Reduction of incubation period by employing OH-terminated Si(001) substrates in the atomic layer deposition of Al2O3

Article abstract of DOI:10.1021/jp048038b

The phenomenon of low initial growth rates in atomic layer deposition (ALD) of various oxides on HF-treated (thus H-terminated) Si(001) substrates, which is termed the incubation effect or incubation period, can be effectively avoided by use of OH-termina

Full text of DOI:10.1021/jp048038b

15128
J. Phys. Chem. B 2004, 108, 15128-15132
Reduction of Incubation Period by Employing OH-Terminated Si(001) Substrates in the
Atomic Layer Deposition of Al2O3
†
†
†
‡
§
,†
Sun S. Lee, Jae Y. Baik, Ki-Seok An, Yung D. Suh, Jin-Ho Oh, and Yunsoo Kim*
Thin Film Materials Laboratory, AdVanced Materials DiVision, and Precision Chemical Technology
Laboratory, AdVanced Chemical Technology DiVision, Korea Research Institute of Chemical Technology,
Yuseong P.O. Box 107, Daejeon 305-600, Korea, and Surface Analysis Laboratory, Nano-Surface Group,
Korea Research Institute of Standards and Science, Yuseong P.O. Box 102, Daejeon 305-600, Korea
ReceiVed: May 7, 2004; In Final Form: July 9, 2004
The phenomenon of low initial growth rates in atomic layer deposition (ALD) of various oxides on HF-
treated (thus H-terminated) Si(001) substrates, which is termed the incubation effect or incubation period,
can be effectively avoided by use of OH-terminated Si(001) substrates. Two ways of preparing OH-terminated
Si(001) were devised in this work: one from atomically clean Si(001) and the other from H-terminated
Si(001). The effect of reducing the incubation period was confirmed in the ALD process of aluminum oxide
2 3
(Al O ) thin films in which trimethylaluminum and water were used as sources of aluminum and oxygen,
respectively. The use of OH-terminated Si(001) substrates has another beneficial effect of producing thin
films with very smooth surface morphology. We propose the use of OH-terminated Si(001) substrates for
growing thin films of many other metal oxides that have shown the incubation period on H-terminated
Si(001) substrates.
I. Introduction
when high-k materials are needed on silicon for the obvious
reason of increasing the effective capacitance. One way out of
In the atomic layer deposition of many oxides on H-
terminated silicon substrates, it is often found that the initial
growth rates of oxide films, for example Al2O3, ZrO2, or HfO2,
this difficulty would be to hydroxylate the silicon surface to
one monolayer thickness only, if possible at all. Preparing
surface hydroxyl groups on silicon substrates is not new as it
are much lower than when the ALD cycles have been executed
can be found in numerous previous works, especially in relation
a few tens of cycles.1
-6
This phenomenon is called the
8-12
to chemisorption of water.
Hydroxyl groups have also been
6
incubation period (effect) or substrate-inhibited growth, and
investigators in the field have tried to eliminate or reduce this
initial delay in growing oxide films by devising various methods
such as formation of a chemical oxide4 by wafer cleaning
10,12
made on a few metal surfaces.
However, it has not been
reported to date that a monolayer of only the hydroxyl groups
of reasonably good quality can be prepared on any silicon
surfaces. In this work, we report on the preparation of surface
hydroxyl groups on Si(001) and the employment of the
hydroxylated Si(001) substrates in the ALD of Al2O3.
-6
7
and excessive use of a metal precursor. The chemical oxide
that is formed on a silicon substrate is known to contain many
surface hydroxyl species3 and they are responsible for the
reduction of incubation period in the case of the ALD of Al2O3
in which trimethylaluminum (Me3Al, TMA) and water were
-6
Previously Bergerson et al.13 devised a way of making a
monolayer of an amine by reacting the amine with a chlorinated
Si(001) substrate in a high-vacuum chamber. The amine took
the place of chlorine on the surface by removing it as hydrogen
chloride, resulting in a self-assembled monolayer. A similar
reaction was thought feasible between water and a chlorinated
Si(001) substrate. A chlorinated Si(001) surface can be prepared
in a high-vacuum chamber by adsorbing chlorine on an
atomically clean Si(001) substrate. In the semiconductor indus-
try, however, it is common to use H-terminated Si substrates,
and therefore, we designed two kinds of hydroxylation schemes
as shown in Figure 1: the first is to start with an atomically
clean Si(001) surface to induce chlorine adsorption followed
by the reaction with water, and the second is to use a
H-terminated Si(001) surface and make a reaction of the surface
with chlorine to have chlorine adsorption similar to the first
case, followed by subsequent reaction with water. In this paper
we describe the realization of the above ideas for the preparation
of OH-terminated Si(001) surfaces and their application in the
ALD of Al2O3 thin films to reduce the incubation period by
the enhanced chemical reactivity due to the presence of the
hydroxyl groups.
1
used as sources for aluminum and oxygen, respectively. It is
also believed true that the ALD of HfO2 with HfCl4 as the Hf
source is initiated by the hydroxyl groups present on the
substrate surface.5,6 Another important aspect of having surface
hydroxyl groups on Si substrates is that the film growth becomes
5
more homogeneous than when the Si surfaces are H-terminated,
for which island growth is initially prevailing. Using an
excessive amount of a metal precursor can be a way of avoiding
the incubation period; however, this may not become a viable
solution for many practical cases. Although the wafer cleaning
method has been improved to such a level that the thickness of
the surface oxide formed during the cleaning process is only a
4
few nanometers, it is still desirable not to have this oxide layer
*
To whom correspondence should be addressed: Tel 82-42-860-7350;
fax 82-42-861-4245; e-mail yunsukim@krict.re.kr.
†
Thin Film Materials Laboratory, Advanced Materials Division, Korea
Research Institute of Chemical Technology.
‡
Precision Chemical Technology Laboratory, Advanced Chemical
Technology Division, Korea Research Institute of Chemical Technology.
§
Korea Research Institute of Standards and Science.
1
0.1021/jp048038b CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/03/2004

Reduced Incubation Period with OH-Terminated Si
J. Phys. Chem. B, Vol. 108, No. 39, 2004 15129
Figure 2. XP peak intensity variations (in atomic percent) of Cl 2p
and O 1s core levels plotted against H O exposure in Langmuir.
2
Water was further purified by several freeze-pump-thaw
cycles before use on the clean Si(001)2×1 surface. Dosing of
chlorine or water was performed from a backfilled gas line into
the appropriate chambers. The exposure was recorded in
-
6
langmuirs (1 L ) 10 Torr s). The purity of Cl2 (99.5%,
Aldrich) was checked by use of the QMA.
Figure 1. Schemes for hydroxylation of Si(001) surfaces: Hydroxy-
lation can be performed with either a clean Si(001) surface or a
H-terminated Si(001) surface.
III. Results and Discussion
Chemisorption of chlorine on a clean Si(001)2×1 surface was
performed at room temperature and the exposure for full
II. Experimental Section
-
8
saturation was found to be about 5 L of chlorine at 1.0 × 10
A small reaction chamber attached to the preparation chamber
of our ESCALAB MK II (VG Scientific, Ltd.) was employed
for chemisorption of chlorine and water and/or ALD experi-
ments. The analysis chamber was equipped with a Mg/Al dual-
anode X-ray source, a differentially pumped helium resonance
lamp, and a spherical sector electron energy analyzer with a
Torr. For a H-terminated Si(001) surface, chlorination was
8
carried out at 80 °C with chlorine exposure of 6.0 × 10 L (for
10 min at 1.0 Torr). Figure 2 shows the intensity variation of
the Cl 2p and O 1s photoelectron peaks starting from a Cl-
saturated Si(001)2×1 surface as the exposure to water increased
with the substrate temperature kept at ∼230 °C. Water was
introduced at ∼1.6 Torr. The peak intensities were calculated
in atomic percent from normalized peak areas including that of
the Si 2p signal. Similar to the report by Klaus et al.,14 in which
SiCl4 and H2O were used for the ALD of SiO2, replacement of
surface chlorine atoms with hydroxyl groups by water exposure
-
10
base pressure lower than 1.5 × 10
Torr. The preparation
-
9
chamber (with a base pressure of ∼2.0 × 10 Torr), equipped
with a quadrupole mass analyzer (QMA), was used for chemi-
sorption of chlorine on clean Si(001). The small reaction
chamber, independently pumped by a turbomolecular pump, had
dosing lines for Cl2 and H2O with a base pressure of ∼1.5 ×
9
required more than 10 L of H2O without using a catalyst such
-
7
1
0
Torr. For the X-ray photoelectron spectroscopy (XPS)
as pyridine. Although not shown, the XP survey spectra of a
clean Si(001)2×1 surface, its Cl-terminated state, and the final
OH-terminated surface revealed that only the relevant peaks of
Si, Cl, and O appeared during the process of hydroxylation.
Carbon contamination was not detectable within the sensitivity
limit of the XPS technique.
measurements, a Mg X-ray (hν ) 1253.6 eV) was used. A He
I line (hν ) 21.22 eV) was used for angle-integrated ultraviolet
photoelectron spectroscopy (UPS) measurements. Synchrotron
radiation experiments were conducted at beamline 7 (hν ) 130
eV) of Hiroshima Synchrotron Radiation Center (HiSOR) at
Hiroshima University. Atomic force microscopy (AFM) (Mul-
timode AFM with Nanoscope IV, Veeco Instruments Inc.) was
used to investigate surface morphology and roughness. Typical
In parallel with the XPS experiments, we took angle-
integrated ultraviolet photoelectron spectra of the chlorinated
Si(001)2×1 surface at increasing water exposures as shown in
Figure 3. For comparison, the spectrum of the clean Si(001)-
2×1 surface is included (Figure 3a). In the spectrum, the surface
state (originated from the dangling bonds) of the clean Si(001)-
2×1 is clearly seen at the binding energy of ∼0.8 eV. Upon
saturation of the surface with chlorine atoms, the UV spectrum,
Figure 3b, changes drastically, showing the appearance of the
structures pertaining to chlorine adsorption at 5.8, 6.8, and 8.4
eV below the Fermi level, mainly attributed to the Cl-bonding
combinations of Cl-3px and Cl-3py orbitals, and the Si-3pz-
2
scan condition was 1 Hz scan rate, 200 × 200 nm scan size in
contact mode.
A boron-doped Si(001) wafer (resistivity 5-12 Ω cm;
3
dimensions 5 × 20 × 0.8 mm ) was chemically etched with a
5% aqueous HF solution and rinsed with deionized water before
it was loaded into the preparation chamber of the ESCALAB.
The wafer was degassed at 800 °C for several hours and
repeatedly annealed at 1250 °C by resistive heating to obtain a
clean 2×1 surface. Cleanliness of the 2×1 surface was checked
by XPS and UPS, which showed no traces of contamination
and a high intensity of the surface state (the dimer dangling
bond state), respectively. The temperature of the substrate was
monitored by infrared and optical pyrometers. Temperatures
below 250 °C were estimated by extrapolation of the relationship
between the current applied to the substrate and the measured
temperature.
1
5,16
Cl-3pz orbital, respectively,
and disappearance of the
surface state of the clean Si(001)2×1. From spectra c-g, gradual
changes of the peak shapes are seen as water reacted with
adsorbed chlorine, removing chlorine as HCl and leaving OH
groups on the surface. The structures at the binding energies of
6.9 and ∼8 eV with a broad feature at 11.8 eV develop with
decreasing the structures related to the Cl 3p orbitals. The

15130 J. Phys. Chem. B, Vol. 108, No. 39, 2004
Lee et al.
Figure 3. Angle-integrated UP spectra for the process of hydroxylation
of a Si(001)2 × 1 surfaces(a) clean Si(001)2×1 surface, (b) after its
chlorinationsand the surfaces obtained by exposing the chlorinated
Si(001)2×1 surface to H O at ∼230 °C for (c) 2 min, (d) 5 min, (e) 8
2
min, (f) 15 min, and (g) 25 min at ∼1.6 Torr.
general features of the UV spectra, Figure 3f and g, closely
resemble those of water adsorbed on Si(001) reported by other
investigators,1
7-19
although the spectral interpretations varied.
These spectra are not clearly distinguished from those of water
adsorbed on the Si(001)2×1 surface. In angle-integrated UPS
measurements with the photon source of 21.22 eV, the Si-H
bond is not clearly observed because of its relatively small cross-
section. However, the difference between the water adsorption
on the Cl-terminated Si(001)2×1 surface and that on the clean
Si(001)2×1 surface will be clearly shown in the Si 2p core level
spectra by SRPES in the following.
Replacement of chlorine adsorbed on the Si(001)2×1 surface
by water and subsequent formation of hydroxyl groups on the
surface were examined more closely by synchrotron radiation
photoemission spectroscopy (SRPES). Figure 4a shows a Si 2p
core level spectrum for a clean Si(001)2×1 surface as a
reference. In this and subsequent figures (Figure 4b-e), the
experimental data are denoted by black dots and the synthesized
curves resulting from curve-fitting are denoted by orange-red
solid curves running through the black dots. The photon energy
and the emission angle were taken at hν ) 130 eV and 60°,
respectively, to obtain surface-sensitive Si 2p spectra. In the
figure, the Si 2p components due to the up (U; binding energy
shift relative to the bulk component ∆BE ) -0.54 eV, in
orange) and down (D; ∆BE ) +0.06 eV, in olive green) dimer
Si atoms having dangling bonds, the Si atoms in the second
layer (2nd, ∆BE ) +0.25 eV, black solid curve), symmetric
dimers and/or defects (C, ∆BE ) -0.22 eV, brown solid curve),
and the bulk (B, blue solid curve) Si atoms are resolved. The
Figure 4. Synchrotron radiation photoemission spectra for the process
of hydroxylation of a Si(001)2×1 surfaces(a) clean Si(001)2×1
surface, (b) after its chlorinationsand the surfaces obtained by exposing
2
the chlorinated Si(001)2×1 surface to H O at ∼230 °C for (c) 2.0 ×
9
9
10 L and (d) 7.2 × 10 L (complete replacement of Cl). A spectrum
for a H
2
O-chemisorbed Si(001)2 ×1 surface is given (e) for comparison
with the OH-terminated Si(001) surface.
9
line shape of the spectrum and the fitting results are quite
× 10 L at ∼230 °C) changed the spectral shape of the Si 2p
consistent with previous reports.2
0,21
The spectrum for a fully
peaks as shown in Figure 4c. This change was caused by the
partial removal of chlorine atoms from the Si surface and
chemisorption of hydroxyl groups in their places as clearly
resolved in the figure (lime for Si-Cl and red for Si-OH, with
∆BE ) +0.99 eV). When all the chlorine atoms were replaced
chlorine-saturated Si(001)2×1 surface (by 5 L of Cl2 exposure)
is shown in Figure 4b, which is typical of Cl-adsorbed Si(001)
as previously reported.1
6,22
The complete disappearance of the
up and down dimer components and the appearance of a new
component (in lime, ∆BE ) +0.94 eV) due to chemisorption
of chlorine indicate full coverage of the Si(001)2×1 surface by
chlorine atoms. Up to this stage, the LEED (low-energy electron
diffraction) pattern (not shown) of the Si(001) surface did not
show a noticeable change in its 2×1 structure. Exposure of this
Cl-covered Si(001)2×1 surface to water vapor (to 2.0
9
by hydroxyl groups (at 7.2 × 10 L of water), the Si 2p spectrum
changed further to that shown in Figure 4d, in which a peak
2
+
probably due to Si species (in gray, ∆BE ) +1.71 eV) is
resolved together with that due to the hydroxylated Si atoms
+
(Si ). There also appeared relatively small components corre-
3
+
4+
sponding to the Si (in black, ∆BE ) +2.62 eV) and Si (in

Reduced Incubation Period with OH-Terminated Si
J. Phys. Chem. B, Vol. 108, No. 39, 2004 15131
purple, ∆BE ) +3.60 eV) species23,24 which indicate partial
surface oxidation. The identity of the Si2 species cannot be
deciphered at the moment, although it is speculated that it may
be due to the oxidized Si species such as partial 2(OH)-Si
species at the top Si layer with partial breaking of the Si dimer
bonds and/or the Si back-bonds between the top and second Si
layers, etc. The larger intensity of the final O 1s peak (∼29 at.
+
%
) than that of the initial Cl 2p peak (∼18 at. %) shown in
Figure 2 may be attributed to partial oxidation of the top and/
or the second layer. Greater attenuation for the Si atoms covered
by OH groups than those covered by Cl atoms may also have
contributed to the difference in intensities.
To distinguish the OH-terminated Si(001) surface from a
H2O-chemisorbed Si(001) surface, SRPES of the latter was taken
(Figure 4e, for 10 L exposure of water at RT) and compared
Figure 5. Al 2p XP peak intensities obtained for the OH-terminated
Si(001) substrate [prepared by using a clean Si(001)2×1 surface] (b)
and the H-terminated Si(001) substrate (O).
with that of the former given in Figure 4d. In the case of the
H2O-chemisorbed Si(001) surface, the Si 2p component at +0.31
eV due to the Si-H bonds (in yellow) contributes much to the
intensity in the valley between the bulk Si 2p1/2 and Si 2p3/2
2
5
peaks whereas for the OH-terminated Si(001), due to the
absence of the Si-H bonds, the corresponding region shows a
decisively deeper valley.
The SRPES experiments demonstrated that we have success-
fully prepared an OH-terminated Si(001) surface starting from
a clean Si(001)2×1 surface. On the other hand, for H-terminated
Si(001) surfaces prepared by a wet process, XPS measurements
showed similar spectral results except that carbon contamination
was always present and the O 1s signals were much larger than
those for OH-terminated Si(001) surfaces obtained from clean
Si(001)2×1 substrates. This is understandable, for the wet
process involves etching, which roughens the surface, and by
26,27
this process many dihydride species are produced,
resulting
in more chlorine atoms on the Si surface and thus subsequently
more hydroxyl groups. It was also noted by XPS that, starting
from a H-terminated Si(001) surface obtained by a wet process,
the final OH-terminated Si(001) surface was somewhat oxidized
judging from the appearance of a Si 2p peak corresponding to
the formation of the Si4 species at ∼3.6 eV higher binding
energy than the bulk Si 2p peak.
+
To demonstrate the effect of the OH-terminated Si(001)
surfaces, which are expected to have enhanced surface reactivity
compared to the H-terminated Si(001) surface, ALD of alumi-
num oxide thin films was performed on two representative
Si(001) surfaces: (i) a OH-terminated Si(001) surface obtained
from a clean Si(001)2×1 surface and (ii) a H-terminated
Si(001) surface. The ALD of Al2O3 was carried out with
trimethylaluminum and water as sources for aluminum and
oxygen, respectively. This ALD reaction has been known to
show an incubation period on H-terminated Si(001) or Si(111).
To compare the initial growth rates of Al2O3 films on these
surfaces, only 12 ALD cycles of TMA-purge-H2O-purge
steps were performed. The substrates were kept at 200 °C for
the ALD processes. This deposition temperature was chosen
Figure 6. AFM images for Al O films deposited on OH-terminated
Si(001) (a) and on H-terminated Si(001) (b) with 12 ALD cycles.
2
3
Figure 5 shows the intensities of the Al 2p photoelectron
peaks taken for the two substrates, plotted against the number
of ALD cycles. After the first ALD cycles, it is obvious that
the Al 2p intensity is much greater for the OH-terminated
Si(001) substrate than for the H-terminated Si(001) substrate.
As the number of ALD cycles is increased, the peak intensity
of the OH-terminated Si(001) substrate consistently shows
greater values than the H-terminated Si(001) substrate. The
differences would have been more striking had we used a much
28,29
by considering the uses of TMA by other investigators.
TMA
and water were each supplied at 500 mTorr for 40 s. After each
ALD cycle was completed, the substrate was transferred to the
analysis chamber of the ESCALAB for the measurement of the
Al 2p photoelectron peak. The takeoff angle of the photoelec-
trons was set at 20° to make the measurements more surface-
sensitive. Although the intensity of the Al 2p photoelectron peak
is not directly proportional to the thickness of a growing film,
it can clearly serve as a way of comparing the initial growth
rates of the Al2O3 films when the films are very thin.
4
shorter water supply time as in an actual ALD situation, because
a longer water supply tends to produce more surface active sites.
The OH-terminated Si(001) substrate made from a H-terminated
Si(001) surface and the SiO2-formed Si(001) substrate showed

15132 J. Phys. Chem. B, Vol. 108, No. 39, 2004
Lee et al.
Al 2p signal intensities comparable to that of the OH-terminated
Si(001) substrate made from a clean Si(001)2×1 surface. Their
Al 2p intensities are not shown in Figure 5 for clarity. It was
noted that the OH-terminated Si(001) substrate made from a
H-terminated Si(001) surface showed highest reactivity toward
the ALD of Al2O3. Thus, the in situ ALD experiment clearly
demonstrated that hydroxylation of Si(001) surfaces has the
effect of reducing the incubation period in the initial growth of
Al2O3 films.
The Al2O3 thin films were then examined by atomic force
microscopy as shown in Figure 6. It is clearly seen that the
film (Figure 6a) deposited on the OH-terminated Si(001)
substrate has much smoother surface morphology than the film
References and Notes
(
1) Copel, M.; Gribelyuk, M.; Gusev, E. Appl. Phys. Lett. 2000, 76,
36-438.
(2) Bender, H.; Conard, T.; Nohira, H.; Petry, J.; Richard, O.; Zhao,
4
C.; Brijs, B.; Besling, W.; Detavernier, C.; Vandervorst, W.; Caymax, M.;
De Gendt, S.; Chen, J.; Kluth, J.; Tsai, W.; Maes, J. W. Extended Abstracts
of International Workshop on Gate Insulators, IWGI 2001, IEEE catalog
no. 01EX537, pp 86-92.
(3) Cosnier, V.; Bender, H.; Caymax, M.; Chen, J.; Conard, T.; Nohira,
H.; Richard, O.; Tsai, W.; Vandervorst, W.; Young, E.; Zhao, C.; De Gendt,
S.; Heyns, M.; Maes, J. W. H.; Tuominen, M.; Rochat, N.; Olivier, M.;
Chabli, A. Extended Abstracts of International Workshop on Gate Insulators,
IWGI 2001, IEEE catalog no. 01EX537, pp 226-229.
(4) Tsai, W.; Carter, R. J.; Nohira, H.; Caymax, M.; Conard, T.;
Cosnier, V.; De Gendt, S.; Heyns, M.; Petry, J.; Richard, O.; Vandervorst,
W.; Young, E.; Zhao, C.; Maes, J.; Tuominen, M.; Schulte, W. H.;
Garfunkel, E.; Gustafsson, T. Microelectron. Eng. 2003, 65, 259-272.
(5) Green, M. L.; Ho, M.-Y.; Busch, B.; Wilk, G. D.; Sorsch, T.;
Conard, T.; Brijs, B.; Vandervorst, W.; R a¨ is a¨ nen, P. I.; Muller, D.; Bude,
M.; Grazul, J. J. Appl. Phys. 2002, 92, 7168-7174.
(Figure 6b) deposited on the H-terminated Si(001) substrate.
The root-mean-square surface roughness values were measured
to be 0.070 and 0.135 nm, respectively, for the Al2O3/OH/
Si(001) and Al2O3/H/Si(001). This is a good indication that the
OH-terminated Si(001) surface helps the Al2O3 film grow
homogeneously on the substrate surface as well as eliminating
the incubation period.
(
6) Puurunen, R. L. J. Appl. Phys. 2004, 95, 4777-4786.
(7) Frank, M. M.; Chabal, Y. J.; Green, M. L.; Delabie, A.; Brijs, B.;
Wilk, G. D.; Ho, M.-Y.; da Rosa, E. B. O.; Baumvol, I. J. R.; Stedile, F.
C. Appl. Phys. Lett. 2003, 83, 740-742.
(8) Ibach, H.; Wagner, H.; Bruchmann, D. Solid State Commun. 1982,
4
2, 457-459.
IV. Summary
(
9) Chabal, Y. J.; Christman, S. B. Phys. ReV. B 1984, 29, 6974-
6
976.
The OH-terminated Si(001) surfaces have been prepared from
both a clean Si(001)2×1 surface and a H-terminated Si(001)
surface obtained by a wet process. An OH-terminated Si(001)
surface was compared with a H2O-chemisorbed Si(001) surface
by SRPES in which the state of the surface OH species was
clearly distinguished from the mixed state of H and OH species
of H2O-chemisorbed Si(001). The OH-terminated Si(001)
surfaces have been found to show the effect of reducing the
incubation period in the initial stage of the ALD of Al2O3 thin
films. It has also been found that the Al2O3 films grown on
OH-terminated Si(001) are much smoother than those grown
on H-terminated Si(001). The method of preparing OH-
terminated Si(001) surfaces described in this paper may have
general applicability for hydroxylation of other semiconductors
(
10) Thiel, P. A.; Madey, T. E. Surf. Sci. Rep. 1987, 7, 211-385 and
references therein.
(
11) Waltenburg, H. N.; Yates, J. T., Jr. Chem. ReV. 1995, 95, 1589-
1673 and references therein.
(12) Henderson, M. A. Surf. Sci. Rep. 2002, 46, 1-308 and references
therein.
(13) Bergerson, W. F.; Mulder, J. A.; Hsung, R. P.; Zhu, X.-Y. J. Am.
Chem. Soc. 1999, 121, 454-455.
(14) Klaus, J. W.; Sneh, O.; George, S. M. Science 1997, 278, 1934-
1936.
(
15) Lindsay, R.; Wincott, P. L.; Thornton, G.; Harrison, N. M. Surf.
Sci. 1998, 398, 301-307.
16) Mart ´ı n-Gago, J. A.; Rom a´ n, E.; Refolio, M. C.; L o´ pez-Sancho, J.
(
M.; Rubio, J.; Hellner, L.; Comtet, G. Surf. Sci. 1999, 424, 82-93.
(17) Schmeisser, D.; Himpsel, F. J.; Hollinger G. Phys. ReV. B 1983,
2
7, 7813-7816.
30
31
32
33
(18) Ranke, W.; Xing, Y. R. Surf. Sci. 1985, 157, 339-352.
such as Ge and GaAs and such metals as Ru, Rh, and
(19) Larsson, C. U. S.; Flodstr o¨ m, A. S.; Nyholm, R.; Incoccia, L.; Senf,
3
4,35
10,12
Pt
and many other materials
that have been reported to
F. J. Vac. Sci. Technol. A 1987, 5, 3321-3324.
(20) Landemark, E.; Karlsson, C. J.; Chao, Y.-C.; Uhrberg, R. I. G.
Phys. ReV. Lett. 1992, 69, 1588-1591.
(21) Pi, T.-W.; Cheng, C.-P.; Hong, I.-H. Surf. Sci. 1998, 418, 113-
21.
show dissociative chemisorption of water, for the purpose of
surface functionalization, including self-assembled mono-
3
6,37
layers.
Presently, we are investigating the possibility of
1
reducing the incubation period for the HfCl4-H2O ALD system
for producing HfO2 thin films.
(
22) Pi, T.-W.; Tsai, S.-F.; Ouyang, C.-P.; Wen, J.-F.; Wu, R.-T. Surf.
Sci. 2001, 488, 387-392.
23) Himpsel, F. J.; McFeely, F. R.; Taleb-Ibrahimi, A.; Yarmoff, J.
A.; Hollinger, G. Phys. ReV. B 1988, 38, 6084-6096.
24) Holl, M. M. B.; McFeely, F. R. Phys. ReV. B 1993, 71, 2441-
444.
25) McGrath, R.; Cimino, R.; Braun, W.; Thornton, G.; McGovern, I.
(
Acknowledgment. This research has been supported by the
Ministry of Science and Technology of Korea through the
National Research Laboratory (NRL) Program and the National
Program for Tera-level Nanodevices (TND) as one of the 21st
Century Frontier Programs. We thank the staff of HiSOR,
particularly Dr. H. Sato and Dr. M. Nakatake, for their technical
assistance in carrying out the SRPES experiments. Y.D.S. thanks
Ms. B. S. Kang for technical assistance and Dr. H. Y. Kim for
valuable discussions.
(
2
(
T. Vacuum 1988, 38, 251-255.
(26) Grundner, M.; Schulz, R. AIP Conf. Proc. 1988, 167, 329-337.
(27) Chabal, Y. J.; Higashi, G. S.; Raghavachari, K.; Burrows, V. A. J.
Vac. Sci. Technol. A 1989, 7, 2104-2109.
28) George, S. M.; Ott, A. W.; Klaus, J. W. J. Phys. Chem. 1996, 100,
(
1
3121-13131.
(29) Matero, R.; Rahtu, A.; Ritala, M.; Leskel a¨ , M.; Sajavaara, T. Thin
Solid Films 2000, 368, 1-7.
(
30) Kuhr, H. J.; Ranke, W. Surf. Sci. 1987, 187, 98-111.
Supporting Information Available: (1) Hydroxylation of
a clean Si(001)2×1 surface, (2) LEED pattern of a Cl-terminated
Si(001)2×1 surface, (3) hydroxylation of a H-terminated Si-
(31) Webb, C.; Lichtensteiger, M. J. Vac. Sci. Technol. 1982, 21, 659-
6
62.
(
32) Feibelman, P. J. Science 2002, 295, 99-102.
(001) surface, (4) comparison of Si 2p photoelectron peaks of
(33) Feibelman, P. J. Phys. ReV. Lett. 2003, 90, 186103.
OH-terminated Si surfaces made in 1 and 3, (5) Comparison of
O 1s photoelectron peaks of OH-terminated Si surfaces made
in 1 and 3, and (6) XP survey spectra for different OH/Si(001)
surfaces (PDF). This information is available free of charge via
the Internet at http://pubs.acs.org.
(34) Creighton, J. R.; White, J. M. Surf. Sci. 1982, 122, L648-L652.
(35) Seitsonen, A. P.; Zhu, Y.; Bed u¨ rftig, K.; Over, H. J. Am. Chem.
Soc. 2001, 123, 7347-7351.
(
(
36) Petrucci, M. G. L.; Kakkar, A. K. AdV. Mater. 1996, 8, 251-253.
37) Petrucci, M. G. L.; Kakkar, A. K. Chem. Mater. 1999, 11, 269-
276.