Band edges determination of CuInS2 thin films prepared by electrodeposition

Band edges determination of CuInS2 thin films prepared by electrodeposition

Materials Chemistry and Physics 88 (2004) 417–420 Band edges determination of CuInS2 thin films prepared by electrodeposition A.M. Martineza , L.G. A...

105KB Sizes 0 Downloads 10 Views

Materials Chemistry and Physics 88 (2004) 417–420

Band edges determination of CuInS2 thin films prepared by electrodeposition A.M. Martineza , L.G. Arriagab,∗ , A.M. Fern´andeza , U. Canob b

a Departamento de Materiales Solares, CIE-UNAM, Av., Xochicalco s/n, Col. Centro, 62580 Temixco, Mor., M´ exico Instituto de Investigaciones El´ectricas, Gerencia de Energ´ıas No Convencionales Av., Reforma 113, Col. Palmira, 62490 Cuernavaca, Mor., M´exico

Received 4 May 2004; received in revised form 9 August 2004; accepted 18 August 2004

Abstract A CuInS2 (CIS) semiconductor thin film was growth by electrodeposition on a stainless steel substrate. In order to improve the polycrystallinity the samples were annealed in a N2 atmosphere. The films were characterized by electrochemical techniques and X ray diffraction and their band gaps were determined by photocurrent spectroscopy. When the electrolytic bath has the same concentration [Cu2+ ] = [In3+ ] the resulting film was of the n-type, while for different concentrations of Cu and In ions the film was of the p-type. A depletion zone during capacitance–voltage measurements at 10 kHz frequency was seen over the voltage range used. Using C–V plots in the depletion zone, flat-band potentials and the energetic position of band edges were calculated. © 2004 Elsevier B.V. All rights reserved. Keywords: Semiconductors; Thin films; Electrodeposition; X-ray diffraction

1. Introduction There is an increasing interest in semiconductor/electrolyte systems in connection with their application as phototelectrolytic energy conversion devices (e.g. hydrogen evolution). One of the most interesting semiconductor materials is the CuYX2 semiconductors, where Y = In, Ga and X = S, Se. Theses materials offer a direct band gap with a range, between 1 and 2.4 eV and high light absorption coefficients [1]. For example, materials with Eg -values near 1.5 eV, are adequate for solar cell devices, and when Eg -values are higher than 1.6 eV, such as CuGaSe2 , CuGaS2 , they can be good prospects for splitting of water [2]. The energetic position of the band edges of any semiconductor/electrolyte interface system is controlled by the charge transfer of the Helmholtz inner layer [3,4], and any change in the charge modifies the flat band potential Ufb . The position of conduction and valence band edges at the semiconductor



Corresponding author. Tel.: +52 777 3623811; fax: +52 777 3623808. E-mail address: [email protected] (L.G. Arriaga).

0254-0584/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.matchemphys.2004.08.009

surface Ecs and Evs , respectively, can be obtained by measurements of capacitance C as a function of applied voltage E. A linear relationship C−2 versus E is expected to hold over a certain voltage range, when there is not charge transfer. An extrapolation of the plot (C−2 versus E) in x-axis permits to obtain the flat-band potential Ufb , when C−2 → 0. CuInS2 is a semiconductor material with a direct band gap close to 1.5 eV with a polycrystalline structure and it exhibits a high light absorption coefficient in the solar spectrum range, so it can be used as an absorber layer in solar cells [5]. Bhattacharya [6] showed n-type samples-electrodes of CuInS2 using a polysulphide solution under illumination. In the same way Herrero et al. [7] used a chopper and white light to determine n- and p-type CuInS2 layers using a polysulphide solution with NaOH. There are different techniques to prepare the compound, both physical and chemical methods. In this work, we used the electrodeposition technique to grow a Cu–In–S precursor thin film on a stainless steel substrate, then it was annealed in N2 atmosphere to improve its polycrystallinity. The experimental conditions for growing these films are reported elsewhere [8]. However, in this paper we used three

418

A.M. Martinez et al. / Materials Chemistry and Physics 88 (2004) 417–420

different electrolytic baths to prepare n- and p-type CuInS2 . The resulting films were analyzed by X-ray diffraction and their flat band potential and band edges were determined by capacitance–voltage measurements. The band gap was calculated by photocurrent measurements. The electrochemical characterization was done at pH = 3.

2. Experimental 2.1. Thin film preparation Cu–In–S layers were prepared using a plating bath of CuCl2 ·2H2 O, InCl3 ·4H2 O, Na2 S2 O3 ·5H2 O and LiCl, we used three different electrolytic baths as shown in Table 1. All solutions were prepared from analytical grade reagents that were dissolved in deionized water (18 M) with a buffer solution of pH = 3. A standard three-electrode cell was used; the working electrode was the Cu–In–S thin film, on its 316 stainless steel substrate (for Eg estimates the films were deposited on ITO/glass substrates), as counter-electrode a Pt mesh (5 cm2 ) was used, and a saturated calomel electrode (SCE) was used as reference electrode. The electrodeposition was controlled using an EG&G Mod. 263 Potentiostat/Galvanostat. The applied potential was 1 V (versus SCE) during 1800 s. The temperature was held at 30 ◦ C using a recirculation bath. The thickness of the samples was ∼1.5 ␮m average and was determined using an Alpha-Step Instrument model 100. After electrodeposition precursor thin films were annealed in a vacuum chamber during 1 h in a N2 atmosphere at 350 ◦ C.

sample edges. The total area expose of the electrode was 0.5 cm2 . The capacitance-voltage measurement was carried out at 10 kHz using a Solartron SI 1260 Impedance/Gain-Phase Analyzer connected with a Solartron 1287 Electrochemical Interface. The current–voltage and capacitance–voltage measurements were carried out in dark conditions. The photocurrent measurements were carried out using a 250 W Oriel tungsten halogen quartz lamp, a chopper, a monochromator and optical filters. The photosignal data were recorded by a PAR 5204 lock-in amplifier connected to a computer.

3. Results and discussion Fig. 1 shows the XRD pattern for annealed samples z1, z2 and z13, each prepared by a different bath (see Table 1). Most of diffraction peaks shown, correspond to the CuInS2 (JCPDS-270159), main peaks lie at 2θ = 27.32◦ , which is the (1 1 2) orientation, also it is important to note the presence of diffraction peaks of the stainless steel substrate. Diffractographs of the three films correspond to the roquesita phase, with a main intensity peak in the (1 1 2), (2 0 4) and (3 2 1) planes. The thickness of the samples was approximately of 1.5 ␮m.

2.2. Thin film characterization The X-ray diffraction (XRD) characterization was done by X-ray Difractometer Rigaku using a Cu K␣ radiation (λ ˚ The data obtained by XRD were compared with = 1.54 A). JCPDS file cards in order to determine the phases of the films. The preparation of the electrodes for electrochemical studies was as follows: Cu–In–S annealed thin film samples were covered with a silver epoxy solution and heated to 40 ◦ C for 4 h. The electric contacts of the samples were done by a copper wire, which was attached to the film with conductive silver paint. The body of the wire was encased in glass tubing and insulated using an epoxy coating that also covered the

Table 1 Bath composition used to prepare precursor Cu–In–S thin films by electrodeposition. All chemicals were dissolved in a buffer solution of pH 3 Sample no.

z1 z2 z13

Bath composition (mM) Cu2+

In3+

S2 O3 2−

LiCl

12.50 12.50 18.75

12.50 18.75 12.50

125 125 125

250 250 250

Fig. 1. XRD patterns of samples electrodeposited from solutions with various compositions of Cu2+ , and In3+ , and annealed in a controlled N2 Flow of 300 mTorrs and 350 ◦ C during 1 h before XRD analysis.

A.M. Martinez et al. / Materials Chemistry and Physics 88 (2004) 417–420

419

Fig. 2. Cyclic voltammograms of buffer solutions (pH = 3) using different annealed- electrodeposited thin films as working electrodes. Scan rate, 10 mV s−1 .

The current–voltage (I–E) plots in dark conditions of CuInS2 samples are shown in Fig. 2. From this figure it is possible to note that in the voltage range applied there is not any electrochemical reaction (oxidation–reduction), this is an indication that the films z1, z2 and z13 are chemically stable in this buffer solution. Fig. 3 shows Mott–Schottky (M–S) plots (1/C2 versus E), from this figure it is possible to note that there is a linear relationship over the voltage range applied. The slope of this line can be used to calculate the charger carrier density and the extrapolated linear portion where the line crosses the xaxis allows us to determine the flat band potential UfB . Also, it is possible to associate the slope of the same line with the conductivity type of the CuInS2 films, if the slope is negative, then the conductivity is p-type and for positive values we have an n-type material. Table 2 shows the flat band potential, carrier concentration, band gap and type of conductivity of z1, z2 and z13 films. Values of Ufb and carrier concentration are in good agreement with those reported previously [1]. The conductivity associated to the films was as follows. If we maintain the concentration of [S2 O3 2− ] constant, and change [Cu2+ ] or [In3+ ], the type of conductivity changes too:

Fig. 3. Mott–Schottky plots (f = 10 kHz) at pH value of 3 for different annealed-electrodeposited thin films.

(ii) When the concentration of copper ion in ternary bath solution was higher than Indium, for example [Cu2+ ]/[In3+ ] = 1.5, or Indium concentration is higher than copper; [In3+ ]/[Cu2+ ] = 1.5, the conductivity is ptype. These results agree with those reported before [9]. The optical transition of the films was obtained according to the well-known relationship α∝ (hν − Eg )n , where α is the absorbance, ␯ the frequency, h is the Planck constant, Eg is the optical band gap, and n varies from 0.5 to 2.0 depending on the nature of the band gap. To determine α we used values of specular reflectance (R) and transmittance (T) obtained by spectrophotometer equipment. In order to calculate the band gaps of the films the following relationship: α = (1/d) ln [(100 − %R)/T%] was used, where d is the film thickness. Here the reflection loss at the air-film interface is compensated. Direct band gaps were found from the extrapolation to zero of α2 versus hν as shown in Fig. 4, and they are reported in Table 2. The band gap of z1 is near to 1.53 eV a value reported before [1], however for sample z2 and z13 the band gaps are close

(i) When the [Cu2+ ] = [In3+ ], the type of the conductivity is −n.

Table 2 Values of flat-band potential, carrier concentration, band-gap and type of conductivity of CuInS2 thin films prepared by electrodeposition Sample no. z1 z2 z13

Flat-band potential Ufb vs. NHE (volts)

Carrier concentration (cm−3 )

Eg band gap (eV)

Type of conductivity

0.45 −0.66 −1.04

3.14 × 9.37 × 1016 3.9 × 1015

1.55 1.47 1.45

p n n

1016

420

A.M. Martinez et al. / Materials Chemistry and Physics 88 (2004) 417–420

that UfB is almost equal to the conduction band Ecs [3], in the point of zero charge. For a p-type, UfB ≈ the valence band Evs . The band edge potential of the films (Ecs and Evs ) and the redox potential for hydrogen reduction and oxygen evolution are shown in Fig. 5. This figure shows that any of the films match the oxygen evolution reaction, but the film z2 has closer values for oxygen evolution reaction. Perhaps minor modifications (e.g. pH, etc.) could give adequate conditions for splitting of water. 4. Conclusion

Fig. 4. Plots of α2 (108 cm−2 ) vs. hν (eV) for different annealedelectrodeposited thin films grown on ITO/glass substrates. Intercept lines show Eg -values.

to 1.45 eV, this might be due to differences in the atomic composition of the films and the polycrystallinity. The flat band potential and band gap values were used to calculate the conduction (Ecs ) and valence (Evs ) band edges of the films. For n-type films a good approximation used is

Precursor CuInS2 thin films were grown by electrodeposition technique on a stainless steel substrate. In order to improve their polycrystallinity the films were annealed in a vacuum chamber at different temperatures. According to the XRD patterns, the predominant phase corresponds to CuInS2 roquesite phase. It was found that p or n-type conductivity of CuInS2 thin films can be grown by changes in the concentration of electrolyte bath. If the ratio [Cu2+ ]/[In3+ ] equals 1, the conductivity is of the n-type, however if [Cu2+ ]/[In3+ ] or [In3+ ]/[Cu2+ ] is more than 1.5 the conductivity is p-type. The band gaps values found were from 1.45 to 1.55 eV, depending on the composition of the electrolytic bath. The occurrence of a depletion layer was seen and shown in the Mott–Schottky plot. The band edge position for z2 film has closer values for oxygen evolution reaction, being then a material of potential use in water splitting. Acknowledgements This work partially supported by DGAPA-UNAM under the project IN105399. Also one of the authors expresses his gratitude to DGEP-UNAM and CONACyT for the economical support. Part of this work was carried out within the Electrical Research Institute’s Hydrogen and Fuel Cells Program (IIE-11889). References

Fig. 5. Band edges for z1, z2, and z13 films and redox potential for hydrogen and oxygen evolution at pH 3.

[1] O. Madelung (Ed.), Data in Science and Technology, Semiconductors Other than Group IV Elements and III–V compounds, SpringerVerlag, Berlin, Heidelberg, New York, 1992. [2] T. Ohta, Solar-Hydrogen Energy Systems, Yokohama National University, Japan, 1979. [3] S.R. Morrison, Electrochemistry at Semiconductor and Oxide metal Electrodes, 1st edn., Plenum Press, New York, 1980. [4] J.A. Turner, J. Chem. Ed. 60 (1983) 327. [5] P. Rajaman, A.K. Sharma, A. Raza, O.P. Agnithori, Thin Solid Films 100 (1983) 111. [6] R.N. Bhattacharya, D. Cahen, G. Hodes, Solar Energy Materials 10 (1984) 41–45. [7] J. Herrero, C. Guillen, J. Appl. Phys. 69 (1991) 429–432. [8] A.M. Martinez, A.M. Fern´andez, L.G. Arriaga, The Electrochem. Soc., in press. [9] S.P. Grindle, C.W. Smith, S.D. Mittleman, Appl. Phys. Lett. 35 (1979) 24.