Area-selective atomic layer deposition of platinum on YSZ substrates using microcontact printed SAMs

Article abstract of DOI:10.1149/1.2789301

Using (methylcyclopentadienyl)trimethylplatinum (MeCpPtMe3) and oxygen as precursors, Pt has been deposited by atomic layer deposition (ALD) on the surfaces of yttria-stabilized zirconia (YSZ), a solid oxide electrolyte, as well as on oxide-covered silicon. Ex situ analyses have been carried out to examine the properties of both as-deposited and postannealed Pt films. X-ray photoelectron spectroscopy measurements demonstrate that there are no detectable impurities in the as-deposited Pt films, and four-point probe measurements show that the resistivity for a 30.2 nm film is as low as 18.3 μ cm. The use of area-selective ALD to deposit patterned Pt has also been investigated. By coating these same substrates with octadecyltrichlorosilane (ODTS) self-assembled monolayers (SAMs), Pt ALD can be successfully blocked. Furthermore, it is shown that by transferring the ODTS SAMs to the substrates by microcontact printing (μCP) using patterned stamps, platinum thin films are grown selectively on the SAM-free surface regions. Features with sizes as small as 2 μm have been deposited by this combined ALD-μCP method; the resolution is limited by the printed pattern and, likely, can be achieved at dimensions significantly smaller than a micrometer.

Full text of DOI:10.1149/1.2789301

D648
Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
0013-4651/2007/154͑12͒/D648/9/$20.00 © The Electrochemical Society
Area-Selective Atomic Layer Deposition of Platinum on
YSZ Substrates Using Microcontact Printed SAMs
Xirong Jianga and Stacey F. Bentb,z
aDepartment of Physics and bDepartment of Chemical Engineering, Stanford University, Stanford,
California 94305, USA
Using ͑methylcyclopentadienyl͒trimethylplatinum ͑MeCpPtMe3͒ and oxygen as precursors, Pt has been deposited by atomic layer
deposition ͑ALD͒ on the surfaces of yttria-stabilized zirconia ͑YSZ͒, a solid oxide electrolyte, as well as on oxide-covered silicon.
Ex situ analyses have been carried out to examine the properties of both as-deposited and postannealed Pt films. X-ray photo-
electron spectroscopy measurements demonstrate that there are no detectable impurities in the as-deposited Pt films, and four-point
probe measurements show that the resistivity for a 30.2 nm film is as low as 18.3 ␮⍀ cm. The use of area-selective ALD to
deposit patterned Pt has also been investigated. By coating these same substrates with octadecyltrichlorosilane ͑ODTS͒ self-
assembled monolayers ͑SAMs͒, Pt ALD can be successfully blocked. Furthermore, it is shown that by transferring the ODTS
SAMs to the substrates by microcontact printing ͑␮CP͒ using patterned stamps, platinum thin films are grown selectively on the
SAM-free surface regions. Features with sizes as small as 2 ␮m have been deposited by this combined ALD-␮CP method; the
resolution is limited by the printed pattern and, likely, can be achieved at dimensions significantly smaller than a micrometer.
© 2007 The Electrochemical Society. ͓DOI: 10.1149/1.2789301͔ All rights reserved.
Manuscript submitted April 18, 2007; revised manuscript received August 15, 2007. Available electronically October 11, 2007.
ing, and metallic surfaces. Their potential applications include the
control of wetting and adhesion, tribology, chemical sensing,
ultrafine-scale lithography, and protection of metals against
corrosion.19-21 In our area-selective ALD process, the SAM is used
to modify the chemical properties of the substrate surface such that
it resists the atomic layer deposition chemistry.
In this paper, we briefly describe the characterization of the Pt
ALD process, demonstrate the properties of the as-deposited and
postannealed Pt thin film grown on the substrates of both Si and
YSZ, and illustrate an area-selective deposition process to achieve
spatially patterned Pt on the same substrates. Our approach to area-
selective deposition consisted of two key steps. First, the patterned
SAMs were formed using microcontact printing ͑␮CP͒. Second, the
Pt thin films were selectively deposited onto the SAM-patterned Si
and YSZ substrates by ALD.
Pt has enormous application prospects in catalysis and
microelectronics.1,2 For example, due to its chemical stability in
both oxidative and reductive environments, and its excellent electri-
cal properties at high temperatures,3 Pt is used as the electrode ma-
terial in nonvolatile ferroelectric random access memory devices
and high dielectric capacitors. It also exhibits excellent catalytic
activity for a number of reactions, including the O2 reduction reac-
tion at the cathode of a solid oxide fuel cell ͑SOFC͒, and is espe-
cially useful at the lower operating temperatures ͑below 600°C͒ that
are desired for integratable fuel cell systems.4
We are exploring whether the technique of atomic layer deposi-
tion ͑ALD͒ can be used to deposit Pt5 on an yttria-stabilized zirconia
͑YSZ͒ solid oxide electrolyte for the catalyst in SOFCs. ALD is a
thin-film deposition method that can achieve excellent thickness
control down to the subnanometer scale and may provide the ability
to deposit ultrathin Pt films required for complex fuel cell
geometries.6 In addition, we are working to apply area selective
ALD to deposit patterned thin films of Pt for use as the current
collector grid of the SOFC. The ALD method relies on sequential
saturated surface reactions, which result in the formation of up to
one monolayer each cycle. The mechanism of successive self-
terminating growth steps in ALD, and the pulse sequence, which
eliminates gas-phase reactions, confines the reaction to the substrate
surface.6,7 For example, the Pt ALD process includes two self-
limiting chemical reactions, repeated in the alternating ABAB se-
quences shown below
Experimental
All chemical reagents, including octadecyltrichlorosilane
͑ODTS͒ ͑97%͒, tridecafluoro-1,1,2,2-tetrahydrooctyltrichlorosilane
͑FOTS͒ ͑97%͒, toluene ͑anhydrous, 99.8%͒, and chloroform ͑99%͒,
used to form SAMs, were purchased from Aldrich ͑Milwaukee, Wis-
consin͒. Poly͑dimethylsiloxane͒ ͑PDMS͒ ͑Sylgard 184͒ was pur-
chased from Dow Corning. The ͑methylcyclopentadienyl͒tri-
methylplatinum ͑MeCpPtMe3͒ ͑99%͒ precursor for ALD was ob-
tained from Strem. All these reagents and gases were used directly
without further treatment. Nitrogen gas ͑99.9995%͒ was obtained
from Praxair and purified by a gas filter from Nanochem ͑model no.:
PF-25͒ before use.
Si substrates were cut from Si͑100͒ wafers purchased from Si-
Tech, Inc. ͑p-type with boron dopant; resistivity of 1.0–10.0 ⍀ cm͒.
Si substrates were cleaned by a standard piranha process, resulting
in a chemical oxide-coated silicon substrate. Thin-film YSZ sub-
strates, deposited by ALD on Si͑100͒, were provided by Friedrich
Prinz’s group at Stanford University.22 Two steps were used to clean
the YSZ substrates.23 First, the substrates were rinsed ultrasonically,
twice, for 5 min in ethyl alcohol in order to remove grease. After the
first rinse, the ethyl alcohol was renewed. Following the ethyl alco-
hol rinses, the YSZ substrates were exposed to rf oxygen plasma for
10 min to make the surface hydrophilic.
Pt͑s͒ + O2 → Pt–O*
͓1͔
CH3C5H4Pt͑CH3͒3 + Pt–O* → Pt*͑s͒ + CO2͑g͒↑ + H2͑g͒↑
+ other by-products↑
͓2͔
where the asterisks represent the surface species.5,8,9 Because the
ALD process is very sensitive to the surface functional groups, it
offers an ideal method for selective deposition of thin films on pat-
terned substrates in which the substrate surface properties are spa-
tially modified.10,11
We have shown in earlier work that microcontact printed ͑␮CP͒
self-assembled monolayers ͑SAMs͒ can be used to achieve spatially
patterned deposition of HfO2 by area selective ALD.12 Others have
used a similar approach for selective ALD of films such as TiO2,
ZnO, and Ru.13-18 SAMs are thin organic films that spontaneously
form on solid surfaces and are well known for modifying the physi-
cal, chemical, and electrical properties of semiconducting, insulat-
The samples were loaded into a custom-built, flow-type ALD
system. Pt thin films were deposited onto Si and YSZ substrates
using MeCpPtMe3 and air, providing a source of oxygen, as ALD
precursors. The air was dried through a desiccant air dryer to a dew
point temperature of −45.5°C before use. Nitrogen was used as both
a purging gas and a carrier gas. The MeCpPtMe3 and air were kept
at 50°C and 21°C, respectively. A regular ALD cycle consisted of a
z E-mail: sbent@stanford.edu
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
D649
The growth rate was determined by dividing the film thickness
measured by ellipsometry by the number of ALD cycles. In the
investigation of growth rate dependence on ALD conditions, mul-
tiple locations on at least two samples obtained from separate 100
cycle ALD processes were probed to determine thickness. To probe
the ODTS film degradation in the ALD Pt process, we used XPS to
determine the resulting ODTS film thickness; after curve fitting of
the XPS C 1s, O 1s, and Si 2s core level peaks, an electron escape
depth calculation was carried out to estimate the film thickness.
2 s exposure to MeCpPtMe3, a 12 s N2 purge, a 2 s exposure to dry
air, and a 12 s N2 purge. The total flow rate of the N2 was 15 sccm.
For the area-selective area ALD processes, the purging time was
extended to 30 s to ensure removal of any unreacted precursor from
the substrate. Except for the study of substrate temperature depen-
dence, for which the temperature was varied, Pt thin films were
grown with the substrate temperature at 285°C at a reactor pressure
of 700 mTorr for a varying number of cycles.
All procedures for SAM formation were performed in a dry
glove box at room temperature. Previous studies10,11 have shown
that the deactivating effect of ODTS SAMs depends on its quality
and that densely packed SAMs are required for good deactivation.
For preparation of a densely packed, continuous ͑unpatterned͒ film,
the cleaned Si and YSZ substrates were dipped in 10 mM solutions
of ODTS in toluene for two days without disturbance. Shorter dip-
ping times were also examined to study the dependence of the resist
quality on the ODTS SAM formation time. The unpatterned ODTS-
coated samples were examined using water contact angle measure-
ments and ellipsometry before loading them into the ALD reactor.
It was found that the working environment was very important
for the SAM formation, as previously reported in the literature.24-26
In our initial experiments, we used a simple dry-air-purged glove
box. However, with this system, we found there was a high likeli-
hood of forming ODTS multilayers instead of monolayers, due to
the water content in the glove box, leading to bulk polymerization of
the silane before it reacted with the surface. We subsequently
switched to a nitrogen-purged commercial glove box ͑from Vacuum
Atmosphere Co.͒ that had an oxygen content below 0.1 ppm and a
water content below 0.7 ppm. Using this glove box, however, only
partial growth or sometimes no growth of ODTS SAMs was ob-
tained. The poor growth of the SAM in the higher-quality glove box
was attributed to the lack of water because small amounts of water
are necessary in the silylation and cross-linking reactions.19,27-29
Thus, for the remaining SAM experiments, we used the simpler
glove box, but took greater care in eliminating sources of water.
Patterned monolayers on the Si and YSZ substrates were made
using standard ␮CP procedures.30 Masters were fabricated from pat-
terned silicon wafers by conventional photolithography. The masters
had a mesh structure with lines and spacings of dimensions from
2 ␮m to 50 ␮m and a line depth of 7 ␮m. The PDMS stamps were
cast according to a previously reported procedure.30 The prepared
PDMS stamps were then wetted with a 10 mM ODTS or 10 mM
FOTS solution, both with toluene as the solvent, and dried by nitro-
gen flow. The dimensions of the ODTS soaked stamp will change
with wetting time because the PDMS stamp swells in the toluene
solution; consequently, the wetting time was held to 1 min, a time
that we have found can retain the patterns on the stamp. Subse-
quently, the stamp was placed in contact with the Si or YSZ sub-
strate at room temperature for varying lengths of time. After peeling
off the stamp, the samples were rinsed with toluene and chloroform,
sequentially, and dried by nitrogen flow, leaving behind patterned
SAMs formed on the substrates. These substrates were then loaded
into the ALD reactor for Pt deposition.
Results and Discussion
Atomic layer deposition of platinum thin films
.— Pt thin-film
deposition by ALD was initially studied by Aaltonen et al.5 Using
the same precursors, MeCpPtMe3 and air, we deposited Pt thin films
by ALD on both Si and YSZ substrates and explored the optimal
ALD conditions that were characteristic of our system.31 Here, we
highlight only a few significant points. First, we indentified the op-
timal deposition temperature in our system to be 285°C. Based on
the previously reported data,5 we investigated substrate temperatures
between 225 and 345°C. We observed that from 225 to 285°C, the
Pt film growth rate increases with temperature, stays constant at
ϳ0.4 Å/cycle from 285 to 300°C, then increases again in the tem-
perature range from 310 to 345°C. The growth rate behavior is in
good accordance with that reported by Aaltonen et al.5 The larger
growth rate at higher substrate temperatures suggests partial decom-
position of the precursor and formation of nonvolatile cracking
products; this decomposition must be avoided in the ALD process
because it destroys the self-saturation growth mechanism and may
lead to Pt films with higher impurity content. Furthermore, it may
also result in the failure of the area selective ALD process. For the
subsequent experiments, a deposition temperature of 285°C was
used to allow a relatively high growth rate while minimizing the
possibility of precursor decomposition.
We also probed the optimal precursor feed temperature. The Pt
precursor ͑melting pointing 31°C͒ as purchased was an off-white
powder. Measurements at room temperature ͑ϳ21°C͒ revealed no
observable pressure increase when the precursor line was opened to
the ALD reactor, and no growth was measured. Hence, in a diver-
gence from the setup of Aaltonen et al.,5 we elevated the platinum
precursor temperature in all of our ALD processes. The Pt precursor
temperature range investigated here was from 32 to 75°C. The
growth rate was observed to increase with temperature when the
metal precursor temperature was Ͻ50°C and reached a constant
value at temperatures Ͼ50°C. Thus, the optimal metal precursor
temperature was set at 50°C, and this temperature was also used for
further experiments.
We observed an interesting effect of the length of the air precur-
sor pulse. In studies of the growth rate dependence on the air pulse
time over the range of 1.5–3 s, the growth rate was observed to rise
with air pulse time, reaching a peak value of 0.43 Å per cycle when
the air pulse time is at 2 s, then decrease at longer air pulse times.
The correlation of growth rate with the air pulse time is different
from that observed by Aaltonen et al., who reported saturation at
longer air pulse times.5 There are several possible explanations for
the decreasing growth rate observed here. The growth rate may de-
crease if the adsorption process of MeCpPtMe3 is reversible, with Pt
precursor desorbing from the substrate with time. This explanation
is supported by the fact that the film growth rate did not reach
saturation, but continued to decrease from 0.5 Å/cycle to
0.29 Å/cycle, when the N2 purging time after the Pt precursor dose
was increased from 7 s to 60 s. Another possible reason is that im-
purities in the dry air source may deactivate active sites on the
substrate.32-36
The film thickness of Pt was measured by ellipsometry and X-ray
reflectometry ͑XRR͒. A single-wavelength ellipsometer was used for
the measurement of Pt thin films at thicknesses below ϳ5 nm. At
larger thicknesses ͑Ͼ5 nm͒, absorption becomes significant which
may lead to inaccurate measurements. For the X-ray reflection mea-
surements, we used a PANalytical X’Pert PRO X-ray diffraction
system to do reflectometry analysis and obtain the film thickness.
The Pt film density was also characterized by XRR by averaging
multiple points on a sample at each film thickness studied. Scanning
electron microscopy ͑SEM͒ and atomic force microscopy ͑AFM͒
were used to determine the Pt film morphology. X-ray photoelectron
spectroscopy ͑XPS͒ analysis was performed to investigate the film
composition, and Auger electron spectroscopy ͑AES͒ ͑Evans Ana-
lytical Group͒ was used for elemental mapping of the patterned
structure. Four-point probe measurement was used to determine the
Pt film resistivity.
The composition of the Pt films deposited by ALD was investi-
gated by X-ray photoelectron spectroscopy ͑XPS͒. XPS spectra of a
Pt film deposited by 500 ALD cycles both as-deposited and after
mild sputtering are shown in Fig. 1. The as-deposited film was found
to consist of 46% platinum, 46% carbon, and 8% oxygen. The pres-
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D650
Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
Figure 3. ͑Color online͒ AFM images of Pt thin films vs the number of ALD
cycles. ͑a͒–͑c͒ are Pt films on native oxide coated Si wafers, ͑d͒–͑f͒ are Pt
films on YSZ substrates. The rms roughness values of Pt films on Si sub-
strates are 1.5–2.8 nm, and the rms roughness of Pt films on YSZ substrates
are 2.2–3.2 nm. The Z range of the AFM images is 8 nm.
sensitivity. X-ray reflectometry was employed to investigate the as-
deposited Pt film density as a function of film thickness. Figure 2
shows that the film density increases with the film thickness and is
saturated at 21.3 g/cm3 for films thicker than 18.3 nm. Because the
bulk density is 21.4 g/cm3, the results indicate that the Pt films
thinner 18.3 nm are discontinuous and porous. This discontinuous
structure may be formed due to the island growth mechanism, in
which islands of fixed lateral dimension first grow to a given height
and then spread to cover the surface.33,34
Evolution of the surface morphology of the as-deposited Pt films
was examined by AFM as the number of ALD cycles was increased.
AFM images collected after 240, 450, and 700 ALD cycles are
shown in Fig. 3 for both Si and YSZ substrates. The rms roughness
of the Pt films on both substrates, determined by AFM, are listed in
Table I, together with Pt film thickness measured by X-ray reflecto-
metry. The AFM study indicates that the rms roughness decreases
with increasing number of ALD cycles. RMS roughness for a 30 nm
Pt film was as low as 2.2 nm on YSZ, indicating that the Pt films are
quite smooth. The Pt films deposited on the Si substrates were
slightly less rough than those deposited on the YSZ substrates, likely
reflecting the smoother starting surface of Si compared to YSZ
͑0.3 nm vs 3.7 nm͒. In addition to morphology, the resistivity of the
films was also measured with the four-point probe method, and the
results are included in Table II, which shows that the resistivity
decreases with increasing film thickness, with a resistivity of
18.3 ␮⍀ cm measured for the 30 nm thick film. This resistivity
compares reasonably well to the bulk Pt value ͑10.8 ␮⍀ cm͒.37
To investigate the temperature dependence of the film morphol-
ogy and resistivity, a postannealing process was carried out in an O2
environment for 30 min at each temperature. Table I lists the rms
Figure 1. XPS spectrum of a 20 nm Pt film on a Si substrate ͑a͒ before argon
ion sputtering and ͑b͒ after removing a ϳ1 nm superficial surface layer by
argon ion sputtering. ALD conditions: substrate temperature: 285°C; precur-
sor temperature: 50°C; dosing time: 2 s; purge time: 12 s; ALD cycles: 500
cycles.
ence of carbon and oxygen can arise from environmental contami-
nation occurring after the ALD process. To confirm this, we used in
situ argon ion sputter etching to remove the topmost layers. Fig. 1b,
taken after sputtering ϳ1 nm of material from the surface, reveals
that the film is pure Pt with no observable contaminants to 0.1%
Table I. RMS roughness of as-deposited and postannealed Pt
films measured by AFM. The film thickness was determined by
XRR. Note: rms roughness of YSZ substrate and Si substrate was
3.7 nm and 0.3 nm, respectively, prior to deposition.
RMS of Pt films
on YSZ substrate
͑nm͒
RMS of Pt films
on Si substrate
͑nm͒
Pt film
thickness
͑nm͒
As-
As-
deposited 350°C 550°C deposited 350°C 550°C
Figure 2. The film density of the as-deposited Pt film vs film thickness. ALD
conditions: substrate temperature: 285°C; precursor temperature: 50°C; dos-
ing time: 2 s and purge time: 12 s. The error bars reflect one standard de-
viation.
11.2
18.3
30.2
3.2
2.5
2.2
3.5
2.9
2.6
3.5
3.5
2.6
2.8
1.8
1.5
2.1
1.6
1.5
1.6
1.9
4.8
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
D651
Table II. Resistivity „␮⍀ cm… of as-deposited and postannealed
Pt films. Dash (–) indicates no measurement was made.
Resistivity of
Pt films on YSZ
substrate
Resistivity of
Pt films on Si
substrate
͑␮⍀ cm͒
͑␮⍀ cm͒
Pt film
As-
As-
thickness ͑nm͒
deposited
350°C
550°C
deposited
11.2
18.3
30.2
31.3
22.6
18.3
23.3
18.8
17.2
22.5
–
15.8
–
41.9
16.3
roughness of both the as-deposited and the postannealed films at two
different annealing temperatures ͑350 and 550°C͒. Measurements
were made on two separate samples for each entry in Table I. The
data show that the rms roughness of the postannealed films depos-
ited on YSZ is slightly increased compared to that of the as-
deposited films. An exception is that of the 30.2 nm Pt film on Si,
for which there is a large increase in the rms roughness on annealing
to 550°C; this increase in the film rms is likely due to agglomeration
of Pt islands at elevated temperatures. Table II lists the resistivity of
both as-deposited and postannealed Pt films deposited by ALD.
Post-ALD annealing studies show that the resistivity of the Pt film
remains low for temperatures up to at least 550°C, which suggests
that Pt is an excellent catalyst and electrode material for SOFCs
operated at low temperature ͑Ͻ600°C͒.
Figure 4. Time dependence of ODTS film formation and its ALD blocking
effectiveness on 100 cycles of Pt ALD. Water contact angle measured on the
ODTS film ͑᭿͒, ODTS film thickness ͑᭺͒, Hf atomic percent ͑᭡͒ and Pt
atomic percent ͑*͒. Data for HfO2 taken from Ref. 18.
Another set of ODTS-coated substrates prepared in the same
process was introduced into the ALD reactor for 100 cycles of
atomic layer deposition of Pt. XPS studies of these substrates after
the Pt ALD process show that the Pt atomic percent ͑Fig. 4͒ is
inversely correlated with the contact angle and thickness of the
monolayer resist. The Pt atomic percent dropped very quickly to
below the XPS sensitivity limit when the dipping time was above
12 h. This selectivity was likely achieved because there were no
active sites presented on the ODTS-covered surface, preventing the
nucleation of Pt precursors and subsequent film growth.5,8
Area-selective ALD of platinum.— In order to carry out area-
selective ALD of Pt, a good resist system is needed to block the
substrate surface active sites. In our studies, ODTS SAMs were
explored as the ALD resist. In the formation of the ODTS SAM, the
chlorine atoms in the ODTS molecules are hydrolyzed by reacting
with trace water. The hydrolysis step is followed by the reaction
between the molecular hydroxide groups and the silanol groups
present on the surface to form a covalent Si–O–Si bond, resulting in
a monolayer containing molecules that are both covalently linked to
the substrate as well as to their neighbors. Some reaction between
precursors is thought to occur in solution, leading to small molecular
clusters that then react to form Si–O–Si bonds at the surface.19,27-29
However, too much water present in the solution will result in bulk
polymerization of the silane before it reacts with the surface. Under
this condition, ODTS films will grow as islands, which have a high
density of defects, instead of as a smooth, densely packed
monolayer.19,38,39 Thus, care was taken to control the amount of
water in the glove box used for the SAM experiments. All Si–Cl
bonds were fully reacted in the hydrolysis process, as indicated by
the absence of chlorine in an XPS analysis of the ODTS-coated
substrates.
Following the ODTS SAM formation, samples were character-
ized by ellipsometry and water contact angle. A systematic study of
the deactivating effect of octadecyltrichlorosilane ͑ODTS͒ SAMs
toward Pt ALD was carried out. The results are shown in Fig. 4,
which also includes information on the silylation time dependence
for ODTS SAM formation by the measurement of film thickness and
contact angle of the SAM. The film thickness of the ODTS SAM
increased with SAM-formation time until a plateau was reached at
values of 26 Å, and the contact angle was found to exhibit similar
behavior, reaching a maximum value of 110°, which is consistent
with previous studies.10,19,39,40 The contact angle measurement re-
veals the hydrophobicity of the monolayer, and the change in the
contact angle roughly indicates the extent of the monolayer surface
coverage as well as the variation in the surface chemical composi-
tion of the substrate.41-44 Approximately, 2 h were required for the
water contact angle of the ODTS SAM to reach a plateau, indicating
that the formation of a closely packed monolayer takes at least 2 h.
In an earlier set of studies, we investigated the ability of ODTS
to block HfO2 ALD.10,11 To compare the ODTS blocking effect on
HfO2 vs Pt ALD, the corresponding Hf atomic percent data as a
function of SAM formation time is also included in Fig. 4. Although
the trend is similar, it takes 42 h to fully block the HfO2 ALD
process while only 12 h are needed to fully block the Pt ALD
growth. Thus, the Pt ALD growth is more easily deactivated by the
SAM, as indicated by the absence of observable Pt on ODTS films
with shorter silylation times. Here, we speculate as to two possible
reasons for the difference between Pt and HfO2. The first explana-
tion is based on the difference of the sizes of the two metal precur-
sors. For the growth of HfO2 and Pt, hafnium tetrachloride and
͑methylcyclopentadienyl͒trimethylplatinum were used as the metal
precursors, respectively. The ligands of MeCpPtMe3 are larger com-
pared to that of the hafnium tetrachloride. For loosely packed ODTS
film, there are defects, or pinholes, present in the surface.11 If the
size of the pinholes is larger than that of the hafnium precursors,
then the precursor can diffuse through the pinholes and react at the
substrate to nucleate HfO2 growth. If the pinhole size is smaller than
that of the MeCpPtMe3 molecule, then the Pt precursor will not
reach the surface active sites due to steric hindrance. Thus, Pt ALD
growth can be deactivated even by an imperfectly packed ODTS
SAM. Besides the geometric point of view, the inherent chemical
reactivity of the precursors used in the two processes may account
for the difference. The ability to block ALD with shorter silylation
times for the Pt process suggests that the Pt precursor and dry air
may not be as reactive as hafnium tetrachloride and water toward
the underlying oxide-coated Si substrate. In addition, the difference
of oxygen affinity between the Hf and Pt may play a role here. In a
study by Park et al.,17 Ru nucleation was also found to be less
sensitive to the quality of the monolayer surface than deposition of
Hf or Zr oxide11 and Ti based films.45 This effect was attributed to
the lower oxygen affinity of Ru, making the Ru precursor less likely
to penetrate the monolayer and react with oxide present at the
Si/monolayer interface. All these arguments noted above support the
conclusion that the conditions for blocking the Pt ALD are less
stringent compared to those needed to block HfO2 growth.
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D652
Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
Figure 5. ͑a͒ Deactivating effectiveness of a densely packed ODTS film
against Pt ALD as measured by Pt atomic percent and ͑b͒ film degradation
during the ALD process indicated by the change of its film thickness. Dotted
line indicates the thickness of a densely packed ODTS monolayer. Error bars
show one standard deviation.
To investigate how well the ODTS SAMs block Pt ALD for
higher numbers of ALD cycles, studies were carried out up to 700
cycles. In these studies, densely packed SAMs were formed by im-
mersion into ODTS solution for 48 h.23 Then, the ODTS-coated
substrates were introduced into the reactor for varying numbers of
ALD cycles. An XPS study was performed to examine the resulting
Pt atomic percent and estimate the degradation in the SAM film
thickness. The results are shown in Fig. 5. Figure 5a shows that
below 400 cycles, there is no observable Pt within the detection
limit of the XPS spectrometer. This result indicates that the Pt ALD
reaction is completely deactivated by ODTS when the number of
ALD cycles is Ͻ400 under the given ALD conditions. Figure 5b
shows the ODTS film thickness estimated with XPS after the given
number of ALD cycles. Each measurement was repeated twice. The
thickness of the initial ODTS film, with no exposure to the ALD
reactor, is shown as a dashed line. It is clear that the ODTS film
thickness decreased relative to the control even for the lowest num-
ber of ALD cycles studied. Furthermore, the degradation seems to
get worse with increasing number of ALD cycles. ODTS monolay-
ers are reported to start degrading at 200°C, and the stability de-
creases with increasing temperature.25,46,47 Thus, the observed loss
of ODTS film thickness is likely due to the thermal decomposition
of the SAM at the elevated temperature of 285°C. However, the
degraded ODTS film still retained the ability to block ALD at the
surface, as is evident by the XPS measurements shown in Fig. 5a.
In addition to Si, area-selective ALD experiments were also car-
ried out on YSZ. Figure 6 shows XPS survey scans of the samples
after 50 cycles of Pt ALD. For the native oxide-coated Si ͑Fig. 6a͒
and YSZ ͑Fig. 6c͒ substrates, there are apparent Pt peaks and the Pt
atomic percent are 29.3 and 33.0%, respectively, while there is no
Figure 6. Deactivating effectiveness of the fully dense ODTS SAMs. ͑a͒ and
͑c͒ are XPS spectra of samples following Pt ALD on ODTS-free Si and YSZ
substrates, respectively; ͑b͒ and ͑d͒ are spectra of samples following Pt ALD
on ODTS-coated Si and YSZ substrates, respectively. ALD conditions: sub-
strate temperature: 285°C; precursor temperature: 50°C; dosing time: 2 s;
purge time: 60 s; ALD cycles: 50 cycles.
evidence of Pt on the ODTS-coated Si ͑Fig. 6b͒ and YSZ ͑Fig. 6d͒
substrates. Thus, excellent deactivation can be obtained by using the
ODTS SAM as a resist on both Si and YSZ. For ALD to occur, the
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
D653
value of 109° after 2 min, while the ODTS film takes 4 min to reach
110°. These two limiting contact angle values are consistent with the
literature.11
Although the contact angle trends are similar between ODTS and
FOTS, differences can be seen in the film thickness measurements.
The ODTS film thickness, shown in Fig. 7a, increased with contact
time then leveled off by 4 min at a thickness of 25.6 Å. The FOTS
film thickness, shown in Fig. 7b, peaked at 17.5 Å with a contact
time 4–4.5 min, then decreased for longer contact times. We note
that the saturated ODTS film thickness and the peak thickness value
of the FOTS films are in good agreement with the literature.11 Al-
though we do not fully understand the decrease in the thickness of
FOTS SAMs with increased stamping time, the results suggest that
the FOTS film may become partially bound to the PDMS stamp at
longer times, so that when the stamp is removed, it leaves behind at
the substrate an incomplete film with high defect density. The dif-
ferences may stem from the higher reactivity of FOTS than ODTS
due to presence of the fluorocarbon tail, since the CF3 substitution
on the ␥–C exerts a large inductive effect on the silicon atom of the
incoming head group, making it a better electrophile for nucleo-
philic attack by the oxygen atom of SiOH groups.49 We speculate
that the measured water contact angles of these incomplete FOTS
films remain high due to the highly hydrophobic fluorocarbon tail,
as well as from increased film roughness.11,19,41
The Pt concentrations measured after ALD on the films follow
the film thickness behavior. The XPS studies of these substrates after
the Pt ALD process show the Pt atomic percent on the FOTS coated
substrates drops more quickly than it does on the ODTS-coated
substrates, and reaches a minimum at 2 min. This trend suggests that
if densely packed FOTS films could be achieved, they may have the
potential to deactivate Pt ALD growth in a shorter time compared to
ODTS films. However, among the microcontact printed SAMs,
ODTS serves as a better ALD resist, shown by its higher deactiva-
tion ability. The minimum Pt composition achieved on the FOTS
and ODTS coated substrates are 0.3% and Ͻ0.1%, respectively. The
increase of the Pt atomic percent after the minimum at 2 min on the
FOTS-coated substrates may be due to a high density of defects in
the films formed with too long a stamping time.
Figure 7. Deactivating effect of ␮CP-prepared ͑a͒ ODTS and ͑b͒ FOTS
films on Pt ALD on Si substrates. Water contact angles ͑᭿͒ and thickness
͑᭺͒ measured on the ODTS and FOTS films; Pt atomic percent ͑᭡͒ on
ODTS- and FOTS-coated Si.
It is interesting to compare the two SAM fabrication approaches:
microcontact printing and solution phase deposition. First, films fab-
ricated in either method have similar saturation film thickness and
contact angles. For ODTS SAMs, the maximum film thickness and
contact angle are ϳ26 Å and ϳ110°, respectively, both by micro-
contact printing and by solution deposition. This suggests that the
ODTS molecules may be of similar order and tilt angles in the final
phase of the films fabricated in both methods.39 Second, to achieve
SAMs of the same quality, the characteristic time is significantly
shorter when microcontact printing is used. We can use the blocking
effect on ALD as a sensitive probe of SAM quality. To fully deac-
tivate the Pt ALD growth, ODTS SAMs developed in the solution
phase need 12 h formation time, whereas the microcontact printed
SAMs need under 4 min, as shown in Fig. 4 and 7a, respectively.
The shorter formation time of the microcontact printed ODTS film
may be attributable to the following effects. Both methods started
with an ODTS solution of the same concentration ͑10 mM͒. How-
ever, the “ink” on the PDMS stamp is dried by nitrogen flow to
evaporate the toluene solvent, resulting in an ODTS solution with
significantly higher concentration. This elevated concentration will
facilitate both stages of the formation of the ODTS SAM, including
the surface-head group reaction and the surface crystallization
process.19,39 Thus, fabrication of SAMs through ␮CP is recom-
mended as a more efficient way of depositing the ALD resist.
solid substrate must contain reactive sites and the reactants must be
sufficiently reactive with those sites at the reaction temperature. Ac-
cording to the proposed Pt ALD growth mechanism by Aaltonen et
al.9 Pt precursors need to nucleate onto the oxygen present on the
substrate to facilitate the film growth. On both native oxide-coated
Si and YSZ substrates, Pt deposition occurs due to the presence of
oxygen on the substrate. In contrast, the ODTS-coated substrates are
terminated with methyl group and do not have any active functional
group to react with the Pt precursor. Thus, as expected, there was no
Pt growth on the ODTS-coated substrates.5,9,11,15,48
Once the facility of ODTS as a resist was confirmed, patterned
octadecylsiloxane SAMs were made on the Si and YSZ substrates
using ␮CP. In this process, the contact time was found to be crucial
to the integrity of the pattern transfer. When the stamp was removed
too quickly, loss of pattern resolution occurred because the ODTS
was still wet and spread. However, long contact times between the
PDMS and the substrate sometimes resulted in the breakage of the
PDMS stamp such that parts of the stamp remained on the substrate,
or alternatively the “ink” remained on the stamp. Thus, before in-
vestigating the area selective ALD process, we first explored the
optimal contact time by transferring a nonpatterned film onto the
substrate using a blank PDMS stamp. For this study, we examined
both 10 mM ODTS and 10 mM FOTS solutions to form the ALD
resists. Figure 7 shows the time dependence of ODTS and FOTS
film formation using water contact angle and ellipsometry analysis
and the resulting Pt atomic percent after ALD on the corresponding
SAMs. Similar to the results for the SAMs formed through solution
phase reaction ͑Fig. 4͒, the measured contact angles in ␮CP increase
with contact time until they reach a plateau for both ODTS and
FOTS. The contact angle of the FOTS film reaches a saturation
Patterning.— The PDMS stamps used in the procedure of area
selective ALD consisted of a grid pattern with feature sizes ranging
from 2 ␮m to 50 ␮m. Inked with an ODTS solution, the PDMS
stamps were placed in contact with the Si and YSZ substrates for
5 min. The result is a negative, hydrophobic pattern of the original
master transferred onto the substrates. Figure 8a shows an SEM
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D654
Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
After ␮CP using a 4 ␮m ϫ 2 ␮m patterned stamp ͑generating
4 ␮m squares of ODTS SAMs separated by 2 ␮m lines͒, Pt thin
films were selectively deposited onto the patterned Si and YSZ sub-
strates by ALD using MeCpPtMe3 and air as ALD precursors. The
selective ALD of the Pt thin films was carried out at a substrate
temperature of 285°C, and the number of ALD cycles in all the
processes was kept at 100.
Figures 8b and c show an SEM image and an AFM image, re-
spectively, of the micropatterned YSZ substrates after area selective
ALD of Pt thin films. The contrast inverts in the SEM image of Fig.
8b relative to that of Fig. 8a, and the raised part of the Pt grid
structure can be seen in the AFM image of Fig. 8c, indicating that Pt
deposition occurs in the regions that are not stamped with ODTS.
The rounding of the edges observed in the pattern arises from the
photolithography step during the fabrication of the silicon master at
these small size scales. In addition to SEM and AFM, Auger elec-
tron spectroscopy ͑AES͒ was used to look at the micropatterned Pt
grid structures on YSZ. The AES instrument used in these studies
has a spatial resolution of ϳ10 nm and a sensitivity of 0.1% for Pt.
Figures 9 and 10 illustrate the AES analysis of the patterned mesh
structures at high spatial resolution. Auger survey spectra, which
provide information on the relative concentrations of elements, were
collected in both the ODTS-coated and the uncoated regions of the
YSZ substrate. The survey spectra, shown in Fig. 9, reveal that in
the ODTS-coated YSZ regions ͑Fig. 9a͒, the Pt signal is below the
AES detection limit ͑0.1%͒, whereas in the ODTS-free YSZ regions
͑Fig. 9b͒, the Pt atomic percent is 24%, a value that is comparable to
that for nonchemically treated YSZ reference samples in the same
ALD run. This result indicates that the Pt thin films are selectively
deposited only onto the ODTS-free regions of the substrate and are
blocked at the regions coated by ODTS.
Auger line scans were used to measure the peak intensity of the
Pt and C and Zr Auger peaks as a function of position along the
defined line displayed in Fig. 8b. In Fig. 9c, the Pt, C, and Zr Auger
peaks show the expected intensity alternation. The edges of the lines
are not extremely sharp, likely because the edge of ODTS films
contains a higher density of defects in contrast with interior of the
film.50-52 This high density of defects will enable Pt precursors to
nucleate at the substrate and thus allow atomic layer deposition of
Pt. Note that the edge effect laterally occupies Ͻ0.5 ␮m, suggesting
that area-selective ALD of Pt can be achieved at submicrometer
spatial resolution using this combined ALD-␮CP technique.
AES elemental mapping of Pt, C, and Zr were also conducted
based on the test pattern. Scanning Auger maps ͑Fig. 10͒ were ac-
quired by measuring the peak intensity of the selected elements as a
function of beam position. The scanned area is defined in the SEM
image in Fig. 8b. The maps show the relative elemental distributions
͑peak intensity͒ as a pixel intensity; that is, the greater the Auger
peak intensity, the brighter the pixel value. Figures 10a-c display the
elemental mapping of Pt, C, and Zr, respectively. In Fig. 10a, the
bright regions indicate the presence of Pt while the dark areas indi-
cate the absence of Pt. Pt clearly is deposited in the desired grid
pattern. Figure 10b showed that more C is present in the square
areas due to the coated ODTS monolayer, whereas Fig. 10c shows
that less Zr ͑proportionately͒ is in the ODTS-free areas due to the
presence of Pt at the surface. Both the SEM image ͑Fig. 8b͒ and the
AES image ͑Fig. 10͒ have the same shape, which indicates success-
ful confinement of Pt deposition to the ODTS-free region only. All
the results from SEM, AFM, and AES confirm that the Pt patterns
were well defined and directed with high selectivity by the patterned
SAMs generated with microcontact printing.
Figure 8. ͑Color online͒ Microscopy images of stamped substrates designed
for a 2 ␮m wide grid line with 4 ␮m spacing. ͑a͒ and ͑b͒ are SEM images of
patterned ODTS on YSZ before and after area-selective ALD, respectively;
͑c͒ is the AFM image of the ODTS-patterned YSZ substrate after Pt ALD. In
͑a͒, the brighter squares are where ODTS was stamped. For the AFM image,
the Z range is 5 nm and the Pt film within the grid regions is ϳ3.5 nm thick.
The method presented here could, in principle, provide high se-
lectivity, high spatial resolution, and compatibility with more com-
plex starting substrates than the flat substrates tested in this work.
We plan to integrate these Pt thin films as electrode, catalyst, and
current collect grid into SOFC’s in studies to improve fuel cell per-
formance at lower operating temperature.
image of the patterned SAMs on YSZ produced from a stamp with
a mesh structure containing 4 ␮m squares separated by 2 ␮m
spaces. The bright square regions are those of ODTS-coated YSZ
areas, whereas the dark regions are the ODTS-free YSZ areas. The
patterned monolayers retain the dimensions of the masters without
noticeable chemical spreading.
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
D655
Figure 9. ͑Color online͒ Auger survey spectra of the areas of ͑a͒ an ODTS-
coated region and ͑b͒ an ODTS-free region of YSZ, for a 4 ␮m ϫ 2 ␮m
patterned substrate. ͑c͒ Defined AES line scan showing alternation of Pt, C,
and Zr signals.
Figure 10. Auger elemental maps for ͑a͒ platinum, ͑b͒ carbon, and ͑c͒ zir-
conium on the micropatterned grid structure after area selective ALD of Pt
on YSZ.
Conclusion
In this work, Pt films were deposited using ALD on YSZ sub-
strates for application in solid oxide fuel cells. Films of high purity
and low resistivity were obtained, and studies show that after an-
nealing the films in oxygen, the films retain similar resistivities and
rms roughness from the deposition temperature up to at least 550°C.
High-quality ODTS SAMs were fabricated through both solution
phase and ␮CP, and the latter approach was illustrated to be much
more efficient in forming a SAM of good quality with shorter sily-
lation times. ODTS SAMs were employed as an ALD resist and Pt
ALD growth was shown to be more easily deactivated by ODTS
SAMs compared to the HfO2 ALD process. Furthermore, patterned
Pt thin films with high spatial resolution have been achieved using
atomic layer deposition on microcontact printed Si and YSZ sub-
strates. ODTS SAMs were used here as an ALD resist to direct the
area selective deposition of the Pt thin film. The Pt thin-film selec-
tively grows only on the ODTS-free region because the octadecyl-
siloxane monolayers act to block reactive functional groups at the
substrate. We plan to integrate this Pt grid structure into SOFC fuel
cells for use as the current collect grid, electrode, and catalyst.
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D656
Journal of The Electrochemical Society, 154 ͑12͒ D648-D656 ͑2007͒
22. J. Shim, C. Chao, H. Hong, and F. B. Prinz, Chem. Mater., 19, 3850 ͑2007͒.
23. M. De Ridder, R. G. Van Welzenis, and H. H. Brongersma, Surf. Interface Anal.,
Acknowledgments
The authors thank Sandeep Giri and Dr. Jeffrey S. King for their
help with the ALD of Pt.
33, 309 ͑2002͒.
24. U. Scrinivasan, M. R. Houston, R. T. Howe, and R. Maboudian, J. Microelectro-
mech. Syst., 7, 252 ͑1998͒.
25. G. J. Kluth, M. M. Sung, and R. Maboudian, Langmuir, 13, 3775 ͑1997͒.
26. B. C. Bunker, R. W. Carpick, R. A. Assink, M. L. Thomas, M. G. Hankins, J. A.
Voigt, G. Sipola, M. P. de Boer, and G. L. Gulley, Langmuir, 16, 7742 ͑2000͒.
27. M. J. Geerken, T. S. van Zanten, R. G. H. Lammertink, Z. Borneman, W. Nijdam,
C. J. M. van Rijn, and M. Wessling, Adv. Eng. Mater., 6, 749 ͑2004͒.
28. R. R. Rye, G. C. Nelson, and M. T. Dugger, Langmuir, 13, 2965 ͑1997͒.
29. R. D. Peters, P. F. Nealey, J. N. Crain, and F. J. Himpsel, Langmuir, 18, 1250
͑2002͒.
30. Y. Xia and G. M. Whitesides, Angew. Chem., Int. Ed., 37, 550 ͑1998͒.
31. X. Jiang and S. F. Bent, ECS Trans., 3͑15͒, 249 ͑2007͒.
32. T. Suntola, Mater. Sci. Rep., 4, 261 ͑1989͒.
33. R. L. Puurunen, J. Appl. Phys., 97, 121301 ͑2005͒.
34. R. L. Puurunen, Chem. Vap. Deposition, 9, 327 ͑2003͒.
35. R. L. Puurunen, S. M. K. Airaksinen, and A. O. I. Krause, J. Catal., 213, 281
͑2003͒.
References
1. J. Sinfelt, Catalysts Discoveries, Concepts, and Applications, Wiley, New York
͑1983͒.
2. S. M. Sze, Editor, VLSI Technology, 2nd ed. McGraw-Hill, New York ͑1988͒.
3. R. E. Jones and S. B. Desu, MRS Bull., 21, 55 ͑1996͒.
4. U.S. Department of Energy, Fuel Cell Handbook, 7th ed., Morgantown, West Vir-
ginia ͑2004͒.
5. T. Aaltonen, M. Ritala, T. Sajavaara, J. Keinonen, and M. Leskelä, Chem. Mater.,
15, 1924 ͑2003͒.
6. M. Ritala and M. Leskelä, Handbook of Thin Film Materials, Academic Press, New
York ͑2002͒.
7. T. Suntola, Handbook of Thin Film Process Technology, B1.5:1, Institute of Phys-
ics, Bristol, U.K. ͑1995͒.
8. T. Aaltonen, M. Ritala, Y. L. Tung, Y. Chi, K. Arstila, K. Meinander, and M.
Leskela, J. Mater. Res., 19, 3353 ͑2004͒.
9. T. Aaltonen, A. Rahtu, M. Ritala, and M. Leskela, Electrochem. Solid-State Lett.,
6, C130 ͑2003͒.
10. R. Chen, H. Kim, P. C. McIntyre, and S. F. Bent, Appl. Phys. Lett., 84, 4017
͑2004͒.
11. R. Chen, H. Kim, P. C. McIntyre, and S. F. Bent, Chem. Mater., 17, 536 ͑2005͒.
12. R. Chen, D. W. Porter, H. Kim, P. C. McIntyre, and S. F. Bent, Mater. Res. Soc.
Symp. Proc., 917, E11–05 ͑2006͒.
13. J. P. Lee, Y. J. Jang, and M. M. Sung, Adv. Funct. Mater., 13͑11͒, 873 ͑2003͒.
14. M. M. Sung, in EUROCVD-15. 15th European Conference on Chemical Vapor
Deposition, A. Devi, R. A. Fischer, H. Parala, M. D. Allendorf, and M. L. Hitch-
man, Editors, PV 2005-09, p. 203, The Electrochemical Society Proceedings Se-
ries, Pennington, NJ ͑2005͒.
15. E. K. Seo, J. W. Lee, H. Sung-Suh, and M. M. Sung, Chem. Mater., 16, 1878
͑2004͒.
16. M. H. Park, Y. J. Jang, H. M. Sung-Suh, and M. M. Sung, Langmuir, 20, 2257
͑2004͒.
17. K. J. Park, J. M. Doub, T. Gougousi, and G. N. Parsons, Appl. Phys. Lett., 86,
051903 ͑2005͒.
18. M. Yan, Y. Koide, J. R. Babcock, P. R. Markworth, J. A. Belot, T. J. Marks, and R.
P. H. Chang, Appl. Phys. Lett., 79, 1709 ͑2001͒.
19. A. Ulman, Chem. Rev. (Washington, D.C.), 96, 1533 ͑1996͒.
20. J. D. Swalen, D. L. Allara, J. D. Andrade, E. A. Chandross, S. Caroff, J. Israelach-
vili, T. J. McCarthy, R. Murray, R. F. Pease, J. F. Rabolt, K. J. Wynne, and H. Yu,
Langmuir, 3, 932 ͑1987͒.
36. M. Putkonen, T. Sajavaara, L. S. Johansson, and L. Niinisto, Chem. Vap. Deposi-
tion, 7, 44 ͑2001͒.
37. Handbook of Chemistry and Physics, 87th ed., D. R. Lide, Editor, Taylor and
Francis, Boca Raton, FL ͑2007͒.
38. F. Schreiber, Prog. Surf. Sci., 65, 151 ͑2000͒.
39. D. K. Schwartz, Annu. Rev. Phys. Chem., 52, 107 ͑2001͒.
40. S. R. Wasserman, Y. Tao, and G. M. Whitesides, Langmuir, 5, 1074 ͑1989͒.
41. S. H. Chen and C. W. Frank, Langmuir, 5, 45 ͑1989͒.
42. G. J. Sagiv, J. Colloid Interface Sci., 112, 457 ͑1986͒.
43. D. L. Allara and R. G. Nuzzo, Langmuir, 1, 45 ͑1985͒.
44. W. A. Zisman, Adhesion and Cohesion, Elsevier, New York ͑1962͒.
45. A. S. Killampalli, P. F. Ma, and J. R. Engstrom, J. Am. Chem. Soc., 127, 6300
͑2005͒.
46. M. M. Sung, G. J. Kluth, and R. Maboudian, J. Vac. Sci. Technol. A, 17, 540
͑1999͒.
47. W. R. Ashurst, C. Carraro, R. Maboudian, and W. Frey, Sens. Actuators, A, 104,
213 ͑2003͒.
48. R. Puurunen, Chem. Vap. Deposition, 9, 249 ͑2003͒.
49. C. P. Tripp, R. P. N. Veregin, and M. L. Hair, Langmuir, 9, 3518 ͑1993͒.
50. R. B. A. Sharpe, B. J. F. Titulaer, E. Peeters, D. Burdinski, J. Huskens, H. J. W.
Zandvliet, D. N. Reinhoudt, and B. Poelsema, Nano Lett., 6, 1235 ͑2006͒.
51. A. J. Black, K. E. Paul, J. Aizenberg, and G. M. Whitesides, J. Am. Chem. Soc.,
121, 8356 ͑1999͒.
52. J. Aizenberg, A. J. Black, and G. M. Whitesides, Nature (London), 394, 868
͑1998͒.
21. G. K. Jennings and P. E. Laibinis, Colloids Surf., A, 116, 105 ͑1996͒.
Downloaded on 2015-03-09 to IP 129.100.58.76 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).