Field evaporation of gold atoms onto a silicon dioxide film
by using an atomic force microscope
Hajime Koyanagi,
a)
Sumio Hosaka, and Ryo Imura
Advanced Research Laboratory, Hitachi Ltd., Kokubunji, Tokyo 185, Japan
Masataka Shirai
Central Research Laboratory, Hitachi Ltd., Kokubunji, Tokyo 185, Japan
͑Received 20 June 1995; accepted for publication 5 September 1995͒
To investigate whether field evaporation of gold atoms is responsible for dot formation in an atomic
force microscope ͑AFM͒ gold-coated tip/vacuum/SiO
2
film/p-type Si substrate configuration, we
have performed elemental analysis of the dots and measured the dependence of the threshold voltage
on SiO
2
thickness with both polarities for the dot formation. The experiments demonstrate that it is
feasible to form gold dots on SiO
2
films 17–107 Å thick by adjusting the pulsed voltages applied
to the gold-coated AFM tip. Energy dispersive x-ray spectroscopy ͑EDX͒ shows that the dots
include gold. The threshold voltages increase almost linearly with the SiO
2
thickness. Furthermore,
the voltage with negative polarity is lower than that with positive polarity. These results provide
evidence that the dot formation on the SiO
2
film using AFM occurs by field evaporation. © 1995
American Institute of Physics.
Considerable attention has been devoted to the scanning
probe microscope ͑SPM͒ as a promising tool for locally
modifying surfaces. This modification technique is expected
to be available for achieving high-density memories and
electronic quantum devices. In particular, modification by
field evaporation using the scanning tunneling microscope
͑STM͒ has already been studied.
1–4
In early work, it was
demonstrated that atomic-scale deposition can be produced
onaGe͑111͒ surface by raising the tip voltage.
5
Subse-
quently, it was experimentally suggested that the electric
field between the tip and the sample is a relevant parameter
for obtaining the deposition and that the mechanism involves
a field evaporation process.
1,2
Field evaporation in the STM is a subject of great physi-
cal importance, but it is difficult to apply it to an insulating
surface. Field evaporation on an insulating surface is also of
great importance from the viewpoint of practical use such as
the fabrication of the metal-oxide-semiconductor ͑MOS͒
structure and recording in air on a disk made of a SiO
2
/Si
substrate. We have investigated field evaporation of gold at-
oms from a gold-coated tip to a thin SiO
2
film/Si substrate
using the atomic force microscope ͑AFM͒. In our previous
studies,
6–8
it was theoretically clarified that field evaporation
in the AFM is possible on thin insulating film. Moreover, the
experiments demonstrated that gold dots 15 nm in diameter
can be formed on SiO
2
film 17 Å thick. In STM studies of
field evaporation of gold atoms between a gold tip and a gold
sample, nearly the same threshold field is observed with both
positive and negative polarity
9
although it is theoretically
predicted that the negative threshold field is distinctly
smaller than the positive one and that gold atoms are most
easily field evaporated as doubly charged negative ions.
10,11
In this letter, in order to determine whether the dot formation
on a SiO
2
/Si substrate using the AFM is caused by field
evaporation of gold atoms from a gold-coated AFM tip, we
report on the elemental analysis of the dots and the depen-
dence of the threshold voltages of both polarities on the
SiO
2
thickness for the dot formation.
Our AFM
7,12
is a conventional one with an optical-beam
deflection system for detecting the atomic force on the tip.
The AFM includes a tripod-type scanner which has a maxi-
mum stroke of about 7
m in three orthogonal directions.
The field evaporation source was a gold-coated AFM tip.
Gold was selected because it is inert and it has a relatively
small threshold field in the field ion microscope ͑FIM͒.
13
The
tip was prepared by vacuum deposition of a gold film with an
average thickness of about 50 nm over a SiO
2
bird-beak-type
cantilever
12
fabricated using the silicon microprocess. The
cantilever has a spring constant of about 0.85 N/m and a
mechanical resonant frequency of about 40 kHz. A pulse
generator was connected to the tip. Pulsed voltages of either
polarity under 100 V for 0.001–10 ms can be applied to the
tip in a predesigned pattern which consists of 1024 pixels
synchronized with the xy address. The sample was a piece of
1–2 ⍀ cm p-type Si wafer with SiO
2
films 17, 34, 75, and
107 Å thick. The 17-Å-thick film was a natural oxide, and
the others were thermal oxides formed by dry oxidation. The
thickness of these films was measured by conventional ellip-
sometry. Both the dot formation and the AFM imaging were
performed in air with a constant repulsive force of the order
of 10
Ϫ9
N. After the dot formation by applying pulsed volt-
ages, elemental analysis of the dots was carried out with a
scanning electron microscopy ͑SEM͒-energy dispersive
x-ray spectroscopy ͑EDX͒ system. The EDX analysis was
performed with accelerating voltage of 20 kV and beam cur-
rent of 10
A.
The dots were formed on SiO
2
films 17–107 Å thick by
adjusting the pulsed voltages. In the AFM images, the dots
appeared as mounds with typical sizes of 100 nm in diameter
and about one tenth of their diameter in height. Figure 1
shows an AFM image of a dot array on a 34-Å-thick SiO
2
film. Each dot was formed by a single 10-ms-pulse of Ϫ30 V
a͒
Electronic mail: koyanagi@crl.hitachi.co.jp
2609Appl. Phys. Lett. 67 (18), 30 October 1995 0003-6951/95/67(18)/2609/3/$6.00 © 1995 American Institute of Physics