Quantum limits on squeezing
Xin Zhou, Francesco Massel · 2026
Read Quantum limits on squeezing by Xin Zhou, Francesco Massel (2026) 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-2604-22500.txt.
Introduction
Zhou and Massel ask how far squeezing can go in linked quantum oscillators. Basic commutation rules plus stability alone set a floor on total steady squeezing. That floor is reached at strong coupling for schemes using dissipation, not direct driving. Adding a separate drive per mode shifts the noise-gain balance, giving a best bound near one half. The same rules recast a three-mode inseparability test as one parameter-based measure, close to reach in present mechanical tests at room temperature.
Catalog record
| CREATOR | Xin Zhou, Francesco Massel |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22411 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.22500
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quantum limits on squeezing Xin Zhou, Francesco Massel Abstract In our work, we show how, for a network of bosonic modes, canonical commutation relations constrain the coefficients relating input and internal modes. Based on these constraints, we derive a lower bound on the total steady‑state squeezing achievable in reservoir‑engineered (dissipative) squeezing schemes, quantified by the sum of mode‑optimal quadrature variances normalized to its corresponding input variance. The bound follows solely from canonical commutation relations and stability, and is saturated in the strong‑coupling limit at 1. Furthermore, we show that adding independent parametric driving terms for each mode changes the quantum noise–gain balance and yields a distinct optimum bound, approaching 1/2. In addition, we show how these constraints allow us to reformulate the Duan inseparability criterion for a three-mode bosonic system in terms of a single parameter-dependent figure of merit. Our results apply directly to current electromechanical and nanomechanical experiments and indicate that the two-mode bounds can be experimentally approached even at room temperature. Constraints imposed by the laws of quantum mechanics can shape the macroscopic behavior of physical systems, from the Pauli exclusion principle underlying the chemical properties of matter, to the fundamental limits on measurement and control 23 ; 36 . In linear bosonic devices, noncommutativity enforces unavoidable tradeoffs, exemplified by the quantum limit on added noise in amplification 11 ; 3 ; 2 and by the 3 dB limit for stable intracavity parametric squeezing 6 ; 7 ; 5 ; 28 . Here we show that, for a stable physically realizable Markovian network of N bosonic modes, canonical commutation relations (CCRs) constrain the transfer coefficients between input and internal modes through simple commutator-budget sum rules. For passive (number-conserving) linear networks, including reservoir-engineered squeezing schemes that become passive in an appropriate Bogoliubov frame, these commutator budgets also act as noise budgets, directly constraining steady-state quadrature variances and thus the squeezing achievable by engineered dissipation. While our framework applies to generic linear bosonic networks, we focus on examples motivated by cavity optomechanics and electromechanics 1 , where controllable radiation-pressure interactions enable reservoir engineering of dissipation and interactions 26 ; 17 ; 18 ; 25 . On the theory side, engineered dissipation provides a versatile route to prepare nonclassical steady states, including ground-state cooling 10 , steady-state mechanical squeezing 13 , and stationary optomechanical and mechanical–mechanical entanglement 15 ; 33 ; 29 ; 35 . Experimentally, reservoir-engineering techniques have enabled ground-state cooling of nanomechanical and micromechanical resonators 30 ; 4 , observation of steady-state mechanical squeezing 34 ; 14 ; 24 , and stabilized entanglement involving mechanical degrees of freedom 22 ; 20 . These constraints yield a lower bound on the total steady-state squeezing achievable by two-mode reservoir engineering 13 ; 35 : the sum of the two mode-optimal quadrature variances (each normalized to its corresponding input variance) cannot be reduced below 1 , and approaches this limit in the strong-coupling regime. We further show that adding local degenerate parametric driving modifies the noise–gain balance and leads to an optimized bound approaching 1/2 within the stability region. Finally, we apply the same commutator-budget framework to the three-mode reservoir-engineering scheme of Refs. 35 ; 20 and recast the Duan inseparability criterion 8 in terms of a single commutator-transfer parameter, enabling a direct assessment of entanglement performance from system parameters. Our starting point is the description of an N -mode bosonic system coupled to an external environment, which we will describe with standard input–output theory 9 . We suppose that each mode is independently coupled to its bath, which determines noise and dissipation properties of the corresponding bath, and consider a time evolution governed by a linear time-invariant (LTI) system of equations which, in the Bogoliubov (doubled) space, can be written as \dot{\xi}=\bar{A}\xi+D\xi_{\mathrm{in}} 12 ; 19 , where \xi=\left[a_{1},a_{2}\dots a_{N};a^{{\dagger}}_{1},a^{{\dagger}}_{2}\dots a^{{\dagger}}_{N}\right] ( \xi_{\mathrm{in}}=\left[a_{1,\mathrm{in}},a_{2,\mathrm{in}}\dots a_{N,\mathrm{in}};a^{{\dagger}}_{1,\mathrm{in}},a^{{\dagger}}_{2,\mathrm{in}}\dots a^{{\dagger}}_{N,\mathrm{in}}\right] ), \bar{A} and D are the drift and input matrices, respectively. We assume all dissipation is accounted for by these explicit Markov bath couplings (i.e., no additional unobserved/internal loss channels), so that input channels are in one-to-one correspondence with damped modes. Under these assumptions, the relations derived below follow solely from stability ( \operatorname{Re}\lambda(\bar{A})<0 ) and the preservation of CCRs, which, following 12 , we term physical realizability (PR) 27 . For a PR system, it is possible to define a 2\mathrm{N}\times 2\mathrm{N} matrix W_{i}=\int_{0}^{\infty}e^{\bar{A}t}(D_{i}\Sigma D_{i}^{\dagger})e^{\bar{A}^{\dagger}t}dt (1) where \Sigma=[I_{N},\,0;0,\,\text{-}I_{N}] , and D_{i} denotes the columns of D associated with input channel i . {(W_{i})}_{jk} quantifies the contribution of input channel i to the commutator budget of the commutator \left[\xi_{j},\xi^{{\dagger}}_{k}\right] for modes j and k in Bogoliubov space. From Eq. ( 1 ), it is possible to show that, for stable \bar{A} , W_{i} is the unique solution to the Lyapunov equation \bar{A}W_{i}+W_{i}\bar{A}^{\dagger}+D_{i}\Sigma D_{i}^{\dagger}=0 . We can restrict W_{i} to the annihilation sector defining the ( \mathrm{N}\times\mathrm{N} ) commutator channel associated with input j as the projection K_{i}=P_{a}W_{i}P_{a}^{\dagger} , where P_{a}=[I_{N},0] (size \mathrm{N}\times 2\mathrm{N} ). PR for internal modes \left[a_{j},a_{k}^{\dagger}\right]=\delta_{jk} immediately implies that \sum_{i}\penalty\ {(K_{i})}_{jk}=\delta_{jk} . In addition, for passive systems, it is possible to show that K_{i} is a positive matrix. If we further consider the usual situation in which dissipation is diagonal (i.e. the annihilation sector A+A^{\dagger}=-\mathrm{diag}(\gamma_{1},\dots,\gamma_{N}) ), and internal dissipation is absent, we have that \sum_{j}\gamma_{j}(K_{i})_{jj}=\gamma_{i} . Upon further restricting to real couplings (i.e. to reciprocal systems 18 ) we have \gamma_{j}(K_{i})_{jj}=\gamma_{i}(K_{j})_{ii} which can be considered as a manifestation of Onsager reciprocity 21 ; 27 . For the N=2 case, any (generally complex) Hermitian matrix can be transformed to a real one, making the reciprocity condition redundant. This can be confirmed by showing how, for N=2 , reciprocity can be directly deduced from \sum_{i}(K_{i})_{jk}=\delta_{jk} and \sum_{j}\gamma_{j}(K_{i})_{jj}=\gamma_{i} . If the LTI system is solved explicitly we can write the lowering operators associated with the internal modes as \displaystyle a_{i}=\sum_{j}\int\frac{d\omega}{2\pi}M_{i,j}(\omega)a_{\mathrm{in},j}+L_{i,j}(\omega)a^{\dagger}_{\mathrm{in},j}, (2) from which we can deduce (K_{j})_{ik}=\int\frac{d\omega}{2\pi}\left[{M^{*}}_{i,j}(\omega){M}_{k,j}(\omega)-{L^{*}}_{i,j}(\omega){L}_{k,j}(\omega)\right] . For passive systems ( L_{i,j}=0 ), it immediately follows that, for uncorrelated inputs, the steady-state equal-time variance \Delta X_{i}^{2}=\braket{X_{i}^{2}}-\braket{X_{i}}^{2} , where X_{i}(\theta)=\left(a^{\dagger}e^{i\theta}+ae^{-i\theta}\right)/\sqrt{2} ( X_{i}\equiv X_{i}(\theta) , P_{i}\equiv X_{i}(\pi/2) ), can be expressed in terms of the inputs as \Delta X_{i}^{2}=\sum_{j}(K_{j})_{ii}\Delta X_{j,\mathrm{in}}^{2} . The passivity of the system therefore allows us to establish a crucial link between CCR constraints and power/noise gain constraints in a linear bosonic system. While the full matrices K_{i} encode diagonal and off-diagonal commutators [a_{j},a_{k}^{\dagger}] and are required when treating inter-mode correlations. In the following we focus on the their diagonal terms, which we denote I_{ij}=(K_{j})_{ii} . I_{ij} determines the contribution of input j to the equal-time commutator (and, for passive systems with uncorrelated inputs, to the local quadrature variance of mode i ). Dissipative squeezing The first case we consider is a coupled two-mode system in the presence of diagonal dissipation. We will show how the relations established above allow us to determine nontrivial limits on the quadrature variances in a dissipative squeezing scenario. The Hamiltonian we are considering describes a standard reservoir engineering setup in which two bosonic modes are coupled through the following (interaction) Hamiltonian \displaystyle H=G_{+}a_{1}^{\dagger}a^{\dagger}_{2}+G_{-}a_{1}^{\dagger}a_{2}+\mathrm{h.c.}. (3) Eq. ( 3 ) represents, for instance, the prototypical optomechanical setup leading to squeezing of mechanical motion 34 ; 24 ; 14 . The Hamiltonian ( 3 ) can be transformed as H=\mathcal{G}a_{1}^{\dagger}\alpha+\mathrm{h.c.} where the