Optical Nanocapillary Fiber

• Optical Nanocapillary Fiber (NCF) is a subwavelength silica fiber with a hollow or liquid-filled core that embeds quantum emitters for efficient guided-mode channeling.
• NCF designs achieve channeling efficiencies up to 52% in bare fibers and up to 87% with composite photonic crystal cavities, significantly enhancing photon collection.
• Key advancements include optimized inner/outer diameters, polarization matching, and emitter placement tolerance, making NCFs ideal for fiber-integrated quantum devices.

Searching arXiv for recent NCF-related papers and the cited work to ground the article. arxiv_search(query="optical nanocapillary fiber single photons guided modes 2024 2025", max_results=10) Optical nanocapillary fiber (NCF) denotes a subwavelength fiber platform in which a silica nanofiber contains a hollow or liquid-filled core, so that a single quantum emitter can be placed inside the guided structure rather than on its surface. In the 2024 numerical study that established the basic single-photon channeling picture, the NCF is formed of a liquid core optical nanofiber with inner and outer diameters, and the central result is a maximum channeling efficiency of $52\%$ for a radially polarized single dipole source (SDS) at the center of a water-filled NCF with $d^{in}=100\ \text{nm}$, $d^{out}=360\ \text{nm}$, and emission wavelength $620\ \text{nm}$ (Elaganuru et al., 2024). Subsequent cavity-QED extensions use the same NCF concept as the core photonic medium and combine it with defect-mode gratings to realize channeling efficiencies up to $87\%$ in a symmetric composite cavity and about $80\%$ into one-sided guided modes in an asymmetric composite cavity (Gadde et al., 8 Jul 2025, Gadde et al., 7 Aug 2025).

1. Definition and structural model

An optical nanocapillary fiber is described as a silica nanofiber with a hollow or liquid-filled subwavelength core (Elaganuru et al., 2024). In the bare-fiber single-photon channeling study, the NCF consists of an outer silica cylindrical shell and an inner core that is either vacuum-filled or water-filled, with particular emphasis on the water-filled liquid-core NCF because it is more experimentally relevant for quantum dots in solution (Elaganuru et al., 2024). In the cavity-QED formulation, the NCF is treated as a three-layer nanophotonic fiber with a subwavelength hollow core filled with water and a silica outer cladding, parameterized by inner and outer diameters $d_{in}$ and $d_{out}$ (Gadde et al., 8 Jul 2025).

This geometry distinguishes the NCF from a conventional optical nanofiber (ONF). In the NCF, the emitter is placed inside the capillary; in the ONF reference case, the emitter remains on the surface. That structural distinction is central to the reported enhancement in guided-mode coupling (Elaganuru et al., 2024). The NCF is also treated as effectively infinitely long in the cavity studies, so that longitudinal confinement can be introduced independently by an external photonic-crystal-like grating rather than by truncation of the waveguide itself (Gadde et al., 8 Jul 2025).

The relevant material configuration for the optimized non-cavity case is explicit: outside the cylinder is vacuum, the wall is silica, and the inner core is filled with water (Elaganuru et al., 2024). The later composite-cavity studies retain the water-filled capillary concept and use $d_{in}=125\ \text{nm}$ and $d_{out}=515\ \text{nm}$ as the optimized NCF dimensions for cavity operation near $d^{in}=100\ \text{nm}$0, matching common single quantum emitters such as quantum dots and NV centers (Gadde et al., 8 Jul 2025, Gadde et al., 7 Aug 2025).

2. Bare-fiber channeling physics

The fundamental operating principle is guided-mode channeling of spontaneous emission from an internal emitter. In the non-cavity study, the SDS represents a single quantum emitter such as a quantum dot or another single-photon emitter, and the optimized quantity is the channeling efficiency

$d^{in}=100\ \text{nm}$1

where $d^{in}=100\ \text{nm}$2 is the power coupled into the guided mode(s) and $d^{in}=100\ \text{nm}$3 is the total power emitted by the SDS in the presence of the fiber (Elaganuru et al., 2024).

The numerical analysis is performed with finite-difference time-domain (FDTD) simulations in Ansys Lumerical. The simulation region is $d^{in}=100\ \text{nm}$4, bounded by perfectly matched layers (PMLs). The fiber length is $d^{in}=100\ \text{nm}$5, the SDS is placed $d^{in}=100\ \text{nm}$6 from the PML, and the power monitor is $d^{in}=100\ \text{nm}$7 from the SDS and also $d^{in}=100\ \text{nm}$8 from the PML (Elaganuru et al., 2024).

For emission at $d^{in}=100\ \text{nm}$9, the optimized water-filled NCF has inner diameter $d^{out}=360\ \text{nm}$0 and outer diameter $d^{out}=360\ \text{nm}$1, yielding a maximum channeling efficiency of $d^{out}=360\ \text{nm}$2 for a centered, radially polarized SDS (Elaganuru et al., 2024). In vacuum, with the inner diameter fixed at $d^{out}=360\ \text{nm}$3, the maximum is lower, $d^{out}=360\ \text{nm}$4, at $d^{out}=360\ \text{nm}$5 (Elaganuru et al., 2024). The same study uses ONFs as a reference and reports markedly lower maxima for surface-coupled emitters: $d^{out}=360\ \text{nm}$6 at $d^{out}=360\ \text{nm}$7 in vacuum and $d^{out}=360\ \text{nm}$8 at $d^{out}=360\ \text{nm}$9 in water, both for radial dipoles (Elaganuru et al., 2024).

The interpretation is organized partly by the standard single-mode condition

$620\ \text{nm}$0

with $620\ \text{nm}$1 the fiber radius, $620\ \text{nm}$2, and $620\ \text{nm}$3, $620\ \text{nm}$4 the refractive indices of core and cladding (Elaganuru et al., 2024). This criterion is used to explain why the ONF optimum depends strongly on whether the surrounding medium is vacuum or water, and why multimode behavior appears at larger diameters (Elaganuru et al., 2024).

3. Polarization, symmetry, and emitter-position dependence

The coupling is strongly polarization dependent because the guided-mode electric-field distribution is not isotropic (Elaganuru et al., 2024). The paper compares radial, azimuthal, and axial dipole orientations. In the cylindrical geometry, radial and axial orientations are explicitly relevant, and the azimuthal case behaves similarly to the radial case when the dipole is centered because of symmetry (Elaganuru et al., 2024).

For the optimized water-filled NCF, radial polarization gives the largest efficiency, while axial polarization gives the smallest efficiency (Elaganuru et al., 2024). The stated reason for the weak axial coupling is that the axial electric-field component $620\ \text{nm}$5 is small near the center of the NCF; the centered radial dipole instead has strong overlap with the guided-field distribution (Elaganuru et al., 2024). The physical picture presented in the paper is that the guided mode field is strongest near the core region, a radially oriented dipole aligns with that field distribution, and a centered source in the liquid core maximizes symmetry and overlap (Elaganuru et al., 2024).

Emitter-placement tolerance is treated as a practical issue because experimental positioning inside the NCF is uncertain. The radial position $620\ \text{nm}$6 is varied from $620\ \text{nm}$7 at the center to $620\ \text{nm}$8 near the inner wall. For the optimum water-filled NCF $620\ \text{nm}$9, the channeling efficiency remains almost unchanged as the emitter position varies, especially for radial polarization (Elaganuru et al., 2024). By contrast, axial polarization remains very low, about $87\%$0 (Elaganuru et al., 2024). This robustness is one of the experimentally significant features of the liquid-core design.

A plausible implication is that the bare NCF does not require nanometric placement accuracy at a single sharply tuned hotspot in order to retain high guided-mode collection, at least in the optimized water-filled geometry. The source paper supports that inference by identifying a broad optimum region in which modest emitter-position uncertainty does not strongly degrade performance (Elaganuru et al., 2024).

4. Composite photonic crystal symmetric cavities on NCFs

The next stage in NCF development is the composite photonic crystal symmetric cavity (CPCSC), which combines the NCF waist region with a symmetric defect mode nano-grating (DMG) placed externally around the fiber (Gadde et al., 8 Jul 2025). In that work, the NCF is not only a guided-mode channel for spontaneous emission but also the confinement-enhanced medium required for cavity QED (Gadde et al., 8 Jul 2025). The channeling efficiency is written as

$87\%$1

where $87\%$2 is the emission rate into guided modes, $87\%$3 is the total emission rate, $87\%$4 is the coupled power, and $87\%$5 is the total emitted power (Gadde et al., 8 Jul 2025).

The DMG is described by grating period $87\%$6, defect width $87\%$7, slat thickness $87\%$8, slat height $87\%$9, and slat number $80\%$0 (Gadde et al., 8 Jul 2025). With optimized NCF dimensions $80\%$1 and $80\%$2, the cavity is designed to operate near $80\%$3 (Gadde et al., 8 Jul 2025). Because the DMG is symmetric about the defect center, the cavity field has standing-wave nodes and anti-nodes, and the single quantum emitter must be placed at the anti-node position to maximize coupling (Gadde et al., 8 Jul 2025).

The cavity field intensity decays exponentially away from the center and has an effective cavity length of about $80\%$4 (Gadde et al., 8 Jul 2025). The reported regime is the Purcell regime of cavity QED, using

$80\%$5

with $80\%$6 the Purcell factor, $80\%$7 the cooperativity, $80\%$8 the emitter-cavity coupling rate, $80\%$9 the cavity field decay rate, and $d_{in}$0 the emitter spontaneous emission rate (Gadde et al., 8 Jul 2025). For the optimized cavity, the paper gives $d_{in}$1, $d_{in}$2, and an estimate $d_{in}$3 for $d_{in}$4, identified as an NV center value (Gadde et al., 8 Jul 2025).

The central performance result is a maximum channeling efficiency of $d_{in}$5 when the single quantum emitter is placed at the anti-node position of the CPCSC (Gadde et al., 8 Jul 2025). At the anti-node, $d_{in}$6-polarized emitters couple best, and the strongest coupling occurs near $d_{in}$7 (Gadde et al., 8 Jul 2025). The cavity is polarization sensitive, with $d_{in}$8 for $d_{in}$9-polarization and $d_{out}$0 for $d_{out}$1-polarization, while the scattering-limited decay rate is $d_{out}$2 (Gadde et al., 8 Jul 2025).

The following values summarize the progression from bare NCFs to cavity-assisted NCFs:

Configuration	Key optimized geometry	Reported channeling efficiency
Water-filled NCF (Elaganuru et al., 2024)	$d_{out}$3, $d_{out}$4, $d_{out}$5	$d_{out}$6
CPCSC on NCF (Gadde et al., 8 Jul 2025)	$d_{out}$7, $d_{out}$8, $d_{out}$9	$d_{in}=125\ \text{nm}$0
One-sided composite cavity on NCF (Gadde et al., 7 Aug 2025)	$d_{in}=125\ \text{nm}$1, $d_{in}=125\ \text{nm}$2, $d_{in}=125\ \text{nm}$3	$d_{in}=125\ \text{nm}$4 into one-sided guided modes

5. One-sided composite cavities and directional channeling

A further extension is the one-sided composite cavity, formed by combining an optical NCF and an asymmetric defect mode grating (ADMG) (Gadde et al., 7 Aug 2025). The optimized NCF again uses $d_{in}=125\ \text{nm}$5 and $d_{in}=125\ \text{nm}$6, with the inner region treated as water and the outer wall as silica (Gadde et al., 7 Aug 2025). The grating retains period $d_{in}=125\ \text{nm}$7, defect width $d_{in}=125\ \text{nm}$8, duty cycle $d_{in}=125\ \text{nm}$9, slat thickness $d_{out}=515\ \text{nm}$0, and slat height $d_{out}=515\ \text{nm}$1 (Gadde et al., 7 Aug 2025).

The asymmetry is introduced through different slat numbers on the two sides of the defect, $d_{out}=515\ \text{nm}$2 and $d_{out}=515\ \text{nm}$3, so that the cavity channels the emitter’s radiation predominantly to a single output direction (Gadde et al., 7 Aug 2025). In this work, the channeling efficiency is defined as

$d_{out}=515\ \text{nm}$4

where $d_{out}=515\ \text{nm}$5 is the power coupled into NCF-guided modes and $d_{out}=515\ \text{nm}$6 is the total power emitted by the emitter in the cavity. The Purcell factor is

$d_{out}=515\ \text{nm}$7

with $d_{out}=515\ \text{nm}$8 the free-space emission power (Gadde et al., 7 Aug 2025).

The cavity is analyzed in over-coupling, critical-coupling, and under-coupling regimes using

$d_{out}=515\ \text{nm}$9

where $d^{in}=100\ \text{nm}$00 is the input coupling rate, $d^{in}=100\ \text{nm}$01 the scattering or intra-cavity loss rate, and $d^{in}=100\ \text{nm}$02 the total decay rate (Gadde et al., 7 Aug 2025). With $d^{in}=100\ \text{nm}$03, the reported cases are $d^{in}=100\ \text{nm}$04 for over-coupling, $d^{in}=100\ \text{nm}$05 for critical coupling, and $d^{in}=100\ \text{nm}$06 for under-coupling (Gadde et al., 7 Aug 2025).

Optimization over grating asymmetry yields a maximum channeling efficiency at $d^{in}=100\ \text{nm}$07 and $d^{in}=100\ \text{nm}$08, with $d^{in}=100\ \text{nm}$09 into one-sided NCF-guided modes (Gadde et al., 7 Aug 2025). For that design, the paper reports a quality factor $d^{in}=100\ \text{nm}$10, finesse $d^{in}=100\ \text{nm}$11, one-pass loss $d^{in}=100\ \text{nm}$12, and effective cavity length $d^{in}=100\ \text{nm}$13 (Gadde et al., 7 Aug 2025). Additional values are $d^{in}=100\ \text{nm}$14, $d^{in}=100\ \text{nm}$15, and an estimated $d^{in}=100\ \text{nm}$16 assuming $d^{in}=100\ \text{nm}$17 for NV centers in nanodiamonds (Gadde et al., 7 Aug 2025).

This directional design changes the function of the NCF from high-efficiency collection alone to high-efficiency routing. A plausible implication is that the relevant figure of merit is no longer only the total fraction entering guided modes but also the asymmetry of the exit channels, which is why the one-sided cavity is positioned as a route to fiber-based deterministic single-photon sources (Gadde et al., 7 Aug 2025).

6. Fabrication context and related nanofiber photonic structures

The core NCF channeling and cavity studies are numerical, but several related optical nanofiber works establish fabrication and photonic-structuring practices that are relevant to NCF-style devices. These studies concern optical nanofibers rather than nanocapillaries in the fluidic sense, yet they are directly informative for subwavelength-waist, low-loss, cavity-enabled fiber photonics (Nayak et al., 2012, Li et al., 2017, Su et al., 18 Mar 2026).

One route is direct photonic-crystal formation on an optical nanofiber by femtosecond laser ablation. Thousands of periodic nano-craters can be fabricated with a single femtosecond laser pulse, with the nanofiber itself acting as a cylindrical lens that focuses the beam on its shadow surface (Nayak et al., 2012). Using a Talbot interferometer with phase-mask pitch $d^{in}=100\ \text{nm}$18, the interference period is $d^{in}=100\ \text{nm}$19, which sets the crater spacing (Nayak et al., 2012). The resulting 1-D photonic crystal exhibits a broad reflection band centered at $d^{in}=100\ \text{nm}$20, a stop band approximately $d^{in}=100\ \text{nm}$21, FWHM $d^{in}=100\ \text{nm}$22, and reflectivity $d^{in}=100\ \text{nm}$23 in one sample; polarization-dependent stop bands are also reported in another sample (Nayak et al., 2012).

A second route is focused-ion-beam structuring of periodic air-nanohole arrays on an ONF waist. In the reported Type III morphology, a triplex periodic air-cube structure combines 1-D photonic crystal and Bragg grating behavior (Li et al., 2017). For a cavity with $d^{in}=100\ \text{nm}$24, cavity length $d^{in}=100\ \text{nm}$25, $d^{in}=100\ \text{nm}$26, $d^{in}=100\ \text{nm}$27, and $d^{in}=100\ \text{nm}$28, the experimental characterization yields a cavity quality factor $d^{in}=100\ \text{nm}$29 and a mode volume of $d^{in}=100\ \text{nm}$30 (Li et al., 2017).

Low-loss submicron-waist fabrication is addressed directly in a 2026 heat-and-pull study of silica ONFs. Using a multi-hole torch tip that provides a relatively large and uniform heating region, the authors achieve optical transmission above $d^{in}=100\ \text{nm}$31 for waist diameters as small as $d^{in}=100\ \text{nm}$32 for a $d^{in}=100\ \text{nm}$33-mm waist length and $d^{in}=100\ \text{nm}$34 for a $d^{in}=100\ \text{nm}$35-mm waist length (Su et al., 18 Mar 2026). The same work emphasizes cleaning, splicing, enclosure, and torch geometry as practical controls over nanofiber loss and long-term stability (Su et al., 18 Mar 2026). This suggests that any future experimental NCF implementation will likely depend not only on capillary geometry but also on the broader nanofiber fabrication discipline already developed for smooth-waist and grating-structured ONFs.

7. Applications, scope, and common misconceptions

The immediate significance of the NCF is as a fiber-integrated interface between a single quantum emitter and a guided optical mode. The non-cavity study explicitly positions the water-filled NCF as a route for generating single photons in quantum technologies and for detecting single cells in bio-sensing (Elaganuru et al., 2024). The cavity studies extend that role to single-photon collection, emitter-to-fiber interface engineering, quantum networking, fiber-coupled quantum light sources, and controlled photon routing or manipulation (Gadde et al., 8 Jul 2025, Gadde et al., 7 Aug 2025).

A central misconception is that an NCF is simply an ONF placed in liquid. The quantitative comparison does not support that equivalence: the surface-coupled ONF in water reaches a maximum $d^{in}=100\ \text{nm}$36, whereas the optimized water-filled NCF with the emitter inside the capillary reaches $d^{in}=100\ \text{nm}$37 (Elaganuru et al., 2024). The difference is tied to emitter placement inside the guided structure, subwavelength inner and outer diameters, and high field overlap with guided modes, especially for radial dipoles (Elaganuru et al., 2024).

Another misconception is that any nanoscale tapered fiber should be classified as an NCF. The micro-lensed single-mode fiber with a spherical or hemispherical tip is related to NCF ideas in spirit and geometry, but it is not an NCF in the strict materials or transport sense because it is a solid silica micro-lensed fiber with no capillary channel (Kato et al., 2013). Likewise, the multiplexed neural-recording optical fiber is described as close in spirit to an optical nanocapillary fiber for neural interfacing, but it does not propose a nanocapillary fiber in the strict fluidic sense (Rodriques et al., 2015). These distinctions matter because NCF research is defined not only by nanoscale waveguiding but by the presence of an internal capillary region that can host emitters or liquid loading.

Within that stricter scope, the NCF literature presented here defines a clear trajectory. The bare water-filled NCF numerically reaches $d^{in}=100\ \text{nm}$38 guided-mode channeling without a cavity (Elaganuru et al., 2024). The symmetric composite photonic crystal cavity on an NCF reaches $d^{in}=100\ \text{nm}$39 when the emitter is placed at the anti-node (Gadde et al., 8 Jul 2025). The asymmetric one-sided cavity reaches about $d^{in}=100\ \text{nm}$40 into one propagation direction with $d^{in}=100\ \text{nm}$41, $d^{in}=100\ \text{nm}$42, and one-pass loss $d^{in}=100\ \text{nm}$43 (Gadde et al., 7 Aug 2025). This progression suggests that the defining NCF advantage is the combination of internal-emitter placement, subwavelength confinement, and external cavity engineering on a fiber-compatible platform.