Simulating Universal Quantum Gate Sets on Photonic OAM Qubits: Single-Qubit and Multi-Qubit Operations via Spatial Light Modulator Phase Holography

Technical Summary: Simulating Universal Quantum Gate Sets on Photonic OAM Qubits

Problem Statement
Despite significant experimental progress in photonic quantum information processing using Orbital Angular Momentum (OAM), a critical gap exists in the literature regarding the prediction of gate fidelity based on specific hardware imperfections. Existing simulation frameworks often treat device errors as a single, unexplained lumped parameter or are tailored to custom architectures (e.g., deep neural networks) that do not generalize to standard, single-plane holographic gates used in laboratory settings. Consequently, researchers utilizing commercial Spatial Light Modulators (SLMs) lack a principled, datasheet-traceable method to predict process fidelity or identify the dominant noise sources limiting their systems before experimental implementation. Furthermore, the impact of phase-stroke constraints in transmissive twisted-nematic (TN) SLMs, particularly regarding phase-wrap clipping at longer wavelengths, remains unquantified in the context of OAM quantum gates.

Methodology
The authors developed a comprehensive, hardware-grounded simulation framework for the HOLOEYE LC 2012 transmissive SLM. The methodology is structured around three core components:

1. Hardware Noise Modeling: Instead of free-parameter fitting, the authors derived a three-channel analytical noise model directly from the manufacturer's datasheet. This model comprises:

   • 8-bit Quantisation Noise: Calculated based on the effective phase levels (230 levels at 532 nm) and the 1.8π maximum phase retardation.
   • TN Electronic and Thermal Noise: Modeled as zero-mean Gaussian noise ($\sigma_{TN} = 80$ mrad), consistent with characterisation data for TN-LC technology.
   • Phase-Wrap Clipping Error: Derived analytically for pixels where the ideal holographic phase exceeds the hardware's maximum retardation ($\Phi_{max}$). This error is wavelength-dependent, being zero at 450 nm but significant at 532 nm and higher.
   • The total noise is combined in quadrature, yielding $\sigma_{total} = 92.4$ mrad at 532 nm.

2. Simulation Pipeline: The framework simulates operations on a $512 \times 512$ computational grid scaled to the SLM's active area. It models Laguerre-Gaussian (LG) modes where OAM states encode qubits (single-qubit: $\ell=0, +1$; two-qubit multiplexed: $\ell=-3, -1, +1, +3$). The pipeline applies ideal phase masks (fork gratings for mode conversion, uniform masks for phase shifts) followed by the sequential application of quantisation, clipping, and Gaussian noise.

3. Gate Set and Metrics: The simulation implements the complete universal single-qubit gate set $\{X, Y, Z, S, T, H\}$ and two-qubit entangling gates $\{CNOT, CZ, SWAP\}$. Performance is evaluated using state fidelity ($F = |\langle \psi_{ideal} | \psi_{noisy} \rangle|^2$), diffraction efficiency, and noise component isolation.

Key Contributions
The paper presents five principal contributions:

• Datasheet-Derived Noise Model: A rigorous analytical noise model for the HOLOEYE LC 2012 covering quantisation, TN noise, and phase-wrap clipping across four wavelengths (450, 532, 633, 800 nm) without empirical fitting.
• Universal Gate Simulation: A full simulation of the universal single-qubit and two-qubit gate sets on an OAM-encoded platform, providing per-gate fidelity analysis.
• OAM Multiplexing Framework: A demonstration of encoding two logical qubits into four orthogonal OAM modes on a single SLM aperture using composite holographic masks, showing no additional fidelity penalty compared to single-qubit operations.
• Bell State Preparation: A step-by-step simulation of Bell state generation ($\Phi^+$) via an H-CNOT circuit, verifying the OAM amplitude spectrum and predicting fidelity after two SLM interactions.
• Literature Benchmarking: A quantitative comparison of the simulation results against six published experimental studies, placing the theoretical predictions within the context of existing experimental fidelities (78%–99.6%).

Results

• Gate Fidelity: Under the full noise model at 532 nm, fork grating gates (X, Y, H) achieved a fidelity of $F = 0.9914$, limited primarily by phase-wrap clipping and TN noise. Phase-only gates (Z, S, T) achieved higher fidelity ($F = 0.9936$) because their required phase shifts map exactly to integer grey levels, eliminating quantisation error and clipping.
• Noise Dominance: Isolation analysis revealed that TN electronic noise is the dominant fidelity-limiting mechanism, accounting for approximately 76% of the total infidelity. Quantisation noise contributed less than 0.01%, confirming 8-bit addressing is sufficient.
• Wavelength Dependence: Performance degrades sharply at longer wavelengths due to reduced phase stroke. At 800 nm, the phase stroke drops to 1.0π, causing a fidelity collapse to $F = 0.7661$ and a diffraction efficiency drop to 13.6%. The optimal operating regime is identified as 450–532 nm.
• Multi-Qubit Operations: The CNOT, CZ, and SWAP gates achieved simulated fidelities of $F \approx 0.9915$. The Bell state preparation circuit yielded $F(\Phi^+) = 0.9914$.
• Efficiency: The first-order diffraction efficiency is limited to 32.6% at 532 nm, primarily due to the 58% fill factor of the device.

Significance and Claims
The paper claims to provide the first unified, hardware-grounded simulation pipeline that integrates specific device imperfections with full-field LG mode-coupling models to predict gate fidelity for commercial SLMs. By demonstrating that predicted fidelities (99.1%–99.4%) sit at the upper end of the range of published experimental results (78%–99.6%), the authors argue that their model accurately captures the intrinsic hardware limits of the HOLOEYE LC 2012.

The work establishes that the HOLOEYE LC 2012 is capable of supporting a universal gate set with high fidelity, provided the system operates in the 450–532 nm range. The authors note that the primary limitations are the TN electronic noise and the fill factor, suggesting that upgrading to LCoS technology could significantly improve performance. The simulation validates the feasibility of OAM multiplexing for multi-qubit photonic processing, offering a scalable approach to encoding multiple qubits within a single optical beam without inter-qubit cross-talk in the ideal case. The paper concludes that while the simulation provides a well-defined upper bound on achievable fidelity, experimental validation is required to account for additional real-world factors such as wavefront aberrations and mode-matching losses.