Quantum illumination with nonzero-mean signal–idler states via noise-enhanced heterodyne work extraction

Mustafa Gündoğan, Mehmet Emre Tasgin · 2025

Read Quantum illumination with nonzero-mean signal–idler states via noise-enhanced heterodyne work extraction by Mustafa Gündoğan, Mehmet Emre Tasgin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-08248.txt.

Introduction

Mustafa Gündoğan and Mehmet Emre Tasgin propose a new receiver for quantum illumination, a method for spotting a faint target in noise, published in 2025. At room heat, in microwave and low terahertz links, strong thermal noise makes faint linked signals hard to recover after loss. Their scheme uses work extraction to turn a hard signal into an easy one.

The returned mode is measured, and that result is fed forward to a stored partner mode. The score from the stored mode then shows whether a target was present. The authors show the score matches the leading result of an ideal prior receiver, but with a linear, direct measure. Heat noise in the source can even help when it is linked before sending.

Catalog record

CREATORMustafa Gündoğan, Mehmet Emre Tasgin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-08248

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Quantum illumination with nonzero-mean signal–idler states via noise-enhanced heterodyne work extraction Mustafa Gündoğan, Mehmet Emre Tasgin Abstract Room-temperature microwave and low-THz links exhibit large thermal occupations, making phase-sensitive signal–idler correlations difficult to recover after loss. We introduce a work-extraction-based quantum-illumination receiver in which the returned mode a ^ R \hat{a}_{R} is measured via heterodyne detection and the outcome is fed forward to a locally stored, possibly displaced idler. For a noisy two-mode–squeezed resource, the receiver is characterized by the heterodyne correlation parameter x h = η ​ c 2 / [ a ⁡ ( b + ν h ) ] x_{\rm h}=\eta c^{2}/[a(b+\nu_{\rm h})] . The calibrated displaced-idler work score has Chernoff exponent ξ h = x h / 4 + O ⁡ ( x h 2 ) \xi_{\rm h}=x_{\rm h}/4+O(x_{\rm h}^{2}) , which becomes linear in the target transmissivity η \eta in the weak-return, background-dominated regime, matching the leading-order performance of an ideal OPA receiver, but achieved here via a linear and directly measurable correlation mechanism. Unlike OPA-based schemes, the present protocol does not require zero first moments and does not rely on weak-probability nonlinear detection. In our scheme, extracted work converts hard-to-measure second-order moment correlation information into an accessible first-moment signal . Moreover, preparation noise n ¯ p \bar{n}_{p} , naturally present at room temperature in the microwave and THz regimes, can be directly harnessed when correlated prior to transmission, whereas a classical coherent signal cannot utilize such incoherent thermal photons without first converting them into usable signal energy. Introduction .—Quantum illumination (QI) uses correlations between a transmitted signal and a retained idler to discriminate a weakly reflecting target from a bright noisy background. Its notable feature is that useful correlation information can remain even when the return-idler state is separable after propagation through an entanglement-breaking environment [ 26 , 39 , 37 , 46 , 45 , 27 , 6 , 22 , 41 , 29 , 12 , 18 , 20 , 32 , 43 ] . In the microwave domain the problem is especially severe: at a few GHz and room temperature, the Planck occupation of the background can be 10^{3} – 10^{4} photons per mode, so the returned marginal is dominated by thermal noise [ 6 , 7 , 10 , 14 , 13 ] . The central experimental challenge is therefore to read out the small residual correlations without relying on highly specialized phase-conjugating or nonlinear joint receivers [ 19 , 49 , 50 , 48 , 33 , 22 ] based on weak-probability nonlinear processes. We develop a work-extraction readout for this setting. The receiver heterodynes the returned mode \hat{a}_{R} and feeds the result forward to a local operation on the idler. The measurement information reduces the conditional uncertainty of the idler and can be converted into extractable work in the standard Gaussian measurement-and-feedback benchmark [ 8 , 28 , 35 , 30 ] . The scheme is written only in terms of first and second moments, so it is naturally compatible with calibrated coherent offsets and noisy Gaussian preparation. This point is important because conventional Gaussian QI is usually formulated with zero first moments, \langle\hat{a}_{I}\rangle=\langle\hat{a}_{R}\rangle=0 , so that the transmitter advantage is not confused with ordinary coherent illumination [ 39 , 37 ] . Here the receiver may instead use a known idler displacement as a local energetic reference. A displacement applied only to, or already present in, the retained idler does not interrogate the target and does not alter the covariance matrix, the Gaussian entanglement criterion, or the normalized return-idler covariance. Thus nonzero displaced two-mode entangled states are allowed in the protocol. The displacement is a removable baseline and, in the calibrated work channel, supplies a practical phase reference for a signed linear-response observable , see SM Sec. 5. In contrast to optical parametric amplifier (OPA) receivers, which use a weak nonlinear interaction to access phase-sensitive correlations such as \langle\hat{a}_{R}\hat{a}_{I}\rangle and are normally analyzed for zero-mean fields, the present scheme operates at the level of heterodyne data and a local idler work readout. The protocol converts the residual correlation into a deterministic, displacement-tolerant work score. This makes it a direct alternative for high-noise microwave and low-THz settings, where implementing a low-noise nonlinear joint receiver is difficult. More importantly, work extraction provides a mechanism for mapping second order signal–idler correlations onto a signed, phase-sensitive “first-order” measurement outcome. The same framework also provides a simple route to exploit trusted preparation noise. If thermal noise is present before the two-mode squeezing operation, it is amplified together with the signal–idler covariance, whereas channel noise is added only to the returned marginal. At fixed squeezing, this asymmetry enhances the normalized heterodyne correlation in a background-dominated regime. The advantage does not arise from a per-photon energy gain, but from operating in a hardware-limited setting where room-temperature thermal excitations are already present and can be directly harnessed by correlating them prior to transmission. Preparation noise \bar{n}_{p} is naturally available at room temperature in microwave and THz devices and can be used directly in this noise-enhanced scheme. A classical coherent signal cannot directly exploit those uncontrolled thermal photons as a phase-stable signal. It must first convert them into usable signal energy, which costs work. Hybrid optical–microwave generation schemes and optomechanical transduction platforms provide possible routes to preparing and probing such noisy correlated states in room-temperature or hybrid architectures [ 23 , 21 , 24 , 34 , 1 , 4 , 47 ] . Gaussian model .—We use quadratures \hat{x}=(\hat{a}+\hat{a}^{\dagger})/\sqrt{2},\qquad\hat{p}=(\hat{a}-\hat{a}^{\dagger})/(i\sqrt{2}), (1) so that the vacuum variance is 1/2 and [\hat{x},\hat{p}]=i . The transmitter prepares two modes by applying two-mode squeezing of strength r to modes with identical trusted preparation noise \nu_{p}=\bar{n}_{p}+\frac{1}{2}. (2) After the squeezing operation, the idler-signal covariance matrix is V_{IS}=\begin{pmatrix}a\,\mathbb{I}_{2}&c\,\mathbb{Z}\\ c\,\mathbb{Z}&a\,\mathbb{I}_{2}\end{pmatrix},\qquad\mathbb{Z}=\mathrm{diag}(1,-1), (3) with a=\nu_{p}\cosh(2r),\qquad c=\nu_{p}\sinh(2r). (4) The signal propagates through a thermal-loss channel, \hat{a}_{R}=\sqrt{\eta}\,\hat{a}_{S}+\sqrt{1-\eta}\,\hat{a}_{\rm ch}, (5) where \eta is the transmissivity and \nu_{\rm ch}=\bar{n}_{\rm ch}+1/2 is the channel quadrature variance [ 40 ] . The post-channel idler-return covariance matrix is V_{IR}=\begin{pmatrix}a\,\mathbb{I}_{2}&\sqrt{\eta}\,c\,\mathbb{Z}\\ \sqrt{\eta}\,c\,\mathbb{Z}&b\,\mathbb{I}_{2}\end{pmatrix},\qquad b=\eta a+(1-\eta)\nu_{\rm ch}. (6) Possible coherent displacements are collected in the first-moment vector \bm{d}_{IR}=(d_{I,x},d_{I,p},d_{R,x},d_{R,p})^{T}. (7) Local displacements change \bm{d}_{IR} but not V_{IR} , so they leave all covariance-based correlation and entanglement formulas unchanged [ 40 , 36 ] . Heterodyne conditioning .—For a Gaussian measurement on the returned mode, the conditional idler covariance is the Schur complement [ 17 , 15 , 40 , 2 , 44 , 25 ] \displaystyle V_{I|\pi_{R}} \displaystyle=V_{I}-C(V_{R}+\Gamma_{\pi_{R}})^{-1}C^{T}, (8) \displaystyle\bm{d}_{I|\bm{y}_{R},\pi_{R}} \displaystyle=\bm{d}_{I}+C(V_{R}+\Gamma_{\pi_{R}})^{-1}(\bm{y}_{R}^{\rm(h)}-\bm{d}_{R}). (9) Here \bm{y}_{R}^{({\rm h})}=(X_{R}^{({\rm h})},P_{R}^{({\rm h})})^{T} is the heterodyne outcome for the returned mode. In the present quadrature convention, ideal heterodyne detection has Gaussian seed covariance \Gamma_{\rm h}=\nu_{\rm h}\mathbb{I}_{2},\qquad\nu_{\rm h}=\frac{1}{2}. (10) Using Eq. ( 6 ) in Eq. ( 8 ) gives V_{I|{\rm h}}=\left(a-\frac{\eta c^{2}}{b+\nu_{\rm h}}\right)\mathbb{I}_{2}. (11) Work extraction .—In the reversible Gaussian measurement-and-feedback benchmark, the dimensionless work w=W/k_{B}T is the positive available work associated with a decrease of the idler nonequilibrium free energy. We use the dimensionless free-energy functional {\color[rgb]{0,0,0}f_{I}(\rho)\equiv\frac{F_{I}(\rho)}{k_{B}T}=\frac{\langle\hat{H}_{I}\rangle_{\rho}}{k_{B}T}-S(\rho).} (12) The additional work extractable due to the information gained from the measurement, corresponding to the entropy reduction term T\,S , contributes as {w}_{\rm cov}=\frac{1}{2}\ln\!\left(\frac{\det V_{I}}{\det V_{I|{\rm h}}}\right)=-\ln(1-x_{\rm h}), (13) where x_{\rm h}\equiv\frac{\eta c^{2}}{a(b+\nu_{\rm h})},\qquad\rho_{\rm h}\equiv\sqrt{x_{\rm h}}=\frac{\sqrt{\eta}\,c}{\sqrt{a(b+\nu_{\rm h})}}. (14) The number \rho_{\rm h} is the Pearson coefficient

Read the complete text