There are a few problems:
- The online calculator gives invalid results (negative ZL), most likely because the aspect ratio (line width / line thickness) is too extreme
- That very wide line has a lot of shunt capacitance, so treating it as a series inductance is inaccurate.
Line impedance of a lossless line is
ZL=sqrt(L'/C')
You can get the free Sonnet Lite EM solver and simulate the line as a flat (zero thickness) conductor, which is quite accurate at this width. Then, you can export a SPICE pi model (L + C).
I did a quick setup (model attached) and get ~ 7.6 Ohm line impedance.
Sonnet Lite model:
View attachment sl.zip
A simple equivalent circuit with series L and shunt C is valid if the line length is less than 1/20 wavelength. For the line length of
10mm that I simulated, I get this equivalent circuit:
C_C1 1 0 3.991385pf
C_C2 2 0 3.993408pf
L_L1 1 2 0.590898nh
If the line is longer than 1/20 wavelength at your frequency of interest, you need to cascade multiple short segments (cascaded LC ladder network, optional R and G are conductor losses and dielectric losses).