Low temperature electron transport on semiconductor surfaces
The low temperature electron transport on semiconductor surfaces has been studied using an ultra high vacuum, variable temperature Scanning Tunneling Microscope (STM). The STM I(V) spectroscopy recorded at various temperatures has enabled to investigate the temperature dependence (300 K to 35 K) of...
Gespeichert in:
| Datum: | 2003 |
|---|---|
| Hauptverfasser: | , , , , |
| Format: | Artikel |
| Sprache: | English |
| Veröffentlicht: |
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України
2003
|
| Schriftenreihe: | Физика низких температур |
| Schlagworte: | |
| Online Zugang: | http://dspace.nbuv.gov.ua/handle/123456789/128815 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Low temperature electron transport on semiconductor surfaces / M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, G. Dujardin // Физика низких температур. — 2003. — Т. 29, № 3. — С. 263-269. — Бібліогр.: 9 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
irk-123456789-128815 |
|---|---|
| record_format |
dspace |
| spelling |
irk-123456789-1288152018-01-15T03:02:50Z Low temperature electron transport on semiconductor surfaces Lastapis, M. Riedel, D. Mayne, A. Bobrov, K. Dujardin, G. Electronically Induced Phenomena: Low Temperature Aspects The low temperature electron transport on semiconductor surfaces has been studied using an ultra high vacuum, variable temperature Scanning Tunneling Microscope (STM). The STM I(V) spectroscopy recorded at various temperatures has enabled to investigate the temperature dependence (300 K to 35 K) of the surface conductivity of three different semiconductor surfaces: highly doped n-type Si(100), p-type Si(100), and hydrogenated C(100). Low temperature freezing of specific surface electronic channels on the higly doped n-type Si(100) and moderately doped p-type Si(100) surfaces could be achieved whereas the total surface conductivity on the hydrogenated C(100) surface can be frozen below only 180 K. 2003 Article Low temperature electron transport on semiconductor surfaces / M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, G. Dujardin // Физика низких температур. — 2003. — Т. 29, № 3. — С. 263-269. — Бібліогр.: 9 назв. — англ. 0132-6414 PACS: 72.20.Jv http://dspace.nbuv.gov.ua/handle/123456789/128815 en Физика низких температур Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| topic |
Electronically Induced Phenomena: Low Temperature Aspects Electronically Induced Phenomena: Low Temperature Aspects |
| spellingShingle |
Electronically Induced Phenomena: Low Temperature Aspects Electronically Induced Phenomena: Low Temperature Aspects Lastapis, M. Riedel, D. Mayne, A. Bobrov, K. Dujardin, G. Low temperature electron transport on semiconductor surfaces Физика низких температур |
| description |
The low temperature electron transport on semiconductor surfaces has been studied using an ultra high vacuum, variable temperature Scanning Tunneling Microscope (STM). The STM I(V) spectroscopy recorded at various temperatures has enabled to investigate the temperature dependence (300 K to 35 K) of the surface conductivity of three different semiconductor surfaces: highly doped n-type Si(100), p-type Si(100), and hydrogenated C(100). Low temperature freezing of specific surface electronic channels on the higly doped n-type Si(100) and moderately doped p-type Si(100) surfaces could be achieved whereas the total surface conductivity on the hydrogenated C(100) surface can be frozen below only 180 K. |
| format |
Article |
| author |
Lastapis, M. Riedel, D. Mayne, A. Bobrov, K. Dujardin, G. |
| author_facet |
Lastapis, M. Riedel, D. Mayne, A. Bobrov, K. Dujardin, G. |
| author_sort |
Lastapis, M. |
| title |
Low temperature electron transport on semiconductor surfaces |
| title_short |
Low temperature electron transport on semiconductor surfaces |
| title_full |
Low temperature electron transport on semiconductor surfaces |
| title_fullStr |
Low temperature electron transport on semiconductor surfaces |
| title_full_unstemmed |
Low temperature electron transport on semiconductor surfaces |
| title_sort |
low temperature electron transport on semiconductor surfaces |
| publisher |
Фізико-технічний інститут низьких температур ім. Б.І. Вєркіна НАН України |
| publishDate |
2003 |
| topic_facet |
Electronically Induced Phenomena: Low Temperature Aspects |
| url |
http://dspace.nbuv.gov.ua/handle/123456789/128815 |
| citation_txt |
Low temperature electron transport on semiconductor surfaces / M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, G. Dujardin // Физика низких температур. — 2003. — Т. 29, № 3. — С. 263-269. — Бібліогр.: 9 назв. — англ. |
| series |
Физика низких температур |
| work_keys_str_mv |
AT lastapism lowtemperatureelectrontransportonsemiconductorsurfaces AT riedeld lowtemperatureelectrontransportonsemiconductorsurfaces AT maynea lowtemperatureelectrontransportonsemiconductorsurfaces AT bobrovk lowtemperatureelectrontransportonsemiconductorsurfaces AT dujarding lowtemperatureelectrontransportonsemiconductorsurfaces |
| first_indexed |
2025-07-09T09:57:47Z |
| last_indexed |
2025-07-09T09:57:47Z |
| _version_ |
1837162896266625024 |
| fulltext |
Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269
Low temperature electron transport on semiconductor
surfaces
M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, and G. Dujardin
Laboratoire de Photophysique Moleculaire, Batiment 210
Universite Paris-Sud, 91405 Orsay Cedex, France
E-mail: gerald.dujardin@ppm.u-psud.fr
Received October 2, 2002
The low temperature electron transport on semiconductor surfaces has been studied using an
ultra high vacuum, variable temperature Scanning Tunneling Microscope (STM). The STM I(V)
spectroscopy recorded at various temperatures has enabled to investigate the temperature
dependence (300 K to 35 K) of the surface conductivity of three different semiconductor surfaces:
highly doped n-type Si(100), p-type Si(100), and hydrogenated C(100). Low temperature freezing
of specific surface electronic channels on the higly doped n-type Si(100) and moderately doped
p-type Si(100) surfaces could be achieved whereas the total surface conductivity on the
hydrogenated C(100) surface can be frozen below only 180 K.
PACS: 72.20.Jv
1. Introduction
The operation of nanoscale electronic devices
requires an efficient electronic decoupling from the
substrate on which the devices have been built ���.
This can be best achieved either on insulating or on
low temperature semiconducting substrates. For
example, the electronic transport of gold nanowires
deposited on a Si(111) surface has been recently
measured at low temperature (4 K) �1�. At such a low
temperature, all the measured current flows through
the gold nanowire as long as the applied bias voltage
is within the silicon electronic band gap. Indeed, the
electronic channels through the silicon surface whose
energy are located within the band gap are assumed to
be frozen at such a low temperature. Low temperature
semiconductor surfaces have several advantages over
insulators. First, they can still be conducting at low
temperature, when sufficiently doped �2�. This allows
the use of experimental surface techniques requiring
some surface conductivity such as the Scanning
Tunneling Microscope (STM). Second, the preparation
of high quality surfaces having few defects at the
atomic-scale is much easier for semiconductors.
The low temperature electron transport properties
of bulk semiconductors are rather well understood �2�,
mainly in terms of the decreased number of free
carriers. However, when dealing with surface
properties, the situation is much more complicated.
Indeed, the interface between the semiconductor and the
vacuum can produce specific surface (or sub-surface)
electronic states inducing charges or holes confined at
the surface �3,4�, which may completely modify the low
temperature electron transport properties.
The low temperature electron transport properties of
semiconductor surfaces have never been investigated
before at the atomic-scale, except a recent study of the
Ge(111) surface ���. In this paper, we report the
temperature dependence of the electron transport
properties of three different semiconductor surfaces:
highly doped n-type Si(100), p-type Si(100), and
hydrogenated C(100). We used the I(V) spectroscopy
using the STM performed at various temperatures to
explore the conductivity of these surfaces. The
observed behavior is very different in these three cases.
It will be shown that specific surface conductivity
channels can be frozen at low temperature for highly
doped n-type Si(100) and p-type Si(100), whereas the
surface conductivity is completely frozen at low
temperature for hydrogenated C(100).
2. Experimental
The electron transport properties of semiconductor
surfaces have been measured in an ultra high vacuum
(UHV) chamber (base pressure 2�10–11 torr) using a
variable temperature Scanning Tunneling Microscope
(STM). The sample can be cooled down to 30 K using
a helium liquid flow cryostat.
© M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, and G. Dujardin, 2003
Three types of semiconductor samples have been
studied:
— highly doped n-type Si(100): the silicon sample
(6�2�0.1 mm) was n doped with Arsenic (bulk
resistivity of 0.004 to 0.007 � � cm). Clean and well
ordered Si(100)—(2�1) reconstructed surfaces were
obtained by first outgassing for at least 12 hours at
700°C in UHV before flashing to 1080°C to remove the
oxide layer. The duration of each flash was adjusted
such that the pressure remained below 1·10–9 torr
(usually < 2���
–10 torr).
— p-type Si(100): the silicon sample (6 2� �
�0 25. )mm was p doped with Boron (bulk resistivity
0.7 to 1.3 � � cm). Clean and well ordered
Si(100)—(2�1) reconstructed surfaces were obtained
as described before.
— hydrogenated C(100): the diamond sample
(6�1�0.2 mm) is a natural single crystal of (100)
orientation. It is a weakly Boron p-type doped sample.
Prior to insertion into the UHV chamber, the
diamond sample was ex-situ saturated with hydrogen
in a microwave hydrogen plasma at 800°C for 1 hour.
Details on the hydrogenation procedure can be found
elsewhere �6�. Clean and well ordered hydrogenated
C(100)–(2�1):H surfaces were obtained by
outgassing in UHV for a few minutes at 300°C to
remove any physisorbed species from the surface.
During the STM experiments, the edges of the
sample were sandwiched between two molybdenum
plates to which the voltage was applied.
Experiments were performed as follows at a given
sample temperature between 30 K and 300 K. The
sample surface was first imaged at a constant tunnel
current with the STM. For some samples and some
temperatures, no tunnel current could be established
between the STM tip and the sample. In such cases,
the STM tip crashed on the surface. Where the STM
imaging was possible, the STM tip was located at
fixed positions across the surface and I(V)
spectroscopy curves were recorded at a fixed
tip-surface distance.
3. Results and discussions
Before we discuss the results, one should note that
the I(V) STM spectroscopy curves are usually
considered to be due only to the STM junction
conductivity. Indeed, the conductance between the
electronic surface states located under the STM tip
and the sample metallic holder to which the voltage is
applied is assumed to be very high. In such a case, the
I(V) curves only reflect the coupling between the
surface electronic states of the tungsten tip with those
of the sample surface � �. However, at low
temperature, where the resistivity of the semicon-
ductor sample cannot be neglected anymore, the
analysis of the I(V) curves is more complicated. The
measured I/V conductivity is the combination of the
STM junction conductivity itself, the coupling
between the electronic surface states of the sample and
the electronic channels through the sample and the
conductivity of these electronic channels as far as the
sample molybdenum holders. At this point, the
electronic channels through the sample can be either
surface or bulk electronic channels.
3.1. Highly doped n-type Si(100)
The highly doped n-type Si(100) surface could be
easily imaged with the STM at any temperature
between 30 K and 300 K under usual sample voltage
(V = –1.5 V) and tunnel current (I = 0.5 nA)
conditions (see Fig. 1). This is not surprising since
such a highly doped silicon sample is known to have
an almost constant conductivity in this temperature
range �
�. More astonishing are the I(V) curves
recorded as a function of the temperature (Fig. 2). As
seen in Fig. 2,b, the dI/dV curve at 300 K shows a
narrow band gap of about 0.2 eV whereas the dI/dV
curve at 35 K shows a band gap of about 1 eV, i.e.,
equal to the bulk electronic band gap of silicon � �.
This temperature effect cannot be explained by any
freezing of the bulk conductance since such a highly
doped silicon sample is known to have a bulk
conductivity which is almost constant between 300 K
and 35 K �
�. A similar opening of the surface band
gap at low temperature has been observed previously
on the Ge(111) surface �5�. It has been assigned to low
264 Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269
M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, and G. Dujardin
Fig. 1. Scanning Tunneling Microscope topography (390 Å
by 240 Å) of the highly doped n-type Si(100)—2�1 surface
recorded at 35 K. The sample voltage is VS = –1.5 V and
the tunnel current I = 0.5 nA.
temperature freezing of specific surface electronic
channels whose energies are located within the
electronic bulk band gap �5�. Recent measurements of
the total surface conductivity of the Si(100) surface
�7� have shown that, on the contrary, the surface
conductivity increases at low temperature. One
cannot completely exclude that the conductivity of
the specific surface electronic channels whose energies
are located within the bulk band gap would decrease
at low temperature whereas the total surface
conductivity would increase. However such an
explanation seems very unlikely. Therefore, another
explanation needs to be considered as illustrated in
Fig. 3. At room temperature, the tunnel current
flowing through occupied surface states located
within the bulk band gap requires some transport of
electrons from the bulk conduction band (CB) to the
occupied surface states (SS). This can be achieved by
the combination of two effects, (i) the electron
transport through the energy barrier of the upward
surface band bending caused by the «pinning» of the
surface states at the Fermi level � � and (ii) the release
of electron energy to reach the lower lying surface
states. This latter effect would require some coupling
between electrons and phonons. Considering that the
transmission over the energy barrier as well as the
phonon population are thermally activated, both
effects are expected to be much less efficient at low
temperature, thus explaining the freezing of the
electronic channels through the surface states located
within the bulk band gap.
Low temperature electron transport on semiconductor surfaces
Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269 265
Fig. 2. dI/dV curves recorded on the highly doped
n-type Si(100)–(2�1) surface at various temperatures. The
scanning conditions for recording these curves are VS =
= –1.5 V and I = 0.5 nA.
Fig. 3. Schematic energy diagram of electrons flowing
from the highly doped n-type Si(100)–(2�1) sample to the
STM tip. At 300 K (a), the thermal energy can (i)
activate the transport of electrons from the sample Fermi
level EF through the sub-surface barrier due to the
upward band bending and (ii) activate the coupling with
phonons and surface states (SS). At 35 K (b), the loss of
thermal activation freezes the electron current.
3.2. p-type Si(100)
At room temperature, the p-type Si(100) sample
shows a STM topography and I(V) spectroscopy
curves similar to the highly doped n-type Si(100) (see
Fig. 4). However, as soon as the sample temperature is
lowered to about 180 K, both the STM topography and
I(V) curves are markedly modified. At 180 K, stable
STM topographies with atomic resolution can hardly
be obtained. At room temperature, STM topographies
could easily be obtained at relatively small negative
(–2 V, I = 0.5 nA) sample voltages. At 180 K, STM
topographies could be obtained only at high negative
sample voltages (–5 V, I = 0.5 nA) (see Fig. 4) and
were found to be very unstable, suggesting some local
charging occuring after a few minutes of tunneling. The
I(V) curves are also strongly modified (see Fig. 5),
with a much reduced conductivity at negative sample
voltages. At 35 K, it was impossible to obtain any STM
topography at any negative sample voltage and the
I(V) curves show a weak conductivity extending at
even larger negative sample voltages (see Fig. 5). The
conductivity is zero for sample voltages between
–3 V and +1 V.
These results are somewhat surprising since they
cannot be simply ascribed to the temperature
dependence of the bulk conductivity of p-type
Si(100). Indeed, from 300 K to 180 K, the
conductivity of the sample is considered to increase
�
�. Obviously this cannot explain the shift from –2 V
to –5 V of the sample voltage for imaging, since this
shift (associated with a tunnel current of 0.5 nA)
would correspond to a spreading resistance of 6·109
�.
In fact, these results can be well understood by
considering a tip induced band bending as
schematically shown in Fig. 6. The «pinning» of the
surface states at the Fermi level produces a downward
band bending � �. At negative sample bias, charge
carriers (h� , holes) are prevented from flowing from
the tip to the sample by the downward band bending
of the valence bands. At room temperature, the holes
can acquire enough thermal energy to overcome this
energy barrier. However, at low temperature the hole
266 Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269
M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, and G. Dujardin
50 Å
40 Å300 Ê, -2 V, 0.5 nA
.
180 Ê, -5 V, 0.5 nA
Fig. 4. Scanning Tunneling Microscope topographies of the
p-type Si(100) surface at 300 K (top – 160�160 Å) and
180 K (bottom – 250�250 Å).
Fig. 5. dI/dV curves recorded on the p-type Si(100)-2�1
surface at various temperatures. The scanning conditions
for recording these curves are (a) –2 V, 0.5 nA; (b)
–6 V, 0.5 nA; (c) – 7 V, 0.5 nA.
transport through this energy barrier is frozen. Since
the conductivity decreases, the STM tip tends to
approach the surface to maintain the tunnel current
constant. As a consequence, the electric field between
the STM tip and the surface is strongly increased,
inducing a more pronounced downward band bending
as shown in Fig. 6. This explains the large shift from
–2 V (300 K) to –5 V (180 K) for imaging the surface
as well as the shape of the I(V) curves (Fig. 5). One
should also mention here the possible influence of the
surface states on the penetration of the electric field
into the silicon sample. Indeed, the reduced charge
population of the surface states at low temperature may
decrease the screening effect and thus facilitate the
penetration of the electric field produced by the tip.
This tip-induced band bending was not observed with
the previously studied n-type sample due to its high
concentration of dopants which prevents the electric
field from penetrating inside the sample.
3.3. Hydrogenated C(100)
At room temperature, the hydrogenated C(100)
diamond surface can be imaged at the atomic-scale
with the STM at both positive (V = +1.5 V, I = 1 nA)
(see Fig. 7) and negative (V = –1.5 V, I = 1 nA)
sample bias. The corresponding I(V) curve is shown
in Fig. 8. When lowering the sample temperature, the
I/V conductivity decreases, especially at positive
sample bias (see Fig. 8), so that below 180 K no STM
imaging is possible anymore. Around 150 K, the I(V)
curve becomes «metallic» in form, i.e. linear. For
temperatures below 150 K, no tunneling current could
be established whatever the sample bias. This
behavior is quite different from the two previous cases
of the silicon surface and requires a specific
explanation.
The clean natural (weakly doped) C(100) diamond
surface is known to be insulating ���. However, when
hydrogenated, the C(100) diamond surface becomes
conductive and STM imaging in the usual tunneling
mode can be performed �9�. It has been demonstrated
recently �9� that this surface conductivity requires the
presence of both the hydrogen atoms on the surface
and sub-surface oxygen and hydrogen. These
sub-surface species, which are produced during the
preparation procedure of the hydrogenated sample,
result in an upward surface band bending (see Fig. 9)
and a concentration of holes in the sub-surface region.
As seen in Fig. 9, this enables a flow of holes at both
negative and positive sample bias. When reducing the
Low temperature electron transport on semiconductor surfaces
Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269 267
Fig. 6. Schematic energy diagram of electrons showing the
flow of positive holes (h+) from the STM tip to the p-type
Si(100) sample. At 300 K (a), the thermal energy can
activate the transport of holes through the energy barrier
due to the surface state induced band bending. At 35 K
(b), the tip induced band bending is too large to allow
the flow of hole current.
40 Å
Fig. 7. 175�175 Å statement topography of the diamond
C(100)-2�1 surface recorded at room temperature (VS
�15. ,V I = 1.5 nA)
sample temperature, the activation as dopants of the
sub-surface species should rapidly decrease if their
activation energy is high enough, of the order of
0.6 eV. This explains the much reduced surface
conductivity. Around 150 K, the sub-surface
conductivity is so low that the STM tip needs to be in
full contact with the sample, thus leading to a
«metallic» I(V) curve. In summary, the freezing of the
sub-surface conductivity explains why the hydrogenated
diamond surface becomes fully insulating below 150 K.
4. Conclusions
Investigating the low temperature conductivity of
semiconductor surfaces with the STM is a very
interesting problem since a large variety of phenomena
can be encountered.
268 Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269
M. Lastapis, D. Riedel, A. Mayne, K. Bobrov, and G. Dujardin
Fig. 8. I(V) curves recorded on the diamond C(100)—(2�1)
surface as a function of the temperature. The scanning condi-
tions for recording these curves are VS
�2 V, I = 0.1 nA.
Fig. 9. Schematic energy diagram of electrons showing the
flow of positive holes (h+) for two polarities of the
voltage on the diamond C(100)–(2�1) surface.
For highly doped n-type Si(100), lowering the
sample temperature down to 35 K enables the freezing
of the electronic channels whose energies are within
the bulk band gap. Other electronic channels (outside
the bulk band gap) seem to be accessible at low
temperature, enabling the STM imaging of the sample
down to 35 K. Due to the high concentration of
dopants, tip-induced surface band bending can be
neglected.
For moderately doped p-type Si(100), the surface
band bending also freezes electronic channels at
negative sample voltages. This effect is amplified by
tip-induced surface band bending which freezes at low
temperature all the electronic channels from –2 V
down to –5 V at 35 K.
For hydrogenated C(100), the surface conductivity
is due to the presence of sub-surface species (oxygen
and hydrogen). The whole surface conductivty is
strongly dependent on the temperature since below
180 K, the hydrogenated C(100) surface becomes
fully insulating.
These three examples offer an interesting range of
applications for nanoelectronics. The highly doped
n-type Si(100) surface can be advantageously used at
low temperature to decouple the electronic channels of
a nanoscale device from the substrate only in the energy
range located within the bulk band gap. The advantage
is that the substrate is still conducting and STM
imaging can still be performed at such a low
temperature (35 K). The hydrogenated C(100) surface
can be used, below room temperature at T = 180 K, to
fully decouple the electronic channels of a nanoscale
device from the substrate. However, in this case, the
substrate is completely insulating and STM imaging is
no longer possible. The moderately doped p-type
Si(100) surface offers another potentially interesting
application where surface tip-induced band bending
can be used at low temperature (35 K) to freeze all the
electronic channels across the surface whose energy is
between –5 V and +1 V.
We wish to thank the European IST-FET
«Bottom-up-Nanomachines» (BUN) and the
European «Atomic and Molecular Manipulation; a
new tool In Science and Technology» (AMMIST)
network.
1. U. Ramsperger, T. Uchihashi, and H. Nejoh, Appl.
Phys. Lett. 78, 85 (2001).
2. S.M. Sze, Physics of Semiconductor Devices, John
Wiley and Sons, 26 (1981).
3. J.A Stroscio and R. M. Feenstra, in : Scanning
Tunneling Microscopy, J.A. Stroscio and W.J. Kaiser
(eds.) Academic Press 27 (1993).
4. W. Mönch, Semiconductor Surfaces and Interfaces,
Springer-Verlag 15 (1993).
5. G. Dujardin, A.J. Mayne, and F. Rose, Phys. Rev.
Lett. 89, 36802 (2002).
6. K. Bobrov, H. Shechter, M. Folman, and A. Hoffman,
Diamond Relat. Mater. 7, 170 (1998).
7. K. Yoo and H.H. Weitering, Phys. Rev. Lett. 87,
026802 (2001).
8. K. Bobrov, A.J. Mayne, and G. Dujardin, Nature
413, 616 (2001)
9. K. Bobrov, A.J. Mayne, G. Comtet, G. Dujardin, L.
Hellner, and A. Hoffman, in preparation.
Low temperature electron transport on semiconductor surfaces
Fizika Nizkikh Temperatur, 2003, v. 29, No. 3, p. 263–269 269
|