Open Access
Add to BibSonomy
Abstract
We present the design and construction of a fiber combiner comprising 7 × 1 35/50 µm input fibers with thin cladding and an output fiber with 125 µm diameter to maintain the high beam brightness. A transmission higher than 98% between the input fibers and the tapered fiber bundle is achieved. The coupling efficiency between the bundle and the output fiber exceeds 95%. The-near field beam profiles of the fabrication steps are measured and analyzed. Due to the thin cladding of the input fibers, we analyze optical tunneling effects with respect to the numerical aperture of the light.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
In contrast to solid-state lasers and gas lasers, diode lasers are characterized by a high electro-optical efficiency and a high cost efficiency. Power scaling by means of geometrical multiplexing leads to a decreased beam quality, and for this reason diode lasers are typically not used for direct industrial applications. In the context of the BRIDLE (Brilliant Industrial Diode LasEr) [1] project, direct diode laser metal processing shall be enabled. A laser source based on diode lasers with an output power of > 2 kW from a 100 µm optical fiber and an efficiency >40% shall be designed [1]. A fiber-coupled diode-laser module is developed, based on dense wavelength division multiplexing around 1064 nm [2]. For the first time, this enables coupling of high power diode laser radiation (P > 10 W) into a fiber with a core diameter below 50 µm.
However, the output power of a single module is still too low for many practical applications. Therefore, a 7 × 1 fiber combiner is developed to scale the output power of the high brightness diode-laser module. By using this combiner seven diode-laser modules can be coupled into one fiber. The target is to maintain the beam brightness of the individual modules, while keeping the power losses below 10 %.
Commercially, 7 × 1 multimode fiber combiners are available from many manufacturers. For the input fibers typically 105/125 µm core/cladding diameter fibers with a numerical aperture of 0.22 are used, while the input efficiency is typically over 90%. Other multimode pump combiner have been demonstrated, with 200/220 µm core/cladding diameter fibers, a pump power handling of 3.01 kW and an efficiency of 99.4% [3] or a 7 × 1 fiber combiner with 105/125 µm input fibers and an 300/320 µm output fiber with an efficiency of 99.4% [4]. However, the brightness conservation of these multimode fiber combiners is not sufficient for the task within the BRIDLE project.
Other 7 × 1 signal fiber combiners have been demonstrated, which typically have a higher input beam brilliance. 7 × 1 fiber combiner with the combination of seven single-mode fibers into one 50 µm core diameter multimode fiber [5], seven 17 µm core diameter fiber into one 100 µm multimode fiber [6] or the use of seven 20/130 µm fibers into one 100 µm core diameter fiber with an efficiency close to 99% [7] have been presented, which also are not applicable to the parameters of the BRIDLE module output fiber.
Therefore, we design a fiber combiner with an input fiber core diameter below 50 µm, which matches the beam brightness of the BRIDLE module. The corresponding beam parameter product is 2.8 mm mrad [2]. To maintain the high beam brightness, fibers with a thin cladding are used. Even though the light is guided within the fiber core, a fraction of the light, which is called the evanescent field, extends into the fiber cladding. If the cladding is too thin, the evanescent field might extend even beyond its boundary. In the case of an optical fiber combiner, this means that a fraction of the optical power within one fiber can be coupled into another fiber or the surrounding capillary, which may result in additional losses. Therefore, this effect, which is called the optical tunneling effect, has to be analyzed.
In section 2 of this paper we present the design and the production of the fiber combiner taking also the theoretical background into account. In section 3 an analysis according to optical tunneling is performed, because of the small cladding thicknesses of circa 4 µm. In section 4 and 5 we present the experimental results and the analysis of the manufacturing steps.
2. Design and production
The production of the 7 × 1 combiner includes the taper process, the cleaving of the fiber bundle and the output fiber splice, which are explained within this chapter.
The fibers used exhibit a 35/50 µm core/cladding diameter fiber with a thin cladding and a numerical aperture ${A_{N,F}}$ of 0.22. The output fiber is a 125 µm core diameter fiber with a numerical aperture of 0.48. Such a fiber is typically used for a fiber laser, where the light guided by the fiber bundle acts as the pump light. The higher numerical aperture is important due to the increasing numerical aperture of the light caused by the taper process.
The schematic principle of the combiner is shown in Fig. 1. For the manufacturing the seven input fibers are arranged in a symmetric hexagonal pattern within a capillary, as shown in Fig. 2.
For the conservation of the high beam brightness it is important that the ratio ${V_R}$ of the total area of the cores of the seven fibers in comparison to the inner total cross section of the capillary is high. To maximize this ratio, expressed in
a fiber with a thin cladding is used, where ${d_{core}}$ is the diameter of the fiber core and ${d_{clad}}$ is the outer diameter of the fiber cladding. The usage of fibers with a small cladding may result in additional losses due to optical tunneling, which will be discussed in chapter 3. For the used fibers the ratio of the core area in comparison to the full area is equal to 38%.To produce a fiber bundle this arrangement is tapered to a smaller outer diameter, so that the fibers and the capillary can be fused together and the bundle’s outer diameter matches the diameter of the output fiber core. For that it is important to take into account the numerical aperture of the light, which is guided within the fiber. It is assumed that the beam parameter product BPP for the light guided within a fiber can be expressed as
where ${\omega _0}$ is the beam waist, $\theta $ the far-field divergence angle and ${A_{N,L}}$ the numerical aperture of the light which is guided within a fiber. ${A_{N,L}}$ has to be lower than the numerical aperture of the guiding fiber$\; {A_{N,F}}$, in order to guide the light within the fiber’s core.The beam parameter product remains constant during the taper process. Therefore the numerical aperture of the light has to increase inversely proportional, if the core diameter decreases. The numerical aperture of the tapered fiber itself remains constant. Within the approximation that the beam parameter product remains constant during the taper process, it is possible to formulate the relation between the numerical aperture of the light before and after the taper, which can be expressed by the equation
The taper process is the most important production step. Our experiments have shown that the beam quality and the degree of transmission highly depend on the quality of the taper. For the handling of the fiber bundle it is important that the input fibers fuse during the taper process, so that the fiber bundle can be cleaved. Our experiments have shown that this condition is fulfilled for a taper ratio ${V_{Taper}}\; \,$ 0.58. Because of the conservation of the beam parameter product, light with a numerical aperture of 0.12 should remain within the fiber’s core, even if the fiber is tapered. Many parameters have an influence on the transmission characteristic of the fiber combiner. Some of the most important parameters are the length of the taper region and the geometrical properties of the fibers in the tapered fiber bundle. Other important technical parameters are for example the taper velocity and the applied heating power. These tapering parameters are experimentally optimized in order to minimize the optical losses within the fiber combiner.
After the taper process, the fiber bundle has to be cleaved. In Fig. 3 the schematic and the actual facet of the tapered fiber bundle are presented. The input fibers are arranged in the desired hexagonal formation. Microscopic measurements show that the target parameters are reached. This shows that the taper process works well geometrically. The movement of the fibers out of the capillary is not possible anymore, because they fused with the capillary.
Fig. 3. Schematic representation (left) and microscopic measurement of the facet of the tapered fiber bundle (right).
In the final step the output fiber is spliced onto the fiber bundle. To get a splice without losses it is elementary that the beam parameter product of the outgoing fiber is higher than that of the incoming fiber, expressed in
A 125 µm fiber with a numerical aperture of 0.48 is used. The resulting beam parameter product of the fiber bundle equals $BP{P_{in}}$ = 19.3 mm mrad, whereas the beam parameter product of the output fiber equals $BP{P_{out}}$ = 62.6 mm mrad. Therefore the light should be guided within the output fiber.3. Optical tunneling
Within an optical fiber the total internal reflection between fiber core and fiber cladding is responsible for the light guidance. When light hits the interface at an angle higher exceeding the critical angle ${\vartheta _T}$ total internal reflection occurs. According to Maxwell’s equations another effect called the evanescent field occurs. Due to the thin cladding of the used input fiber, power losses may occur because of this effect.
The evanescent field extends into the fiber cladding and may extend even farther outside the cladding, too. In our case this means that a fraction of the optical power within one fiber can be absorbed or coupled into the surrounding capillary or another fiber, if the penetration depth is in the order of the cladding thickness. The penetration depth dp is defined by [9]
Due to the taper process described in section 2, the cladding of the tapered fibers has a thickness of roughly 4 µm. Therefore this effect might generate power losses. This effect will be analyzed in the following.
Figure 4 (left) shows the penetration depth with respect to the numerical aperture of the light propagating within the used fiber. For small numerical apertures the penetration depth is 3 to 4 times smaller than the cladding thickness. In Fig. 4 (right) the expected percentage loss per total reflection with respect to the cladding thickness for different numerical apertures of the light is plotted. This shows that as long as the numerical aperture of the light is small enough the light should be guided within the fiber’s core. For higher values light couples into the capillary, leading to power losses due to the small overlap between the capillary and the output fiber.
Fig. 4. Penetration depth with respect to the numerical aperture of light (left) and expected percentage losses per total reflection with respect to the cladding thickness (right).
For the quantification of this model the expected losses are calculated based on the measured gaussian-like far field distribution with a numerical aperture of 0.22. The fiber in this model has a cladding thickness of 4 µm and a length of 3 cm, which is a typical taper length. The expected number of total internal reflections and losses due to the optical tunneling effect is calculated with respect to the numerical aperture of the light. Following these calculations roughly 5% of the light is not guided within the fiber’s core anymore.
4. Transmittance of the fiber bundle
Based on these considerations fiber bundles are produced and characterized. For the characterization of the fiber bundle the transmission has to be determined. Therefore the optical input power and the output power of the fiber bundle are compared to each other.
Due to the limited availability of the BRIDLE modules, alternative diode laser beam sources are used for the following characterization. These modules have a reduced optical power of several 100 mW but the same geometrical beam properties, i.e. 35 µm core and a numerical aperture of 0.2 ± 0.01, according to measurements of the light’s numerical aperture by using 90/10 knife-edge techniques [10]. Here, the near field intensity profile is approximated as a top-hat function. It is observed that a degree of transmission over 98 ± 1.5% can be achieved, even though the numerical aperture of the light is higher than the transmission limit of 95% described in section 2.
The near field intensity profile of the fiber bundle facet is shown in Fig. 5. Only six input ports are measured to detect the scattered light in the reverse direction. More than 95% of the light is guided within the fibers, roughly 5% couples into the capillary, which corresponds to the expected losses within chapter 3. This effect might occur because of the enlargement of the numerical aperture of the light, due to the taper process.
5. Transmittance of the combiner
For the measurement of the transmittance for the fiber combiner the transmitted power of the output fiber has to be measured additionally and compared to the fiber bundle. The output fiber is a 125 µm fiber with a numerical aperture of 0.48. By using this fiber no power losses are expected, as the output fiber’s BPP is larger than that of the radiation in the tapered bundle.
The measurement of the transmission of the combiner is performed by comparing the transmitted power of the fiber bundle to the transmitted power of the combiner. Measured data of different fibers are shown in Fig. 6 (left). A transmission over 95 ± 1.5% is achieved. The light within the capillary can be considered as a power loss. Because of the capillary’s small overlap with the output fiber’s core. Therefore only a small fraction of the light within the capillary can be coupled into the output fiber’s core. Considering the measured near field intensity of the fiber bundle a power loss of 2.8% is expected. This results in a possible transmission of 93 ± 2% for the whole fiber combiner.
Fig. 6. Measurement of the transmission of a fiber bundle for a numerical aperture of 0.2 (left) and near field intensity profile of the combiner output fiber (right).
The far-field intensity profile corresponds to a two-dimensional, rotationally symmetrical distribution. The numerical aperture of this light is 0.27 ± 0.01, as is the numerical aperture of the radiation in the fiber bundle. Therefore the splice process is not expected to decrease the beam quality.
In Fig. 6 (right) the near field intensity profile of the fiber facet is shown. In comparison to Fig. 5 the intensity profile has been homogenized by the output fiber.
6. Summary
We presented the design and construction of a 7 × 1 35/50 µm fiber with thin cladding into a 125 µm core diameter fiber combiner. Based on the measurements, we estimate a total transmission over 93%.
For the output fiber splice a transmission over 95% is achieved by comparing the transmission between the fiber bundle and the output fiber. For the fabrication a 125 µm fiber is spliced onto the fiber bundle. This results in an overall transmission of 93% for the whole fiber combiner. The target of further improvements is the utilization of a 90 µm core diameter fiber with a numerical aperture of 0.28 as the output fiber, which corresponds to a beam parameter product of 25.5 mm mrad, to maintain a higher beam brightness.
Analysis according to the optical tunneling effect has shown that no power losses should occur due to this effect as long as the numerical aperture of the light is smaller than 0.16 which would represent a loss of less than 0.1% per reflection.
In further experiments the BRIDLE laser module which emits light with a numerical aperture of 0.12 shall be used to achieve a much better beam brightness, by using other fiber combiner output fibers. It is expected that less radiation propagates into the cladding and the capillary. Therefore a higher transmission, which should only depend on the splice quality, may be reached. Additional analysis regarding to the conservation of the beam parameter product in tapered fibers can be done to get information about the theoretically possible beam brilliance.
The presented fiber combiner may for example be used for direct diode laser metal processing or the pumping of rare-earth-doped fibers. This has many benefits due to the use of cost effective and high efficiency diode lasers.
Funding
FP7 Information and Communication Technologies (ICT) (314719).
Acknowledgments
This work was funded in part by the European Commission within the BRIDLE project in the seventh framework program under grant number 314719 (www.bridle.eu).
References
1. “BRIDLE” http://www.bridle.eu.
2. U. Witte, M. Traub, A. Di Meo, M. Hamann, D. Rubel, S. Hengesbach, and D. Hoffmann, “Compact 35 μm Fiber Coupled Diode Laser Module Based on Dense Multiplexing,” Proc. SPIE 9733, 97330H (2016). [CrossRef]
3. Q. Xiao, H. Ren, X. Chen, P. Yan, and M. Gong, “Tapered Fiber Bundle 7 × 1 End-Pumping Coupler Capable of High Power CW Operation,” IEEE Photon. Technol. Lett. 25(24), 2442–2445 (2013). [CrossRef]
4. G. Zhu, “High Efficiency Pump Combiner Fabricated by CO2 Laser Splicing System,” Proc. SPIE 10513, 105131C (2018). [CrossRef]
5. M. Plötner, O. de Vries, T. Schreiber, R. Eberhardt, and A. Tünnermann, “High power incoherent beam combining by an all-glass 7:1 fiber coupler with high beam quality,” Advanced Solid State Lasers (2014).
6. D. Noordegraaf, M. D. Maack, P. M. W. Skovgaard, J. Johansen, F. Becker, S. Bekle, M. Blomqvist, and J. Lægsgaard, “All-fiber 7 × 1 signal combiner for incoherent laser beam combining,” Proc. SPIE 7914, 79142L (2011). [CrossRef]
7. H. Zhou, Z. Chen, X. Zhou, J. Hou, and J. Chen, “All-fiber 7 × 1 signal combiner for high power fiber lasers,” Appl. Opt. 54(11), 3090–3094 (2015). [CrossRef]
8. B. Wang and E. Mies, “Review of Fabrication Techniques for Fused Fiber Components for Fiber Lasers,” in Proceedings of SPIE Photonic West ‘09, Fiber Lasers VI: Technology, Systems, and Applications, (2009).
9. D. Meschede, Optik, Licht und Laser (Teubner, 2008).
10. A. E. Siegman, IEEE Fellow, M. W. Sasnett, and T. F. Johnston Jr., “Choice of Clip Levels for Beam Width Measurements,” IEEE J. Quantum Electron. 27(4), 1098–1104 (1991). [CrossRef]
