MIM capacitor design flow

Hi for all,
I am planning to start a new design, it is about MIM capacitor for DRAM applications.
In fact, i am facing a delima :? , I can't figure out the best design flow/tools.
the project should cover all aspects of the process; mechanical structure, physical proprieties, device integration on top of substrate ..
So, In the first place, I have to take decision on the design flow ; In ower univ we have a lot of tools : ADS, COMSOL, HFSS, Silvaco, MatLab, HSPICE, Cadence Virtuoso ...

In my opinion, i will fellow the next :

1) Do Mechanical design and the thermal analysis in COMSOL MultiPhysics.
2) Export the design and doing Electromagnetic analysis in HFSS.
3) Export the design from HFSS as a GDS2 file and do the electrical analysis in Silvaco.
4) Extract a Spice Model of the design for frequency response using HSPICE.
5) Device integration and layout using Virtuoso.

I know that is a challenging job that i put myself into, but it worth the effort :shock:
But my problem is :
1* compatibility between tools,
2* when exporting, if i assign a particular material in COMSOL (like HfO2) , other tools will recognize it ? GDS2 format describe material properties (like dielectric constant ..) ?
3* accuracy ...

please tell me about the best design flow, i will be very thankful.

Malek Lamari said:

2* ... GDS2 format describe material properties (like dielectric constant ..) ?

Click to expand...

No, just layer-dependent geometrical data.

could you please tell me other options for exporting data between tools.

Malek Lamari said:

could you please tell me other options for exporting data between tools.

Click to expand...

Possibly with the EDIF 3 0 0 format. Start here or see there.

It seems to me that you're really overthinking the problem.
It's a simple (linear) dielectric albeit higher K, parallel plate
problem.

The valuable details are not the X and Y extents, which is
what you would get from an IC design (layout) tool. It's the
vertical stuff, how many layers, how planar, up from the
substrate, how the explicit dielectric thickness reacts to
underlying topography (if any / if allowed), details of how
you step up out of thin dielectric and how you oversize the
plates to the MIM cut and how that sidewall looks, etc.

I think what you really want are some SEM cross-sections
for the closest relative you can find (figuring the foundry
would reuse unit processes in dep and litho), and a device /
process simulator tool with a simple graphical entry interface
where you can reconstruct what you see in the third
dimension.

Malek Lamari said:

2) Export the design and doing Electromagnetic analysis in HFSS.
3) Export the design from HFSS as a GDS2 file and do the electrical analysis in Silvaco.

Click to expand...

If you do the EM modelling in step 2, you will get the complete electrical response with good accuracy and can create a broad band model from there.
So what is the purpose of step 3?

thank you for your valuable time,

At first, let me explain the situation further. In fact my problem is that the used dielectric layer is a particular material that i could not find it in any one of the libraries of the above tools. another approach consist in manually adding a custom material and manually assigning his properties (permittivity, dielectric cons ...).
and here i doubt that this approach gives me an accurate results.
Second, i want to know the inter-tools compatibility; by that i mean when exporting, does material specific properties are conserved ?

@ erikl: my biggest problem, is i have to concentrate in the physical aspect of the phenomena, with extensive analysis of the electric field.
@ ****_freebird: exactly ^^' , i am overthinking the problem, but please tell me; can i simply replace the particular dielectric materiel by a numeric model ? for example its dielectric constant ?
@ volker_muehlhaus: yes as you say, step 2 will allow me to do a full electrical analysis, but step three will offer me some of the interesting features of Silvaco tools, for example carriers mobility and density...

Malek Lamari said:

In fact my problem is that the used dielectric layer is a particular material that i could not find it in any one of the libraries of the above tools. another approach consist in manually adding a custom material and manually assigning his properties (permittivity, dielectric cons ...).

Click to expand...

If you know the material properties (epsilon_r is most relevant, tan_delta can be assumed zero) then adding this material is easy. I have designed a couple of technology files (including MIM layers) for EM analysis in Sonnet. Your HFSS setup is not very different for the material properties.

Malek Lamari said:

and here i doubt that this approach gives me an accurate results.

Click to expand...

From my experience, the MIM are very robust in terms of EM accuracy. If you know how to use your EM tools, and if you have the correct stackup data, you will get accurate results.

Malek Lamari said:

Second, i want to know the inter-tools compatibility; by that i mean when exporting, does material specific properties are conserved ?

Click to expand...

Layout exchange formats lile GDSII do not include material properties.

Malek Lamari said:

@ volker_muehlhaus: yes as you say, step 2 will allow me to do a full electrical analysis, but step three will offer me some of the interesting features of Silvaco tools, for example carriers mobility and density...

Click to expand...

Is that relevant for a linear passive device like the MIM, which is already completely charaterized by the EM analysis?

volker_muehlhaus said:

Is that relevant for a linear passive device like the MIM, which is already completely charaterized by the EM analysis?

Click to expand...

thanks for your quick response,
If this analysis is done, this will be a plus for the project; as the project is a purely research, and we look to enhance the structure through deep analysis.