Fig. 2 compares the estimated distributions of layers doped with GeO at two wavelengths, with a nominal width of 65438 0.5 μ m and doping compositions of 3.5%, 5.0% and 7.5%. Fig. 2 compares the estimated (refractive index) distribution of the germanium dioxide doped time-sensitive layer at two wavelengths, and its nominal width is 65438±05 μm m.. While the doping compositions are 3.5, 5.0 and 7.5. There is evidence that in all cases, the refractive index transition from Shi Ying substrate to layer is not a simple step. There is a gradual change in the layer doped with ge02, and its refractive index increases with the distance from the interface. Obviously, in all cases, the transition of refractive index from substrate to doped layer is not a simple step. On the contrary, the doped layer of germanium dioxide has a gradual change, and the refractive index increases gradually from the interface. At this time, it is impossible to evaluate whether this is only the result of the stress field caused by the change of lattice parameters or whether it is the result of the conditions in the growth chamber [I]. At this time, it is impossible to estimate whether this is only the result of the stress field caused by the change of lattice parameters, or whether this is the result of the growth chamber conditions. In any case, it seems that waves are made in every case, but the refractive index distribution is not a simple step function. However, it seems that waveguides are made in every case, and the refractive index distribution is not just a step function. _-

In order to compare with a more familiar example, this analysis technique is applied to the proton-exchanged lithium niobate waveguide produced by GEC Marconi Company. In order to compare with similar examples, this analysis technique is applied to the proton exchange lithium niobate waveguide produced by GEC Marconi Company. Fig. 3(a) shows the experimental dark mode position measured at 0.488 pm. The experimental point error given by triangle is within-f 0.00 0 1. Of the measured 17 dark lines, only six appear as clear real modes, and the rest 1 1 are broadband modes called "base", but they are still valuable in calculating contours. Fig. 3(a) shows the dark mode position of the experiment measured at 0.488 pm. The triangle gives some experimental points and the error is within 0.05. Among the measured dark lines of 17, only 6 lines show steep actual patterns, and the rest 1 1 are wide strips, which are called "base" patterns, but they are still valuable in calculating the distribution. Although the six real patterns mean an almost square well profile, since the data in Figure 3(a) is on a straight line, using all 17 patterns enables a more complete analysis of exponential changes. Although the six actual modes mean a nearly square potential well distribution, because the data in Figure 3(a) are all on a straight line, the change of refractive index can be analyzed more completely by using all 17 modes. Note that the distribution form of LiNbO 3:H is different from that of epitaxial Shi Ying layer (Figure 1). Figure 3(b). The calculated refractive index distribution of three optional functions in the form of refractive index changing with depth is given Please note that the refractive index distribution of LiNbO _ 3: H is different from that of the epitaxial growth time layer (Figure 1). Fig. 3(b) shows the refractive index distribution calculated by three substitution functions in the form of refractive index changing with depth. The choice is simple exponential change of interface layer, exponential plus inclined base and free contour. The choice is simple exponential refractive index change of interface layer, exponential plus inclined substrate and free distribution. Through the minimum evaluation of the minimum square error in the calculation, all three show the full reproduction of the mode. All three options show that the model can be correctly represented, just as it is estimated by minimizing the least square error in calculation. Because the contour is close to the step function, in the analysis using only the "real" (that is, bright) mode of si x, the slight change among them will not be obvious. However, the calculation stage of refractive index distribution caused by protocol exchange includes a step gradient at the boundary. Because the distribution is close to the step function, in the analysis of only six "real" (that is, bright) modes, the slight change between them will not be obvious. However, the calculation emphasizes that the refractive index distribution caused by proton exchange contains a steep gradient at the boundary.