Vacuum 92 (2013) 44e51
Contents lists available at SciVerse ScienceDirect
Vacuum
journal homepage: www.elsevier.com/locate/vacuum
Post-selenization of stacked precursor layers for CIGS
Zs. Baji a, *, Z. Lábadi a, Gy. Molnár a, B. Pécz a, A.L. Tóth a, J. Tóth c, A. Csik c, I. Bársony a, b
a
Institute of Technical Physics and Matrials Science (MFA), Research Centre for Natural Sciences, Hungarian Academy of Sciences, P.O. Box 49, H-1525 Budapest, Hungary
Doctoral School of Molecular- and Nanotechnologies, Faculty of Information Technology, University of Pannonia, Egyetem u.10, H-8200 Veszprém, Hungary
c
Institute of Nuclear Research of the Hungarian Academy of Sciences (ATOMKI), P.O. Box 51, H-4001 Debrecen, Hungary
b
a b s t r a c t
In this study the possibility of the fabrication of CIGS layers from stacked precursors with selenization is
examined. Different sequences of precursor layers and two different selenization methods were applied,
in order to establish the optimal order of Cu, In and Ga layers in the precursor layer stack. The obtained
CIGS films were studied by different micro- and surface analysis methods (TEM, SEM, EDS, XRD, SNMS,
XPS). Since the evaporation of a Se layer and post-annealing does not result in a homogeneous CIGS layer,
the appropriate selenization must be accomplished in Se-vapour.
1. Introduction
Chalcopyrites are perhaps the most promising materials as
absorbers in thin film solar cells. The band gap of CuInSe2 is
Eg ¼ 1 eV, that of CuGaSe2 Eg ¼ 1.7 eV. Therefore, a manipulation of
the band gap between these two extremes, thus the preparation of
a wide variety of absorber materials is possible by simply adjusting
the Ga/In ratio [1].
CuInxGa1xSe2 (CIGS) films have been fabricated by a number of
methods. The most successful, but also the most complex technique
proved to be the co-deposition of all components from individual
sources [2]. This approach resulted in the highest reported
conversion efficiency solar cells with 20.3% [3]. On the other hand
this method also implies the most complex and expensive procedure, where an economical usage of the source materials is
impossible. Another disadvantage is that it requires an incredibly
well controlled processing technique [4].
Post-selenization of precursors is a low cost alternative that
does not require expensive apparatus. It can be achieved by the use
of Se vapour, diethylselenide or H2Se. The latter is highly toxic,
therefore its use rises environmental and health concerns and it is
very complicated to handle [5]. Copper, gallium and indium
precursors may be selenized to create CIGS material. In that case it
is of great advantage that no precise control of the parameters is
needed, as the metallic precursors only absorb the amount of
selenium required by the stoichiometry [6e8]. An additional
advantage of the technique is that Se diffuses into the Mo backcontact of the solar cell. Thereby MoSe2 is formed at the
boundary of the active layer and the back-contact, which leads to an
improved contact between the two materials [9].
The chemical reactions that occur during the selenization of
precursors may take a number of routes, which also depend on the
precursor structure. In the case of selenization of metallic copper,
indium and gallium depending on time and temperature a certain
order of the chemical reactions can be observed. First the binary
phases form, then the CuSe phase, and finally the CuInSe2 and
CuGaSe2 phases that react to form Cu(InGa)Se2 [10e12].
It was indicated [13,14] that the morphology of the as deposited
CIG layer may have a crucial role in the resulting morphology, thus
the overall quality of the final CIGS film after selenization. Both
evaporated and sputtered CIG layers are inhomogeneous with
a rough surface, which would deteriorate the quality of the
resulting films.
Another often mentioned disadvantage is that gallium tends to
diffuse near the back-contact at the high temperatures during
selenization, while the top layers become richer in indium. This
finally results in an adverse phase separation of CIS and CIGS
material, which reduces the device efficiency.
2. Motivation
* Corresponding author. Tel.: þ36 1 3922222; fax: þ36 1 3922226.
E-mail address: baji.zsofi[email protected] (Zs. Baji).
DOI: 10.1016/j.vacuum.2012.11.012
The purpose of our research is to study the CIGS material and the
method of post-selenization, as the detailed knowledge of the
Zs. Baji et al. / Vacuum 92 (2013) 44e51
processes that take place during layer preparation is prerequisite to
the fabrication of device quality absorber material. In our previous
studies [15] we examined flash evaporated and post-selenized
layers and found that even a rough morphology of precursor
metals results in homogeneous CIGS layers after selenization. On
the other hand these layers showed low mechanical stability and
adhesion to the substrate underneath. The aim of the present work
is to analyse whether sequentially evaporated precursors could also
result in homogeneous layers with perhaps improved mechanical
properties.
Another disadvantage of the flash evaporation method [15] was
that the thickness of the deposited precursor material was
<400 nm due to the limited size of the evaporation source, which
resulted in ca. 800 nm of CIGS material. In the present work we
attempted to determine how thick precursor layers could be selenized reliably and what the optimal selenization method is.
3. Experimental
CIGS layers were deposited by the selenization of metallic
precursors. To examine the effect of the order of the layers in the
stack we deposited them in all possible deposition sequences. Si
and glass substrates were used, both cleaned in cc.HNO3 and high
purity water before layer deposition.
The sputtering and evaporation took place in a flexible research
tool optimized for the manufacturing of Cu(InGa)Se2 based solar
cell modules. Mechanical movement, gas pressure and composition, DC electric supply and additional pulse parameters are
computer controlled.
Mo layers were deposited and target characteristics monitored
by pulsed DC magnetron sputtering (deposition chamber manufactured by Energosolar, Hungary), with 10 ms period length and
10% duty factor from a 114 440 mm2 sized metallic Mo target at
a 60 mm working distance. Under the target of the substrate was
moved at a speed of 50 mm/s bidirectionally. The electric supply
unit can work both in DC and adjustable pulse mode. Before
opening the Ar valves the pressure in the vacuum chamber was
typically 8$107 mbar. Depositions and target characterization
were made at 6$103 mbar working pressure (measured by
Edwards WRG e S type Gauge) with a total gas inlet of 50 sccm. The
target power in power controlled mode was 750 W and the target
voltage 300e315 V. We applied no substrate bias and no preheating.
Cu layers were deposited from a Cu target in the same magnetron sputtering equipment under the same conditions. The target
power was 250 W, the target voltage 290 V.
The temperature during In evaporation was T ¼ 1040e1060 C
and the pressure p ¼ 2$105 mbar, In the case of the Ga evaporation T ¼ 1115e1141 C and p ¼ 1.7$105 mbar. Table 1 summarizes
the prepared samples.
Two different approaches were applied for the selenization of
the consecutively deposited precursors. The first consisted of
evaporation of selenium on top of the layer stack and an
Mo
Mo
Mo
Mo
Mo
Mo
Mo
Mo
In
Ga
Cu
Cu
In
Ga
Cu
Cu
Ga
In
In
Ga
Ga
In
In
Ga
Cu
Cu
Ga
In
Cu
Cu
Ga
In
additional annealing so that the Se could diffuse into the layers,
and the precursors could transform into CIGS layer. This process
took place in a vacuum chamber designed for this purpose. The
background pressure was p w 106 mbar. The Se-source was
a resistivity heated W-boat. The estimated temperature of the
source was T ¼ 800 C. Se pellets were completely evaporated
during the ramp-up and hold period (1 minone minute each) at
the final boat temperature. Thereby the thickness of the “selenium source layer” on the differently stacked precursors is expected to be constant. The substrate temperature was T ¼ 430 C
during evaporation, followed by a 20 min annealing step at the
same temperature.
The other method involved the sealing of the samples into glass
ampoules evacuated to ca. 1 Pa and a subsequent annealing of the
precursors in Se atmosphere for 15 min at T ¼ 500 C, where the
partial pressure of Se is about 1 Pa.
The X-ray diffraction (XRD) measurements were performed
with Cu K radiation using a Bruker AXS D8 Discover horizontal Xray diffractometer equipped with Göbel mirror and a two dimensional GADDS detector system. The beam diameter was about
500 m.
The scanning electron microscopy (SEM) and energy dispersive
spectroscopy (EDS) measurements were carried out in a LEO 1540
XB microscope equipped with a Schottky field emission gun,
Gemini lens system and in-lens detector.
The depth profiles were measured with an INA-X type
secondary neutral mass spectroscopy (SNMS) equipment (product
of Specs GmbH, Berlin). For the sample bombardment and post
ionization Ar plasma was used.
Transparent specimens for transmission electron microscopy
(TEM) investigations were prepared by mechanical cutting and
polishing, followed by Ar þ ion milling at 10 keV ion energy. Cross
sectional TEM samples were inserted into special Ti discs of 3 mm,
and thinned from both sides. In order to remove the thin surface
layer damaged during high energy ion milling from both sides of
the specimens, the process was finished at 2.5e3 keV ion energy.
Further details about sample preparation can be found in [16].
Conventional electron microscopy was carried out in a Philips CM
20 microscope at 200 keV, which is equipped with a NORAN EDS.
The XPS measurements were performed using conventional
excitation of non-monocromated Al X-rays. The XPS spectra originated from a few representative depth point of the CIGS layers
etched in the SNMS plasma. The specimens from the SNMS
measurement chamber to the XPS measurement chamber were
moved in vacuum. The XPS machine (product of SPECS, Germany)
was working in the fixed analyser transmission mode. The chemical
states of the CIGS layers were assigned by the help of the XPS
binding energy database from the National Institute of Standards
and Technology (USA). The XP spectrometer energy scale was
calibrated with the relative calibration method measuring Cu, Au
and Ag XPS lines and using standard binding energy values determined at the National Physical Laboratory (UK).
4. Results and discussion
4.1. Precursor morphology after evaporation
Table 1
Summary of sample parameters.
1
2
3
4
5
6
7
8
45
Selenization
Selenization
Selenization
Selenization
Selenization
Selenization
Selenization
Selenization
with
with
with
with
with
with
with
with
Se
Se
Se
Se
Se
Se
Se
Se
evaporation
evaporation
evaporation
evaporation
vapour
vapour
vapour
vapour
and
and
and
and
annealing
annealing
annealing
annealing
Fig. 1 shows the SEM micrographs (left) and the Energy
Dispersive Spectra, the EDS-maps of the precursors (right). It is
evident on almost all samples that the precursors completely
covered the substrates. The average grain size of the samples is
100e200 nm. The importance of the deposition order of the
precursor metals is revealed by the comparison of the micrographs.
The first row in Fig. 1 shows the morphology of sample 1. The
surface is scattered with 0.5e1 mm sized islands. On a larger scale
46
Zs. Baji et al. / Vacuum 92 (2013) 44e51
Fig. 1. SEM micrographs at 5000 magnification (for sample 1) and 50000 magnification (for the other samples) as well as element maps of the evaporated precursors for
samples 1, 2, 3 and 4, in the respective rows.
(see the elemental map beside), 10 mm sized islands can be found
on the surface. From the component analysis the smaller objects
appear to be In-grains, whereas the larger ones Ga-droplets. By
looking at the origin of this morphology we have to consider that
In, when evaporated, forms droplets on the substrate surface. In the
case of this sample the Ga was evaporated afterwards on this
surface, and finally the Cu was sputtered. During sputtering the
substrate temperature exceeded the melting point of Ga, which
then diffused to the top of the layers. Thus, although the Cu sputtering took place after the Ga evaporation, the gallium droplets can
be found on top of the evenly sputtered copper layer. This applies to
the cases where the evaporations were followed by the sputtering.
This effect may be even of advantage, as a higher Ga content in the
top layer of the film results in an ideal band gap profile with a wider
band gap at the active region and a Ga rich layer near the surface.
The SEM results of sample 2 can be seen in the 2nd row of Fig. 1.
Once again the Ga droplets are apparent, but no In islands could be
found according to the larger scale image.
The morphology of sample 3 is presented in the 3rd row. In this
case the In layer was evaporated on top of the sputtered copper
layer. This time the Ga was evaporated last, therefore the
morphology is slightly different, the Ga-islands are smaller.
Zs. Baji et al. / Vacuum 92 (2013) 44e51
The only sample where no Ga droplets could be found is sample
4 (4th row). The sole inhomogeneities in this layer are the mm size
In grains. The elemental distribution of the sample reveals
extended regions of In-deficiency.
4.2. Post-annealing of stacked precursors
The SNMS results of the layers can be seen in the first column of
Fig. 2. The results must be interpreted with some caution though, as
due to the rough surface morphology of the layers the intensities
from - at least the surface layers - may be somewhat misleading.
This is also the reason why no depth scale can be calculated from
these results.
In all cases the increase of the Mo-yield indicates the bottom of
the precursor layer. From the depth profiles the following
47
observations can be made: The layers of the precursor metals are
still visible in the original order of the deposition. Therefore, the
metals did not mix with each other and the alloy formation was
negligible at the temperature of the selenization. On the other
hand, as the Se diffused into the layer, a mixing of the components
started, which can be concluded from the way the In layer always
diffuses towards the surface. The In concentration rapidly increases
in the Se-containing top layer. Therefore, the local minima in the In
depth profile indicate the chemical reactions.
Comparing these results with the cross sectional SEM images in
the right column of Fig. 2, the thickness of the selenized CIGS layers
and that of the unreacted precursor film, which remained underneath, may be defined.
For sample 1 this means 1.4e1.7 mm thick CIGS on top of the
500e550 nm thick unreacted metals. In sample 2: 1e1.3 mm CIGS
Fig. 2. SNMS depth profiles (left) and corresponding SEM cross sections (right) of the samples selenized with Se evaporation.
48
Zs. Baji et al. / Vacuum 92 (2013) 44e51
on 550e600 nm metallic layer. Sample 3 is the most uneven layer,
with 1 mm high CIGS islands on a w500 nm thick CIGS layer. There
is also a ca. 300 nm thick metallic layer visible underneath. Sample
4: w1 mm thick CIGS on w400 nm precursor layer.
It is evident that the selenization was more effective in the
samples where the Cu was at the top. Samples 3 and 4 exhibit
a much rougher morphology, with more unreacted precursors
remaining at the bottom.
The XRD results are shown in Fig. 3, and the main characteristics
of the layers are summarized in Table 2.
In conclusion, besides the phases CIGS and CIS, selenides were
formed and a portion of the metallic precursors remained
unreacted at the bottom of the layer stack as evidenced by the SEM
micrograph in Fig. 2. This can probably be ascribed to the Sedeficiency, i.e. to the insufficient Se evaporated on the layers.
Moreover, at the annealing temperature a portion of this unreacted
Se - instead of diffusing into it - probably even evaporates from the
layer stack.
4.3. Selenization of precursors in Se vapour
The composition of the layers according to the EDS analysis is
shown in Table 3.
Fig. 4 shows SEM micrographs of the samples after the selenization. All the samples became laterally uniform in composition.
Samples 5 and 6 have a generally similar morphology (see Fig. 4a),
the circular (“cauliflower-like”) grains usually observed on CIGS
films cannot be found here. This could result in a denser, more
uniform surface, which is generally considered more appropriate
and yields a better efficiency. Samples 7 and 8 (Fig. 4b) have
a similar morphology, except for the hexagonal crystallites scattered all over the surface.
Fig. 5 presents the SNMS results and the cross sectional SEM
images of the samples. All the layers are built up from two sublayers
of different morphologies: a bottom layer with a finer structure,
and a top layer with larger grain size. In the films where the Cu was
sputtered first (i.e. sample 7 and 8 in 3rd and 4th row of Fig. 5.) this
difference is even more prominent. According to the EDS analysis
the bottom sublayer is richer in Ga. On the other hand the SNMS
depth profiles (left column in Fig. 5.) show that the Se diffused
throughout the full depth of all the layers, although the depth
Table 2
Phases identified by XRD in the samples 1e4.
Phases found by XRD
1
2
3
4
CuIn0.7Ga0.3Se2 and CuInSe2 phases with metallic In
CuInSe2 phase dominates, with some CuIn0.7Ga0.3Se2, and some
metallic In
CuInSe2 phase with (213) orientation apparent, but with small
grains, and some Cu and In.
CuInSe2 phase with In and Cu
profiles are not homogeneous. For an understanding of the background of this phenomenon, the XRD results need to be analysed.
The XRD spectra (in Fig. 6.) revealed that all the layers contained
chalcopyrite phase, some also had one of the metallic precursors in
excess. A summary of the crystalline quality of the layers is given in
Table 4.
The absence of binary phases and the CuInGaSe2 phase in case of
the samples 5 and 6 indicates that the selenization reaction was
complete. The two different layers apparent in the cross sectional
images may be the CIS and the CIGS phases, therefore a phase
separation must have taken place with the CIS at the top and the
CIGS dominantly at the bottom of the layer stack.
In samples 7 and 8, however, CuSe is still present. It is he
hexagonal phase CuSe that forms the crystallites occurring in the
SEM micrographs in Fig. 4b. This was also supported by EDS
elemental analysis. Since in these samples copper was the bottom
layer in the layer structure, which is probably why some CuSe was
left in the samples. Therefore, this layer structure is either less
favourable, or it requires a longer annealing time. Sample 8 also
reflected a separation of the CIGS and CIS phases, which could also
be seen from EDS analysis. It is obvious from the XRD results that
where Ga is lying deeper in the layer stack than In, the phase
separation is more pronounced.
Transmission electron microscopy was performed on samples 5
and 7. The TEM results of sample 7 are presented in Fig. 7, the
results of sample 5 were very similar to these. Fig. 7 a shows an
overview of the grown layer in a bright field image. The insert is
a selected area diffraction pattern in which diffraction rings originating from a polycrystalline material are visible. This is clearly
identified as the sputtered Mo layer on the substrate. The layer
exhibits a typical columnar structure, otherwise it is uniform in
thickness and composition. Fig. 7b is a dark field image, revealing
the large crystallites of the CIGS layer. The micrometre sized grains
form a relatively rough surface, which to a certain extent can even
be advantageous in solar cell applications. Here we have to note,
that apparently some parts of the layer were sputtered off by the Ar
ions during ion beam thinning. In the insert in Fig. 7a the individual
spots close to the middle of the image and their regular network
represent large CIGS grains. These diffraction patterns and the
lattice spacing calculated from them fit the CIGS phase (JCPDS card
40e1488) very well. The closest diffraction spot to the 000 is the
112 reflection (this is the most intensive diffraction line of this
phase) of the CIGS phase. Fig. 7 b is the dark field image taken with
Table 3
EDS elemental composition of samples 5e8 along with the stoichiometric
composition.
Fig. 3. XRD spectra of the CIGS layers selenized by Se evaporation and annealing
(samples 1e4).
Sample no./component
Cu (at%)
In (at%)
Ga (at%)
Se (at%)
1
2
5
8
Ideal composition
35.7
35.5
26.9
16.5
25
15.5
15.7
17.8
29.5
17.5e20
5.1
2.6
6.6
2.5
5e7.5
43.7
46.1
48.8
51.5
50
Zs. Baji et al. / Vacuum 92 (2013) 44e51
49
Fig. 4. Typical SEM morphology of samples 5 and 6 (a) at 2000 magnification, and that of samples 7 and 8 (b) at 20 000 magnification.
Fig. 5. SNMS depth profiles (left) and corresponding SEM cross sections (right) of the samples selenized in Se vapour: samples 5, 6, 7 and 8 shown in the rows 1e4, respectively.
50
Zs. Baji et al. / Vacuum 92 (2013) 44e51
Table 5
XPS binding energies of Se, In and Cu measured on sample 8 and on reference
foreign CIGS samples.
Sample 8
Reference foreign CIGS
sample (own
measurement)
Ref. [16] CuInSe2
Cu metal (own
measurements)
In metal
Se
Se 3d
In 4d 5/2
Cu 2p 3/2
54.4 eV
54.45 eV
444.5 eV
444.6 eV
932.5 eV
e
54.3e54.5 eV
444.6e444.8 eV
932.1e932.6 eV
932.35 eV
443.9 eV
55.2 eV
In conclusion, by this selenization method more homogeneous
depth profiles and CIGS layer formation could be registered.
5. Conclusions
Fig. 6. XRD spectra of the CIGS layers selenized in Se vapour (samples 5, 6, 7 and 8).
that reflection showing one of the large CIGS grains with bright
contrast. Not only the diffraction pattern, but also the EDS analysis
carried out on the same sample in the same microscope confirmed
the formation of CIGS phase.
Fig. 7c is a dark field image of the Mo layer taken with the 110
reflection of Mo (corresponding to the first ring of the insert in
Fig. 7a). The thin layer observed above the Mo and below the CIGS
was found by EDS to be a gallium rich CIGS layer.
Two different post-selenization methods on the consecutively
evaporated components of the quaternary CIGS structure were
compared. We found that the evaporation of Se and the subsequent
annealing is not sufficient to complete all the reactions needed to
result in a homogeneous CIGS layer. Therefore, the appropriate
selenization must be performed by an annealing in Se-vapour.
Layers with Cu on top provide better CIGS compositions with
both selenization methods. The reason for that is probably that in
a Cu rich environment both In and Ga diffuse faster [10].
It is interesting to note, that with In atop of Ga more CIS is
present. On the other hand the CGS phase was never identified in
the layers, not even with Ga on top. This is explained in all cases by
the diffusion of In towards the Se rich surface, resulting in a more
homogeneous precursor structure.
We determined the optimum sequence in the deposition of the
precursor metals. Cu sputtering as top layer is most favourable for
subsequent selenisation, as to some extent it ensures a mixing of
the layers. During sputtering of Cu namely Ga-outdiffusion can take
place, on the other hand In has to be covered by Ga in the layer
stack. Therefore the optimal sequence is: In, Ga, Cu followed by
post-selenization in Se vapour.
4.4. X-ray photoelectron spectroscopy (XPS) analysis of the samples
Acknowledgements
XPS binding energies were evaluated using the NIST XPS binding
energy database and summarized in Table 5. The accuracy of our
own measurement is in the range of 0.1 eV (similar to the accuracy of [17]). The XPS binding energies are in good agreement with
the XRD data and prove the presence of CuInSe2 and CuIn0,7Ga0.3Se2
phases in sample 8.
The help of Zs.E. Horváth of MFA with XRD and K. Vad of MTA
ATOMKI with SNMS is gratefully acknowledged. The authors wish
to thank for the support of the Hungarian National Science Fund
OTKA by the grant No. NK73424. This work was also supported by
the National Development Agency grant TÁMOP-4.2.2/B-10/12010-0025.
Table 4
Phases in the samples 5e8 identified by XRD.
Phases found by XRD
5 Predominantly CuIn0.7Ga0.3Se2 with a little CuInSe2 phase present.
6 Both CuInSe2 and CuIn0.7Ga0.3Se2 phases present, the CuInSe2 phase is more
dominant.
7 Only one chalcopyrite phase present, that of CuIn0.9Ga0.1Se2 with some
hexagonal CuSe.
8 CuInSe2 and CuIn0.7Ga0.3Se2 phases with a little hexagonal CuSe present.
Fig. 7. Cross sectional TEM images taken from sample 7: bright field image with an insert of the diffraction pattern (a), a dark field image taken with the 112 reflection of CIGS (b)
dark field image taken with the Mo 110 reflection (c).
Zs. Baji et al. / Vacuum 92 (2013) 44e51
References
[9]
[1] Klenk R, Lux-Steiner MC. Chalcopyrite based solar cells. In: Poortmans J,
Arkhipov V, editors. Thin film solar cells. Wiley; 2006.
[2] Powallaa M, Dimmler B. Development of large-area CIGS modules. Sol Energy
Mater Sol Cells 2003;75:27e34.
[3] Jackson P, Hariskos D, Lotter E, Paetel S, Wuerz R, Menner R, et al. New world
record efficiency for CIGS thin film solar cells beyond 20%. Prog Photovolt: Res
Appl 2011;19:894e7.
[4] Klenk M, Schenker O, Alberts V, Bucher E. Preparation of device quality
chalcopyrite thin films by thermal evaporation of compound materials. Semic
Sci Technol 2002;17:435e9.
[5] Li W, Sun Y, Liu W, Zhou L. Fabrication of Cu(InGa)2 thin films solar cell by
selenization process with Se vapor. Sol Energ 2006;80:191e5.
[6] Ganchev M, Kois J, Kaelin M, Bereznev S, Tzvetkova E, Volobujeva O, et al.
Preparation of Cu(In, Ga)Se2 layers by selenization of electrodeposited CueIne
Ga precursors. Thin Solid Films 2006;511-512:325e7.
[7] Caballero R, Maffiotte C, Guillén C. Preparation and characterization fo
CuIn1xGaxSe2 thin films obtained by sequential evaporations and different
selenization processes. Thin Solid Films 2005;474:70e6.
[8] Chen GS, Yang JC, Chan YC, Yang LC, Huang W. Another route to fabricate
single-phase chalcogenides by post-selenization of CueIneGa precursors
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
51
sputter deposited from a single ternary target. Sol Energy Mater Sol Cells
2009;93:1351e5.
Abou-Ras D, Kostorz G, Bremaud D, Kälin M, Kurdesau FV, Tiwari AN, et al.
Formation and characterisation of MoSe2 for Cu(In, Ga)Se2 based solar cells.
Thin Solid Films 2005;480e481:433e8.
Lundberg O, Lu J, Rockett A, Edoff M, Stolt L. J Phys Chem Solids 2003;64:1499.
Kim WK, Payzant EA, Kim S, Speakman SA, Crisalle OD, Anderson TJ. J Cryst
Growth 2008;310:2987.
Kim WK, Payzant EA, Anderson TJ, Crisalle OD. Thin Solid Films 2007;215:
5837.
Su CY, Ho WH, Lin HC, Nieh CY, Liang SC. The effects of the morphology on the
CIGS thin films prepared by CuInGa single precursor. Sol Energy Mater Sol
Cells 2011;95:261e3.
Liu W, Tian J-G, He Q, Li F-Y, Sun Y. Effect of metallic precursors on the thin
film thickness and reaction resistances in the selenization process. Curr Appl
Phys 2010. http://dx.doi.org/10.1016/j.cap.2010.07.028.
Baji Zs, Lábadi Z, Molnár Gy, Tóth AL, Bársony I. CIGS films by flash evaporation and post selenization, in press.
Barna Á, Radnóczi G, Pécz B. In: Amelinckx S, van Dyck D, van Landuyt J, van
Tendelo G, editors. Handbook of microscopy, vol. 3. VCH Verlag; 1997. p. 751e
801. Chapter II/3.
Kohiki S, Nishitani M, Negami T, Wada T. X-ray photoelectron spectroscopy of
CuInSe2. Phys Rev B 1992;45:9163.