PCB integrated waveguides Launching light into highly multi mode structures cd.

3. EXPERIMENTS

The introduced set-up was used to measure the loss at different positions of the coupling fiber. Figure 3 shows the absolute detected power (integrated over entire detecting endface) for each position indicated by indices of the measurement matrix.



This 2 dimensional scan could be understood as a superposition of coupling and launching conditions. When dividing by the reference power a loss graph could be derived which in principle would look the same.

3.1. SET-UP CHARACTERIZATION

In order to evaluate the accuracy and repeatability of the aforementioned measurements the deviations between consecutive scans were analyzed. In figure 4 an example of a normalized differential plot of scans with MMF can be seen. Remarkable is that the biggest relative failures occur at the boundary of the core area. As the



repeatability of the stages lies in the range of the spatial sampling a misalignment is expected. This leads to misreadings and therefore higher deviations at the gradient. For coupling with the SMF@850 nm it depicts differently. The active area exhibits a floor besides the already mentioned effect at the margin. It is assumable that this statistical noise is due to mode depending effects usually named as modal noise. Figure 5 summarizes the described effects as relative occurrence in certain failure classes with the respective cumulative percentage. For coupling with MMF the relative failure is mostly around 2% and maximum of 10% while the distribution for SMF is wider and ranges up to 20%. One could assume a superposition of two distributions. One possibly caused by the positioning stages the other one by modal noise. It can be concluded that launching with a MMF results in measurements with high repeatability that cancels out mode depending deviations and is slightly deteriorated by the repeatability of the stages. Relief to this issue would be the usage of piezo actuators instead of the DC motors to obtain submicron resolution.



3.2. LAUNCHING CONDITION

By means of the spatial scanning we can observe how the coupling efficiency changes due to the different launching positions of the coupling fiber and how that situation changes for various source fibers (SMF@850 nm, MMF50). By using the setup shown in figure 2 a spatial scan of coupling conditions was performed. The analysis of such a scan can be lead into three directions:
- a: axial, lateral characterization,
- b: surface characterization,
- c: spatial behavior.

As figure 6 illustrates, coupling in the middle position (in the smallest possible distance from surface of waveguide) does not guarantee the biggest coupling efficiency. The efficiency can become greater with increasing distance because the numerical aperture of SM fiber assures that the much wider area of coupled surface is illuminated. the numerical aperture of SM fiber assures that the much wider area of coupled surface is illuminated.



a: axial and lateral displacement (figure 6 and 7) was realized for the following conditions: two kinds of coupling fibers: SMF@850 nm wavelength and standard MMF (50 μm core); start distance between integrated waveguide and coupling fiber 27 μm; measurement step 1 μm, source wavelength 850 nm. The procedure of this part of the experiment included: detecting the position for minimal coupling losses and line scans for characterization of the coupling efficiency in function of the axial and lateral displacement. For comparison the same characteristic from middle of waveguides cross-section has been taken. The axial displacement from the start position was 1 mm. It is important to notice that for MMF there is a minor difference (~10 μm) between middle position and the position of maximal coupling efficiency. That means that SMF is more sensitive to imperfections of waveguides coupling faces for example dust or poor grinding while in case of MMF it is integrated as with a low pass filter. Therefore surrounding places with much better transmission parameters can be taken in the consideration when explaining that aspect.
b: Surface characterization allows to obtain general information of the quality of the waveguide surface for a specified coupled source. Figure 3 shows the waveguide’s coupling side scanned with a MMF (core 50 μm). That kind of surface scan can be used for example to trace point of maximum coupling efficiency, for quick active alignment or near field scans of detecting endface from the waveguide.



c: With the scanning set-up (figure 2) it is also possible to achieve a full 3D spatial distribution of coupling conditions. Figure 8 presents the visualization of volume data taken during the 3D measurement. It allows observing the general behaviour of coupling conditions and of course can be also performed at the detecting side.



3.3. COUPLING CONDITION

Looking at an ideal endface the transmitted power for different fiber positions would solely depend on the launching condition. Real waveguides though reveal imperfect coupling endfaces. Herewith they are subject to deviations due to the coupling conditions. The surface quality, index matching etc. impacts this condition. Figure 9 demonstrates a backlit image of a waveguide endface with residual scratches from the grinding process. For the ease of comparison the image was already mirrored, because of the perspective from the detector side. This endface has been measured once conventionally as said above and a second time when it was immersed into water. This immersion liquid acts both as refractive index matching and smoothes out the surface roughness. Since this reduces reflection losses it enhances the coupling efficiency. Consequently the normalized absolute deviation of the two matrices displays as in figure 10. This visualization enables one to characterize the surface property and its influence on the coupling. The scratches (compare figure 9) yield significantly increased deviation.



This endface has been measured once conventionally as said above and a second time when it was immersed into water. This immersion liquid acts both as refractive index matching and smoothes out the surface roughness. Since this reduces reflection losses it enhances the coupling efficiency. Consequently the normalized absolute deviation of the two matrices displays as in figure 10. This visualization enables one to characterize the surface property and its influence on the coupling. The scratches (compare figure 9) yield significantly increased deviation.



3.4. NEAR FIELD SCAN

Focusing on multi mode structures requires to handle with effects generated by mutual interactions between modes also influence of environment for those effects cannot be underestimated. Modal noise was taken into concern [2]. In each cross-section of wave guiding structure (and in the endface) a light pattern can be observed. Modes with different delays are interfering with each other. Due to the phase shifts modes add constructively or destructively. That is the reason why the speckle pattern can be detected. The effect is strongly dependent on the spectral width of the light source. The shorter time coherence of the laser source is the bigger is the probability of mode interference to occur. Therefore when the spectral width of the source is narrowed the effect of speckle pattern becomes more fundamental [4]. Because of interference character of the effect the speckle pattern is very sensitive to changes of wavelength (laser chirp), temperature, micro bendings of the fiber and all those effects leading to phase shift between modes. Speckle pattern is the cause of modal noise in all the elements in optical transmission line that are providing spatial filtration for example connectors, mode mixers, couplers etc. One part of the investigation was devoted to the described effect. The speckle pattern at the endface of the integrated waveguide for two coupling fibers (SMF@850 nm, MMF) was observed (figure 11a,b) with a CCD camera. The dependence of speckle distribution from the numerical aperture (NA) of the coupling fiber can be noticed. For SMF@850 bright speckles occur around the center of the waveguide.



Coupled with MMF speckles are equally distributed within the whole cross-section of the waveguide. It was attempted to measure the speckle pattern by a near field scan performed with a tapered fiber [3]. The tapered fiber with a submicron active area guided the detected signal to the integration sphere. The scan was provided with the smallest resolution of 1 μm, in the distance of 10 μm from the guiding structure. The result of the scan is presented in figure 12. The characteristic granulation as a result of mode interference (pattern) is clearly to be seen. That kind of distribution in case of misalignment causes power fluctuation, which can manifest with already mentioned modal noise.