Journal of The Electrochemical Society, 150 ͑7͒ G371-G375 ͑2003͒
G375
undoped film growth. This means that the reaction temperature of
hydrogen desorption from Si-H is shifted lower for boron-doped
films than that for undoped films in an opposite manner for the
phosphorus-doped film, as predicted by the result of temperature
programmed desorption ͑TPD͒ of hydrogen from each doped
layer.10 In our study, when B H was supplied along with Si H ,
growth, when the temperature was higher than 700°C. As a result,
the PDB structure for I-SIT was achieved successfully by Si MLE
with doping at 510°C.
Conclusion
2
6
2
6
We achieved Si molecular layer epitaxy using our doping method
at the process temperature of 510°C by intermittent injection of
Si H ͑99.99%͒ and dopant precursor, PH3 or B H , into the
doping efficiency was low; however, when B H was supplied al-
2
6
ternately with Si H , the efficiency increased by about one order of
2
6
2
6
2
6
magnitude. These facts clearly show that the doping method used in
MLE is advantageous for impurity doping technology.
Si͑100͒ surface. At the process temperature of 525°C, an n-type
layer doped with phosphorus was obtained up to the carrier concen-
1
9
Ϫ3
As an application, device structures of nanometer scale were
fabricated by Si MLE using the doping method. Figure 7 shows the
tration of 2.8 ϫ 10 cm with three-dimensional growth, however
a smooth surface by two-dimensional growth was obtained when the
ϩ
19
Ϫ3
depth profiles of boron in a planar doped structure, ͑a͒ i-p -i and ͑b͒
doping level was under 2 ϫ 10 cm . The higher concentration
of phosphorus seems to prevent not only the surface migration for
the two-dimensional growth, but also the surface reaction rate of
Si-H species for Si growth.
ϩ
ϩ
ϩ
n -i-p -i-n , fabricated by Si MLE with doping at 510°C. The i
layer indicates an intrinsic undoped epitaxial layer by SiMLE. The
measurement was performed using TOF-SIMS. The parameters for
ϩ
the Ar primary ion beam for TOF analysis were an energy of 10
The carrier concentration of the p-type layer doped with boron
2
0
Ϫ3
keV and current of 0.85 pA with 10 ϫ 10 m raster scan, flooding
reached almost 5 ϫ 10 cm
while maintaining an atomically
Ϫ4
O at 1 ϫ 10 Pa to emphasize the detection sensitivity of boron
smooth surface by two-dimensional growth at 510°C. As an appli-
cation, nanometer-scale devices with n-i-p-i-n structure, which is the
source-drain structure of an I-SIT were fabricated by Si MLE using
the doping method. The doping profile measured by TOF-SIMS in-
dicated a sufficiently sharp distribution on the nanometer scale.
2
ϩ
in Si. The parameters for sputtering were Ar , 1 keV, 7.5 nA, and
100 ϫ 100 m raster scan with sputtering rate of 0.71 nm/min for
Si. The result in Fig. 7a shows that the designed multilayer structure
ϩ
was nearly achieved by Si MLE, and the boron-doped p layer
designed to be 5 nm thick at the carrier concentration of 1.1
ϫ 1020 cm was measured to be 6.8 nm ͓full width at half maxi-
mum ͑fwhm͔͒.
Semiconductor Research Institute assisted in meeting the publication
costs of this article.
Ϫ3
References
Figure 7b shows the planar doped barrier ͑PDB͒ structure
n -i-p -i-n , which is the basic building block, namely, the source-
ϩ
ϩ
ϩ
1. J. Nishizawa, H. Abe, and T. Kurabayashi, J. Electrochem. Soc., 132, 1197 ͑1985͒.
2
3
4
5
6
7
. J. Nishizawa, K. Aoki, S. Suzuki, and K. Kikuchi, J. Electrochem. Soc., 137, 1898
1990͒.
. S. Imai, T. Iizuka, O. Sugiura, and M. Matsumura, Thin Solid Films, 225, 168
͑1993͒.
drain, of an ideal static induction transistor ͑I-SIT͒ as fabricated in
͑
1
1,12
the GaAs process.
The channel length of the I-SIT is defined as
ϩ
the length of i-p -i in the structure. Therefore, the PDB structure in
Fig. 7b shows the channel length of 21.2 nm, which is one of the
. D. Lubben, R. Tsu, T. R. Bramblett, and J. E. Greene, J. Vac. Sci. Technol. A, 9,
3
003 ͑1991͒.
smallest transistor structures recorded. In this structure,
a
. S. Imai, S. Takagi, O. Sugiura, and M. Matumura, Jpn. J. Appl. Phys., Part 1, 30,
3646 ͑1991͒.
. K. Ohtsuka, A. Murai, T. Oizumi, T. Yoshida, T. Kurabayashi, K. Suto, and J.
Nishizawa, J. Vac. Sci. Technol. A, 18, 48 ͑2000͒.
. J. Nishizawa, A. Murai, T. Ohizumi, T. Kurabayashi, K. Ohtsuka, and T. Yoshida, J.
Cryst. Growth, 209, 327 ͑2000͒.
ϩ
phosphorus-doped n layer at the carrier concentration of 1
ϫ 1019 cm was inserted at the top and the bottom of the i-p -i
structure. The growth temperature in each layer of the PDB structure
was chosen as 510°C to reduce the deterioration of the doping pro-
file at a higher growth temperature. The result in Fig. 7b shows that
the design multilayer structure was nearly achieved by Si MLE, and
Ϫ3
ϩ
8. M. Suemitsu, Y. Tsukidate, H. Nakazawa, and Y. Enta, J. Vac. Sci. Technol. A, 16,
772 ͑1998͒.
1
ϩ
9. I. Goldfarb, J. H. G. Owen, D. R. Boeler, C. M. Goringe, P. T. Hayden, K. Miki, D.
G. Pettifor, and G. A. D. Briggs, J. Vac. Sci. Technol. A, 16, 1938 ͑1998͒.
10. B. Doris, J. Fretwell, J. L. Erskine, and S. K. Banerjee, Appl. Phys. Lett., 70, 2819
͑1997͒.
the boron-doped p layer designed to be 1.2 nm thick at the carrier
concentration of 1.1 ϫ 1020 cmϪ3 was measured to be 5.7 nm
͑
fwhm͒. In this structure, the most probable diffusion source is bo-
ϩ
ϩ
11. P. Plotka, T. Kurabayashi, Y. Oyama, and J. Nishizawa, Appl. Surf. Sci., 82-83, 91
ron from the p layer by the electric field toward the n region
during a high-temperature growth process. By process simulation,
the fatal diffusion of boron occurred in the entire i layer during
͑
1994͒.
1
2. Y. X. Liu, P. Plotka, K. Suto, Y. Oyama, and J. Nishizawa, IEE Proc.-G: Circuits,
Devices Syst., 146, 31 ͑1999͒.