Factors affecting the “non-cleanables”
measurement of resist outgas testing:
interpreting the null results
1
S. Hill, N. S. Faradzhev, B. Berg, T. B. Lucatorto
NIST, Gaithersburg, MD, 20899 USA
M. Barclay, H. Fairbrother
Johns Hopkins University, Baltimore, MD 21218, USA
D. Alvarado, M. Upadhyaya, Y. Kandel, G. Denbeaux
CNSE, State University of New York, Albany 12203,USA
Special Thanks to Tonmoy Chakraborti , SEMATECH
EUVL Symposium, Toyama, Oct. 10, 2013
ASML Resist-Outgas Testing Protocol at NIST
(1) Determine E0 (2) Co-expose witness sample & resist-
coated 200 mm wafer to E0 in 1 hr.
(3) Measure C-thickness with
spectroscopic ellipsometry
and scale to 300 mm wafer (4) Clean C with atomic H
<3nm
(5)
Measure amount
of residual non-C
{S, P, F, I, Cl, …}
with XPS.
2
• Referred to as “non-cleanables,” yet no resist has ever failed due to non-C
residuals.
• Typically XPS does not detect atomic concentrations significantly above
background levels of S.
• F is rarely observed (in contamination spot) despite being common resist
component.
• Resist developers may avoid elements like iodine with high PAG quantum
efficiency due to potential contamination risk.
• Systematic study of actual contamination threat posed by non-C outgas
species is needed.
EUVL Symposium, Toyama, Oct. 10, 2013
EUV/e-beam exposure of polymer films
Spin coat
Ru-MLM substrate
1) Spin coat <10 nm film of polymer onto Ru-cap MLM substrate & characterize
2) Perform EUV/e-beam exposures with varying dose
3) Characterize changes with spectroscopic ellipsometry (SE) and XPS
4) Subject to atomic-H (AH)
5) Characterize with SE and XPS
SE & XPS
EUV AH cleaning
SE & XPS Analyze native
film SE & XPS
1 2 3 4 5
3
S-containing: (C10H18S)n
Poly(3-hexylthiophene)
P3HT or
Repeat to determine cleaning rate
F-containing: (C2H2F2)n
Polyvinylidene fluoride
PVDF
EUVL Symposium, Toyama, Oct. 10, 2013
EUV interaction with P3HT and PVDF
4
XPS
• ~ 40% of S is rapidly desorbed by EUV (<10 J/mm2)
• ~60% of S is resistant to desorption by highest doses
• No evidence of C desorption
P3HT: (C10H18S)n
WS
outgas
test dose
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
S is partially desorbed by EUV
EUVL Symposium, Toyama, Oct. 10, 2013
Consistent with x-ray-induced desorption of F but not S during XPS measurements
(observed by EIDEC and confirmed in collaboration with NIST)
5
F is strongly desorbed by EUV S is partially desorbed by EUV
PVDF: (C2H2F2)n
XPS
P3HT: (C10H18S)n
WS
outgas
test dose
• ~50% of F is rapidly desorbed by low doses
• F continues to desorb with increasing dose
• No evidence of C desorption
• ~ 40% of S is rapidly desorbed by EUV (<10 J/mm2)
• ~60% of S is resistant to desorption by highest doses
• No evidence of C desorption
EUV interaction with P3HT and PVDF
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
WS
outgas
test dose
EUVL Symposium, Toyama, Oct. 10, 2013
Compare EUV & e-beam desorption of F from PVDF
• Dose response from correlation of XPS maps and dose distributions across exposure spots
• Similar trends in EUV and e-beam data suggest two-step desorption process
• Electrons appear to desorb F ~100x more efficiently than EUV
• Does this mean that e-beam-based outgas tests are inherently insensitive to F contamination?
6
Slower second reaction:
Cross sec. ~ 5×10-20 cm2
Fast initial reaction:
Cross sec. ~ 5×10-19 cm2
Slower second reaction:
Cross sec. ~ 2×10-18 cm2
Fast initial reaction:
Cross sec. ~ 2×10-17 cm2
PVDF exposed to EUV PVDF exposed to 2keV e-beam
Johns Hopkins Univ. NIST
EUVL Symposium, Toyama, Oct. 10, 2013
7
XPS
EUV interaction with PVDF
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
PVDF: (C2H2F2)n
F:C before AH cleaning
EUVL Symposium, Toyama, Oct. 10, 2013
8
• The presence of F significantly slows overall cleaning rate of PVDF
• Areas exposed to high EUV doses have lower F:C ratios and hence clean fastest
Atomic-H cleaning of PVDF: C removal
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
XPS
PVDF: (C2H2F2)n
F:C before AH cleaning
EUV-deposited C
EUVL Symposium, Toyama, Oct. 10, 2013
9
Atomic-H cleaning of PVDF: F removal
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
XPS
PVDF: (C2H2F2)n
F:C before AH cleaning
EUVL Symposium, Toyama, Oct. 10, 2013
• F is removed by AH at very slow rate
• The amount of F detected in XPS portion of outgas test could be artificially lowered by
Excessively long AH cleaning times (cleaning until all C is gone)
Witness sample intensity much higher than that expected on lowest-intensity NXE optic
10
• AH removed all P3HT after first short cleaning interval
• AH cleaning rate for P3HT (S:C~10%) is at least as fast as typical EUV-deposited C.
• S is removed at least as quickly as C and does not affect C cleaning rate
AH cleaning of P3HT (S-containing polymer)
Ide
al s
toic
hio
me
try
Un
ex
po
se
d
XPS
PVDF: (C2H2F2)n
F:C before AH cleaning
P3HT lower bound
EUV-deposited C
EUVL Symposium, Toyama, Oct. 10, 2013
Studying the effectiveness of cleaning
• After over 100 customer samples, none have failed the non-
cleanables specification
• We did more in depth measurements of one sample – it started
with 15 nm of contamination growth
Measured XPS of the contamination spot PRIOR to cleaning
Partially cleaned multiple times with XPS after each clean cycle to
measure the composition within the contamination, during cleaning, and
after cleaning
XPS XPS XPS XPS
Partial clean Partial clean Partial clean
ellipsometry ellipsometry ellipsometry ellipsometry
Atomic hydrogen cleaning
EUVL Symposium, Toyama, Oct. 10, 2013 11
Composition between cleaning steps
• Primarily carbon contamination spot, after cleaning primarily ruthenium substrate
• Oxygen is present in the contamination, but at a higher level on the ruthenium substrate
carbon ruthenium
oxygen
thickness
0
2
4
6
8
10
12
14
16
18
0
10
20
30
40
50
60
70
80
90
100
0 5 10 15 20 25
Thic
kne
ss (
nm
)
Ato
mic
co
nce
ntr
atio
n (
%)
Cleaning time (hours)
Nitrogen, sulfur, silicon
EUVL Symposium, Toyama, Oct. 10, 2013 12
Composition between cleaning steps
• Sulfur is present in both the contamination spot and the final cleaned ruthenium surface
• Sulfur in contamination is elemental, in/on ruthenium surface is sulfur oxide
thickness
0
2
4
6
8
10
12
14
16
18
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 5 10 15 20 25
Thic
kne
ss (
nm
)
Ato
mic
co
nce
ntr
atio
n (
%)
Cleaning time (hours)
silicon
nitrogen
sulfur
5 - 130929A_CG150_CG151_CG103.104.spe: 130929A_CG150_CG151_CG103 PHI
2013 Oct 1 Al mono 64.9 W 200.0 µ 45.0° 224.00 eV 1.8187e+004 max 6.73 min
S2p/9/1
1551601651701751.73
1.74
1.75
1.76
1.77
1.78
1.79
1.8
1.81
1.82x 10
4 5 - 130929A_CG150_CG151_CG103.104.spe
Binding Energy (eV)
c/s
1 - CG103 & CG106.102.spe: CG103 & CG106 PHI
2013 Apr 15 Al mono 44.6 W 200.0 µ 45.0° 224.00 eV 3.8033e+003 max 6.73 min
S2p/3/1
1551601651701752200
2400
2600
2800
3000
3200
3400
3600
3800
40001 - CG103 & CG106.102.spe
Binding Energy (eV)
c/s
Sulfur in
contamination
primarily
elemental
Sulfur on
witness plate
in partially
oxidized state
EUVL Symposium, Toyama, Oct. 10, 2013 13
Summary
14
• S-containing polymers and outgas contamination appear to clean at rate
similar to pure EUV-deposited C.
• Atomic-H cleaning rate of C is significantly slowed by presence of F.
• F is rarely observed in outgas testing because it is efficiently desorbed by
EUV & electrons not because it is efficiently cleaned by atomic H.
• Electrons desorb F (from polymer PVDF) ~100x more efficiently than EUV
• Is e-beam outgas testing inherently insensitive to F contamination?
• NIST and EIDEC systems use EUV on witness sample and have reported small
amounts of F (< 1 atomic %) before AH cleaning.
• Has F ever been seen in any e-beam test?
EUVL Symposium, Toyama, Oct. 10, 2013