CMC Subcommittee on MOS Varactor Model

1
CMC Subcommittee on MOS Varactor Model

• 3Q 2006 Meeting
• Maastricht, The Netherlands

2
Beta Model Generation

• Decisions from 2Q meeting on model functionality
  were incorporated into beta model
• New external node in the well (bi)
• Addition of vertical resistance through gate
• Gate current has been added
• Voltage dependent doping density

3
Beta Testing

• Jim Victory supplied the beta model in August for
  testing
• Two telecons to discuss results (19th September
  and 5th October)
• Some good preliminary results, but more extensive
  testing by more subcommittee members would be
  very helpful

4
CV Tests

• Nominal CV tests show good results by several
  groups
• Voltage dependent doping variation has desired
  effect of increasing capacitance from deep
  depletion value. There is however an unphysical
  upturn in capacitance further into deep
  depletion.
• Side discussions about how to use this model for
  differential MOS varactor devices

5
Temperature Effects on CV

• Temperature effects are in model but the spread
  in capacitance in the knee in accumulation is too
  large. Ted handles this issue with temperature
  dependence in tox value.

6
Transient simulations

• Richard reports a kind of hysteresis in transient
  simulations. Starting from deep depletion (for
  NFET in Nwell device) and moving towards
  accumulation, he notes strange results from
  model as opposed to starting in accumulation and
  moving into deep depletion in which he finds the
  expected results from model.

7
Tau parameter extraction

• Plamen has been try to develop a method to
  extract tau, the minority carrier generation
  rate. In dark, he observes deep depletion, but
  with a bright light on sample, the capacitance
  shows the typical flat line CV curve in deep
  depletion.
• Jim is working on a simulation that more
  accurately reflects our experiments of slowly
  moving into deep depletion (transient) and
  applying a small ac voltage in contrast to the
  standard ac test that is essentially an operating
  point measurement.

8
Gate Current

• Gate current has been a bit problematic.
• Needs to be brought into full alignment with PSP
• Parameter values are hard to determine, though
  Plamen has had success
• Plamen does report that for PFET in Pwell, the
  gate current has an effect on the series
  resistance values.

9
Mixed Polarity Device

• Ted reported on attempts to fit CV curve of
  device with p-type gate and n-type well. The fit
  was fairly good, but the data shows an increasing
  capacitance in accumulation that could not be fit
  by the model, even with a variation in quantum
  mechanical effects.

10
The Future

• Time-line Final model by July 07 ???
• Aligned with PSP, same parameter names,
  independent self-standing macros, shared modules
  with PSP ???
• Documentation Part of funding proposal
• QA Needs to be generated, possible volunteer???
• Vendors Ted to inquire if they are on board
  with implementing the model once it is voted as a
  CMC standard model.