251904-3
Xu et al.
Appl. Phys. Lett. 89, 251904 ͑2006͒
TABLE I. Surface composition of 150 nm SiNx on TEOS SiOx buffer ana-
lyzed by XPS.
the annealing temperature Tanneal. For the as-grown sample,
there is insufficient O content to provide enough luminescent
centers ͑Si oxide layers͒, and hence the PL is weak. When
Tanneal ͑°C͒
͓Si͔ ͑%͒
͓O͔ ͑%͒
͓N͔ ͑%͒
As-grown
650
800
1200
T
annealϾ650 °C, the O content increases with increasing
Tanneal, leading to an increment of the number of the lumi-
nescent centers. Meanwhile, the Si content providing with
photoexcited carriers keeps almost constant. As a result, the
PL intensity increases during this process. It should be noted
that the N content decreases monotonously as Tanneal in-
creases. As Tanneal is further increased above 800 °C, the O
content keeps increasing while the N content decreases
sharply. In this case, N cannot effectively passivate the
Si/SiOx interface. This means that the probability of photo-
excited carriers transferred to the PL emission center are sub-
stantially decreased. As such, the PL intensity starts to de-
crease when TannealϾ800 °C.
53.67
14.00
32.33
53.90
20.21
25.89
54.44
24.38
21.18
51.12
31.60
17.28
are clearly seen. A quantitative composition analysis ͑Table
I͒ shows that below 800 °C the concentration of Si atom
slightly increases with increasing Tanneal. Above this tempera-
ture, this trend is reversed. On the other hand, the concentra-
tion of N decreases monotonously with increasing Tanneal
,
whereas the opposite is true for the concentration of O.
In order to explain the origin of the PL of the SiNx/SiO2
film, a pure Si film is deposited with the same thickness as
that of SiNx film by sputtering a pure Si target with Ar
plasma at a flow rate of 35.2 SCCM. The PL intensity of
Si/SiO2 films ͑see Fig. 1͒ is found to be even weaker than
that in TEOS SiO2 films. Interestingly, the PL spectra of
Si/SiO2 films exhibit a similar shape to that in SiNx/SiO2
films under high-temperature annealing ͑not shown͒. This
fact indicates that after annealing the PL of SiNx/SiO2 films
is most likely originated from the Si oxide layer which ap-
pears to be formed due to the updiffusing of O atoms in the
derstand if there indeed exist PL centers in the SiNx particles,
a pure SiNx film is deposited directly onto the Si substrate
and subsequently annealed under the same conditions as
mentioned above. However, no obvious PL peak in the pure
SiNx film ͑whether as-grown or annealed at a temperature
above 650 °C͒ has been observed. This further excludes the
possibility of forming Si oxide layer with luminescent cen-
ters from SiNx surface. Another comparison with the PL of
TEOS SiO2 films is highlighted in Fig. 1. A significant in-
crease of the PL intensity in the SiNx/SiO2 films is observed
in comparison to Si/SiO2 films or TEOS SiO2 films. As
shown in Table I, there indeed exists rich Si in all samples.
Therefore, the PL mechanism of the annealed samples may
be understood as follows: the photoexcited carriers are first
generated in the rich Si in the SiNx film. These carriers pass
through the surrounding amorphous SiNx layer and move
into the Si oxide to emit light. In fact, Chen et al.21 have
reported that the defect passivation in multicrystalline Si
could be realized by depositing SiO2/SiN coatings, which
significantly increased the bulk lifetime and decreased the
surface recombination velocity. Therefore, the PL intensity is
remarkably enhanced in the SiNx/SiO2 film in comparison to
the other films presented above. In the recent years, there
have been a number of reports on the PL from the silicon
cluster-containing film in SiNx matrix or Si oxynitride, in
which emission has been conveniently assigned to the pres-
ence of nanometric amorphous silicon or Si–N bonds. In our
work, the photoexcited carriers appear to take place in a
Si-rich SiNx film and effectively move to the luminescent
centers in the Si oxide layer, resulting in strong PL emission.
Further work is needed to confirm if there indeed exists a
blue-violet PL in such system and hence to explore the pos-
sibility of strengthening the PL intensity.
In conclusion, we report the PL enhancement of the SiNx
films prepared by inductively coupled plasma assisted rf
magnetron sputtering by using a TEOS SiO2 buffer and high-
temperature annealing in N2 ambient. The PL intensity
reaches a maximum at the annealing temperature of 800 °C.
The composition and structure analyses revealed that the
strong PL is relevant with the content of the Si–N and Si–O
bonds. Sufficient oxidation and moderate nitridation can ef-
fectively enhance the photoluminescence.
This work is supported by Education Bureau of Sichuan
Province, People’s Republic of China ͑Project No.
2005A092͒ and A*STAR, Singapore ͑Project No. 042-101-
0079͒.
1L. T. Canham, Appl. Phys. Lett. 57, 1045 ͑1990͒.
2T. S. Iwayama, S. Nakao, and K. Saitoh, Appl. Phys. Lett. 65, 1814
͑1994͒.
3K. D. Hirschman, L. Tsybeskov, S. P. Duttagupta, and P. M. Fauchet,
Nature ͑London͒ 384, 338 ͑1996͒.
4L. Pavesi, L. D. Negro, C. Mazzoleni, G. Franzo, and J. P. Prolo, Nature
͑London͒ 408, 440 ͑2000͒.
5H. Takagi, H. Ogawa, Y. Yanazaki, and T. Nakagiri, Appl. Phys. Lett. 56,
2397 ͑1990͒.
6S. Hayashi and K. Yamamoto, J. Lumin. 70, 352 ͑1996͒.
7T. Shimizu-Iwayama, K. Fujita, S. Nakao, K. Saitoh, T. Fujita, and N.
Itoh, J. Appl. Phys. 75, 7779 ͑1994͒.
8H. Z. Song and X. M. Bao, Phys. Rev. B 55, 6988 ͑1997͒.
9A. Kassiba, M. Makowska-Janusik, J. Bouclé, J. F. Bardeau, A. Bulou,
and N. Herlin-Boime, Phys. Rev. B 66, 155317 ͑2002͒.
10K. Ostrikov, Rev. Mod. Phys. 77, 489 ͑2005͒.
11B. H. Augustine, Y. Z. Hu, E. A. Irene, and L. E. Mcneil, Appl. Phys. Lett.
67, 3694 ͑1995͒.
12A. Aydinli, A. Serpenguezel, and D. Vardar, Solid State Commun. 98, 273
͑1996͒.
13L. S. Liao, Z. H. Xiong, X. Zhou, X. B. Liu, and X. Y. Hou, Appl. Phys.
Lett. 71, 2193 ͑1997͒.
14T. Qiu, X. L. Wu, G. J. Wan, Y. F. Mei, G. G. Siu, and Paul K. Chu, Appl.
Phys. Lett. 86, 193111 ͑2005͒.
15M. Molinari, H. Rinnert, and M. Vergnat, Appl. Phys. Lett. 77, 3499
͑2000͒.
16Z. Pei, Y. R. Chang, and H. L. Hwang, Appl. Phys. Lett. 80, 2839 ͑2002͒.
17M. Xu, S. Xu, Y. C. Ee, Clare Yong, J. W. Chai, S. Y. Huang, and J. D.
Long, Mater. Sci. Eng., B 128, 89 ͑2006͒.
18S. Xu, K. N. Ostrikov, Y. Li, E. L. Tsakadze, and I. R. Jones, Phys.
Plasmas 8, 2549 ͑2001͒.
19T. V. Torchynska, A. Diaz Cano, M. Morales Rodriguez, and L. Yu.
Khomenkova, Physica B 340-342, 1113 ͑2003͒.
20Y. C. Fang, W. Q. Li, L. J. Qi, L. Y. Li, Y. Y. Zhao, Z. J. Zhang, and M.
Lu, Nanotechnology 15, 494 ͑2004͒.
21Z. Chen, A. Rohatgi, R. O. Bell, and J. P. Kalejs, Appl. Phys. Lett. 65,
2078 ͑1994͒.
Combining the structural and composition results, one
could establish a connection between the variation of PL and