Stabilization of ambient sensitive atomic layer deposited lanthanum aluminates by annealing and in situ capping

Article abstract of DOI:10.1063/1.3557501

We have studied the effect of air exposure on lanthanum aluminates (LaAlO_x) deposited by atomic layer deposition. Fourier transform infrared spectroscopy and thermal desorption spectroscopy of as-deposited LaAlO_x showed that H₂ O is absorbed during air exposure and that the amount of absorbed H₂ O increases with increasing La atomic percent. C was found to be incorporated already during deposition in the form of carbonates. H₂ O and CO₂ are outgassed during postdeposition annealing in an inert atmosphere. After a 700 °C postdeposition anneal, the LaAlO_x becomes resistant against H ₂ O absorption due to film densification. Alternatively, in situ capping of the LaAlO_x with a ～2 nm thin Al₂ O ₃ film protects the LaAlO_x against H₂ O absorption, but it also hinders the outgassing of the C contaminants during a postdeposition anneal.

Full text of DOI:10.1063/1.3557501

Stabilization of ambient sensitive atomic layer deposited lanthanum aluminates by
annealing and in situ capping
J. Swerts, S. Gielis, G. Vereecke, A. Hardy, D. Dewulf, C. Adelmann, M. K. Van Bael, and S. Van Elshocht
Citation: Applied Physics Letters 98, 102904 (2011); doi: 10.1063/1.3557501
View online: http://dx.doi.org/10.1063/1.3557501
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/98/10?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Effect of annealing and electrical properties of high- κ thin films grown by atomic layer deposition using
carboxylic acids as oxygen source
J. Vac. Sci. Technol. B 27, 230 (2009); 10.1116/1.3058743
Thermal stability of surface and interface structure of atomic layer deposited Al 2 O 3 on H-terminated silicon
J. Appl. Phys. 102, 094503 (2007); 10.1063/1.2803727
Characteristics of atomic-layer-deposited thin Hf x Zr 1 − x O 2 gate dielectrics
J. Vac. Sci. Technol. B 25, 845 (2007); 10.1116/1.2734978
Improvement of thermal stability and composition changes of atomic layer deposited Hf O 2 on Si by in situ O
3 pretreatment
Appl. Phys. Lett. 90, 152906 (2007); 10.1063/1.2721845
Annealing behavior of atomic layer deposited hafnium oxide on silicon: Changes at the interface
J. Appl. Phys. 99, 094102 (2006); 10.1063/1.2191434
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:
137.149.200.5 On: Mon, 10 Aug 2015 22:15:32

APPLIED PHYSICS LETTERS 98, 102904 ͑2011͒
Stabilization of ambient sensitive atomic layer deposited lanthanum
aluminates by annealing and in situ capping
J. Swerts,1,a͒ S. Gielis,1,2 G. Vereecke,1 A. Hardy,3 D. Dewulf,3 C. Adelmann,1
M. K. Van Bael,3 and S. Van Elshocht1
1IMEC vzw, B-3001 Leuven, Belgium
2Department of Chemistry, K.U. Leuven, B-3001 Leuven, Belgium
3Inorganic and Physical Chemistry, IMO, and IMEC, Division IMOMEC, Hasselt University,
B-3590 Diepenbeek, Belgium
͑Received 6 October 2010; accepted 27 January 2011; published online 8 March 2011͒
We have studied the effect of air exposure on lanthanum aluminates ͑LaAlOx͒ deposited by atomic
layer deposition. Fourier transform infrared spectroscopy and thermal desorption spectroscopy of
as-deposited LaAlOx showed that H2O is absorbed during air exposure and that the amount of
absorbed H2O increases with increasing La atomic percent. C was found to be incorporated already
during deposition in the form of carbonates. H2O and CO2 are outgassed during postdeposition
annealing in an inert atmosphere. After a 700 °C postdeposition anneal, the LaAlOx becomes
resistant against H2O absorption due to film densification. Alternatively, in situ capping of the
LaAlOx with a ϳ2 nm thin Al2O3 film protects the LaAlOx against H2O absorption, but it also
hinders the outgassing of the C contaminants during a postdeposition anneal. © 2011 American
Institute of Physics. ͓doi:10.1063/1.3557501͔
Many dielectrics with high permittivity, including the
class of lanthanide oxides and aluminates, have been studied
in order to replace SiO2 in future transistor and capacitor
devices.1 These materials show interesting properties such as
a high permittivity and band gap. However, because the lan-
thanide oxides are known to be hygroscopic, exposure to
ambient moisture can potentially affect their dielectric
quality.2,3 In addition to moisture, C incorporation by reac-
tion with CO2 has also been reported.4,5 Hence, an in situ
metal gate process would be required when integrating the
lanthanide oxides in future complementary metal oxide semi-
conductor applications.6 Besides in situ capping, mixing with
a less hygroscopic oxide such as Al2O3 is proposed to stabi-
lize the lanthanide oxide.7 LaAlO3 is reported to be much
less susceptible to moisture: it combines the advantages of
the high-k of La2O3 and the chemical and thermal stability of
Al2O3.8
Atomic layer deposition ͑ALD͒ has become the pre-
ferred industrial deposition technique for high-k oxides be-
cause of its low deposition temperature and its excellent
thickness and composition control and conformal growth, as
it is based on self-limiting surface reactions.9 The purity and
density of as-deposited ALD films is often somewhat lower
than those deposited by molecular beam epitaxy or physical
vapor deposition, which makes them more susceptible to ab-
sorption of species from the ambient. Despite the existence
of several reports on the deposition and characterization of
ALD LaAlOx,10,11 studies of the absorption and desorption of
contaminants during air exposure are lacking. In this letter,
we describe the sensitivity to air exposure of as-deposited
ALD LaAlOx films and we study the origin of H2O and
organic in-film contaminants. Furthermore, we compare the
effectiveness of two potential methods to protect the layer
against the effects of air exposure: in situ capping versus a
postdeposition anneal.
We have deposited LaAlOx films on 200 or 300 mm
Si wafers by ALD from tris͑2,2,6,6-tetramethyl-3,5-
heptanedionato͒lanthanum ͓La͑thd͒ ͔, trimethylaluminum
3
͑TMA͒, and O3 in an ASM Pulsar®3000 ALD reactor.12 The
deposition temperature was 300 °C. The composition was
varied by changing the subcycle values ͑x,y͒ in the
͓TMA/O3͔x/͓La͑thd͒3/O3͔y ALD sequence. ͑1:0͒, ͑1:2͒,
͑1:5͒, and ͑1:10͒ resulted in 0, 28, 50, and 65 at. % La/͑La
+Al͒, respectively, as determined by x-ray photoelectron
spectroscopy ͑XPS͒. Prior to the ALD, the native oxide of
Si͑100͒ substrates was stripped in HF and the bare Si surface
reoxidized in an O3/H2O solution to a nominal thickness of
ϳ1 nm ͑IMEC clean͒.13 On one set of wafers, a ϳ2 nm
Al2O3 capping layer was deposited in situ after the deposi-
tion of the LaAlOx. A second set of samples was annealed in
N2 at 700 °C in an ASM Levitor®4300 or at 500 °C in Ar in
a tube furnace. Thickness measurements were done with a
KLA-TENCOR Aleris spectroscopic ellipsometer. Mass
measurements of the deposited films were performed on a
300 mm Metrix Mentor SF3-Single FOUP system. X-ray re-
flectivity ͑XRR͒ was used to determine the film density. The
chemical structure of as-deposited LaAlOx films was mea-
sured by grazing angle attenuated total reflectance Fourier
transform infrared spectroscopy ͑GATR-FTIR͒ on a Harrick
Bruker Vertex 70 tool.14 Thermal desorption spectroscopy
͑TDS͒ measurements were carried out in a rapid thermal pro-
cessor ͑RTP͒ coupled to an atmospheric pressure ionization
mass spectrometer.15
Figure 1͑a͒ and 1͑b͒ show the effect of air exposure on
the thickness and mass of LaAlOx layers, respectively. The
initial thickness of the films was ϳ10 nm. As can be seen,
the thickness of the LaAlOx film increases much more as
compared to both the Al2O3 and Al2O3 capped LaAlOx lay-
ers. Since the capped LaAlOx film shows identical aging
behavior as Al2O3, one can assume that the absorption of
species from air is limited to surface adsorption on the Al2O3
surface. Thus, 2 nm Al2O3 is an effective protective layer for
a͒
Electronic mail: johan.swerts@imec.be.
0003-6951/2011/98͑10͒/102904/3/$30.00
98, 102904-1
© 2011 American Institute of Physics
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:
137.149.200.5 On: Mon, 10 Aug 2015 22:15:32

102904-2
Swerts et al.
Appl. Phys. Lett. 98, 102904 ͑2011͒
0.6
0.5
0.4
0.3
0.2
0.1
0
1
0.8
0.6
0.4
0.2
0
(a)
AAl2l2OO3
[La] ~350%
(a)
[La] ~ 65%
[La] ~ 50%
[La] ~ 50% + cap
[La] ~ 50% + cap
[La] ~ 25%
Si/SiO2
0
12
24
36
48
Air exposure afterdeposition (h)
0.1
3600
3100
2600
2100
1600
1100
3.4
3.2
3
2.8
2.6
2.4
2.2
2
1 h after dep
(b)
Wavenumber (cm-1)
0.4
24 h after dep
.4
.3
.2
.1
24 h after 700C PDA
[La]~ 50%
+ cap
[La]~ 50%
(b)
(c)
As dep
As dep
0.3
0.2
0.1
Afteranneal
Afteranneal
1600 1450 1300
1600 1450 1300
no cap
2O3 cap
Al2O3
Wavenumber (cm-1)
Wavenumber (cm-1)
FIG. 1. ͑a͒ Thickness increase due to air exposure in clean room environ-
ment ͑40% humidity͒ measured by ellipsometry of 10 nm as-deposited
Al2O3 ͑᭡͒, LaAlOx ͑᭿͒, and LaAlOx with 2 nm Al2O3 cap ͑b͒. ͑b͒ Mass
measurement of an uncapped and a capped 10 nm LaAlOx film as-deposited,
24 h after deposition, and 24 h after 700 °C N2 anneal.
FIG. 3. ͑a͒ GATR-FTIR spectra of as-deposited LaAlOx films of various
compositions. The spectrum of a Si/SiO2 ͑1 nm͒ substrate and a 50%
LaAlOx film with 2nm cap are also shown; ͑b͒ and ͑c͒ show the carbonate
doublet and SiO2 LO band peak immediately before and after a 500 °C Ar
anneal for an as-deposited LaAlOx film without and with Al2O3 cap,
respectively.
the bulk LaAlOx film. Note that the observed thickness in-
crease also depends on the composition of the LaAlOx films:
more Al atomic percent results in a lower thickness increase
for a similar air exposure time ͑data not shown͒. Hence, the
addition of aluminum reduces the sensitivity of a lanthanide
oxide layer toward air exposure, but clearly not to a negli-
gible level. This is in contrast to what would have been ex-
pected for stoichiometric or Al-rich LaAlOx layers. The ab-
sorption of species from air for uncapped LaAlOx layers is
confirmed by a significant mass increase ͓Fig. 1͑b͔͒. After
anneal, the mass of the uncapped film is slightly lower than
that of the as-deposited mass, indicating that the LaAlOx has
outgassed absorbed species and does not reabsorb them. For
the capped film, the air exposure and anneal do not have a
significant impact on the mass, suggesting that no significant
absorption or desorption occurred. The stabilization of the
annealed uncapped film can likely be attributed to film den-
sification. XRR measurements showed that the density in-
creases from 4.5 to 5.1 g/cm3 by annealing, whereas the
thickness is reduced from ϳ18 to ϳ14 nm ͑see Fig. 2͒. Note
that the as-deposited LaAlOx films have a ϳ2.5 nm low
density top layer ͑ϳ3.2 g/cm3͒. The thickness of this top
layer decreases with increasing Al%, but is independent of
thickness in the studied range of 5–20 nm suggesting the
layer is formed by absorption of species from the ambient.
After anneal, this low density top layer disappeared.
We have used GATR-FTIR and TDS to study the origin
of the absorbed and desorbed species in 20 nm LaAlOx films.
The films were exposed to air for more than 7 days before
the measurements. For as-deposited films ͓see Fig. 3͑a͔͒, the
GATR-FTIR substrate spectrum shows one sharp peak
around 1240 cm−1 that can be assigned to the longitudinal
optical band of SiO2. Besides, one can observe a broad peak
at 3100–3600 cm−1 that can be attributed to hydrogen
bonded OH stretching vibration. The OH band decreases in
intensity with decreasing atomic percent La in the LaAlOx.
When comparing OH signals for the sample with and with-
out cap, it becomes clear that the cap significantly reduces
the formation of the broad OH peak. Consequently, this OH
peak must be caused by absorption of H2O during air expo-
sure. The more La, the more H2O can be absorbed due to the
hygroscopic nature of La2O3. We do not find evidence of
hydroxide formation, i.e., there are no sharp peaks around
3450 and 3610 cm−1 in contrast to what has been reported
earlier for ALD La2O3 films after air exposure.5
5
0E+00
10
As deposited
3.2 g/cm3
LaAlOx
4
0E-02
10
4.5 g/cm3
Si
0E-043
10
0E-062
10
Annealed
0E-081
10
LaAlOx
5.1 g/cm3
Si
100
0E-10
0
5000
10000
θ (arcsec)
The doublet with maxima around 1470 and 1340 cm−1
can be identified as being due to unidentately bound La car-
bonate species, i.e., La2O2CO3.5,16 Similarly as for the OH
FIG. 2. XRR spectra ͑symbols͒ of an air explosed LaAlOx ͑50 at. % La͒
layer as deposited and after 700 °C anneal in N2. The full lines show the
simulation of the XRR spectra. The structure of the stacks to obtain a good
fit with the experimental data is schematically shown.
peak, the intensity of the carbonate peak decreases with de-
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:
137.149.200.5 On: Mon, 10 Aug 2015 22:15:32

102904-3
Swerts et al.
Appl. Phys. Lett. 98, 102904 ͑2011͒
Figure 4͑b͒ shows the CO2 thermal desorption. No peaks
(a)
[La]~ 50%
4.0E+05 5
are observed for Al2O3 and the annealed LaAlOx. The as-
4.0 x 10
Al2O3
[La]~ 50% + cap
deposited LaAlOx shows
a significant peak around
400–600 °C, whereas for the capped LaAlOx layer, only a
slightly enhanced signal can be observed. The TDS observa-
tions are in agreement with GATR-FTIR measurements that
were carried out on capped and uncapped LaAlOx films after
a 500 °C argon anneal. In Fig. 3͑b͒, the carbonate peak be-
fore and after anneal is shown for an uncapped 50% LaAlOx
film, revealing a strong reduction after anneal. To the con-
trary, only a very small reduction of the carbonate intensity is
observed after a similar anneal for the capped LaAlOx film
͓see Fig. 3͑c͔͒. This would again be consistent with Al2O3 as
barrier working in both directions, i.e., not only keeping CO2
out during air exposure, but unfortunately also keeping CO2
in during anneal.
In summary, we have shown that as-deposited ALD
LaAlOx absorbs a large amount of H2O when exposed to
atmospheric ambient, whereas C is already incorporated as
carbonates during the deposition process. We demonstrated
that a 2 nm thin Al2O3 in situ cap can protect the layer and
inhibit H2O absorption, but at the same time also prevent that
in-film C and excessive O impurities can be outgassed during
a postdeposition anneal. This compromises an approach of
clustered processing to avoid air exposure. In this respect, a
postdeposition anneal of the LaAlOx layer seems to be a
more viable approach when integrating ALD LaAlOx layer in
device stacks. On the one hand, it leads to outgassing of H2O
and CO2, while on the other hand it densifies the layer avoid-
ing reabsorption of H2O or CO2 during subsequent air expo-
sure.
[La]~ 50% + PDA
22.0.0E+x01505
Si/SiO2
400
Mass 18
800
0.0 x 100
0.0E+00
0
200
600
Temperature (°C)
1.5E+05 5
1.5 x 10
Mass 44
1.0E+05 5
1.0 x 10
[La]~ 50%
[La]~ 50%+ cap
Al2O3
5.0E+04 4
5.0 x 10
[La]~ 50%
+ PDA
(b)
0.0 x 100
0.0E+00
0
200
400
600
800
Temperature (°C)
FIG. 4. Amount of ͑a͒ H2O ͑mass 18͒ and ͑b͒ CO2 ͑mass 44͒ desorbing
from ϳ15% nm LaAlOx films with and without 2 nm Al2O3 cap as function
of temperature ͑obtained at a constant heating rate of 60 °C/min͒. One of
the films got a 700 °C N2 post deposition anneal. An Al2O3 film and a bare
Si/SiO2 wafer are added as reference. Due to an artifact in the measurement
of the as-deposited uncapped LaAlOx, a dip in intensity in the range of
450–650 °C is observed. For this reason, a dotted interpolated line is shown
in the figure in this range.
creasing atomic percent La. This is consistent with time-of-
flight secondary ion mass spectroscopy data where the C
content was found to decrease with decreasing La content
͑data not shown͒. Note that the amount of C in the LaAlOx
films was in all cases below the detection limit of XPS ϳ1%.
More importantly and opposed to the OH band, in situ cap-
ping of the LaAlOx layer does not affect the intensity of the
carbonate doublet. Because ALD Al2O3 is known as a good
CO2 barrier,17 our data suggest that the carbonates must
solely have formed during the LaAlOx deposition process as
has been observed earlier for ALD La2O3 processes5,14 and
not also by reaction with CO2 during air exposure as reported
for La2O3 and other rare earth oxides deposited by chemical
vapor deposition.4,18 This also implies that the aging behav-
ior as shown in Fig. 1͑a͒ for the case of a noncapped LaAlOx
film is only caused by absorption of H2O and not CO2.
Next, the impact of a thermal postdeposition treatment is
discussed. TDS was used to monitor H2O and CO2 desorp-
tion. For H2O ͓Fig. 4͑a͔͒, one can clearly observe a first,
relatively sharp peak around 100–200 °C originating from
the physisorbed H2O that desorbs from the surface:
LaAlOxϾLaAlOx/Al2O3ХAl2O3ХLaAlOx+PDA
1Rare Earth Oxide Thin Films: Growth, Characterization, and Applica-
tions, Applied Physics, Vol. 106, edited by M. Fanciulli and G. Scarel
͑Springer, Berlin, 2007͒.
2D. Niu, R. W. Ashcraft, and G. N. Parsons, Appl. Phys. Lett. 80, 3575
͑2002͒.
3S. Jeon and H. Hwang, J. Appl. Phys. 93, 6393 ͑2003͒.
4T. Gougousi, D. Niu, R. W. Ashcraft, and G. N. Parsons, Appl. Phys. Lett.
83, 3543 ͑2003͒.
5M. Nieminen, M. Putkonen, and L. Niinisto, Appl. Surf. Sci. 174, 155
͑2001͒.
6Y. Zhao, M. Toyama, K. Kita, K. Kyuno, and A. Toriumi, Appl. Phys.
Lett. 88, 072904 ͑2006͒.
7Y. Zhao, K. Kita, and A. Toriumi, Appl. Phys. Lett. 96, 242901 ͑2010͒.
8T. D. Manning, Y. F. Loo, A. C. Jones, H. C. Aspinall, P. R. Chalker, J. F.
Bickley, L. M. Smith, and G. W. Critchlow, J. Mater. Chem. 15, 3384
͑2005͒.
9T. Suntola, Mater. Sci. Rep. 4, 261 ͑1989͒.
10M. Nieminen, T. Sajavaara, E. Rauhala, M. Putkonen, and L. Niinisto, J.
Mater. Chem. 11, 2340 ͑2001͒.
11K. Kukli, M. Ritala, V. Pore, M. Leskala, T. Sajavaara, R. I. Hedge, D. C.
Gilmer, P. J. Tobin, A. C. Jones, and H. Aspinall, Chem. Vap. Deposition
12, 158 ͑2006͒.
12Pulsar® 3000 is a registered trademark of ASM International NV.
13M. Meuris, P. W. Mertens, A. Opdebeeck, H. F. Schmidt, M. Depas, G.
Vereecke, M. M. Heyns, and A. Philipossian, Solid State Technol. 38, 109
͑1995͒.
ӷSi/SiO2. The as-deposited LaAlOx clearly is most sensi-
tive to H2O adsorption, whereas the annealed, i.e., the stabi-
lized, LaAlOx film is the least. H2O desorption is very simi-
lar for Al2O3 and the Al2O3-capped LaAlOx. The latter
confirms our earlier conclusion based on thickness measure-
ment that for Al2O3-capped LaAlOx, the absorption of H2O
is limited to surface adsorption on the Al2O3 top surface. In
contrast to the GATR-FTIR measurement, TDS reveals the
presence of chemisorbed H2O in the case of the as-deposited
LaAlOx; a second, broad peak occurs around 300–450 °C.
The Si/SiO2 line reflects the H2O background caused by
14A. Hardy, C. Adelmann, S. Van Elshocht, H. Van den Rul, M. K. Van
Bael, S. De Gendt, M. D’Olieslaeger, M. Heyns, J. A. Kittl, and J. Mul-
lens, Appl. Surf. Sci. 255, 7812 ͑2009͒.
15G. Vereecke, E. Kondoh, P. Richardson, K. Maex, and M. Heyns, IEEE
Trans. Semicond. Manuf. 13, 315 ͑2000͒.
16J. Kwon, M. Dai, M. D. Halls, E. Langereis, Y. J. Chabal, and R. G.
Gordon, J. Phys. Chem. C 113, 654 ͑2009͒.
17M. D. Groner, F. H. Fabreguette, J. W. Elam, and S. M. George, Chem.
Mater. 16, 639 ͑2004͒.
18X. L. Li, D. Tsoutsou, G. Scarel, S. C. Capelli, S. N. Volkos, L. Lamagna,
outgassing of the RPT chamber.
and M. Fanciulli, J. Vac. Sci. Technol. A 27, L1 ͑2009͒.
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:
137.149.200.5 On: Mon, 10 Aug 2015 22:15:32