Surface passivation of Cu(In,Ga)Se2 using atomic layer deposited Al2O3

Article abstract of DOI:10.1063/1.3675849

With Al₂O₃ passivation on the surface of Cu(In,Ga)Se₂, the integrated photoluminescence intensity can achieve two orders of magnitude enhancement due to the reduction of surface recombination velocity. The photoluminescenc

Full text of DOI:10.1063/1.3675849

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Surface passivation of Cu(In,Ga)Se2 using atomic layer deposited Al2O3
W.-W. Hsu, J. Y. Chen, T.-H. Cheng, S. C. Lu, W.-S. Ho, Y.-Y. Chen, Y.-J. Chien, and C. W. Liu
Citation: Applied Physics Letters 100, 023508 (2012); doi: 10.1063/1.3675849
View online: http://dx.doi.org/10.1063/1.3675849
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/100/2?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Charge carrier dynamics and recombination in graded band gap CuIn1−xGaxSe2 polycrystalline thin-film
photovoltaic solar cell absorbers
J. Appl. Phys. 114, 154505 (2013); 10.1063/1.4825211
Arsenic decapping and pre-atomic layer deposition trimethylaluminum passivation of Al2O3/InGaAs(100)
interfaces
Appl. Phys. Lett. 103, 071602 (2013); 10.1063/1.4818330
Changes in the dominant recombination mechanisms of polycrystalline Cu ( In,Ga ) Se 2 occurring during growth
J. Appl. Phys. 94, 5584 (2003); 10.1063/1.1613804
Band bending independent of surface passivation in ZnO/CdS/Cu ( In,Ga )( S,Se ) 2 heterojunctions and Cr/Cu (
In,Ga )( S,Se ) 2 Schottky contacts
Appl. Phys. Lett. 82, 3559 (2003); 10.1063/1.1576500
Microstructure of surface layers in Cu(In,Ga)Se 2 thin films
Appl. Phys. Lett. 81, 1008 (2002); 10.1063/1.1498499
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: 129.2.19.100
On: Thu, 25 Dec 2014 06:32:10

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
APPLIED PHYSICS LETTERS 100, 023508 (2012)
Surface passivation of Cu(In,Ga)Se2 using atomic layer deposited Al2O3
W.-W. Hsu,1 J. Y. Chen,1 T.-H. Cheng,2 S. C. Lu,2 W.-S. Ho,1 Y.-Y. Chen,1 Y.-J. Chien,3
and C. W. Liu4,a)
1Graduate Institute of Electronics Engineering and Department of Electrical Engineering,
National Taiwan University, Taipei, Taiwan
2Graduate Institute of Photonics and Optoelectronics and Department of Electrical Engineering,
National Taiwan University, Taipei, Taiwan
3Advanced Technology Department, TF RD & System Dev. Div. AU Optronics Corporation, Taichung, Taiwan
4Graduate Institute of Electronics Engineering, Graduate Institute of Photonics and Optoelectronics,
Department of Electrical Engineering, Center for Condensed Matter Sciences, and Center for Emerging
Material and Advanced Devices, National Taiwan University, Taipei, Taiwan and National Nano Device
Laboratories, Hsinchu, Taiwan
(Received 28 October 2011; accepted 17 December 2011; published online 11 January 2012)
With Al2O3 passivation on the surface of Cu(In,Ga)Se2, the integrated photoluminescence intensity
can achieve two orders of magnitude enhancement due to the reduction of surface recombination
velocity. The photoluminescence intensity increases with increasing Al2O3 thickness from 5 nm to
50 nm. The capacitance-voltage measurement indicates negative fixed charges in the film. Based
on the first principles calculations, the deposition of Al2O3 can only reduce about 35% of interface
defect density as compared to the unpassivated Cu(In,Ga)Se2. Therefore, the passivation effect is
mainly caused by field effect where the surface carrier concentration is reduced by Coulomb
C
V
repulsion. 2012 American Institute of Physics. [doi:10.1063/1.3675849]
The Cu(In,Ga)Se2 (CIGS) thin film solar cell has a
potential to reduce production cost for photovoltaic modules.
The record efficiency of the CIGS thin film solar cells is
ꢀ20%.1 To solve the additional surface recombination at the
edge of the damage area by laser scribing,2,3 a good passiva-
tion layer for CIGS is necessary. In this letter, atomic layer
deposited (ALD) Al2O3 is investigated to passivate the
CIGS, and the effect of the passivation is measured by pho-
toluminescence (PL). The effectiveness of passivation is due
to both the low interface defect density (Qit) at the CIGS/
Al2O3 interface and the negative fixed charges (Qf) in bulk
Al2O3 for field effect passivation.
is multiplied by a factor of 100. The PL emission shoulder at
1.15 eV is due to the band-impurity (BI) recombination
between electrons in the conduction band and holes at
acceptor levels4 (the inset of Fig. 1). The emission peak at
1.09 eV is due to the recombination from donor-to-accepter
pair (DAP) between donor levels and acceptor levels.4,5 The
Al2O3 passivated CIGS have 200 times integrated PL inten-
sity of the unpassivated CIGS due to the low effective sur-
face recombination velocity (Seff). Al2O3 can serve as a good
passivation layer for CIGS films.
The correlation between the Seff and the PL intensity is
simulated using 50 ns and 250 ns lifetime of bulk CIGS,
which covers typical reported lifetime of CIGS films
(Fig. 2).6–8 The electron and hole concentration profiles of
the CIGS films are simulated under one sun illumination
CIGS films with a thickness of ꢀ2 lm were deposited
on ꢀ1 lm Mo on glass by coevaporation at 500 ꢁC. The
Al2O3 layers were deposited by the ALD at 250 ꢁC, lower
than the CIGS deposition temperature. The alternating pulses
of trimethyl aluminum (Al(CH3)3, TMA) and H2O in an Ar
carrier gas were used to deposit the Al2O3 layers. Each cycle
includes TMA for 0.03 s, Ar purge for 10 s, H2O for 0.06 s,
and Ar purge for 10 s. The thickness of the deposited Al2O3
film was determined by the number of ALD cycles. The
chamber of the Al2O3 deposition is different from the cham-
ber of the CIGS deposition. During the transportation, a vac-
uum bag is used to minimize the exposure in air. The PL
spectra of the same sample were measured before and after
the Al2O3 deposition by the lock-in technique with the laser
excitation wavelength of 671 nm and the pumping power of
100 mW. Note that the Al2O3-deposited CIGS has longer air
exposure time than the as-grown CIGS.
Figure 1 shows the PL spectra at room temperature of
the coevaporated CIGS film with and without Al2O3 (50 nm)
passivation, where the PL signal of the unpassivated sample
FIG. 1. (Color online) PL spectra of coevaporated CIGS films with and
without 50 nm thick Al2O3 passivation at room temperature. The Al2O3 pas-
sivated CIGS has 200 times integrated PL intensity of the unpassivated
CIGS. The inset shows the recombination mechanisms.
a)Author to whom correspondence should be addressed. Electronic mail:
chee@cc.ee.ntu.edu.tw.
C
V
0003-6951/2012/100(2)/023508/3/$30.00
100, 023508-1
2012 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: 129.2.19.100
On: Thu, 25 Dec 2014 06:32:10

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
023508-2 Hsu et al.
Appl. Phys. Lett. 100, 023508 (2012)
with the most representative density of the amorphous
oxide13 (Fig. 3(a)) in our calculation. Fig. 3(b) shows the simu-
lated DOS at the surface layer of the CIGS for unpassivated
CIGS and the Al2O3 passivated CIGS. Al2O3 passivation can
reduce about 35% of the Dit as compared to unpassivated sur-
face. Considering the reduction of the Dit effect only, the Seff
can be only reduced about 35% approximately, which is not
sufficient to explain the experimental enhancement.
Al2O3 is well-known to have excess defects with nega-
tive fixed charges in oxygen-rich condition, while to have
excess defects with positive fixed charges in aluminum-rich
condition.12 As shown in the inset of Fig. 4, a positive flat
band shift of the capacitance-voltage curve from a 10 nm
Al2O3 device to a 50 nm Al2O3 device is observed, indicat-
ing negative fixed charge in bulk Al2O3. In Figure 4, the cor-
relation between Seff and Qf with Dit ¼ 2 ꢂ 1012 and Dit ¼ 4
ꢂ 1012 cmꢃ2 is simulated. The energy distribution of inter-
face trap density, electron capture cross section rn, and hole
capture cross section rp are taken from Refs. 14 and 15. The
recombination rate (Us) is obtained by an integration over
the band gap,
FIG. 2. (Color online) Simulated Seff and the PL intensity using 50 and 250
ns lifetime of bulk CIGS. Each curve is normalized with respect to the Seff
of the unpassivated surface, which is assumed to be 1 ꢂ 106 cm/s.
(100 mW/cm2, AM 1.5 G), and the radiative recombination
is proportional to the product of the electron and hole con-
centrations.8 The parameter used in the simulation such as
refractive index, extinction coefficient, mobility, effective
density of state (DOS), and dielectric constant are adapted
from Refs. 9 and 10. Each curve is normalized with respect
to the Seff of the unpassivated surface, which is assumed to
be 1 ꢂ 106 cm/s.8,11 At Seff larger than 104 cm/s, the PL in-
tensity is independent on life time (sb), while the dependence
is stronger for Seff smaller than 104 cm/s. With the deposition
of 50 nm thickness Al2O3, the PL intensity enhances by 200
times as compared to the unpassivated surface. Our calcula-
tion indicates the Seff is reduced to 14 to 44 cm/s for 50 nm
Al2O3 passivation using sb ¼ 50 ns to 250 ns.
Us ¼ ðnsps ꢃ n2i Þmth
ð
Ec
DitðEtÞ
ꢂ
dEt: (1)
ðns þ n1Þ=rpðEtÞ þ ðps þ p1Þ=rnðEtÞ
Ev
The definition of each term in Eq. (1) is referred to Ref. 16.
Seff is then determined by Us/Dn, where Dn is the light gener-
ated excess minority concentration. Seff decreases with
increasing fixed charge density for both negative charge and
positive charge. With negative fixed charge smaller than 1
ꢂ 1012 cmꢃ2, the field effect passivation is not effective,
while the dependence is stronger for negative fixed charge
larger than 1 ꢂ 1012 cmꢃ2. Since rp of 6 ꢂ 10ꢃ18 cm2 is
smaller than rn of 5 ꢂ 10ꢃ16 cm2, the Seff at positive Qf is
smaller than that at negative Qf with the same charge density.
The negative charge density of 50 nm Al2O3 on CIGS is larger
than 3 ꢂ 1012 cmꢃ2 to have Seff ¼ 14 to 44 cm/s (Fig. 4).
The negative fixed charge is consistent with the case of
To have a clear picture of interface defect density (Dit)
and field effect on Al2O3 passivation, the DOS is calculated
using first principal calculation. Al2O3 has a variety of
phases, but is amorphous in Al2O3/CIGS structure. Simula-
tion of amorphous phases using finite supercells is difficult
due to the uncertainty of structures and the lack of long-
range order.12 For simplicity, we have chosen j-phase Al2O3
FIG. 3. (Color online) (a) The structure used in
the first principle calculation. (b) The calculated
density of states of the unpassivated CIGS and
the Al2O3/CIGS interface. The Dit of passivated
CIGS is 35% lower than the surface state den-
sity of the unpassivated CIGS.
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: 129.2.19.100
On: Thu, 25 Dec 2014 06:32:10

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
023508-3 Hsu et al.
Appl. Phys. Lett. 100, 023508 (2012)
FIG. 4. (Color online) The simulated Seff as a function of fixed charge. Seff
decreases with the increasing fixed charge density for both negative charge
and positive charge. The inset shows the capacitance voltage measurement
of CIGS with Al2O3 thickness of 10 nm and 50 nm.
FIG. 5. (Color online) The integrated PL intensity as a function of Al2O3
thickness and annealing time. The PL intensity after N2 and forming gas
annealing is the same as the original CIGS.
This work is partially supported by National Science
Council of ROC under Contract No. 100-3113-E-002-011.
The support of AU Optronics Corp. (AUO) is also highly
appreciated.
Al2O3 deposited on Si.12 Note the incremental fixed charge
density from 10 nm Al2O3 to 50 nm Al2O3 is about 1 ꢂ 1012
cmꢃ2 from the CV measurement.
Fig. 5 shows the integrated PL intensity for different
thickness of Al2O3 and annealing time with nitrogen and
forming gas. The integrated PL intensity increases with
increasing thickness of Al2O3, indicating the larger density
of fixed charge in thicker Al2O3 has better passivation. The
effects of thermal budget (250 ꢁC, 34 min–168 min) in N2
and forming gas are also investigated. The PL intensity after
N2 and forming gas annealing with the same thermal budget
has the same intensity as the unpassivated CIGS. Therefore,
the Al2O3 is the root of cause responsible for the surface
passivation.
1I. Repins, M. A. Contreras, B. Egaas, C. DeHart, J. Scharf, C. L. Perkins,
B. To, and R. Noufi, Prog. Photovoltaics. 16, 235 (2008).
2K. Matsunaga, T. Komaru, Y. Nakayama, T. kume, and Y. Suzuki, Sol.
Energy Mater. Sol. Cells 93, 1134 (2009).
3M. Powalla and B. Dimmler, Thin Solid Films 361, 540 (2000).
4R. Bacewicz, P. Zuk, and R. Trykozko, Opto-Electron. Rev. 11, 277 (2003).
5S. B. Zhang, S.-H. Wei, and A. Zunger, Phys. Rev. B 57, 9642 (1998).
6W. K. Metzger, I. L. Repins, and M. A. Contreras, Appl. Phys. Lett. 93,
022110 (2008).
7S. Shirakata and T. Nakada, Thin Solid Films 515, 6151 (2007).
8W. K. Metzger, I. L. Repins, M. Romero, P. Dippo, M. Contreras, R.
Noufi, and D. Levi, Thin Solid Films 517, 2360 (2009).
9S.-H. Han, F. S. Hasoon, A. M. Hermann, and D. H. Levi, Appl. Phys.
Lett. 91, 021904 (2007).
Surface passivation of CIGS using ALD Al2O3 is dem-
onstrated for the first time. Even with the thickness of only 5
nm, the passivation effect is still observable. The effective-
ness of passivation is mainly due to the negative fixed
charges in bulk Al2O3 for field effect passivation. Al2O3 can
be used for the edge passivation of CIGS module after p3
scribing to reduce the surface recombination velocity at
edge. This can increase the shunt resistance, short circuit cur-
rent, and open circuit voltage of the cell based on our
simulation.
10M. Gloeckler, Ph.D. thesis, Colorado State University, 2002.
11K. Bothe, G. H. Bauer, and T. Unold, Thin Solid Films 403, 453 (2002).
12J. R. Weber, A. Janotti, and C. G. Van de Walle, J. Appl. Phys. 109,
033715 (2011).
13D. R. Jennison, P. A. Schultz, and J. P. Sullivan, Phys. Rev. B 69, 041405
(2004).
14J. T. Heath, J. D. Cohen, and W. N. Shafarman, J. Appl. Phys. 95, 1000
(2004).
15M. Igalson, M. Bodegard, and L. Stolt, J. Phys. Chem. Solids 64, 2041
(2003).
16A. G. Aberle, S. Glunz, and W. Warta, J. Appl. Phys. 71, 4422 (1992).
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: 129.2.19.100
On: Thu, 25 Dec 2014 06:32:10