Open Access
Add to BibSonomy
Abstract
We demonstrate a novel few-moded ultralarge mode area chalcogenide glass photonic crystal fiber for mid-infrared high power applications. The numerical simulation indicates that the fiber has ultralarge mode areas of ∼10500 µm2 and ∼12000 µm2 for the fundamental mode LP01 and the lowest higher-order mode LP11, respectively. Dual-moded operation is confirmed experimentally at 2 µm, in good agreement with the numerical simulation. By selectively launching technique, low bending loss of 0.7 dB/m, equivalent to 0.55 dB/turn, has been observed in the fiber with a small bending radius of ∼12 cm, indicating excellent bending resistance of the few-moded fiber with such a large mode area. The fiber has been demonstrated to sustain an incident power density up to 150 kW/cm2 under 2-µm CW laser irradiation, showing the potential of the fiber for high-power applications in mid-infrared.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The 2-20 µm mid-infrared (mid-IR) spectral region includes two important windows (3-5 µm and 8-12 µm), where the Earth’s atmosphere is highly transparent over multi-kilometer air paths [1]. Also most chemical and biological molecules undergo characteristic absorptions in the mid-IR domain. Mid-IR photonics [1–6] are therefore developed rapidly for a number of applications such as remote detection, medical treatment, security, metrology, and high-precision IR spectroscopy.
Due to the advantages including the compactness, low-weight, low cost per unit length, and excellent beam quality, silica fiber optical components have been demonstrated successful for high-power fiber laser and signal or power delivery in the near-IR range [7,8]. However, because of the strong fundamental vibration of hydroxyl groups at 2.7 µm and the multi-phonon absorption arising from the Si-O network, silica fibers show inferior transmission performance beyond 2 µm. Non-silica glasses, such as tellurite (based on TeO2), fluoride (based on AlF3, ZrF4, GaF3, or InF3), and chalcogenide (based on chalcogen elements S, Se, and Te) [9–11], show excellent transmission properties in the range of 0.5-5 µm, 0.4-6 µm and 1-16 µm, respectively. Mid-IR optical fibers based on these glasses are therefore important candidates as for mid-infrared applications over the conventional silica glass.
Amongst the mid-IR glass fibers, chalcogenide glass (ChG) fiber is highly transparent within the broadest wavelength domain up to 16 µm. Second, it possesses high n2 of 100-1000 × 10−20 m2/W, which is higher than that of conventional silica glass by 2-3 orders of magnitude [12]. ChG fiber has been considered as an ideal medium (i) for mid-IR signal or power delivery [6,13–15] and (ii) for mid-IR nonlinear frequency conversion [16–20]. For most of these applications, the capability of handling decently high power (i.e., multiple tens of watts) of mid-IR lasers is required. Unfortunately, ChG fiber has low damage threshold, which is in general lower than that of silica by 2-3 orders of magnitude. For example, for As-S glass fibers, which is the most commonly used ChG fiber, the maximum incident power density is 12 MW/cm2 under a 2-µm CW thulium-doped silica fiber laser [21], 2.8 MW/cm2 under a 2.5-µm CW Cr:ZnSe laser [21], 0.23 MW/cm2 under a 4.1-µm CW Fe:ZnSe laser [21], and 0.18 ± 0.05 MW/cm2 under a 5.4-µm CW CO laser [14,22], respectively. Either low laser damage threshold or the longwave transmission of ChG fiber is due to the same origin, i.e., the weak chalcogen-related chemical bonds forming the glass network. Note that the longwave transmission limit of an optical glass can be explained by the Hooke’s law using a simple two-mass spring model [23]. It can therefore be deduced that the enhancement of the damage threshold of ChG by optimizing the glass composition is very likely at the cost of losing the longwave transmission performance and the effect of such compositional optimization is limited.
Large-mode-area (LMA) fiber technology [24,25] becomes the only effective approach to enhance the capability of power handling of ChG fiber, while still maintaining the beam quality. For example, when the effective mode field diameter (MFD) is enhanced from the typical 10 µm to 100 µm, the power handling capability can be scaled up by 2 orders of magnitude. But for a single-mode ultralarge-mode-area (ULMA) fiber with a MFD of 100 µm, the index difference between the core and the cladding should be precisely controlled down to the order of 10−4, corresponding to a numerical aperture (NA) <0.02, which is beyond the limit of the common method by tailoring the compositions of the core and cladding glasses.
Large-mode-area photonic crystal fiber (PCF) [26–30], also known as ‘endless single-mode’ PCF, provides the ideal solution to realize a single-mode fiber with precisely controlled ultralow NA. In principle, an index-guided LMA PCF with an optimized microstructured cladding is without cutoff wavelength for single mode operation, and is single-moded for all the wavelengths and for all the scaled core diameter. So far, for silica fibers, the reported largest MFDs are 135 µm and 205 µm in active and passive PCFs [28,29] respectively. For mid-IR ChG fibers, the largest MFD of 80 µm has been reported in a single-mode all-solid chalcogenide PCF based on germanium-arsenic-selenium/sulfur glasses [30].
As one of the most common effects to influence the guiding mode performance of an ULMA fiber, the bending effect on large-mode area fibers has been comprehensively studied [31–33]. Essentially, the bending leads into an asymmetric index shift across the core and push the light propagating towards the outside of the bend. Therefore, bending distorts the mode profile and narrows the intensity distribution (i.e., the reduction of the mode area). It is known that the scaling up of the mode-area requires the bending radius to be scaled proportional to NA−3 [33]. For example, the conventional fiber is with a MFD of 10 µm and NA of ∼0.15, and its corresponding bending radius is at centimeter-level. This means that for an ULMA fiber with a MFD of 50-100 µm and a NA ≤0.02, the bending radius should be scaled up to multiple meters. Such a cubic behavior of the bending radius requires a straight rod-type configuration for a single-mode ULMA fiber in the practical usage, in order to avoid the significant bending-induced loss. It actually gives up the fiber flexibility, which is one of the most important advantages of the fibers for compact photonic devices.
The key to achieve effective single-moded ULMA PCF is to enhance the discrimination of confinement loss of the fundamental mode and other higher-order modes. On the other hand, single-mode operation has also been demonstrated in few-moded LMA fibers, for example a dual-moded fiber, by coiling the fiber to induce significant bend loss for all higher-order modes but the lowest-order fundamental mode [25,34], because the bending losses of higher-order modes are significantly higher than that of the fundamental mode [56]. Hence, few-moded ULMA fibers are widely used in high-power silica fiber lasers, because (i) the MFDs of a higher-order modes are typically larger than the fundamental mode in a few-moded ULMA PCF, and (ii) a few-moded ULMA fiber has the freedom to have the near-diffraction-limit, high beam-quality output by selectively launching and/or suitably bending. But so far few-moded ULMA fiber technology has not been applied to mid-IR chalcogenide fibers for high-power usage yet.
In this paper, a few-moded ChG PCF with ULMA for mid-IR high power applications has been designed and fabricated. By adjusting the microstructure parameters, a mid-IR chalcogenide PCF with dual-moded operation is obtained. The numerical simulation indicates that the fiber has the recorded largest mode areas of >10000 µm2, i.e., MFD >115 µm, for the fundamental mode LP01 and the lowest higher-order mode LP11. Dual-moded operation has been confirmed experimentally at 2 µm, in good agreement with the numerical simulation. By launching with lens with long focal length, the fiber shows improved bending loss, indicating the controllable power fraction launching into the different modes. It proves excellent bending resistance of the few-moded chalcogenide ULMA PCF with MFD >100 µm by selective launching. Finally the maximum sustainable incident power density of the fiber is measured to be 150 kW/cm2 under CW laser irradiation at 2 µm.
2. Design and fabrication of few-moded chalcogenide PCF with ULMA
For an ULMA PCF, effective single mode can be realized by introducing sufficiently large confinement loss (CL) contrast between the fundamental mode LP01 and the higher-order modes. Empirically, when the ratio d/Λ is < 0.4-0.45, where d is the hole diameter of the triangular arranged ‘holey’ cladding and pitch Λ is the periodic center-to-center spacing between the neighbouring holes respectively, the PCF exhibits effective single-mode behaviour for all the core sizes in a wide wavelength range [26,27]. Mind that the ratio d/Λ of 0.4-0.45 makes the glass gap between the neighbouring holes close to the size of single lobe of the kidney-like lowest higher-order mode LP11. The effective single-mode fiber is therefore realized by controlling the leakage loss of fundamental mode and the higher-order modes through the geometric microstructure. Moreover, when the ratio d/Λ is slightly above 0.45, the first higher-order mode LP11 will be trapped in the guiding core and the fiber becomes a dual-moded fiber [27].
It has been demonstrated that for a single-mode ULMA PCF, two or three rings of microstructures around the core are enough for having decently low confinement loss (CL) for the fundamental mode and high discrimination of the CLs between the fundamental mode and the higher-order modes [27,29,30]. Here the number of the ring is chosen as 2. As the sketch of the designed PCF structure shown in Fig. 1(a), the microstructured fiber is based on a high-index Ge12As24Se64 (Ge-As-Se) glass and a low-index Ge10As24S66 (Ge-As-S) glass. The glass transition temperatures Tgs of the two glasses are 205°C and 202°C respectively, and they can be thermally co-drawn into fibers [30]. Here we only enlarge the hole-to-spacing ratio (d2/Λ) of two holes on the X’-X direction in the second ring surrounding the core (see Fig. 1(a)) and the hole-to-spacing ratio of all other holes (d1/Λ) is fixed at 0.40, in order to lower the CL of the first higher-order mode LP11.
Fig. 1. (a) Schematic design of Ge-As-Se glass ULMA PCF with two rings of Ge-As-S rods. (b) Measured refractive index curves of the selenide and sulfide glasses over 2-12 µm range. (c) Calculated confinement losses of the lowest order modes LP01 and LP11 at (c) 2 µm and (d) 4 µm when d2/Λ varies between 0.4 and 0.70 and Λ increases from 70 µm to 100 µm. Calculated Aeffs of LP01 and LP11 modes at (e) 2 µm and (f) 4 µm with d2/Λ of 0.40, 0.55 and 0.70 when Λ increases from 70 µm to 100 µm. Note that the Aeff curves of LP01 modes for three d2/Λ ratios are overlapped.
All-solid microstructured cladding is chosen here to (i) stop the collapse of holey microstructure during the thermal fabrication process, (ii) avoid the uncertain attenuation due to the gas impurities in the surrounding air by the evanescent field [35], and (iii) prevent the moisture contamination from the atmosphere [36].
The refractive indices (n) of the two glasses were measured by a J. A. Woollam IR-VASE ellipsometer. The fitted Sellmeier equations are shown below:
The numerical simulation shows that CLs of LP01 and LP11 modes drop with increasing Λ for all wavelengths; at 2 µm, with the increase of Λ from 70 µm to 100 µm, for the fiber with d2/Λ of 0.40-0.70, CLs of LP01 and LP11 modes drop quickly by 1-2 orders of magnitude (see Fig. 1(c)); in comparison, the CLs at 4 µm decrease by a factor of only ∼3 with increasing d2/Λ (see Fig. 1(d)), indicating that the impact of d2/Λ on the CLs becomes much less at longer wavelengths. In our experimental experience, when the CL is > 10 dB/m, the mode is difficult to be observed even in a meter-level long straight fiber; when the CL is < 0.1 dB/m, the mode is easily to be excited. Therefore, by increasing the d2/Λ from 0.4 to 0.55, we can guarantee the fiber dual-moded in the mid-IR region. But when d2/Λ exceeds 0.70, the guidance of more higher-order modes can be supported in the ChG ULMA PCF.
It is seen from Fig. 1(e) and (f), the effective mode areas Aeffs of both LP01 and LP11 modes increase with increasing Λ, and when Λ reaches 90 µm, Aeffs of the two lowest-order modes exceed 10000 µm2 beyond the wavelength of 2 µm; when the d2/Λ ratio varies from 0.40 to 0.55, the effective mode area of the LP01 mode shows almost unchanged, but the Aeff of the LP11 mode decreases due the better confinement within the core; finally it can be seen that the LP11 mode has a larger Aeff than the LP01 mode by ∼10% at both 2 µm and 4 µm.
High purity elemental germanium (5N), arsenic (7N), selenium (6N) and sulfur (6N) were used for glass melting. The batched materials were prepared in a dry-N2 filled glove box and then sealed into low-OH silica ampoules by flame. The materials were then melted in a rocking furnace for more than 12 hours. The melting temperature was 850 °C. The selenide and sulfide glass rods or billets were obtained by quenching the melts in water and then annealing the formed glasses at 200 °C for 2 hours.
Figure 2 plots the transmission spectra of Ge-As-Se and Ge-As-S glasses, measured by a Perkin-Elmer Lambda950 spectrophotometer and a Bruker Tensor 27 Fourier-transform infrared (FTIR) spectrometer. Each glass sample was double-side polished. It is seen that the sulfide glass is highly transparent from 1 µm to 8 µm except the absorption bands of the impurities, while the selenide sample is highly transparent up to 15 µm, indicating these two glasses promising hosts for mid-IR usage.
Fig. 2. Transmission spectra of prepared Ge-As-Se (thickness: 4.2 mm) and Ge-As-S glasses (thickness: 6.8 mm).
The preform was prepared by the stacking technique. The fiber fabrication followed the procedure described in Ref. [30]. Extrusion technique [37] was employed to fabricate a Ge-As-Se tube with ∼5.0 mm inner diameter (ID) and ∼12.3 mm OD (see Fig. 3(a)), and a Ge-As-S rod with ∼5.0 mm OD (see Fig. 3(b)). The set of rod-in-tube (see Fig. 3(c)) was elongated into a ∼1.4 mm OD thin cane. For the side two canes with the enlarged ratio d/Λ of 0.55, a selenide glass tube with ∼12.3 mm OD and ∼6.9 mm ID and a sulfide glass rod with ∼6.9 mm OD were used.
Fig. 3. Photographs of (a) extruded selenide glass tube with ∼12.3 mm OD and ∼5.0 mm ID, (b) extruded sulfide glass rod with ∼5.0 mm OD, (c) set of rod-in-tube before elongation, and (d) extruded ∼12.0 mm-OD selenide jacket tube with an inner hexagonal hole. (e) SEM photo of ChG ULMA PCF
A ∼12.0 mm-OD selenide tube with an inner hexagonal hole (see Fig. 3(d)) was also extruded. The inner hexagon had a side length of ∼3.5 mm. Note that the glass billets used for the above extrusion were in a diameter of 30 mm and a height of 30 mm. The glasses were extruded at the temperature of ∼330 °C.
The selenide/sulfide canes were stacked inside the selenide glass jacket tube. Thermoplastic polyetherimide (PEI) film was wrapped out of the stacked preform as the protection. The preform was then drawn into the fiber with glass OD of 785 µm. The drawing temperature was around 350 °C in the dry nitrogen atmosphere. The total yield of the uniform fiber was ∼10 meters.
Figure 3(e) gives the scanning electron microscopic image of the fabricated ChG PCF. The core diameter is ∼145 µm and the pitch Λ is 90 ± 4 µm. The low-index rods are with the diameter d1 of 36 ± 2 µm and d2 of 50 ± 2 µm, giving the ratios of d1/Λ and d2/Λ of ∼0.40 and ∼0.55, respectively.
Figure 4(a) and (b) illustrate the simulated mode profiles of LP01 and LP11 modes of the fabricated few-ChG ULMA PCF (Λ=90 µm, d1/Λ=0.40, and d2/Λ=0.55) at 2 µm. Note the frame is with a scale of 150 × 150 µm. Figure 4(c) plots the calculated confinement losses of LP01 and LP11 modes of the fabricated fiber in the range of 2-8 µm. It shows that the CL of LP01 mode is lower than 0.1 dB/m and that of LP11 mode is at dB/m level, respectively, indicating that such a straight fiber is a dual-moded fiber within such a two-octave range. Note that the simulation also indicates the CLs of other higher-order modes are with enough mode discrimination between these two lowest-order modes.
Fig. 4. Simulated mode profiles of (a) LP01 and (b) LP11 modes of the fabricated ChG ULMA PCF. Note the frame is with a scale of 300 × 300 µm. (c) Calculated CLs of LP01 and LP11 modes of the fabricated fiber (Λ=90 µm, d1/Λ=0.40, and d2/Λ=0.55) in the range of 2-8 µm, in comparison with the ‘endless single-mode’ PCF with uniform d/Λ=0.40 (d1=d2) and Λ=90 µm. (d) Calculated Aeffs of LP01 and LP11 modes of the fabricated fiber (Λ=90 µm, d1/Λ=0.40, and d2/Λ=0.55) in the range of 2-8 µm, in comparison with the ‘endless single-mode’ PCF with uniform d/Λ=0.40 (d1=d2) and Λ=90 µm. Note that the Aeff curves of LP01 modes for both d2/Λ ratios are overlapped.
Figure 4(d) plots the calculated effective mode areas Aeffs of LP01 and LP11 modes of the fabricated fiber. The Aeff of LP01 mode ranges from 10400 to 10700 µm2 (e.g. corresponds to a MFD of 115 µm at 2 µm) in 2-8 µm, while the Aeff of LP11 mode ranges from 11300 to 12200 µm2. These are the largest Aeffs in mid-IR chalcogenide fibers reported so far.
3. Bending characterization of few-moded ChG ULMA PCF at 2 µm
Due to the lack of a consensus of bending loss measurement of a LMA PCF particularly in the mid-infrared wavelengths, the macro-bending losses measurement followed the principle of the formal industry standard, IEC 60793-1-47, which was used for determining the bending loss of a single-mode or a multimode fiber. According to such a standard, the bending is carried out on the fiber sample making an equivalent number of partial turns in the open air or around suitable supports.
Figure 5(a) illustrates the schematic experiment setup of bending characterization of the fabricated ChG ULMA PCF at the wavelength of 2 µm. A home-made laser-diode pumped single-mode thulium-doped silica fiber master-oscillator-power-amplifier (MOPA) emitting at 1.998 µm was used as the laser source. A piece of ChG ULMA PCF with the length of 67 cm was used in the experiment. The MFD of the thulium-doped fiber (TDF) was ∼25 µm. The collimating lens #1 after the TDF was with a focal length f1 of 25 mm. The focal length f2 of the focusing objective lens #2 in front of the ChG ULAM PCF was chosen to be 75 mm or 100 mm, providing a factor of 3 or 4 for beam expanding, respectively. In the case of using the lens #2 with the f2 of 75 mm, the laser spot size launched onto the input end of the ChG PCF was ∼75 µm, which fills ∼65% of the MFD of LP01 mode for the ULMA PCF. In the case of using the focal lens with the f2 of 100 mm, the laser spot size launched onto the input end of the ChG ULMA PCF was ∼100 µm, which fills ∼87% of the MFD of LP01 mode. A Tigris-640 MCT-BB high-speed MWIR (1.5-6 µm) camera was used to monitor the output mode profile from the PCF. Once the guidance was confirmed within the fiber core, a power meter with thermal power sensor head was used to measure the output power from the PCF.
Fig. 5. (a) Schematic of experimental setup for bending characterization of fabricated ChG ULMA PCF at 2 µm. R: bending radius of the curved fiber. (b) Photograph of ChG ULMA PCF (the black line) with R of 12.7 cm in measurement.
Initially the PCF was maintained straight for the mode observation. The PCF was then bent with a reducing radius. Because the OD of the PCF with the PEI coating was ∼1 mm and the bending radius is in the range of meter to sub-meter level, the bending radius R (i.e., the reciprocal of the curvature) was measured by photographing the bent fiber (see Fig. 5(b)), rather than using the method by progressively wrapping the fiber around a mandrel with a fixed radius [31]. The bending-induced loss under a certain bending radius R was determined through dividing the reduction of the transmitted power by the curved fiber length. During the measurement, the fiber has been bent approximately on the X-X’ plane, by observing the output mode profile with the IR camera. In addition, the IR camera was used for confirming that the guiding was still well confined inside the core after the fiber was bent.
Figure 6(a) illustrates the relation between the bending loss (in dB/m) at 2 µm of the ChG ULMA PCF and the bending radius, using the objective lens with f2 of 75 mm. It is seen that the bending loss increases significantly with the curvature (1/R) of the bent fiber and the bending loss is enhanced to >2.5 dB/m, equivalent to 2.0 dB/turn, when the bending radius R is 12.7 cm, corresponding to the curvature 1/R of 7.9 m−1. Figure 6(b) & (c) show the observed near-field mode profiles from the fiber output when the PCF was straight and bent with the minimum radius of 12.7 cm. It is seen that the first higher-order mode LP11 was excited when the fiber was straight and only the fundamental mode LP01, which was pushed to one side on the X-X’ plane, remained when the bent radius R was 12.7 cm.
Fig. 6. Measured bending loss and fitting with exponential decay function of few-moded ChG ULMA PCF using objective lens with f2 of (a) 75 mm and (b) 100 mm at the wavelength of 2 µm.
On the other hand, Fig. 6(d) illustrates the relation between the bending loss (in dB/m) at 2 µm of the ChG ULMA PCF against the bending radius, using the objective lens with f2 of 100 mm. The bending loss increases slowly with the curvature (1/R) of the bent fiber and the bending loss only increases to ∼0.7 dB/m when the bending radius R is 12.7 cm. Similar to Fig. 6(b) & 6(c), the first higher-order mode LP11 was excited when the fiber was straight; and only the fundamental mode LP01, which was pushed to one side on the X-X’ plane, can be seen when the bent radius R was 12.7 cm (as shown in Fig. 6(e) & 6(f)). In addition, the few-moded ULMA PCF was observed insensitive to the polarization of the incident 2-µm laser. Consequently no polarization-dependent bending characterization was done for this few-moded ULMA PCF.
Considering that the dual modes exhibit different behaviors under bending, the bending loss can be fitted with the sum of two single exponential decay functions [38,39] of ${A_1} \cdot \exp ( - R/{R_1}) + {A_2} \cdot \exp ( - R/{R_2}) + c$, where A1, A2 and c are constants. Table 1 summarizes the fitting parameters A1, R1, A2, and R2 using lenses with focal lengths f2 of 75 and 100 mm respectively. The bending performance of the few-moded ChG ULMA PCF is seen highly dependent to the launching condition. When the focused laser spot size is ∼ 2/3 of the MFD of the fundamental mode, more laser power has been coupled into the LP11 mode and the bending loss increases significantly with increasing bending curvature (1/R). When the focused laser spot size is close to the MFD of the LP01 mode, more laser power has been coupled into the LP01 mode and the few-moded ChG ULMA PCF shows highly bending resistibility; the bending loss is < 0.7 dB/m, equivalent to 0.55 dB/turn, even when the bending radium R is as small as 12.7 cm.
Table 1. Summary of fitting parameters of bending loss using lens #2 with different focal length f2
4. Laser damage characterization of ChG ULMA PCF
Measurements of high power delivery and laser damage of the ChG ULMA PCF were carried out at 2 µm using the same set-up as show in Fig. 5(a), while the objective lens was with f2 of 100 mm and the fiber was maintained straight. The 2-µm thulium doped silica fiber power amplifier was selected because it was the only available high-power mid-IR source with >10 W output in our lab.
Figure 7 shows the loss curve of an unclad Ge-As-Se fiber measured by the cutback method. The bare fiber was drawn from a glass rod. The loss spectrum was measured on the Bruker Tensor 27 FTIR system, equipped with an MCT detector and a fiber coupling stage. The unclad fiber has the loss lower than ∼2 dB/m in 2-9 µm, and 2-4 dB/m on shortwave (1.5-2 µm) and longwave (9-9.5 µm) sides. The selenide glass fiber is therefore shown promising for short-length mid-IR fiber devices. Note that the Se-H impurity remaining in Ge-As-Se glass exhibits absorption bands at 2.32, 3.53, 4.12, and 4.57 µm [40,41].
The 2-µm loss of the PCF was measured to be 7.8 ± 0.2 dB/m (see the scatter point in Fig. 7), using four cutbacks from 67 cm to 47 cm. The extra loss of the PCF over the bare fiber could be due to (i) the imperfections created inside the glass in the multiple-stage thermal processing; (ii) the defects formed on the interstitial interfaces from stacking; and (iii) some influence due to the bending. It should be expected that the loss of the ChG PCF can ultimately be lowered down to the level of 0.5-1 dB/m, by further optimizing the fiber design, purifying the glasses, and eliminating possible contaminations in the fabrication.
The 2-µm laser with high power was launched into the ChG ULMA PCF with a length of 47 cm, left from the cutback measurement. The coupling efficiency was estimated to be ∼11% according to the linear fitting of the cutback data. The fiber was kept straight during the measurement. The power of the 2-µm thulium doped silica fiber laser increases gradually from multiple tens mW-level to multiple watts, until the fiber input was damaged. The incident laser power, which was measured before the fiber input end, and the output power from the exit of the fiber have been recorded. Under each incident power, the fiber was irradiated for ∼ 30 seconds. Under the maximum incident power of 11.8 W, the input end of the fiber began to burn immediately (<1 second) after the power reached the set value, indicating that the facet failure was due to the thermal induced breakdown.
Figure 8(a) shows the photograph of the ChG ULMA PCF sandwiched between two aluminium plates with forced water-cooling. Thermally conductive adhesive was put onto the fiber between the plates for increasing the thermal conductivity between the glass fiber and the aluminum metal. Figure 8(b) shows the photograph of incident end of the ChG ULMA PCF. The fiber is only ∼1 mm out of the aluminum plates. Figure 8(c) shows the photograph of the bottom aluminum plate contaminated by the burnt fiber. It is seen the first ∼1 cm part of the fiber was burnt. Figure 8(d) shows the relation between the incident power and output power through 47-cm long ChG ULMA PCF at 2 µm, under forced water cooling condition. It is seen that the output power increases almost linearly with the increase of the incident power, until the fiber facet was damaged under the incident power of 11.8 W. The forced water cooling has shown significant influence on the power handling. In an additional experiment without using any forced cooling condition, the fiber facet of the ChG ULMA PCF with the similar length was damaged when the incident power was only 5.1 W. Note that the PEI layer out of the glass fiber was dissolved for the first centimeter fiber for the this measurement.
Fig. 8. Photographs of (a) ChG ULMA PCF sandwiched between two aluminum plates with forced water-cooling, (b) incident end of PCF, and (c) bottom aluminum plate after input facet of PCF was damaged. (d) Relation between incident and output power under forced water cooling condition. Used PCF length is 47 cm.
The laser spot launched onto the fiber facet was ∼100 µm, the maximum incident laser power density of Ge-As-Se/Ge-As-S ULMA PCF under 2-µm CW laser irradiation is thus estimated to be 150 kW/cm2. The ULMA PCF is expected to sustain higher incident power density in 2-9 µm (except around 4.6 µm), where the fiber shows lower loss than that at 2 µm (see Fig. 7).
Due to lack of the data of the maximum CW power density of the Ge-As-Se glass bulk at 2 µm, an extra high-power facet damage test was carried out for bulk ChG glasses using the same set-up. Three ChG glass samples with the compositions of Ge12As24Se64, Ge10As24S66, and the well-studied As2S3 (As-S) were prepared. The samples were in the disc shape with an OD of 20 mm and a thickness of 6 mm. All the samples have one side polished and another side with mat surface. The polished face of the sample under testing was facing the laser irradiation at the focal position. The laser power increases step by step with a step of ∼0.5 W, until the sample surface was observed damaged. Under each power, the bulk was irradiated for ∼30 seconds.
Figure 9(a)-(e) illustrate the optical microscopic photographs of the bulk surface damaged irradiated by 2-µm laser with the focused spot of ∼100 µm. It is seen that the front facet of Ge-As-Se glass sample was with a hole of ∼98 µm (see Fig. 9(a)), when the power of the 2-µm laser was raised to 6.0 W (corresponding to an incident power density of ∼75 kW/cm2), indicating the clear thermal-induced damage under the CW laser irradiation. It is seen that our measurements on the bulk is in good agreement with the result on the fiber. The heat accumulates at the very short beginning part of the glass or fiber (see Fig. 8(c) and Fig. 9(a)). For the fiber, since the heat distributes along the long fiber, forced cooling with air or water can take the heat away quickly and the maximum incident power density is doubly higher than the one from the bulk measurement.
Fig. 9. Optical microscopic photographs of bulk surface damaged irradiated by 2-µm CW laser with focused spot of ∼100 µm: (a) Ge-As-Se glass under 6.0 W (using reflection mode); (b) front surface (using reflection mode) and ∼3 mm below from surface (using transmission mode) of Ge-As-S glass under 15.9 W; (d) front surface (using reflection mode) and (e) ∼3 mm below from surface (using transmission mode) of As2S3 glass under 10.4 W.
However, for Ge-As-S and As-S samples, no front facet damage was observed on the reflection mode of the microscope, under the incident power of 15.9 W (corresponding to an incident power density of ∼200 kW/cm2) and 10.4 W (corresponding to an incident power density of ∼130 kW/cm2), respectively (see Fig. 9(b) & 9(c)). Under the transmission mode of the microscope, by moving the focal position from the surface into the bulk by ∼3 mm, it is seen that the samples were damaged inside with sub-mm hole. This is believed to be arising from (i) the low absorbance of Ge-As-S and As-S samples at 2-µm and (ii) the long waist of the objective lens with 100-mm long focal length. Essentially, the heat was generated from the absorbed laser power and accumulated inside the glass at the beginning of the optical path in the sample. The ambient air cools the surface and therefore causes significant temperature gradient from surface to inner body in the case of using a focusing lens with long focal length. Once the temperature is enhanced close to the glass transition temperature, which is 205 °C for Ge-As-Se glass, 202 °C for Ge-As-S glass, and 196 °C for As-S glass [14] respectively, thermal induced break-down occurs on the laser irradiated side. The maximum incident power densities Ge-As-S and As-S bulks are seen higher than that of Ge-As-Se glass by the factor of 3 and 2, respectively, indicating that the higher chemical bonding strength of a chalcogenide glass, the higher power handling capability of the glass.
Figure 10 summarizes the reported maximum incident power density of the ChG glass bulks and fibers under mid-IR CW laser irradiation, according to Refs. [14,21,22,42] and this work. In the case of As-S fibers, when at 2 µm and 2.5 µm, where the fibers show low loss of ≤1 dB/m, the fiber without antireflection coating can work under CW laser irradiation with high power density of MW/cm2 level; but at 4.1 µm, where the fiber has loss close to 4 dB/m due to the S-H impurity, the fiber can only handle the incident power density of ∼200 kW/cm2 [21]; while at 5.4 µm, the fiber can handle the incident power density of ∼230 kW/cm2, although the fiber has low loss of 0.45 dB/m [14]. In the case of Ge-As-Se-Te fibers, when the fiber has the loss of ∼1 dB/m at 10.6 µm, the fiber can sustain a maximum incident power density of 54 kW/cm2 [22]; when the fiber has the loss of ∼6 dB/m, the fiber can only work under CW CO2 laser irradiation with the power density of 5.4 kW/cm2 [42]. We can therefore deduce that the reason why our As-S sample can sustain the incident laser power density of only ∼130 kW/cm2, which is much lower than the reported 12 MW/cm2 at 2 µm, probably is because the optical quality of our ChG bulk is lower than the one used in Ref. [21].
It is known that the fiber loss mainly arises from the scattering defects and/or the absorbing impurities in the fiber matrix, and the facet damage of a ChG glass or fiber under CW laser irradiation is thermal-induced failure. It is easy to understand that the impurity absorbs the laser power and increases the temperature around the incident section and leads into the thermal failure. Hetero-phase inclusions, which are commonly existing scattering defects in ChG, are also another type of impurities [40]. Such inclusions consist of substances (e.g., carbon and silicon) which are not readily soluble in the molten chalcogenide, and forms hetero-phase structures inside the glass matrix with sizes of tens of nanometers to several microns. Although these hetero-phase inclusions typically might not show distinguishable absorption bands in the transmission spectrum, they possess very different thermal behavior (e.g., the thermal expansion coefficient and viscosity) from the glass matrix. Under CW laser irradiation with high power, the existence of high-content hetero-phase inclusions will reduce the power handling capability of a ChG glass by a few orders of magnitude, in comparison with the glass matric with low-content hetero-phase inclusions [40].
All in all, it can be deduced that the coupling efficiency, glass forming bonding strength, fiber attenuation, and impurity-induced hetero-phase inclusion defects are the current major factors limiting our ChG ULMA PCF reaching the recorded MW/cm2-level incident power density at 2 µm. Besides using high purity ChG glass, carefully tailoring the laser power distribution on both cross-sectional and longitudinal directions of the fiber will be the key to enhance the power handing capability of the ChG ULMA PCF. Technical solutions for enhancing the coupling efficiency of the ChG ULMA PCF, further purifying the glasses, and reducing the defects formed in the fiber are currently being investigated.
5. Conclusion
In summary, we have demonstrated a novel few-mode chalcogenide photonic crystal fiber with an ultralarge mode area >10000 µm2 for mid-IR high-power usage. The few-moded ChG ULMA PCF shows the excellent bending resistance even with a small bending radius of tens of centimeters. The facet damage characterization under high power 2-µm laser indicates that the promise of using such a type of few-moded ChG ULMA PCFs for mid-IR 10-100 W high-power applications.
Funding
National Natural Science Foundation of China (61575086); Priority Academic Program Development of Jiangsu Higher Education Institutions; Jiangsu Collaborative Innovation Centre of Advanced Laser Technology and Emerging Industry.
Disclosures
The authors declare no conflicts of interest.
References
1. A. Schliesser, N. Picqué, and T. W. Hänsch, “Mid-infrared frequency combs,” Nat. Photonics 6(7), 440–449 (2012). [CrossRef]
2. Y. Yao, A. J. Hoffman, and C. F. Gmachl, “Mid-infrared quantum cascade lasers,” Nat. Photonics 6(7), 432–439 (2012). [CrossRef]
3. S. D. Jackson, “Towards high-power mid-infrared emission from a fibre laser,” Nat. Photonics 6(7), 423–431 (2012). [CrossRef]
4. F. Maes, V. Fortin, S. Poulain, M. Poulain, J. Carrée, M. Bernier, and R. Vallée, “Room-temperature fiber laser at 3.92 µm,” Optica 5(7), 761–764 (2018). [CrossRef]
5. S. B. Mirov, V. V. Fedorov, D. Martyshkin, I. S. Moskalev, M. Mirov, and S. Vasilyev, “Progress in mid-IR lasers based on Cr and Fe-doped II–VI chalcogenides,” IEEE J. Sel. Top. Quantum Electron. 21(1), 292–310 (2015). [CrossRef]
6. A. Ventura, F. B. Slimen, J. Lousteau, N. White, A. Masoudi, P. Janicek, and F. Poletti, “Flexible Mid-IR fiber bundle for thermal imaging of inaccessible areas,” Opt. Express 27(15), 20259–20272 (2019). [CrossRef]
7. D. J. Richardson, J. Nilsson, and W. A. Clarkson, “High power fiber lasers: current status and future perspectives,” J. Opt. Soc. Am. B 27(11), B63–B92 (2010). [CrossRef]
8. G. Agrawal, Nonlinear Fiber Optics, 4th ed. (Academic, 2009).
9. V. A. G. Rivera and D. Manzani, eds. Technological Advances in Tellurite Glasses: Properties, Processing, and Applications, 1st ed. (Springer Series in Materials Science, 2017).
10. M. R. Majewski, R. I. Woodward, J. Carreé, S. Poulain, M. Poulain, and S. D. Jackson, “Emission beyond 4 µm and mid-infrared lasing in a dysprosium-doped indium fluoride (InF3) fiber,” Opt. Lett. 43(8), 1926–1929 (2018). [CrossRef]
11. B. Eggleton, B. Luther-Davies, and K. Richardson, “Chalcogenide photonics,” Nat. Photonics 5(3), 141–148 (2011). [CrossRef]
12. X. Feng, A. K. Mairaj, D. W. Hewak, and T. M. Monro, “Nonsilica Glasses for Holey Fibers,” J. Lightwave Technol. 23(6), 2046–2054 (2005). [CrossRef]
13. S. Qi, B. Zhang, C. Zhai, Y. Li, A. Yang, Y. Yu, D. Tang, Z. Yang, and B. Luther-Davies, “High-resolution chalcogenide fiber bundles for longwave infrared imaging,” Opt. Express 25(21), 26160–26165 (2017). [CrossRef]
14. S. Sato, K. Igarashi, M. Taniwaki, K. Tanimoto, and Y. Kikuchi, “Multihundred-watt CO laser power delivery through chalcogenide glass fibers,” Appl. Phys. Lett. 62(7), 669–671 (1993). [CrossRef]
15. A. Sincore, J. Cook, F. Tan, A. F. Abouraddy, M. C. Richardson, and K. L. Schepler, “Practical limits of power transmission through single-mode chalcogenide fibers,” Opt. Eng. 57(11), 1 (2018). [CrossRef]
16. M. Bernier, V. Fortin, N. Caron, M. El-Amraoui, Y. Messaddeq, and R. Vallée, “Mid-infrared chalcogenide glass Raman fiber laser,” Opt. Lett. 38(2), 127–129 (2013). [CrossRef]
17. K. Hu, I. V. Kabakova, T. F. S. Büttner, S. Lefrancois, D. D. Hudson, S. He, and B. J. Eggleton, “Low-threshold Brillouin laser at 2 µm based on suspended-core chalcogenide fiber,” Opt. Lett. 39(16), 4651–4654 (2014). [CrossRef]
18. C. R. Petersen, U. Møller, I. Kubat, B. Zhou, S. Dupont, J. Ramsay, T. Benson, S. Sujecki, N. Abdel-Moneim, Z. Tang, D. Furniss, A. Seddon, and O. Bang, “Mid-infrared supercontinuum covering the 1.4–13.3µm molecular fingerprint region using ultra-high NA chalcogenide step-index fibre,” Nat. Photonics 8(11), 830–834 (2014). [CrossRef]
19. K. Jiao, J. Yao, X. Wang, X. Wang, Z. Zhao, B. Zhang, N. Si, J. Liu, X. Shen, P. Zhang, S. Dai, Q. Nie, and R. Wang, “1.2–15.2 µm supercontinuum generation in a low-loss chalcohalide fiber pumped at a deep anomalous-dispersion region,” Opt. Lett. 44(22), 5545–5548 (2019). [CrossRef]
20. M. Zhang, L. Li, T. Li, F. Wang, K. Tian, H. Tao, X. Feng, A. Yang, and Z. Yang, “Mid-infrared supercontinuum generation in chalcogenide fibers with high laser damage threshold,” Opt. Express 27(20), 29287–29296 (2019). [CrossRef]
21. A. Sincore, J. Cook, F. Tan, A. Halawany, A. Riggins, S. McDaniel, G. Cook, D. V. Martyshkin, V. V. Fedorov, S. B. Mirov, L. Shah, A. F. Abouraddy, M. C. Richardson, and K. L. Schepler, “High power single single-mode delivery of mid mid-infrared sources through chalcogenide fiber,” Opt. Express 26(6), 7313–7323 (2018). [CrossRef]
22. J. S. Sanghera, I. D. Aggarwal, L. B. Shaw, L. E. Busse, P. Thielen, V. Nguyen, P. Pureza, S. Bayya, and F. Kung, “Applications of chalcogenide glass optical fibers at NRL,” J. Optoelectron. Adv. Mater. 3, 627–640 (2001).
23. S. Wartewig, IR and Raman Spectroscopy: Fundamental Processing (Spectroscopic Techniques: An Interactive Course) (Wiley-VCH, 2003).
24. D. Taverner, D. J. Richardson, L. Dong, J. E. Caplen, K. Williams, and R. V. Penty, “158-µJ pulses from a single-transverse-mode, large-mode-area erbium-doped fiber amplifier,” Opt. Lett. 22(6), 378–380 (1997). [CrossRef]
25. H. L. Offerhaus, N. G. Broderick, D. J. Richardson, R. Sammut, J. Caplen, and L. Dong, “High-energy single-transverse-mode Q-switched fiber laser based on a multimode large-mode-area erbium-doped fiber,” Opt. Lett. 23(21), 1683–1685 (1998). [CrossRef]
26. P. S. J. Russell, “Photonic crystal fibers,” Science 299(5605), 358–362 (2003). [CrossRef]
27. L. Dong, H. A. McKay, and L. Fu, “All-glass endless single-mode photonic crystal fibers,” Opt. Lett. 33(21), 2440–2442 (2008). [CrossRef]
28. F. Stutzki, F. Jansen, A. Liem, C. Jauregui, J. Limpert, and A. Tünnermann, “26 mJ, 130 W Q-switched fiber-laser system with near-diffraction-limited beam quality,” Opt. Lett. 37(6), 1073–1075 (2012). [CrossRef]
29. A. Steinkopff, C. Jauregui, F. Stutzki, J. Nold, C. Hupel, N. Haarlammert, J. Bierlich, A. Tünnermann, and J. Limpert, “Transverse single-mode operation in a passive large pitch fiber with more than 200 µm mode-field diameter,” Opt. Lett. 44(3), 650–653 (2019). [CrossRef]
30. H. Ren, S. Qi, Y. Hu, F. Han, J. Shi, X. Feng, and Z. Yang, “All-solid mid-infrared chalcogenide photonic crystal fiber with ultralarge mode area,” Opt. Lett. 44(22), 5553–5556 (2019). [CrossRef]
31. J. C. Baggett, T. M. Monro, K. Furusawa, V. Finazzi, and D. J. Richardson, “Understanding bending losses in holey optical fibers,” Opt. Commun. 227(4-6), 317–335 (2003). [CrossRef]
32. J. M. Fini, “Bend-resistant design of conventional and microstructure fibers with very large mode area,” Opt. Express 14(1), 69–81 (2006). [CrossRef]
33. R. T. Schermer, “Mode scalability in bent optical fibers,” Opt. Express 15(24), 15674–15701 (2007). [CrossRef]
34. J. P. Koplow, D. A. V. Kliner, and L. Goldberg, “Single-mode operation of a coiled multimode fiber amplifier,” Opt. Lett. 25(7), 442–444 (2000). [CrossRef]
35. P. Toupin, L. Brilland, G. Renversez, and J. Troles, “All-solid all-chalcogenide microstructured optical fiber,” Opt. Express 21(12), 14643–14648 (2013). [CrossRef]
36. P. Toupin, L. Brilland, D. Méchin, J. Adam, and J. Troles, “Optical Aging of Chalcogenide Microstructured Optical Fibers,” J. Lightwave Technol. 32(13), 2428–2432 (2014). [CrossRef]
37. H. Ren, Y. Yu, C. Zhai, B. Zhang, A. Yang, K. Tian, X. Feng, Z. Yang, P. Wang, and B. Luther-Davies, “Chalcogenide glass fibers with a rectangular core for polarized mid-infrared supercontinuum generation,” J. Non-Cryst. Solids 517, 57–60 (2019). [CrossRef]
38. W. S. Wong, X. Peng, J. M. McLaughlin, and L. Dong, “Breaking the limit of maximum effective area for robust single-mode propagation in optical fibers,” Opt. Lett. 30(21), 2855–2857 (2005). [CrossRef]
39. A. W. Snyder and J. D. Love, Optical Waveguide Theory (Chapman and Hall Ltd., 1983).
40. G. E. Snopatin, V. S. Shiryaev, V. G. Plotnichenko, E. M. Dianov, and M. F. Churbanov, “High-purity chalcogenide glasses for fiber optics,” Inorg. Mater. 45(13), 1439–1460 (2009). [CrossRef]
41. Z. Tang, S. Shiryaev, D. Furniss, L. Sojka, S. Sujecki, T. Benson, A. Seddon, and M. Churbanov, “Low loss Ge-As-Se chalcogenide glass fiber, fabricated using extruded preform, for mid-infrared photonics,” Opt. Mater. Express 5(8), 1722–1737 (2015). [CrossRef]
42. J. Su, S. Dai, L. Jiang, C. Lin, C. Yang, N. Zhang, and Y. Yuan, “Fabrication and bending strength analysis of low-loss Ge15As25Se40Te20 chalcogenide glass fiber: a potential mid-infrared laser transmission medium,” Opt. Mater. Express 9(7), 2859–2869 (2019). [CrossRef]
