Localized fabrication
implantation
of Si nanostructures
by focused
ion beam
A. J. Steckl, H. C. Mogul, and S. Mogren
University of Cincinnati/AJanoelectronics
Laboratory
Cincinnati, Ohio 45221-0030
(Received 7 November 1991; acceptedfor publication 11 February 1992)
Si nanostructures have been fabricated by focused ion beam implantation (FIB) followed by
etching in KOH/IPA. The FIB implantation into Si at a sufficiently high dose ($1015/cm2)
renders the local Si region much less susceptible to chemical etching. This effect has been
observed for FIB implantation with Ga, Au, and Si ions. After etching, the implanted layer
forms a cantilever structure whose thickness is a function of the implantation energy. At low
energies ( < 30 keV) nanometer-scaleSi structures can be formed using this technique.
Recent reports of visible photoluminescencefrom micro-crystalline Si structures’ and porous Si2 have greatly
stimulated both experimental and theoretical investigations
of quantum confinement effects in Si. In the caseof porous
Si, the specific structure is difficult to measure and to reproduce. A major challengelies in the fabrication of regular arrays of quantum wires in Si with controlled size and
distribution, which would considerably enhancethe understanding of the physics of emission from these structures.
In this letter we present an approach to meeting this
challenge through the use of selective focused ion beam
(FIB) implantation to locally control the size and spacing
of these nanostructures. FIB implantation of heavy group
HI (cf. Ga, In) has been shown3to result in shallow layers
and p-n junctions in Si with excellent electrical properties
upon rapid thermal anneal at low temperatures (e.g., 30 s
at 600 “C). Low energy Ga + FIB implantation has
resulted4in the fabrication of p-n junctions with less than
20 nm depth.
High dose Ga + ( > 1X 10”/cm2) implantation into Si
has been reported’ to render the implanted region relatively immune to subsequent certain chemical etchants,
such as KOH. Furthermore, mixtures of potassium
hydroxide/isopropyl alcohol (KOH/IPA) are known”8 to
be anisotropic etches with a low etch rate along the Si
{ill) planes.
The starting substrates were (100) n-Si wafers, Pdoped with a background concentration of 2~ 1015/cm3.
The FIB implants were performed perpendicular to the Si
surface using a NanoFab 150 system.4The patterns implanted consistedof arrays of squaresranging from 4 x 4 to
1 X 1 ,um”, with a nominal center-to-center spacing of 6
pm. The Ga implants were performed primarily at 30 keV,
using a 69Ga beam current of 12 pA. For the cases of
197A~+ and 28Si+ +, implants were performed at energies
of 30 and 60 keV with beam currents of 7.8 and 2.4 pA,
respectively.
Anisotropic etching was performed on the as-implanted samples at 85 “C using KOH, IPA, and water in
the ratio of 5:4:16 by weight, respectively. A typical example of Si nanostructures obtained with a 30 keV Ga FIB
implantation at a dose of 1 X 1016/cm2after a 5.5 min etch
is shown in Fig. 1. The figure shows several4 x 4 and 2 X 2
1833
Appl. Phys. Lett. 60 (15), 13 April 1992
,um2 implanted regions, at the top of truncated pyramids
formed primarily by inclined { 111) planes. Noticeable in
Fig. 1(b) is the semitransparencyof the implanted region
cantilever to the 25 kV electron beam of the scanning electron microscope @EM). The undercut of the implanted
region is measuredto be approximately 1 ,um.
The evolution of the Si nanostructure with etch time is
shown in Fig. 2 for the 2 X2 pm’ features. The lateral
progressionof the slowly etching { 111) planesresults in an
increasing undercut, of dimension (L- W)/2, of the implanted region, of dimension L. This is illustrated in the
SEM microphotographs (Fig. 2 inset) which were obtained from structures etched for 1.5 and 3 min, respectively. As seen by comparing the 1 X 1 pm2 structures re-
FIG. 1. SEM microphotographs of a 4X4 pm* nanostructure fabricated
by 30 keV/l x 10’6/cm2 FIB Ga+ implantation followed by 1 min etch:
(a) 10 kx; (b) 50 kx.
0003-6951/92/151833-03$03.00
@ 1992 American Institute of Physics
1833
Downloaded 07 Oct 2003 to 129.137.210.120. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
YE+
2
!z
2
2
2
ifi
‘5
5
1
.J
0
0
1
2
3
4
5
6
Etch Time (min)
FIG. 2. Lateral dimensions as a function of etching time for Si nanostructures implanted with 30 keV Ga+ at a dose of 1 X 10’6/cmZ.
sulting from a 3 min etch (Fig. 3)) the erosion of the edge
of the implanted region appears enhancedin the 20 kV
implanted structure over that occurring in the 30 kV structures.
At the dosesused the FIB-implanted region is amorphized, requiring recrystallization by heat treatment. In
addition, by activating the Ga implanted ions, ap-n junction is formed between the nanostructure and the substrate, which can provide quantum confinement. Rapid
thermal annealing was performed at 600 “C/30 s in N2
ambient. These conditions have been previously used’to
FIG. 4. SEM microphotographs of a 2X2 pm* nanostructure fabricated
with 30 keV/l X 10i6/cm2 FIB Ga+ implantation following a 5.5 min
etch: (a) as-implanted; (b) after 600 “C/30 s N, anneal.
FIG. 3. SEM microphotographs of 1 X 1 pm’ nanostructure after 3 min
etch fabricated by 1X 10Lb/cmz FIB Gaf implantation at: (a) 30 (b) 20
keV.
produce high quality ultrashallow p-n junctions by low
energy Ga FIB implantation. The effect of rapid thermal
annealing (RTA) is illustrated in Fig. 4 where SEM microphotographs of a 2 X2 pm” feature before and after
anneal are shown. As can be seen,the post-etching anneal
does not result in any significant morphological or dimensional changes.Raman spectroscopyon the annealedsample showed indistinguishable signals from the implanted
and unimplanted regions, indicating that the implanted
section has beenepitaxially regrown. Interestingly, the preetching RTA of the FIB-implanted samplesresulted in the
formation of similar structures.
We have also briefly investigatedthe effect of implanting other speciesinto Si substrates, namely 139A~+ and
28Si+ + . A similar selective etching effect was observed
after FIB implantation with a dose of 3X 1015/cm2and
etch time of 2.5 min, as illustrated in Figs. 5(a) and 5(b)
for Au and Si.
Direct measurementof the nanostructure thickness,
tcap,is difficult to accomplish.To estimateits value we have
used TRIM-generated”implant profiles, in conjunction
with the observationof Berry and Caviglia’that the critical
Ga+ peak concentration for an effective etch-stop is
- 1 X 1021/cm3at energiesranging from 10 to 50 keV. tcaP
is then approximatedas the thickness of the region with Ga
concentration in excess of 1X 102i/cm3. At a dose of
1 X 10r6/cm2,fcapvalues of -20 and 30 nm are calculated
at 20 and 30 keV, respectively.For an implantation energy
1 a34
Appl. Phys. Lett., Vol. 60, No. 15, 13 April 1992
Steckl, Mogren, and Mogul
1834
Downloaded 07 Oct 2003 to 129.137.210.120. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp
of 2 keV, we calculate tap values of 4 and 10 nm for doses
of 1X 10” and 1 X 1016/cm2,respectively.
In our experiments, the etch-stop characteristic appears to be insensitive to the chemical nature of the implanted species.Hence, we tentatively conclude that the
mechanism at work is causedby the implantation-induced
strain. This conclusion is similar to that reachedby Palik et
al” for B- and P-diffused Si anodically etched in a
KOH/H20 system. They report a sharp drop in etch rate
at doping levels greater than w 1x 10i9/cm3 for ( 100) Si
doped with either impurity.
The authors acknowledge many useful discussions
with I. Berry and J. Zavada.
FIG. 5. SEM microphotographs of 1 X 1 pm’ nanostruc tures fabricated
by FIB implantation with a dose of 3 X 10’5/cm2 and a 2. 5 mm etch: (a)
60keVSi++;
(b) 30keVAuf.
‘S. Furukawa and T. Miyasato, Jpn. J. Appl. Phys. 27, L2207 (1988).
‘L. T. Canham, Appl. Phys. Lett. 57, 1046 (1990).
‘C. M. Lin, A. J. Steckl, and T. P. Chow, J. Vat. Sci. Technol. B 6, 977
(1988).
‘A. J. Steckl, H. C. Mogul, and S. Mogren, J. Vat. Sci. Technol. B 8,
1937 (1990).
‘1. L. Berry and A. L. Caviglia, J. Vat. Sci. Technol, B 1, 1059 ( 1983).
“K. E. Bean, IEEE Trans. Electron Devices ED-25, 1185 ( 1978).
‘E. Bassous, IEEE Trans. Electron Devices ED-25, 1178 ( 1978).
s K. E. Petersen, Proc. IEEE 70, 420 (1982).
‘A. J. Steckl, H. C. Mogul, and S. Mogren, J. Vat. Sci. Technol. B 9,
2718 (1991).
“J. F. Ziegler, J. P. Biersack, and U. Littmark, Stopping and Rnnge of
Ions in Matter (Pergamon, New York, 1988). Vol. I.
“E . D . Pal&, J. W. Faust, H. F. Gray, and R. F. Greene, J. Electrochem.
Soc.129,2051
(1982).
1835
Steckl, Mogren, and Mogul
1835
Appl. Phys. Lett., Vol. 60, No. 15, 13 April 1992
Downloaded 07 Oct 2003 to 129.137.210.120. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/aplo/aplcr.jsp