Before a BPM simulation can begin, the optical field must be known for every point on the input plane. This usually requires a mode solver. OptiBPM will call one if set that way. The above message is sent when that modesolving fails, either because of incorrect settings in the modesolver or because the mode does not exist. Using Gaussian fields for the input plane will help, but you’ll have to supply a value for the reference index – OptiBPM will call the mode solver to find that parameter if you don’t specify it.
Yes, that’s right. Segment is not the same as period. Segment means only that within the segment the grating is taken to be uniform – same period, same amplitude. OptiGrating calculates non-uniform gratings by connecting many short uniform ones.
Profile0 is a plot of the grating shape: the refractive index variation with Z over one grating period. That explains why 0 < z < .5338 for that plot. OptiGrating will display that curve in its own interface if you right click on the Profile tab and select Shape…
Plots 1, 2, and 3 display local period, apodization, and average refractive index. Those are defined over the length of the grating, 0 < z < 50,000. The number of points for display comes from the number of segments specified in the Grating Definition dialog box. The file will show the value at the beginning and the end of each segment: two values per segment. Therefore if there are 100 segments, there will be 200 points.
Unfortunately User Function Profile returns only real numbers. There is no way to add loss as a calculated imaginary part of refractive index. On the other hand, putting an additive User Function Profile on a lossy dielectric is a clever trick. That should be a good workaround for most cases.
In the usual case, modes in a waveguide are independent, meaning that the light in one mode will propagate without affecting the light in another mode. The case where light goes from propagation in one mode to propagation in another is exceptional, usually due to some special circumstances, such as a periodic disturbance or possibly an environmental influence. Therefore mode conversion is sometimes the basis for a sensor, making it interesting for practical applications.
I’m afraid OptiFiber will not help with this simulation. OptiFiber uses modal analysis, but this method won’t work well for this sample. Your sample fibre has a radius of 980 µm, wavelength .65 µm, core and cladding refractive index 1.49 and .13725. The V number is therefore 5,494. The number of modes for high V number is estimated from 0.5*V^2, which is more than 15 million. I’m afraid that is too many modes! OptiFiber can perform multimode analysis. It will work reliably even if there are dozens of modes, but 15 million is just too many!
You can probably get meaningful results from a propagation kind of simulation, like BPM, but modal analysis is simply the wrong method for this problem.
The Beam Propagation Method (BPM) is a paraxial method. This means it applies only in the case where there is an optical axis. The U bend is not such a case. On the other hand, accurate results have been demonstrated when BPM is used together with a conformal mapping. (see, for example, IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. 43, NO. 10, pg 899-909 OCTOBER 2007)
OptiBPM has a conformal mapping region. Please use the conformal mapping region of OptiBPM to calculate the details of optical propagation in a U bend.