how to get the layout of a Silicon-Embedded transformer in momentum

[mustangyhz]

I know how to get the layout of a tranditional front-side on-chip transformer in momentum. But, I don't know how to get the layout of a Silicon-Embedded transformer in momentum, because the fabrication of the transformer is based on some special process including trench formation, thermal oxidation, trench filling, pad formation, backside metal removal and BCB layer formation.

help me!
thanks!

 

[volker@muehlhaus]

This requires "normal" Momentum stackup definition - except for the through silicon vias. These TSV need to be isolated from the conducting silicon substrate by some vertical oxide, and that is not possible with normal Momentum vias. But from what I have seen, Keysight have added some TSV functionality for Momentum in ADS2015. You should contact Keysight support and ask for more information on TSV simulation in Momentum.

~

Wait ... I see that vias and conductors are both embedded in silicon, with only a thin oxide layer as isolation. That is a lot of area and the capacitance is more critical then, compared to normal TSV cases. I don't know if the TSV approach in Momentum is accurate enough for this case. You should consider using the ADS 3D FEM solver instead of Momentum.

The oxide layers could be auto-generated from the Metal layers by using "derived layers". I had shown the use of these derived layers for auto-generating passivation here:
https://muehlhaus.com/support/ads-application-notes/ads-fem-3d-passivation

 

[mustangyhz]

This requires "normal" Momentum stackup definition - except for the through silicon vias. These TSV need to be isolated from the conducting silicon substrate by some vertical oxide, and that is not possible with normal Momentum vias. But from what I have seen, Keysight have added some TSV functionality for Momentum in ADS2015. You should contact Keysight support and ask for more information on TSV simulation in Momentum.

~

Wait ... I see that vias and conductors are both embedded in silicon, with only a thin oxide layer as isolation. That is a lot of area and the capacitance is more critical then, compared to normal TSV cases. I don't know if the TSV approach in Momentum is accurate enough for this case. You should consider using the ADS 3D FEM solver instead of Momentum.

The oxide layers could be auto-generated from the Metal layers by using "derived layers". I had shown the use of these derived layers for auto-generating passivation here:
https://muehlhaus.com/support/ads-application-notes/ads-fem-3d-passivation

thanks for your reply!
can you give the substrate file in ads2011_10 for this work?

 

[mustangyhz]

This requires "normal" Momentum stackup definition - except for the through silicon vias. These TSV need to be isolated from the conducting silicon substrate by some vertical oxide, and that is not possible with normal Momentum vias. But from what I have seen, Keysight have added some TSV functionality for Momentum in ADS2015. You should contact Keysight support and ask for more information on TSV simulation in Momentum.

~

Wait ... I see that vias and conductors are both embedded in silicon, with only a thin oxide layer as isolation. That is a lot of area and the capacitance is more critical then, compared to normal TSV cases. I don't know if the TSV approach in Momentum is accurate enough for this case. You should consider using the ADS 3D FEM solver instead of Momentum.

The oxide layers could be auto-generated from the Metal layers by using "derived layers". I had shown the use of these derived layers for auto-generating passivation here:
https://muehlhaus.com/support/ads-application-notes/ads-fem-3d-passivation

my substrate

my transformer


Cu_mml resistivity of 3.9 μΩ-cm
conductivity

simulation results


obiviously, it's wrong,
the resistance is too big! Do I set the wrong conductivity?
the primary coil is a capacitor! why?

 

[mustangyhz]

This requires "normal" Momentum stackup definition - except for the through silicon vias. These TSV need to be isolated from the conducting silicon substrate by some vertical oxide, and that is not possible with normal Momentum vias. But from what I have seen, Keysight have added some TSV functionality for Momentum in ADS2015. You should contact Keysight support and ask for more information on TSV simulation in Momentum.

~



Wait ... I see that vias and conductors are both embedded in silicon, with only a thin oxide layer as isolation. That is a lot of area and the capacitance is more critical then, compared to normal TSV cases. I don't know if the TSV approach in Momentum is accurate enough for this case. You should consider using the ADS 3D FEM solver instead of Momentum.

The oxide layers could be auto-generated from the Metal layers by using "derived layers". I had shown the use of these derived layers for auto-generating passivation here:
https://muehlhaus.com/support/ads-application-notes/ads-fem-3d-passivation

Can you tell me what's wrong with my project in ads2011_10 momentum?
Thanks!

 

[mustangyhz]

I see, I set the wrong conductivity
Cu_mml
resistivity : 3.9 μΩ-cm = 0.039 μΩ-m


conductivity: 2.546e7 siemens/m