ICP Etching Recipes

From UCSB Nanofab Wiki
Jump to navigation Jump to search

Back to Dry Etching Recipes.

DSEIII_(PlasmaTherm/Deep_Silicon_Etcher)

Process Control Data (DSEiii)

Si Etching C4F8/SF6/Ar (PlasmaTherm DSEiii)

  • Recipe: STD_Bosch_Si (⭐️Production), on 100mm Si Wafer with ~50% open area, photoresist mask, ~40µm deep

Edge-Bead Removal (DSEiii)

Make sure to remove photoresist from edges of wafer, or PR may stick to the top-side wafer clamp and destroy your wafer during unload!

  • Edge Bead Removal via Photolithography: use a custom metal mask to pattern the photoresist with a flood exposure.
    • If you are etching fully through a wafer, remember that removal of edge-bead will cause full etching in the exposed areas. To prevent a wafer from falling into the machine after the etch, you can mount to a carrier wafer using wax.
  • Manual PR Edge-Bead Removal - using swabs and EBR100. This is prone to error and easy to accidentally leave a blob of PR on the edge - so be extra careful to ensure NO PR is left on the edges!

High Rate Bosch Etch (DSEIII)

  • Bosch Process Recipe and Characterization - Standard recipe on the tool.
    Example SEM image
    Example of 100µm Deep Bosch Etched Silicon posts with hard mask. Close inspection shows the horizontal "scalloping" from the cycling nature of the etch. (Image Credit: Demis D. John, 2021-07)
    • STD_Bosch_Si (⭐️Production) - Developed 2024-10
      • Old Recipe Name: "Plasma-Therm Standard DSE" - lower EtchA, less tolerant
    • Standard Bosch Process for high aspect-ratio, high-selectivity Silicon etching.
      • Cycles between polymer deposition "Dep" / Polymer etch "Etch A" / Si etch "Etch B" steps. Step Times gives fine control.
      • To reduce roughening/grassing ("black silicon"), Increase "Etch A" time by ~50%. Alternatively, reduce "Dep" step time by ~20%.
    • Patterns with different exposed/etched areas will have different "optimal" parameters.
    • This recipe has 2s Etch A time compared to "Plasma-Therm Standard DSE" (which has 1.5s Etch A) below - this reduced the undercut of mask to ~1% of the etch depth and the effect of aspect ratio on etch rate. All other recipe parameters are the same.
    • Selectivity to UV6 Photoresist ~60-80.
    • Selectivity to SiO2 should be higher, not yet measured.
    • Selectivity to Al2O3 is extremely high, >9000. See below TSV process for processing notes with Al2O3 hardmask.
    • If you need to pattern all the way to the edge of the wafer, PR won't work because you have to remove the edge-bead of photoresist (see above). Instead use hardmask process (See "Through Silicon Via" etch below).
    • <1% center to edge variability in etch rate.
    • Larger open area → lower selectivity & lower etch rate.
    • Thick PR's approx ≥10µm tend to burn, avoid thick PR's. They also make edge-bead removal very difficult.
Tilted SEM of 40um deep etch
~40µm deep Silicon etch, run as Process Control "EtchCal" (Process Development and Image: Noah Dutra, 2024-10-07)
Example SEM image
Example of 16.32µm Deep Etched Silicon with UV6 Photoresist mask, 2µm Pitch. (Image Credit: Noah Dutra 2024-08)

Through Silicon Via (TSV) etch (DSEiii)

Since the topside clamp requires the removal of photoresist on the outermost ~5-7mm of the wafer, this makes PR incompatible with through-silicon etching (as the outer edges would be etched-through, dropping the inner portion into the chamber). In addition, in practice we have found that thick PR often roughens and burns during long ~30-60min etches, making removal very difficult.

Instead, we recommend the following process with Al2O3 hardmask: 

NOTE: We have recently found that the wax-mounting process process can leave wax on the wafer clamp, causing the next user's wafer to get stuck and fail transfer!  DO NOT RUN the wax-mounting process without discussing with staff first. (Through-wafer process with no wax is still acceptable.) -- Demis 2024-03-11
Process for Through-Wafer Silicon Etching
Process to etch through ~550µm Silicon Demis D. John & Biljana Stamenic 2022-11-11. Please consider our publication policy if you use/modify this process.
Deposit 150nm Al2O3 on either: May need to do dep. rate check beforehand.
Deposit ~3nm SiO2, in situ (same machine as above) This improves adhesion to photoresist and prevents developer attacking the Al2O3.
Lithography - your preferred method. Needs approx. ≥500nm thick PR.
Etch the Al2O3 in Panasonic ICP 1/2 Use 50W version. Overetch by ~20%, will also etch through the thin SiO2 layer.
Strip PR - either in situ, or via NMP 80°C soak followed by PEii Technics ashing. In situ PR strip appears to give better + faster results.
If pieces of the wafer are at risk of falling into the chamber, mount the product wafer to a carrier wafer:

Logitech Wax Mounting Recipe - Bulk Crystal Bond


If you are only etching small holes through the wafer (majority of wafer is intact), then wax-mounting is not necessary.

CONTACT STAFF before attempting this step!


Critical - Ensure no wax is present on either side or edge of wafer prior to DSE etching, or wafer may break in the DSE during robot unload!

Use POLOS spinners with ACE/ISO to clean front and back of wafer.

IMPORTANT for wax-mounting, to ensure wax does not stick your wafer to the DSE clamp.

Observe carefully for any wax protruding from between wafers - redo spin-clean as needed.

Also make sure wax thickness is not too thick, of long etches could cause wax to seep out from between the wafers.
DSEiii etch - reduce Dep step to eliminate grassing:
  • Bosch Cycles: Dep: 1.2sec / Etch A: 1.5sec / Etch B: 2.0sec
  • Rate ≈ 4.25µm / min
 Can use Lasermonitor and/or Camera to observe when etch is fully through. Trenches may get black/rough, but then clear up when fully etched.
  • Record Helium FLOW during recipe run, for next step (if He leaks).
  • Ok to remove wafer, observe/measure, and re-load for etching.
  • If see black grass and etch rate drops, may need to run an O2 plasma with Technics PEii few min to remove polymer, Increase EtchA step time (eg. by 50-100%) and then continue the etch.
If you did not wax-mount your wafer, the recipe will eventually fail for Helium Pressure/Flow out of compliance. This is because the cooling Helium leaks through the wafer when the openings get fully etched through.

Once this happens,

  • Leave your wafer in the chamber, then
  • Edit recipe to set Helium Cooling (first step only) to "Flow Control Only".
  • Set to typical flow of normal process from above (Something like ~6sccm? Not sure. Exact value is not critical)
  • Re-run the recipe with He on Flow-control only until etch is complete.
Strip Al2O3/SiO2 either with Buffered HF, or same Pan1/2 dry etch as above.

BHF: Eg. ~2min to fully remove SiO2 + Al2O3, with overetch.

See etch BHF rates of the thin-films on this table.
IF wax-mounted - either
  • dissolve in Acetone overnight (make sure to excess-fill enough and cover tightly with tinfoil so it doesn't dry up), complete with ACE/ISO rinse

OR

  • place wafer on tinfoil-covered hotplate at 150°C, and slide product wafer off, then
    • ACE/ISO clean (eg. POLOS) to remove wax.
If you publish using the above process, please consider our publication policy. This process was developed by Biljana Stamenic and Demis D. John, 2022.

Single-Step Low Etch Rate Smooth Sidewall Process (DSEIII)

PlasmaTherm/SLR Fluorine Etcher

Recipe Tips

  • RF1: Bias Power (with DCV readback)
  • RF2: ICP Power
  • For trouble igniting ICP plasma, add 15 to 75 W of bias power during ignition step. Typical ignition pressures 5 to 10 mT.

Process Control Data (Fluorine ICP Etcher)

SiO2 Etching with CHF3/CF4 (Fluorine ICP Etcher)

Si Etching (Fluorine ICP Etcher)

Si Etching C4F8/SF6/CF4 (Fluorine ICP Etcher)

  • SiVertHFv2 (⭐️Production) - Full Wafer Si etching with ~50% open area and resist mask
    • This recipe has 2x gas flow compared to "SiVertHF" below - this reduced the loading effect (dependence on % etched area).

Process Notes/Observations

  • Due to high selectivity against SiO2, it may be necessary to run a ~10sec 50W SiO2 etch (below) to remove native oxide on Si. This can be performed in situ before the Si etch.
  • We have observed that full-wafers with small open area in photoresist masks might require a recalibration of the C4F8/SF6 ratio in order to prevent very low etch rates.

SiO2 Etching (Fluorine ICP Etcher)

  • SiO2 Etching using Ruthenium Hardmask - Full Process Traveler
    • Ning Cao & Bill Mitchell, 2019-06
    • High-selectivity and deep etching using sputtered Ru hardmask and I-Line litho.
    • Etch also works well with PR masking
    • Chemistry: CHF3/CF4
    • Variations in SiO2 etch Bias Power: 50 / 200 / 400W bias.
    • Ru etch selectivity to PR: 0.18 (less than 1): 150nm Ru / 800nm PR
    • 50W Bias: (recommended)
      • Selectivity to photoresist: 1.10–1.20
      • SiO2 selectivity to Ru: 36
      • SiO2 etch rate: 263nm/min
    • 200W Bias:
      • SiO2 selectivity to Ru: 38
      • SiO2 etch rate: 471nm/min
    • This etch is detailed in the following article: W.J. Mitchell et al., JVST-A, May 2021

Si3N4 Etching (Fluorine ICP Etcher)

Developed by Bill Mitchell. Please see publication policy.

  • ICP = 950/75W
  • Pressure = 5mT
  • Low Polymer Dep: CF4 = 60sccm
    • Etch Rate = 420nm/min (PECVD Si3N4)
  • Higher verticality: CF4 = 35 / CHF3 = 25 sccm
    • Etch Rate = 380nm/min (PECVD Si3N4)

Photoresist & ARC (Fluorine ICP Etcher)

Chain multiple Recipes in a Flow, to allow you to to do in situ BARC etching, and follow up with in situ Photoresist Strip.

PR/BARC Etch (Fluorine ICP Etcher)

  • Etching DUV42P-6 Bottom Anti-Reflection Coating
    • ~60nm thick (2500krpm)
    • O2=20sccm / 10mT / RF1(bias)=100W / RF2(icp)=0W
    • 1min

Photoresist Strip/Polymer Removal (Fluorine ICP Etcher)

    • O2=100sccm / 5mT / RF1(bias)=10W / RF2(icp)=825W
    • 75W Bias can be helpful for difficult to remove polymers, eg. 2min
    • Use laser monitor to check for complete removal, overetch to remove Fluorocarbon polymers.

Cleaning Procedures (Fluorine ICP Etcher)

To Be Added

ICP Etch 1 (Panasonic E646V)

Process Control Data (Panasonic 1)

SiO2 Etch with CHF3/CF4 - Process Control Data (Panasonic 1)

SiO2 Etching (Panasonic 1)

Recipes

Recipe Variations

Use these to determine how each etch parameter affects the process.

SiNx Etching (Panasonic 1)

Al Etch (Panasonic 1)

Cr Etch (Panasonic 1)

Ta Etch (Panasonic 1)

Ti Etch (Panasonic 1)

W-TiW Etch (Panasonic 1)

GaAs-AlGaAs Etch (Panasonic 1)

GaN Etch (Panasonic 1)

Photoresist and ARC Etching (Panasonic 1)

Please see the recipes for Panasonic ICP#2 - the same recipes apply.

Etching of DUV42P at standard spin/bake parameters also completes in 45 seconds.

SiC Etch (Panasonic 1)

Sapphire Etch (Panasonic 1)

Cleaning Recipes

To Be Added

Old Deleted Recipes

Since there are a limited number of recipe slots on the tool, we occasionally have to delete old, unused recipes.

If you need to free up a recipe slot, please contact the tool's Supervisor and they'll help you find an old recipe to replace. We take photographs of old recipes, and save them in case a group needs to revive the recipe. Contact us if your old recipe went missing.

ICP Etch 2 (Panasonic E626I)

Recipes starting points for materials without processes listed can be obtained from Panasonic1 recipe files. The chambers are slightly different, but essentially the same, requiring only small program changes to obtain similar results.

Process Control Data (Panasonic 2)

SiO2 Etch with CHF3/CF4 - Process Control Data (Panasonic 2)

SiO2 Etching (Panasonic 2)

Recipes

Recipe Variations

Use these to determine how etch parameters affect the process.

SiNx Etching (Panasonic 2)

Al Etch (Panasonic 2)

Al2O3 Etching (Panasonic 2)

ALD Al2O3 Etch Rates in BCl3 Chemistry (click for plots of etch rate)

Contributed by Brian Markman, 2018

  • BCl3 = 30sccm
  • Pressure = 0.50 Pa
  • ICP Source RF = 500
  • Bias RF = 50W or 250W (250W can burn PR)
  • Cooling He Flow/Pressure = 15.0 sccm / 400 Pa
  • Etch Rate 50W: 39.6nm/min (0.66nm/sec)
  • Etch Rate 250W: 60.0nm/min (1.0 nm/sec)

GaAs Etch (Panasonic 2)

Photoresist and ARC etching (Panasonic 2)

Basic recipes for etching photoresist and Bottom Anti-Reflection Coating (BARC) underlayers are as follows:

ARC Etching: DUV-42P or AR6 (Panasonic 2)

  • O2 = 40 sccm // 0.5 Pa
  • ICP = 75W // RF = 75W
  • 45 sec for full etching (incl. overetch) of ~60nm DUV-42P (same as for AR6; 2018-2019, Demis/BrianT)

Photoresist Etch/Strip (Panasonic 2)

Works very well for photoresist stripping

  • O2 = 40 sccm // 1.0 Pa
  • ICP = 350W // RF = 100W
  • Etch Rate for UV6-0.8 (DUV PR) = 518.5nm / 1min (2019, Demis)
  • 2m30sec to fully remove UV6-0.8 with ~200% overetch (2019, Demis)

Ru (Ruthenium) Etch (Panasonic 2)

Oxford ICP Etcher (PlasmaPro 100 Cobra)

Process Control Data (Oxford ICP Etcher)

Process Control Data for "Std InP Ridge Etch" Cl2/CH4/H2/60°C

Calibration / Process testing data taken using the "InP Ridge Etch" process: Cl2/CH4/H2 @ 60°C, 1cm piece with ~50% SiO2 hardmask.

Process Control Data for "GaN Etch" (Cl2/BCl3/Ar/200°C)

Recipe: , on 1cm x 1cm ~1.2µm deep GaN etch with Cl2/BCl3/Ar at 200°C. Sapphire substrate with SiO2 mask for GaN.

InP Ridge Etch (Oxford ICP Etcher)

Low-Temp (60°C) Process

  • Low-Temp InP Ridge Etch Characterization - Ning Cao, 2021-09-08
    • InP etches were characterized with no mounting adhesive used, 1/4-wafer of 50mm wafer placed on blank Silicon carriers (rough side up).
    • Recipe: Cl2/CH4/H2 - 60°C
    • NOTE: Rates in these 2021-09 characterizations are lower than current due to a software timing bug, fixed in 2022-01
  • See Operating Procedure for full traveler and post-cleaning.

Sample Size effect on Etch Rate

See the above table for data showing effect on sample size/exposed etched area.

InP Grating Etch (Oxford ICP Etcher)

  • InP/InGaAsP Grating Etch Characterization - Ning Cao, 2021-08-26
    • InP/InGaAsP etches were characterized with no mounting adhesive used, 1/4-wafer of 50mm wafer placed on Silicon carriers (rough side up).
    • Recipe: Cl2/CH4/H2/Ar - 20°C
    • NOTE: Rates in these 2021-09 characterizations are lower than current due to a software timing bug, fixed in 2022-01
  • See Operating Procedure for full traveler and post-cleaning.

GaAs Etch (Oxford ICP Etcher)

This recipe also provides a starting point for GaSb-based etches.

  • GaAs-based materials - etch recipe available on tool - provided by Oxford, not yet qualified internally
  • See Operating Procedure for full traveler and post-cleaning.

GaN Etch (Oxford ICP Etcher)

  • GaN-based materials - etch recipe available on tool - provided by Oxford, not yet qualified internally
  • See Operating Procedure for full traveler and post-cleaning.

GaN Atomic Layer Etching (Oxford ICP Etcher)

GaN-ALE Recipe written and tested by users - contact supervisor for use.

Cleaning Recipes (Oxford ICP Etcher)

To Be Added: Required cleaning time & recipes

Retrieved from "https://wiki.nanofab.ucsb.edu/w/index.php?title=ICP_Etching_Recipes&oldid=162645"