Thermal stability of (HfO2)x(Al2O 3)1-x on Si

Article abstract of DOI:10.1063/1.1519733

The kinetics of the interfacial layer (IL) growth between Hf aluminates and the Si substrate during high-temperature rapid thermal annealing (RTA) in either N₂ (～10Torr) or high vacuum (～2×10 ^-5Torr) is studied by high-resolution x-r

Full text of DOI:10.1063/1.1519733

Thermal stability of ( HfO 2 ) x ( Al 2 O 3 ) 1−x on Si
H. Y. Yu, N. Wu, M. F. Li, Chunxiang Zhu, B. J. Cho, D.-L. Kwong, C. H. Tung, J. S. Pan, J. W. Chai, W. D.
Wang, D. Z. Chi, C. H. Ang, J. Z. Zheng, and S. Ramanathan
Citation: Applied Physics Letters 81, 3618 (2002); doi: 10.1063/1.1519733
View online: http://dx.doi.org/10.1063/1.1519733
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/81/19?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Thermal stability of nitrogen in nitrided HfSiO 2 / SiO 2 / Si ( 001 ) ultrathin films
J. Appl. Phys. 104, 103520 (2008); 10.1063/1.3021051
In situ investigation of growth and thermal stability of ultrathin Si layers on the Hf O 2 ∕ Si (100) high- κ dielectric
system
Appl. Phys. Lett. 86, 191904 (2005); 10.1063/1.1923158
Thermal stability and decomposition of the HfO 2 – Al 2 O 3 laminate system
Appl. Phys. Lett. 84, 28 (2004); 10.1063/1.1637955
Thermal decomposition behavior of the HfO 2 / SiO 2 / Si system
J. Appl. Phys. 94, 928 (2003); 10.1063/1.1578525
Thermal stability and structural characteristics of HfO 2 films on Si (100) grown by atomic-layer deposition
Appl. Phys. Lett. 81, 472 (2002); 10.1063/1.1487923
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.160 On: Fri, 21 Nov 2014 22:54:07

APPLIED PHYSICS LETTERS
VOLUME 81, NUMBER 19
4 NOVEMBER 2002
Thermal stability of „HfO2…x„Al2O3…1Àx on Si
H. Y. Yu, N. Wu, M. F. Li,a) Chunxiang Zhu, and B. J. Cho
Silicon Nano Device Lab, Department of Electrical and Computer Engineering, National University
of Singapore, Singapore 119260
D.-L. Kwong
Department of Electrical and Computer Engineering, The University of Texas, Austin, Texas 78752
C. H. Tung
Institute of Microelectronics, Singapore 117685
J. S. Pan, J. W. Chai, W. D. Wang, and D. Z. Chi
Institute of Materials Research and Engineering, 3 Research Link, Singapore 117602
C. H. Ang and J. Z. Zheng
Chartered Semiconductor Manufacturing Limited, Singapore 738406
S. Ramanathan
Genus Incorporated, Sunnyvale, California 94089
͑Received 18 July 2002; accepted 13 September 2002͒
The kinetics of the interfacial layer ͑IL͒ growth between Hf aluminates and the Si substrate during
high-temperature rapid thermal annealing ͑RTA͒ in either N2 (ϳ10 Torr) or high vacuum (ϳ2
ϫ10Ϫ5 Torr) is studied by high-resolution x-ray photoelectron spectroscopy and cross-sectional
transmission electron microscopy. The significant difference of the IL growth observed between
high vacuum and relatively oxygen-rich N2 annealing ͑both at 1000 °C) is shown to be caused by
the oxygen species from the annealing ambient. Our results also show that Hf aluminates exhibit
much stronger resistance to oxygen diffusion than pure HfO2 during RTA in N2 ambient, and the
resistance becomes stronger with more Al incorporated into HfO2 . This observation is explained by
the combined effects of ͑i͒ smaller oxygen diffusion coefficient of Al2O3 than HfO2 , and ͑ii͒ higher
crystallization temperature of the Hf aluminates. © 2002 American Institute of Physics.
͓DOI: 10.1063/1.1519733͔
Recently, HfO2-based high-k gate dielectrics including
HfO2 ,1,2 Hf silicates,3,4 and Hf aluminates,5–9 as alternates to
SiO2 have been extensively studied for future generation of
metal–oxide–semiconductor transistors to address the exces-
sive high leakage current concern. Pure as-deposited amor-
phous HfO2 crystallizes during postdeposition annealing ͑for
example ϳ400 °C),6 which may induce grain-boundary
leakage current and nonuniformity of the film thickness.10
More importantly, HfO2 is transparent to the oxygen
diffusion.3,11 Annealing in an oxygen-rich ambient will lead
to fast diffusion of oxygen through the HfO2 , causing the
growth of uncontrolled low-k interfacial layers ͑either SiOx
or SiOx-containing layer͒.10,12 The uncontrolled low-k layer
poses a serious limitation to further scaling of the equivalent
oxide thickness for HfO2 gate dielectrics. Several research
groups have demonstrated that alloying HfO2 with Al results
in the increase of crystallization temperature of HfO2
films.6–8 Furthermore, due to their reasonable k value and
the band offset values to Si,5,7,8 hafnium aluminates are thus
regarded as a promising candidate for high-k gate dielectrics
application. However, the impact of alloying Al into Hf on
oxygen diffusivity through Hf aluminates has not been stud-
ied. In this letter, we demonstrate that the resistance to the
oxygen diffusion in HfO2 films can be greatly enhanced by
the incorporation of Al.
(HfO2)x(Al2O3)1Ϫx (0рxр1) films of four different
Al2O3 compositions were deposited by atomic layer deposi-
tion at 300 °C onto 8 in. p-type ͑100͒ Si substrates. Trim-
ethyl aluminum ͓TMA, Al(CH)3], hafnium tetrachloride
(HfCl4), and water (H2O) were used as precursors, and ni-
trogen was employed as carrier and purge gas. Hf–Al–O
composite films were formed by switching between
the metal precursors, TMA and HfCl4 , while an H2O
pulse follows every metal pulse.8 Rapid thermal annealing
͑RTA͒ was conducted in either 10 Torr of N2 or in a
high vacuum (ϳ2ϫ10Ϫ5 Torr) at several temperatures
(800 °C–1000 °C) for 20 s. Thick films (ϳ20 nm) were
prepared for x-ray diffraction ͑XRD͒ and x-ray photo-
electron spectroscopy ͑XPS͒ study. The XPS results show
that the compositions of these samples are HfO2 ,
(HfO2)0.85(Al2O3)0.15 , (HfO2)0.67(Al2O3)0.33 , and Al2O3 .
Thin films were prepared using the same process recipe and
characterized using XPS and cross-sectional transmission
electron microscopy ͑XTEM͒. Thicknesses for the as-
deposited high-k thin films determined by transmission
electron microscopy are ϳ5.5 nm (HfO2), ϳ4.4 nm
(HfO )0.85(Al2O3)0.15 , ϳ4.2 nm (HfO )0.67(Al2O3)0.33 ,
͔
͓
͔
͓
2
2
and ϳ4.6 nm (Al2O3).
High-resolution XPS was used to quantitatively describe
the growth of the interfacial layer ͑IL͒ between
(HfO2)x(Al2O3)1Ϫx films and Si substrate during RTA pro-
cessing. All of the high-resolution scans were taken at a pho-
a͒
Electronic mail: elelimf@nus.edu.sg
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2002/81(19)/3618/3/$19.00 3618 © 2002 American Institute of Physics
131.193.242.160 On: Fri, 21 Nov 2014 22:54:07

Appl. Phys. Lett., Vol. 81, No. 19, 4 November 2002
Yu et al.
3619
FIG. 2. The ratio of IOxy /ISi for various (HfO2)x(Al2O3)1Ϫx samples versus
the different annealing conditions based on XPS spectra in Fig. 1. The
change in this ratio directly correlates with the variation of the IL growth.
thicker the IL. In the case of annealing in N2 , IL thickness
increases with increasing temperature for all of the four
samples. For a given annealing temperature, the extent of IL
growth is determined versus Al%, with HfO2 film ͑0% Al͒
showing the largest growth, and Al2O3 film the smallest.
Doping of HfO2 film with Al slows down the IL growth
during annealing. Based on these XPS results, one can draw
the conclusion that the ability to block oxygen diffusion
through HfO2 films is greatly enhanced by the incorporating
of Al, and the ability becomes stronger when more Al is
incorporated.
The ratio of IOxy /ISi corresponding to a high vacuum
(ϳ2ϫ10Ϫ5 Torr) annealing at 1000 °C is also plotted in
Fig. 2. Although the changes of IOxy /ISi for these four
samples show the same trend as that of annealing in N2 ,
these ratios are significantly smaller than those of annealing
in N2 at the same temperature. This result implies that the
active source of oxygen in an N2 ambient, not the oxygen
species present in the high-k films themselves, is responsible
for the IL growth during RTA.
It should also be noted that no Hf silicide (HfuSi bond-
ing͒ was detected from the Hf 4f core-level scans ͑data not
shown͒ for all samples after annealing in a high vacuum at
1000 °C. This is because the pressure for annealing in our
study is not low enough to trigger the formation of SiO spe-
cies at the high-k/IL region, which is the key step for silicide
formation.14
FIG. 1. XPS Si 2p core-level spectra recorded from various
(HfO2)x(Al2O3)1Ϫx samples ͑the thin ones͒ of as-deposited ͑solid lines͒,
after 800 °C/N2 annealing ͑dashed lines͒, 900 °C/N2 annealing ͑dotted
lines͒, 1000 °C/N2 annealing ͑dashed dotted lines͒, and 1000 °C/high
vacuum annealing ͑dashed lines with open triangles͒. The peak located at
ϳ99.3 eV is assigned to SiuSi bonds from the substrates, and the one at
ϳ103.0 eV to SiuO bonds from IL. The intensities for XPS peaks of
SiuSi bonds have been normalized for comparison.
The high-resolution XTEM micrographs of two samples
toelectron take-off angle of 90° and with a pass energy of 20
eV. The Si 2p core-level XPS spectra for the as-deposited
and after various RTA annealed (800 °C/N2 , 900 °C/N2 ,
1000 °C/N2 , and 1000 °C/high vacuum͒ samples are shown
in Fig. 1. The peak located at ϳ99.3 eV is attributed to
SiuSi bonds from Si substrates, and the one at ϳ103.0 eV
to SiuO bonds from the interfacial layer.13 For the sake of
comparison, the intensities of XPS peaks of SiuSi bonds
have been normalized. From Fig. 1, it is obvious that an IL
exists for all of the as-deposited samples. For better under-
standing, the ratio of IOxy /ISi for each sample with the dif-
ferent annealing conditions is plotted in Fig. 2. IOxy and ISi
are determined by integrating the SiuO and SiuSi peak
area, respectively, after a Shirley background subtraction.
The change in the ratio of IOxy /ISi directly reflects the varia-
HfO and (HfO2)0.85(Al2O3)0.15 films͔ before and after
͓
2
900 °C annealing in N2 are presented in Fig. 3. After anneal-
ing, it is observed that the growth of IL is greater for the
HfO2 sample compared to the Al-doped HfO2 , consistent
with the XPS results shown in Fig. 1. The composition of the
interfacial layer is likely to be Hf͑Al͒ silicate,12,15 which is
supported by XPS measurements shown in Fig. 1. More evi-
dence for silicate formation is the high-k film thickness de-
crease after annealing, as shown by XTEM images, which
may be due to the consumption of high-k films through the
reaction with IL15 and/or film densification.16 Further study
of the IL composition by high spatial resolution electron en-
ergy loss spectroscopy is underway.
As reported previously,6–8 HfO2 film crystallization tem-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
tion of the interfacial layer growth: The higher the ratio, the
perature is increased by alloying with Al. Our XRD study of
131.193.242.160 On: Fri, 21 Nov 2014 22:54:07

3620
Appl. Phys. Lett., Vol. 81, No. 19, 4 November 2002
Yu et al.
closely correlated with the Al percentage in the films. This
observation is explained by ͑i͒ Al2O3 has much lower oxy-
gen diffusion coefficient than HfO2 at high temperature; ͑ii͒
doping HfO2 by Al raises the film crystallization temperature
of HfO2 and thus drastically reduces the oxygen diffusion
along the grain boundaries during annealing. The difference
of IL growth between N2 ͑10 Torr͒ and a high vacuum
(ϳ2ϫ10Ϫ5 Torr) 1000 °C annealing indicates that the ac-
tive oxygen species from the annealing ambient is the cause
for the IL growth during RTA in N2 . A reduced oxygen at-
mosphere is therefore required to suppress the IL growth for
(HfO2)x(Al2O3)1Ϫx films on Si during the postdeposition
process.
This work was supported by the Singapore grant ͑No.
NSTB/EMT/TP/00/001.2͒ and the National University of
Singapore grant ͑No. R263-000-221-112͒. The authors thank
Professor J. Y. Lin of the Dept. of Physics at NUS for use of
the XRD apparatus.
FIG. 3. HRTEM images for of ͑a͒ the as-is HfO2 sample, ͑b͒ the 900 °C/N2
annealed HfO2 sample, ͑c͒ the as-is (HfO2)0.85(Al2O3)0.15 sample, and ͑d͒
the 900 °C/N2 annealed (HfO2)0.85(Al2O3)0.15 sample.
1 Y. Harada, M. Niwa, S. J. Lee, and D. L. Kwong, Symp. VLSI Tech. Dig.
3-3-1, ͑2002͒.
2 L. Kang, K. Onishi, Y. Jeon, B. H. Lee, C. Kang, W. J. Qi, R. Nieh, S.
Gopalan, R. Choi, and J. C. Lee, Tech. Dig. - Int. Electron Devices Meet.
2000, 35 ͑2000͒.
20 nm thick (HfO2)x(Al2O3)1Ϫx films after 900 °C anneal-
ing in N2 also confirms this point, as depicted in Fig. 4.
These spectra show that (HfO2)0.67(Al2O3)0.33 remains
amorphous after 900 °C annealing. In addition, the crystalli-
zation temperatures increases with more Al incorporated.6–8
These results are well correlated with the experimental ob-
3 G. D. Wilk, R. M. Wallace, and J. M. Anthony, J. Appl. Phys. 87, 484
͑2000͒.
4 S. Gopalan, K. Onishi, R. Nieh, C. S. Kang, R. Choi, H. J. Cho, S.
Krishna, and J. C. Lee, Appl. Phys. Lett. 80, 4416 ͑2002͒.
5 H. Y. Yu, M. F. Li, B. J. Cho, C. C. Yeo, M. S. Joo, D. L. Kwong, J. S.
Pan, C. H. Ang, J. Z. Zheng, and S. Ramanathan, Appl. Phys. Lett. 81, 376
͑2002͒.
servation
regarding
oxygen
diffusion
through
6 W. Zhu, T. P. Ma, T. Tamagawa, Y. Di, J. Kim, R. Carruthers, M. Gibso,
and T. Furukawa, Tech. Dig.-Int. Electron Devices Meet. 2001, 20.4.1
͑2001͒.
(HfO2)x(Al2O3)1Ϫx films ͑Figs. 1 and 2͒: the higher Al con-
centration, the higher the crystallization temperature and,
hence, the lower the oxygen diffusion along the grain bound-
aries of the high-k films which, in turn, reduces the IL
growth. Another reason for the reduced rate of oxygen diffu-
sion by the addition of Al2O3 is that the Al2O3 is known to
have much lower oxygen diffusion coefficient compared to
HfO2 at high temperature.17
7 G. D. Wilk, M. L. Green, M. Y. Ho, B. W. Busch, T. W. Scorsch, F. P.
Klemens, B. Brijs, R. B. van Dover, A. Kornblit, T. Gustafsson, E. Gar-
funkel, S. Hillenius, D. Monroe, P. Kalavade, and J. M. Hergenrother,
VLSI Tech. Dig. 9-4-1, ͑2002͒.
8 A. R. Londergan, S. Ramanathan, K. Vu, S. Rassiga, R. Hiznay, J. Win-
kler, H. Velasco, L. Matthysse, T. E. Seidel, C. H. Ang, H. Y. Yu, and M.
F. Li, ͑unpublished͒.
9 Y. T. Hou, M. F. Li, H. Y. Yu, Y. Jin, and D. L. Kwong, Tech. Dig.–Int.
Electron Meet. ͑to be published͒.
In conclusion, we report that both the thermal stability
and the resistance to oxygen diffusion of HfO2 are improved
by adding Al to form Hf aluminates, and the improvement is
10 G. D. Wilk, R. M. Wallace, and J. M. Anthony, J. Appl. Phys. 89, 5243
͑2001͒.
11 A. Kumar, D. Rajdev, and D. L. Douglass, J. Am. Chem. Soc. 55, 439
͑1972͒.
12 C. Hobbs, H. Tseng, K. Reid, B. Tayhor, L. Dip, L. Hebert, R. Garcia, R.
Hedge, J. Grant, D. Gilmer, A. Frake, V. Dhandapani, M. Azrak, L.
Prabhu, R. Rai, S. Bagchi, J. Conner, S. Backer, F. Dumbuya, B. Nguyen,
and P. Tobin, Tech. Dig. - Int. Electron Devices Meet. 2001, 30.1.1 ͑2001͒.
13 J. F. Moulder, Handbook of X-ray Photoelectron Spectroscopy, 2nd ed.
͑Physical Electronics, Eden Prairie, MN, 1992͒.
14 T. S. Jeon, J. M. White, and D. L. Kwong, Appl. Phys. Lett. 78, 368
͑2001͒.
15 M. Cho, J. Park, H. B. Park, C. S. Hwang, J. Jeong, and K. S. Hyun, Appl.
Phys. Lett. 81, 334 ͑2002͒.
16 Y. Kim, G. Gebara, M. Freiler, J. Barnett, D. Riley, J. Chen, K. Torres, J.
Kim, B. Foran, F. Shaapur, A. Agarwal, P. Lysaght, G. Brown, C. Young,
S. Borthakur, H. Li, B. Nguyen, P. Zeitzoff, G. Bersuker, D. Derro, R.
Bergmann, R. Murto, A. Hou, H. Huss, E. Shero, C. Pomarede, M. Giv-
ens, M. Mazanec, and C. Werkhoven, Tech. Dig. - Int. Electron Devices
Meet. 2001, 20.2.1 ͑2001͒.
FIG. 4. XRD characteristics of various (HfO2)x(Al2O3)1Ϫx samples
(ϳ20 nm) after 900 °C/N2 annealing.
17 M. L. Green, E. P. Gusev, R. Degraeve, and E. L. Garfunkel, J. Appl.
Phys. 90, 2057 ͑2001͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.160 On: Fri, 21 Nov 2014 22:54:07