Hardware Realization of a Hamiltonian Simulation Algorithm for Time-Domain Maxwells Equations

This paper presents the first quantum-hardware implementation of a Schrödingerisation-based algorithm for simulating time-domain Maxwell's equations, demonstrating accurate retrieval of electromagnetic field amplitudes and directions on an IonQ QPU for both benchmark problems and scattered fields.

The Gist

Imagine you are trying to predict how a ripple moves across a pond, but instead of water, the "pond" is the invisible space around us filled with electricity and magnetism. In the real world, these ripples (electromagnetic waves) follow strict rules called Maxwell's equations. Solving these rules on a normal computer is like trying to count every single grain of sand on a beach while the tide is coming in—it gets incredibly slow and expensive as the beach gets bigger.

This paper describes a team's attempt to solve this problem using a quantum computer, which is a special kind of machine that uses the weird rules of quantum physics to process information. Here is a simple breakdown of what they did:

1. The Problem: The "Non-Unitary" Puzzle

Quantum computers are like dancers; they are great at performing specific, reversible moves (called "unitary" operations). However, the math describing how electric and magnetic fields change over time is a bit messy and "non-reversible" (non-unitary) when you break it down into small steps. It's like trying to teach a dancer to walk backward through a wall—the standard dance moves don't fit.

2. The Solution: "Schrödingerisation" (The Magic Elevator)

To fix this, the authors used a trick called Schrödingerisation.

• The Analogy: Imagine you have a messy, tangled ball of yarn (the non-unitary math) that you can't untangle. Instead of trying to untangle it directly, you put the whole ball into a special elevator (the Schrödingerisation process) that lifts it up to a higher floor where the rules are different. On this higher floor, the tangled yarn magically becomes a neat, reversible dance routine that a quantum computer can handle perfectly.
• Once the computer finishes the dance, they take the result back down the elevator to get the answer they need.

3. The Dance Moves: Bell-Basis Decomposition

Even with the elevator trick, the dance routine was still too long and complicated for today's quantum computers.

• The Analogy: Think of the math as a massive instruction manual for a dance. The authors found a way to rewrite the manual using a special shorthand called Bell-basis decomposition. Instead of writing out every single step in a long, boring list, they grouped the steps into efficient "blocks" (like choreographed moves in a musical). This made the dance routine much shorter and faster to perform.

4. The Tricky Part: Reading the Signs

Quantum computers have a weird quirk: when you look at the result, you can see how strong a wave is, but you often lose track of which way it's pointing (positive or negative). It's like seeing a car's speedometer but not knowing if it's driving forward or backward.

• The Fix: The team invented a clever measurement trick. They added a tiny, known "offset" (like adding a constant weight to one side of a scale) to the starting electric field. This forced the computer to keep the numbers positive during the dance. After the dance was over, they simply subtracted that weight back out. This allowed them to figure out not just the strength of the field, but also its direction (the "sign"), which is crucial for understanding physics.

5. The Results: From Simulation to Real Hardware

• The Test Drive: First, they ran the algorithm on a simulator (a fake quantum computer running on a regular laptop). It worked perfectly, matching the known math answers for 2D and 3D scenarios, including cases with obstacles (like a wall inside the pond).
• The Real Deal: Then, they ran it on a real quantum computer made by IonQ (a machine that uses trapped ions, like tiny charged atoms, as qubits).
  • The Challenge: The original dance routine was too deep (too many steps) for the real machine to handle without getting confused by noise.
  • The Compression: They used a smart tool called ADAPT-AQC to "compress" the dance. It's like taking a 40,000-step instruction manual and condensing it into a 200-step version that still teaches the same dance, just with fewer moves.
  • The Outcome: Even with the real machine's noise and imperfections, the results looked very similar to the perfect math solutions. They successfully measured the electric and magnetic fields at specific points, proving that a quantum computer can simulate these physical waves.

Summary

In short, this paper is the first time anyone has successfully taken a complex physics problem (how light and radio waves move), translated it into a language a quantum computer can speak, compressed the instructions so it fits on today's machines, and actually run it on real hardware to get the correct answer. They didn't just simulate the math; they figured out how to read the "direction" of the waves, which is a major step forward for using quantum computers to solve real-world engineering problems.

Technical Summary

1. Problem Statement

The paper addresses the challenge of simulating time-domain electromagnetic fields governed by Maxwell's equations on quantum hardware. Key difficulties include:

• Vector-Valued Nature: Maxwell's equations describe coupled vector fields (Electric $\mathbf{E}$ and Magnetic $\mathbf{H}$) with magnitude and direction, whereas quantum measurements typically yield state information only up to a global phase, making the recovery of physical signs (directions) non-trivial.
• Non-Unitary Dynamics: Spatial discretization (e.g., Finite-Difference Time-Domain, FDTD) and boundary conditions often result in non-unitary evolution operators, which are incompatible with standard unitary quantum gates.
• Scalability: Existing quantum PDE solvers often suffer from high circuit depths, classical optimization bottlenecks (in variational approaches), or are limited to scalar equations rather than vector systems.
• Measurement Overhead: Full state tomography to reconstruct the entire field is inefficient for large grids; a method to extract local observables is required.

2. Methodology

The authors propose a framework combining Schrödingerisation, Bell-basis decomposition, and Trotterization to map the problem onto a gate-based quantum computer.

A. Discretization and Formulation

• FDTD Scheme: Maxwell's curl equations are discretized on a staggered Yee grid using central finite differences, converting the PDEs into a system of coupled ordinary differential equations (ODEs): $\frac{d\mathbf{u}}{dt} = A\mathbf{u}$, where $\mathbf{u}$ contains stacked $\mathbf{E}$ and $\mathbf{H}$ components.
• Boundary Conditions: Perfect Electric Conductor (PEC) and Perfect Magnetic Conductor (PMC) conditions are enforced by modifying discrete curl operators and using ghost-field extensions (antisymmetric for PEC, symmetric for PMC).
• Padding: Field vectors are padded with zeros to ensure dimensions are powers of 2, facilitating quantum encoding.

B. Schrödingerisation

To handle the non-unitary generator matrix $A$:

• The generator is decomposed into Hermitian parts: $A = H_1 + iH_2$.
• An auxiliary real variable $p$ is introduced to "lift" the system into a higher-dimensional space, transforming the non-unitary dynamics into a family of unitary Schrödinger equations: $\frac{d\tilde{v}}{dt} = i(\xi H_1 + H_2)\tilde{v}$.
• This allows the use of standard Hamiltonian simulation techniques.

C. Bell-Basis Decomposition

To reduce circuit depth compared to standard Pauli decompositions:

• The Hermitian matrices $H_1$ and $H_2$ are decomposed into tensor-product blocks.
• Instead of Pauli strings, the authors use Bell-basis decomposition. This involves mapping rank-one operators to Bell states using a unitary operator $O$, reducing the complexity of derivative operators.
• The evolution $e^{iSt}$ is implemented as $O (CRZ) O^\dagger$, where $CRZ$ is a multi-controlled RZ gate. This significantly lowers the circuit depth for derivative-based matrices.

D. Sign Reconstruction (Measurement Strategy)

To recover the physical direction (sign) of the fields:

• A constant offset $C$ is added to a reference field component (e.g., $E_z \to E_z + C$) such that the value remains strictly positive throughout the simulation.
• After simulation, the offset is subtracted to restore the original sign.
• Relative phases between the reference field and other components are measured to determine the signs of all field components at specific grid points, avoiding full state reconstruction.

E. Hardware Optimization

• Circuit Compression: Due to the depth of Trotterized circuits, the authors use ADAPT-AQC (Adaptive Approximate Quantum Compiling). This variational algorithm iteratively adds two-qubit unitaries to approximate the target circuit, reducing depth while maintaining fidelity.
• Error Mitigation: The implementation on IonQ hardware utilizes the device's native debiasing techniques.

3. Key Contributions

1. First Hardware Implementation: This is the first demonstration of a quantum algorithm producing signed vector-field solutions for time-domain Maxwell's equations on actual quantum hardware.
2. Schrödingerisation + Bell-Basis: A novel combination of Schrödingerisation for non-unitary dynamics and Bell-basis decomposition for efficient circuit construction, specifically tailored for vector-valued PDEs.
3. Sign-Recovery Protocol: A practical measurement strategy that retrieves both magnitudes and physical directions of electromagnetic fields at selected points without global tomography.
4. Boundary Condition Handling: Successful integration of PEC and PMC boundary conditions, including internal scatterers, within the quantum framework.

4. Results

The study validates the algorithm through classical simulation and hardware execution:

• Classical Simulations (Qiskit):

  • 2D & 3D: Simulations on 2D (32x32 grid) and 3D (16x16x16 grid) grids showed good agreement with analytical solutions.
  • Error Scaling: The $\ell_2$-norm error followed the expected first-order Trotter scaling ($O(dt)$), increasing linearly with time.
  • Scatterers: The algorithm successfully simulated wave propagation with internal PEC scatterers, capturing reflections and diffraction.

• Hardware Execution (IonQ Forte):

  • Setup: A 16x16 2D grid was simulated on a 36-qubit trapped-ion device.
  • Compression: ADAPT-AQC reduced the circuit depth from $\sim 40,000$ (naive Trotter) to $\sim 200$ gates, with CNOT counts below 150.
  • Performance: While hardware results were affected by noise, they qualitatively reproduced the correct temporal evolution of $E_z$, $H_x$, and $H_y$ at the grid center. The results matched the noise-free simulation trends, demonstrating feasibility on NISQ devices.

5. Significance and Future Work

• Foundational Step: This work lays the groundwork for solving computational electromagnetics problems on quantum computers, moving beyond scalar approximations to full vector-field simulations.
• Efficiency: The Bell-basis approach offers a scalable path for derivative operators, which are common in physics simulations.
• Practicality: The focus on local measurements rather than full state reconstruction makes the approach viable for larger problem sizes where global tomography is impossible.

Future Directions identified by the authors include:

• Extending to more complex geometries and dielectric bodies.
• Incorporating explicit source terms.
• Improving circuit efficiency via higher-order Trotterization or qubitization techniques.
• Scaling to larger spatial grids and longer time evolutions.