909 FTP
909 FTP
909 FTP
COMMUNICATION
DOI: 10.1002/adma.200700646
fine fabrication of oxide films with nanoscale control still remains remarkably challenging. At this moment, the lower limit for the fine fabrication of oxides using photolithography is
only ca. 1 lm, notwithstanding the remarkable advances in
processing silicon surfaces with better than 100 nm resolution.
Recently, there have been several reports in the literature
dealing with the nanofabrication of transition metal oxides
using top-down approaches such as focused ion beam (FIB) lithography,[57] atomic force microscopy (AFM) lithography,[418] and electron-beam (e-beam) lithography.[19,20] These
methods have been used to achieve feature sizes less than ca.
100 nm. However, each of these top-down methods suffers
from some serious drawback. For example, FIB and AFM lithography are suitable for making holes and grooves in selected regions within an oxide film, but are not as useful for
defining isolated oxide islands and lines, because such patterns need to be defined by removing all sections of the film
that reside outside the pattern. However, for the application
of oxide nanopatterns to integrated memories and sensors, island and dot morphologies tend to be preferred over holes
and grooves. Although e-beam lithography is a feasible alternative for defining not only holes and grooves but also islands
and lines, it is generally necessary to perform some kind of
wet or dry etching after the patterning of the resist. In general,
it can be incredibly challenging to process oxides by most conventional wet or dry etching approaches because these materials are usually very hard and chemically stable. As a result,
the etching conditions depend strongly on the materials used.
Notably, it has been almost impossible to process some oxide
films (e.g., Pb(Zr,Ti)O3, SrTiO3) by wet etching.
The fabrication of oxide nanostructures by lift-off techniques, essentially involving some combination of top-down
and bottom-up approaches (i.e., preparation of an area-selective film by deposition using a patterned mask, followed by removal of the masked area), presents some intriguing possibilities by obviating the need to etch the oxide film. For example,
Donthu et al.[21,22] have recently developed a scheme to fabricate oxide nanostructures using soft e-beam lithography combined with a lift-off technique. These authors have patterned a
polymer e-beam resist by e-beam lithography and cast a solution of sol-gel precursors to the oxides onto the pattern by spin
coating. The sol-gel solution has then been dried by heat treatment at low temperature (less than 150 C). The patterned oxide has been obtained by removing the organic resist with acetone. The residual pattern has been annealed at 700 C in order
to crystallize the film. Donthu et al. have successfully prepared
ca. 40 nm oxide line patterns by this approach. However, the
909
COMMUNICATION
910
www.advmat.de
COMMUNICATION
AFM t i p
(b)
3
Mo
O
Mo
ALO
ALO
10
2
3
4
Distance/ m
ALO
Height / nm
Height / nm
Height / nm
20
0
0
(c)
FM
O
(a)
-10
0
2
3
4
Distance/ m
10
0
0
2
3
4
Distance/ m
Figure 1. Mo lift-off process used for the fabrication of oxide nanostructures. The top panel schematically illustrates the steps involved in the process.
The middle panel shows topographic images of line patterns fabricated by this process. The bottom panels depict the cross sections of the line patterns shown in the middle panels along the yellow lines.
Figure 2. 3D AFM images of a FMO pattern fabricated by the Mo lift-off process. The top panel
shows an enlarged AFM image of the demarcated area in the bottom image.
www.advmat.de
911
gap
COMMUNICATION
10m
FMO lines
Pt electrode
ALO substrate
Pt electrode
Figure 4. Optical microscopy image of FMO stripes between Pt electrodes fabricated by the Mo lift-off process. The locations of the FMO
lines are indicated by arrows.
Figure 3. a) Scanning electron microscopy (SEM) image of a FMO pattern fabricated by the Mo lift-off process; b) SEM image of the same pattern obtained while performing electron probe microanalysis (EPMA)
measurements. Spatial distribution of c) Mn and d) Fe derived from
EPMA measurements.
resistivity/ cm
10 1
10 0
150
912
www.advmat.de
200
250
temperature/K
300
Experimental
The FMO patterns were fabricated as follows: first, a ca. 20 nm Mo
layer was deposited onto an ALO substrate. Next, a MoO3 pattern
was formed by the local anodic oxidation of Mo induced by a voltagebiased AFM tip. The MoO3 regions were then removed by immersion
in water for 3 min. After slight dry etching with CF4+ to eliminate the
residual Mo layer, a FMO film with a thickness of ca. 10 nm was deposited onto the pattern at an elevated temperature and then cooled
down to room temperature. The Mo layer was subsequently removed
by ultrasonic treatment with 10 vol% H2O2 solution for 5 min.
ALO substrates were supplied by Furuuchi Chemicals and used
after ultrasonic cleaning in ethanol without any further treatment.
The Mo layer was deposited by magnetron sputtering using a Quick
Coater SC-701HMC system (Sanyu Electron). During deposition, argon
was used as the sputtering gas. The purity of the Mo target was 99.9%.
AFM lithography was performed with a Dimension 3100 AFM microscope (Veeco) using Olympus OMCL-AC240TM AFM tips with a
resonant frequency of ca. 70 kHz and a spring constant of 2 N m-1. In
order to achieve good uniformity of the pattern, the lithography process was performed with a constant current.[14] Control over the current was obtained by connecting the AFM tip and the sample to an
external setup comprising a picoammeter and a feedback circuit.
The FMO thin film was deposited by PLD (arc excimer laser;
k = 193 nm). The target was prepared by sintering a FMO pellet
synthesized by a conventional solid reaction process. The PO2 was set at
1.0104 Pa, and the temperature of the substrate was kept at 340 C.
In order to measure the spatial atomic distribution, EPMA measurements were performed using an electron probe microanalyzer JXA8800R (JEOL). For this measurement, ca. 50 nm thick FMO was deposited on a ca. 70 nm thick Mo mask to improve the sensitivity of EPMA.
The resistivity of FMO stripes fabricated as described above was
also measured. FMO stripes were prepared between a ca. 10 lm gap
between Pt electrodes on an ALO substrate. The Pt electrodes were
fabricated by sputtering Pt onto a substrate wound with a 10 lmU
tungsten wire and removing the wire after completion of the sputtering process. In order to obtain measurable conductivity, 100 FMO
lines were fabricated between the Pt gap electrodes. The resistivity of
the system was measured using a Physical Property Measurement System Model 6000 (Quantum Design).
Received: March 18, 2007
Revised: November 13, 2007
Published online: January 30, 2008
[1]
[2]
www.advmat.de
COMMUNICATION
913