Scaling up multi-mode entanglement generated by mode swapping

AI summary

Overview

• Research area: Quantum physics — photonic quantum information, multi-photon interference in linear optical circuits, and multi-mode entanglement generation.
• Technical level: Intermediate. The paper is written in the formalism of creation/annihilation operators and Fock states, but the physical picture (modes interfering, photons being redistributed, post-selection) is accessible with undergraduate quantum optics background.
• Scope (one sentence): The paper extends a mode-swapping operation from two-mode to three-mode local systems and analyzes, for a specific asymmetric three-photon input state, how the swap redistributes photons and generates entanglement between the two spatially separated systems.

What This Paper Is About

The authors study a linear-optical interferometer made of two spatially separated three-mode systems, A and B, in which one mode from each side is exchanged ("swapped") in the middle of the circuit. The core problem is how this single swap changes the output photon-number statistics and creates non-local correlations, since the local transformations applied before and after the swap cancel each other when no swapping occurs. The goal is to build a systematic framework for scaling mode-swapping entanglement generation up to three or more modes, and to show that the choice of local photon-number inputs is an additional handle on the resulting interference.

Key Contributions

1. Extension of mode swapping to three-mode systems. The earlier mode-swapping scheme is generalized from a pair of two-mode-like systems to two three-mode systems, A and B, each with modes indexed 0, 1, 2 (M = 3).
2. A simplified operator-level description of the swap. The authors decompose each local creation operator into a part contained in the swapped mode and a part orthogonal to it, and show that the swap's action is confined to the two-mode subspace spanned by the swapped modes, appearing only through the difference of the two swapped-mode operators.
3. Explicit calculation of a three-photon output state. For the asymmetric input |200⟩_A|100⟩_B, the full output state and the joint photon-number distribution conditioned on the post-selection (N_A, N_B) = (2, 1) are derived.
4. Demonstration of entanglement between unequal photon numbers. The post-selected state entangles one photon in three possible states of system B with two photons in six possible states of system A, with many output combinations having exactly zero probability due to destructive interference.

Main Findings

• The swap acts only on two modes. Using a local discrete Fourier transform (DFT), the exchanged modes are chosen as unbiased, equal superpositions of the three input modes, ŝ_A† = (1/√M)Σ ȃ_m† and ŝ_B† = (1/√M)Σ b̑_m†. The mode transformation reduces to ȃ_m† → ȃ_m† + (1/√M)(ŝ_B† − ŝ_A†) for system A and b̑_m† → b̑_m† − (1/√M)(ŝ_B† − ŝ_A†) for system B, so the physics depends only on ŝ_B† − ŝ_A†.
• Two processes contribute to the (2,1) post-selection. The post-selected state is a sum of a "no photon in the swapped modes" amplitude and a "one photon in each swapped mode, exchanged reciprocally" amplitude.
• Post-selection probabilities. The no-swap process has probability P_no-swap = 8/27 and the reciprocal-exchange process P_swap = 4/27, giving a total post-selection success probability P_2,1 = 4/9. The authors note that even though only one of three local modes is swapped, the probability that each system keeps its initial photon number is smaller than one half.
• The original input distribution dominates but does not win outright. Conditioned on (2,1), the probability of recovering the input distribution |200⟩_A|100⟩_B is 4/9 (entry 16/36 in the joint probability table). All other allowed distributions have much smaller probabilities, mostly 2/36, with 1/36 for the symmetric B-mode/moved A-mode combinations.
• Zero-probability outcomes reveal coherent interference. Several entries of the joint distribution are exactly 0: |200⟩_A with |010⟩_B or |001⟩_B, |020⟩_A or |002⟩_A with |100⟩_B, and |110⟩_A or |101⟩_A with |100⟩_B. These zeros arise from complete destructive interference between the no-swap and swap amplitudes, not from a lack of contributing diagrams.
• Strong conditional correlations. If both photons in A are detected in mode 0, the photon in B is necessarily detected in mode 0. If the B photon is detected in mode 0, the photon number in mode 0 of A is either one or two but never zero — so both A photons cannot be found in mode 1 or mode 2.
• Symmetry between modes 1 and 2. The distribution is symmetric under exchange of input modes 1 and 2 in each system, because only mode 0 is populated initially while the swapped mode is an equal superposition of all modes.
• Input choice is a new degree of freedom. The authors emphasize that selecting different local photon-number inputs adds a new degree of freedom to the non-local interference effects seen in output photon-number statistics, which is the basis for strategies to efficiently generate entanglement with three or more modes.

Methodology in Plain English

The authors design a thought-experiment circuit rather than an experiment. Each of the two three-mode systems first passes through a local discrete Fourier transform, which mixes its three modes into equal superpositions. One superposition mode from A and one from B are then exchanged in the middle of the circuit, and the inverse Fourier transform is applied afterward. Because the forward and inverse transforms cancel when no swap occurs, any difference between input and output can be attributed purely to the swap.

To compute what happens, the authors use the standard trick of tracking how the operators that create photons transform. They split each photon-creation operator into a piece that lies inside the swapped mode and a piece orthogonal to it; the orthogonal piece passes through untouched, while the swapped piece is replaced by the corresponding mode of the other system. Because the swap only touches two modes, the algebra stays simple even with multiple photons. They then apply the transformed operators to a chosen input state — two photons in mode 0 of A and one photon in mode 0 of B — expand the resulting state in the photon-number basis, and keep only the terms where A still holds two photons and B still holds one. Squaring the amplitudes gives the joint probability table and the success probability.

Why This Matters

Mode swapping offers a way to entangle two matter-like or photonic multi-mode systems using a single exchange of one optical path, which is far simpler than building a large entangling gate network. Understanding how the interference scales with the number of modes and with the input photon-number pattern is a prerequisite for using such circuits as entanglement sources.

• Optical quantum computing: Multi-mode multi-photon interference in linear optical circuits is a core resource for photonic processors, and a compact swap-based entangler could simplify how correlations are generated between registers.
• Quantum communication: Non-local correlations between spatially separated systems are the raw material for entanglement-based communication protocols, and the swap is naturally a between-two-locations operation.
• Quantum metrology: Multi-photon entangled states with correlated photon-number statistics can improve phase estimation, which the authors list among the technologies demanding larger linear-optical circuits.
• Foundational characterization of multi-photon interference: The zero-probability outcomes provide clean, testable signatures that distinguish genuine multi-photon interference from classical mixing.

Industry relevance: Photonic quantum hardware is attractive because it operates at room temperature and is compatible with integrated waveguide fabrication. A scheme whose resource is "one swapped waveguide mode plus post-selection" is comparatively easy to map onto integrated photonic chips, making this analysis relevant to hardware groups designing multi-mode interferometers.

Future Directions

1. Scaling to more modes and photons. The authors state that the framework is intended as the basis for extending the analysis to optical quantum circuits with larger numbers of modes and photons; the general M-mode form of the transformation is already written, but only M = 3 is evaluated.
2. Surveying other input states. The paper analyzes one asymmetric three-photon input, |200⟩_A|100⟩_B, and notes that a wide variety of input states are possible and that most have highly non-trivial characteristics. Which inputs maximize or optimize the generated entanglement is left open.
3. Efficient entanglement-generation strategies. The abstract frames the work as considering strategies for efficient entanglement generation with three or more modes; turning the observed correlations into quantified entanglement (and comparing strategies) is a natural next step.
4. Experimental realization. No experimental implementation or measured data is reported; testing the predicted zero-probability outcomes and the 4/9 post-selection probability in an actual three-mode interferometer would directly validate the model.

Target Audience

Researchers and graduate students in photonic quantum information and quantum optics who work on multi-photon interference, linear-optical entanglement sources, or mode-swapping schemes; also useful for experimentalists designing integrated multi-mode interferometers, who will find the concrete probability table (post-selection success 4/9, input-recovery probability 4/9, and the exact zeros) directly comparable to their count statistics. Readers without comfort in second-quantized notation will need to work through the operator transformations to follow the derivations.

Authors’ abstract

Entanglement between two local multi-photon multi-mode systems can be generated by swapping a pair of modes between the two local multi-mode systems. In this presentation, we consider possible strategies for the efficient generation of entanglement between systems with three or more modes. It is shown that the choice of photon number inputs in the local systems adds a new degree of freedom to the non-local interference effects observed in the output photon number statistics.