Photonic QC

Photonic Qubits

Photonic quantum computing uses photons as qubits, encoding information in:

• Polarization: , (horizontal/vertical)
• Path: which arm of an interferometer
• Time-bin: early or late arrival time
• Orbital angular momentum: different OAM modes

Photons are naturally robust to decoherence (they don't interact with the environment easily), but are hard to make interact with each other.

Linear Optical QC

Linear optical quantum computing (LOQC) uses only passive optical elements (beam splitters, phase shifters) and photon detectors.

The KLM protocol (Knill-Laflamme-Milburn) shows that LOQC is universal using:

• Single-photon sources
• Beam splitters
• Phase shifters
• Single-photon detectors
• Feed-forward based on measurement outcomes

The key resource is photonic entanglement created via beam splitters.

Photon Sources

Single-photon sources are critical for photonic QC:

• Spontaneous parametric down-conversion (SPDC): nonlinear crystal converts one photon to two entangled photons
• Quantum dots: semiconductor sources that emit single photons on demand
• Nitrogen-vacancy centers: solid-state single-photon sources
• Attenuated lasers: weak coherent pulses (simpler but probabilistic)

The indistinguishability of photons is critical for interference-based gates.

Measurement-Based QC

Measurement-based quantum computing (MBQC) or cluster state QC:

1. Prepare a highly entangled cluster state
2. Perform single-qubit measurements in adaptive bases
3. The measurement pattern determines the computation

Cluster states can be created using:

• Photonic crystal waveguides
• SPDC sources + beam splitters
• Trapped ion arrays

MBQC is well-suited to photonic platforms because entanglement is created first, then consumed by measurements.

Gaussian Boson Sampling

Gaussian boson sampling (GBS) is a photonic quantum advantage task:

1. Input squeezed vacuum states into a linear optical network
2. Measure photon counts at the output
3. The output distribution is hard to sample classically

GBS is related to the permanent and hafnian of matrices, which are #P-hard to compute. This is used by Xanadu's photonic quantum computers.

Photonic Challenges

• Photon loss: every optical component has loss
• Photon-photon interaction: weak for linear optics
• Detection: single-photon detectors have finite efficiency
• Indistinguishability: photons must be identical for interference

Solutions: nonlinear optical materials, atomic interfaces, photonic crystal cavities.

Python: Linear Optics

import numpy as np

def beam_splitter(theta):
    # Beam splitter transformation matrix.
    return np.array([
        [np.cos(theta), 1j*np.sin(theta)],
        [1j*np.sin(theta), np.cos(theta)]
    ], dtype=complex)

def hong_ou_mandel(bS1, bS2):
    # Hong-Ou-Mandel interference: two photons on a beam splitter.
    # Two-photon input |2,0>
    # Output should be |2,0> + |0,2> (bunching)
    BS = beam_splitter(np.pi/4)  # 50:50
    # Output amplitudes for |2,0>, |1,1>, |0,2>
    print("HOM interference at 50:50 BS:")
    print("Input: |2,0>")
    print("Output: |2,0> + |0,2> (photons bunch)")

hong_ou_mandel(1, 0)

Photonic Quantum Computing Platforms

Photonic Challenges and Solutions

Summary

This topic covers the fundamental concepts and applications in quantum computing. Understanding these concepts is essential for advancing in the field and applying quantum techniques to real-world problems. The mathematical framework provides the foundation for analyzing quantum algorithms and hardware implementations.

Key takeaways include the importance of quantum coherence, the role of entanglement as a resource, and the tradeoffs between different quantum computing architectures. As the field progresses from NISQ to fault-tolerant devices, these foundational concepts will continue to underpin new developments and applications.

Further study should include hands-on implementation using quantum programming frameworks, analysis of recent research papers, and exploration of the connections between quantum computing and other fields such as machine learning, optimization, and simulation.