WO2017032454A1

WO2017032454A1 - Hollow-core fibre and method of manufacturing thereof

Info

Publication number: WO2017032454A1
Application number: PCT/EP2016/001424
Authority: WO
Inventors: Philip Russell; Patrick Uebel; Michael Henoch Frosz
Original assignee: MAX-PLANCK-Gesellschaft zur Förderung der Wissenschaften e.V.
Priority date: 2015-08-26
Filing date: 2016-08-24
Publication date: 2017-03-02
Also published as: US20220146907A1; TWI699570B; IL290524A; CN112305663A; CA2995736A1; US11733451B2; CN108351465B; EP3136143A1; JP2018533042A; CN112305663B; CN108351465A; US20190377131A1; CA2995736C; US20180267235A1; US11269135B2; EP3136143B1; US10393956B2; TW201712381A; DK3136143T3; EP3341771A1

Abstract

A hollow-core fibre (100) of non-bandgap type comprises a hollow core region (10) axially extending along the hollow-core fibre (100) and having a smallest transverse core dimension (D), wherein the core region (10) is adapted for guiding a transverse fundamental core mode and transverse higher order core modes, and an inner cladding region (20) comprising an arrangement of anti-resonant elements (AREs) (21, 21A, 21B) surrounding the core region (10) along the hollow-core fibre (100), each having a smallest transverse ARE dimension (d_i) and being adapted for guiding transverse ARE modes, wherein the core region (10) and the AREs (21, 21A, 21B) are configured to provide phase matching of the higher order core modes and the ARE modes and the ARE dimension (d_i) and the core dimension (D) are selected such that a ratio of the ARE and core dimensions (d_i/D) is approximated to a quotient of zeros of Bessel functions of first kind (u_lm,_ARE/u_lm,core), multiplied with a fitting factor in a range of 0.9 to 1.5, with m being the m-th zero of the Bessel function of first kind of order 1, said zeros of the Bessel functions describing the LP_lm ARE modes and LP_lm higher order core modes, respectively. Furthermore, an optical device (200) including the hollow-core fibre (100) and a method of manufacturing the hollow-core fibre are described.

Description

Hollow-core fibre and method of manufacturing thereof The invention relates to a hollow-core fibre (HCF, or: hollow-core photonic crystal fibre, HC-PCF) of non-bandgap type (or: hollow-core anti-resonant-reflecting fibre, HC-AF), in particular having an axial hollow core region and an inner cladding region comprising an arrangement of anti-resonant elements (AREs) surrounding the core region. Furthermore, the invention relates to an optical device including at least one HC-AF and to a method of manufacturing a HCF of non-bandgap type. Applications of the invention are available in the fields of data transmission, in particular low-latency data transmission, high-power beam delivery systems, in particular for material processing, modal filtering, gas-based nonlinear optics, in particular supercontinuum generation from the ultraviolet to the infrared or generation of ultrashort pulses, fibre gyroscopes or chemical sensing.

In the present specification, reference is made to the following prior art illustrating the technical background of light-guiding fibres, in particular HCF' s of bandgap or non- bandgap type.

[1] T.A. Birks et al. in "Optics Letters" 1997. 22(13): p. 961-963;

[2] N. A. Mortensen et al. in "Optics Letters" 2003.

28 (20) : p. 1879-1881;

[3] US 2015/0104131 Al;

[4] P. J\ Roberts et al. in "Optics Express" 2005. 13(1): p. 236-244;

[5] J. K. Lyngso et al. in "22nd International Conference on Optical Fiber Sensors" Pts 1-3, 2012. 8421; [6] J. M. Fini et al. in "Nature Communications" 2014. 5: article number 5085;

[7] F. Couny et al. in "Optics Letters" 2006. 31(24): p. 3574-3576;

[8] B. Debord et al . in "Optics Letters" 2014. 39(21): p. 6245-6248;

[9] W. Belardi et al. in "Optics Express" 2014. 22(8) : p. 10091-10096;

[10] W. Belardi et al. in "arXiv : 1501.00586v2" 2015: p.

[physics . optics ] ;

[11] P. Jaworski et al. in "Optics Express" 2015. 23(7): p. 8498-8506;

[12] A. Hartung et al. in "Optics Express" 2014. 22(16);

[13] Benabid, A.F., et al., Hollow-core photonic crystal fibre, in US patent 8306379 B2, GLOphotonics SAS;

[14] F. Poletti in "Optics Express" 2014. 22(20): p. 23807- 23828;

[15] T. G. Euser et al. in "Optics Express" 2008. 16(22): p. 17972-17981;

[16] E. A. J. Marcatili et al. in "Bell Systems Technical Journal" 1964. 43: p. 1783-1809;

[17] J. M. Fini et al. in "Optics Express" 2013. 21(5) : p. 6233-6242;

[18] C. Wei et al . in "Optics Express" 2015. 23: p. 15824; and

[19] A. Hartung et al. in "Optics Letters" 2015. 40(14): p. 3432.

Solid-core fibres are generally known and broadly used e.g. in data communication applications. Solid-core fibres can be designed for low-loss, single-mode transmission in a broadband transmission range of the fibre material, like quartz glass. So-called endlessly single-mode guidance (ESM, i.e. all higher order modes, HOMs, are leaky while the fundamental LPoi mode is fully confined) is achieved in solid-core pho^¬ tonic crystal fibres (PCFs) by engineering a cladding structure surrounding the solid-core, as shown in Figure 7A (prior art) , such that the diameter d of channels in the cladding structure and their centre-centre spacing (pitch) A fulfils the geometrical condition d/Λ < 0.41 ([1, 2]) . However, due to light guiding in the solid fibre material, disadvantages exist in terms of increased latency in data transmission, optically non-linear effects resulting in new light frequen- cies, and relatively low damage thresholds.

Through their ability to guide light in a non-solid core region, which is evacuated (vacuum core) , filled with a gas or filled with a liquid, hollow-core photonic crystal fibres, HC-PCFs, have unique advantages compared to solid-core fibres, resulting in application areas such as low-latency data transmission, high-power beam delivery, gas-based nonlinear optics, light guiding with ultralow nonlinearities and chemical sensing. HC-PCFs are typically divided into two classes depending on the physical guidance mechanism: hollow-core photonic bandgap fibres (HC-PBFs) and hollow-core anti- resonant-reflecting fibres (HC-AFs) .

Figures 7B to 71 (prior art) shows a selection of scanning electron micrographs of different types of conventional HC- PCFs. Figures 7B and 7C show HC-PBFs that confine modes inside a central hollow core by means of a photonic bandgap in the cladding [4 - 6] . These types of PCF typically have relatively low loss (ca. < 20 dB/km) at telecommunication wave- lengths. However, due to the wavelength-specific effect of the photonic bandgap, they guide light over a relatively narrow bandwidth (ca. < 15 THz) only. Although in general HC- PBFs support HOMs, Fini et al . ([6], [3]) have shown that bended HC-PBFs can be made effectively single-mode by includ- ing "satellite" hollow cores in the cladding (Fig. 7B) . These satellites strongly suppress HOMs in the core by phase- matching to them, causing high HOM loss. HC-PBGs can also be made truly single-mode over a narrow spectral range (ca. < 7 THz) if a small enough core is used (see Fig. 7C) [5], but this results in fabrication difficulties and significantly higher loss for the desired fundamental mode.

Figures 7D to 71 show a selection of HC-AF structures, i. e. fibres having guidance mechanism based mostly on anti- resonant effects. Figures 7D and 7E have a Kagome-lattice cladding [7, 8, 13] and Figures 7F and 7G have one ring of single (F) or nested (G) anti-resonant elements (AREs). Figure 7H shows a HC-AF with a square core [12] and Fig. 71 de- picts a HC-AF with guiding properties in the ultraviolet

[19] . Compared to HC-PBFs, the loss of HC-AFs is in general larger because of the non-ideal confinement, but the transmission window is broader. F. Poletti showed numerically that by engineering the radial distance between nested AREs, a HC-AF can be made effectively single-mode over a limited wavelength range [14], but this has not been demonstrated experimentally. In particular, mode suppression has been suggested in [14] for a certain centre wavelength and a certain structural parameter z/R only, wherein z is a diameter difference of AREs and nested structures in the AREs and R is the radius of the hollow core. This theoretical finding cannot be extended to a general design and manufacturing of HC-AFs.

Also Wei et al. ([18]) and A. Hartung et al. ([19]) showed that HOM suppression can be enhanced. This was achieved in HC-AFs with touching AREs, but the overall HOM suppression was relatively low. According to [18], the effect of varying the thickness t of the touching AREs on the effective refractive index of the hollow core and the AREs has been investi^¬ gated. An optimum thickness has been found wherein HOMs are suppressed. As a disadvantage of this approach, HOM suppres- sion has been shown for a certain wavelength only, but not for a broad wavelength range.

Compared to solid-core fibres, one particular drawback of conventional HC-AFs is their inability to be purely single- mode, i.e. higher order modes (HOMs) are supported for relatively long distances. As a consequence the output beam quality is degraded, which is undesirable in many applications as it introduces modal beating, blurring the focal spot and - if the fibre experiences change in stress - causes power fluctu- ations. Another disadvantage of conventional HC-AFs results from limitation in manufacturing thereof, in particular positioning the AREs at specific azimuthal locations in a sufficient stable and reproducible manner. The objective of the invention is to provide an improved hollow-core fibre of non-bandgap type, which is capable of avoiding disadvantages of conventional HC-AFs. In particular, the HC-AF is to be provided with an extended wavelength range of single-mode transmission, increased loss of HOMs, and/or increased ratio in loss between the highest-index core HOM and the fundamental, e. g. LPoi mode. Furthermore, the objective of the invention is to provide an improved optical device being equipped with at least one HC-AF and avoiding limitations of conventional optical devices, in particular in terms of low loss single mode guidance, damage-free high power light delivery and targeted creation of optically nonlinear effects. Furthermore, the objective of the invention is to provide an improved method of manufacturing a hollow- core fibre of non-bandgap type resulting in the improved in- ventive HC-AF and avoiding limitations of conventional manufacturing methods.

These objectives are solved with a hollow-core fibre of non- bandgap type, an optical device and a method of manufacturing a hollow-core fibre of non-bandgap type, comprising the fea^¬ tures of the independent claims, resp.. Advantageous embodi^¬ ments and applications of the invention are defined in the dependent claims .

According to a first general aspect of the invention, the above objective is solved by a hollow-core fibre of non- bandgap type, comprising a hollow core region axially arranged along the longitudinal extension of the hollow-core fibre and an inner cladding region comprising an arrangement of multiple anti-resonant elements, AREs, surrounding the core region along the length of the hollow-core fibre. The core region is adapted for guiding a transverse fundamental core mode and- transverse higher order core modes of a light field coupled into the HC-AF. The core region has a smallest transverse core dimension (D), which is the smallest distance between AREs on diametrically opposite sides of the core region. Preferably, the smallest transverse core dimension is constant along the length of the HC-AF. Each of the AREs is adapted for guiding transverse ARE modes, and the i-th ARE has a smallest transverse ARE dimension (di) . Preferably, the smallest transverse ARE dimension also is constant along the length of the HC-AF. The core region and -the AREs are configured to provide phase matching of the higher-order core modes of the core region to the ARE modes of the AREs. In other words, the higher order core modes and the ARE modes have refractive indices which are equal or approximated to each other so that the higher-order core modes can resonantly couple to the ARE modes, i.e. the ARE modes can be excited by the higher-order core modes.

The hollow core region of the inventive HC-AF is formed by an inner space of the hollow-core fibre, which is empty (evacu^¬ ated) or filled with a gas, in particular at least one of air, noble gas and hydrogen, a liquid, and/or a material hav^¬ ing a non-linear optical response, like e.g. at least one of the above gases or a Rydberg gas. Accordingly, the term "hol- low core region" covers any longitudinal waveguide structure which is free of solid materials in its inner space. The core region has a first refractive index determined by the materi^¬ al thereof, while the refractive (or effective) index of the core mode is determined by the material refractive index, the shape of the core region and the excitation wavelength.

In radial directions, the core region is surrounded by the AREs, which are arranged in a non-touching manner. In other words, the inner cladding region comprises an arrangement of single AREs without an azimuthal contact thereof. The term

"anti-resonant element", ARE (or: "tube" or "cladding capillary"), refers to any hollow wave-guiding element having a smaller diameter than the hollow core region and extending along the length of the HC-AF. Preferably, the wall thickness of the AREs of the inner cladding region is smaller than 20 % of the core dimension D, e. g. < 5 μπι. All AREs can have the same smallest inner dimension, or each ARE can have a different smallest inner dimension. The inner cladding region surrounding the core region has a refractive index (a second re- fractive index) higher than the first refractive index of the core region. The hollow AREs are empty (evacuated) or filled like gas or liquid, e. g. like the core region. An even number of AREs, e. g. 4 or 6, or an odd number of AREs, e. g. 3 or 5 or 7,. is provided, which can be arranged with an even or odd-numbered symmetry delimiting the core region. Preferably, the AREs are non-nested AREs, and/or the AREs are made of a glass, in particular silica, a plastic ma^¬ terial, in particular polymer, composite, in particular ZBLAN fiber composite, metal or crystalline material. Preferably, for holding the AREs, an outer cladding region is provided, wherein the AREs are attached to an inner surface of the out- er cladding region.

According to the invention, each of the ARE dimensions (di) and the core dimension (D) are selected such that a ratio of the ARE and core dimensions (di/D) is approximated to a quo- tient of zeros of Bessel functions of first kind

( Ulm, ARE/ Ulm, core ) , multiplied with a fitting factor in a range of 0.9 to 1.5, preferably 0.96 to 1.20, with m being the m-th zero of the Bessel function of first kind of order 1, said zeros of the Bessel functions describing (modelling) the LPi_m ARE modes and LPim higher order core modes, respectively. Advantageously, the inventors have found a modal filtering effect depending on only one dimensionless geometrical parameter (di/D) , akin to the well-known d/Λ parameter for endlessly single-mode solid-core PCF.

The inventors have found that modal refractive indices of the LPim ARE modes and LPi_m higher order core modes can be represented by an analytical model based on modes of coupled capillary waveguides, said modes being approximated by a Mar- catili-Schmelt zer expression, which depends on the zeros of the Bessel function providing the modes and the fitting factor. Phase-matching the LPi_m ARE modes and LPi_m higher order core modes, i. e. matching the modal indices thereof can be obtained if the ratio of the ARE and core dimensions (di/D) is adapted to the quotient of zeros of Bessel functions of first kind (ui_m,ARE/ui_m,core) , multiplied with the fitting fac^¬ tor . Preferably, the fitting factor is a model parameter, which is obtained from matching the analytical model to a vectorial finite element (FE) modelling of the LPi_m ARE modes and LPim higher order core modes of the HC-AF. In particular, the fitting factor results from a division from a first factor matching the analytical model to the FE model of the LPi_m higher order core modes and a second factor matching the analytical model to the FE model of the LPi_m ARE modes, as further described below with reference to equation (4) . The terms "approximating" (or "adapting") include an equality of the ratio of the ARE and core dimensions and the quotient of zeros of Bessel functions of first kind, multiplied with the fitting factor, or a difference between the ratio of the ARE and core dimensions and the quotient of zeros of Bessel functions of first kind, multiplied with the fitting factor, being minimized such that a ratio in loss (dB/m) between the highest-index core HOM and the LPoi core mode is > 5, in particular > 10 or even > 25, preferably > 50. Advantageously, with the above condition for selecting the ratio of the ARE and core dimensions, the inventors have found a new design parameter for the HC-AFs allowing a broadband extension of the HOM coupling to ARE modes, e.g. up to all wavelengths within the transparency^ window of its funda- mental core mode. According to the invention, refractive indices of the core region and the AREs are equal or approximated to each other in a broad wavelength range, preferably covering at least 10 THz, particularly preferred covering at least 20 THz. All modes except the fundamental are strongly suppressed, resulting in effectively single-mode behaviour of the HC-AF (effectively endlessly single-mode behaviour, eESM) . Contrary to (effectively) single-mode HC-PBFs [17], the proposed HC-AF structure provides a much larger bandwidth at relatively low loss while providing eESM behaviour. Furthermore, contrary to the HC-AF described in [18], wherein the thickness of the glass capillaries in the cladding has been optimized only, the invention provides the new design parameter for the ratio of the ARE and core dimensions. This new design provides real phase-matching, whereas in [18] phase-matching was not achieved (see Figure 2a of [18]). Contrary to the SM guidance for HC-PBFs [6] where shunts have been included in the bandgap cladding to enhance HOM suppression for a relatively narrow wavelength range, the inventive HC-AF supports HOM suppression for an extended frequency range.

The ratio of the ARE and core dimensions is selected on the basis of a model which has been found by the inventors for describing tubular AREs (ARE capillaries) with a circular transverse cross-section as outlined below. In this case, the smallest transverse ARE dimension (di) is the inner diameter of the ARE. As a further advantage of the invention, the se- lected design parameter is valid not only for tubular AREs, but also for AREs having other cross-sections, like elliptical or even polygonal cross-sections. With the latter variants, the smallest transverse ARE dimension (d±) is the smallest inner cross-sectional dimension of the ARE. The in- ventors have found that the model based on tubular AREs still provides an excellent approximation for HOM suppression in non-circular AREs, if they can be approximately modelled like a tubular ARE. According to a preferred embodiment of the invention, the AREs have a first smallest transverse ARE dimension (di) and the ratio of the first smallest transverse ARE dimension and the core dimension (di/D) is approximated to a quotient of zeros of Bessel functions of first kind (uoi,ARE/un_Core) , multi^¬ plied with the fitting factor, said zeros (UO I , ARE) , (un,_Core) describing the LPoi ARE modes and the LPn core mode, respec^¬ tively. Accordingly, the most pronounced HOM of the core re^¬ gion, the LPn core mode, is coupled to the ARE modes. With this embodiment, preferably all AREs have the same smallest transverse ARE dimension. As an example, all AREs are tubular AREs having the same inner diameter. According to particularly preferred variants of this embodiment, the ratio of the first ARE dimension and the core dimension (di/D) is selected in a range from 0.5 to 0.8, preferably from 0.60 to 0.75, in particular in a range from 0.62 to 0.74. Advantageously, a HC-AF is provided with these ranges, having low-loss guidance for the fundamental mode at all wavelengths within its transmission window, strongly suppressing the LPn core mode and also all higher-order modes.

According to a further advantageous embodiment of the invention, the AREs comprise a first group of AREs with the first ARE dimension (di) and a second group of AREs with a second smallest transverse ARE dimension (d2) smaller than the first ARE dimension (di) of the first group of AREs. With this embodiment, the ratio of the second ARE dimension and the core dimension (d2/D) is approximated to a quotient of zeros of Bessel functions of first kind (uoi, ARE/u2i,core) , multiplied with the fitting factor, said Bessel functions (U O I , ARE ) ,

( u2i,core) describing the LPoi ARE modes and the LP21 core mode, respectively. AREs having the first ARE dimension and AREs having the second ARE dimension can be arranged in an alter- nating fashion, surrounding the hollow core region of the HC- AF.

According to particularly preferred variants of this embodi- ment, the ratio of the second ARE dimension and the core di^¬ mension (d₂/D) is selected in a range from 0.3 to 0.7, preferably from 0.43 to 0.59, in particular in a range from 0.45 to 0.54. Advantageously, a HC-AF is provided with these rang^¬ es, having a further improved suppression of the HO s .

Advantageously, multiple variants are available for designing the arrangement of AREs, which provide the inner cladding re^¬ gion of the inventive HC-AF. These variants can be selected as alternatives or in combination in dependency on the par- ticular application of the invention. Firstly, the arrangement of AREs can have a three-fold symmetry. Alternatively, the arrangement of AREs can have two-fold symmetry and causes optical birefringence. Furthermore, the AREs can be arranged such that the cross-sections thereof are distributed on a single ring surrounding the core region. Alternatively, the

AREs can be arranged such that the cross-sections thereof are distributed on multiple, e. g. two, three or more, coaxial rings surrounding the core region. Advantageously, lower loss HC-AFs with improved HOMS suppression can be realized with these variants of the invention.

The outer cladding region for holding the AREs has the shape of a hollow sleeve extending along the HC-AF and having an inner surface for attaching the AREs and an outer surface which can be exposed as the surface of the HC-AF or covered with further layers, e. g. protective layers, like a polymer coating, or an opaque layer. According to a variant of the invention, the outer cladding region has an inner transverse cross-section with a regular polygonal shape, e.g. a hexago- nal shape, and the AREs are fixed to corners of the polygonal shape. As an advantage, each ARE is fixed along two contact lines with the inner surface of the outer cladding region, thus increasing the mechanical stability of the inner clad- ding region (AREs) when preparing the preform. Furthermore, the AREs are fixed with a regular azimuthal spacing by arranging them in the corners of the polygonal shape. According to an alternative variant of the invention, the outer cladding region has an inner transverse cross-section with a curved, in particular circular shape, and the AREs are evenly distributed in the curved shape.

According to a second general aspect of the invention, the above objective is solved by an optical device, including at least one hollow-core fibre according to the above first general aspect of the invention. Preferably, the optical device comprises at least one of a modal filtering device, a light source, in particular a laser, an optical amplifier, a beam delivery system, a data communication system, a frequency converter, in particular for supercontinuum generation and a pulse shaper, in particular for pulse compression.

According to a third general aspect of the invention, the above objective is solved by a method of manufacturing a hol- low-core fibre according to the above first general aspect of the invention. Preferably, the method of manufacturing a hollow-core fibre comprises the steps of providing ARE preforms and a hollow jacket preform, fixing the ARE preforms on an inner surface of the jacket preform in a distributed manner, and heating and drawing the jacket preform including the ARE preforms until the final ARE and core dimensions are set.

Optionally, the jacket preform including the ARE preforms are - in a first heating and drawing step - first drawn to a cane and - in a second heating and drawing step - then drawn to a fibre until the final ARE and core dimensions are set.

Preferably, the heating and drawing steps include applying a vacuum or an increased fluid pressure to at least one of the jacket preform and the ARE preforms for setting the ARE and core dimensions, respectively. Advantageously, this allows a precise adjustment of the smallest transverse core dimension (D) of the core region and the smallest transverse ARE dimen- sion(s) (di) of the AREs .

According to a preferred embodiment of the invention, the ARE and core dimensions are selected, preferably set during the final heating and drawing step, such that the phase matching of the higher order core modes and the ARE modes is obtained over a frequency range above 20 THz. This can be achieved by theoretical considerations or reference measurements.

According to further advantageous embodiments of the inven- tion, a post-processing step can be provided, wherein at least one of the core region and the anti-resonant elements can be filled with at least one of a gas, in particular air, noble gas and/or hydrogen, a liquid, and/or a material having a non-linear optical response. To this end, at least one por- tion of the hollow core fiber can be enclosed into a cell. The cell and all of the hollow core regions can be filled with a material from an external reservoir, or some of the hollow core regions can be filled with a material from a first external reservoir and some of another material of a second external reservoir. This post-processing step can be implemented just before or during the application of the fiber, e., g. in an optical experiment. Further post-processing steps can follow for adapting the inventive fibre to the ap- plication thereof, e. g. in data transmission or non-linear optics.

Further advantages and details of the invention are described in the following reference to the attached drawings, which show in:

Figure 1: cross-sections of HC-AFs according to preferred embodiments of the invention;

Figure 2: graphical illustrations of coupling HO s of the core region with ARE modes and numerical simulations when varying d/D to find the optimum ratio between ARE dimension d and core dimension D;

Figure 3: graphical illustrations of numerical simulations to illustrate the scalability of the inventive HC- AF;

Figure 4: illustrations of applying the inventive model for designing a HC-AF;

Figure 5: an optical device according to an exemplary embodiment of the invention;

Figure 6: steps of a method of manufacturing an inventive

HC-AF; and

Figure 7: cross-sections of conventional solid or hollow

core fibres (prior art) .

In the following, exemplary reference is made to choosing a proper geometry of the HC-AF, in particular the diameter of the core and AREs, like glass capillaries in the inner clad- ding region. Implementing the invention is not restricted to the indicated examples of geometric quantities, like the dimensions D, d and t, but rather possible with varied values providing the inventive design parameters. Features of HC- AFs, the manufacturing thereof and the application thereof are not described, as far as they are known from prior art.

Practical examples of HC-AFs Figures 1A to IE show transverse cross-sections of practical examples of inventive HC-AFs 100 (perpendicular to the axial, extension thereof) . The bright circles represent the solid material of the AREs or outer cladding region, like quartz glass or silica, while the dark portions are free of solid materials (evacuated or filled with gas or liquid) . The geometric design of the HC-AFs 100 is selected as outlined in the model section below.

Each HC-AF 100 comprises a hollow core region 10 (represented in Figure 1A by a dotted circle) , an inner cladding region 20 with multiple AREs 21 and an outer cladding region 30. The hollow core region 10 is the empty space, between the AREs 21, extending along the longitudinal length of the HC-AF 100 and having a smallest transverse core dimension D. The AREs 21 of the inner cladding region 20 comprise capillaries having a wall thickness t and a smallest transverse ARE dimension d. The AREs 21 are fixed to the inner surface of the outer cladding region 30, e.g. as described below with reference to Figure 6. The outer cladding region 30 comprises a larger ca- pillary being made of e.g. glass and providing a closed cladding of the HC-AF 100.

HC-AF 100 of Figure 1A illustrates an embodiment wherein the AREs 21 comprise a single-ring of six thin-wall capillaries with a circular transverse cross-section (inner diameter d = 13.6 μιτι and wall thickness t = 0.2 μιτι) arranged within the larger capillary of the outer cladding region 30 in six-fold symmetric pattern so as to create a central hollow core of diameter D (the shortest distance between diametrically oppo^¬ site AREs 21) , with D = 20 μιη. The outer cladding region 30 has an outer diameter of 125 pm and a cladding thickness of 38 pm. Alternatively, the core dimension D can be selected in a range from 10 μπι to 1000 μιη, wherein the other geometrical parameters (like d, t) are scaled accordingly.

Figure IB shows a modified embodiment with multiple, in particular two coaxial rings of AREs 21 (d = 13.6 μπι, t = 0.2 μηα, and D = 20 μκι arranged within the outer cladding region 30 with six-fold symmetry. For holding the inner and outer rings of AREs 21, a support tube 22 is included in the HC-AF 100. The support tube 22 is made of e.g. silica with a diameter of e.g. 48 μπι. Figures 1C and ID illustrate a three-fold symmetry and Figure IE shows a two-fold symmetry of the inventive HC-AF 100, advantageously offering enhanced eESM effects.

According to Figure 1C, the AREs 21A, 21B comprise a first group of AREs 21A with a first, larger ARE dimension di (e. g. 13.6 μπι) and a second group of AREs 21B with a second, smaller ARE dimension d∑ (e. g. 10.2 μπι) , both with a wall thickness of e.g. 0.2 μιη. On the inner surface of the outer cladding 30, longitudinal protrusions 31 are provided, which have an axial extension along the HC-AF 100. In the cross- sectional view of Figure 1C, the protrusions 31 are shown as blobs. Preferably, the protrusions 31 are made of the same material like the AREs, e. g. glass. Each of the smaller AREs 21B is fixed to the top of one of the protrusions 31. The AREs 21A, 21B surround the core region 10 having a diameter D = 20 μπ. The radial height of the protrusions 31 preferably is selected so as to yield the most circularly-symmetric guided mode in the central core region 10. Alternatively, the protrusions 31 could be omitted, yielding a simpler struc^¬ ture, as shown in Figure ID.

By arranging the AREs 21A, 21B so as to form a two-fold symmetric structure (Figure IE), a birefringent polarization- maintaining HC-AF is obtained, displaying eESM behaviour.

The examples of inventive HC-AFs 100 as shown in Figure 1 can be modified, in particular with regard to the shape of the AREs 21, 21A, 21B, which can have e. g. an elliptic or polyg- onal cross-section; the inner shape of the outer cladding 30, which can have e. g. a polygonal cross-section (see Figure 6) ; the solid materials of the AREs 21, 21A, 21B, which may comprise e. g. plastic material, like P MA, glass, like silica, or soft-glass, like ZBLAN; the dimensions of the AREs 21, 21A, 21B; the number of rings of AREs, e. g. three or more; the number of AREs, e. g. 4 or 5 or 7 or more; and the symmetry of the ARE arrangement.

Model for HC-AF design

The ARE dimension (di) and the core dimension (D) of the inventive HC-AFs 100 are selected such that a ratio of the ARE and core dimensions (di/D) is approximated to a quotient of zeros of Bessel functions of first kind ( ui_m,ARE/ui_m, core) , mul- tiplied with the fitting factor, as defined above. If all

AREs have the same ARE dimension (di) , the ratio of the ARE dimension and the core dimension (di/D) preferably is approximated to a quotient of zeros of Bessel functions of first kind (uoi,ARE/ uncore ) , multiplied with the fitting factor, wherein the zeros ( UO I , A E ) , (un, _Core ) describe the LPoi ARE modes and the LPn core mode, respectively. If further AREs have a second, smaller ARE dimension, the ratio of the second ARE dimension and the core dimension (d₂/D) preferably is ap- proximated to a quotient of zeros of Bessel functions of first kind ( uoi, ARE/u2i, core ) , multiplied with the fitting fac^¬ tor, wherein the Bessel functions (UO I , ARE) , (u₂i, core ) describe the LPoi ARE modes and the LP₂i core mode, respectively. These design conditions are found on the basis of the theoretical considerations and numerical simulations illustrated in the following with reference to Figures 2 to 4. These theoretical considerations and numerical simulations can be correspondingly extended to the coupling of core modes higher than the LP21 core mode to the ARE modes.

The central core region 10 of the HC-AF 100 supports several transverse core modes each with a characteristic modal refractive index and leakage loss. The inventive structure is provided in such a way that the LPoi mode (with the highest effective index) has a loss that is much lower than any of the core HOMs . This is achieved by designing the AREs 21, 21A, 21B and the gaps between them so that they support a band of leaky modes (or states) that phase-match to HOMs in the core region 10, making them highly leaky. This strong loss discrimination can be made broadband enough for obtaining eESM behaviour.

Figure 2A shows the HC-AF 100 of Figure 1A (left) illustrat- ing the fundamental LPoi core mode (centre) and the LPn core mode (right) as well as the leaky ARE modes (right) , to which the LP11 core mode and higher core modes are resonantly coupled. Figure 2B shows curves of the finite-element (FE) modelling computed effective refractive index in dependency on the parameter d/D, and Figure 2C shows curves of leakage loss for changing ARE dimension d but constant core dimension D. The central curve in Figure 2C shows the corresponding HOM suppression. The dashed lines show the fully vectorial com- puted values for the LPoi mode of a free standing tube with diameter d and thickness t. Exemplary structure parameters are t/D = 0.01 and D/λ = 20, n_gi_ass = 1.45 and n_COre = 1.

The structure shown in Figure 2A supports several leaky transverse core modes formed by anti-resonant reflection at the walls of the AREs 21 with negative curvature. The AREs 21 support a cladding photonic bandstructure provided by the inner cladding region 20, and they can be approximated by quasi free-standing, thin-walled' tubes. The thick surrounding glass wall of the outer cladding region 30 does not appreciably influence the modal properties, but is provided to physically support the AREs 21.

Figure 2B shows the effective index distribution of the two highest-index core modes LPoi and LPn when varying the ARE diameter (for fixed D) , effectively changing the cladding photonic bandstructure. The index of the LPoi core mode is high enough to avoid resonant coupling to the fundamental ARE modes, and remains almost independent of d/D, whereas the LPii core mode undergoes a strong anti-crossing with the fundamental ARE mode at d/D * 0.68. As one moves away from this anti-crossing, the even and odd eigenmodes evolve asymptotically into uncoupled LPn core and fundamental ARE modes.

Core modes of even higher order (not shown in Figure 2B) have lower indices and couple to highly leaky modes of the ARE ring, some of which are concentrated in the gaps between the AREs 21. Figure 2C plots the calculated leakage loss of the LPoi mode and the two hybrid LPn/AREoi modes. Over the range shown the LPoi core mode has a relatively constant loss with a minimum value of 0.17 dB/m at d/D * 0.65. For smaller ARE 21 diame- ters the loss increases, closely matching the value for an isolated thick-walled dielectric capillary in the limit d/D → 0 (not shown in Figure 2C) . This limit was used to cross-check the FE calculations with analytical results [16], in particular the accuracy of the perfectly matched layers (PMLs). At the anti-crossing point the loss of the two hybrid LPn/AREoi modes strongly increases, almost reaching the value for an isolated capillary in vacuum (dashed brown line) , which was calculated by solving Maxwell' s equations in full vectorial form. This provides further confirmation that the PMLs were set up correctly.

The HOM suppression increases strongly at the anti-crossing, peaking at a value of about 1200. Far from the anti-crossing it drops to less than 5, which is similar to values typically achieved in kagome-PCF [15]. For a comprehensive analysis, the HOM suppression of all the higher-order core modes must be calculated. FE modelling reveals that the HOM with the next-lowest loss after the LPn core mode is the four-lobed LP21 core mode, with a HOM suppression of ~ 70 at d/D « 0.68 and an anti-crossing with the fundamental ARE mode at

d/D « 0.51. In experiments, however, this particular core mode is less likely to be excited by end-fire illumination or by stress- and bend-induced scattering from the LPoi core mode (the index difference is some two times larger than for the LPn core mode) . FE modeling shows that LPi_m core modes of even higher order do not affect the overall HOM suppression because they phase-match to modes of the ARE ring (some of which are concentrated in the gaps between the AREs 21), resulting in strong leakage loss. Figure 3A plots, versus D/λ (λ: wavelength) , the difference Anim between the refractive indices of the LPi_m core modes and the fundamental ARE mode at constant d/D = 0.68 and t/D = 0.01. Δηοι decreases with increasing D/λ but overall remains positive. As a consequence, the LPoi core mode is anti- resonant with the fundamental ARE mode and remains confined to the core (see also left panel of the inset in Figure 3A) . In contrast, Ann is much smaller, reaching values as small as 10^~6 at D/λ * 66.

At certain values of D/λ, anti-crossings appear between the LPoi mode and the q-th order transverse mode in the glass walls of the AREs 21, foll wing the simple relationship:

The vertical dotted lines in Figure 3 are centred at the first two of these resonances, for t/D = O.'Ol. In the vicinity of these points the LPoi core mode leaks rapidly through the resonant AREs 21 into the solid glass jacket 30, yielding loss values that are close to those of an isolated thick- walled dielectric capillary [16]; the result is a strong reduction in the HOM suppression (see Figure 3B) . Away from these narrow regions, however, the HOM suppression remains relatively high - a consequence of the fact that the indices of the LPii core and fundamental ARE modes remain close to each other. The result is very strong LPn core mode suppression over all the ranges of LPoi mode transmission.

To explain why maximum HOM suppression occurs at d/D = 0.68 for all wavelengths (except in the vicinity of ARE wall reso nances, see Eq. 1) , the inventors have applied an analytical model in which the real structure with the core 10 and the AREs 21 are treated as thick-walled capillaries (see Figure 4A) . The modal indices of the LPi_m modes in a thick-walled capillary can be approximated by the modified Marcatili- Schmeltzer expression [16]:

where ui_m is the m-th zero of the Bessel function Ji and d± is the inner diameter of the capillary. The parameter f_s (which has a value close to 1, s = co represents the core 10 and s = ARE the AREs 21) is used to heuristically fit the analytical values from the model equation to the results of FE simula^¬ tions. It corrects for the non-circular core and the finite wall thicknesses of core 10 and AREs 21.

Figure 4B plots the effective index of the two LPn/AREoi hybrid modes together with the fitted values for the LPn mode (zero line) computed using Eq. (2) with fit parameters f_Co = 1.077 for the core 10 and fARE = 0.990 for the ARE 21. The convenient analytical form of Eq. (2) allows one to derive a simple expression for the d/D value at which the LPn core and AREoi modes couple o timally:

Eq. (3) provides a convenient rule-of-thumb for designing robustly single-mode eESM PCFs. To a first approximation it depends neither on the refractive indices nor on the absolute physical dimensions of the fibre, making the design scalable. This means that, provided the ratio d/D is maintained, it becomes possible to design large-core eESM PCFs and to deliver losses of some dB/km in multiple transmission windows, the broadest of which spans more than one octave.

By using Eq. (2) one can also easily find structural parameters where higher order core modes (e.g. the LP21 core mode) are effectively suppressed. Also by adjusting the physical dimensions, the resonance bands can be blue/red shifted (for smaller/thicker wall thickness t) and the minimum transmission loss of the LPoi core mode can be adjusted (for changing core diameter) .

Eq. (2) can be also used to find appropriate geometrical pa^¬ rameters for designing a HC-AF with an enhanced eESM effect, i.e., a fibre where _. the first two HOMs of the core couple to resonances in the AREs. This yields the conditions: O.51 (4)

A fibre structure with such an enhanced eESM effect is depicted in Figures 1C to IE consisting of one or several of single rings of AREs with different inner diameters d\ and d₂ .

Embodiments of optical device

The inventive HC-AF 100 has multiple applications for light guiding, e.g. for beam delivery, data transmission or fre- quency conversion purposes. Accordingly, an optical device, which represents a further subject of the invention, comprises at least one inventive HC-AF 100 and further optical components, monitoring components, detector components and/or control components, which are selected in dependency on the particular application of the optical device.

Figure 5 schematically illustrates an optical device 200 according to an embodiment of the invention, which is adapted for a high power beam delivery, e.g. for surface processing purposes. The optical device 200 comprises a light source 210, like a laser source, and the HC-AF 100. The output of the light source 210 is optically coupled with an input side of the HC-AF 100, while an output side thereof is directed to a location of beam delivery (see arrow) .

With alternative applications of the invention, the light source 210 comprises a laser source for driving a frequency conversion process, in particular a supercontinuum generation process or pulse compression, inside the HC-AF 100. According to yet further applications, the light source 210 may comprise an optical transmitter of a data communication system, which is coupled via the HC-AF 100 with an optical receiver (not shown) .

It is noted that Figure 5 represents a schematic drawing only. Details of an optical device including at least one in- ventive HC-AF 100 can be implemented as it is known from conventional optical devices.

Method of manufacturing HC-AFs Figure 6 schematically illustrates the main steps of manufacturing an inventive HC-AF 100. Figure 6 is a schematic illustration only, which presents the main steps of providing ARE preforms 23 and a hollow jacket preform 32 (Figure 6A) , fixing the ARE preforms 23 on an inner surface of the jacket preform 32 (Figure 6B) and heating and drawing the preform jacket 32 including the ARE preforms 23 for obtaining the HC- AF 100 (Figure 6C) . Optionally, the heating and drawing step may .include a first step of heating and drawing the jacket preform 32 with the ARE preforms 23 to a cane, and a second step of heating and drawing the cane until the ARE and core dimensions are set. Details of the steps in Figure 6A, 6B and 6C can be implemented as far as they are known from conventional fibre manufacturing methods. According to Figure 6A, the jacket preform 32 is a hollow tubular form, made of glass, which has an inner transverse cross-sectional regular hexagonal shape. The outer diameter of the preform jacket 32 is e.g. 28 mm, while the inner transverse dimension is about 22 mm. The longitudinal length of the preform jackets 32 and the ARE preforms 23 is about 120 cm.

With the fixing step of Figure 6B, the ARE preforms 23 are fixed to the corners of the inner hexagonal shape of the jacket preform 32. This is obtained by applying heat resulting in a physical connection between ARE preforms and jacket preform. Subsequently, the composite of the jacket preform 32 and the ARE preforms 23 is drawn during the application of heat until the ARE and core transverse dimensions are obtained. The ARE and core transverse dimensions can be influenced by applying a vacuum or an increased pressure to the jacket preform 32 and/or the ARE preforms 23 during the heating and drawing steps.

Applying a vacuum or an increased pressure during the heating and drawing steps is schematically illustrated in Figure 6B. The hollow inner space of the jacket preform 32 is connected with a first external reservoir 41, like a source of pressur- ized nitrogen. Furthermore, the ARE preforms 23 are connected with at least one second external reservoir 42, like e.g. an' external source of pressurized nitrogen. If all AREs are to be produced with the same inner transverse dimension, all AREs can be connected with a common external reservoir. Oth- erwise, e.g. two groups of AREs are connected to two different external reservoirs for creating different inner transverse dimensions, as shown e.g. in Figure IE. The final heating and drawing step is conducted e.g. in a furnace selected in dependency of the material of the jacket preform 32 and the ARE preforms 23 (e.g. around 2000°C for silica components) .

After obtaining the final HC-AF 100, it can be filled with, a gas, like air or a noble gas or hydrogen, or a liquid, like water, and the input and output sides of the HC-AF 100 are enclosed by a cell withstanding high fluidic pressure and which is partially transmissive, e.g. by including a glass plate, for optical radiation, e.g. from a laser source.

The features of the invention disclosed in the above description, the drawings and the claims can be of significance individually, in combination or sub-combination for the implementation of the invention in its different embodiments.

Claims

Claims 1. Hollow-core fibre (100) of non-bandgap type, compris^¬ ing :

- a hollow core region (10) axially extending along the hollow-core fibre (100) and having a smallest transverse core dimension (D), wherein the core region (10) is adapted for guiding a transverse fundamental core mode and transverse higher order core modes, and

- an inner cladding region (20) comprising an arrangement of anti-resonant elements (AREs) (21, 21A, 21B) surrounding the core region (10) along the hollow-core fibre (100), each hav- ing a smallest transverse ARE dimension (di) and being adapted for guiding transverse ARE modes,

characterized in that

- the core region (10) and the AREs (21, 21A, 21B) are configured to provide phase matching of the higher order core modes and the ARE modes, and

- the ARE dimension (di) and the core dimension (D) are selected such that a ratio of the ARE and core dimensions

(di/D) is approximated to a quotient of zeros of Bessel functions of first kind ( uim, ARE/ Ulm, core ) , multiplied with a fitting factor in a range from 0.9 to 1.5, with m being the m-th zero of the Bessel functions of first kind of order 1, said zeros of the Bessel functions describing the LP_im ARE modes and LPi_m higher order core modes, respectively.

2. Hollow-core fibre according to claim 1, wherein

- the AREs surrounding the core region are arranged in a non- touching manner.

3. Hollow-core fibre according to one of the foregoing claims, wherein

- the AREs (21, 21A) have a first smallest transverse ARE di^¬ mension (di) , and

- the ratio of the first ARE dimension and the core dimension (di/D) is approximated to a quotient of zeros of Bessel functions of first kind (UOI,ARE/ Ullcore ) , multiplied with the fit^¬ ting factor, said zeros (UOI,ARE) , (un,_∞re) describing the LPoi ARE modes and the LPu core mode, respectively.

4. Hollow-core fibre according to claim 3, wherein

- the ratio of the first ARE dimension and the core dimension (di/D) is selected in a range from 0.5 to 0.8.

5. Hollow-core fibre according to claim 4, wherein

- the ratio of the first ARE dimension and the core dimension (di/D) is selected in a range from 0.62 to 0.74.

6. Hollow-core fibre according to one of the claims 3 to 5, wherein

- each of the AREs (21) has the first ARE dimension (di) .

7. Hollow-core fibre according to one of the claims 3 to

5. wherein

- a first group of the AREs (21A) has the first ARE dimension (di) , and

- a second group of the AREs (21B) has a second smallest transverse ARE dimension (d₂) smaller than the first ARE dimension (di) of the first group of the AREs (21A) , and

- the ratio of the second ARE dimension and the core dimension (d₂/D) is approximated to a quotient of zeros of Bessel functions of first kind ( uoi,ARE/U₂I, core ) , multiplied with the fitting factor, said Bessel functions (UOI,ARE) , (u₂i,_COre) de- scribing the LPoi ARE modes and the LP21 core mode, respectively.

8. Hollow-core fibre according to claim 7, wherein

- the ratio of the second ARE dimension and the core dimension (d₂/D) is selected in a range from 0.3 to 0.7.

9. Hollow-core fibre according to claim 7, wherein

- the ratio of the second ARE dimension and the core dimen- sion (d₂/D) is selected in a range from 0.45 to 0.54.

10. Hollow-core fibre according to one of the foregoing claims, wherein

- the number of AREs (21, 21A, 21B) is 3, 4, 5, 6 or 7.

11. Hollow-core fibre according to one of the foregoing claims, wherein the arrangement of AREs (21, 21A, 21B) has at least one of the features:

- the arrangement of AREs (21, 21A, 21B) has three-fold sym- metry,

- the arrangement of AREs (21, 21A, 21B) has two-fold symmetry and causes optical birefringence,

- the AREs (21, 21A, 21B) are arranged such that the cross- sections thereof are distributed on a single ring surrounding the core region (10), and

- the AREs (21, 21A, 21B) are arranged such that the cross- sections thereof are distributed on multiple rings surrounding the core region (10) .

12. Hollow-core fibre according to one of the foregoing claims, wherein the AREs (21, 21A, 21B) have at least one of the features :

- the AREs (21, 21A, 21B) have a circular, elliptic or polygonal transverse cross-section, and - the AREs (21, 21A, 21B) are made of a glass, in particular silica or ZBLAN, polymer, in particular PMMA, composite, met^¬ al or crystalline material.

13. Hollow-core fibre according to one of the foregoing claims, wherein

- at least one of the core region (10) and the AREs (21, 21A, 21B) is evacuated or filled with at least one of a gas, in particular at least one of air or noble gas or hydrogen, a liquid, and a material having a non-linear optical response.

14. Hollow-core fibre according to one of the foregoing claims, comprising

- an outer cladding region (30) surrounding the inner cladding region (20) along the hollow-core fibre (100), wherein

- the outer cladding region (30) has a transverse cross- section with a polygonal shape, and

- the AREs (21, 21A, 21B) are located in corners of the polygonal shape.

15. Hollow-core fibre according to one of the claims 1 to 13, comprising

- an outer cladding region (30) surrounding the inner cladding region (20) along the hollow-core fibre (100) , wherein

- the outer cladding region (30) has a transverse cross- section with a curved, in particular circular shape, and

- the AREs (21, 21A, 21B) are evenly distributed in the curved shape.

16. Hollow-core fibre according to one of the foregoing claims, wherein

- the core region (10) and the AREs (21, 21A, 21B) are configured to provide the phase matching of the higher order core modes and the ARE modes in a broadband wavelength range.

17. Hollow-core fibre according to claim 16, wherein

- the core region (10) and the AREs (21, 21A, 21B) are con^¬ figured to provide the phase matching of the higher order core modes and the ARE modes in a wavelength range covering up to all wavelengths within a hollow-core fibre transparency window of the fundamental core mode.

18. Hollow-core fibre according to claim 16 or 17, wherein - the core region (10) and the AREs (21, 21A, 21B) are configured to provide the phase matching of the higher order core modes and the ARE modes in a wavelength range covering at least 10 THz.

19. Optical device (200), including at least one hollow- core fibre (100) according to one of the foregoing claims.

20. Optical device according to claim 19, comprising at least one of

- a modal filtering device,

- a light source (210), in particular a laser,

- an optical amplifier,

- a beam delivery system,

- a pulse shaper, in particular for pulse compression,

- a data communication system, and

- a frequency converter, in particular for supercontinuum generation .

21. Method of manufacturing a hollow-core fibre (100) of non-bandgap type, comprising the steps of:

- providing a hollow core region (10) axially extending along the hollow-core fibre (100) and having a smallest transverse core dimension (D), wherein the core region (10) is adapted for guiding a transverse fundamental core mode and transverse higher order core modes, and

- providing an inner cladding region (20) comprising an ar^¬ rangement of anti-resonant elements (AREs) (21, 21A, 21B) surrounding the core region (10) along the hollow-core fibre (100), each having a smallest transverse ARE dimension (di) and being adapted for guiding transverse ARE modes, wherein

- the core region (10) and the AREs (21, 21A, 21B) are configured to provide phase matching of the higher order core modes and the ARE modes,

characterized in that

(di/D) is approximated to a quotient of zeros of Bessel func- tions of first kind (ui_m,ARE/ui_m,_Core) , multiplied with a fitting factor in a range from 0.9 to 1.5, with m being the m-th zero of the Bessel functions of first kind of order 1, said zeros of the Bessel functions describing the LPi_m ARE modes and LPi_m higher order core modes, respectively.

22. Method according to claim 21, wherein

- the hollow-core fibre (100) is manufactured with the features of the hollow-core fibre (100) according to one of the claims 1 to 17.

23. Method according to claim 21 or 22, wherein

- the ARE dimension (di) is selected by applying an analytical model in which the core (10) and the AREs (21, 21A, 21B) are treated as capillaries, wherein modal indices of the LPim modes in the capillaries are approximated by

wherein ui_m is the m-th zero of the Bessel function Ji, di is the inner diameter of the capillary, and the parameter f_s is a heuristic fit parameter.

24. Method according to one of the claims 21 to 23, com^¬ prising the steps of

(a) providing ARE preforms (23) and a hollow jacket preform (32) ,

(b) fixing the ARE preforms (23) on an inner surface of the jacket preform (32) in a distributed manner,

(c) heating and drawing the jacket preform (32) including the ARE preforms (23) until the ARE and core dimensions are set.

25. Method according to claim 24, wherein step (c) in- eludes

(cl) heating and drawing the jacket preform (32) including the ARE preforms (23) to a cane, and

(c2) heating and drawing the cane until the ARE and core dimensions are set.

26. Method according to claim 25, wherein step (c) includes

- applying a vacuum or an increased pressure to at least one of the jacket preform (32) and the ARE preforms (23) or hol- low regions of the cane for setting the ARE and core dimensions .

27. Method according to one of claims 25 or 26, comprising a post-processing step of

- filling at least one of the core region (10) and the AREs (21, 21A, 21B) with at least one of a gas, in particular air or noble gas or hydrogen, a liquid, and a material having a non-linear optical response.