optical power may be impractical for dynamic systems, sensor materials, or analytes with damage thresholds.Source: §1 · p. 16, B. Shot-noise limited sensing · p. 17

1. Shot-noise limited intensity sensing

The section models intensity sensing as estimating a beam splitter’s transmittance and phase sensing as estimating a relative phase in an interferometer. Coherent probes yield shot-noise-limited precision, motivating quantum probes for intensity sensing.

1. Shot-noise limited intensity sensing

  • For coherent-state intensity sensing, the sample mean is an optimal estimator and intensity measurement reaches the quantum Cramér–Rao bound, with uncertainty scaling as N^-1/2.This is the shot-noise limit for the mean input photon number N.Source: §1 · pp. 17–18

2. Shot-noise limited phase sensing

  • In a Mach–Zehnder interferometer, coherent-state phase uncertainty is shot-noise limited; with equal arm losses, a 50:50 beam splitter is optimal and ∆φCR = (νηN)^-1/2.The quantum Cramér–Rao bound remains shot-noise limited regardless of beam-splitter transmittance or measurement type.Source: §2 · p. 18

C. Sub-shot-noise intensity sensing

  • The section asks whether optimizing the probe state can lower the quantum Cramér–Rao bound below the standard quantum limit for intensity and phase sensing.These quantum techniques are subsequently applied to plasmonic sensors.Source: C. Sub-shot-noise intensity sensing · p. 18

1. Quantum-enhanced intensity sensing

For a fixed mean photon number, reducing probe photon-number variance lowers intensity-estimation uncertainty. A Fock-state probe provides a quantum enhancement over a coherent probe, although its uncertainty still scales as N^-1/2.

1. Quantum-enhanced intensity sensing

  • At fixed mean photon number and given η and T, lower probe photon-number variance reduces transmitted-intensity variance and estimation uncertainty.The Fock state has zero photon-number variance, the minimum among the described probes.Source: §1 · p. 19
  • A Fock-state probe gives an intensity-sensing enhancement by a factor of (1−ηT) over a coherent probe, while retaining N^-1/2 scaling.Its quantum Cramér–Rao bound equals the classical Cramér–Rao bound, and photon-number states attain the ultimate quantum limit in intensity sensing.Source: §1 · p. 19
  • Fock-state and coherent-state probes yield binomial and Poisson transmitted photon statistics, respectively, with equal mean ηTN; the Fock-state strategy minimizes the quantum Cramér–Rao bound.The binomial variance is smaller than the Poisson variance for the same mean.Source: §1 · p. 19

2. Multiparameter or multimode intensity sensing

Multiparameter intensity sensing extends loss estimation to several channels, potentially with multiple modes per channel, using signal probes entangled with lossless ancillas. The review describes an ultimate quantum bound, measurement and probe-state strategies that attain it, and an application to circular dichroism sensing.

2. Multiparameter or multimode intensity sensing

  • The framework connects multiparameter loss estimation to image sensing, thermal-source resolution, and probing intensity changes across frequencies or temporal modes.Measuring image intensity at individual pixels of an amplitude mask is framed as estimating multiple loss parameters.Source: §2 · p. 19
  • The ancilla-assisted strategy sends signal modes through K lossy channels while keeping entangled ancillary modes lossless, and permits multiple probe modes per channel under individual energy constraints.The task is to estimate the vector of channel transmittivities.Source: §2 · pp. 19–20
  • The maximum QFIM for estimating multiple channel transmittivities exhibits a quantum enhancement over the QFIM for a product coherent probe, accounting for unwanted channel losses η_k.A product probe of number-diagonal-signal states is shown to achieve this maximum.Source: §2 · p. 20
  • A joint measurement in the Schmidt bases of the ancilla and signal modes makes the FIM equal to the QFIM for any set of transmittivities, and the multiparameter bounds are simultaneously achievable.The simultaneous attainability follows because the associated SLDs commute.Source: §2 · p. 20
  • Circular dichroism sensing estimates the transmittivity difference between left- and right-circularly polarized light through a chiral medium, with additional losses considered for both modes.Quantum modeling has been used to identify the ultimate uncertainty bound and investigate optimal input states and measurements.Source: §2 · p. 21, Figure 13 · p. 21

3. Quantum noise reduction in intensity measurements

Intensity-difference measurements compare signal and reference modes to suppress common excess noise, with quantum performance assessed through noise reduction and signal-to-noise measures. The section compares correlated and classical probes, showing that their advantage depends on the state and channel conditions.

3. Quantum noise reduction in intensity measurements

  • Reference-assisted transmission or absorption spectroscopy keeps one mode unchanged and uses intensity-difference measurement to remove common excess noise between signal and reference.This measurement technique has also been used in classical plasmonic sensing to reach the shot-noise limit.Source: §3 · p. 21
  • The noise reduction factor is below one for nonclassical light, and photon-number-correlated probes can reach zero, compared with one for product coherent light.The twin Fock and TMSV states are examples of photon-number-correlated probes; TMSV is generated through spontaneous parametric down conversion.Source: §3 · p. 22
  • TMSV twin beams support quantum imaging of weakly absorbing objects through spatial correlations, while their average photon number is typically small because of weak nonlinearities.The cited demonstration compares TMSV differential-intensity imaging with product-coherent direct-intensity imaging.Source: §3 · p. 22, Figure 14 · p. 23
  • TMSD probes can raise individual-mode intensity by increasing pump power, while their noise reduction factor remains below one; a 85Rb-vapor-cell experiment measured about 87% noise reduction, or −9 dB.The bright twin beams were used in quantum sensing, imaging, and quantum plasmonics.Source: §3 · p. 22
  • The twin Fock state always achieves sub-shot-noise measurement, while TMSV and TMSD probes do so when G(Tη_a − η_b)^2 satisfies the stated condition.With balanced effective transmissions, Tη_a = η_b, all considered states reach sub-shot-noise levels for any T; experiments have demonstrated this regime.Source: §3 · p. 23
  • For symmetric probes, the SNR depends on the Mandel Q-parameter and input noise reduction factor, and increases as either quantity decreases.The twin Fock state reaches Q_M = −1 and σ_in = 0.Source: §3 · p. 24
  • The twin Fock state gives quantum-enhanced SNR over a balanced product coherent probe regardless of transmittance and channel efficiencies.Its Mandel Q-parameter and input noise reduction factor attain their minima, −1 and 0, respectively.Source: §3 · p. 24
  • TMSV probes provide SNR enhancement only when (Tη_a − η_b)^2N < 2Tη_aη_b, making them more useful at lower N or when Tη_a is close to η_b.Pair coherent and finite-dimensional photon-number-entangled states are also noted as useful candidates.Source: §3 · p. 24

1. Phase sensing in Mach-Zehnder interferometers

In a Mach-Zehnder interferometer, injecting squeezed vacuum alongside a coherent input can improve phase estimation beyond the shot-noise-limited classical setup. The attainable precision depends on photon number and losses, and several measurement schemes can approach the quantum bound.

1. Phase sensing in Mach-Zehnder interferometers

  • With the optimal phase between displacement and squeezing, the lossless coherent-plus-squeezed-vacuum probe can reach Heisenberg scaling when the two inputs contribute approximately N/2 photons each.Too little or too much squeezing does not achieve Heisenberg scaling, though sub-shot-noise sensing remains possible.Source: §1 · p. 25
  • Squeezed vacuum is the optimal state for mode b when a coherent state enters mode a in a lossless MZI, and the resulting bound is practically attainable with photon counting or parity detection.The approach originated as a modification for gravitational-wave interferometers and contributed to improvements in LIGO and Virgo.Source: §1 · p. 25
  • For a single phase shift in one MZI arm, valid precision analysis requires a reference beam whose resources are counted when determining scaling with N.Optical phase is defined only relative to a reference phase.Source: §1 · p. 25
  • With equal arm losses, the phase-estimation quantum bound worsens as transmission efficiency η decreases, and the lossless Heisenberg scaling deteriorates.The lossless bound is recovered at η = 1.Source: §1 · p. 25
  • A weak-local-oscillator linear-optics setup with photon counting achieves the lossy quantum bound, while homodyne detection approaches it in the large-power regime.TMSV robustness against photon loss also supports sub-shot-noise phase sensing over a wide phase range without pre- or post-selection.Source: §1 · p. 25

2. Phase sensing with NOON states

NOON states exploit large photon-number-difference variance for precise phase estimation, but are highly vulnerable to loss. Optimized definite-photon-number probes and single-mode squeezed vacuum provide alternatives for achieving precision beyond classical benchmarks.

2. Phase sensing with NOON states

  • NOON-state generation is difficult and has been limited to N = 5 in photonic systems, while even moderate photon loss can sharply increase estimation uncertainty.Adding other photon-number components can improve loss robustness while retaining sub-shot-noise behavior.Source: §2 · pp. 25–26
  • For a NOON probe, the quantum Cramér–Rao bound scales as ∆φ = 1/√(νN), exhibiting Heisenberg scaling.The scaling follows from the maximal photon-number-difference variance of the state.Source: §2 · p. 25
  • Optimized definite-photon-number distributions can beat the SQL or SNL at any loss level, and experiments demonstrated optimized probes and unconditional NOON-state enhancement.The optimized-probe demonstration used loss-dependent probe preparation; the unconditional NOON result was obtained in a nearly lossless setting with increased efficiency and visibility.Source: §2 · p. 26, Figure 16 · p. 26
  • For two-photon NOON sensing, super-sensitivity requires average transmission above 0.8, corresponding to total loss below 20% on average.The criterion is based on measured visibility exceeding its threshold.Source: §2 · p. 27

3. Single-mode phase sensing

  • Single-mode phase sensing assumes a fixed reference phase and therefore requires phase-sensitive detection, such as homodyne or general-dyne measurement.Photon-number counting alone cannot extract phase information in this setup.Source: §3 · p. 27
  • A single-mode squeezed vacuum has a QCR bound that scales with N^-1, beating the coherent-state SQL at equal energy.Homodyne detection reaches this bound for a pure squeezed-vacuum probe without loss; realistic thermal contributions require an SLD-eigenstate measurement to attain the bound.Source: §3 · p. 28

4. Multiple-phase sensing

Multiple-phase sensing compares classical benchmarks, individual quantum estimation, and simultaneous estimation, with advantages depending on the parameters and probe states. Entanglement can help for global or distributed parameters, but does not guarantee improvement in every setting.

4. Multiple-phase sensing

  • The total uncertainty across phases is characterized by the covariance matrix, and comparisons distinguish classical, individual quantum, and simultaneous quantum strategies.Whether quantum resources beat the SQL or simultaneous estimation improves on individual estimation depends on the parameters and sensing scenario.Source: §4 · p. 28
  • For simultaneous estimation of d phases, a coherent superposition of N photons across modes gives an O(d) advantage over individual estimation among definite-photon-number probes.The considered probe is optimal within that state class.Source: §4 · p. 28
  • For d = 2 without loss or decoherence, Holland–Burnett and NOON states are optimal among N-particle photonic probes, but entanglement does not always outperform individual estimation.For Gaussian inputs, the simultaneous-estimation improvement is at most a factor of 2; suitable mode-separable inputs can match or exceed simultaneous-estimation precision across broader state classes.Source: §4 · p. 28
  • For non-commuting generators, simultaneous estimation can yield a three-fold improvement over individual strategies, but precision tradeoffs remain among the estimators [387].More precise estimation of one parameter can mean less precise estimation of others.Source: §4 · p. 29
  • Entanglement can enhance estimation of global parameters formed from multiple phases, including distributed sensing, compared with estimating phases individually and combining them classically.The cited work identifies entanglement as more significant for global than for individual phase estimation.Source: §4 · p. 29
  • Squeezed-vacuum schemes for linear combinations of phases and displacements have been experimentally demonstrated, while continuous-variable error correction can reinstate Heisenberg scaling under moderate loss or decoherence.The demonstrations used homodyne detection and covered phase and displacement parameters.Source: Figure 18 · p. 29, §4 · p. 30
  • For Gaussian probes estimating an average of independent phase shifts, partially entangled states are optimal, whereas maximal entanglement can be detrimental.Multiparameter squeezing has been proposed as an operational concept for identifying metrologically useful states and estimation strategies.Source: §4 · p. 30

5. Quantum sensing with SU(1,1) interferometers

SU(1,1) interferometers use nonlinear amplifiers to achieve phase precision approaching Heisenberg scaling, with gain and loss shaping their performance. The section also reviews apparent sub-Heisenberg strategies and explains why accounting for total resources limits their scaling.

5. Quantum sensing with SU(1,1) interferometers

  • SU(1,1) interferometers replace beam splitters with nonlinear amplifiers and can approach 1/N phase-precision scaling with vacuum, coherent, or combined coherent and squeezed-vacuum inputs.Their precision enhancement scales with amplifier gain, but conventional configurations require a near-zero internal phase difference to retain that advantage.Source: §5 · p. 30
  • Increasing the second amplifier’s gain makes an unbalanced SU(1,1) interferometer increasingly insensitive to external loss, reaching insensitivity in the arbitrarily large-gain limit.The second amplifier can compensate for external loss through phase-sensitive amplification.Source: §5 · p. 30, Figure 19 · p. 31
  • Balanced SU(1,1) interferometers are robust to small internal losses, but increasing loss pushes their precision toward the SNL.Internal losses include sensor interactions and losses within the nonlinear amplifiers.Source: §5 · p. 31
  • Truncated nonlinear interferometers can retain conventional SU(1,1) phase precision while simplifying the design and allowing high-power local oscillators to raise measurement SNR without adding sensor power.No truncated nonlinear interferometric plasmonic sensor had been demonstrated in the literature covered by the review.Source: §5 · pp. 31–32

E. Quantum sensing beyond the Cramér-Rao bound

  • Probe states suggesting arbitrary sub-Heisenberg scaling are limited to Heisenberg scaling when prior information and repeated measurements needed for the asymptotic QCR bound are counted as resources.This total-resource accounting resolves the apparent sub-Heisenberg cases discussed in the section.Source: E. Quantum sensing beyond the… · p. 32
  • With highly non-Gaussian likelihoods and small samples, quantum Ziv–Zakai bounds can be more appropriate than QCR bounds, and sub-Heisenberg strategies require substantial prior information.With large prior information, guessing from the prior without measurement can achieve comparable precision to the corresponding ZZ bound.Source: E. Quantum sensing beyond the… · p. 32

IV. QUANTUM-ENHANCED PLASMONIC SENSORS

Quantum plasmonic sensors combine plasmonic transducers with quantum inputs and measurements to pursue high-sensitivity, high-precision sensing below the diffraction limit. The section distinguishes these sensors from approaches that rely on quantum effects inherent to plasmonic materials, and introduces intensity-sensitive sensing as a method that uses quantum light and measurements.

IV. QUANTUM-ENHANCED PLASMONIC SENSORS

  • Combining a plasmonic transducer with quantum input states and measurements brings together sub-diffraction sensitivity and the potential for sub-shot-noise sensing.The review describes quantum plasmonic sensors as seeking combined benefits not obtained by either classical plasmonic or conventional quantum optical sensors alone.Source: §IV · p. 33
  • Figure 20 classifies sensors by probe state, measurement scheme, and transducer, showing classical sensors upgraded to plasmonic, quantum, or quantum plasmonic types.The quantum plasmonic type occupies the classification's top-right position.Source: Figure 20 · p. 33, §IV · p. 33
  • Quantum plasmonic sensing can also exploit inherent effects such as electron tunneling, quantum size effects, or quantum surface response without external quantum resources.Whether these effects enable sub-shot-noise sensing or merely improve sensitivity requires careful analysis.Source: §IV · p. 33

A. Quantum plasmonic intensity sensing

1. Intensity sensing with discrete variable states

The section reviews discrete-variable quantum approaches to intensity-sensitive plasmonic sensing, including heralded-photon experiments and two-mode schemes. Reported results include reduced estimation uncertainty against classical benchmarks, while practical gains depend on system losses and source characteristics.

1. Intensity sensing with discrete variable states

  • The prism experiment heralded each signal photon with its idler and normalized measured transmittance against air at each incident angle.The air reference addressed angle-dependent misalignment; refractive-index estimates were obtained by fitting the measured average transmittance to the prism reflectance model.Source: Figure 21 · p. 34, §1 · p. 35
  • For equal per-mode photon number, twin-mode and coherent-state probes have the same intensity-difference signal, but their measurement noise and SNR can differ.The twin-Fock state gives unconditional enhancement; the two-mode squeezed vacuum's enhancement is conditional, approaching the twin-Fock case when ηa ≈ ηb and T ≈ 1.Source: §1 · p. 34
  • 10–20% lower refractive-index estimation uncertainty than the classical benchmark was demonstrated for BSA solutions using heralded single photons.The measurements used an incidence angle of 67.5° and BSA concentrations from 0% to 2% in 0.25% steps; the quantum resource reduces uncertainty, not the mean concentration–index calibration relation.Source: Figure 22 · p. 35, §1 · p. 36
  • In the prism experiment, quantum enhancement depended on total transmittance, making reduced channel loss and higher detection efficiency relevant to improving it.Source: §1 · p. 36
  • Heralded single photons in a tapered hetero-core fiber SPR sensor reached the theoretically predicted estimation uncertainty below the shot-noise limit.The setup used a gold-coated tapered fiber; the cited work demonstrated sensing of BSA concentration and salinity.Source: §1 · p. 36, Figure 23 · p. 36
  • Heralded photons require high total detection efficiency for genuine enhancement, and fair classical comparisons count all photons incident on the sample.Post-selected photons are not true single-photon states; their antibunched character degrades at higher flux.Source: §1 · p. 36
  • Current deterministic single-photon sources have not simultaneously met the stated requirements of purity, indistinguishability, flux, and fidelity.The review describes low demonstrated photon flux as a barrier to practical use.Source: §1 · pp. 36–37
  • Squeezing was maintained after photons excited long-range surface plasmon polaritons in a gold waveguide.Source: Figure 24 · p. 37

2. Intensity sensing with continuous variable states

Two-mode squeezing generated by four-wave mixing can survive plasmonic transmission and support sensing below the shot-noise limit. The review describes intensity-sensing demonstrations in extraordinary optical transmission, Kretschmann sensors, and ultrasonic sensing, while identifying multi-pixel readout as a remaining need for quantum-enhanced SPR imaging.

2. Intensity sensing with continuous variable states

  • Four-wave mixing generates a two-mode squeezed state whose intensity-difference noise is below one shot-noise unit for η > 0 and G > 1.With symmetric moderate losses, the predicted SNR increase over shot-noise-limited detection is 2η(G−1)/(2G−1).Source: §2 · pp. 37–38
  • Localized surface plasmon-mediated extraordinary optical transmission transmitted spatial information in probe images while conserving quantum noise reduction.A beam-splitter loss model matched the measured squeezing as a function of transmission.Source: §2 · p. 39, Figure 27 · p. 39, Figure 28 · p. 40
  • 4.6 dB of noise reduction below the shot-noise limit was measured on the left shoulder of the Kretschmann sensor resonance, with squeezing at virtually all data points.The reported minimum was 0.3 dB for n=1.300 and n=1.301, while the n=1.305 dataset reached the shot-noise limit.Source: §2 · p. 39, Figure 30 · p. 41
  • Twin-beam sensing of ultrasonic pressure waves used transmission changes from localized surface plasmons to transduce refractive-index shifts into an intensity-modulation signal below the shot-noise limit.One twin beam passed through the sensor and the other served as the reference.Source: §2 · p. 40, Figure 31 · p. 41
  • 5.5 × 10^-9 RIU versus 9.6 × 10^-9 RIU was the minimum refractive-index modulation resolution for twin beams and coherent states, respectively, a 56% precision enhancement in ultrasonic sensing.The twin-beam noise floor was 4 dB lower, and the sensor discerned shifts of O(10^-9) RIU.Source: §2 · p. 40, Figure 32 · p. 41
  • Quantum-enhanced SPR imaging has not been reported, and multi-pixel readout of plasmonic sensors is identified as a key ingredient for using spatial quantum correlations.Localized surface plasmon-mediated transmission preserves spatial information, unlike the cited SPP-mediated process; multi-spatial-mode sources and readout schemes are among the proposed building blocks.Source: §2 · p. 42

3. Intensity sensing robust to thermal noise

Twin-beam quantum spectroscopy is reported to distinguish plasmonic resonance spectra more clearly than classical spectroscopy under added thermal noise. The section also introduces nanowires as a compact setting for quantum phase sensing through refractive-index-dependent phase shifts.

3. Intensity sensing robust to thermal noise

  • The comparison measured glycerin-water concentrations around a gold nanoparticle array using TMSV coincidence spectroscopy or conventional lamp-based spectroscopy, with added thermal counts controlled in the monochromator.The tested noise levels were Ns = 10^2, 2 × 10^4, and 7 × 10^4 c/s.Source: §3 · pp. 42–43
  • Quantum spectroscopy produced more distinguishable resonance curves than classical spectroscopy under increasing thermal noise, including when noise photocounts were 70 times the signal photocounts.The authors attribute this robustness to the TMSV state's strong photon-number correlation, absent from the classical scheme.Source: §3 · p. 43, Figure 34 · p. 43

B. Quantum plasmonic phase sensing

  • A metal nanowire offers a compact, sub-diffraction-limit setting for quantum phase sensing, where refractive-index changes alter the optical phase rather than transmission amplitude.The passage also notes potential modification into more advanced nanophotonic circuitry.Source: B. Quantum plasmonic phase sensing · p. 44

1. Phase sensing with discrete variable states

The review examines discrete-variable quantum phase sensing with plasmonic nanowires, from theoretical precision gains to a two-photon NOON-state experiment. It also notes that loss constrains practical performance and that squeezed interferometric plasmonic sensors remain underdeveloped.

1. Phase sensing with discrete variable states

  • In the N=4 model, quantum phase sensing with a plasmonic nanowire improved precision over the classical case and the dielectric-nanowire comparison.The theoretical framework separates quantum measurement-and-state effects from the classical sensitivity set by nanowire length and propagation constant.Source: §1 · p. 45
  • Optimized probe states were reported to provide phase-sensing precision beyond the shot-noise interferometric limit even with moderate nanowire loss, for arbitrary photon number N.The coefficients cn are optimized to maximize the quantum Fisher information and minimize the quantum Cramér–Rao bound.Source: §1 · p. 45
  • The two-photon NOON-state experiment showed phase oscillations over half the phase interval of the classical signal, but its measured visibility of approximately 0.880 did not overcome the loss-adjusted precision threshold.The experiment applied the phase shift outside the nanowire; the authors state that the threshold could not be surpassed without reducing setup loss, even at unit visibility.Source: §1 · p. 46

2. Phase sensing with continuous variable states

  • Squeezed interferometric plasmonic sensors remain little developed, in part because plasmon absorption leaves little squeezing in phase-sensitive operation.The review identifies loss-resilient designs and new SU(2) and SU(1,1) approaches as opportunities for further work.Source: §2 · p. 47

C. Quantum plasmonic sensing based on emitter-plasmon coupling

This section surveys emitter–plasmon coupling routes to quantum plasmonic sensing, including spectral readout and emitter-based detection. It also establishes the shot-noise limit for classical coherent states and the need for a full quantum description of plasmonic systems.

C. Quantum plasmonic sensing based on emitter-plasmon coupling

  • Strong coupling between emitters and localized surface plasmons is sensitive to analyte structural and material parameters, motivating multiple sensor designs based on a full quantum treatment.The plasmonic structures are modeled as lossy, small-mode-volume resonators; the work builds on plasmon-enhanced field effects described for PEF and SERS sensors.Source: C. Quantum plasmonic sensing based on… · p. 47
  • Emitter–plasmon emission readout has been proposed for refractive-index sensing, single-molecule recognition, molecular localization and tracking, and TNT detection.The review notes that further analysis is needed to determine whether these approaches yield sub-shot-noise sensing, improved sensitivity, or both.Source: C. Quantum plasmonic sensing based on… · p. 47
  • Nearly 15-fold sensitivity enhancement over conventional shifting-type label-free plasmonic sensors was reported for splitting-type sensing with quantum-emitter labels in a nanoplasmonic dimer.Changing analyte–emitter complex surface density shifts the extinction spectrum and produces multiple resonance peaks through strong emitter–LSP coupling; the proposal targets single-analyte detection.Source: C. Quantum plasmonic sensing based on… · p. 47, Figure 37 · p. 48
  • A three-level emitter coupled to a plasmonic cavity can form an ‘embedded superstate’ with unboundedly narrow emission lines, reducing scattering-spectrum noise and enabling significantly improved sensitivity.The effect relies on quantum interference among multiple atomic transition pathways together with strong coupling.Source: C. Quantum plasmonic sensing based on… · p. 48
  • The nanofiber heterodyne experiment demonstrated quantum-noise-limited evanescent single-molecule biosensing with four orders of magnitude less optical intensity for state-of-the-art sensitivity.A related proposal found that maximally entangled nitrogen-vacancy centers optimize β-factor estimation at initial times, while a product state is optimal at long times.Source: C. Quantum plasmonic sensing based on… · p. 48, Figure 37 · p. 48
  • A plasmonic groove waveguide can collect fluorescence from nitrogen-vacancy centers for room-temperature optically detected magnetic resonance, while a metasurface proposal uses infrared absorption readout for enhanced magnetic-field sensitivity.The infrared approach addresses low photon-collection efficiency and modulation contrast in standard fluorescence-based ODMR.Source: C. Quantum plasmonic sensing based on… · pp. 48–49

GLOSSARY

  • Shot noise is statistical noise from random fluctuations in a resource or measured electric signals, not fundamental noise.In coherent light, photons follow a Poisson distribution in photon number, producing shot noise when intensity is measured.Source: GLOSSARY · p. 52
  • The shot-noise limit is the best possible precision obtained using a classical coherent state.Source: GLOSSARY · p. 52
Loading 2012.01525v2…