Multimode polymer optical waveguide evaluations by using a combination of multimode optical fibers (MMF) and Light Emitting Diode (LED) give very often inconsistent experimental result. It is due to the over-filled properties of the launch light created by this configuration. We propose an optimized simple launch configuration to overcome the problems by using similar configuration. The optimized launch configuration creates an optimal-filled launch light where its encircled flux (EF) profile satisfies the EF template provided by the International Electrotechnical Commission (IEC). We have used the standard optical fibers and optical fiber adaptors to create the optimized simple launch configuration with high stability. Therefore, our proposed launch configuration is suitable for realizing a cost effective optimized launch configuration for standardization of multimode polymer optical waveguide evaluations. We demonstrated reliability of the proposed launch configuration by examining the reproducibility of the insertion loss (IL) measurements. We have used two waveguide samples that have different characteristics for this purpose. It was found that the insertion loss (IL) measurements of the two samples are consistent with the largest variation of less than 5%. This variation is better than the proposed value given by the IEC that is 10%.
© 2010 OSA
Coherent light sources such as Laser Diode (LD) and Vertical Cavity Surface Emitting Laser (VCSEL) in combination with multimode optical fibers (MMF) are often used in practical applications including as input launch in multimode polymer optical waveguide evaluations. On the other hand, it is well known that the transmission characteristics of MMF strongly depend on the launch conditions. Combination of coherent light sources and MMF as input launch usually provides consistent experimental results. Unfortunately, this launch configuration gives underestimate experimental results. It is due to the under-filled launch light created by the coherent light sources which produces misleading and over optimistic evaluation results . Applying a combination of incoherent light source such as LED and MMF as input launch gives over-filled launch light. It results over-estimate, unrepeatable and unreliable loss readings . The IEC proposed several techniques written in several categories of the IEC recommendations to overcome the problems by creating an optimal-filled launch light.
There are several previous reports have been presented on the EF category by others up to now [2–4]. The methods, however, proposed complex configurations consisting of the special fibers and fiber adaptors. Recently, we reported a simple launch configuration by implementing the Step Index (SI)-Graded Index (GI)-Step Index (SI) fibers to create the SGS mode scrambler. The SGS launch configuration with incoherent LED light source provided very stable mode distributions and became a promising candidate for simple evaluation of multimode waveguides, especially multimode polymer optical waveguides .
In the current report, we have chosen the EF category (category A1) that is one of the IEC recommendations to achieve consistent and reliable experimental results. We implemented the IEC recommendation about fiber-optic communication subsystem test procedure-Part 4.1: Installed cable plant: Multimode attenuation measurement . The document recommends so called control radii for the EF profile shown in Fig. 1 . The EF targets and its tolerance limit are only quantified by 4 discrete radii denoted by the triangles up (red) and the triangle down (blue) at the radii of 10, 15, 20 and 22 μm. The values of these control radii are given in a table that is inserted in Fig. 1. We used a standard experimental setup including the use of standard optical fibers and standard optical fiber adaptors for our experiments. Therefore, the proposed launch configuration is useful for realizing a cost effective launch condition. In order to demonstrate the reliability of the proposed launch configuration, we examined the reproducibility of the insertion loss (IL) measurement results by using two various samples with different core sizes, waveguides materials and substrate materials.
2.1 Near-Field Pattern (NFP) Experiment
We have used standard Near-Field Pattern (NFP) experimental setup shown schematically in Figure 2 to capture the NFP created by the launch GI fiber. We used a LED with a central wavelength of about 850 nm as a light source. In order to have similar condition between the NFP experiment and the IL experiment that would be performed latter, in the current NFP experiment the LED was connected to the 1 X 2 coupler.
However, only one branch of the coupler has been used in the NFP experiment. This branch was connected to the mode scrambler that was finally connected to the launch GI fiber. In Fig. 2 we have skipped the 1 X 2 coupler to avoid misunderstanding.
The mode scrambler consists of a combination of the standard GI and SI fibers. These fibers have the same core size of 50 μm and a NA of 0.22. They are connected by using the standard optical fiber adaptors to create the GSG mode scrambler.
The length of each fiber is 3 meters. The launch fiber is a standard GI fiber that has a length of 2 meter, a core size of 50 μm and a NA of 0.22 manufactured by Tonichikyosan. We have used a standard charge-coupled device (CCD) camera to capture the created NFP. The launch GI fiber and the CCD camera have been put on the similar xyz stages that can be adjusted in the xyz directions. Moreover, the stage of the launch GI fiber could also be rotated in a limited area of horizontal and vertical directions to enable better alignment.
The launch GI fiber produces the NFP. The NFP is projected to the CCD camera by an objective lens together with the NFP optical system. The CCD camera has been connected to the computer to enable image processing. Figure 3a shows an example of the NFP created by an optimized simple launch configuration captured in the NFP experiment. Meanwhile the NFP obtained by using a coherent LD is given in Fig. 3b where the speckles on the NFP indicate unstable characteristics of the LD. Therefore, incoherent light source is much more appropriate for multimode waveguide evaluations. The EF is defined as a fraction of the total amount of optical power from the centroid to a ring radius, r divided by the total optical power as formulated in Eq. (1) [2,3,6,7]
In order to obtain the EF profile, the NFP captured by the experiment is processed by calibrated commercial image processing software to obtain raw data. This raw data is two-dimensional Near-Field intensity distributions. We calculated the EF by collapsing the two-dimensional Near-Field intensity distributions into one-dimensional radial function [2,6,7]. Further, we implemented the EF definition in Eq. (1) to obtain the EF profile.
2.2 Insertion loss (IL) Experiments
We have modified the input part of the experimental setup in Fig. 2 for the insertion loss (IL) experiment. The modified experimental set up is given in Fig. 4 . We have used a LED with a central wavelength of 850 nm as a light source.
We used a standard 1 X 2 coupler to divide the power of the LED. One branch of the coupler goes to the reference port of the optical power meter. The other branch goes to the optimized GSG mode scrambler. The last optical fiber adaptor of the GSG mode scrambler was connected to the launch GI fiber that acts as an input fiber.
The output fiber is a silica core with hard polymer cladding SI optical fiber with a core size of 200 μm and a NA of 0.48. This large core size and NA fiber was intended to enable collecting all the light from the output waveguide. We put the input and output fibers on the identical 5-D stages. These stages could be adjusted in xyz directions and be rotated in a limited area of horizontal and vertical directions to have better alignment of the fibers to the waveguides. We put the device under test (DUT) on a table that can be rotated. We used appropriate matching oils in our IL experiments to minimize the coupling loss. Suppose the light injected to the input waveguide has a power of Pin (dBm) and the light come out from the output waveguide and captured by the output fiber has a power of Pout (dBm) then the insertion loss (IL in dB) is defined by
The experimental setup shown in Fig. 4 has been used to measure the IL of two chips that consist of only straight waveguides with different core sizes, waveguides materials, and substrate materials.
An acrylate polymer waveguide sample (chip A, in Table 1 ) that has a core size of about 50 x 50 μm2 and a silicone polymer waveguide sample (chip B, in Table 1) that has a core size of about 40 x 40 μm2 have been used. We manufactured the both chips by implementing direct photo-patterning. We measured the three channels of two samples twice and calculated the insertion loss (IL) by using Eq. (2) to demonstrate the reliability of the created launch configuration. We compared the two IL measurement results and calculated the variations of the insertion loss (IL) values to examine the reproducibility of the IL experimental results. The IL variation (ΔIL) was obtained by taking the absolute value of the difference of the IL measurement results.
3. Results and discussions
In our first experiment, we have implemented the SM fiber as an input launch. This configuration gives stable output  and shows good repeatability of the IL measurement results. Unfortunately, this input launch gives underestimate IL values due to the under-filled characteristics of the SM fiber output into the waveguides. We have used the waveguide samples structured by using molding technique shown in Table 2 for this purpose. Table 2 shows large differences of the IL measurement data obtained by using the GSG mode scrambler and the SM fiber, especially for larger core size of multimode waveguides. These results indicate that scattering loss factor by surface roughness on the boundary between core and cladding is not taken into account for SM input launch. Further, we have used multimode fibers to create an appropriate input launch configuration. Unfortunately, multimode fiber creates over-filled launch condition . It seems that we should treat the launch configuration such as filtering the higher order modes inside the optical fibers to obtain the appropriate input launch condition.
The multimode SI fiber has an abrupt refractive index change between the core and the cladding . The light that is guided by the total internal reflection should satisfy the critical angle. The light with the angle of incident at the core-cladding interfaces is larger than the critical angle will penetrate the cladding and remove several (higher order) modes [8,9]. Moreover, the modes in the SI fiber experience different optical path length and causes spatial broadening [8,10]. On the other hand, the multimode GI fiber has a gradually refractive index change of the core from the fiber axis to the cladding. Therefore, the modes in the GI fiber experience similar optical path length and there is almost no spatial broadening [8,10]. This description shows that filtering the higher order modes can be done more efficiently by bending the SI fiber than by bending the GI fiber. However, we will probably need the GI fiber for beam shape recovery . Therefore, it opens the possibility to apply combinations of SI and GI fibers to obtain the appropriate input launch condition.
Figure 5 (top) and (bottom) show experimental results of the intensity evolution inside the GI and SI fibers for the GSG mode scrambler with the bending diameter of 6 mm, 6 mm, 6 mm and 25 mm, 25 mm, 25 mm respectively that confirms the theoretical explanation given previously. Meanwhile, we did not do any treatments to the GI launch fiber. Figure 5 (top, 1) shows the NFP when we use only one GI fiber that has a bend diameter of 6 mm. The NFP diameter becomes larger when we attached the additional SI fiber to the previous GI fiber shown by Fig. 5 (top, 2). It is due to the small bending diameter that is 6 mm then there are significant higher orders modes penetrate the cladding. Figure 5 (top, 3) shows the similar NFP diameter to the NFP diameter given in Fig. 5 (top, 1). The beam shape has been recovered by the second GI fiber with bending diameter of 6 mm attached to the previous SI fiber. Figure 5 (bottom) shows the GSG mode scrambler with the bending diameter of 25 mm, 25 mm, 25 mm and using similar configuration to the Fig. 5 (top). We obtained the similar NFP diameter for all configurations. It means that the bending diameter of 25 mm is not sufficient to filter the higher order modes significantly, especially in the SI fiber. Besides, modes selection happens in every optical fiber adaptors due to the mode mismatch and misalignment.
Firstly, we have adopted the SGS mode scrambler that has been investigated previously . We bent the fibers with the bending diameter of 25 mm, 25 mm, 25 mm to control the launch light. Although the SGS mode scrambler could provide very stable mode distribution , the EF profile of this configuration did not satisfy the EF template provided by the IEC. It is probably due to the higher order mode filtering in the last SI fiber where the higher order mode will penetrate the cladding and create larger beam diameter. Therefore, we obtained the over-filled launch condition shown in Fig. 6 . Later, we have used the GSG mode scrambler with the bending diameter of 25 mm, 25 mm, 25 mm to create a stable and optimal-filled launch light. We compared the EF profile obtained by using the SGS and the GSG mode scramblers to the EF template.
Figure 6 shows that the EF profile obtained by the GSG mode scrambler that is much more promising to satisfy the EF template than the EF profile obtained by the SGS mode scrambler. It is probably due to the beam shape recovery that has been done by the last GI fiber in the GSG mode scrambler. Based on this fact, we then continued to work with the GSG mode scrambler as an alternative. However, the appropriate treatment should be applied such that the EF profile satisfies the EF template. For this purpose, the GI and SI fibers of the GSG mode scrambler have been rolled on the various rods diameter to manage the higher order modes inside the optical fibers. We optimized the excited light by controlling their rods diameter.
Figure 7 shows the EF profile obtained by using several rods diameter combinations for optimizing the launch configurations. The triangle up (▲) expresses the lower limit and the triangle down (▼) expresses the upper limit of the EF, respectively provided by the IEC. Rods diameter given in the legend of Fig. 7 represents the bending diameter of the GSG fibers. The first two launch configurations with the bending diameter of the GSG mode scrambler of 25 mm, 25 mm, 25 mm and 6 mm, 6 mm, 6 mm respectively give the EF profile shown in the black dash and green dot curves. They show the values below the lower limit of the EF at the control radii of 20 μm and 22 μm. The EF profile given in the red dash-dot curve for the bending diameter of the GSG mode scrambler of 6 mm, 6 mm, 5 mm has the EF value below the lower limit of the EF at the control radius of 22 μm. The EF profile given in the brown full-line for the bending diameter of the GSG mode scrambler of 6 mm, 6 mm, 4 mm confirmed the EF values located between the upper and lower limit of the EF template. Only at the control radius of 22 μm, the EF value located on the lower limit of the EF template. The inset of Fig. 7 enlarges the EF profile at the control radii of 20 μm and 22 μm.
We demonstrated reproducibility of the experimental results by measuring the IL of the various waveguide samples. We applied the optimized launch configuration that satisfies the EF template as given by brown full-line in Fig. 7. Table 1 gives a comparison between two insertion loss (IL) measurement results of two samples. It was found from the experiments that the insertion loss (IL) measurement results by implementing the proposed launch configuration show good reproducibility. The IL variation is between 0.005 dB and 0.016 dB or less than 5%. This variation value is better than the proposed value by the IEC in  that is 10%.
We have succeeded in creating the optimized simple launch configuration that produced optimal-filled launch light where its EF profile satisfies the EF template provided by the IEC. We demonstrated reliability of the optimized launch configuration by examining reproducibility of the insertion loss (IL) results. We have used two waveguide samples with different properties for the IL experiments. It was found from the experiments that the IL results show good reproducibility. The IL variation of better than 5% has been achieved. The proposed launch configuration is also useful for realizing a stable and cost effective input launch condition for multimode optical waveguide evaluation.
This work has been supported by the Ministry of Economy, Trade and Industry (METI) Japan in Standardization Project for Polymer Optical Waveguide Evaluation and the Japan Science and Technology Agency (JST) program for Strategic Promotion of Innovative Research and Development (S-Innovation).
References and links
1. V. C. Hicks, “Multimode fiber certification: Light source launch conditions and encircled flux standard,” EXFO Application Notes 196.
2. B. Lane, A. Brunsting, and R. Pimpinella, “Insertion Loss Performance Testing of 10 Gb/s Fiber Patch Cords for High-Speed Networks,” Proc. 57th IWCS, November 2008.
3. J. B. Schlager, M. J. Hackert, P. Pepeljugoski, and J. Gwinn, “Measurement for Enhanced Bandwidth Performance Over 62.5 μm Multimode Fiber in Short-Wavelength Local Area Network,” J. Lightwave Technol. 21(5), 1276–1285 (2003). [CrossRef]
4. A. G. Hallam, D. A. Robinson, and I. Bennion, “Mode Control for Emerging Link Performance Standards,” IET Optoelectron. 2(5), 175–181 (2008). [CrossRef]
5. O. Sugihara, K. Takayama, J. S. Selvan, S. Shibata, T. Kaino, K. Hirano, T. Ushiwata, M. Morimoto, S. Yagi, T. Makino, Y. Matsui, and K. Tajiri, “Polymer Optical Waveguide Chip: New Tool for Simple Evaluation of Optical Parameters of Waveguide Elements,” IEEE Photon. Technol. Lett. 22(10), 703–705 (2010). [CrossRef]
6. IEC 61280–4-1 Ed. 2.0 (2009–06–10, 86C/ 820/ Publication Issued), Fibre-optic communication subsystem test procedures-Part 4–1: Installed cable plant-Multimode attenuation measurement.
7. S. Takahashi, T. Noda, and Y. Koike, “Restricted mode launch condition for graded index plastic optical fibers,” the 17th International Conference on Plastic Optical Fibers (POF 2008), pp. 124, Santa Clara, CA, USA.
8. H. J. R. Dutton, Understanding Optical Communications (IBM 1998), Chap. 2.
9. R.G. Hunsperger, Integrated Optics, Theory and Technology (Springer 1995), Chap. 5.
10. Y. Koike, “Status of Photonics Polymers for Fiber to the Display”, slide presentation in Finnish-Japanese Workshop on Functional Materials, Espoo and Helsinki, Findland, May 25, 2009.