Hi Michael,
I looked at your script and the results, and I don't see any obvious issues with the simulation - the settings for the solver are all reasonable, and the results are essentially independent of any of the grid-related parameters. Unfortunately, the paper you linked to provides zero information about how they ran their simulations (it doesn't even say what software was used), so its impossible to tell why their result might be different.
Regards,
Ray