Light Damages the Transistors

Discussion in 'Cadence' started by Su, Jun 22, 2006.

  1. Su

    Su Guest

    Hello, Everyone:

    I need to do on-wafer measurement on my high frequency
    amplifier. But I find the transistors are very sensitive to
    microscopic light.
    Sometime after I switch off the light and then apply the DC bias,
    there will be large leakage current on transistor gate.

    Do you have idea on why this happens and any suggestion to avoid this
    problem?

    thanks,

    suyu
     
    Su, Jun 22, 2006
    #1
  2. Su

    sheldonbrand Guest

    Su,

    Have you tried putting a light shield over the block? For example,
    have you placed a piece of top-level metal over the amp? The
    reason this occurs is the same reason you can build CCDs and
    CMOS Image Sensors.

    Best
    Regards,


    Sheldon
     
    sheldonbrand, Jun 26, 2006
    #2
  3. Su

    sheldonbrand Guest

    Su,

    One other question, if you turn the power off and the light
    off does the leakage go away?

    If the leakage goes away when both power and the light
    are off then it maybe that the circuit is latching up. If the
    leakage does not go away, then you maybe injecting hot
    carriers into the oxide over the B-E junction.

    What happens when you don't bias the circuit while the
    light is on?

    Best Regards,

    Sheldon
     
    sheldonbrand, Jun 26, 2006
    #3
  4. Su

    syime1974 Guest

    HI, Sheldon:

    I did not put top-level metal to block the light. Instead, I've used
    "dummy-block" to block any dummy metal layer which could be inserted by
    foundry to reduce the parasitic capacitance.
    THIs problem has become more severe when we used shorter-channel
    device.
    But once again, how does the light damage the transistor gate?

    Many thanks to your reply,

    suyu
     
    syime1974, Jun 26, 2006
    #4
  5. Su

    syime1974 Guest

    HI, Sheldon:

    Our measurement process is:
    1. Landing the probles on the pads.
    2. Turn off the microscopic light.
    3. BIas the circuit by DC power supply.

    At this point, we observed large gate leakage current.

    If we did not turn off the light, the circuit was fine when biased.

    thanks,

    suyu
     
    syime1974, Jun 26, 2006
    #5
Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.