Hello David,

The use of nitride instead of oxide is a good idea. However, if we do it like you described, then the lower electrode will be such poly that had been excluded from doping, that is particularly a resistor, giving the cap a high series resistance and a very low Q-factor. We can even have that by using a poly-well or poly-nbase cap that is already available with our current process. The aim of the mim is to reduce the ESR from some Ohms to some ten milliOhms and thus increase the Q, making the cap suitable for RF applications. I would approach this by leaving existing steps as-is and then form a metal2-~40nm nitride-metcap-via2-metal3 stackup. Also, by using metal2 as the lower layer, metal1 and the frontend layers remain free for device formation, so you don't completely waste the space under the cap. Is it possible to deposit nitride on top of metal (cleanliness, adherence to the surface, etc.)?

Regards,

Ferenc

On Mon, Nov 12, 2018 at 7:12 AM David Lanzendörfer <david.lanzendoerfer@o2s.ch> wrote:

Hi Ferec
I've been thinking about reducing the energy of the nimplant/pimplant field
implant steps and then making the implant stop nitride layer thinner on top of
the diode so that we can use that nitride layer as as well as dioelectricum.
So we could have polysilicon, then nitride, and then a metal layer on top.
Would do you say? Nitride is a high-k material with a relative permittivity of
around 7-8 compared to silicon oxide with around 4.
I think a junction with 50nm depth still is good enough?

Cheers
David