These tools use completely different numerical methods and numerical error manifests itself differently in each method.  For this reason it may not be possible to get an exact match.  The only thing you can do is try to use extremely tight convergence conditions.  In theory, since both are rigorous methods both will eventually converge to the same answer.  In practice, this may not be possible due to computer hardware limitations.

There are times where we wish to first calculate the band structure to design the basic unit cell and then simulate a device composed of that unit cell.  In cases where we need an exact match, we will calculate the band structure using the same numerical technique that we use to perform the simulation.  This usually leads to slower band calculations, but we can get exact matches.

Hope this helps!!!

With my experience, its better to use Eigen mode solver for Photonics crystal. I used Lumerical FDTD and compare the results with CST eigen mode solver. I found that CST eigen mode solver provides more accurate results because of eigen mode solver. In my opinion, eigen mode solver will lead to more accurate results even with normal accuracy setting. In FDTD Luemrical, we used several clouds source randomly distributed, sometimes no of sources and mesh may not be enough and lead to accuracy problem. Please note that I have no idea what methodology is used optiFDTD. I guess it must be similar to Lumerical FDTD.  

Thank you Prof. Rumpf for your response. In optiFDTD, there are options for both PWE band solver, and optifdtd analyzer. I'm using the same numerical tool, but the bandgap results are not matching.

Hi Vishal,

I've compared the bandgap results using both optiFDTD and Lumerical FDTD. The results are same. But the moment i'm trying with the 'PWE band solver' its showing something different. I've even tried with changing the no. of bands, mesh and super cells, but unable to point out my mistake.

@Swagata, I believe If  two simulator are giving same results than one can trust the results. Do you have specific purpose to do the simulation using PWE band solver or you just want to study both methodology ? Please upload reference which article you are trying, maybe one can check and give you more reasonable advice. I suppose you must be doing unit cell simulation to plot band-diagram, Am I right ? 

I also suggest duplicating a known result from the literature.  Which ever matches is likely the correct method.  Then you just have to figure out what is going wrong with the other method.

In general, I do not recommend calculating bands using FDTD.  That method has a lot of difficulty distinguishing bands when they are close and can even sometimes miss bands.