Yan 2014
Yan 2014
Yan 2014
(2014) 22(1):37–44
DOI 10.1007/s40534-014-0035-1
Received: 18 November 2013 / Revised: 17 January 2014 / Accepted: 18 January 2014 / Published online: 23 February 2014
Ó The Author(s) 2014. This article is published with open access at Springerlink.com
123
38 Y. Yan et al.
throughout the fabrication process. This is a critical chal- penetration depth calculated by Monte Carlo simulation
lenge for large-scale mass manufacture. Recently, efforts was approximately 300 nm. Surface morphologies and
have been made to fabricate CIGS layers by sputtering. To cross-sectional images were obtained by scanning electron
further reduce the fabrication costs, a one-step sputtering microscopy (SEM, JOEL FESEM 7001). Optical trans-
has been developed [7–10] to deposit CIGS films without missions were measured at the room temperature using a
selenization. Frantz et al. [7] prepared CIGS solar cell with UV/VIS/NIR spectrophotometer (Perkin Elmer LAMBDA
an efficiency of up to 8.9 % through one-step radio fre- 900) in the wavelength range of 600–2,500 nm in 1-nm
quency (RF) sputtering at Tsub of 550 °C. Therefore, the steps. Raman scattering was performed on a HORIBA
validity of one-step sputtering CIGS films at low Tsub needs LabRAM HR Raman spectrometer at the room tempera-
to be investigated. ture. The 633-nm laser line was used as the excitation light
In our previous study [10], we reported on the effect of source, and the diameter of the laser spot is about 1 lm.
Tsub on the CIGS film properties and found that a Tsub of Hall Effect and conductivity measurement were carried out
350 °C produces a single chalcopyrite phase CIGS film. on an ET 9003 using the Van Der Paw method. The con-
Unfortunately, all the 80 W-deposited films are deficient in ductivity of the films was measured in the 50–300 K range.
Se due to Se’s re-evaporation, and thus unsuitable for For the electrical measurement, silver-painted electrodes
absorber layers. In another study [11], we demonstrated were used as electrical contacts.
that the films sputtered at high power do not show Se’s re-
evaporation in the annealing process. Given all this, we
improved the sputtering power from 80 to 100 W to get
high-quality absorber layers and systematically investi- 3 Results and discussion
gated the structure and properties of CIGS films fabricated
at less than 380 °C in this study. XRD, SEM, Raman, 3.1 Morphology and composition
optical and electrical measurements were carried out to
gain a better understanding of the relation between Tsub and Figure 1 shows the surface morphologies and cross-sec-
the film properties. tional images of the CIGS thin films deposited at different
temperatures. A CIGS film, deposited at 200 °C (Fig. 1a),
shows pebble-shaped particles homogeneously distributed
2 Experimental details on the surface. The average size of the particles is in the
range of 1–3 lm. The compositions of the particle and the
CIGS thin films were deposited by a one-step RF magne- back ground film were determined by EDS, and the
tron sputtering process. SLG substrates with the dimen- results are summarized in Table 1. The EDS results reveal
sions of 2.5 cm 9 2.5 cm 9 1 mm were used as substrate. that the grains contain a much higher content of Cu
Before deposition, the glass substrates were ultrasonically (45.4 at.%) than that of the film (25.7 at.%). The Ga and
cleaned sequentially with acetone, alcohol, and deionized In contents in the grains are lower. The compositional
water. A single ceramic quaternary Cu(In0.7Ga0.3)Se2 target difference might be correlated to the faster diffusion
with the composition of Cu, In, Ga, and Se = 25, 17.5, 7.5, velocity of copper than that of indium and gallium [12].
and 50 atom percent ratio (at.%), respectively, was used as As the Tsub is elevated to 260 °C, the particles exhibit
sputtering source. CIGS films were sputtered at the power clear boundary and decreased sizes. The Cu content in the
of 100 W. Sputtering was performed in a high-purity argon particles decreased significantly, and the Se content
atmosphere at a pressure of 0.5 Pa and an argon flow rate increased notably. Meanwhile, a compositional change of
of 20 sccm. To calibrate the real sample temperature and the background film is not so obvious. For the films
the shown Tsub, a second thermocouple was mounted in the deposited at 320 and 380 °C, Fig. 1c, d shows a uniform
front side of SLG. Based on this calibration, the Tsub was and compact morphology, with no particles being dis-
maintained at different temperatures of 200, 260, 320, and tributed on the surface. Both the CIGS films exhibit near-
380 °C,. The total deposition time of the films is 120 min. stoichiometric composition. The composition of the
The crystallinity of the films was measured by glancing 380 °C-deposited film is stoichiometric with Cu23.7In20.8-
incidence X-ray diffraction with the glancing angle of 0.5° Ga5.8Se49.7, which implies a high crystalline quality and a
(GIXRD, PANalytical X’Pert PRO, CuKa radiation). small amount of native defects in the film.
Compositions of the films were determined by energy Figure 1e–h denotes the corresponding cross-sectional
dispersive X-ray spectroscopy (EDS, INCA spectrometer) SEM images of the deposited films. All the CIGS films
at an accelerating voltage of 10 keV. The accelerating exhibit compact structure, an no cracks and pinholes can be
voltage of 10 keV was used to give lower penetration observed throughout the film. The thickness of the films is
depth-work nearer the surface of the samples. The about 1 lm. The detailed microstructure from the film
Fig. 1 Top view (a–d) and cross-sectional (e–h) SEM images of the CIGS films on glass substrates obtained at various temperatures: 200 °C (a,
e); 260 °C (b, f); 320 °C (c, g); and 380 °C (d, h), respectively
surface down to the bottom can be clearly observed in the featureless-shaped grains, and that the 380 °C-deposited
SEM cross-sectional images. Figure 1g, h reveals that the film is composed of polyhedral grains with a uniform size
320 °C-deposited film consists of densely packed of about 200 nm.
Table 1 EDS spectrum of the CIGS film background and particles of the as-deposited film
Tsub (°C) Film Particles
Cu (at.%) In (at.%) Ga (at.%) Se (at.%) CIG Se/M Cu (at.%) In (at.%) Ga (at.%) Se (at.%) CIG Se/M
200 25.7 24.5 7.55 42.3 0.80 0.73 45.4 13.4 3.77 37.4 2.64 0.60
260 24.2 21.4 6.73 48.7 0.76 0.95 29.4 14.5 3.24 52.9 1.66 1.12
320 21.8 20.8 6.0 51.3 0.81 1.05 – – – – – –
380 23.7 20.8 5.8 49.7 0.89 0.99 – – – – – –
CIG is Cu/(In ? Ga) ratio, and Se/M is Se/(Cu ? In ? Ga) ratio
FHWM (o)
0.50
This Cu-rich phase usually acts as recombination center in
the CIGS solar cell device and deteriorates the device
0.45
Intensity (a.u.)
performance. The Cu-rich phase needs to be removed
(112)
during the fabricating process of the CIGS solar cell [13].
0.40
260 320 380
(220/204)
3.2 Crystal structure Temperature ( oC)
(312/116)
380 °C
GIXRD patterns of the CIGS films deposited at different Tsub
are shown in Fig. 2. All the deposited films exhibit three 320 °C
X-ray diffraction peaks located at 26.7°, 44.4°, and 52.8°.
260 °C
These peaks can be indexed to (112), (220)/(204), and (312)/
(116) planes of the CuIn0.7Ga0.3Se2 chalcopyrite phase
(JCPDS 35-1102), respectively. All GIXRD patterns exhibit 20 25 30 35 40 45 50 55 60 65 70 75 80
a prominent diffraction peak corresponding to CIGS (112) 2 θ (°)
plane. The intensity of the CIGS (112) plane diffraction peak
Fig. 2 GIXRD patterns of the CIGS thin films deposited at different
improves with the increasing deposition temperature and temperatures. The inset illustrates the FHWM of (112) peak versus
reaches its maximum value at 380 °C. From the GIXRD temperature variation
patterns, no other complex peak can be observed in the dif-
fraction angles ranging from 20° to 70°, indicating that no
micro-Raman spectra of the deposited CIGS films. All the
secondary phase can be distinguished by XRD.
Raman spectra exhibit a prominent peak located in the range
The inset of Fig. 2 shows the full width at half maxi-
of 175–181 cm-1 which corresponds to the A1 mode of the
mum (FWHM) values of the CIGS films deposited at
chalcopyrite (CH) CIGS phase. When the Tsub is increased,
various temperatures. As Tsub increases, the FWHM value
the intensity of A1 vibration peak increases, and the corre-
decreases, indicating crystalline quality improvement.
sponding FWHM value decreases, indicating the improve-
Furthermore, the grain size of the film can be calculated
ment of the film crystalline quality. Besides the A1 peak,
using Debye–Scherrer formula:
Raman peaks located at 212 and 229 cm-1 can also be
0:94k observed. These two peaks correspond to the B2/E and B2
d¼ ; ð1Þ
B cos hB vibrational modes of the CIGS phase, respectively. The
where d is the crystalline size, k is the wavelength of CuKa Raman and XRD results both indicate that the low-tem-
radiation (k = 1.54 Å), B is the FWHM of the (112) peak, perature-deposited CIGS films in this study by one-step RF
and hB is the Bragg angle. The calculated grain sizes are sputtering exhibit a single-phase polycrystalline chalcopy-
123, 161, and 194 nm for the 260-, 320-, and 380 °C- rite structure without any secondary phases in the films.
deposited CIGS films, respectively. It reveals that the
crystalline size increases with the elevating Tsub. 3.3 Electrical properties
Raman spectroscopy is considered as an appropriate
method to assess the structure of a few material phases, The resistivity (q), conduction type, and carrier concen-
which are difficult to be identified by XRD. Figure 3 depicts tration of the deposited films were measured using the Van
600 o
380 C
A1 CIGS 0.15
0.10
500
0.05
B2 / E 0.00
Intensity (arb.units)
σ (Ω –1.cm–1)
o
380 C 0.15
300 0.10
0.05
200 o
0.00
320 C o
2.5 260 C
2.0
100
1.5
o
260 C 1.0
0 0.5
100 150 200 250 300 350 0.0
-1 50 100 150 200 250 300
Raman shift (cm )
T (K)
Fig. 3 Raman spectra of CIGS films recorded at room temperature
Fig. 4 Temperature dependence of conductivity measurements of the
films
Der Paw method at the room temperature [14]. Table 2
summarizes the results obtained from the Hall measure-
ment, which indicates that all the deposited films exhibit p- energy Ea, which determines the potential barrier height,
type characteristics. As the Tsub is elevated, the resistivity can be calculated from this Arrhenius plots, and it increases
of the film increases, and the carrier concentration is from 27 to 90.6 meV as Tsub increases from 260 to 380 °C
reduced. The resistivity of CIGS films deposited at 320 and (shown in Table 2). The Ea value of 90.6 meV is very close
380 °C are suitable for fabricating solar cell devices. Films to that of the CIGS film deposited by three-stage process
deposited at 260 °C exhibit a lower resistivity and a higher (98.1 meV) [15]. The Ea value of the 260 °C-deposited
carrier concentration, which might be attributed to the film is almost the same as that measured at the room-
presence of a Cu-rich phase on the films surface [4]. temperature thermal energy (25.9 meV), which may be
The temperature (T) dependence of the conductivity (r) attributed to the presence of the trace Cu2Se phase in the
was measured to identify the transport mechanisms in the film.
deposited films. Figure 4 shows that the conductivity of the In the low-temperature range of 50–120 K, the curve of
film increases with the elevating measurement temperature, r versus T can be expressed by
indicating the semiconducting nature of our CIGS films. " #
Figure 5 shows the corresponding depicted curves of r0 T0 1=4
ln r versus 103/T of the deposited CIGS films. The related r ¼ 1=2 exp ; ð2Þ
T T
ln r versus 1,000/T curves in the temperature range of
50–300 K present two quite different slopes, indicating that where T0 is the localization temperature, which is
the conductivity is affected by two different transport associated with the stoichiometry and disorder in the
mechanisms. In the high-temperature range (over 120 K), films; and r0 is the pre-exponential factor. This conduction
all the curves show linear behavior, indicating a thermionic process requires a T0/T ratio much higher than 1. Figure 6
emission transport mechanism of the films. The activation shows the plots of ln(rT1/2) versus T-1/4. The plots are
20
o
2 380 C 380 C
o
o
320 C 18 320 C
o
o
260 C 260 C
o
0 16
ln(σT ) (Ω ·cm ·K )
1/2
14
ln σ (Ω –1.cm–1)
-2
-1
12
-1
-4
10
1/2
-6 8
6
-8
4
-10
0.30 0.31 0.32 0.33 0.34 0.35 0.36 0.37 0.38
2 4 6 8 10 12 14 16 18 20 22 1/4 -1/4
-1
1/T (K )
1000/T (K )
Fig. 6 The depicted curves of ln(rT1/2) versus T-1/4 in the temper-
Fig. 5 The depicted curves of ln r versus 103/T in the temperature ature range of 50–120 K
range of 50–300 K
linear enough allowing us to interpret the measurements Figure 7 shows the transmittance (Tr) spectra of the CIGS
based on our belief that the films are affected by a variable films deposited at various temperatures. Transmittance of
range hopping (VRH) conduction mechanism. As predicted the 260 °C-deposited film is kept at a value of 20 % in the
by Davis–Mott model, conductivity is affected by transport range of 1,500–2,400 nm, which might be ascribed to the
in a band of localized states, where the carriers move absorption tails of the Cu-rich particles on the surface [20].
between states via a phonon-assisted tunneling process A significant improvement of transmittance and a sharp fall
[16]. This transport mechanism has already been reported of transmittance at the band edge can be observed when the
in n-type samples of CIGS [17] in Mott’s derivation as CIGS films are deposited at a higher temperature. The
follows: 380 °C-deposited film exhibits interference patterns in the
transmittance spectrum, indicating a very good thickness
18a3 2 N ðEF Þ 1=2
T0 ¼ ; and r0 ¼ 3e m ; ð3Þ
kNðEF Þ 8p g k
70
where N(EF) is the density of the localized states, k is the
Boltzmann constant, g is the decay constant of the wave o
380 C
60
function of the localized states near the Fermi level, and m o
320 C
o
is the Debye frequency. The two parameters T0 and N(EF) 260 C
Transmittance (%)
50
can be evaluated from the slope of ln(rT1/2) versus T-1/4
curve and the intercept at T-1/4 = 0, respectively. Both
40
the T0 and N(EF) values are consistent with the proposed
model, and they are dominated by Tsub. N(EF) is related to
30
the density of native defects in the films since native
defects generate extra energy levels in the forbidden band,
20
forming the localized energy states. Because our 380 °C-
deposited film has the stoichiometric composition, the
10
corresponding N(EF) value is as low as that of the co-
evaporated polycrystalline CuInSe2 [18, 19], indicating a
low density of the localized state at the Fermi level. 0
Therefore, our low-temperature-deposition method can 600 900 1200 1500 1800 2100 2400
(α h ν) (x10 (eV/cm) )
2
1.05
designated as Tr max and Tr min, which correspond, 8
8
6 260 320 380
spectrum. According to the EM, the refractive index, n, is o
Temperature ( C)
given by
2
h 4
1=2 i1=2
n ¼ N þ N 2 S2 ; ð4Þ o
380 C
2
o
320 C
where N ¼ 2SðTr max Tr min Þ=ðTr max Tr min Þ þ ðS2 þ 1Þ 2; o
260 C
S is the refractive index of the substrate; and Tr max and
Tr min are the transmission maximum and the correspond- 0
ing minimum at a certain wavelength, respectively. The 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
2.72
4 Conclusion
2.70
CIGS films were prepared by a one-step RF magnetron
2.68
sputtering process at low Tsub. It can be concluded that
there exists a minimum suitable Tsub of 320 °C. Cu-rich
n
Natural Science Foundation of China (No. 51271155, 51377138); and 10. Yu Z, Yan Y, Li S et al (2013) Significant effect of substrate
the Science Foundation of Sichuan Province (Nos. 2011JY0031 and temperature on the phase structure, optical and electrical prop-
2011JY0130). erties of RF sputtered CIGS films. Appl Surf Sci 264:197–201
11. Yu Z, Liu L, Yan Y et al (2012) Properties of different temper-
Open Access This article is distributed under the terms of the ature annealed Cu(InGa)Se2 and Cu(InGa)2Se3.5 films prepared
Creative Commons Attribution License which permits any use, dis- by RF sputtering. Appl Surf Sci 261:353–359
tribution, and reproduction in any medium, provided the original 12. Kim H, Horwitz JS, Qadri SB et al (2002) Epitaxial growth of Al-
author(s) and the source are credited. doped ZnO thin films grown by pulsed laser deposition. Thin
Solid Films 420–421:107–111
13. Zhang L, He Q, Jiang W-L et al (2009) Effects of substrate
temperature on the structural and electrical properties of Cu(In-
Ga)Se2 thin films. Sol Energy Mater Sol Cells 93(1):114–118
References 14. Mesa F, Calderón C, Gordillo G (2010) Study of electrical
properties of CIGS thin films prepared by multistage processes.
1. Chopra K, Paulson P, Dutta V (2004) Thin-film solar cells: an Thin Solid Films 518(7):1764–1766
overview. Prog Photovolt Res Appl 12(2–3):69–92 15. Moussa EH, Ariswan GW, Khoury A et al (2002) Fabrication and
2. Repins I, Contreras MA, Egaas B et al (2008) 19.9 %-efficient study of photovoltaic material CuInxGa1-xSe2 bulk and thin films
ZnO/CdS/CuInGaSe2 solar cell with 81.2 % fill factor. Prog obtained by the technique of close-spaced vapor transport. Solid
Photovolt Res Appl 16(3):235–239 State Commun 122(3):195–199
3. Seike S, Shiosaki K, Kuramoto M et al (2011) Development of 16. Essaleh L, Wasim SM (2006) Magnetoresistance and hall
high-efficiency CIGS integrated submodules using in-line depo- mobility in the variable range hopping regime in n-type CuInSe.
sition technology. Sol Energy Mater Sol Cells 95(1):254–256 AIP Conf Proc 850:1470–1471
4. Zhang L, Liu F, Li F et al (2012) Structural, optical and electrical 17. Amara A, Ferdi A, Drici A et al (2006) Electrical and optical study
properties of low-temperature deposition Cu(InxGa1-x)Se2 thin of Cu(InGa)Se2 co-evaporated thin films. Catal Today 113(3):
films. Sol Energy Mater Sol Cells 99:356–361 251–256
5. Lammer M, Klemm U, Powalla M (2001) Sodium co-evaporation 18. Gandotra VK, Ferdinand KV, Jagadish C et al (1986) Effect of
for low temperature Cu(InGa)Se2 deposition. Thin Solid Films excess copper on the electrical properties of polycrystalline thin
387(1):33–36 films of CuInSe2. Physica Status Solidi (a) 98(2):595–603
6. Shafarman WN, Zhu J (2000) Effect of substrate temperature and 19. Naoki K, Takayuki N, Mikihiko N (1995) Wada takahiro prep-
deposition profile on evaporated Cu(InGa)Se2 films and devices. aration of device-quality Cu(InGa)Se2 thin films deposited by co-
Thin Solid Films 361:473–477 evaporation with composition monitor. Jpn J Appl Phys Part 2
7. Frantz JA, Bekele RY, Nguyen VQ et al (2011) Cu(InGa)Se2 thin 34:L1141–L1144
films and devices sputtered from a single target without addi- 20. Manifacier JC, Gasiot J, Fillard JP (1976) A simple method for
tional selenization. Thin Solid Films 519(22):7763–7765 the determination of the optical constants n, k and the thickness of
8. Yu Z, Yan C, Huang T et al (2012) Influence of sputtering power a weakly absorbing thin film. J Phys E 9(11):1002–1004
on composition, structure and electrical properties of RF sput- 21. Alonso MI, Wakita K, Pascual J et al (2001) Optical functions
tered CuIn1-xGaxSe2 thin films. Appl Surf Sci 258(13): and electronic structure of CuInSe2, CuGaSe2, CuInS2, and Cu-
5222–5229 GaS2. Phys Rev B 63(7):075203
9. Yu Z, Yan C, Yan Y et al (2012) Effect of annealing temperature 22. Park KC, Ma DY, Kim KH (1997) The physical properties of Al-
on properties of RF sputtered Cu(InGa)Se2 thin films. Appl Surf doped zinc oxide films prepared by RF magnetron sputtering.
Sci 258(22):8527–8532 Thin Solid Films 305(1):201–209