ICP-RIE of Ultrananocrystalline Diamond...Optimization of ICP-RIE, Oxford 100 NCCAVS Plasma...
Transcript of ICP-RIE of Ultrananocrystalline Diamond...Optimization of ICP-RIE, Oxford 100 NCCAVS Plasma...
Advanced Diamond Technologies, Inc. © 2007. All rights reserved. We haven’t even scratched the surfaceSM
ICP-RIE of Ultrananocrystalline
Diamond
UNCD
OP
NP
CO2
-e
N. Moldovan, H. Zeng, J. A. Carlisle
Ralu Divan
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
Contact: N. Moldovan
www.thindiamond.com
(815) 905-9395
Outline
• Motivation
• Preliminary thoughts on direct etching
• Optimization of diamond etching with ICP-RIE
• Discussion of tip etching results
• Conclusions
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
Ultrananocrystalline diamond (UNCD)
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
Smooth: 3-20 nm rms as grown, <1 nm “Horizon”
Thickness: 40 nm to 20 mm; <5% uniformity
Tunable electrical and thermal conductivities
High Young’s modulus (~900 GPa)
Biocompatibility
Up to 300 mm wafers
Most of bulk diamond mechanical properties
Deposited at 400-900 oC
HFCVD MPCVD
UNCD micro-manufacturability
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
Selective seeding (SiO2)
700-mm-long, 1.5-mm-thick,
2 to 10-mm-width
200 mm
Selective seeding through PRSelective seeding
through photoresist or SiO2 mask
O2-RIE through Al mask
Rounded edges
~1-2 mm resolution
Vertical edges, rough at the nanoscale
~ 1-2 mm UNCD
UNCD nano-manufacturability
E-beam lithography, lift off (Al, Ni), O2-RIE of UNCD, release by KOH (XeF2) etch
Molding
Tip-based nanopatterning of seeds and growth
2004
2006
NaDiaProbes, ADT
2008
2009
Why is molding not good enough?
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
<10nm
It can be spectacular, but…
Why is molding not good enough?
Top look into a V-groove, showing a wedge.
Wedge-shaped tip
Double tip after oxidation-sharpening
Why is molding not good enough?
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
Statistics
Best median value of wedge size: 13.3 nm
Preliminary thoughts on direct etching
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
Two ways:
• Isotropic underetching
Case of Si in XeF2
• Slope amplification by differential etching & sharpening
UNCD ICP-RIE – needs optimization
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
RIE : 1mm in 30 min (33 nm/min)O2:CF4 (50:2 sccm), 300 W, 20 mTorr
ICP-RIE: 1 mm in 10 min (100 nm/min)
O2:CF4 (50:2 sccm), ICP= 2800 W, RIE=100 W, 20mTorr
Our experience:
Literature:
O2-RIE: 30-40 nm/min
ECR O2 plasma, 400 nm/min (Pearton et.al. Electron. Lett. 28, 822 1992.)
Optimization of ICP-RIE, Oxford 100
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
300
250
200
150
100
50
0
Etc
hin
g r
ate
(n
m/m
in)
43210SF6 sccm
ICP=2500 WRIE=100WO2=50 sccmP=20 mTorr
T=20 oC
400
300
200
100
Etc
hin
g r
ate
(n
m/m
in)
25002000150010005000
ICP power (W)
RIE=100 WO2=50 sccm
T=20 oC
P=10 mTorr
500
400
300
200
100
0
Etc
hin
g r
ate
(n
m/m
in)
200150100500
RIE Power (W)
ICP= 2500 WO2=50 sccm
P= 10 mTorr
T= 20 oC
400
300
200
100
0
Etc
h R
ate
(n
m/m
in)
50403020100O2 Pressure (mTorr)
ICP=2500 WRIE=100 WO2=50 sccm
T=20 oC
Thermal oxidation of UNCD in air
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
• Diamond etches at > 600 oC in air(but slowly: 25 nm/min)
1600
1400
1200
1000
800
600
400
200
Th
ickn
ess r
em
oved
(n
m)
1000900800700600500
Temp (oC)
Oxidation in air1 hour
Stationary because entire film is etched
ICP-RIE rate with temperature
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
800
600
400
200
0
Etc
hin
g r
ate
(n
m/m
in)
200150100500-50-100
Temperature (oC)
ICP= 2800 WRIE= 200 WO2= 50 sccm
SF6= 0.5 sccm
P= 9 mTorr
• Almost no temperature dependence
• Isotropy unlikely in this regime
• Machine tripped above 300 oC (cooling problems)
• Do the isotropic etching by thermal oxidation sequentially
Investigating isotropy: test mask
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
• 100 mm x 100 mm field
• 3.5 mm UNCD/350 nm PECVD SiO2
• Exposed with Raith 150, (30 keV, 60 um aperture, 1554 pA, dose 400 mC/cm2)
• Resist: ma-N 1405 (MicroChem) @ 2000 rpm, baked at 100oC for 90 sec, thickness 350 nm, developed 30 sec in mD533S
• SiO2 etched by ICP-RIE (CHF3-Ar plasma)
• UNCD etched half way for rate measurement, then continued to full removal
Etched structures on 3.5 mm films
EIPBN 2009 Marco Island, FL.
Etched structures: best conditions for isotropy
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
Sample etched incompletely at 300oC ICP=2500W, RIE=50 W, P=50 mTorr , O2=50 sccm
• No sign of underetching
• No sign of isotropy
• Ultra-sharp “grass” originating from SiO2 in crevasses (~50 nm) on the surface, generating spikes by differential etching mechanism
• Mostly vertical walls, except the top of structures, where edges are curved by differential etching when the oxide vanishes
Thermal oxidation after ICP-RIE
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
Before After 675oC 30 min, air
Oxidation proceeds by roughing the surface and generation of cavities
Boosting differential etching
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
• 6.5 mm of UNCD
• 350 nm of SiO2• ICP= 2800W, RIE= 200 W, P= 9 mTorr
SF6= 0.5 sccm, O2= 50 sccm, T= 20 oC,
R<10 nm
From
F = 3 mm circle
12 min
Pleasant surprises: Nanowire ending tips
Are not that rare!
14%?
Statistics rules!
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Oc
cu
ren
ce
pro
ba
bil
ity
10009008007006005004003002001000
Wedge size (nm)
Molded tips, circular precursors
ICP-RIE, 3 mm circle precursors
Molded tips, circular precursors (test mask)
Molded tips, square precursors
Median tip radii 13.3 nm 23 nm 48.5 nm120 nm
• Tips considered: from the 6.5 mm batch, 3 mm circles
• ICP-RIE tips are somewhat worse than the best molded tips
• ICP-RIE tips statistics improved by the NW-ending tips
ICP-RIE: etching of pores*
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
100-nm-diameter pores 80 nm pores
1 mm low stress UNCD
Same ICP-RIE recipe
*to be presented at EIPBN 2010, O.Makarova et.al.
Conclusions
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.
• Sharp tips can be obtained by ICP-RIE of UNCD, R<10 nm
• Isotropic underetching does not work
• Differential etching and sharpening work
• Post-RIE Oxidation sharpening does not work – roughness
• ICP-RIE was optimized: DRIE of UNCD?
• Maximum rate (674 nm/min) with ICP=2800W, RIE=200W, P = 9 mTorr, O2 = 50 sccm, SF6 = 0.5 sccm, T = 20 oC
• Technique is good for arrays of tips (field emission?), but still inferior to molding for AFM probes
•Some AFM probes may benefit from feathery shape – DPN?
ACKNOWLEDGMENTSUse of the Center for Nanoscale Materials, Argonne National Laboratory was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357
Q&A
NCCAVS Plasma Applications Group Meeting, San Jose, Febr 10, 2010
.