Specific Process Knowledge/Characterization/Profiler/Optical Profiler (Sensofar) acceptance test

Results from the Optical Profiler (Sensofar) acceptance test

The acceptance test was performed in January 2012 by STInstruments and Sensofar together with Pernille V. Larsen @ Danchip and Berit G. Herstrøm @ Danchip.

This Table shows the acceptance criteria

'	Title	Sample description / measurement	Acceptance criteria
1	Patterned flat sample of silicon	Sample material: Patterned silicon substrate.	Depth 100±2 µm
		Trench depth with aspect ratio 1:10 on a 10 µm wide trench
2	Patterned flat sample of silicon.	Sample material: Patterned silicon substrate.	Depth 91±2 µm
		Trench depth with aspect ratio 1:11 on a 8 µm wide trench
3	Patterned flat sample of silicon	Sample material: Patterned silicon substrate	Depth 85±2 µm
		Trench depth with aspect ratio 1:13 on a 6.4 µm wide trench
4	Patterned flat sample of glass	3D profile of pattern etched down in a quartz sample.	Depth within ±1% from a standard profiler measurement and repeatability (3 successive measurements) within 0.1%
		Pattern size 20 µm x 20 µm, depth 500 nm
5	Flat sample of silicon with thick patterned oxide	Step height of patterned thick (10 µm) oxide on top of a silicon wafer.	Step height must be within ±3% of a SEM profile measurement.
		7 µm deep pattern, trench width 6 µm
6	Flat sample of silicon with thick layer of patterned polymer	Sample material: SU8 on silicon.	Height 69±2 µm
		Pillar heights of 69 µm with 25 µm between pillar edges
7	Free standing structure	Measure bow due to stress of a membrane.	Membrane bow repeatability of 5 successive measurement within 2%
		Membrane material: Si(2 µm)/SiO2(~1 µm)/Au(50 nm).
		Membrane size: honey comb structure approximately 150 µm in diameter.
		Bow < 500 nm
8	Stitching of large area	Stitching 4 mm x 4 mm on same sample as in 7. The area consists of many closed packed honey comb formed membranes.	Membrane bow must the same as on 7 within 2%
9	Narrow trenches and holes	Sample material: Patterned silicon substrate.	Depth 20±2 µm
		2.5 µm wide trench in silicon with a depth of 20 µm
10	Film thickness measurement of transparent thin film	Transparent thin film thickness of 28 nm SiO2 on Si	SiO2 thickness 28±1 nm
11	Measurements of multiple stacks	120 nm nitride on 110 nm oxide on a silicon substrate	Within ±2% on each layer from an ellipsometer measurement.
12	Film thickness measurements of transparent films on small structure	Sample material: Si with 1.5 µm patterned AZ-resist	Within ±1% from a standard profiler measurement.
		Measure thickness of AZ-resist on pillars of 50 µm in diameter
13	Roughness repeatability	Sample material: Si wafer with poly-silicon layer.	Repeatability within 0.2%
		3 successive measurements of the roughness

Results of acceptance test no. 1, 2 and 3

Sample material: Patterned silicon substrate Measurement: Trench depth with aspect ratio 1:10, 1:11 and 1:13 on a 10µm, 8mm and 6.4µm wide trenches respectively Acceptance criteria:Depth 100±2 µm, 91±2µm and 85±2µm. The SEM profile images of the three trenches are shown here:

A1	A2	A3

Test no. 1 was done in two ways:

1. With confocal objective EPI 100x-N: See the results here
2. With Interferometric objective 50X DI: See the results here

Test no. 2 was done using: confocal objective EPI 100x-N: See the results here

Test no. 3 was done using: confocal objective EPI 100x-N: See the results here

Setting for methode no. 1 for test no. 1,2 and 3: confocal

Recipe: Trench

• Operation mode: trench
  • + Raw smoothing confocal
  • + fine shift
• Objective: EPI 100X-N
• Z scan
  • Dual - bottom up
    • top: 8µm
    • Gap: 91µm (the trench depth)
    • Bottom: 8µm
  • Speed factor: 1x
  • + Linear stage
• Threshold: 0.0%
• Light source
  • Levels: 2
    • 900 -> 60 (might need to be set a little different)
    • Gain: default

Setting for methode no. 2 for test no. 1: interferometric

We do not have a recipe for that but we used:

• Objective: Interferometric 50x DI
• Z scan: VSI
• Light souce: increased gain and contrast

Results of acceptance test no. 4

Sample material: Patterned fused silica.

Measurement: Depth of pattern

Standard profiler measurement:335nm

Acceptance criteria:Depth within ±1% from a standard profiler measurement (331.65nm-338.35nm) and repeatability (3 successive measurements) within 0.1%

System settings:

Objective: DI 50x-N

VSI z scan: 10µm

Treshold: 1%

Results

It was repeated 10 Time.

The first measurement is shown here

This table show the depth value for the 10 measurements

'	Measured depth [nm]
1	337.5
2	336.5
3	334.7
4	335.5
5	339.2
6	337.2
7	334.2
8	335.5
9	341.1
10	344.4
Average depth	336.58
Repeatability	3% (the bad repeatability was accepted due to the high noise level in the room)

Results of acceptance test no. 5

Sample: Flat sample of silicon with thick patterned oxide.

Measurement: Step height of patterned thick (10 µm) oxide on top of a silicon wafer.The pattern is aprox. 7µm thick

Acceptance criteria: Step height must be within ±3% of a SEM profile measurement. See the SEM profile here:

A5

Settings:

• Recipe: Acceptance nr. 5 SMR
• Mode: default
• Objective: 50x-DI
• VSI Z scan 18.0µm
• Threshold 2%
• Note: Scan only in the APOX (not in the Si) or else the system will get confused.

See the result here:

Sensofar A5 result

Results of acceptance test no. 6

Sample: Flat sample of silicon with thick layer of patterned SU8.

Measure: step height.

Acceptance criteria: Height 69±2 µm

Settings:

• Recipe: Acceptance nr. 6 SMR
• Mode: application Acceptance 6
  • Raw smoothing confocal
  • Coarse shift single samle
  • Reject multiple reflections
• Objective: Confocal EPI 20x-N
• Z scan range: 104µm
• Threshold: 2.0%

See the result here:

Sensofar A6 result no. 1	Sensofar A6 result no. 2

Results of acceptance test no. 7

Sample: Free standing structure.

Measure: Measure bow due to stress of a membrane.

Acceptance criteria: Membrane bow repeatability of 5 successive measurement within 2%

The sample was not as we suspected. There was almost no bow. The membrane looked broken and it did not really make any sense to measure the bow. The measurement we did are shown here:

Sensofar A7 result