Chalcopyrite thin film solar cells by electrodeposition

Solar Energy 77 (2004) 725–737 www.elsevier.com/locate/solener

Chalcopyrite thin film solar cells by electrodeposition D. Lincot a,*, J.F. Guillemoles a, S. Taunier a, D. Guimard a, J. Sicx-Kurdi a, A. Chaumont a, O. Roussel a, O. Ramdani a, C. Hubert a, J.P. Fauvarque a, N. Bodereau a, L. Parissi a, P. Panheleux a, P. Fanouillere a, N. Naghavi a, P.P. Grand a, M. Benfarah a, P. Mogensen b, O. Kerrec a a

Laboratoire Cellules Solaires en Couches Minces (EDF-CNRS/ENSCP/UMR7575), Plateau CISEL, 6 Quai Watier-BP 49, 78401 Chatou cedex, France b Saint-Gobain Recherche, Aubervilliers, France Received 27 January 2004; accepted 20 May 2004 Available online 30 October 2004 Communicated by: Associate Editor T.M. Razykov

Abstract This paper reviews the state of the art in using electrodeposition to prepare chalcopyrite absorber layers in thin film solar cells. Most of the studies deal with the direct preparation of Cu(In,Ga)Se2 films, and show that the introduction of gallium in the films is now becoming possible from single bath containing all the elements. Electrodeposition can also be used to form precursor films with stacked layer structures, of pure elements or of combinations with binary or even ternary films. Thermal annealing treatments are of dramatic importance to provide suitable electronic quality to the layers. They are often done in the presence of a chalcogen (selenium or sulfur) over pressure and there is a tendency to use rapid thermal processes. Less studies are devoted to complete solar cell formation. Significant progresses have been made in the recent period with several groups achieving cell efficiencies around 8–10% on different substrates. A record efficiency of 11.3% is reported for a cell with an absorber presenting a band gap of 1.47 eV. First results on the manufacturability of the corresponding process to large areas are presented. 2004 Elsevier Ltd. All rights reserved. Keywords: CuInSe2; Cu(In,Ga)Se2; Electrodeposition; Solar cell; Chalcopyrite

1. Introduction It is more and more recognized that photovoltaic conversion of solar energy has to become as soon as possible a major source of the world energy supply in the future.

* Corresponding author. Tel.: +33 1 5542 6377; fax: +33 1 4427 6750. E-mail address: [email protected] (D. Lincot).

The photovoltaic industry will be a major industry producing solar modules with the square km as a reference area unit. All mature technologies, either based on silicon or on thin film chalcogenides (CIGS or CdTe), aim to reach this objective. For this, a key point is to develop large area processing at low module production costs while maintaining, or better, increasing the conversion efficiencies. Junctions based on Cu(In,Ga,Al)(S,Se)2 chalcopyrite absorbers have already demonstrated high conversion efficiencies and compatibility with large area

0038-092X/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.solener.2004.05.024

726

D. Lincot et al. / Solar Energy 77 (2004) 725–737

production of efficient modules at the pre-industrial level (Rau and Schock, 2001). Leading edge results are obtained with deposition and processing steps operating under high vacuum conditions, in particular for the absorber layer, like coevaporation or sputtering. It is recognized that significant breakdown of the production costs can be achieved if these steps could be replaced by less expensive ones based on non vacuum deposition technologies like screen printing or electrodeposition. Electrodeposition is already a major technology for mass production of large area metallic protective coatings in the industry, with impressive figures as for instance in the case of zinc coating by roll to roll processes (several meters per minute, deposition rates of tens of microns per minute, several meter large plates . . .). Applying successfully such a technique for the mass production of solar modules would be extremely well adapted for a large photovoltaic industry. However, even if electrodeposition is also used in the electronic industry (magnetic alloys, Cu plating in the Damascene process for integrated circuits) for metallic layers, its extension for preparing active semiconducting layers is still an open challenge. Semiconducting properties are by far more difficult to obtain than metallic properties since they involve the control of minority carriers at the ppm level, which is a priori much more difficult to achieve from a solution processes than from vacuum ones. Nevertheless researches have been devoted for long time to apply electrodeposition for the synthesis of semiconductors (Pandey et al., 1996; Hodes, 1995). The major breakthrough came with the preparation of CdTe layers for solar cells and modules. In this case about 20 years of research led to the realization of electrodeposited modules of CdTe, with world record efficiency of 10.6% on large area 0.9 m2 modules at BP Solar in 2001. This is a very stimulating example to develop the electrodeposition technology also for chalcopyrite solar cells (Vedel, 1998) or more generally solution technologies (Lincot et al., 1998). However CIGS is much more complicated than CdTe, as a quaternary compound, and the progression is facing many difficulties. Nevertheless, we can assist to a renewed interest for the electrodeposition as demonstrated by the steady growth in publication in the domain (40 papers since 2000, 28 for 1995–99, 20 for 1990–94, 12 for 1985–89, source: Inspec and Current Contents data bases) of CIGS, with some very encouraging results, such as the recent obtention of conversion efficiencies above 10% on small cells (Guimard et al.). In this paper we will give an overview of the state of the art in this domain and present some of the latest results of our own activities.

2. Low cost strategies for CIGS thin film solar cells Possible low cost strategies for CIGS thin film solar cells have been described several times which aim to re-

place the direct deposition of photovoltaic quality films by coevaporation by a two step process as illustrated in Fig. 1. In a first step a precursor layer is prepared by a low cost method, as electrodeposition or screen printing, which does not exhibit suitable electronic properties. These properties are provided in a second ‘‘functionalization’’ step which is based on a specific thermal treatment. The optimization of both the precursor properties (composition, structure) and the thermal annealing step (temperature, pressure, atmosphere, duration) should lead to final electronic p-type quality of the material comparable to those achieved with the one step process. The precursor layer can be already intermixed CIGS matrix or consist of stacked layers, either elemental or binary (Fig. 1). All these approaches have been explored for CIGS cells.

3. Electrodeposition studies in the CIGS system Electrodeposition for chalcopyrite films deals with the electrochemical behavior of four elements (Cu, In, Ga, Se), six if we consider in addition sulfur and tellurium. Each of these elements possesses specific electrochemical properties making the overall system already very complex, even more so when deposition is aimed at with several or all elements together in a single bath. The simplest situation is thus to deal with the preparation of elemental layers from single element solutions for preparing elemental layers (Cu,In,Ga,Se) and to change from one bath to the other for the preparation of precursor film with stacked layer structure. The thickness of each layer being easily controlled by coulometry, the overall composition of the precursor films is thus also controlled. Such an approach has been used for long time for preparing copper/indium bilayers which are then reacted thermally with a selenium atmosphere to form CISe (Kapur et al., 1987) or In/Se bilayers for indium selenide (Hodes et al., 1985). In Fig. 2, are reported schematic views of electrodeposition domains for single layers in acidic solutions. We can note that they possess very different electrodeposition potentials ranging from +0.75 V vs NHE (Normal Hydrogen Electrode) for selenium to 0.34 V for Cu, to 0.34 V for In, to 0.53 V for gallium. If Se and Cu are rather easy to form by reduction (excepted that Se tends to be insulating), indium is more difficult due to a rather negative value and the existence of the reduction of protons in parallel and gallium is much more difficult. Obtaining good quality layers often needs solution optimization (pH, concentration, complexing agent, potential . . .) as described in text books on metal deposition. The next stage is to use solutions containing two elements. The formation of copper indium is the first example of relevant combination (Herrero and Ortega, 1987; Prosini et al., 1996). In that case the precursor film is


727

Fig. 1. Flow chart for low cost strategy for chalcopyrite thin film solar cells. On the left side is the classical is the high cost one step process. On the right side is the low cost strategy, starting with a precursor preparation by electrodeposition or screen printing of a mixed chalcopyrite layer or stacked layers.

In3þ þ 3e ! In Se

Cu2þ þ 2e ! Cu Cu

ðIn=CuÞfilm ¼ kðIn=CuÞsolution CuxSe

In In2Se3

Ga

Ga 2Se3

usual deposition range -0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

Potential (V vs NHE)

Fig. 2. Simplified electrochemical diagrams showing the deposition domains in the Cu–In–Ga–Se system for elemental layers and binary compound layers (CuxSe, In2Se3, Ga2Se3). The solution is supposed to be acidic under standard conditions. The potential are given with respect to the normal hydrogen reference electrode (NHE). The usual potential deposition range is indicated as a dotted area.

intimately mixed ‘‘electrochemically’’ which can be an advantage with respect to stacked elemental layers. The deposition basically takes place for potentials more negative than that of the less noble element, here In. In the simplest case the composition is simply controlled by the concentrations in solution when the partial currents are additives and under diffusion limitation:

The second class of relevant combinations is when the two elements in solution are a metal and selenium (see for instance Massaccesi et al., 1996 for In2Se3; Massaccesi et al., 1993 and Thouin et al., 1993 for CuxSe). In that case in addition to the previous mechanism, we have to consider an additional effect due to the fact that many metallic selenide compounds can be formed due to their large formation energies (386 kJ/mol for In2Se3, 418 kJ/mol for Ga2Se3, 104 kJ/mol for Cu2Se for example). As shown in Fig. 2 this leads to a new deposition domains where these compounds are specifically formed. For instance the redox potential of the reaction: 2In3þ þ 3Se þ 6e ! In2 Se3 is shifted by an amount of DG/6F = +0.65 V with respect to the deposition potential of metallic indium. This means that the deposition of indium selenide can theoretically take place below 0.35 V. Similar mechanism takes place for gallium selenide or copper selenides. This mechanism, known as the KrogerÕs mechanism, is a key point for the electrodeposition of good quality materials, since it allows to self-regulate the composition of the film and is highly beneficial for improving the structural quality (this mechanism is also valid in the vapor phase) (Panicker et al., 1978). It appears to work with indium


for indium selenide (Massaccesi et al., 1996; Vedel, 1998) and copper indium diselenide electrodeposition (Thouin et al., 1994; Vedel, 1998). It works also remarkably well for cadmium telluride, allowing in some case near room temperature epitaxial growth (Lincot et al., 1995). In the case of copper selenides, the formation of well defined compounds is also evidenced, not only Cu2Se but also CuSe, Cu3Se2 (Thouin et al., 1993). Their formation depends on the applied potential and the solution composition. Interestingly the addition rule of partial currents is also valid for the copper selenide electrochemical system, allowing to monitor the layer composition precisely. The possibility to form separately indium selenide and copper selenide layers gives another method to prepare ternary precursor films with stacked binary layers. The interest is that selenium is also present in the film allowing to form the ternary CIS compound directly after annealing. Next approach is to try to form directly CIS precursor films by using a unique bath containing the three elements, as initiated by Bhattacharya (1983). This is by far the most investigated case due to the fact that such a process, often called one step electrodeposition is the most simple in principle, involving only one electrochemical process. The counter part is that the electrochemical aspects are becoming even more complex, with the possibility of forming either the elements in their elemental form, or as binary compounds in addition to the desired ternary CIS phase (Fig. 2). In fact the precursor films formed are generally a mixture of several phases in a nanocrystalline or amorphous state (Thouin and Vedel, 1995; Vedel et al., 1996). This is clearly seen by XRD studies or selective chemical etching studies (Guillen and Herrero, 1994). In addition to CIS, unreacted copper selenide and indium selenide domains can exist together in the precursor film. Only few studies deal with local structural and chemical characterizations by electron transmission measurements (Lincot et al., 1998). This intermixed structure can be used advantageously during the annealing process since it is very reactive. One of the few rational analysis of the correlations between the film composition and the deposition parameters has been published some years ago (Thouin et al., 1994; Thouin and Vedel, 1995). This led to the establishment of electrochemical phase diagrams as shown in Fig. 3. A key point of the proposed deposition mechanism which is widely accepted is that the composition of the film is controlled by the ratio between selenium and copper fluxes (a) to the electrode (addition of partial currents under diffusion limitation) which in turn control the insertion of indium by the KrogerÕs mechanism between excess selenium in the copper selenide film and indium ions. Such a diagram both shows the complexity of the Cu–In–Se system but also the great interest of one step electrodeposition for preparing in a rational way films with specific deviations

α

728

CuInSe2 + In2Se3 Se 2.5 + + In 2Se3 CuInSe2 CuInSe2 2

+

+

Se Se +

CuInSe2

1.5

CuSe

CuInSe2

CuSe

+ 1 0.5

Cu2 Se

CuInSe2 + Cu

Cu2 Se Cu2Se + Cu

0 -1.0

-0.9

-0.8

-0.7 -0.6 potential /V vs MSE

Fig. 3. Electrochemical phase diagram showing the composition of one step electrodeposited films in the Cu–In–Se system from acidic solutions as a function of the electrodeposition potential and the flux ratio between selenium(IV) and copper(II) ionic species in solution. The indium(III) concentration in solution is assumed to be in excess (Thouin et al., 1994; Thouin and Vedel, 1995; Guillemoles et al., 1996). The potentials are referred to the Sulfate Mercurous Electrode (MSE, 0.65 V vs NHE).

of composition for device optimization (Vedel et al., 1996). Similar diagrams were established for In–Se (Massaccesi et al., 1996) and Cu–Se (Thouin and Vedel, 1995) systems. Besides reference bath formulations involve simple sulfate media, many groups have investigated other formulations by replacing sulfate ions or adding other species, mostly organic, in the bath, as done for long time for the electroplating of metals. These species act as complexing agents for metallic ions or surface modifiers which can affect markedly the quality of the films or electrochemical kinetics. This is still an open avenue of optimization. Examples are chloride (often used) (Bhattacharya and Rajeshwar, 1986), nitrate (Bhattacharya et al., 1997) citrate (often used) (Pottier and Maurin, 1989), ammonium (Bhattacharya, 1983), triethanolamine (Bhattacharya, 1983; Bhattacharya and Rajeshwar, 1986), ethylenediamine (Pern et al., 1988), ethylenediaminetetraacetic acid (EDTA) (Ugarte et al., 2000), thiocyanate (Ganchev and Kochev, 1993; Tzvetkova et al., 1997; Kemell et al., 2000; Kemell et al., 2001), glycine (Ugarte et al., 2000). Out of these species, the case of thiocyanate, introduced by Kochev and coworkers (Ganchev and Kochev, 1993), is interesting since it is a strong complexing agent of copper(I). This allows to shift markedly the deposition potential of elemental copper as close as possible to that of indium and to use Cu(I) precursors instead of Cu(II) in solution. LeskelaÕs and coworkers tried to use this further to


improve the direct formation of CIS by extending the KrogerÕs mechanism, besides Indium, to copper deposition too (induced codeposition mechanism). Promising results show that a wide deposition potential range with self-regulated composition can be obtained either for copper selenide or copper indium diselenide depositions (Kemell et al., 2000, 2001). For long time the formation of the quaternary phase with gallium by one step electrodepostion was not reported. Our own experience in the past showed that the insertion of gallium in the standard electrodeposition potential domain (near 0.35 V) was very difficult. This is a direct consequence of the negative shift of the redox potential for gallium deposition as compared to indium which is reflected for gallium selenide vs indium selenide (Fig. 2). However in the recent period this serious bottleneck seems to have been overcomed. Several groups have now reported the successful insertion of gallium in the films up to amounts corresponding to the optimal values for the preparation of high efficiency cells (6– 10%). Bhattacharya et al., were the first to show Ga insertion, but at a very low level (Ga/In = 0.1), from a chloride bath (Bhattacharya et al., 1997). Note that they mentioned the use of a superimposed alternating voltage (at 20 kHz). Moreover the films were very copper rich (Cu/(In + Ga) = 2.1). They realized the breakthrough by adding a pH buffer in the chloride bath, consisting in a mixture of sulphamic acid and potassium biphtalate at pH = 3 (‘‘hydrion’’ buffer) (Bhattacharya and Fernandez, 2001, 2003). By changing the solution composition (for the same deposition potential) they were able to monitor the ratio Ga/In from 0.3 up to 0.7, with only slightly copper rich films, demonstrating the real possibility of preparing one step CIGS layers over a wide range of as grown compositions suitable for efficient solar cells. Fahoume et al. were also able to monitor the gallium insertion up to high levels by using the chloride bath but playing with the deposition potentials towards more negative values, even below the gallium deposition potential value (In/Ga = 0, 0.4, 1 at 0.38 V, 0.68 V, 0.78 V respectively) (Fahoume et al., 2001). Interestingly the formation of CIGS is checked by the shift of XRD peak position. Kampmann et al. succeeded in introduction gallium in CIS from a sulfate bath, and indicate a Ga/In ratio of 0.11 (Kampmann et al., 2000), Zhang et al. also achieved gallium insertion by this way (Zhang et al., 2003). These results are very encouraging for device developments but up to now a precise analysis and interpretation of gallium electrochemical insertion phenomenon is lacking. We can note that CIGS films formation have been reported by using a stacked layer/annealing approach, with first the deposition of a CuGa alloyed film and then a ternary CIS film (Friedfeld et al., 1999). Only a few studies concern the introduction of sulfur in one step electrodeposited CuInS2 films. Bhattacharya,

729

Hodes and Cahen have deposited copper indium disulfide films by using thiourea as the dissolved precursor for sulfur (Hodes et al., 1985; Bhattacharya et al., 1984). Yukawa et al. (1996), Nakamura and Yamamoto (1997) report the possible use of sulfite anions as the precursor for sulfur. This lack of development on one step electrodeposition of sufide chalcopyrites may be due to the specificity of the electrochemical behavior of sulfur, which is more irreversible than for its Te or Se analogues. For sulfide chalcopyrites the sequential route is more developed, which consists on the electrodeposition of Cu/In layers followed by annealing under hydrogen sulfide or sulfur atmosphere (Hodes et al., 1985; Herrero and Ortega, 1990; Nakamura and Yamamoto, 2003). Concerning telluride based chalcopyrite, one step electrodeposition of CuInTe2 has been reported by Ishizaki et al. (2000, 2004). Note that the electrodeposition of CuIn(Se,Te)2 (Bhattacharya and Rajeshwar, 1986; Diaz et al., 2000) and CuIn(S,Se)2 (Edamura and Muto, 1995; Garg et al., 1991) ternaries have been reported. Concerning silver based chalcopyrite, the electrodeposition of AgInSe2 has been reported (Raviendra and Sharma, 1985). It thus appears the one step electrodeposition of chalcopyrites is mainly devoted to copper indium gallium selenides, but that many other related compound have entered the real accessibility by one step electrodeposition.

4. Thermal annealing post-treatments As indicated in the previous paragraph, none of the precursor films present as grown electronic quality. Thus thermal annealing post-treatments are needed to form device quality films. Thermal treatments are thus a very important aspect to deal with. Their characteristics will depend on the nature of the precursor layer. Stacked elemental layers, staked binary layers, single multiphase layer, will obviously not recrystallize and homogeneize by the same mechanisms during the annealing. Annealing under neutral atmosphere or vacuum (400–550 C) has been proved to improve markedly the structural and optical properties of one step electrodeposited films (see for instance Sratieva et al., 1997; Guillen and Herrero, 1994, 1996a,b), however the electronic properties are generally not sufficient (Lincot et al., 1994; Guillemoles et al., 1996). For this, a sufficient activity of selenium in the vapor phase during annealing is needed. This was clearly proved by early experiments in our group for instance where the activity of the selenium was varied while the temperature of the sample was maintained the same (Guillemoles et al., 1994a,b). This resulted in a dramatic improvement of the electronic quality (as shown by electrical transport and luminescence measurements) when the selenium partial pressure was increased. Not only


the required p-type character of the layer was achieved but also the structural quality of the layer was improved dramatically as compared to annealing treatments under neutral atmosphere as also observed in (Tzvetkova et al., 1997; Guillen and Herrero, 1996a,b). The crystallization was also dependent on the composition of the precursor films (Lincot et al., 1994; Guillemoles et al., 1994a, 1996). Nowadays many groups working with annealing treatments of electrodeposited precursor films report the use of a selenium over pressure (Kampmann et al., 2000; Bhattacharya et al., 1999; Calixto et al., 1999; Chaure et al., 2004). Annealings are generally carried out in close chamber (Kampmann et al., 2000; Guillemoles et al., 1996). However some groups are using annealing treatments in an evaporation chamber, since additional amount of In or Ga are coevaporated on the surface to readjust the composition (Bhattacharya et al., 1999; Calixto et al., 1999). Such an approach can be considered as a hybrid approach where the electrodeposition provides most of the material needed for the devices by a non vacuum method, the high vacuum step being restricted to the minimum amount of evaporated material. In the case of stacked metallic layers, annealing under selenium (or H2Se) or sulfur (or H2S) pressure leads to the formation of the desired chalcopyrite film (Kapur et al., 1987; Hodes et al., 1985; Herrero and Ortega, 1990; Nakamura and Yamamoto, 2003). Annealing treatments are generally carried out for temperatures ranging from 450 C to 650 C from a few minutes to one hour or more. The differences can be explained by the differences between the precursor layers. It is anticipated that forming the single phase compound from stacked layers is more demanding in terms of temperature or duration than from a single ‘‘electrochemically’’ pre-mixed layer. The presence of gallium may also affect the recrystallization kinetics as shown by experiments on evaporated layers. An optimization is thus needed for every specific absorber layer. One can note in the recent period the development of rapid thermal annealing processes (RTP) for electrodeposited precursor layers (Chaure et al., 2004; Friedfeld et al., 1999). RTP has been popularized by its use by Shell Solar for module production (Palm et al., 2003).

5. Devices Fig. 4 present the evolution of efficiencies on one step electrodeposited chalcopyrite solar cells. 5.1. One step electrodeposited CIGS layers in substrate configuration The first report on devices made with one step electrodeposited CIS were made in 1988 by Mc Gill University (Qiu and Shih, 1988) and NREL (Pern et al., 1988) groups, with efficiency of 5.2% and 1.9% respectively on

NREL/PVD

14

Efficiency %

730

NREL/PVD NREL/PVD

12

CISEL

10

CISEL NREL/PVD

NREL/PVD CISEL

8

McGill McGill

6

ENSCP

CIS Solartech

Hannover

4 2

Atotech

NREL

1990

2000

Year

Fig. 4. Evolution of efficiencies on one step electrodeposited chalcopyrite solar cells as a function of time. Black symbols correspond to one step electrodeposited layer with metals readjustment by coevaporation.

Mo/glass substrates. We can note that in both cases the annealing was done in an argon atmosphere or vacuum. In 1994 our group reported the use of the annealing treatment for one step electrodeposited CIS under high elemental selenium pressure, with an efficiency of 5.6% (Guillemoles et al., 1994a) and 6.5% (Lincot et al., 1994). This result was highlighted independently as introducing a promising industrial approach (Highligths, Chemistry and Industry, August 1995). In same period the Mc Gill university group also reached 7% efficiency CIS cells, with a non reactive annealing (Qiu et al., 1994, 1995). Surprisingly one had to wait up to 2000 for having new announcements on efficient cells made on one step electrodeposited CIS. The Hannover group reported an efficiency of 4.8% (Kampmann et al., 2000). The interesting point is that they indicated that gallium was present in the film at a level of 10–15%. This was probably the first time that one step electrodeposited CIGS layer led to efficient device, using again annealing under selenium atmosphere (corresponding to a temperature of 280 C while the substrate temperature was between 300 and 400 C). However the open circuit voltage was lower than expected for CIGS with a value of 363 mV. In 2002 an efficiency of 4% was also demonstrated with one step electrodeposited CIS by the Atotech group in collaboration with Shell and HMI (Voss et al., 2002). The thermal annealing treatment was done with the rapid thermal annealing process. Cross-section views show that the absorber layer was with small grains even after annealing. 5.2. The hybrid ‘‘electrodeposition/evaporation’’ approach In 1996 the NREL group made a breakthrough by demonstrating that one step electrodeposited precursor


layers, when post-treated in a classical evaporation chamber, where able to give 9.4% efficient device (Bhattacharya et al., 1996). However final composition of the film was significantly readjusted by evaporating given amounts of In or Cu and Se. This demonstrated that bulk properties of electrodeposited layers were of device quality and that no strong detrimental impurity effect was present from the bath. In 1997, the NREL group reached a new record of 12.3% with the same hybrid approach but with gallium containing CIGS absorbers (Bhattacharya et al., 1997). However only a very small part of the gallium was introduced already from the deposition step, the remaining one was introduced by PVD together with In. In 1998, the record value was 14.1% with copper rich as grown electrodeposited films with still very low gallium content (Bhattacharya et al., 1998). In 1999, this record was further raised up to an efficiency of 15.4%. The amounts of added Ga (300 nm) and In (720 nm) (Bhattacharya et al., 1999). With such levels of added Ga and In, these results can be considered only as proof of concept. Then the group tried to minimize the amount of material added by PVD. The used of buffered solution (pH-hydrion) allowed a significant increase of the gallium concentration in the precursor film. Only 200 nm of In were then added by PVD. The efficiency was 9.4% (Bhattacharya and Fernandez, 2001, 2003). This result is more relevant to the electrodeposition strategy. 5.3. Electrodeposited cells with superstrate configuration Contrary to CdTe cells, CIGS cells are classically with the substrate structure Glass/Mo/CIGS/CdS/ZnO. The development of superstrate configuration for electrodeposited cells is also investigated but to a lesser extent (Abken et al., 1998; Kampmann et al., 1999, 2000). Indium tin oxide substrates with all electrodeposited indium selenide/CIS layers led to 3.2% efficiency in 1998 (Abken et al., 1998). A 3.6% was announced later by the same group (Kampmann et al., 2000). More recently values in the range of 6% were claimed but with a cadmium sulfide buffer layer (Chaure et al., 2004). 5.4. Electrodeposited copper indium disulfide solar cells Contrary to selenide cells, less work has been published on their sulfide counter parts, due probably to the difficulty to prepare CuInS2 precursor films by one step electrodeposition. The preferred route seems to be the sulfidation of alloyed or stacked Cu/In electrodeposited layers. The first approach has been reported recently with annealing in a sulfur atmosphere but the efficiency is low (1.3%) (Herrero and Ortega, 1990; Nakamura and Yamamoto, 1997). An efficiency of 9% was claimed recently (Takeuchi et al., 2003). However we can mention that this process is similar to that developed by the HMI

731

group using sputtering deposition for the Cu and In layers, and RTP annealing, with efficiencies up to 11.4% (Siemer et al., 2001). We can note that electrodeposited stacked Cu/In layer approach has been tested initially for CuInSe2 using an annealing step under hydrogen selenide pressure, the achieved efficiency was 7% (Kapur et al., 1987). This process but with In and Cu deposited by sputtering, further lead to high efficiency cells and modules based on Cu(In,Ga)(S,Se)2 in Shell Solar and Showa Shell (Palm et al., 2003). 5.5. Electrodeposited cells on metal substrates The use of metallic substrate is very attractive for electrodeposited CIGS cells in order to develop a roll to roll industrial process. Electrodeposited In layers on copper tape, with rapid sulfurization is the basis of the CISCuT technology with efficiencies of about 6% (Pendorf et al., 1998). Results were presented in 2003 by the group of CIS Solartechnik. They use Mo, Cu or stainless steel (SS) flexible substrates. They explored the option of stacked elemental layers (Cu, In, Ga by electrodeposition and Se by evaporation) and that of one step CIS. The use of rapid thermal annealing (RTP) under nitrogen at 580620 C was successful. They reach promising efficiencies of 9% for the stacked electrodeposition route (on SS/Mo) and 7.5% for the one step electrodeposition one (on SS/Cr/Mo) (Kampmann et al., 2003). Later an efficiency of 9% was reported on Cu foil/TaN/Mo (Rechid et al., 2003). An efficiency of 10.2% was reported in June 2004 (Kampmann et al., 2004).

6. Results of the CISEL project Even if the hybrid approach presented before can be beneficial in terms of cost reductions, eliminating any post-vacuum deposition step is still the most adapted objective, with the conditions that sufficient efficiencies can be achieved. This is the aim of the CISEL project (Copper Indium Selenide by ELectrodeposition) involving CNRS/ENSCP and EDF, together with St Gobain Recherche with the support of ADEME. Research is directed along to main research lines: • increasing the efficiency of laboratory cells, • transferring the electrodeposition step to large area processing.

6.1. Results on laboratory cells Concerning the work on laboratory cells, an efficiency of 8.8% has been reached in 2002 for pure copper

732


indium diselenide absorbers (Guimard et al., 2002). Fig. 5 shows a SEM view in cross-section of a typical high efficiency electrodeposited cell. One can note that the absorber layer presents large grains of micron size which extend all through the thickness. The best device parameters were (under 100 mW/cm2) Jsc = 32.7 mA/cm2, Voc = 0.416 V and FF = 65% for an area of 0.06 mm2 representing a new record for electrodeposited cells without post-evaporation step. Average efficiency of 4.5% was obtained on 25 cm2 areas. In Fig. 6 is shown the spectral response of the cell, compared to that of a standard coevaporated CIGS Cell from ZSW. The quantum efficiencies in the visible wavelength range are comparable. The absorption onset is shifted towards higher wavelengths due to the small band gap of pure CIS. The analysis of the device by studying the relationship between the open circuit voltage and the short circuit current density allowed to determine a value of 1.7 for the ideality factor, an activation energy of 1.1 eV for the dark current and a value of 1.5 mA/cm2 for the pre-exponential factor of the saturation current. These values are comparable to those measured with standard CIS and CIGS devices. The result suggested the presence of strong inversion at the interface between the CIS and the CdS layers and that recombination losses where taking place in the space charge layer. However, as also observed for pure CIS cells, the open circuit voltage is rather low. To increase it is very important to increase the band gap of the absorber. We succeeded in the elaboration of absorber layers with band gap values between 1 eV and 1.63 eV and using these layers for complete junction formation. In all cases efficiencies in excess to 8% were achieved. Fig. 7 shows the variation of the open circuit voltage of the cells as a function of the band gap value, as compared to that

1 0.8 0.6 QE 0.4 CISEL 0.2 0

coevaporated 0.6

0.8

1 Lambda (m)

1.2

1.4

Fig. 6. Spectral response of a 8.8% pure CIS electrodeposited cell compared to a reference CIGS coevaporated cell from ZSW (Guimard et al., 2002).

obtained on evaporated cells from literature data. There is a good agreement between both set of results, indicating a good quality of the electrodeposited cells. Above Eg = 1.3 eV, the rate of increase of Voc with Eg becomes smaller. This change of slope has been observed as well by others (Rau et al., 2001), and accounted for by a change of the dominant recombination process, from space charge recombination to tunnelling enhanced recombination. With this strategy a new efficiency record value of 10.2 % has been obtained (Guimard et al., 2003a,b). The open-circuit voltage, short-circuit current density and Fill Factor

1 0.9 0.8

Voc (V)

0.7 0.6 0.5 0.4

Electrodeposited

0.3

evaporated

0.2

Voc max

0.1

Voc(1.5)

0 1.00

1.20

1.40

1.60

Gap (eV)

Fig. 5. Typical SEM view in cross-section of efficient electrodeposited CIS cells prepared within the CISEL project, showing the recrystallized absorber layer and CdS/ZnO top layers.

Fig. 7. Comparison of open circuit voltages obtained with electrodeposited (this work) and coevaporated cells (from literature data) over a wide absorber band gap range (Guimard et al., 2003a,b).


6.2. Up-scaling studies Concerning scale up studies, the area of the CuInSe2 precursors have been increased from the 5 · 5 cm2 initial state (laboratory), to an intermediate 10 · 10 cm2 state and finally 30 · 30 cm2. Two reactors were designed and built, in order to process 100 and 900 cm2 substrates.

1.02 5

normalized standard deviation = 1.3%

%Cu (normalized)

were respectively equal to 741 mV, 23.2 mA/cm2 and 59.6%. However the series resistance was high, with a value of 2.4 X cm2, and accounts partly for the low fill factor obtained. Quantum efficiencies from 0.8 to about 0.9 were measured between 0.8 lm and 0.55 lm. The band gap value of this absorber was 1.47 eV. We noted that the ideality factor A (2.37) was slightly higher than for pure CIS cells. Admittance measurements allowed to determine the distribution of traps in the band gap of the material. Two acceptor peaks are found at about 0.27 eV and 0.45 eV from the valence band. In November 2003 a new record efficiency of about 11.3% on 0.1 cm2 has been obtained by optimization of all the procedures. Fig. 8 shows the IV curve and the spectral response of the record cell.

733

10 15

1.01

20 25 30

1

0.99

0.98 5

10

15

20

25

30

30 cm

Fig. 9. Mapping of the Cu content as a function of position on a 30 · 30 precursor. The compositions were measured by XRF, and the data shown here were normalized relative to the average.

Table 1 Lateral homogeneity assessment Standard deviation/average value (%) Thickness Cu content In content Se content

6.9 1.3 2.1 1.2

Standard deviations as a function of position on a 30 cm · 30 cm precursor.

To illustrate the homogeneity of the precursor thickness and composition, Fig. 9 shows a mapping of the Cu content on a 30 · 30 cm2 plate. The compositions were measured by XRF, and the data shown here were

90 130 precursors (100 cm2) : 410 measurements

80 average = 1.91 µm std deviation= 0.15 µm (7.6%)

Occurence

70 60 50 40 30 20 10 0 1.4 1.5 1.6 1.7 1.8 1.9

2

2.1 2.2 2.3 2.4 2.5 2.6

Thickness (µm)

Fig. 8. (Top) I–V characteristic of the record cell with 11.3% efficiency. (Bottom) Spectral response corresponding to a band gap value of the absorber of 1.47 eV.

Fig. 10. Thickness distribution of the 100 cm2 CIS precursors controlled by X-ray fluorescence (XRF). The statistics are made from 410 measurements, on 130 different precursors.

734


Standard deviation/average value (%)

processing. In parallel the transfer of all individual steps to large area processing is very important to demonstrate manufacturability.

3.5 5.6 2.2

Acknowledgments

Table 2 Reproducibility assessment

Cu content In content Se content

2

Standard deviations for 130 precursors (100 cm ).

normalized relative to the average. Table 1 shows the standard deviations with respect to the three elements. The deposited layer is very homogeneous, with relative standard deviations below 5% for the compositions. The reproducibility of the process has been addressed on 100 cm2 plates. Fig. 10 shows an histogram calculated from 130 different precursors, 100 cm2 in size (410 measurements performed). The standard deviation is 0.15 lm, out of an average 1.91 lm, i.e., 7.6% relative. The thickness of the precursor was measured by X-ray fluorescence (XRF), which is very sensitive to small changes in the layer density. This could explain the relatively ‘‘high dispersion’’ of the results. The normalized deviations of the compositions are also given in Table 2. The values are in the percent range, which illustrates the good reproducibility of the process.

7. Conclusion Electrodeposition of CIS cells is experiencing a renewed interest as a possible option for decrease the production costs of CIS modules. Several approaches are developed successfully by different groups, using different types of precursor layers, from stacked elemental layers prepared from different baths to multinary layers prepared from a single bath. Important results are (1) successful introduction of gallium to form CIGS absorbers over wide composition range; (2) cell efficiencies over 11%. Several processes leading to cells around 810%; (3) encouraging results on homogeneity and reproducibility for scaling up the one step electrodeposition process to 30 · 30 cm2. Future directions of investigation should aim first to further increase the efficiency of electrodeposited cells, with as limited and simple preparation steps. This needs a better understanding of electrochemical formation of quaternary CIGS absorbers, of annealing treatments, especially rapid thermal processes, and of complete cell

The authors would like to acknowledge the support of the CNRS (Energy Program, Dele´gation aux Entreprises) and EDF R&D. The French Agency for Environment and Energy Management (ADEME) is also acknowledged for the support to the CISEL project. The Zentrum fu¨r Sonnenenergie und Wasserstoff-Forschung (ZSW) in Stuttgart is acknowledged for providing reference samples.

References Abken, A., Heinemeyer, F., Kampmann, A., Leimku¨hler, G., Rechid, J., Sittinger, V., Wietler, T., Reineke-Koch, R., 1998. Large area electrodeposition of Cu(In,Ga)Se2 precursors for the fabrication of thin film solar cells. In: Schmid, J., et al. (Eds.), Proceedings of the Second World Conference and Exhibition on Photovoltaic Solar Energy Conversion, pp. 1133–1136. Bhattacharya, R.N., 1983. Solution growth and electrodeposited CuInSe2 thin films. J. Electrochem. Soc. 130, 2040– 2042. Bhattacharya, R.N., Fernandez, A.M., 2001. CuInGaSe2 based photovoltaic cells from electrodeposited precursor films. In: Birkmire, R., et al. (Eds.), II–VI compound semiconductor photovoltaic materials, Proceedings Mat. Res. Symp 668, pp. H8.10.1–H8.10.6. Bhattacharya, R.N., Fernandez, A.M., 2003. CuIn1xGaxSe2based photovoltaic cells from electrodeposited precursor films. Solar Energy Mater. Solar Cells 76, 331–337. Bhattacharya, R.N., Rajeshwar, K., 1986. Electrodeposition of CuInX (X = Se, Te) thin films. Solar Cells 16, 237–243. Bhattacharya, R.N., Batchelor, W., Hiltner, J.F., Sites, J.R., 1999. Thin-film CuIn1xGaxSe2 photovoltaic cells from solution-based precursor layers. Appl. Phys. Lett. 75, 1431– 1433. Bhattacharya, R.N., Batchelor, W., Wiesner, H., Hasoon, F., Granata, J.E., Ramanathan, K., Alleman, J., Keane, J., Mason, A., Matson, R.J., Noufi, R.N., 1998. 14.1% CuIn1 xGaxSe2-Based photovoltaic cells from electrodeposited precursors. J. Electrochem. Soc. 145, 3435–3440. Bhattacharya, R.N., Cahen, D., Hodes, G., 1984. Electrodeposition of Cu–In–S layers and their photoelectrochemical characterization. Solar Energy Mater. 10, 41–45. Bhattacharya, R.N., Fernandez, A.M., Contreras, M.A., Keane, J., Tennant, A.L., Ramanathan, K., Tuttle, J.R., Noufi, R.N., Hermann, A.M., 1996. Electrodeposition of In–Se, Cu–Se, and Cu–In–Se thin films. J. Electrochem. Soc. 143, 854–858.

D. Lincot et al. / Solar Energy 77 (2004) 725–737 Bhattacharya, R.N., Wiesner, H., Berens, T.A., Matson, R.J., Keane, J., Ramanathan, K., Swartzlander, A., Mason, A., Noufi, R.N., 1997. 12.3% efficient CuIn1xGaxSe2-based device from electrodeposited precursor. J. Electrochem. Soc. 144, 1376–1379. Calixto, M.E., Sebastian, P.J., Bhattacharya, R.N., Noufi, R., 1999. Compositional and optoelectronic properties of CIS and CIGS thin films formed by electrodeposition. Solar Energy Mater. Solar Cells 59, 75–84. Chaure, N.B., Young, J., Samantilleke, A.P., Dharmadasa, I.M., 2004. Electrodeposition of p-i-n type CuInSe2 multilayers for photovoltaic applications. Solar Energy Materials and Solar Cells 81, 125–133. Diaz, R., Merino, J.M., Rueda, F., Ocon, P., Herrasti, P., 2000. Simultaneous electrodeposition of CuInSeTe thin films. J. Mater. Chem. 10, 1623–1627. Edamura, T., Muto, J., 1995. Fabrication and characterization of quaternary chalcopyrite CuIn(S,Se)2 thin films by pulse plating. J. Mater. Sci. Lett. 14, 1400–1402. Fahoume, M., Chraibi, F., Aggour, M., Delplancke, J.L., Ennaoui, A., Lux-Steiner, M., 2001. Electrodeposition of CuInGaSe2 thin films. In: Nellis, Mc et al. (Eds.), Proceedings of the 17th European Photovoltaic Solar Energy Conference, pp. 1247–1250. Friedfeld, R., Raffaelle, R.P., Mantovani, J.G., 1999. Electrodeposition of CuInxGa1xSe2 thin films. Solar Energy Mater. Solar Cells 58, 375–385. Ganchev, M.G., Kochev, K.D., 1993. Investigation of the electrodeposition process in the Cu–In–Se system. Solar Energy Mater. Solar Cells 31, 163–170. Garg, P., Garg, A., Garg, J.C., 1991. Structural and optical characterization of electrodeposited chalcopyrite CuInSxSe2x thin films. Thin Solid Films 206, 236–240. Guillen, C., Herrero, J., 1994. Effects of thermal and chemical treatments on the composition and structure of electrodeposited CuInSe2 thin films. J. Electrochem. Soc. 141, 225– 230. Guillen, C., Herrero, J., 1996a. Improved selenization procedure to obtain CuInSe2 thin films from sequentially electrodeposited precursors. J. Electrochem. Soc. 143, 493– 498. Guillen, C., Herrero, J., 1996b. Improvement of the optical properties of electrodeposited CuInSe2 thin films by thermal and chemical treatments. Solar Energy Mater. Solar Cells 43, 47–57. Guillemoles, J.F., Cowache, P., Massaccesi, S., Thouin, L., Sanchez, S., Lincot, D., Vedel, J., 1994a. Electrodeposited copper indium diselenide thin films based solar cells with improved efficiencies. Adv. Mater. 6, 379–6381. Guillemoles, J.F., Lusson, A., Cowache, P., Massaccesi, S., Vedel, J., Lincot, D., 1994b. Recrystallisation of electrodeposited copper indium diselenide thin films under selenium pressure. Adv. Mater. 6, 376–379. Guillemoles, J.F., Cowache, P., Lusson, A., Fezzaa, K., Boisivon, F., Vedel, J., Lincot, D., 1996. One step electrodeposition of CuInSe2: improved structural, electronic, and photovoltaic properties by annealing under high selenium pressure. J. Appl. Phys. 79, 7293–7302.

735

Guimard, D., Grand, P.P., Bodereau, N., Cowache, P., Guillemoles, J.F., Lincot, D., Taunier, S., Ben Farah, M., Mogensen, P., Kerrec, O., 2002. Copper indium diselenide solar cells prepared by electrodeposition. In: Proceedings of the 29th IEEE Photovoltaic Specialists Conference, pp. 692–695. Guimard, D., Bodereau, N., Kurdi, J., Guillemoles, J.F., Lincot, D., Grand, P.P., Ben Farah, M., Taunier, S., Kerrec, O., Mogensen, P., 2003a. Efficient and low cost CIS solar cells prepared by electrodeposition without vacuum post deposition step. In: Noufi, R., et al. (Eds.), Compound semiconductor photovoltaics symposium. Proceedings of the Materials Research Society, vol. 763, pp. 281–286. Guimard, D., Bodereau, N., Kurdi, J., Guillemoles, J.F., Lincot, D., Grand, P.P., Ben Farah, M., Taunier, S., Kerrec, O., Mogensen, P., 2003b. Efficient CIS solar cells prepared by electrodeposition. In: Kurohawa, K., et al. (Eds.), Proceedings of the Third World Conference on Photovoltaic Energy Conversion, vol. 1, pp. 515–518. Herrero, J., Ortega, J., 1987. Electrochemical synthesis of photoactive In2Se3 thin films. Solar Energy Mater. Solar Cells 16, 477–485. Herrero, J., Ortega, J., 1990. Electrodeposition of Cu–In alloys for preparing CuInS2 thin films. Solar Energy Mater. 20, 53–65. Hodes, G., 1995. Electrodeposition of II–VI semiconductors. In: Rubinstein, I. (Ed.), Phys. Electrochem.. Marcel Dekker, New York, pp. 195–295. Hodes, G., Engelhard, T., Cahen, D., Kazmerski, L.L., Herrington, C.R., 1985. Electroplated CuInS2 and CuInSe2 layers: preparation and physical and photovoltaic characterization. Thin Solid Films 128, 93–106. Ishizaki, T., Saito, N., Yatei, D., Fuwa, A., 2000. Behavior of Cu–In–Te electrodeposition from acidic solution. In: Proceedings of the Second International Conference on Processing Materials, pp. 281–284. Ishizaki, T., Saito, N., Fuwa, A., 2004. Electrodeposition of CuInTe2 films from acidic solution. Surf. Coat. Technol. 182, 156–160. Kampmann, A., Cowache, P., Lincot, D., Vedel, J., 1999. Junction formation study of one step electrodeposited CuInSe2/CdS. J. Electrochem. Soc. 146, 150–155. Kampmann, A., Sittinger, V., Rechid, J., Reineke-Koch, R., 2000. Large area electrodeposition of Cu(In,Ga)Se2. Thin Solid Films 361, 309–313. Kampmann, A., Rechid, J., Raitzig, A., Wulff, S., Mihhailova, M., Thyen, R., Kalberlah, K., 2003. Electrodeposition of CIGS on metal substrates. In: Noufi, et al. (Eds.), Compound Semiconductors Photovoltaics. Proceedings of the Materials Research Society Symposium, vol. 763, pp. 323– 328. Kampmann, A., Rechid, J., Thyen, R., Wulff, S., Mihhailova, Kalberlah, K., 2004. In: 19th European Photovoltaic Solar Energy Conference and Exhibition, Paris. Kapur, V.J., Basol, B.M., Tseng, E.S., 1987. Preparation of thin films of chalcopyrites for photovoltaics. In: Hwang, H.L., et al. (Eds.), Proceedings Seventh International Conf. Ternary and Multinary Compounds, pp. 219–224.

736


Kemell, M., Ritala, M., Saloniemi, H., Leskela, M., Sajavaara, E., Rauhala, E., 2000. One step electrodeposition of Cu2xSe and CuInSe2 thin films by the induced co-deposition mechanism. J. Electrochem. Soc. 147, 1080–1087. Kemell, M., Saloniemi, H., Ritala, M., Leskela, M., 2001. Electrochemical quartz microbalance study of the electrodeposition mechanism of CuInSe2. J. Electrochem. Soc. 148, C110–C118. Lincot, D., Guillemoles, J.F., Cowache, P., Marlot, A., Lepiller, C., Canava, B., Yousfi, E.B., Vedel, J., 1998. Solution deposition technologies for thin film solar cells: status and perspectives. In: Schmid, J., et al. (Eds.), Proceedings of the Second World Conference and Exhibition on Photovoltaic Solar Energy Conversion, pp. 440–445. Lincot, D., Guillemoles, J.F., Cowache, P., Massaccesi, S., Thouin, L., Fezzaa, K., Boisivon, F., Vedel, J., 1994. Tailoring the electronic properties of one step electrodeposited CuInSe2 films by annealing in Se vapor under controlled activity conditions. Elaboration of efficient solar cells. Proceedings of the first world conference and exhibition on photovoltaic solar energy conversion, pp. 136–139. Lincot, D., Kampmann, A., Mokili, B., Vedel, J., Cortes, R., Froment, M., 1995. Epitaxial electrodeposition of CdTe films on InP from aqueous solutions: role of a chemically deposited CdS intermediate layer. Appl. Phys. Lett. 67, 2355–2357. Massaccesi, S., Sanchez, S., Vedel, J., 1993. Cathodic deposition of copper selenide films on tin oxide in sulfate solutions. J. Electrochem. Soc. 140, 2540–2545. Massaccesi, S., Sanchez, S., Vedel, J., 1996. Electrodeposition of indium selenide In2Se3. J. Electroanal. Chem. 412, 95– 101. Nakamura, S., Yamamoto, A., 1997. Preparation of CuInS2 with sufficient sulfur content and excellent morphology by one step electrodeposition. Solar Energy Mater. Solar Cells 49, 415–421. Nakamura, S., Yamamoto, A., 2003. Electrodeposited CuInS2 based thin film solar cells. Solar Energy Mater. Solar Cells 75, 81–86. Palm, J., Probst, V., Brummer, A., Stetter, W., Tolle, R., Niesen, T.P., Visbeck, S., Hernandez, O., Wendl, M., Vogt, H., Calwer, H., Freienstein, B., Karg, F., 2003. CIS module pilot processing applying concurrent rapid selenization and sulfurization of large area thin film precursors. Thin Solid Films 431–432, 514–522. Pandey, R.K., Sahu, S.N., Chandra, S., 1996. Handbook of Semiconductor Electrodeposition. Marcel Dekker, New York. Panicker, M.P.R., Knaster, M., Kro¨ger, F.A., 1978. J. Electrochem. Soc. 125, 566. Pendorf, J., Winkler, M., Tober, O., Reser, D., Jacobs, K., 1998. CuInS2 thin film formation on a Cu tape substrate for photovoltaic applications. Solar Energy Mater. Solar Cells 53, 285–298. Pern, F.J., Goral, J., Matson, R.J., Gessert, T.A., Noufi, R., 1988. Device quality thin films of CuInSe2 by a one-step electrodeposition process. Solar Cells 24, 81–90. Pottier, D., Maurin, G., 1989. Preparation of polycrystalline thin films of CuInSe2 by electrodeposition. J. Appl. Electrochem. 19, 361–367.

Prosini, P.P., Addonizio, M.L., Antonaia, A., Loreti, S., 1996. Electrodeposition of copper–indium alloy under diffusionlimiting current control. Thin Solid Films 288, 90–94. Qiu, C.X., Shih, I., 1988. Thin film cells on electrodeposited CuInSe2. In: Solomon, I., et al. (Eds.), Proceedings of the Eighth European Photovoltaic Solar Energy Conference, pp. 1051–1055. Qiu, S.N., Li, L., Qiu, C.X., Shih, I., 1994. Heterojunction cells on electrodeposited CuInSe2 films. In: Hill, R., Palz, W., Helm, P. (Eds.), Proceedings of the 12th European Photovoltaic Solar Energy Conference. Stephens Bedford, pp. 1539–1541. Qiu, S.N., Li, L., Qiu, C.X., Shih, I., Champness, C.H., 1995. Study of CuInSe2 thin films prepared by electrodeposition. Solar Energy Mater. Solar Cells 37, 389–393. Rau, U., Schmidt, M., Jasenek, A., Hanna, G., Schock, H.W., 2001. Electrical characterization of Cu(In,Ga)Se2 thin-film solar cells and the role of defects for the device performance. Solar Energy Mater. Solar cells 67, 137. Rau, U., Schock, H.W., 2001. Cu(In,Ga)Se2 solar cells. In: Archer, D., Hill, R. (Eds.), Clean Energy from Photovoltaics. Imperial College Press, London, pp. 277–345, References therein. Raviendra, D., Sharma, J.K., 1985. n-CdS/p-AgInSe2 solar cells by electrodeposition. Phys. Stat. Solidi A 88, 365–368. Rechid, J., Thyen, R., Raitzig, A., Wulff, S., Mihhailova, M., Kalberlah, K., Kampmann, A., 2003. Electrodeposition of CIGS on metal substrates. In: Kurokawa, K., et al. (Eds.), Proceedings of the Third World Conference on Photovoltaic Solar Energy Conversion, pp. 559–561. Siemer, K., Klaer, J., Luck, I., Bruns, J., Klenk, R., Braunig, D., 2001. Efficient CuInS2 solar cells from a rapid thermal process (RTP). Solar Energy Mater. Solar Cells 67, 159– 166. Sratieva, N., Tzvetkova, E., Ganchev, M., Kochev, K., Tomov, I., 1997. Structural and optical properties of electrodeposited CuInSe2 layers. Solar Energy Mater. Solar Cells 45, 87– 96. Takeuchi, K., Takei, K., Nakamura, K., Ichikawa, S., Onuma, Y., 2003. Oral communication at the Third World Conference on Photovoltaic Solar Energy Conversion. Thouin, L., Massaccesi, S., Sanchez, S., Vedel, J., 1994. Formation of copper indium diselenide by electrodeposition. J. Electroanal. Chem. 374, 81–88. Thouin, L., Rouquette-Sanchez, S., Vedel, J., 1993. Electrodeposition of copper–selenium binaries in citric acid medium. Electrochim. Acta 38, 2387–2394. Thouin, L., Vedel, J., 1995. Electrodeposition and characterisation of CuInSe2 thin films. J. Electrochem. Soc. 142, 2996– 3000. Tzvetkova, E., Stratieva, N., Ganchev, M., Tomov, I., Ivanova, K., Kochev, K., 1997. Preparation and structure of annealed CuInSe2 electrodeposited films. Thin Solid Films 311, 101–106. Ugarte, R., Schrebler, R., Cordova, R., Dalchiele, E.A., Gomez, H., 2000. Electrodeposition of CuInSe2 thin films in a glycine acid medium. Thin Solid Films 340, 117–124. Vedel, J., 1998. An electrochemical route for the preparation of chalcopyrite semiconductors. Inst. Phys. Conf. Ser., Section B 152, 261–268.

D. Lincot et al. / Solar Energy 77 (2004) 725–737 Vedel, J., Thouin, L., Lincot, D., 1996. Relation Between Physical Properties and Deposition Conditions of Electrodeposited CuInSe2. J. Electrochem. Soc. 143, 2173– 2180. Voss, T., Muffler, H.J., Schulze, J., Meyer, H., Lux Steiner, M.C., Niesen, T.P., 2002. Electrodeposition of CuInSe2. In: Proceedings of Photovoltaics in Europe, V2.33.

737

Yukawa, K., Kubawara, K., Koumoto, K., 1996. Electrodeposition of CuInS2 from aqueous solution (II) electrodeposition of CuInS2 film. Thin Solid Films 286, 151–153. Zhang, L., Jiang, F.D., Feng, J.Y., 2003. Formation of CuInSe2 and Cu(In,Ga)Se2 films by electrodeposition and vacuum annealing treatment. Solar Energy Mater. Solar Cells 80, 483–490.