International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 2, February 2014)

Dopant Activation in Ion-shower-doped Poly-Si

Jae-Sang Ro
Department of Materials Science and Engineering, Hongik University, Seoul, 121-791, Korea
Abstract— Abnormal behavior of dopant activation was
observed in P +/B+ ion shower doped poly-Si upon post implant
annealing. Phosphorous or boron was implanted by ion
shower doping using a source gas mixture of PH 3/H2 or
B2H6/H2. Activation annealing was conducted using a tube
furnace in the temperature ranges from 350oC to 650oC. Hall
measurement revealed that reverse annealing occurred for
poly-Si implanted with P and B, respectively. It was observed
that reverse annealing starts at 550oC in P+ ion shower doped
poly-Si, while at 350oC in the case of B-doping.
Keywords— Activation, Ion shower doping, LTPS, Reverse
annealing, TFTs
I. INTRODUCTION
Non-mass analyzed ion shower doping (ISD) technique
with a bucket-type ion source has been widely used for
source/drain doping, for LDD (lightly-doped-drain)
formation, and for channel doping in fabrication of low-
temperature polycrystalline silicon (poly-Si) thin-film
transistors [1,2,3]. Phosphorous or boron can be implanted
using ion shower doping with a mixture gas of PH3 or
B2H6, diluted with H2 [4]. Due to non-mass-separation
nature of ISD, hydrogen atoms incorporated into poly-Si
films after ion shower doping may play an important role in
enhanced dopant activation and passivation of defects.
Activation efficiency of dopant atoms in single crystalline
or polycrystalline silicon is therefore higher for ISD than
for mass-separated ion implantation [3]. Greater electrical
activation has been reported in the hydrogeneated samples
at lower temperatures below 600oC [5].
In the case of boron implantation, activation at lower
temperatures is limited by the formation of boron-
interstitial clusters easily formed at concentrations lower
than solid solubility limit. The presence of mobile
hydrogen atoms may assist in the breakup of boron-
interstitial clusters and hydrogen effusion during annealing
may create vacancies. Low energy H+ ion implantation
following As implantation with low temperature annealing
was reported to be effective for damage passivation and
dopant activation [6]. Meanwhile, reverse annealing is well
known phenomenon for B implantation into single
crystalline silicon in the temperature ranges between 500oC
and 600oC. We observed reverse annealing in P+/B+ ion
shower doped poly-Si and report here on its behavior.
II. EXPERIMENTAL PROCEDURE
For the sample preparation SiO2 insulation layer with a
thickness of 3000 Å was formed on a Corning 7059 glass
substrate of 370 mm x 470 mm x 0.7 mm (length x width x
thickness) by means of plasma enhanced chemical vapor
deposition (PECVD). An a-Si thin film with a thickness of
500 Å was formed successively upon the insulation layer
using PECVD. The substrates used were poly-Si produced
by excimer laser crystallization on 500 Å-thick PECVD a-
Si. The glass substrate was broken into pieces of 20 mm x
20 mm, thereby preparing a test piece. Phosphorous or
boron was implanted by ion shower doping with a main ion
source of P2 Hx or B2Hx using a source gas mixture of
PH3/H2 or B2H6/H2 as shown in Fig. 1. Activation
annealing was performed in the temperature ranges from
350oC to 650oC for 30 min using a tube furnace in nitrogen
ambient. Carrier concentration by dopant activation was
determined by Hall measurement using a van der Paw
method. A 4-point-probe was used to check the value of
sheet resistance determined by Hall measurements. The as-
implanted damage induced by ion shower doping and
damage recovery following an activation annealing was
assessed with transmission electron microscopy (TEM).
Fig. 1 Schematic diagram of ion-shower-doping system used in this
work.
III. RESULTS AND DISCUSSION
Acceleration voltage was changed from 3 kV to 15 kV
for P doping and from 2 kV to 6 kV for B doping. Doping
time was kept constant as 1 min for all samples used in this
study. Figure 2 shows sheet resistance as a function of
acceleration voltage after activation annealing.

12
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 2, February 2014)
30 min-isochronal annealing was conducted in the
temperature ranges from 350oC to 650oC. As shown in Fig. 100000 o
400 C
2-(a), for P-doped samples annealed at temperatures greater 80000 o
450 C
than 600oC the sheet resistance gradually decreases as 60000
40000
o
500 C
o
550 C
acceleration voltage increases, while it initially decreases o
600 C
with acceleration voltage and increases again with o
650 C

Rs (/sq)
8000
acceleration voltage beyond a certain value for the samples
annealed at temperatures lower than 500oC. A 6000

concentration profile of implanted atoms becomes broader 4000

and the peak concentration at the projected range gets
2000
lower as the acceleration voltage increases. An excess
concentration of dopant atoms beyond solid solubility at a 0
2 4 6 8 10 12 14 16
given annealing temperature thus becomes large as the Acceleration Voltage (kV)
acceleration voltage decreases. Activation efficiency is
therefore higher as the acceleration voltage increases. This
(a)
may explain the decrease of sheet resistance with
acceleration voltage.
Meanwhile, for the samples implanted with higher 12000
o
400 C
o
acceleration voltage followed by low temperature 450 C
o
500 C
annealing as-implanted damage by high mass ion such as P 10000 o
550 C
o
may not be sufficiently cured. This may explain the 8000
600 C
o
650 C
increase of sheet resistance with acceleration voltage.
Rs (/sq)

Figure 3 shows transmission electron microscopy (TEM) 6000

for the samples implanted with acceleration voltage of 15 4000

kV for 1 min. Figure 3-(a) shows TEM micrographs for as-
implanted sample, Fig. 3-(b) shows the one annealed at 2000

550oC for 30 min, and Fig. 3-(c) shows the one annealed at
0
600oC for 30 min. TEM results indicate that implanted 2 3 4 5 6

damage is not recovered sufficiently even for the sample Acceleration Voltage (kV)

annealed at 550oC.
In contrast to the case of P-doping the sheet resistance (b)
decreases as the acceleration voltage increases in B-doped Fig. 2 Sheet resistance vs. acceleration voltage for the samples doped
sample except for the samples annealed at 650oC. Since with (a) phosphorous and (b) boron, respectively.
boron is a light mass ion damage may not be generated to a
significant amount. If we look at Fig. 2-(b) carefully the
values of sheet resistance have an abnormal relationship
with annealing temperatures. In the case of 6 kV-
implanted samples the sheet resistance decreases as the
annealing temperature increases.

(a)

13
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 2, February 2014)
(b)
(c)
Fig. 3 TEM micrographs for (a) the sample implanted with 15-kV P+,
followed by (b) 550℃-30min annealing, and (c) 600℃-30min
annealing.
Figure 4 shows sheet resistance as a function of
annealing temperature for P-doped samples with
acceleration voltage from 3 kV to 15 kV, and for B-doped
samples with acceleration voltage from 2 kV to 6 kV,
respectively. The sheet resistance decreases as annealing
temperature increases as indicated in Fig. 4-(a) for P-doped
samples. Large increase in sheet resistance for 15 kV-
implanted samples annealed at temperatures blow 500oC
can be explained by the effect of uncured damage as
already indicated in Fig. 3.
The sheet resistance, however, has an unusual
relationship with annealing temperature for B-doped
samples as shown in Fig. 4-(b). Figure 5 shows sheet
carrier concentration determined by Hall measurement
using a van der Paw method as a function of annealing
temperature for the samples implanted with P and B,
respectively. Carrier concentration increases as acceleration
voltage increases at a given annealing temperature as
shown in Fig. 5.
14
Carrier concentration, however, increases as annealing
temperature increases up to 550oC, while it decreases as
annealing temperature increases beyond 550oC, for P-
doped samples as indicated in Fig. 5-(a). It can thus be
said that reverse annealing starts at 550oC in P+ ion
shower doped poly-Si.
Reverse annealing is well known phenomenon for B-
doped single crystalline silicon using a mass separated ion
implanter. Seidel et al. reported that reverse annealing is
observed in single crystalline silicon implanted with boron
in the temperature ranges between 500oC and 600oC [7]. B-
doped samples were observed to exhibit reverse annealing
behavior as indicated in Fig. 5-(b). Reverse annealing,
however, begins at lower temperatures in B-doped poly-Si
samples compared to B-doped single crystalline silicon.
Reverse annealing already starts from 350oC in B+ ion
shower doped poly-Si. Unlike semiconductor processing, in
which a bulk single crystalline silicon is used, LTPS-TFTs
make use of a very thin poly-Si, which serves as an active
layer on the top of a SiO2 dielectric layer. Thus the
structure of LTPS-TFTs is similar to silicon-on-insulator in
that they both include a silicon/buffer oxide interface [8].
As the interface between the poly-Si and SiO2 buffer layer
on top of the glass substrate acts like a diffusion barrier for
the Si self-interstitials, the self-interstitials generated by ion
implantation tend to be supersaturated in a thin poly-Si
layer. Such factors are bound to significantly affect the
activation behavior of ion-implanted poly-Si. This may
explain the reason why we observed reverse annealing at
significantly lower temperatures in B-doped poly-Si
samples. Reverse annealing seems to end at 600oC for the
samples implanted with lower acceleration voltage.
100000
3kV
80000
5kV
60000
40000
10kV
15kV
Rs (/sq)
8000
6000
4000
2000
0
400 450 500 550 600 650
Temperature ( C)
o
(a)
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 2, February 2014)

21000
2.0kV 3kV
3.0kV 1E15 5kV
18000
4.5kV 10kV

Carrier Concentration (cm )

-2
6.0kV 15kV
15000
Rs (/sq)

12000

9000

6000
1E14

3000

0
350 400 450 500 550 600 650
o 400 450 500 550 600 650
Temperature ( C)
o
Temperature ( C)

(b)
(a)
Fig. 4 Sheet resistance as a function of annealing temperature (a)
Phosphorous doping with acceleration voltages from 3 kV to 15 kV
(b) Boron doping with acceleration voltages from 2 kV to 6 kV.
2.0kV
As mentioned earlier reverse annealing is related to Si Carrier Concentration (cm ) 3.0kV
-2

4.5kV
self-interstitials released from the ion induced damage- 6.0kV

region [9-11]. As primary damage such as silicon self- 1E14

interstitials generated by ion implantation is
thermodynamically unstable, they may agglomerate
themselves to form extended defects, or, they may migrate
and bind with boron atoms to form boron-interstitial
clusters. Once boron-interstitial clusters are produced,
1E13
boron atoms become electrically inactive, thus resulting in
the loss of charge carriers. As the diffusion of both boron 350 400 450 500 550 600 650
atoms and silicon self-interstitials is required, the kinetics Temperature ( C)
o

associated with the formation of boron-interstitial clusters

would be a slow process. Based on this argument, damage
recovery would be required in order to release self-
interstitials for formation of boron-interstitial clusters
resulting in reverse annealing. This would explain why
activation efficiency becomes lower as the annealing
temperature increase as demonstrated in Fig. 5-(b).
(b)
Fig. 5 Sheet carrier concentration as a function of annealing
temperature determined by Hall measurement. (a) Phosphorous
doping with acceleration voltages from 3 kV to 15 kV (b) Boron
doping with acceleration voltages from 2 kV to 6 kV.

15
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 4, Issue 2, February 2014)
IV. CONCLUSIONS
The sheet resistance decreases as acceleration voltage
increases since an excess concentration of dopant atoms
beyond solid solubility at a given annealing temperature
becomes small as the acceleration voltage increases with
some exception for the P doped samples implanted with
acceleration voltages exceeding a critical value. In this
work reverse annealing was observed in P+/B+ ion shower
doped poly-Si upon activation annealing. It was found that
reverse annealing starts at 550℃ in P+ ion shower doped
poly-Si, while at 350℃ in the case of B-doping. Compared
to B-doped single crystalline silicon reverse annealing,
however, begins at lower temperatures in B-doped poly-Si
samples due to supersaturation of the self-interstitials since
the interface between the poly-Si and SiO2 buffer layer on
top of the glass substrate acts like a diffusion barrier for the
self-interstitials.
Acknowledgment
This work was partially supported by the IT R&D
program of MOTIE/KEIT. [10043413, Development of
Heat Treatment Processes for Fabrication of Next-
Generation Flat Panel Displays and Semiconductor Devices
Using Rapid Heating Methods] and by EnSilTech
Corporation.
16
REFERENCES
[1] Yoshida, A., Setsune, K., and Hirao, T. 1987 Phosphorus doping for
hydrogenated amorphous silicon films by a low-energy ion doping
technique. Appl. Phys. Lett. 51, 253-255.
[2] Kawachi, G., Aoyama, T., Miyato, K., Ohno, Y., Mimura, A., and
Mochizuki, Y. 1990 Large-Area Ion Doping Technique with Bucket-
Type Ion Source for Polycrystalline Silicon Films. J. Electrochem.
Soc. 137, 3522-3526.
[3] Mishima, Y., and Takei, M. 1994 Non-mass-separated ion shower
doping of polycrystalline silicon. J. Appl. Phys. 75, 4933-4938.
[4] Kim, D.-M., Kim, D.-S., Ro, J.-S. 2005 Dopant activation after ion
shower doping for the fabrication of low-temperature poly-Si TFTs.
Thin Solid Films 475, 342– 347
[5] Vengurleaker, A., Ashok, S., Kanlas, C. E., and Theodre, N. D. 2004
Hydrogen plasma enhancement of boron activation in shallow
junctions. Appl. Phys. Lett. 85, 4052-4054.
[6] Singh, R., Fonash S. J., Rohatgi, A., Chodhury, P. R., and Gigante, J.
1984 A low-temperature process for annealing extremely shallow
As+ -implanted n+/p junctions in silicon. J. Appl, Phys. 55, 867-870.
[7] Seidel, T. E., and MacRae. 1971. In Eisen, F, and Chadderton, L,
Eds., First Intl. Conf. on Ion Implantation, Thousand Oaks, Gordon
and Breach, New York
[8] Saavedra, A. F., Jones, K. J., Law. M. E., Chan, K. K. and Jones, E.
C. 2004 Electrical activation in silicon-on-insulator after low energy
boron implantation. J. Appl. Phys. 96, 1891-1898
[9] Mirabella, S., Bruno, S. E. Priolo, F., De Salvador, D., Napolitani, E.
and Carnera, A. 2003 Dissolution kinetics of boron-interstitial
clusters in silicon. Appl. Phys. Lett. 83 680-682
[10] Solmi, S., Mancini, L., Milita, S. Servidori, M., Mannino, G. and
Bersani, M. 2001 Effects of boron-interstitial silicon clusters on
interstitial supersaturation during postimplantation annealing Appl.
Phys. Lett. 79 1103-1105
[11] Mannino, G., Cowern, N. E. B., Roozeboom, F. and van Berkunm, J.
G. M. 2000 Role of self- and boron-interstitial clusters in transient
enhanced diffusion in silicon J. Appl. Phys. Lett. 76 855-857