Specific Process Knowledge/Characterization/Stress measurement

From LabAdviser
Revision as of 09:42, 17 April 2020 by Reet (talk | contribs) (→‎Stress measurement using XRD: corrected links and added a bit more info)

Feedback to this page: click here

Stress measurement

The stress in a thin film can be quantified with a profilometer by measuring the wafer bow before and after deposition of a thin film. If the then film is deposited on both sides of the wafer, you can measure the bow after deposition and again after removing the film from one of the sides.

If your thin film is are crystalline, you can also measure stress using XRD (see below). It is more time consuming but may be more accurate.

Stress measurement using a profilometer

Using a profilometer, the procedure is as follows:

When a thin film is deposited on one side of the wafer

  1. Make a pre-stress measurement. Measure the wafer bow on one of the stylus profilometers (Dektak XTA_new or Dektak 8). Save the measurement. It is a good idea to measure across most of the wafer (at least along 70% of the wafer diameter) in two directions perpendicular to each other.
  2. Measure the thickness of the wafer
  3. Deposit the thin film
  4. Measure the thickness of the thin film (ex. using the FilmTek or the Ellipsometer).
  5. Make a post-stress measurement: Measure the wafer bow again:
    • On the same profilometer as used for the pre-measurement.
    • With the same recipe as used for the pre-mesurement.
    • On the same position(s) on the wafer as the pre-measurement.
  6. Use the program for stress measurement in the profilometer software. Use both the pre-stress measurement and the post-stress measurement.
    • The profiler program asks for a substrate elasticity (choose the substrate type, e.g. Si(100)), substrate thickness, film thickness and the name of the pre-stress measurement.

When a thin film is deposited on both sides of the wafer

  1. Deposit the thin film
  2. Make a pre-stress measurement. Measure the wafer bow on one of the profilometers (Dektak XTA_new or Dektak 8). Save the measurement. It is a good idea to measure across most of the wafer (at least along 70% of the wafer length) in two directions perpendicular to each other.
  3. Remove the thin film from one side of the wafer. If it is a single side polished wafer, then remove it on the non-polished side.
  4. Make a post-stress measurement. Measure the wafer bow again:
    • On the same profilometer as used for the pre-measurement.
    • With the same recipe as used for the pre-mesurement.
    • On the same position(s) on the wafer as the pre-measurement.
  5. Measure the thickness of the wafer
  6. Measure the thickness of the thin film (ex. using the FilmTek or the Ellipsometer).
  7. Use the program for stress measurement in the profilometer software. Use both the pre-stress measurement and the post-stress measurement.
    • The profiler program asks for substrate elasticity (choose the substrate type, e.g. Si(100)), substrate thickness, film thickness and the name of the pre-stress measurement.

Stress measurement using XRD

After you make a θ-2θ scan of a crystalline thin film, the peak positions can be used to calculate the lattice constant of the film. By comparing the calculated lattice constant to the theoretical one, you can calculate the amount of compressive or tensile stress. It is important to align the thin film carefully to the beam height, as you can in the XRD Smartlab. You may use a standard measurement (e.g. of a crystalline Si sample) to be able to take into account the intrinsic peak broadening of the instrument. Use the Ni-filter to remove the Kβ side peaks and if you have enough signal you can also use the monochromator and analyzer crystals to remove the Kα2 contribution.

Nanolab has the SmartLab Studio software and the Malvern Panalytical HighScore software available for quantifying the stress (ask staff for details).

You may be able to measure the relaxed lattice directly rather than comparing to a theoretical lattice, but you would have to be able to release the film and measure it separately (if this means scraping the thin film off your sample mechanically, you could potentially measure it in the XRD Powder, but you would need a relatively large amount of powder to get a good signal).