ArticlePDF Available

Reaction of a hydrogen-terminated Si(100) surface in UHV with ion-pump generated radicals

Authors:

Abstract and Figures

The authors present scanning tunneling microscopy images and mass spectra that show that dosing gases at pressures in the range of 10-6 Torr in an ion-pumped ultrahigh vacuum (UHV) chamber results in a measurable concentration of reactive molecular radicals and atomic hydrogen ions being created. One source of radicals is the fragmentation of the dosed molecule, while another is atomic hydrogen that is re-emitted from the ion pump itself. The dosing of noble gases such as helium also results in harmful radicals escaping the ion pump. These radicals are able to create new reactive sites on a hydrogen-terminated Si(100) surface; they show that these new dangling bonds result in extra molecular line growth in a 2,3-dimethyl-1,3-butadiene dosing experiment. These results serve as a cautionary note to experimenters working with ion-pumped UHV systems and surfaces that are sensitive to radicals, such as hydrogen-terminated Si.
Content may be subject to copyright.
Reaction of a hydrogen-terminated Si(100) surface in UHV with ion-pump
generated radicals
Janik Zikovsky, Stanislav A. Dogel, Adam J. Dickie, Jason L. Pitters, and Robert A. Wolkow
Citation: J. Vac. Sci. Technol. A 27, 248 (2009); doi: 10.1116/1.3071944
View online: http://dx.doi.org/10.1116/1.3071944
View Table of Contents: http://avspublications.org/resource/1/JVTAD6/v27/i2
Published by the AVS: Science & Technology of Materials, Interfaces, and Processing
Related Articles
Surface chemistry controlled diameter-modulated semiconductor nanowire superstructures
J. Vac. Sci. Technol. B 31, 020603 (2013)
Flash sample heating for scanning tunneling microscopy: Desorption of 1-octanethiolate self-assembled
monolayers in air
J. Vac. Sci. Technol. B 31, 013201 (2013)
New method of calculating adsorption and scattering for Xe-Pt(111) using Direct Simulation Monte Carlo
techniques
J. Vac. Sci. Technol. A 30, 061401 (2012)
Gain and loss mechanisms for neutral species in low pressure fluorocarbon plasmas by infrared spectroscopy
J. Vac. Sci. Technol. A 30, 051308 (2012)
Differences in erosion mechanism and selectivity between Ti and TiN in fluorocarbon plasmas for dielectric etch
J. Vac. Sci. Technol. B 30, 041811 (2012)
Additional information on J. Vac. Sci. Technol. A
Journal Homepage: http://avspublications.org/jvsta
Journal Information: http://avspublications.org/jvsta/about/about_the_journal
Top downloads: http://avspublications.org/jvsta/top_20_most_downloaded
Information for Authors: http://avspublications.org/jvsta/authors/information_for_contributors
Downloaded 19 Jul 2013 to 129.128.50.162. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvsta/about/rights_and_permissions
Reaction of a hydrogen-terminated Si100 surface in UHV with ion-pump
generated radicals
Janik Zikovsky
a
and Stanislav A. Dogel
Department of Physics, University of Alberta, Edmonton, Alberta T6G 2G7, Canada
Adam J. Dickie and Jason L. Pitters
National Institute for Nanotechnology, 11421 Saskatchewan Drive, Edmonton, Alberta T6G 2M9, Canada
Robert A. Wolkow
Department of Physics, University of Alberta, Edmonton, Alberta T6G 2G7, Canada and National Institute
for Nanotechnology, 11421 Saskatchewan Drive, Edmonton, Alberta T6G 2M9, Canada
Received 1 October 2008; accepted 15 December 2008; published 12 February 2009
The authors present scanning tunneling microscopy images and mass spectra that show that dosing
gases at pressures in the range of 10
−6
Torr in an ion-pumped ultrahigh vacuum UHV chamber
results in a measurable concentration of reactive molecular radicals and atomic hydrogen ions being
created. One source of radicals is the fragmentation of the dosed molecule, while another is atomic
hydrogen that is re-emitted from the ion pump itself. The dosing of noble gases such as helium also
results in harmful radicals escaping the ion pump. These radicals are able to create new reactive sites
on a hydrogen-terminated Si100 surface; they show that these new dangling bonds result in extra
molecular line growth in a 2,3-dimethyl-1,3-butadiene dosing experiment. These results serve as a
cautionary note to experimenters working with ion-pumped UHV systems and surfaces that are
sensitive to radicals, such as hydrogen-terminated Si. © 2009 American Vacuum Society.
DOI: 10.1116/1.3071944
I. INTRODUCTION
The controlled environment provided by ultrahigh
vacuum UHV兲共10
−9
Torr base pressure systems has
been fundamental to the advancement of surface science over
the past half-century. The hydrogen-terminated surface of Si,
in particular, has attracted interest since hydrogen can be
desorbed with atomic precision
1
or over broader areas and
used as a resist for further surface modification.
2
Numerous
studies of molecular line growth
3
also rely strongly on high-
quality H-terminated Si surfaces.
4
Radical-initiated molecular line growth on a H-terminated
Si surface was first observed for styrene dosed in the gas
phase in a UHV chamber.
3
Subsequently, many different or-
ganic molecules have been found to form nanostructures via
a self-directed process on the H–Si surface in the same
way.
58
The process begins with a molecule bonding to an
isolated surface dangling bond DB, which leaves an un-
paired electron on one atom of the molecule. The adsorbed
radical then abstracts a nearby hydrogen atom from the sur-
face, passivating the molecule and creating a new DB where
the chain reaction can continue. The extent and direction of
growth depend on both the stability of the adsorbed radical
and the molecular geometric constraints, as well as the flux
of molecules to the reactive site. For example, while styrene
grows straight lines parallel to the 2 1 dimers, benzalde-
hyde grows in double lines
6
and allyl mercaptan grows lines
perpendicular to the dimer rows.
8
It is a standard practice in UHV scanning tunneling mi-
croscopy STM laboratories to use sputter-ion pumps exclu-
sively to maintain vacuum, since such pumps are completely
vibration-free and thus allow high-quality STM imaging. In
this article, we detail how the sputter-ion pump, an essential
component of UHV systems, influences the quality of a
hydrogen-terminated surface and the growth of molecular
structures on silicon when experiments involving modest gas
pressures in the 10
−6
Torr range are performed.
II. METHOD
STM experiments were performed at room temperature in
an UHV chamber with a base pressure below 5
10
−11
Torr equipped with an Omicron STM1. A schematic
of our experimental setup is shown in Fig. 1. The ion pump
used is a 200 l/s diode model from Gamma Vacuum, also
equipped with an integrated titanium sublimation pump. The
ion-pump controller used maintains a constant voltage of
+6700 V throughout the experiments, with the ion current
proportional to the pressure. The system is also equipped
with a 60 l/s turbomolecular pump from Pfeiffer Vacuum
TMU-071P, equipped with a drag stage, ultimate pressure
below 410
−10
Torr. Our sample consists of a 0.1 cm
n-type Si crystal, hydrogen terminated in UHV to obtain a
H–Si100-2 1 surface by exposing it for 120 s to atomic
hydrogen cracked by a hot filament while maintaining the
sample at 320 °C.
9
Several molecules, including 2,3-
dimethyl-1,3-butadiene Aldrich, 99.5%, were subjected
to several freeze pump-thaw cycles prior to dosing through a
variable leak valve. Hydrogen and helium gas 99.999%
pure were purified using a liquid nitrogen cold trap and
dosed through a variable leak valve.
a
Electronic mail: jzikov@phys.ualberta.ca
248 248J. Vac. Sci. Technol. A 272, Mar/Apr 2009 0734-2101/2009/272/248/5/$25.00 ©2009 American Vacuum Society
Downloaded 19 Jul 2013 to 129.128.50.162. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvsta/about/rights_and_permissions
We performed two types of dosing experiments. First, we
used an ion pump throughout the procedure: during sample
preparation, dosing, and imaging—this is a normal procedure
for a molecular dosing experiment. In later experiments, we
used a turbomolecular pump without ion-pump operation.
The ion pump was used during sample preparation to ensure
high surface quality, but was immediately turned off not
gated off after completing hydrogen termination. From that
moment the chamber was pumped with the turbopump ex-
clusively. Although we do not expect significant radical cre-
ation at the hot filament of the ion gauge, it was turned to the
lowest current setting as a precaution. We continued to use
the turbopump during dosing and subsequent imaging, which
introduces an additional source of vibration. Fortunately, in
our setup that influence was barely noticeable in STM im-
ages.
Additionally, mass spectra were obtained using an Accu-
Quad residual gas analyzer RGA mounted on a flange on
the vacuum chamber directly facing the ion pump’s inlet port
see Fig. 1. The STM chamber is located on a side flange on
this vacuum chamber, and so does not have direct line of
sight to the ion-pump inlet port. The filament of the RGA
was turned off during acquisition so as to be exclusively
sensitive to ions created by the ion pump, rather than ions
created thermally by the RGA filament. The partial pressure
of these ions is quite small, requiring us to use the channel
electron multiplier, long acquisition times, and averaging to
obtain the required signal to noise.
III. RESULTS AND DISCUSSION
While studying the line growth of 2,3-dimethyl-1,3-
butadiene molecules, we observed inconsistencies in line
growth. In Fig. 2, occupied-state STM images are shown of
the H Si100-2 1 surface before and after 2,3-dimethyl-
1,3-butadiene exposure. In Fig. 2A, the pristine
H–Si100-2 1 surface is shown with dimer rows clearly
visible and rare DB sites indicated by the arrows. In Fig.
2B, the identical surface is shown after exposure to 250 L
2 10
−6
Torr for 125 s of 2,3-dimethyl-1,3-butadiene va-
por, with an ion pump in operation, as is standard UHV
practice. Extended molecular growth patterns are observed
over the entire surface. The growth of 2,3-dimethyl-1,3-
butadiene on H Si100-2 1 surface requires a DB as a
starting point,
10
however, the growth features in Fig. 2B
occur at many sites where no pre-existing DB was present.
Excluding molecular growth that could have started from an
original DB, 175 new features were counted in the frame
shown. This number of new features yields a density of at
least 1.25 10
13
cm
−2
new reaction sites created during the
dosing procedure, equivalent to 1.8% of the surface Si at-
oms.
It was suspected that the ion pump was in some way
responsible for the new features observed in Fig. 2B, there-
fore the dosing procedure was repeated with the ion pump
turned off throughout, using only a turbomolecular pump to
maintain vacuum. The resulting STM images are shown in
Figs. 2C and 2D. Even with an extended, 2400 L dose of
2,3-dimethyl-1,3-butadiene vapor 2 10
−6
Torr for 1200 s,
nearly ten times the dose performed with the ion pump,
molecular growth was only observed from existing DBs ar-
rows, Fig. 2C. No new initiation sites were created on the
surface, and as a result, the molecular surface coverage is far
lower than in Fig. 2B.
FIG. 1. Schematic of the experimental setup used. 1: ion pump, 2: TSP
filaments, 3: turbomolecular pump, 4: residual gas analyzer, 5: leak valve
used to dose He, 6: leak valve used to dose molecules, 7: ion gauge, 8:
STM tip, 9: sample viewed from the side. Drawing not to scale, distances
shown are in millimeters.
FIG. 2. STM images 共⬃40 35 nm
2
of dosing 2,3-dimethyl-1,3-butadiene
on two n-type medium doped H Si100-2 1 surfaces. Images A 0.08
nA, 1.8 V and C 0.065 nA, 3.0 V were taken before dosing. Image B
0.05 nA, 2.2 V shows the result of a 250 L dose of 2,3-dimethyl-1,3-
butadiene with the ion pump turned on. Image D 0.065 nA, 3.0 V shows
the result of dosing 2400 L of the same molecule, using only a turbomo-
lecular pump see text for details. A white circle highlights the same area in
A and B. In spite of the fact that it contains only a few DBs indicated with
arrows in A, molecular growth occurred at a much larger number of sites in
this area in image B. Overall 175 new features/growth that could not be
assigned to pre-existing DBs were counted in this frame, yielding a density
of 1.25 10
13
new features/ cm
2
. Conversely, the area highlighted in C and
D shows that molecular growth occurred at existing DBs exclusively and no
new sites were initiated.
249 Zikovsky et al.: Reaction of a hydrogen-terminated Si100 surface 249
JVST A - Vacuum, Surfaces, and Films
Downloaded 19 Jul 2013 to 129.128.50.162. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvsta/about/rights_and_permissions
To investigate more directly the quality of the vacuum
during a dosing experiment, we acquired mass spectra with a
RGA Fig. 3A while taking care to leave the RGA filament
off so as to only detect existing ions in the chamber. Multiple
mass spectra were obtained while no molecules were being
dosed and the ion pump alone was operating Fig. 3A, line
1, which show that no peaks are present in the chamber
beyond the noise level of 10
−14
Torr. Similarly, line 2 in
Fig. 3A shows the average of many mass spectra taken
while dosing 10
−5
Torr of 2,3-dimethyl-1,3-butadiene with
only the turbomolecular pump maintaining vacuum. No
peaks are visible beyond our noise level of 10
−13
Torr the
noise level in line 1 is lower because we were able to average
25 times more spectra. Finally, line 3 in Fig. 3A shows the
result of dosing 2,3-dimethyl-1,3-butadiene at 10
−5
Torr
with only the ion pump maintaining vacuum. In addition to a
large peak near 1 amu, three small peaks are visible above
the noise, at masses of 28, 67.3, and 82.6 the noninteger
peak positions are due to the imperfect calibration of masses
in the RGA. We note that the two largest peaks in the ex-
pected mass spectrum Fig. 3B, data from NIST Ref. 11兲兴
at 82 and 67 are observed in our experimental spectra. The
mass peak at 82 corresponds to the mass of the whole
2,3-dimethyl-1,3-butadiene molecule; the peak at 67 can
be attributed to the molecule losing a CH
3
group. These mass
spectra show clearly that at high dosing pressures, the ion
pump ionizes the 2,3-dimethyl-1,3-butadiene molecule,
which then breaks into smaller, reactive fragments. We note
that only ions generated near the inlet of the ion pump will
be able to escape the electric field within; however, neutral
radicals produced by the breakup of ionized molecules will
be unaffected and may escape the ion pump if they are not
ionized and pumped themselves. Therefore, we expect that
the concentration of neutral molecular fragments will be
much higher than those of the ions measured in this way. It is
also noted that because of molecular fragmentation, multiple
neutral fragments can result from a single ionized molecule.
Unfortunately, in a normal RGA measurement with the ion-
izing filament on, the signal from the ionizer-induced frag-
mentation of the parent molecule is large and overlaps the
small signal from the neutral molecular fragments produced
in the pump. In many experiments, both the neutral and ion-
ized molecular fragments that escape the ion pump could
cause undesired reactions to occur.
When the ion pump is operating, Fig. 3A shows a very
large peak near 1 amu, which we attribute to H atoms the
maximum value of the peak is actually at 1.5 amu, but this is
due to a cutoff in RGA response near 1 amu, and the unit’s
inability to record data below 1 amu.H
2
can be excluded as
the source of that peak, since when dosed in the chamber
with the RGA filament on it appears as a peak centered at 2.0
amu. The ionized atomic hydrogen is being released from the
ion pump while it is pumping a modest gas pressure. It is a
known effect that after extended use, ion pumps become
saturated and reach an equilibrium condition where ion bom-
bardment on the cathode causes gas re-emission.
12
This satu-
ration point can be reached in as little as1hatpressures near
10
−6
Torr for reactive gases. Sputtering of the cathode when
exposed to a large gas flux also releases molecules that were
previously pumped, and a measurable amount of hydrogen
and other gases previously adsorbed is released from the
cathode. The mass spectrum in Fig. 3A shows that at least
part of those released gases will ionize and fragment to gen-
erate atomic hydrogen ions. Although we cannot measure
them directly in the RGA, it is likely that neutral hydrogen
atoms are also released by this process. Similarly, the mass
peak at 28 likely corresponds to carbon monoxide CO,a
gas commonly present in background UHV pressure, being
re-emitted from the ion pump. Our STM experiments show
that these reactive H atoms/ions are able to abstract H from
the H Si100-2 1 surface, creating a new DB site for line
growth.
13
In our chamber geometry there is no line of sight
between the ion pump and the sample in the STM; however,
we have found that with substantial probability atomic H can
survive chamber wall collisions to reach a sample. This was
confirmed by performing hydrogen termination with no line
of sight between the hot filament producing atomic hydrogen
and the sample. We found that the surface would still become
H-terminated surface despite atomic hydrogen needing to
bounce off the chamber walls to reach the sample.
FIG.3. Color online兲共A Averaged mass spectra obtained with a RGA with
its filament turned off in three different conditions: Line 1, in blue, the
background signal with the ion pump turned on and no molecules being
dosed average of 1000 spectra; line 2, in green, the mass spectra while
dosing 10
−5
Torr of 2,3-dimethyl-1,3-butadiene with only the turbomolecu-
lar pump operating average of 26 spectra; line 3, in black, the mass spectra
from the same pressure 10
−5
Torr of 2,3-dimethyl-1,3-butadiene with the
ion pump operating average of 31 spectra. The curves have been offset for
clarity; below 10 amu, line 3 has been reduced in height by a factor of 5; the
scales for the three lines are identical otherwise. B Expected fragmentation
pattern for 2,3-dimethyl-1,3-butadiene as measured in a mass spectrometer
data from NIST Ref. 11. C Mass spectra obtained with a RGA with its
filament turned on and the chamber pumped by the ion pump: line 1, in
black, helium at the pressure of 2 10
−6
Torr; line 2, in blue, the back-
ground signal, arbitrarily offset for clarity pressure of 1 10
−10
Torr.
250 Zikovsky et al.: Reaction of a hydrogen-terminated Si100 surface 250
J. Vac. Sci. Technol. A, Vol. 27, No. 2, Mar/Apr 2009
Downloaded 19 Jul 2013 to 129.128.50.162. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvsta/about/rights_and_permissions
From the mass spectra presented above, we conclude that
hydrogen atoms/ions and reactive molecular fragments es-
caping from the ion pump created the surface DBs that re-
sulted in the additional molecular growth features seen in
Fig. 2B. Based on the number of created DB sites, and
assuming an impingement rate of 1 molecule s
−1
for each
surface atom in the STM image, 0.007% of molecules hitting
the surface during exposure are impurities. Such a result is
consistent with a minute fraction of reactive species escaping
the ion pump.
Subsequently, we performed an experiment where a nor-
mal H-terminated Si100 surface was exposed to a pressure
of 2 10
−6
Torr of ultrapure He for 20 min a 2400 L dose
while pumped with the ion pump. Figure 3C, line 1, shows
a normal mass spectrum acquired with the ionizing filament
on during this dose. A large amount of H
2
amu 2 is seen to
be released from the ion pump. Numerous other peaks far
exceeding background levels Fig. 3C, line 2 are visible in
this spectrum—notably peaks at amu 15 and 16, which we
assign to CH
3
and CH
4
, respectively. The height of these
peaks, 10
−8
Torr, is far above the concentration of impu-
rities present in the ultrapure helium used 99.999% mini-
mum purity, giving a total partial pressure of impurities be-
low 2 10
−11
Torr. These peaks are likely due to the
release of methane from the ion pump. We suspect that some
carbon and hydrogen were released by sputtering action in-
side the ion pump, forming CH
4
and CH
3
and releasing these
into the UHV chamber.
12
Registered STM images of the surface taken before and
after dosing He are shown in Fig. 4. These images show that
the He dose created a range of new features on the surface—
first, several DBs were created by atomic hydrogen ions or
other reactive species released by the ion pump. The concen-
tration of new surface DBs, 1.7 10
12
cm
−2
, is within an
order of magnitude of what was observed in Fig. 2B. Sec-
ond, a smaller number DBs found to have been capped by an
incoming H atom. Finally, some DBs were seen to have re-
acted with unknown molecular fragments, imaged as bright
features. These results show that fragments of molecules pre-
viously pumped by an ion pump in our case, H
2
during
hydrogen termination and organic molecules during dosing
can be released and modify reactive surfaces even when a
nonreactive, noble gas such as He is dosed.
Naturally, the results described here spurred reassessment
of recently published reports of styrene and related mol-
ecules undergoing self-directed line growth. Fortunately, in
all of our published work, reaction probability was suffi-
ciently large that the doses employed were smaller than those
required to show significant spurious ion-pump induced ef-
fects.
IV. CONCLUSIONS
In conclusion, we have found that dosing organic mol-
ecules such as 2,3-dimethyl-1,3-butadiene or a noble gas
such as He at modest pressures in a ion-pumped UHV sys-
tem degrades the Si100-2 1-H surface by producing re-
active radicals that either remove H atoms or bond to the
surface. Before-and-after STM images have shown the modi-
fication to the surface from this treatment, which does not
occur if the molecules are dosed at the same pressure but
pumped with a turbomolecular pump. In the case of the 2,3-
dimethyl-1,3-butadiene experiment, the dangling bonds cre-
ated by these reactive impurities resulted in extra growth of
molecular lines. Mass spectra obtained during these doses
confirmed the presence of large amounts of atomic hydrogen
ions re-emitted from the ion pump, as well as fragments of
the dosed molecule. He gas dosed in the same way showed
the creation of new DBs and the capping of existing DBs
with atomic hydrogen ions or other reactive molecular frag-
ments.
These results serve as a cautionary note to UHV experi-
menters using any surface that is reactive to radicals, as
hydrogen-terminated Si is. For line-growth experiments, or if
STM-tip lithography is used with the remaining surface H
atoms to be used as a resist for further surface modification,
care should be taken to avoid exposing the surface to reactive
atomic hydrogen ions or molecular fragments. This may be
overcome in some circumstances through the use of directed
dosing where the local pressure at the surface exceeds that of
the rest of the vacuum chamber and the ion pump.
ACKNOWLEDGMENTS
We thank Dr. Gino DiLabio for helpful discussions. Fund-
ing was provided by iCORE, NSERC, and CIFAR.
1
J. W. Lyding, G. C. Abeln, T. C. Shen, C. Wang, and J. R. Tucker, J. Vac.
Sci. Technol. B 12, 3735 1994.
2
S. R. Schofield, N. J. Curson, M. Y. Simmons, F. J. Ruess, T. Hallam, L.
Oberbeck, and R. G. Clark, Phys. Rev. Lett. 91, 136104 2003.
3
G. P. Lopinski, D. D. M. Wayner, and R. A. Wolkow, Nature 406,48
2000.
4
T. R. Leftwich and A. V. Teplyakov, Surf. Sci. Rep. 63,12008.
5
S. A. Dogel, G. A. DiLabio, J. Zikovsky, J. L. Pitters, and R. A. Wolkow,
J. Phys. Chem. C 111, 11965 2007.
6
J. L. Pitters, I. Dogel, G. A. DiLabio, and R. A. Wolkow, J. Phys. Chem.
FIG. 4. STM images 40 40 nm
2
of n -type medium doped
H–Si100-2 1 surface before left; 0.07 nA, 2.5 V and after right;
0.07 nA, 2.5 V helium was leaked into the chamber pumped with the ion
pump at the pressure of of 2 10
−6
Torr for 20 min 2400 L exposure.
Some changes on the surface are highlighted as follows: the dashed circle
indicates an area where four new DBs appeared, the arrows mark DBs
which are present on the left image but look much brighter and larger than
a usual DB on the right image, and the wedge shows an example of a DB
that disappeared. Overall we counted 27 new DBs, 6 modified DBs, and 2
capped DBs in the displayed area. This yields a density of 1.7
10
12
new DBs/ cm
2
.
251 Zikovsky et al.: Reaction of a hydrogen-terminated Si100 surface 251
JVST A - Vacuum, Surfaces, and Films
Downloaded 19 Jul 2013 to 129.128.50.162. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvsta/about/rights_and_permissions
B 110, 2159 2006.
7
J. Zikovsky, S. A. Dogel, A. B. Haider, G. A. DiLabio, and R. A.Wolkow,
J. Phys. Chem. A 111, 12257 2007.
8
M. Z. Hossain, H. S. Kato, and M. Kawai, J. Phys. Chem. B 109, 23129
2005.
9
J. J. Boland, Surf. Sci. 261,171992.
10
J. Zikovsky, S. A. Dogel, S. Sinha, G. A. DiLabio, and R. A. Wolkow,
Chem. Phys. Lett. 458,1172008.
11
NIST, NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/,
2008.
12
M. H. Hablanian, High–Vacuum Technology: A practical guide, 2nd ed.
Dekker, New York, 1997.
13
I. A. Dogel, S. A. Dogel, J. L. Pitters, G. A. DiLabio, and R. A. Wolkow,
Chem. Phys. Lett. 448, 237 2007.
252 Zikovsky et al.: Reaction of a hydrogen-terminated Si100 surface 252
J. Vac. Sci. Technol. A, Vol. 27, No. 2, Mar/Apr 2009
Downloaded 19 Jul 2013 to 129.128.50.162. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvsta/about/rights_and_permissions
... Ref. [38]. ...
... (C): mass spectra obtained with a RGA with its filament turned on and the chamber pumped by the ion pump: Line 1, in black, helium at the pressure 2 × 10 −6 Torr; Line 2, in blue, the background signal, arbitrarily offset for clarity (pressure 10 −10 Torr). Figure and caption from Ref. [38]. ...
... From Zikovsky et al. [38]: ...
Full-text available
Thesis
Extending electronic devices beyond the limitations of current micro-electronics manufacturing will require detailed knowledge of how to make contacts to semi- conductor surfaces. In this work, we investigated several methods by which such connections to silicon surfaces could be achieved. Scanning tunneling microscopy (STM) was our main experimental tool, allowing direct imaging of the surfaces at the atomic level. First, the growth of self-forming linear nanostructures of organic molecules on silicon surfaces offers a possibility of creating devices with hybrid organic-silicon functionality. We have studied the growth of many different molecules on a variety of hydrogen-terminated silicon surfaces: H-Si(100)-2x1, H-Si(100)-3x1, and H-Si(111)-1x1. We found molecular growth patterns affected by steric crowding, by sample doping level, or by exposure to ion-pump created radicals. We formed the first contiguous "L-shaped" molecular lines, and used an external electric field to direct molecular growth. We attempted to study a novel method for nanoscale information transfer along molecular lines based on excitation energy transfer. The second part of the work focuses on the development and use of a new multiple-probe STM instrument. The design and the custom STM control software written for it are described. Connections to Si surfaces were achieved with a combination of lithographically defined metal contacts and STM tips. Two-dimensional surface conductivity of the Si(111)-7x7 surface was measured, and the effect of modifying the surface with organic molecules was investigated. A novel method, scanning tunneling fractional current imaging (STFCI), was developed to further study surface conductance. This method allowed us to determine, for the first time, that the resistance of steps on the Si(111)-7x7 surface is significantly higher than that of the surface alone.
Article
The paper offers a simulation solution for the cellular membrane in a microelectronics engineering software. The cell membrane presents the first instance for the electrical stimulus in the cellular biology. The electrical transmembranar voltage can be extracted by simulations along the entire cross-section of the membrane. The action potential distribution can be extracted. The insulator properties of the bi-lipidic layer can be introduced as input data in software. Due to the ionophore channels, the membrane selectively allows the ionic electrical charge transport. Therefore, the i.c. and e.c. environments that are wealthy in ionic electric charge were assumed as a semiconductor material.
Article
The reaction mechanism for the adsorption and growth of allylic mercaptan (ALM) at a defect site on the Si(100)-2 × 1 surface has recently been proposed. The adsorbate structure is believed to be a branched or linear ALM molecule forming a bridge across silicon dimer rows on the Si(100)-2 × 1 surface. Subsequent reactions at the radical site formed by an ALM adsorbate have not been studied previously. We have now calculated the reactivity of ALM, acetone, and styrene at radical sites formed by an ALM adsorbate. The reactivity of ALM and acetone is unaffected by adjacent ALM adsorbates. The same is true for styrene reacting adjacent to a linear ALM adsorbate. A branched adsorbate significantly destabilizes a styrene adsorbate, making styrene more likely to desorb than to react further. The origin of this destabilization is the partially broken silicon dimer bond. These results are consistent with available experimental observations and support the proposal of a branched ALM adsorbate bridging dimer rows.Keywords (keywords): allylic mercaptan; styrene; acetone; molecular line growth; nanopattern; Si(100)-2 × 1 surface; molecular wires
Article
The growth of molecular lines on the Si(100)-2 × 1 surface is an area of intense interest due to its possible application to the microelectronics industry. While many molecules have been found to grow along dimer rows of the Si(100)-2 × 1 surface, only allylic mercaptan (ALM) has been shown to grow exclusively across dimer rows. In this work, we compare several possible mechanisms for line growth across silicon dimer rows. We conclude that a reaction mechanism starting from a branched carbon radical adsorbed on the surface and ending with an interdimer bridged structure is the most probable mechanism for initial ALM reactivity.
Article
A method for constructing the potential energy surface for reactions of a molecule with the surface of cleaved non-conducting crystals is reported. The method uses systematic fragmentation to express the total potential in terms of potential energy surfaces which describe reactions of relatively small molecules in the gas phase. The approach is illustrated by an application to the reaction of hydrogen atoms with a hydrogen-terminated silicon(111) surface.
Article
Constant miniaturization of electronic devices and interfaces needed to make them functional requires an understanding of the initial stages of metal growth at the molecular level. The use of metal-organic precursors for metal deposition allows for some control of the deposition process, but the ligands of these precursor molecules often pose substantial contamination problems. One of the ways to alleviate the contamination problem with common copper deposition precursors, such as copper(I) (hexafluoroacetylacetonato) vinyltrimethylsilane, Cu(hfac)VTMS, is a gas-phase reduction with molecular hydrogen. Here we present an alternative method to copper film and nanostructure growth using the well-defined silicon surface. Nearly ideal hydrogen termination of silicon single-crystalline substrates achievable by modern surface modification methods provides a limited supply of a reducing agent at the surface during the initial stages of metal deposition. Spectroscopic evidence shows that the Cu(hfac) fragment is present upon room-temperature adsorption and reacts with H-terminated Si(100) and Si(111) surfaces to deposit metallic copper. Atomic force microscopy (AFM) and scanning electron microscopy (SEM) are used to follow the initial stages of copper nucleation and the formation of copper nanoparticles, and X-ray energy dispersive spectroscopy (XEDS) confirms the presence of hfac fragments on the surfaces of nanoparticles. As the surface hydrogen is consumed, copper nanoparticles are formed; however, this growth stops as the accessible hydrogen is reacted away at room temperature. This reaction sets a reference for using other solid substrates that can act as reducing agents in nanoparticle growth and metal deposition.
Article
Highly ordered hydrogen-terminated silicon surfaces are ideal testing grounds for molecular electronics. However, upon formation of these surfaces it is inevitable that some surface sites are not capped by hydrogen. These remaining dangling bonds can interfere with the chemical and electronic properties of nanostructures formed on the silicon surface. In this work, using scanning tunneling microscopy, high resolution electron energy loss spectroscopy and ab initio computational methods, we explore two chemical approaches to refining the hydrogen termination process. We investigate the utility of diimide (N2H2) and N,N-diethylhydroxylamine (DEHA) as hydrogen atom sources that have the ability to cap dangling bonds.
Article
The nanoscale structuring of molecules on silicon surfaces is one approach for combining the tuneable properties of chemical species with the functionality of semiconductor materials. In this study, we report on the growth characteristics of trimethylene sulfide (TMS) on p- and n-type H−Si(100)-2 × 1. The nanostructures formed by TMS on either surface are indistinguishable by scanning tunneling microscopy (STM). However, high-resolution electron energy loss spectroscopy (HREELS) and modeling by density functional theory indicate that the molecular attachment mechanism differs with dopant type. Our results show that TMS adds to a surface silicon dangling bond through the formation of a Si−S bond on p-type silicon and through the formation of a Si−C bond on n-type silicon. In both cases, the added TMS undergoes ring opening following covalent bond formation with the surface. The different ring-opened radicals are able to abstract a hydrogen atom from one of two neighboring silicon dimers. The overall reaction produces TMS-derived nanostructures that grow via a square-wave pattern on the neighboring edges of two dimer rows.
Article
The formation of 1,3-butadiene and 2,3-dimethyl-1,3-butadiene derived nanostructures on hydrogen-terminated silicon(100)-2×1 was studied using ultra-high vacuum scanning tunneling microscopy and density functional theory modeling. We find that the steric interactions between the methyl groups on 2,3-dimethyl-1,3-butadiene and the surface hydrogen causes 2,3-dimethyl-1,3-butadiene to grow relatively long, linear structures along one side of a dimer row. However, the less sterically encumbered 1,3-butadiene grows short, less orderly, self-terminating structures that involve at least two dimer rows. These results show that steric effects can be used as a parameter in the rational design of organic nanostructures on silicon surfaces.
Article
Nanoscale patterning of the Si(100)‐2×1 monohydride surface has been achieved by using an ultrahigh vacuum (UHV) scanning tunneling microscope (STM) to selectively desorb the hydrogen passivation. Hydrogen passivation on silicon represents one of the simplest possible resist systems for nanolithography experiments. After preparing high quality H‐passivated surfaces in the UHV chamber, patterning is achieved by operating the STM in field emission. The field emitted electrons stimulate the desorption of molecular hydrogen, restoring clean Si(100)‐2×1 in the patterned area. This depassivation mechanism seems to be related to the electron kinetic energy for patterning at higher voltages and the electron current for low voltage patterning. The patterned linewidth varies linearly with the applied tip bias achieving a minimum of ≪10 Å at -4.5 V. The dependence of linewidth on electron dose is also studied. For positive tip biases up to 10 V no patterning occurs. The restoration of clean Si(100)‐2×1 is suggestive of selective area chemical modifications. This possibility has been explored by exposing the patterned surface to oxygen and ammonia. For the oxygen case, initial oxidation of the patterned area is observed. Ammonia dosing, on the other hand, repassivates the surface in a manner different from that of atomic hydrogen. In both cases the pattern resolution is retained and the surrounding H‐passivated areas remain unaffected by the dosing.
Article
To realize nanoscale wiring in two dimensions (2D), the fabrication of interconnected one-dimensional (1D) molecular lines through the radical chain reaction of alkene molecules (CH2=CH-R) on the H-terminated Si(100)-(2 x 1) surface have been investigated using scanning tunneling microscopy (STM) at 300 K. By the judicious choice of R in the CH2=CH-R molecule and by creating a dangling bond (DB) at a desired point using the STM tip, the perpendicularly connected allyl mercaptan (ALM) and styrene lines have been fabricated on the Si(100)-(2 x 1)-H surface. The continuous growth of the styrene line at the end DB of a growing ALM line (or vice versa) does not occur, perhaps, due to steric hindrance or/and interaction between adsorbed molecules.
Article
The self-directed growth of organic molecules on silicon surfaces allows for the rapid, parallel production of hybrid organic-silicon nanostructures. In this work, the formation of benzaldehyde- and acetaldehyde-derived nanostructures on hydrogen-terminated H-Si(100)-2x1 surface is studied by scanning tunneling microscopy in ultrahigh vacuum and by quantum mechanical methods. The reaction is a radical-mediated process that binds the aldehydes, through a strong Si-O covalent bond, to the surface. The aldehyde nanostructures are generally composed of double lines of molecules. Two mechanisms that lead to double line growth are elucidated.
Article
Future nanoscale integrated circuits will require the realization of interconnections using molecular-scale nanostructures; a practical fabrication scheme would need to be largely self-assembling and operate on a large number of like structures in parallel. The self-directed growth of organic molecules on hydrogen-terminated silicon(100) [H-Si(100)] offers a simple method of realizing one-dimensional molecular lines. In this work, we introduce the ability to change the growth direction and form more complex, contiguous shapes. Numerous styrene and trimethylene sulfide L shapes were grown on a H-Si(100)-3x1 surface in parallel with no intermediate surface lithography steps, and similar shapes were also grown using allyl mercaptan and benzaldehyde on H-Si(100)-2x1. Registered scanning tunneling microscopy (STM) images and high-resolution electron energy loss spectroscopy (HREELS) were used to investigate the growth process.
Reaction of a hydrogen-terminated Si " 100… surface 252
  • J Zikovsky
  • S A Dogel
  • A B Haider
  • G A Dilabio
  • R A Wolkow Hossain
  • H S Kato
  • M Kawai
  • S A Dogel
  • S Sinha
  • G A Dilabio
  • R A Wolkow
  • Zikovsky
B 110, 2159 ͑2006͒. 7 J. Zikovsky, S. A. Dogel, A. B. Haider, G. A. DiLabio, and R. A.Wolkow, J. Phys. Chem. A 111, 12257 ͑2007͒. 8 M. Z. Hossain, H. S. Kato, and M. Kawai, J. Phys. Chem. B 109, 23129 ͑2005͒. 9 J. J. Boland, Surf. Sci. 261, 17 ͑1992͒. 10 J. Zikovsky, S. A. Dogel, S. Sinha, G. A. DiLabio, and R. A. Wolkow, Chem. Phys. Lett. 458, 117 ͑2008͒. 11 NIST, NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/, 2008. 12 M. H. Hablanian, High–Vacuum Technology: A practical guide, 2nd ed. ͑Dekker, New York, 1997͒. 13 I. A. Dogel, S. A. Dogel, J. L. Pitters, G. A. DiLabio, and R. A. Wolkow, Chem. Phys. Lett. 448, 237 ͑2007͒. 252 Zikovsky et al.: Reaction of a hydrogen-terminated Si " 100… surface 252 J. Vac. Sci. Technol. A, Vol. 27, No. 2, Mar/Apr 2009 Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 141.212.109.170 On: Mon, 15 Dec 2014 16:09:17
  • J Zikovsky
  • S A Dogel
  • A B Haider
  • G A Dilabio
  • R A Wolkow
J. Zikovsky, S. A. Dogel, A. B. Haider, G. A. DiLabio, and R. A.Wolkow, J. Phys. Chem. A 111, 12257 2007. 8
  • M Z Hossain
  • H S Kato
  • M Kawai
M. Z. Hossain, H. S. Kato, and M. Kawai, J. Phys. Chem. B 109, 23129 2005. 9