Effect of Cu/In molar ratio on the microstructural and optical properties of microcrystalline CuInS2 prepared by solvothermal route

Article abstract of DOI:10.1016/j.materresbull.2007.10.013

Synthesis and characterization of CuInS₂ powder sample prepared by a simple and convenient solvothermal method is reported. The influence of the variation of Cu/In molar ratio from 0.69 to 1.25 on the particle morphology, crystal structure and optical properties of CuInS₂ samples was studied. The X-ray diffraction studies indicated that the samples were polycrystalline in nature. SEM images of the samples revealed that the copper-rich products were uniform microspheres with smooth surfaces, whereas microspheres formed by network of interconnected flakes were obtained for indium-rich products. The optical band gaps (E_g) of the products decreased from 1.60 to 1.43 eV with variation of Cu/In molar ratio. The variation of the Urbach tail width with Cu/In molar ratio indicated that the density of the defects is much higher for the indium-rich CuInS₂, which was clearly revealed from Raman measurements.

Full text of DOI:10.1016/j.materresbull.2007.10.013

Materials Research Bulletin 43 (2008) 2742–2750
www.elsevier.com/locate/matresbu
Effect of Cu/In molar ratio on the microstructural and
optical properties of microcrystalline CuInS2 prepared
by solvothermal route
Kajari Das a, Subhendu K. Panda a, Soma Gorai a, Pratima Mishra b,
Subhadra Chaudhuri a,
*
a Department of Materials Science, Indian Association for the Cultivation of Science, Kolkata 700032, India
b Advanced Materials Technology Department, Regional Research Laboratory, Bhubaneswar 760013, India
Received 1 November 2006; accepted 18 October 2007
Available online 23 October 2007
Abstract
Synthesis and characterization of CuInS2 powder sample prepared by a simple and convenient solvothermal method is reported.
The influence of the variation of Cu/In molar ratio from 0.69 to 1.25 on the particle morphology, crystal structure and optical
properties of CuInS2 samples was studied. The X-ray diffraction studies indicated that the samples were polycrystalline in nature.
SEM images of the samples revealed that the copper-rich products were uniform microspheres with smooth surfaces, whereas
microspheres formed by network of interconnected flakes were obtained for indium-rich products. The optical band gaps (Eg) of the
products decreased from 1.60 to 1.43 eV with variation of Cu/In molar ratio. The variation of the Urbach tail width with Cu/In molar
ratio indicated that the density of the defects is much higher for the indium-rich CuInS2, which was clearly revealed from Raman
measurements.
# 2007 Elsevier Ltd. All rights reserved.
Keywords: A. Semiconductors; C. Raman spectroscopy; D. Optical properties
1. Introduction
Ternary chalcopyrite semiconductors (I-III-IV2) are promising candidates for various applications in optoelectronic
devices, as their band gaps are well matched to the solar spectrum with high absorption coefficients a > 104 cmꢀ1 and
they can be easily doped to have p- and n-type conductivity. Among all I-III-IV2 ternaries, the most extensively studied
compounds are CuInSe2 and Cu(In,Ga)Se2 which in polycrystalline forms produced solar cells with high efficiency
(18.8%) [1,2]. Although the electrical properties of CuInSe2 are suitable for such application, its optical band gap
(1.02 eV) is quite small compared to the optimum value (ꢁ1.4 eV) for solar energy conversion. It is true that the addition
of Ga increases the band gap, but it is difficult to obtain the higher efficiency with Ga due to the reduction of minority
carrier mobility [3]. Also the ternary semiconductor containing selenium is harmful due to the toxicity of the material.
Therefore, the search for an equally important and environmentally benign solar cell material continued which brought
* Corresponding author. Tel.: +91 33 2473 4971; fax: +91 33 2473 2805.
E-mail address: mssc2@iacs.res.in (S. Chaudhuri).
0025-5408/$ – see front matter # 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2007.10.013

K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
2743
CuInS2 to the forefront. Further, CuInS2 has a band gap of 1.55 eV, which is very near to the optimum band gap for solar
cells. Solar cells fabricated with CuInS2 thin films have already exhibited an efficiency of 16% on an area of 1 cm2 [4].
There are various techniques for the preparation of CuInS2, e.g. spray pyrolysis [5–7], rf sputtering [8,9],
sulfurization [10], electrodeposition [11] and chemical process [12,13]. It has been observed that the structural and
optical properties of CuInS2 depend on the preparative conditions such as growth temperature and the molar ratio of
Cu/In. It is also reported that by controlling the molar ratio of the compositional elements, n-type or p-type
conductivity of CuInS2 can be adjusted easily [14,15]. Indium-rich product may lead to the formation of n-type
CuInS2, while copper-rich product may result in p-type CuInS2.
It is well known that the device properties of CuInS2-based solar cells are highly affected by their stoichiometric
composition. In contrast to the physical and chemical vapour deposition techniques, solvothermal method is a simple,
inexpensive technique in which the ratios of the constituents are controlled by adjusting their molar ratios in the
solution. In the present work, we have synthesized CuInS2 by a convenient solvothermal method with variation of the
molar ratio of Cu/In. We have studied their microstructural and optical properties as a function of this molar ratio.
2. Experimental
In a typical experimental procedure, the product CuInS2 was prepared from a mixture of copper chloride (CuCl,
Loba), indium chloride (InCl3) and thiourea (CH4N2S, Merck). Indium chloride was prepared from indium metal and
hydrochloric acid. The precursors were added into a teflon-lined steel autoclave, which was then filled with ethanol
upto 90% of the total volume. The autoclave was then sealed and maintained at 200 8C for 12 h, and then air-cooled to
room temperature. The black precipitate, thus obtained, was centrifuged and washed several times with distilled water
and ethanol and then vacuum dried at 60 8C for 4 h. Five sets of experiments were performed with variation of Cu/In
molar ratios in range 0.69–1.25 and the products were marked as CIS-1, CIS-2, CIS-3, CIS-4 and CIS-5 (as shown in
Table 1), while S/In ratio was always maintained constant (ꢁ3.5) for all the cases.
The crystalline phase of the products were determined by X-ray powder diffraction by using a Seifert 3000P
diffractometer with Cu Ka radiation (l = 1.540 Å´). The morphologies of the samples were studied by a scanning
electron microscope (SEM; Hitachi S-2300). The purity and composition of the prepared samples were also studied by
energy dispersive X-ray analysis (EDAX, Kevex, Delta Class I) attached with SEM and the FTIR spectra recorded in
the range 4000–400 cmꢀ1 with a FTIR spectrometer (Shimadzu-84005). Optical transmittance of the samples was
recorded by a UV–vis–NIR spectrophotometer (Hitachi, U-3410). For the optical transmittance measurement, CuInS2
films were prepared by Doctor’s blade technique using slurry of CuInS2 powders dispersing in ethanol as described in
our previous work [13]. Raman spectra were acquired with the Renishaw inVia Raman spectrometer, which is
amenable to measure samples in the microrange. The excitation source was a 514 nm argon ion laser with 100%
power. The resolution of all the spectra was within ꢂ2 cmꢀ1. All the spectra were analyzed with the software supplied
by Renishaw and background corrections were made when needed.
3. Results and discussion
3.1. XRD studies and EDAX analysis
The effect of the variation of Cu/In molar ratio on the crystallinity of CuInS2 powder samples was studied by the X-
ray diffraction (Fig. 1). All the peaks due to reflections from (1 1 2), (2 0 4) and (3 1 2) planes of the samples CIS-1 to
Table 1
Summarized Cu/In molar ratio of the products and their corresponding band gap values
Sample no.
Molar ratio of Cu/In in starting solutions
Molar ratio of Cu/In (from EDAX)
Band gaps (eV)
Cu (at.%)
In (at.%)
S (at.%)
Cu/In
CIS-1
CIS-2
CIS-3
CIS-4
CIS-5
0.69
0.74
0.80
1.00
1.25
23.28
23.99
25.09
27.23
28.09
29.10
26.95
24.60
20.57
17.67
47.60
49.05
50.29
52.21
54.32
0.80
0.89
1.02
1.32
1.59
1.60
1.57
1.53
1.50
1.43

2744
K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
CIS-5 matched well with the standard JCPDS (27-0159) pattern of tetragonal CuInS2. No impurities were detected for
the sample CIS-3, but the XRD spectra of the indium-rich products (CIS-1 and CIS-2) indicated the formation of b-
In2S3 along with CuInS2. The XRD spectra of the copper-rich products (CIS-4 and CIS-5) revealed the presence of
CuS. It was also observed that the crystallinity of the products increased with increasing Cu/In molar ratio. This
indicated that the crystal growth is faster in the copper rich products [16]. Copper and indium reacted with thiourea at
room temperature producing copper–thiourea and indium–thiourea complexes. At low temperature (ꢁ130 8C)
[17,18], copper–thiourea complex can dissociate to give well crystalline copper sulfide, whereas a high temperature
ꢁ180 8C is necessary for dissociation of indium–thiourea complex to produce well crystalline indium sulfide [19].
Thus the copper-rich solution promotes the formation of CuInS2 with well-defined crystallites at lower growth
temperatures than the indium rich products.
The compositional analysis by EDAX for the all samples is shown in Table 1. It was observed that the
stoichiometric ratio between copper and indium could be accurately controlled from the starting composition of
the precursor. The products exhibited a deficiency of indium and sulfur in comparison to the starting composition
that we have taken. This deficiency of indium and sulfur in the products is a result of the reaction process of
CuInS2. The ternary compound CuInS2 is formed through the thermal decomposition of metal–thiourea complex
salts [20], which was also revealed from the FTIR studies. During this formation, the system easily lost indium
sulfide due to its higher vapour pressure and a deficiency of indium and sulfur was apparent in the products
(Table 1).
3.2. Microstructural analysis
The morphological studies of CuInS2 samples were performed by SEM, which are shown in Fig. 2. The
copper-rich products showed microspheres having diameters in the range of 1.8–3.9 mm as shown in Fig. 2(a). The
surfaces of the copper-rich microspheres were not very smooth. The stoichiometric product (Fig. 2(b)) revealed
microspheres with smooth surfaces having diameters in the range of 2.6–4.7 mm. The indium-rich product
(Fig. 2(c)) represented microspheres having diameters in the range of 1.6–3 mm but they had a different look
compared to the copper-rich and stoichiometric products. Fig. 2(c) showed the microspheres formed by the
network of interconnected flakes of thickness ꢁ50 nm. The magnified part of this network is shown in the inset of
Fig. 2(c).
Fig. 1. XRD of CuInS2 obtained by solvothermal route at different Cu/In molar ratios (Table 1) of the starting solution: (a) CIS-1; (b) CIS-2; (c) CIS-
3; (d) CIS-4; (e) CIS-5.

K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
2745
Fig. 2. SEM micrographs for CuInS2 obtained by solvothermal route at different Cu/In molar ratios (Table 1) of the starting solution: (a) CIS-5; (b)
CIS-3; (c) CIS-1. The inset of Fig. 3(c) shows the interconnected nanoflakes (the scale bar of inset in Fig. 3(c) is 250 nm).
3.3. FTIR study
FTIR spectroscopy has been used to get the information about the organic impurities and the reaction mechanism of
CuInS2 powder samples as shown in Fig. 3. The IR spectra of the samples CIS-1, CIS-3 and CIS-5 revealed that there is
no vibration caused by CuInS2 in the spectral range 4000–400 cmꢀ1. The IR spectra of the samples showed that the
vibrations at 3436 and 1628 cmꢀ1, could be due to NH2 group [21]. The vibration at 2348 cmꢀ1 was due to CO2 and

2746
K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
Fig. 3. FTIR spectra for CuInS2 samples: (a) CIS-3; (b) CIS-5; (c) CIS-1; (d) metal–thiourea complex solvothermally treated at 80 8C for 12 h; (e)
metal–thiourea complex at room temperature.
was not related to the samples. CuInS2 is formed after dissociation of the metal–thiourea complex. This was confirmed
by the FTIR studies of metal–thiourea complex prepared at room temperature and solvothermally treated complex at
80 8C for 12 h. For the preparation of the metal-thiourea complex at room temperature, CuCl and InCl3 were first
dissolved in ethanol under 2 h constant stirring and then thiourea was added to the solution. A white precipitate of
metal–thiourea complex appeared. This precipitate was collected and dried in air at room temperature. The FTIR
spectra of the metal–thiourea complex at room temperature and solvothermally treated complex at 80 8C are shown in
Fig. 3(e and d), respectively. The IR spectrum (Fig. 3e) showed the absorptions at 3374, 3285 and 3182 cmꢀ1, which
could be due to the thiourea and all other absorptions of the spectrum (e) may be due to the metal–thiourea complex
[22]. The peaks due to the absorption by the thiourea and metal–thiourea complex completely disappeared above
80 8C. The IR spectrum (d) showed the vibrations at 1050 and 669 cmꢀ1, which could be coupled with the sulfate
(SO4) group [21]. The spectrum (d) also indicated the absorptions at 1111 cmꢀ1 and 619 cmꢀ1, which are assigned to
CuSO4ꢃH2O and Cu2O, respectively [23].
3.4. Optical transmittance and band gap
Fig. 4 shows the transmittance versus wavelength (%T vs. l) spectra of the CuInS2 films with different Cu/In molar
ratios prepared by Doctor’s blade method.
It is well known that absorption coefficient (a) is related to the incident photon energy (hn) by the following
equation
m
ahn ¼ Aðhn ꢀ EgÞ
(1)
where A is a constant, Eg is the band gap, m is the index indicating the type of the transition. It is known that the values
of m for allowed direct, allowed indirect, forbidden direct, forbidden indirect transitions are 1/2, 2, 3/2 and 3,
respectively.
CuInS2 is a well known direct band gap material, so putting m = 1/2, Eq. (1) can be written as
1=2
ahn ¼ Aðhn ꢀ EgÞ
(2)
The optical band gap Eg can be obtained from the plot of (ahn)2 versus hn, as shown in Fig. 5. It can be observed that
Eg decreases from 1.60 to 1.43 eV when the Cu/In molar ratio increases from 0.69 to 1.25 (Table 1). The value of band

K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
2747
Fig. 4. Optical transmittance spectra of CuInS2 samples at different Cu/In molar ratio of the starting solution: (a) CIS-3; (b) CIS-4; (c) CIS-5; (d)
CIS-2; (e) CIS-1 as indicated in Table 1.
gap 1.53 eV for the stoichiometric product is close to 1.55 eV, which is the reported value for bulk crystalline CuInS2
[3]. The variation of absorption coefficient (a) with Cu/In molar ratio is shown in the plot of the absorption coefficient
(a) versus energy (hn) in Fig. 6. The maximum absorption coefficient (ꢁ8 ꢄ 104 cmꢀ1) is obtained for the
stoichiometric product. It may be noted that the lower Cu/In ratio leads to the increase of the optical band gap [24]. The
deviations from stoichiometry in CuInS2 result in some defects in crystalline lattice, such as copper, indium and sulfur
vacancies (VCu, VIn and VS), interstitial sulfur (Si), substitutional copper in indium sites (CuIn), and also substitutional
indium in copper sites (InCu). When Cu/In molar ratio increases, the CuIn and VIn defect densities may increase which
make the materials p-type degenerate, but when the materials become indium-rich, the most probable defects are VS,
InCu, which can introduce two donor levels and also the defect VCu is responsible for producing an acceptor level. Thus
Fig. 5. A plot of (ahn)2 vs. hn for CuInS2 samples at different Cu/In molar ratio of the starting solution: (a) CIS-5; (b) CIS-4; (c) CIS-3; (d) CIS-2; (e)
CIS-1 as indicated in Table 1.

2748
K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
Fig. 6. A plot of the absorption coefficient (a) vs. photon energy (hn) for CuInS2 obtained by solvothermal route at different Cu/In molar ratio of the
starting solution: (a) CIS-3; (b) CIS-4; (c) CIS-5; (d) CIS-2; (e) CIS-1as indicated in Table 1.
the indium-rich materials become n-type degenerate. Obviously, these defects play very important role in the variation
of the optical band gap of CuInS2 with Cu/In molar ratio [25].
The transmittance spectra in Fig. 4 indicated that the sharpness of the fall of the spectra near the band edge
decreased with the deviation from the stoichiometry. This was also reflected from the plot of the Urbach tail width (e)
versus Cu/In molar ratio as shown in Fig. 7. The Urbach tail width (e) was estimated from the following equation:
ꢀ
ꢁ
hn
a ¼ const:exp
(3)
e
Fig. 7. Variation of the Urbach tail width (e) with Cu/In molar ratio (Table 1) of CuInS2 samples.

K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
2749
Fig. 8. Raman spectra of three representative CuInS2 powder samples: (a) CIS-1; (b) CIS-3; (c) CIS-5.
where e represents the Urbach tail width [26], a is the absorption coefficient and hn is the photon energy. The inverse of
the slope of the ln a versus hn plot gives the value of e (the plot is not shown here). Fig. 7 revealed that the broadening
of the tail width (e) was observed with deviation from the stoichiometry due to the formation of a few secondary phases
for the presence of excess Cu or In. The defect states arising out of different secondary phases in the non-
stoichiometric products would modulate the fall of transmittance due to the transitions occurring through the defect
states. The variation of Urbach tail with the Cu/In molar ratio indicated that the density of the defects on the crystal is
minimum for the stoichiometric product than the copper-rich and indium-rich products, which was clearly revealed
from Raman measurements, discussed below.
3.5. Raman measurements
Fig. 8 shows the Raman spectra of CuInS2 powder samples CIS-1, CIS-3 and CIS-5. It may be observed that there
are five peaks at 240, 266, 290, 320 and 340 cmꢀ1 for all the above samples, which were assigned to the E, B2, A1, E
and B2 modes of the chalcopyrite phase of CuInS2 [27,28]. The Raman peak at 307 cmꢀ1, which was observed only for
the indium-rich sample CIS-1, may be due to the impurity b-In2S3 [29]. In addition, there were some peaks below
160 cmꢀ1 for the copper-rich sample CIS-5, which may be due to the Cu–S vibration [30]. The Raman peak at
290 cmꢀ1 for all the samples were fitted to the Gaussian curve and the full-width at half-maximum (FWHM) were
estimated. It was noted that the FWHM of the peaks increased in the order CIS-1 > CIS-5 > CIS-3, which may be
attributed to the presence of defects. Thus, the density of the defects on the crystal is much higher for the indium-rich
CuInS2 compared to the copper-rich and stoichiometric CuInS2. It may be commented here that the single crystal
CuInS2 and almost defect free CuInS2 thin films were reported to have the FWHM of 3.5 and 5 cmꢀ1, respectively
[29]. Compared to this value, the FWHM of our stoichiometric sample was 8.6 cmꢀ1 indicating almost defect free
CuInS2.
4. Conclusions
CuInS2 powder samples were successfully synthesized by a relatively simple and convenient solvothermal route
using ethanol as the solvent. The stoichiometry of the product could be accurately controlled by adjusting the Cu/In
molar ratio in the starting precursor. It was also observed that the composition of the precursor has a strong influence
on the crystallinity, morphology, and optical properties of the products. The XRD studies of the powder samples
revealed that the stoichiometric and the copper-rich products had better crystallinity than the indium-rich products. In

2750
K. Das et al. / Materials Research Bulletin 43 (2008) 2742–2750
the Raman study, A1 mode at 290 cmꢀ1 of CuInS2 was dominant for all the samples. There was a broadening of the
Raman peak at 290 cmꢀ1 with deviations from the stoichiometry, which may be attributed to the defects in the crystal.
Some impurities such as b-In2S3 and CuS were detected from XRD for the indium-rich and copper-rich products,
respectively. The optical band gaps of the CuInS2 increased from 1.43 to 1.60 eV when the Cu/In molar ratio was
decreased from 1.25 to 0.69. Broadening of the absorption tail with variation of the Cu/In molar ratio was determined
and it passed through a minimum for the stoichiometric product. From the XRD, Raman measurements and optical
transmittance studies of the samples, it was clearly revealed that the stoichiometric CuInS2 is almost defect-free and is
suitable for the photovoltaic applications.
Acknowledgements
The authors like to express their sincere gratitude to the Department of Science and Technology (DST) and Ministry
of Non-Conventional Energy Sources (MNES), Govt. of India, for financial assistance and sanction of a fellowship
during the tenure of the programme.
References
[1] M. Contreras, B. Egaas, K. Ramanathan, J. Hiltner, A. Swartzlander, F. Hasoon, R. Noufi, Prog. Photovoltaics 7 (1999) 311.
[2] B. Li, Y. Xie, J. Huang, Y. Qian, Adv. Mater. 11 (1999) 1456.
[3] J.L. Shay, J.H. Wernick, Ternary Chalcopyrite Semiconductors: Growth, Electronics Properties and Applications, Pergamon Press, New York,
1975.
[4] H.W. Schock, in: Proceedings of the 12th European Photovoltaic Solar Energy Conference, 1994, p. 944.
[5] M. Ortega-López, A. Morales-Acevedo, Thin Solid Films 330 (1998) 96.
[6] I. Ojaa, M. Nanub, A. Katerskia, M. Krunksa, A. Merea, J. Raudojaa, A. Goossens, Thin Solid Films 480/481 (2005) 82.
[7] A. Mere, O. Kijatkina, H. Rebane, J. Krustok, M. Krunks, J. Phys. Chem. Solids 64 (2003) 2025.
[8] V. Alkerts, J.H. Schon, M.J. Witcomb, E. Bucher, U. Ruhle, H.W. Schock, J. Phys. D Appl. Phys. 31 (1998) 2869.
[9] Y.B. He, T. Kramer, A. Polity, R. Gregor, W. Kriegseis, I. Österreicher, D. Hasselkamp, B.K. Meyer, Thin Solid Films 431/432 (2003) 231.
[10] T. Wada, T. Negami, M. Nishitani, J. Mater. Res. 8 (1993) 545.
[11] B. Asenjo, A.M. Chaparro, M.T. Gutiérrez, J. Hrrero, Thin Solid Films 511/512 (2006) 117.
[12] P. Guha, S. Gorai, D. Ganguli, S. Chaudhuri, Mater. Lett. 57 (2003) 1786.
[13] P. Guha, D. Das, A.B. Maity, D. Ganguli, S. Chaudhuri, Sol. Energy Mater. Sol. Cells 80 (2003) 115.
[14] G.C. Park, H.D. Chung, C.D. Kim, H.R. Park, W.J. Jeong, J.U. Kim, H.B. Gu, K.S. Lee, Sol. Energy Mater. Sol. Cells 49 (1997) 365.
[15] T. Yukawa, K. Kuwabara, K. Koumoto, Thin Solid Films 286 (1996) 151.
[16] M. Krunks, V. Mikli, O. Bijakina, H. Rebane, A. Mere, T. Varema, E. Mellikov, Thin Solid Films 361/362 (2000) 61.
[17] S. Gorai, D. Ganguli, S. Chaudhuri, Mater. Lett. 59 (2005) 826.
[18] S. Gorai, D. Ganguli, S. Chaudhuri, Cryst. Growth Des. 5 (2005) 875.
[19] J.Q. Hu, B. Deng, W.X. Zhang, K.B. Tang, Y.T. Qian, Inorg. Chem. 40 (2001) 3130.
[20] M. Krunks, T. Leskela, R. Mannonen, L. Niinisto, J. Therm. Anal. 53 (1998) 355.
[21] M. Krunks, O. Kijatkina, H. Rebane, I. Oja, V. Mikli, A. Mere, Thin Solid Films 403/404 (2002) 71.
[22] P. Bombicz, I. Mutikainen, M. Krunks, T. Leskelä, J. Madarász, L. Niinistö, Inorg. Chim. Acta 357 (2004) 513.
[23] J.G. Dunna, C. Muzenda, Thermochim. Acta 369 (2001) 117.
[24] X. Hou, K.L. Choy, Thin Solid Films 480/481 (2005) 13.
[25] H.Y. Ueng, H.I. Hwang, J. Phys. Chem. Solids 50 (1989) 1279.
[26] F. Urbach, Phys. Rev. 92 (1953) 1324.
[27] D.M. Hwang, C.C. Chen, H.L. Hwang, Chin. J. Phys. 19 (1981) 56.
[28] K. Kondo, S. Nakamura, K. Sato, Jpn. J. Appl. Phys. 37 (1998) 5728.
[29] J. Álvarez-García, J. Marcos-Ruzafa, A. Pérez-Rodríguez, A. Romano-Rodríguez, J.R. Morante, R. Scheer, Thin Solid Films 361/362 (2000)
208.
[30] L. Calvo-Barrio, A. Pérez-Rodríguez, J. Alvarez-Garcia, A. Romano-Rodríguez, B. Barcones, J.R. Morante, K. Siemer, I. Luck, R. Klenk, R.
Scheer, Vacuum 63 (2001) 315.