High-Dimensional Quantum Key Distribution

• High-dimensional QKD is a quantum cryptography method that uses d-level systems to encode information, increasing secret-key rates and noise resilience.
• It employs diverse encoding schemes—including time-bin, frequency-bin, and spatial modes—to support robust communications over fiber and free-space channels.
• Experimental implementations demonstrate enhanced key rates and security through advanced post-processing, error correction, and photonic integration techniques.

High-dimensional Quantum Key Distribution (HD-QKD) generalizes conventional qubit-based QKD by exploiting d-dimensional Hilbert spaces (qudits), enabling increased secret-key rates, improved noise tolerance, and enhanced security. HD-QKD encompasses time-bin, frequency-bin, spatial (OAM, multicore), polarization, and hybrid encodings, supporting versatile quantum communication protocols both in fiber and free-space channels.

1. Mathematical Principles and Protocol Design

The essential feature of HD-QKD is the encoding of information into d-dimensional orthonormal bases $\{\ket{n}\}_{n=0}^{d-1}$ and their mutually unbiased counterparts. In prepare–measure schemes, Alice selects a symbol $k$ from the computational basis or prepares a superposition in a conjugate basis (e.g., Fourier basis $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$). Bob randomly selects a basis and measures; correct basis matches yield raw key bits.

Protocols include:

• Time-bin HD-COW QKD: Alice encodes a single weak coherent pulse in one of $d$ time slots per block; Bob performs direct detection (key) and interferometric monitoring (coherence test) (Sulimany et al., 2021).
• Dispersive optics HD-QKD: Alice and Bob select either direct time-bin measurement or dispersive Fourier conjugate via group-velocity dispersion (GVD), enabling high-rate key exchange with entropic uncertainty-based security (Mower et al., 2012).
• Spatial-mode HD-QKD: Information is encoded in OAM, multicore fiber, or 3D vector-polarized spatial modes, with programmable mode sorters (MPLC, inverse design) enabling up to $d=25$ (Lib et al., 2024, Otte et al., 2023).
• Fourier-qubit HD-QKD: A non-mutually-unbiased protocol, each state is a superposition of two computational basis levels with one of $d$ possible phases, simplifying preparation and measurement while retaining dimensional security (Scarfe et al., 4 Apr 2025).
• Restricted basis HD-QKD: Protocols requiring only one full basis and a single test state from a second (e.g., 3-state HD-BB84), facilitating implementations with limited quantum-control (Iqbal et al., 2023).
• Round-robin differential-phase-shift (RRDPS) HD-QKD: Security does not rely on disturbance monitoring; high-dimensional phase encoding allows flexible trade-offs between key rate and noise tolerance (Stasiuk et al., 2023).

Common to these schemes is the use of $d$-ary Shannon entropy $h_d(e) = -e\log_2(e/(d-1))-(1-e)\log_2(1-e)$ as the fundamental metric of noise and information.

2. Secret Key Rate Formulation and Security Analysis

The asymptotic secret-key rate per sifted photon often takes the form:

$R(d, e) \geq \log_2 d - H_d(e) - e\log_2(d-1)$

where $e$ is the quantum bit error rate (QBER) arising from cross-talk, dark counts, or ambient noise. In protocols with two or more mutually unbiased bases, $k$0 appears twice (once per basis), further penalizing noise.

Security proofs are based on:

• Entropic uncertainty relations: Guarantee bounds on Eve’s information via the overlap of POVMs in conjugate bases; the overlap parameter $k$1 for rank-1 projectors on coherent states (Sulimany et al., 2021, Mower et al., 2012).
• De Finetti reduction and composable security: Protocols permute blocks for i.i.d. reduction, allowing reduction of general coherent attacks to collective attacks with finite-size smoothing gaps (Kanitschar et al., 6 May 2025).
• Holevo bound and SDP techniques: For continuous-variable protocols, mutual information $k$2 and Eve’s Holevo information $k$3 are computed from covariance matrices of measured statistics (Mower et al., 2012, Liu et al., 2022, Islam et al., 2019).
• Decoy-state methodology: Multi-photon emission is bounded using one or more classical intensities. Secure key rates are achieved with only one or two decoy states at distances $k$4 km and multiple secure bits per photon (Bunandar et al., 2014).

Finite-size security is addressed by entropic uncertainty relations over recommendable acceptance sets based on experimentally accessible observables, with variable-length privacy amplification providing strictly higher expected rates under rapid channel fluctuations (Kanitschar et al., 6 May 2025, Niu et al., 2016).

3. Experimental Realizations and System Architectures

HD-QKD has been experimentally demonstrated across diverse platforms:

• Time-bin systems: Standard COW hardware (pulsed laser, intensity modulator, unbalanced MZI, two detectors) supports $k$5-dimensional encoding without hardware changes; SKR enhanced by log$k$6 factor (Sulimany et al., 2021).
• Dispersive-optics schemes: CW SPDC, heralded single-photon sources, and GVD modules (fiber Bragg gratings, silicon PICs) enable frames of up to $k$7, with bits per photon up to 4 (Mower et al., 2012).
• Multicore fiber/PICs: Silicon photonics with integrated MZIs and VOAs enable robust $k$8 logical encoding over MCF, keeping QBER < 19% at long reach (Ding et al., 2016).
• Spatial mode sorting / MPLC: Ten-plane SLMs programmed via wavefront-matching sort $k$9–$$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$0 spatial modes into multiple MUBs. Block-biased error structure enables robust, scalable high-dimensional key rates ($$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$1–$$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$2 bits/photon) (Lib et al., 2024).
• OAM encoding / quantum dot SPS: Deterministic room-temperature SPS yields $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$3 OAM encoding, experimentally delivering $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$4 bit/photon secure key rate (Halevi et al., 2024).
• Vector beam inverse design: On-chip SOI nanophotonic antenna prepares/measures full 3D-polarized spatial MUBs, almost doubling achievable key rates and halving QBER compared to 2D-only schemes (Otte et al., 2023).

Resource-efficient detection has been achieved via temporal Talbot effect for time–phase HD-BB84, requiring only one detector per basis (Ogrodnik et al., 2024). Two-photon interference (quantum-controlled measurement) in time–phase protocols further eliminates interferometric scaling bottlenecks (Islam et al., 2019).

4. Noise Resilience, Range, and Channel Integration

HD-QKD protocols demonstrate enhanced tolerance to detector and ambient noise:

• QBER threshold for positive key rate increases with dimension: 11% for $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$5, 19% for $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$6, 24% for $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$7 (Zhang et al., 12 Dec 2025).
• Environmental robustness: In hybrid quantum–classical access networks, $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$8 time–phase encoding significantly outperforms $$|\psi_m\rangle = \frac{1}{\sqrt{d}} \sum_{n=0}^{d-1} e^{2\pi i m n/d} |n\rangle$$9 under high Raman and ambient noise; maintains Mbps rates at 10 km fiber even with substantial background (Elmabrok et al., 2022).
• Long-distance feasibility: Dispersive optics HD-QKD and decoy-state HD-QKD protocols demonstrate multi-bit/sifted photon rates over loss budgets up to 200 km (fiber) and $d$050 dB channel attenuation (Bunandar et al., 2014, Mower et al., 2012, Liu et al., 2022).
• Spatial mode self-healing: Bessel–Gaussian hybrid spin–orbit encoding enables secure transmission through line-of-sight obstacles, maintaining QBER up to three times lower than standard LG modes (Nape et al., 2018).

Composable finite-size security proofs for HD-QKD show keys become positive at block sizes $d$1–$d$2, and variable-length privacy amplification yields $d$3–$d$4 higher average key rates under fluctuating loss/noise, critical for satellite/free-space channels (Kanitschar et al., 6 May 2025).

5. Classical Post-Processing and Information Reconciliation

Reconciliation efficiency directly impacts the HD-QKD system throughput and saturates at or near the Slepian–Wolf bound:

• Nonbinary LDPC codes over GF($d$5): Achieve $d$6–$d$7 for $d$8, with computational cost $d$9 (Mueller et al., 2023).
• Generalized Cascade protocols: Interactive, high-throughput variants exploit symbol-wise parity exchanges and “partner bits” yielding $d=25$0–$d=25$1 in $d=25$2–$d=25$3 (Mueller et al., 2023).
• HD information reconciliation enables $d=25$4% higher throughput and up to $d=25$5 dB additional channel loss tolerance over binary methods in $d=25$6 time-bin systems.

6. Scaling, Integration, and Future Directions

HD-QKD scales advantageously with dimension, but practical implementation faces detector dark counts, intermodal cross-talk, and complexity constraints:

• Scaling with dimension: Information per photon $d=25$7; maximum tolerable QBER increases with $d=25$8, but detector noise scales as $d=25$9 (Mower et al., 2012, Zhang et al., 12 Dec 2025).
• Programmability and photonic integration: MPLC (O($d$0) complexity for tailored MUBs), on-chip vector beam decoders, and silicon PIC-based schemes support dynamic MUB switching, low-loss transformation, and large-scale integration (Lib et al., 2024, Otte et al., 2023, Ding et al., 2016).
• Hybrid and novel coding: Spin–orbit, vector-polarization, multicore SDM, and OAM–time-bin hybrid states extend the accessible alphabet and redundancy against loss/turbulence (Zhang et al., 12 Dec 2025, Elmabrok et al., 2022, Nape et al., 2018).

Variable-length key distillation and dual-security frameworks accommodate highly fluctuating free-space and satellite QKD links (Kanitschar et al., 6 May 2025).

7. Experimental Performance Overview

Recent HD-QKD systems achieve secure key rates up to 100 Mbps at short range (fiber, $d$1–$d$2), maintain multi-bit/photon efficiency over $d$3200 km, and outperform qubit protocols under comparable loss and finite-key conditions:

Protocol	d	Range (km)	Key Rate (bits/photon)	QBER
Dispersive Optics	8–64	0–200	2–4 (short), 0 (200)	<15%
MPLC Spatial Modes	5,25	lab	1.57, 0.8	11–32%
OAM–QD SPS	3	lab	1.03±0.10	<4%
Multicore Fiber	4	1.2–25	0.41, 0.2	13%
Access Network	4	0–10	Mbps-scale	<10%

These results highlight the operational feasibility and key-rate enhancement of HD-QKD over traditional QKD, especially in challenging noise and loss regimes.

---

High-dimensional QKD thus constitutes a mature, versatile, and scalable extension to quantum cryptographic systems. The protocol landscape encompasses entanglement-based, prepare–measure, restricted-basis, and hybridized frameworks. Advanced encoding, resource-efficient detection, photonic integration, and robust post-processing increasingly position HD-QKD for deployment in metropolitan, access, satellite, and hybrid quantum networks (Sulimany et al., 2021, Mower et al., 2012, Lib et al., 2024, Kanitschar et al., 6 May 2025, Zhang et al., 12 Dec 2025).