Using HFSS to simulate Metamaterials, something wrong

Hello,

I am using HFSS to simulate Metamaterials.
I am using the paper attached for reference.
I traced the values of the constitutive parameters (Figure 3) and saved the traces into the attached files
Epsr.tab
MiuR.tab
Etan_del.tab
Mtan_del.tab
The traces are not perfect, but a good approximation.
Than, I built the HFSS file attached, which basically consists of a ‘homogeneous’ cube with a dispersive material defined by the four datasets above. I used vacuum extensions ports (with the proper deembedding) to make sure that only the two modes are relevant for the port definitions.
The S-parameters I get are WAY different from the ones published in the paper. The difference between the datasets values and the published constitutive parameters do not justify the difference in the S-parameters I get and the one published in the paper.
I suppose the paper is correct and I do something wrong in HFSS.
Can anybody take a look?

Thank you

Kato