Photoresist Characterization

1
Photoresist Characterization

• Spin Speed vs. Thickness
• Nate Hamm, Steve Kelly, Brian MacFarland, John
  Yarbrough, Jeff Flint

2
Introduction

• Photolithography is an important procedure in
  semiconductor processing
• For our processes in the lab, having the proper
  thickness coating is essential
• This is controlled by the spin speed on the spin
  coater

3
Procedure

• To determine the spin speed vs. thickness, we
  used Shipley 1813 photoresist and spun it on
  with varying dwell speeds
• 3000 rpm 5000 rpm
• 3500 rpm 5500 rpm
• 4500 rpm 6000 rpm

4
Analysis

• Once the resist was spun on, we took thickness
  measurements with the ellipsometer and
  profilometer
• On the ellipsometer we used two angle
  measurements to obtain the correct thickness
• For the profilometer we first put a glass slide
  over part of the resist and used the RIE to etch
  the exposed region

5
Thickness vs. Spin Speed

• With the ellipsometer, we took a series of
  measurements across the samples as shown
• This demonstrates the resist uniformity across
  the sample vs. spin rate

6
Thickness vs. Spin Speed (cont.)

• Then portrayed against theoretical data, the
  experimental set looks accurate

7
Etch rate on RIE

• To make measurements on the profilometer, we
  placed a slide over part of the photoresist and
  etched away the exposed portion with an oxygen
  plasma
• Summarized in this figure are our findings on
  etch rate vs duration at 125 W

8
Profilometer Mishap?

• The data we gathered from the profilometer is not
  close to the theoretical set

9
Profilometer Mishap? (cont.)

• Possible reasons for the error were
• Outside of optimal spin range
• Uneven coating of the wafer by the photoresist
• User error interpreting the data obtained by the
  profilometer
• RIE etching process
• Gnomes?

10
Conclusions

• The ellipsometer data shows a good agreement with
  the Shipley data for the thickness as a function
  of spin rate
• The thickness was shown to be inversely
  proportional to the square root of the spin rate
• The profilometer data did not agree well with
  either the ellipsometer data or the Shipley data,
  especially at low spin speeds
• This may have been a result of uneven coating of
  the wafer by the photoresist or outside the
  optimal range

11
Conclusions (cont.)

• The etch depth vs. etch time displayed a linear
  relationship.
• From this we found the etch rate for an oxygen
  plasma in the reactive ion etcher at 125 W to be
  37.90Å/sec