---
title: Yurke-Type SU(1,1) Interferometer
url: https://www.emergentmind.com/topics/yurke-type-su-1-1-interferometer
type: topic
---

# Yurke-Type SU(1,1) Interferometer

A Yurke-type SU(1,1) interferometer is a nonlinear, quantum-enhanced interferometric configuration in which passive beam splitters are replaced by two-mode parametric amplifiers tunable via the SU(1,1) algebra. This design provides phase sensitivity that can surpass the shot-noise limit (SQL) and approach Heisenberg scaling, by leveraging quantum correlations between modes rather than classical intensity division. The canonical implementation consists of two cascaded parametric amplifiers (PAs) with a controllable phase shift inserted in one mode between them. Multimode implementations, dispersion engineering, seeding strategies, detection protocols, and photon subtraction at the output further optimize sensitivity and metrological performance.

## 1. SU(1,1) Algebraic Structure and Operational Principle

The essential operators of a Yurke-type SU(1,1) interferometer are the SU(1,1) generators: $K_+ = a_s^\dagger a_i^\dagger$, $K_- = a_s a_i$, and $K_0 = \frac{1}{2}(a_s^\dagger a_s + a_i^\dagger a_i + 1)$. They satisfy $[K_0, K_+] = +K_+$, $[K_0, K_-] = -K_-$, $[K_-, K_+] = 2K_0$. The core building block is the unitary two-mode squeezer $U_P(g, \theta) = \exp[g e^{i\theta} K_+ - g e^{-i\theta} K_-]$, where $g$ is the gain and $\theta$ the pump phase.

A Yurke-type SU(1,1) interferometer comprises two PAs (with gain parameters $g_1$, $g_2$), separated by a phase shifter $\phi$ applied to one of the modes (typically the idler). The total unitary sequence is $U_\mathrm{tot}=U_P(g_2, 0) \cdot e^{i\phi a_i^\dagger a_i} \cdot U_P(g_1, 0)$. In the plane-wave, single-frequency limit, this implements a two-mode squeezing and rotation that can surpass conventional shot-noise scaling [2012.03751].

## 2. Input–Output Relations and Phase Sensitivity

The Heisenberg-picture input–output relations for signal ($a_s$) and idler ($a_i$) through the interferometer, for equal gain in both PAs, are:
\[
\begin{aligned}
a_s^{(\text{out})} &= a_s^{(0)} \cosh g_1 \cosh g_2 + a_i^{(0)\dagger} \sinh g_1 \cosh g_2 + e^{-i\phi} [ a_i^{(0)\dagger} \cosh g_1 \sinh g_2 + a_s^{(0)} \sinh g_1 \sinh g_2 ], \\
a_i^{(\text{out})} &= \text{... (similarly) }
\end{aligned}
\]
The measured observable is typically the signal photon number $N_s = a_s^{(\text{out})\dagger} a_s^{(\text{out})}$, whose mean and variance as functions of $\phi$ are:
\[
\langle N_s(\phi) \rangle = 2 \sinh^2 g [ \cosh^2 g - \sinh^2 g \cos \phi ], \quad \text{Var}(N_s) = \sinh^2 g\, \cosh^2 g [1 + \cos \phi ].
\]
Phase sensitivity is quantified as $\Delta\phi = \Delta N_s / |d\langle N_s\rangle/d\phi|$. For large $g$, $\Delta\phi \simeq 1 / N$ (Heisenberg scaling), where $N$ is the total probe photon number [2012.03751].

## 3. Single-Mode vs. Highly Multimode Interferometry

Realistic sources generate highly multimode biphoton states with joint spectral amplitude (JSA) $F(\omega_s, \omega_i) = \sum_k \sqrt{\lambda_k} u_k(\omega_s) v_k(\omega_i)$. Schmidt mode operators $A_k, B_k$ diagonalize the mode structure, with each $(A_k, B_k)$ pair squeezed independently by gain $\gamma_k = G\sqrt{\lambda_k}$. For $K$ Schmidt modes,
\[
\langle N_s \rangle = \sum_k \sinh^2 \gamma_k, \quad \mathrm{Var}(N_s) = \tfrac{1}{4} \sum_k \sinh^2(2\gamma_k).
\]
Single-mode operation ($\lambda_1 = 1$) achieves ideal Heisenberg scaling, while strong multimode occupation degrades phase sensitivity due to leakage “spilling out” from the dark fringe, especially at higher gains [2012.03751, 1909.05026].

## 4. Integrated, Spectrally Multimode Designs and Dispersion Engineering

Recent architectures have focused on integrated waveguide platforms (e.g., KTP) with polarization converters and CW pump to realize compact, multimode Yurke-type SU(1,1) interferometers [2012.03751, 2202.05025]. Key features include:

- Use of a mid-line polarization converter to compensate group-velocity mismatch, swapping o/e polarizations after the first PDC stage.
- The output JSA takes the form $F(\omega_s, \omega_i) \propto \alpha(\omega_s+\omega_i) \sinc[\Delta\beta L/2] \cos(\phi/2) e^{i...}$.
- Near-perfect destructive interference occurs at the “dark fringe” when $\phi_{dark} = \pi$ ($\cos(\phi/2)=0$).

Spectral filtering around the pump central frequency and dispersion suppression via careful polarization converter placement maximize phase sensitivity and visibility, enabling sub-SQL operation at photon numbers up to $10^4$ and visibilities $\gtrsim98\%$ [2012.03751, 2202.05025].

## 5. Detection Protocols and Seeding Strategies

Detection modalities include direct photon-counting and homodyne detection:

- Photon-counting: Optimal for moderate gain and vacuum seeding; phase estimation uses the classical error-propagation bound.
- Homodyne detection: Local oscillator matched to dominant Schmidt mode yields quadrature measurement $H = |\beta_\mathrm{lo}| [ e^{i\theta} A_1^{(\mathrm{out})} + e^{-i\theta} A_1^{(\mathrm{out}) \dagger} ]$. This can outperform counting at low gains [2012.03751, 1610.06891].
- Seeding: Vacuum input already beats SQL. Bright seeding or single-mode seeding disrupts signal–idler balance or multimode correlations. Distribution of weak seeds across Schmidt modes preserves two-mode quantum advantage.

Filtering of the output further restricts phase estimation to best-compensated modes, eliminating deleterious side modes [2012.03751].

## 6. Enhanced Phase Sensitivity via Output Photon Subtraction

Photon subtraction at the output port of the interferometer (operator $M_m = a^m$) is an advanced protocol that enhances phase sensitivity and quantum Fisher information (QFI). For input $|0\rangle_a \otimes |\alpha\rangle_b$, subtraction of $m$ photons induces a non-Gaussian transformation that tightens both the SQL and HL, especially in super-Poissonian regimes:

- Phase sensitivity $\Delta\phi$ and QFI $F_Q$ improve monotonically with $m$ (number of subtracted photons).
- Internal photon losses (within the interferometer) have a more significant adverse effect than external losses. Photon subtraction partially compensates internal loss, maintaining sub-SQL operation up to $T \approx 0.6$.
- For $m \ge 2$, the scheme can approach HL over a wide range of losses.
- The protocol is robust to input (vacuum vs. coherent state), and the refinement is mathematically explicit in error propagation and QFI expressions [2410.17612].

## 7. Comparative Analysis and Future Directions

The Yurke-type SU(1,1) interferometer has been implemented across photonic (integrated and fiber-based [2012.03751, 2202.05025]), atomic [1711.04552, 2309.12980], and optomechanical platforms [2509.22248]. It exceeds the SQL for phase sensitivity by leveraging two-mode quantum correlations and active nonlinear operations.

Key technical challenges remain:

- Achieving single-Schmidt-mode dominance in multimode platforms to preserve Heisenberg scaling.
- Suppressing internal loss, which is more detrimental than detection inefficiency.
- Implementing time-reversal protocols (matched gain, dark fringe operation) for robust performance against technical noise and instabilities.

The future of Yurke-type SU(1,1) interferometry lies in integrated photonics, robust multimode dispersion engineering, advanced quantum-enabled detection, and hybrid matter-light platforms, with immediate applications in quantum metrology, imaging, and simulation.

Source: https://www.emergentmind.com/topics/yurke-type-su-1-1-interferometer