Hollow-Core Anti-Resonant Fibers (NC-ARF)

• NC-ARF are microstructured optical fibers featuring a hollow core surrounded by thin capillary walls that use anti-resonant reflection for light confinement.
• Optimized capillary geometry and material choices yield high mode purity, unique polarization control, and broad spectral operation ideal for high-power and sensing applications.
• Precision engineering of wall thickness, nested designs, and asymmetry enables dispersion tailoring, ultrafast nonlinear dynamics, and robust bend insensitivity.

Hollow-Core Anti-Resonant Fibers (NC-ARF) are a high-performance class of microstructured optical fibers in which light is guided by the anti-resonant reflection of thin dielectric membranes surrounding a central hollow core. By tailoring capillary geometry and leveraging the anti-resonance mechanism, NC-ARFs enable ultra-low propagation loss, high mode purity, unique polarization properties, and broad spectral operation. This versatility has driven the emergence of NC-ARFs as leading platforms for high-power pulse delivery, nonlinear optics, fiber-based sensing, micromachining, and advanced photonics.

1. Geometric Architectures and Anti-Resonance Principle

The defining feature of a hollow-core anti-resonant fiber is a central evacuated (or gas-filled) region, surrounded by a cladding composed of one or more rings of thin-walled capillaries or tubes. Each capillary acts as a Fabry–Pérot cavity that, when off-resonant at the operating wavelength, reflects incident field back into the core, suppressing leakage. The anti-resonance condition for minimal coupling into the wall of thickness $t$ and refractive index $n$ is given by

$$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$

whereas strong leakage (resonance) occurs at

$$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$

The negative curvature of the capillary boundary suppresses coupling to leaky cladding states, enhancing confinement and lowering propagation loss.

Fiber cross-sections can be:

• Simple single-ring (“negative curvature,” e.g., six or eight capillaries) (Chang et al., 1 Feb 2025)
• Double or multi-nested arrangements for further suppression of higher-order or surface modes (Gao et al., 2024)
• Polygonal or non-circular (e.g., triangular) (Chen et al., 2016)
• Multi-core configurations, each core individually surrounded by antiresonant tubes (Mears et al., 2024)
• Hybrid, combining distinct wall materials such as silica and chalcogenide glasses (Herring et al., 2023)

Design parameters such as capillary count, wall thickness, gap, and geometric ratios allow tuning of loss, modal content, and spectral bandwidth.

2. Modal Guidance, Loss, and Birefringence

The fundamental mode in an NC-ARF resembles the lowest-order hybrid mode of a dielectric capillary. Away from resonance, the effective index can be approximated by

$$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$

where $a$ is the core radius and $u_{01}\approx 2.405$ for LP$_{01}$. The propagation loss is dominated by confinement (leakage), scaling approximately as

$$\alpha \propto \left(\frac{\lambda}{a}\right)^p, \quad p \approx 3-4$$

Higher-order modes (HOMs) exhibit increased spatial overlap with the cladding, hence suffer greater leakage. Suppression of HOMs can be greatly improved by increasing the capillary-to-core diameter ratio ($d/D$), tuning anti-resonant conditions, or introducing double-nested/truncated geometries; for example, a 4T-DNANF achieves fundamental mode loss $n$0 dB/km and HOM loss $n$16 dB/m, yielding mode extinction ratios $n$250,000 (Gao et al., 2024).

Birefringence is engineered by breaking cross-sectional symmetry, such as employing tubes of differing wall thicknesses or hybrid tube materials (silica/As$n$3Se$n$4) (Yerolatsitis et al., 2019, Herring et al., 2023). Phase birefringence values $n$5 of order $n$6 and polarization extinction ratios PER $n$7550 (27 dB) are achievable (Herring et al., 2023, Wang et al., 2 Jan 2026).

3. Polarization Control and Filtering

Conventional NC-ARFs support degenerate orthogonally polarized modes, challenging environmental stability and polarization-sensitive applications. Polarization control can be introduced by:

• Hybridizing the cladding with high-index materials in select tubes. A silica/As$n$8Se$n$9 design enables suppression $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$0 dB/m for the blocked polarization, low loss ($$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$1 dB/m) for the pass axis, and PER $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$2550 (Herring et al., 2023).
• Dual-ring, “polarization-filtering” architectures, where a subset of capillaries are made thicker and aligned with the block axis. This introduces a polarization-selective resonance, yielding PER $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$3 21 dB over 10 m (Wang et al., 2 Jan 2026).
• Reduced symmetry by varying wall thickness or tube size (including triangular or compressed capillaries) (Yerolatsitis et al., 2019).

Polarization extinction is robust under mechanical bending (PER reduces moderately with radius). Such fibers are targeted for in-fiber polarizers, gyroscopes, and quantum optical devices requiring stable single-polarization transmission.

4. Dispersion Engineering and Ultrafast Nonlinear Dynamics

The group-velocity dispersion (GVD) and higher-order dispersion in NC-ARFs are highly engineerable by adjusting core size, wall thickness, and the number/nesting of tubes. Away from resonance, GVD is well captured by empirical or analytical models, e.g., (Hasan et al., 2017, Bache et al., 2018): $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$4 Key dispersion features:

• Broad low-GVD windows (e.g., $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$5–$$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$6 ps/nm/km flat dispersion at $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$7–$$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$8 μm (Chen et al., 2016))
• Rapid GVD oscillations and “cusps” in proximity to anti-resonance, critically affecting soliton dynamics and modulational instability (MI) (Hemsworth et al., 27 Oct 2025)
• Dispersion flattening achievable with nested or multi-shelled designs (Gao et al., 2024)

Ultrafast pulse compression, spectral broadening, and multi-stage dispersive wave generation in gas-filled NC-ARFs are controlled by fine-tuning the wall thickness to position anti-resonant bands and managing MI via dispersion profile engineering (Hemsworth et al., 27 Oct 2025, Habib et al., 2017, Lekosiotis et al., 2023). MI gain can be suppressed by selecting capillary wall thickness to keep anti-resonances outside the broadened spectrum, enabling high-peak-power and high-energy pulse stability.

5. Bend Sensitivity, Structural Robustness, and Tolerance Analysis

Low bending loss is crucial for practical deployment. Mechanisms influencing bend loss include coupling of the core mode to lossy cladding and Fano-type resonances at structural “nodes.” Nodeless designs (non-touching capillaries) eliminate such resonances, achieving $$\lambda_{AR,m} = \frac{2 t n}{m + 1/2},\quad m = 0, 1, 2, ...$$9 dB/m at a $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$0 cm radius and maintaining performance across $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$1–$$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$2 nm (Gao et al., 2016). Nested or multi-ring architectures can further reduce losses but trade off increased outer diameter and complexity.

Monte-Carlo studies highlight the importance of geometric tolerances: RMS thickness variation of $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$3 nm and tube angle offsets $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$4 limit loss penalties to $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$5 at $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$6 μm, while moderate angle randomness can even improve HOM suppression (HOMER scales as $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$7, $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$8/degree) (Petry et al., 2022). Bend amplifies the effect of imperfections, potentially increasing FM loss by $$\lambda_{R,m} = \frac{2 t n}{m},\quad m = 1,2,3,...$$9 at $$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$0 cm radius compared to straight fiber.

6. Applications and Integration

The ability of NC-ARFs to combine ultra-low loss and broad bandwidth in a hollow (or gas-filled) core drives their use in a range of advanced photonics contexts:

• High-peak-power pulse delivery: sub-50 fs, mJ-level pulses have been delivered through tens-of-meter evacuated and tightly coiled nested ARFs, with $$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$120 GW output and %%%%37$u_{01}\approx 2.405$38%%%% on-target intensity (Lekosiotis et al., 2023).
• Ultrafast laser micromachining: delivery of 20 W ps pulses in purely single-mode regime enables pointing stability and machining quality equal to free-space delivery, with greatly enhanced system flexibility (Chang et al., 1 Feb 2025).
• Quantum optics and fiber-based gyroscopes: robust single-polarization guidance and high mode extinction ratios support quantum state preservation and precision metrology (Wang et al., 2 Jan 2026, Herring et al., 2023).
• Sensing and nonlinear optics: rapid gas sensing designs (e.g., with lateral cuts for fast filling) and multi-core or modal architectures support distributed measurements and space-division multiplexing (Belardi, 2015, Mears et al., 2024).

Ongoing advances target kilometer-length scaling, sub-$$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$4 dB/km loss, bend insensitivity, and architectures tailored for specific nonlinear, sensing, laser, or quantum photonic applications.

7. Design Guidelines, Trade-Offs, and Outlook

Optimal NC-ARF performance requires precise engineering of structure and materials. Guidelines distilled from experimental and modeling data include:

• Select wall thickness $$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$5 so that anti-resonant bands accommodate the operational wavelength (avoid resonance overlap).
• Increase capillary-to-core ratio ($$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$6) to strongly suppress HOMs while balancing core size for loss and mode area.
• Employ double-nested and/or truncated geometries to maximize HOM extinction with minimal FM loss (Gao et al., 2024).
• Implement hybrid or asymmetric elements for polarization discrimination, targeting $$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$720–30 dB extinction at $$n_{\text{eff}}(\lambda) \approx 1 - \frac{u_{01}^2}{2 a^2 k_0^2}$$8 dB/m loss (Herring et al., 2023, Wang et al., 2 Jan 2026).
• Enforce wall thickness and gap tolerances at the sub-nm and sub-μm level for ultra-high mode purity and low loss (Petry et al., 2022, Gao et al., 2024).
• For low bend loss, avoid structural nodes, choose moderate core sizes, and employ nested elements judiciously (Gao et al., 2016, Mears et al., 22 Jan 2025).

Emerging directions include further material hybridization (e.g., wide-band IR/UV guidance, high-damage-threshold coatings), integration with acousto-optic functionalities (Silva et al., 16 Mar 2025), and expansion into multi-core/space-division multiplexed systems with ultra-low inter-core crosstalk (Mears et al., 2024). NC-ARFs thus delineate a rapidly evolving frontier in photonic fiber research, with a robust theoretical framework and maturing fabrication methodologies supporting their deployment in high-impact technological domains.