Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 4 (2017) pp. 428-433
© Research India Publications. http://www.ripublication.com
428
Studying Composition of Al2O3 Thin Films Deposited by Atomic Layer
Deposition (ALD) and Electron-Beam Evaporation (EBE) upon Rapid
Thermal Processing
Alexey Alexeevich Dronov, Darya Alexeevna Dronova, Elena Petrovna Kirilenko,
Igor Makarovich Terashkevich, and Sergei Aleksandrovich Gavrilov
National Research University of Electronic Technology (MIET)
Bld. 1, Shokin Square, Zelenograd, Moscow, 124498, Russia.
Abstract
This article discusses integrated analysis of composition of
aluminum oxide layers deposited onto substrates of
monocrystalline Si by atomic layer deposition (ALD) and
electron-beam evaporation (EBE) using Auger spectroscopy and
time-of-flight secondary ion mass spectrometry (ToF SIMS).
Peculiar features of surface morphology and in-depth
compositions of layers deposited by these methods upon rapid
thermal annealing in argon environment were investigated.
Detailed comparative specimens analysis of these procedures
have been performed upon deposition of thin passivating layers of
aluminum oxide on monocrystalline Si substrates before and after
thermal post-processing. Possible reasons of the processes
occurring in the considered structures upon thermal processing
have been determined. Reasonability of deposition of atomic
layer deposition of technological layers upon generation of
passivating and antireflection coatings in photovoltaics has been
substantiated.
Keywords: Al2O3 films, silicon plates, atomic layer deposition,
electron-beam evaporation, Auger spectroscopy, time-of-flight
secondary ion mass spectrometry, surface morphology, thermal
annealing, element distribution profile.
INTRODUCTION
Nowadays surface passivation procedures become more and more
important in production of high efficient solar cells. Substrate
thickness is made thinner in order to decrease prime costs. While
volume to surface area ratio decreases the surface recombination
probability with numerous imperfections, namely dangling bonds,
increases. Currently some materials for passivation on the basis of
silicon are already studied, for instance, SiO2, SiNx, SiC and a-Si.
Various deposition procedures of passivating coatings of various
compositions are applied, in particular: chemical vapor deposition
(CVD) [1], metalorganic chemical vapor deposition (MOCVD)
[2], spray pyrolysis [3], thermal evaporation [4], magnetron
sputtering [5]. In the latter procedure one of the most promising
passivating materials is Al2O3.
Thus, in [6] aluminum oxide films were obtained by electron
beam evaporation under ultrahigh vacuum. The authors studied
microstructure, optical and dielectric properties of Al2O3 thin
films as a function of substrate temperature and annealing
temperature.
At the same time atomic layer deposition (ALD) of Al2O3
became a promising approach to researches due to superior
passivating properties for Si of p- and n-type [7-8]. Such
passivating properties are related with high negative charge
which generates charged layer on Si surface in order to
decrease recombination [9-11]. Al2O3 coatings deposited by
ALD are characterized by extended charge lifetime and low
rate of surface recombination on Si substrates of n- and p-
type [12-15]. In addition it should be mentioned that ALD
is based on the principles of precursor adsorption with its
subsequent transformation in oxide, thus, the Al2O3 layer
deposited on silicon substrates can decrease -surface
imperfections density of silicon crystalline structure. Such
decrease in the surface imperfections density is also related
with high passivation level of aluminum oxide.
However, production of solar cells involves high
temperature procedures upon formation of metal contacts.
These procedures can impair passivating properties of
aluminum oxide layers. Thus, some works were published
devoted to investigations into degradation and
imperfections occurrence of Al2O3 passivating layers after
thermal processing [16-17].
This article investigates into peculiar features of
composition variations of Al2O3 thin films deposited onto
monocrystalline Si substrates by ALD and electron beam
evaporation upon rapid thermal processing.
EXPERIMENTAL
In both cases silicon plates Si (100) with the diameter of
100 mm were used as substrates with the layer native oxide
SiO2. Deposition was carried out by ALD and electron
beam evaporation.
Deposition of aluminum oxide by atomic layer deposition
was carried out on a QT-ALD 250 facility of our own
design. The temperature in reaction chamber was
maintained at 200°C. Trimethylaluminum (TMA, Sigma-
Aldrich) and deionized water were used as precursors
maintained at ambient temperature. Nitrogen was used as
carrier and purging gas. The following parameters were
applied as optimum deposition mode:
• Temperature of working chamber: 200°C;
• Vacuum range in working chamber: 0.01-0.1 mbar;
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 4 (2017) pp. 428-433
© Research India Publications. http://www.ripublication.com
429
• Duration of TMA feeding into reactor: 0.5 s;
• Duration of water injection into reactor: 0.3 s;
• Duration of purging: 5 s;
• Nitrogen consumption upon purging: 0.6 l/min;
• Number of cycles: 200.
Deposition of aluminum oxide by electron beam evaporation was
carried out on an Angstrom EvoVac facility with the following
process variables:
• Pressure in vacuum chamber: 2.6*10-5 mbar;
• Cleaning was performed by Ar+ ions in 30 seconds;
• Substrate temperature: 20ºC;
• Sapphire plates were used as sputtering material.
Thermal processing was carried out on an AS-One rapid thermal
annealing (RTA) facility (ANNEALSYS, France). Annealing was
performed at 800ºC and 1000ºC in argon environment in 5 min.
Heating rate was 20ºC/s.
Profile analysis of composition and structure of the obtained
Al2O3 films before and after thermal processing was performed by
Auger spectroscopy using a PHI-670 xi Physical Electronics
spectrometer and by time-of-flight secondary ion mass
spectrometry using a ToF-SIMS.5 instrument.
Profile analysis by Auger spectroscopy was carried out at the
accelerating voltage of primary electron beam of 5kV, primary
current of 10 nA at the surface of ~Ø75 ÷100 µm in order to
prevent artifacts of electron beam impact on specimen (reduction
of oxide, interlayer diffusion, and so on). Sputtering upon profile
analysis was performed by Ar+ ions with the accelerating voltage
of 2 kV, current of 0.5 µA at 30° to specimen surface. Herewith,
the ion beam was unfolded into a raster of 1.5×2 mm (in order to
adjust sputtering rate). Concentration of elements was calculated
according to the model of homogeneous distribution using
relative coefficients of reverse element sensitivity. The element
sensitivity coefficients were determined by reference specimens
analysis: sapphire plates and high temperature SiO2 on Si.
Time-of-flight secondary ion mass spectrometry was performed
using a ToF-SIMS.5 instrument. Analysis was performed by Bi
ions. Layer-by-layer etching was performed by Cs ions.
Accelerating etching voltage of 500 V was selected as optimum
parameter for in-depth resolution, since in such etching mode
additional relief propagation does not occur. Etching area was
300×300 µm. Analysis area: 100×100 µm. Such parameters were
selected in order to eliminate the influence of etching crater walls
on final result.
Recalculation of etching time to the depth was performed after
depth measurement of ion etching crater using an Alpha-Step D-
120 KLA Tencor stylus profilometer.
RESULTS AND DISCUSSION
In both cases the thickness of deposited aluminum oxide films
about 20 nm.
Figure 1 illustrates the profiles of element in-depth
distribution in initial Al2O3, films deposited onto
monocrystalline Si plate by ALD and electron beam
evaporation obtained by Auger spectroscopy.
a)
b)
Figure 1. Element in-depth distribution profile in initial
Al2O3 films deposited onto monocrystalline Si substrate by
ALD (a) and electron beam evaporation (b).
Already on initial ALD sample the O peak is highlighted at
Si-Al2O3 interface which differs from the oxygen peak in
Al2O3 film. On initial Al2O3 sample obtained by electron
beam sputtering in vacuum in addition to the oxygen peak,
which differs from the peak in Al2O3 film, the interface
contains the Si(O)–78 eV peak. It means that upon
deposition the Si surface already become oxidized. The
width of transient layer upon profile analysis is determined
not only by true layer density but also by average surface
roughness and thickness nonuniformity on the analysis
surface area. Since the roughness of both initial samples is
in fact the same and less than 0.3 nm, it is obvious that on
ALD samples the layer thickness with oxygen, bonded not
only with Al or in free form, amounts to several Å. In
Al2O3 samples (electron beam) SiOx layer formation begins
at the interface with the thickness about several nm.
While increasing the annealing temperature on interface of
both samples Si(O)–78 eV content increases, which
evidences continuous oxidation of Si at the interface and
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 4 (2017) pp. 428-433
© Research India Publications. http://www.ripublication.com
430
thickness increasing of this oxidized layer. On ALD samples the
thickness of this layer at equal annealing temperatures is more
than by 1.5 times lower than on the Al2O3 samples formed by
electron beam evaporation.
Figure 2 illustrates the images of deposited films surfaces
in secondary electrons of Al2O3 films of both types before
and after annealing.
a b
c d
e f
Figure 2. SEM images of surface of Al2O3 films deposited onto monocrystalline Si plate by ALD (a, c, e) and electron beam
evaporation (b, d, f) before (a, b) and after (c-f) thermal annealing at 800ºC (c, d) and 1000ºC (e, f).
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 4 (2017) pp. 428-433
© Research India Publications. http://www.ripublication.com
431
As it can be seen from the presented images the surface of initial
samples is very smooth. However, upon thermal processing
variations of surface morphology can be observed. While
comparing these variations it is obvious that the film structure
upon rapid annealing varies differently.
The detected variations were analyzed by Auger
spectroscopy in element distribution profiles of Al2O3 films
obtained by ALD and electron beam evaporation after
annealing at 800ºC and 1000ºC (Fig. 3).
a b
c d
Figure 3. Element distribution profile obtained by Auger spectroscopy in Al2O3 films after annealing at 800°С (a, b) and 1000°С
(c, d) generated by ALD (a, c) and electron beam evaporation (b, d).
It follows from the obtained results that with annealing
temperature increasing the Si(O) -78 eV content increases at the
interface of both sample types, which evidences continuous
oxidation of Si layer on the interface and thickness increasing of
this oxidized layer. Herewith, on ALD samples the thickness of
this SiOx layer at equal annealing temperatures is significantly
lower than on Al2O3 specimens deposited by electron beam
evaporation.
For more detailed profile analysis of the obtained layers and their
evolution upon rapid thermal annealing the ion fragmentation was
analyzed by time-of-flight secondary ion mass spectrometry.
Figure 4 illustrates the distribution profiles of ion fragments of
Al2O3 layers on Si substrate formed by ALD and electron beam
evaporation techniques before and after thermal processing at
800ºC and 1000ºC.
As follows from the obtained results, upon Al2O3 film deposition
by ALD the layer-substrate interface is sharper than in the case of
film deposited by electron beam evaporation (Fig. 4a,b).
Probably, this feature can be attributed to the fact that upon Al2O3
deposition by electron beam evaporation preliminary
cleaning procedure was applied which can lead to increase
amount of dangling bond imperfections on the substrate
surface and substrate heating upon deposition, which highly
increases the chemical activity of Si, thus, leading to Si
surface layer oxidation upon interaction with high energy
aluminum oxide particles. The obtained results are in good
agreement with the results of Auger spectroscopy. On the
other hand, in a samples formed by ALD it was possible to
detect Si-O-Al bonds on the transient layer. Existence of
such bonds can be attributed to possible chemical surface
reactions occurring upon contact between SiO2 and
Al(CH3)3 couples [18]:
Si-OH + Al(CH3)3 Si-O-Al(CH3)2 + CH4
2 Si-OH + Al(CH3)3 (Si-O)2=Al(CH3) + 2CH4
(Si-O)2 + Al(CH3)3 Si-O-Al(CH3)2 + Si-CH3
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 4 (2017) pp. 428-433
© Research India Publications. http://www.ripublication.com
432
which, probably, evidences incomplete decomposition of TMA
on substrate surface upon interaction with water at initial stages
of deposition.
Distributions of CH groups in both cases are nearly identical. The
SiO2 signal in ALD is, most probably, related with natural oxide
of substrate and is insignificant. The absence of the SiO2
signal in ALD Al2O3 in profile distributions obtained by
Auger spectroscopy is related with higher sensitivity of
ToF-SIMS. Similarly, there is no carbon signal obtained by
Auger spectroscopy in profile distributions of structures
formed by electron beam evaporation.
a b
c d
e f
Figure 4. Distribution profiles of ion fragments of Al2O3 layers deposited by ALD (a, c, e) and electron beam evaporation (b, d, f)
before (a, b) and after thermal processing at 800ºC (c, d) and 1000ºC (e, f).
International Journal of Applied Engineering Research ISSN 0973-4562 Volume 12, Number 4 (2017) pp. 428-433
© Research India Publications. http://www.ripublication.com
433
Upon rapid thermal annealing at 800ºC CH groups in a sample
obtained by ALD are localized, which is a consequence of
increased roughness of the deposited layer in comparison with
non-annealed specimen, and CH groups in a sample obtained by
electron beam evaporation are still uniformly distributed across
interface. Increasing of the silicon oxide layer thickness at the
interface upon thermal oxidation is also obvious.
Upon further annealing temperature increasing up to 1000ºC CH
groups at the interface of sample obtained by ALD are almost
completely absent, and the silicon dioxide signal propagates from
the transient layer into the depth of the substrate, which is most
probably related with peculiar features of ToF-SIMS analysis
method of the samples with highly rough surface (Fig. 2e). In
samples formed by electron beam evaporation CH groups remain
distribution across the analyzed surface on the transient layer
even upon annealing at 1000ºC with uniform. Continuous silicon
oxide layer thickness increasing is also observed in both sample
types.
CONCLUSIONS
Therefore, in current work peculiar features of such passivating
layers deposition procedures for micro- and functional electronics
as atomic layer deposition and electron beam evaporation has
been studied in details. These procedures has been compared
upon deposition of aluminum oxide thin passivating layers on
monocrystalline Si substrates before and after processing of
samples upon rapid thermal annealing (RTA) in Ar environment
at 800ºC and 1000ºC. Auger spectroscopy and time-of-flight
secondary ion mass spectrometry have been applied for profile
analysis of the obtained samples. The obtained results
demonstrate that the surface morphology of initial samples is
smooth. However, upon samples thermal processing the surface
morphology varies and becomes rougher. It is demonstrated that
upon Al2O3 deposition by electron beam evaporation Si substrate
already starts to oxidize. Such feature can put additional
contribution to optical and electrophysical properties of the
functional structures. In addition, existence of CH groups at
Al2O3-Si interface has been detected in both cases. As a
consequence of thermal processing Si is oxidized and SiOx
thickness under Al2O3 layer increases, however, when ALD is
used for deposition the oxidized silicon thickness is lower by
more than 1.5 times comparing to the sample obtained by electron
beam evaporation technique. In addition, upon analysis of
obtained samples by ToF-SIMS the signal of CH groups
disappear in the samples with aluminum oxide layer deposited by
ALD after annealing at 1000ºC contrary to EBE where CH
groups still existed in Al2O3/Si transient layer. Such impurities
can make additional uncontrolled contribution to the properties of
produced layers. Hence, it is demonstrated that ALD is one of the
most advanced deposition procedures of passivating and
antireflection functional layers in photovoltaics.
ACKNOWLEDGEMENTS
This work was supported by Federal targeted program titled
Development of science and technology in Russia for the years
2014–2020, Agreement No. 14.578.21.0019 (unique ID: ПНИ
REMEFI57814X0019).
REFERENCES
[1] T. Murayama, T. Nakai, Appl. Phys. Lett., vol. 58, p.
2079, 1991.
[2] A.C. Jones, D.J. Hoalton, S.A. Rushworth, J. De
Phys., 5, p. 557, 1995.
[3] M. Aguilar-Frutis, M. Garcia, C. Falcony, G. Plesch,
S. Jimenez-Sandoval, Thin Solid Films, vol. 389, p.
200, 2001.
[4] S. Mansour, G.N. Al-Robaee, K. Subbanna, N. Rao,
S. Mohan, Vacuum, vol. 45, no. 1, p. 97, 1994.
[5] S. Maniv, W.D. Westwod, J. Vac. Sci. Technol., vol.
17, pp. 743, 1980.
[6] S. Guha, E. Cartier, N.A. Bojarczuk, J. Bruley, L.
Gignac, J. Karasinski, J. Appl. Phys., vol. 90, no. 1, p.
512, 2001.
[7] B. Hoex, “Silicon surface passivation by atomic layer
deposited Al2O3,” J. Appl. Phys., vol. 104, p. 044903,
2008.
[8] P. Saint-Cast et al., “High-efficiency c-Si solar cells
passivated with ALD and PECVD aluminum oxide,”
IEEE Electron Device Lett., vol. 31, no. 7, pp. 695–
697, Jul. 2010.
[9] G. Agostinelli, A. Delabie, P. Vitanov, Z. Alexieva,
H.F.W. Dekkers, S.D. Wolf, G.Beaucarne, “Very low
surface recombination velocities on p-type silicon
wafers passivated with a dielectric with fixed negative
charge,” Sol. Energy Mater. Sol. Cells, vol. 90, pp.
3438–3443, 2006.
[10] B. Hoex, S.B.S. Heil, E. Langereis, M.C.M. van de
Sanden, W.M.M. Kessels, “Ultralow surface
recombination of c-Si substrates passivated by
plasma-assisted atomic layer deposited Al2O3,” Appl.
Phys. Lett., vol. 89, pp. 042112, 2006.
[11] J. Schmidt, A. Merkle, R. Brendel, B. Hoex, M.C.M.
van de Sanden, W.M.M. Kessels, “Surface
passivation of high-efficiency silicon solar cells by
atomic-layer-deposited Al2O3,” Prog. Photovolt. Res.
Appl., vol. 16, pp. 461–466, 2008.
[12] N. M. Terlinden, G. Dingemans, M. C. M. van de
Sanden, W. M. M. Kessels, Appl. Phys. Lett., vol. 96,
p. 112101, 2010.
[13] F. Werner, B. Veith, D. Zielke, L. Kuhnemund, C.
Tegenkamp, M. Seibt, R. Brendel, J. Schmidt, J.
Appl. Phys., vol. 109, p. 113701, 2011.
[14] G. Dingemans, W. M. M. Kessels, J. Vac. Sci.
Technol., vol. 30, p. 040802, 2012.
[15] B. Hoex, M. C. M. van de Sanden, J. Schmidt, R.
Brendel, W. M. M. Kessels, Phys. Status Solidi RRL,
vol. 6, p. 4, 2012.
[16] J. Benick, A. Richter, M. Hermle, S. W. Glunz, Phys.
Status Solidi RRL, vol. 3, p. 233, 2009.
[17] G. Dingemans, P. Engelhart, R. Seguin, F. Einsele, B.
Hoex, M. C. M. van de Sanden, W. M. M. Kessels, J.
Appl. Phys., vol. 106, p. 114907, 2009.
[18] I. Gorelkin “Atomic layer deposition as a way of
nanostructuring materials,” Current Research
Innovation: Science and Practice, no. 2, p. 12, 2010
(in Russian).