shifted frequency in band diagram calculation

Dear prof Steven and All MPB users,


I had used MPB for more than half year, however, i still very confused with 
the results i obtained as compare to reported publication.

I was trying to calculate the band diagram for line-defect photonic crystal 
waveguide, triangular lattice air holes in planar photonic crystal. I am 
using supercell method to obtain the band diagram.

-*-*-*-
I found that my calculated frequency(a/lambda) for the guide mode in the 
line defect is always lower than those reported in journal.

for example:
at wavevector,k=0, my calculated frequency is 0.285(a/lambda), However, the 
one reported in journal(i had tried reproduce the banddiagram for at least 3 
journals) is higher around 0.295(a/lambda).

What is the reason for the frequency shifted to lower value in my band 
diagram? is it due to the supercell method itself? Any idea how can i solve 
it?


I am waiting for your help.

thank you.

Regards,
DC