Mechanistic aspects of preheating effects of electrodeposited metallicprecursors on structural and photovoltaic propertiesof Cu2ZnSnS4 thin films

Yixin Lin a,b, Shigeru Ikeda a,n, Wilman Septina a, Yoshihito Kawasaki a, Takashi Harada a,Michio Matsumura a

a Research Center for Solar Energy Chemistry, Osaka University, 1-3 Machikaneyama, Toyonaka Campus, Osaka 560-8531, Japanb School of Energy and Power Engineering, Huazhong University of Science & Technology, Wuhan 430074, PR China

a r t i c l e i n f o

Article history:Received 19 September 2012Received in revised form22 August 2013Accepted 6 September 2013

Keywords:Cu2ZnSnS4ElectrodepositionEffects of preheatingSolar cell performance

a b s t r a c t

Effects of 350 1C preheating of electrochemically stacked Cu–Sn–Zn layers in an inert atmosphere onstructural and morphological properties of quaternary kesterite Cu2ZnSnS4 (CZTS) films obtained bysulfurization of the metallic layers were investigated. When the CZTS film was fabricated without the350 1C preheating, the film contained a secondary Cu2SnS3 phase and other impurities. The CZTS filmalso showed poor adhesion to the bottom molybdenum (Mo) substrate with many crevices and voids inthe film; a very thick MoS2 layer was formed at the top part of the Mo layer. These structural failures ofthe CZTS film were significantly reduced when the metallic precursor was preheated at 350 1C beforesulfurization for 20 min: the CZTS film thus-obtained exhibited no secondary phases, good adhesion tothe bottom Mo layer, flat and densely packed surface morphology, no appreciable crevices or voids, andformation of a relatively thin MoS2 layer. Based on results of studies on structural transformations during350 1C preheating and successive sulfurization, such differences in properties of resulting CZTS films areattributed to differences in formation pathways. Moreover, since 350 1C preheating for a longer durationwas found to be detrimental for the properties of resulting CZTS films, there should be an optimumduration of preheating to achieve a good CZTS film for solar cell application. As expected from theseproperties, the solar cell with an Al:ZnO/CdS/CZTS/Mo structure derived from the metallic precursorpreheated at 350 1C for 20 min showed the best conversion efficiency of 5.6%.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Copper zinc tin sulfide (Cu2ZnSnS4, CZTS) is one of the mostpromising materials as an alternative low-cost absorber for thin-film photovoltaic devices. The material has a high optical absorp-tion coefficient of 104 cm�1 and a direct band gap of 1.5 eV [1–3]:the theoretical limit of power conversion efficiency is as high as32% based on Shockley-Queisser photon balance calculations [4].Moreover, compared to other commercial CdTe and CIGS absor-bers, it is advantageous that all of the constituent elements in CZTSare naturally abundant and non-toxic, thus facilitating large-scaleproduction [5,6].

A literature survey indicated that CZTS-based solar cells havebeen primarily fabricated by vacuum technologies, e.g., CZTSdevices with high conversion efficiencies of 8.4% and 6.7% havebeen prepared by evaporation [7] and sputtering [8], respectively.

However, these vacuum processes suffer from relatively lowthroughput, low material utilization, and difficulties associatedwith large-scale production [9]. To overcome these drawbacks,many researchers have adopted solution-based non-vacuumdeposition approaches, such as spray pyrolysis [10], printing[11,12], spin coating [13,14], and electrodeposition [15–17],because of their suitability for use in large-scale substrates, highthroughput, and efficient material usage. Recently, a selenosulfidederivative, Cu2ZnSn(S,Se)4 (CZTSSe), fabricated by spin coating of ahydrazine-based precursor solution containing required elements,has reached more than 11% efficiency [14], though the CZTSSematerial contains toxic selenium and the use of explosive hydra-zine is necessary for its synthesis.

Among the above-mentioned non-vacuum methods, electro-deposition of a CZTS film has been shown to be desirable becauseof its advantages, including low equipment cost, negligible wasteof chemicals with utilization efficiencies close to 100%, andpossible formation of a compact film required for solar cellapplication. In the pioneering work by Ennaoui et al., a CZTS-based solar cell prepared from an elecrodepositied Cu–Sn–Zn alloy

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/solmat

Solar Energy Materials & Solar Cells

0927-0248/$ - see front matter & 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.solmat.2013.09.006

n Corresponding author. Tel.: þ81 6 6850 6696; fax: þ81 6 6850 6699.E-mail address: sikeda@chem.es.osaka-u.ac.jp (S. Ikeda).

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i

Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎

followed by sulfurization at 550 1C for 2 h showed appreciablephotovoltaic properties with a conversion efficiency of 3.4% [15].Scragg et al. also reported 3.2% conversion efficiency using a CZTS-based solar cell prepared by electrodeposition of a Cu/Sn/Cu/Znmetallic stack followed by sulfurization at 575 1C for 2 h [16].One of the main drawbacks of these methods reported in theliterature is the requirement of long sulfurization duration, whichis unsuitable for an industrial process in terms of both time andenergy usage.

Recently, Ahmed et al. have shown a significant effect ofpreheating of an electrodeposited metallic stack before sulfuriza-tion on photovoltaic properties: a CZTS-based solar cell exhibitingmore than 7% conversion efficiency was obtained by sulfurizationof the precursor in a short period (12 min) [17]. Although theprimary factor of the preheating effect has not yet been revealed ina preliminary study, these results motivated us to find optimumconditions of such preheating for achieving better photovoltaicperformance as well as to examine mechanistic aspects of thepreheating effects. In our work, therefore, the structural andmorphological variations of an electrodeposited Cu–Sn–Zn stackinduced by preheating and the growth mechanisms of CZTS filmsfrom these metallic precursors with and without preheating wereexamined in detail. Based on the results, their effects on photo-voltaic properties of the thus-obtained CZTS-based cell are alsodiscussed.

2. Experimental

Metallic layers of copper (Cu), tin (Sn), and zinc (Zn) weredeposited successively on a molybdenum-coated soda lime glass(Mo/glass) substrate by using a conventional three-electrode cellwith a Pt foil as a counter electrode and an Ag/AgCl reference

electrode under potentiostatic control using a Hokuto Denko HSV-110 potentiostat. A Cu layer was firstly deposited on Mo/glass at�0.4 V (vs. Ag/AgCl) in an aqueous solution containing 0.05 MCuSO4, 0.02 M citric acid and 0.04 M tri-sodium citrate. An Sn layerwas deposited on the Cu-covered Mo/glass substrate at �0.55 V(vs. Ag/AgCl) in an aqueous solution with 0.05 M Sn(SO3CH3)2 and1 M CH3SO3H. Finally, a Zn layer was deposited at �1.20 V (vs. Ag/AgCl) in a pH 3 buffer solution containing ZnSO4 (0.1 M) and K2SO4

(0.5 M). During depositions of these metallic layers, depositionbaths were kept at 25 1C by a thermostatted water bath. The thus-obtained metallic stack, labeled CTZ(0), was placed in an evac-uated Pyrex ampule (ca. 160 cm3). The ampule was then put in a350 1C-heated electric furnace; after keeping it for 20 min or40 min, the ampule was ejected from the furnace for cooling.Samples after the 350 1C preheating for 20 min and 40 min werelabeled CTZ(20) and CTZ(40), respectively. In order to facilitatesulfurization to form CZTS films, these metallic precursors with orwithout the 350 1CC preheating (i.e., CTZ(0), CTZ(20), and CTZ(40))were again put in the evacuated Pyrex ampule with 5 mg ele-mental sulfur. The ampule was then put in a 580 1C-heated electricfurnace; after keeping it for 10 min, the ampule was ejected fromthe furnace for cooling. The CZT(0) sample was also sulfurized at585 1C for 12 min for comparison.

Crystal structures of the films were analyzed by X-ray diffrac-tion (XRD) using a Rigaku MiniFlex X-ray diffractometer (Cu Kα, Nifilter). Crystalline phase compositions were determined by Ramanspectroscopy using a JASCO NRS-3100 Laser Raman Spectrophot-ometer with a laser wavelength of 532 nm. Morphology of thefilms was examined using a Hitachi S-5000 FEG field emissionscanning electron microscope (FE-SEM) at a voltage of 20 kV.Atomic compositions were determined by energy-diffuse X-ray(EDX) analysis using a Hitachi S-2250N EDX analyzer.

As-prepared CZTS films were processed to complete with an Al:ZnO/CdS/CZTS/Mo/glass structure [18,19]. The CdS buffer layer wasdeposited by chemical bath deposition (CBD). Then a transparentconductive oxide (TCO) layer of Al-doped ZnO (Al:ZnO) wasdeposited on the CdS surface by radio-frequency (RF)-magnetronsputtering. Current density–voltage (J–V) characteristics undersimulated AM1.5 irradiation (100 mW cm�2) were measured witha Bunkoh-Keiki CEP-015 photovoltaic measurement system.

3. Results and discussion

Fig. 1 shows XRD patterns of the electrodeposited metallicprecursor film (CTZ(0)) and the films heated at 350 1C in anevacuated ampule for 20 min (CTZ(20)) and 40 min (CTZ(40)).As expected, the XRD pattern of CTZ(0) can be indexed to be amixture of Cu (JCPDS 04-0836), Zn (JCPDS 65-3358), and Sn (JCPDS65-2631) (Fig. 1a). On the other hand, the CTZ(20) sample showedXRD patterns of Cu–Sn and Cu–Zn alloys with chemical composi-tions of Cu6Sn5 (JCPDS 65-2303), Cu5Zn8 (JCPDS 41-1435) andCu3Sn (JCPDS 65-4653), and reflections of pure metals disap-peared, as shown in Fig. 1b. The fact that the color of the filmchanged from white-grey of the Zn top layer of the precursor togrey-gold is attributable to formation of such alloys. Due to thelimited mutual solubility between Sn and Zn (i.e., less than 0.5 wt%) [20], there was no observation of any Sn–Zn alloys. Indeed, noexperimental evidence for any stable solid phases between Sn andZn has been reported [21]. In addition, prolonged heat treatmentfor 40 min induced an increase in intensities of reflections ofCu6Sn5 and Cu5Zn8, indicating crystalline growth of these alloys(Fig. 1c).

Corresponding surface SEM images of metallic precursors areshown in Fig. 2. The SEM image of the CTZ(0) film shows island-shaped morphology: there are many crevices between their grains

Fig. 1. X-ray diffraction patterns of (a) CTZ(0), (b) CTZ(20), and (c) CTZ(40). *notassigned.

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎2

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i

(Fig. 2a). Based on the fact that the Cu bottom layer formedcompact morphology (data), the observed island-shaped surface isderived from inhomogeneous electrodeposition of the second Snlayer. Although shapes of grains of CTZ(20) and CTZ(40) filmsbecame somewhat angular, island-shaped morphologies stillremained (Fig. 2b and c). These results suggest the requirementto find a condition for homogeneous deposition of an Sn layer inorder to obtain a homogeneous/flat metallic precursor layer.As can be expected from the above XRD results, moreover, angulargrains in the CTZ(40) films were larger than those in the CTZ(20)films. It is also notable that there are appreciable voids formed inthe CTZ(40) film: such a structural failure may affect properties ofthe final CZTS film (see below).

Sulfurization of these metallic precursors in a sulfur atmo-sphere at 580 1C for 10 min led to alterations of diffractionpatterns, as shown in Fig. 3. With or without preheating treat-ment, the XRD patterns of thus-obtained films matched thekesterite structure of CZTS (JCPDS 26-0575): reflections corre-sponding to (101), (110), (112), (103), (200), (202), (211), (105),(220), (312), (224) and (008) planes were detected without anysignificant reflections of other compounds except for those derivedfrom the bottom Mo layer (i.e., Mo and its sulfurized form). Asshown in Fig. 4, corresponding Raman spectra showed an intensepeak centered at 338 cm�1 along with two weak peaks located at289 cm�1 and 369 cm�1; these Raman signals are in good agree-ment with those derived from CZTS [8,22,23]. Other than theseCZTS peaks, no appreciable peak was observed on sulfurized filmsprepared from the CTZ(20) and CTZ(40) metallic precursors,indicating high phase purity of these CZTS films (Fig. 4b and c).On the other hand, the presence of an additional peak centered at303 cm�1, which is assignable to the ternary phase of Cu2SnS3[24,25], indicates insufficient phase purity of the CZTS filmobtained from CTZ(0), as shown in Fig. 3a. Since the Cu2SnS3compound shows almost the same diffraction patterns as those ofCZTS, the presence of this secondary phase was not identifiedby the above XRD analysis. It should be noted that the Cu2SnS3phase disappeared when the CTZ(0) metallic precursor wassulfurized at a higher temperature, e.g., 585 1C for 12 min,

indicating incomplete reaction of the CTZ(0) precursor when itwas sulfurized at 580 1C for 10 min (see below).

As determined by EDX analyses, chemical compositions ofthese sulfurized films were almost the same, with or withoutthe 350 1C preheating. When compared to the stoichiometriccomposition (i.e., Cu/Zn/Sn/S¼0.25/0.125/0.125/0.5), all of thefilms were Cu-poor (Cu content¼ca. 0.23), leading to the forma-tion of Cu vacancies that work as shallow acceptors [26,27]. With

Fig. 2. Top-view SEM images of (a) CTZ(0), (b) CTZ(20), and (c) CTZ(40).

Fig. 3. X-ray diffraction patterns of films sulfurized at 580 1C for 10 min that wereobtained from (a) CTZ(0), (b) CTZ(20), and (c) CTZ(40).

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i

regard to Zn and Sn compostions, the CZTS films were Zn-richrelative to Sn (Zn/Sn¼ca. 1.2) when compared to the stoichio-metric ratio (Zn/Sn¼1). Since such Cu-poor and Zn-rich composi-tions have been shown empirically to have superior photovoltaicperformance [26,28], this metallic composition of the films is anapproximate composition, though the precise optimum conditionhas not yet been clarified. In addition, sulfur compositions of all ofthe films are close to the stoichiometric composition (S con-tent¼ca. 0.50), implying samller amounts of detrimental donorlevels formed by S vacancies. It should be noted that there appearto be other compounds included in the CZTS film derived from CTZ(0) because of the presence of the Cu2SnS3 secondary phasedespite having almost the same chemical composition as that ofthe single-phase CZTS films obtained from CTZ(20) and CTZ(40).Based on the fact that no reflection or peak due to other suchimpurity phases was observed in the XRD pattern and Ramanspectrum of the CZTS film, these contaminants are likely to bepresent as less-crystalline/amorphous forms in the film.

Fig. 5 shows surface and cross-sectional SEM images of CZTSfilms obtained from CTZ(0), CTZ(20), and CTZ(40). The surface SEMimage of the CZTS film obtained by sulfurization of the CTZ(0) metallic precursor at 580 1C for 10 min indicates that the filmconsists of angular-shaped solids with particle sizes ranging fromseveral hundreds of micrometers to a few micrometers, as shownin Fig. 5a. Micrometer-sized voids are also frequently observedbetween these solids. As expected from the surface morphology,

Fig. 4. Raman spectra of films sulfurized at 580 1C for 10 min that were obtainedfrom (a) CTZ(0), (b) CTZ(20), and (c) CTZ(40).

Fig. 5. Top-view and cross-sectional SEM images of sulfurized CZTS films obtained from (a, d) CTZ(0), (b, e) CTZ(20), and (c, f) CTZ(40).

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎4

the corresponding cross-sectional view shows the formation ofmany crevices in the CZTS film (Fig. 5d). Based on the fact that apart of the CZTS film thus-obtained was usually peeled off,formation of these voids and crevices resulted in insufficientadhesion of the CZTS film to the Mo/glass substrate. Even thoughpure CZTS phase was obtained by sulfurization of the CTZ(0) pre-cursor at 585 1C for 12 min, the thus-obtained CZTS film hadsimilar structural and morphological failures. On the other hand,there was no appreciably large void on the surface of the CZTS filmobtained from the CTZ(20) metallic precursor: round-shapedparticles were densely deposited to form a relatively smoothsurface morphology, as shown in Fig. 5c. From its cross-sectionalview, the CZTS film showed good adhesion to the Mo/glasssubstrate without any crevices (Fig. 5d). On the other hand, furtherextension of preheating to 40 min (i.e., use of the CTZ(40)precursor) resulted in the formation of a CZTS film with less-packed morphology. It should be noted that angular- and round-shaped particles were mixed in the film. As shown in Fig. 5f,morphological inhomogeneity was also observed in its crosssection, i.e., there were two different types of particles: round-shape particles with relatively large sizes and angular-shapedparticles with sizes of around a few hundred nanometers. Inaddition to the inhomogeneity, there were several voids, especiallyat the boundary between the CZTS film and the Mo layer. Hence, itwould be an optimum duration of preheating to achieve ahomogeneous and densely-packed CZTS film.

As shown in Fig. 5d–f, cross-sectional SEM images clearlyindicate darkening of the upper parts of the Mo layer due to theformation of molybdenum disulfide (MoS2) during sulfurization.A thick MoS2 layer (ca. 500 nm in thickness) was observed in the

sample prepared from CTZ(0) (Fig. 5d). The thickness of the layerin the sample obtained from CTZ(40) was moderate (Fig. 5f), andthe MoS2 layer with the smallest thickness was formed in thesample derived from CTZ(20) (Fig. 5e). Since the thickness of theMoS2 layer tended to increase with increase in amounts of voidsand crevices formed in the CZTS layer, the reaction of the Mo layerand sulfur vapor should be enhanced by peeling-off of the uppermetallic and/or CZTS layer during sulfurization. Formation of anMoS2 interfacial layer at the CZTS/Mo interface is known to becritical in facilitating quasi-ohmic electrical contact [9] but notwould be detrimental when it forms with an appreciable thick-ness. Thus, the CZTS film derived from CTZ(20), having the least-thick MoS2 layer, is favorable for photovoltaic application.

For further evaluation of the effects of preheating on propertiesof thus-obtained CZTS films, morphologies and phase composi-tions of CTZ(0), CTZ(20) and CTZ(40) films after sulfurization forshort periods (1 and 2 min) were compared. Typical cross-sectional SEM images and Raman spectra of these films are shownin Figs. 6 and 7, respectively. The 1-min sulfurized film obtainedfrom the CTZ(0) metallic precursor shows formation of a layeredcompound along with appreciable crevices, as can be seen inFig. 6a. The corresponding Raman spectrum presented in Fig. 7ashows an intense peak centered at 473 cm�1, assignable to coppersulfide (CuxS) [29,30], together with the 338 cm�1 peak derivedfrom CZTS. Since the CuxS compound usually forms a plate-likeshape, the observed layered morphology should reflect the for-mation of the compound. Although the 2-min sulfurized filmobtained from CTZ(0) shows an increase in the peak intensity ofCZTS, i.e., facilitation of CZTS formation (Fig. 7d), the peak of CuxSstill appeared, suggesting difficulty in complete diffusion of

Fig. 6. Cross-sectional SEM images of (a, d) CTZ(0), (b, e) CTZ(20), and (c, f) CTZ(40) films after sulfurization for (a, b, c) 1 min and (d, e, f) 2 min.

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i

metallic components due to their inhomogeneous distributions.This should result in the formation of a ternary by-product(Cu2SnS3) after prolonged sulfurization (see Fig. 4a). Anothernotable point for the 2-min sulfurized film is that large creviceswere still present in the film (Fig. 6d). This results in poor adhesionof the CZTS film to the Mo/glass substrate, as also shown in Fig. 5d.With regard to the 1-min sulfurized film obtained from the CTZ(20) metallic precursor, micrometer-sized lumps were formedwithout appearance of a layered structure, as shown in Fig. 6c.Although the corresponding Raman spectrum showed an appreci-able peak of the CuxS compound (Fig. 7c), its peak intensity wassignificantly small compared to that for the 1-min sulfurized filmobtained from CTZ(0). Moreover, complete disappearance of the

CuxS compound to form a single-phase CZTS film was achievedafter 2-min sulfurization, as shown in Fig. 7e. The cross sectionindicates much better adhesion to the bottom Mo layer than thatof the 2-min sulfurized film derived from CTZ(0) (see above),while the film was still composed of small particles of ca. 100 nmin size (Fig. 6e). The 1-min and 2-min sulfurized films obtainedfrom the CTZ(40) precursor showed similar morphologicalchanges when compared to those observed on the sulfurized filmsfrom the CTZ(20) precursor, as shown in Fig. 6c and f. Correspond-ing Raman spectra also indicated similar results (Fig. 7c and f).These results indicate the effectiveness of preheating to facilitatecrystallization of the CZTS film without formation of secondaryphases and failure of adhesion to the Mo bottom layer.

Since the metallic stack used in this study was deposited in theorder of Cu, Sn and Zn, the Cu layer was on the bottom of the as-deposited metallic stack. As mentioned above, sulfurization of themetallic stack without preheating induced significant formation ofthe CuxS compound. Based on these facts, preferable reactionbetween Cu and sulfur vapor was likely to be induced by out-diffusion of the bottom Cu component to the surface; this wouldlead to inhomogeneous distribution of the Cu component to formthe Cu2SnS3 secondary phase and formation of many voids andcrevices in the resulting CZTS film. When the 350 1C preheatedprecursor was employed, however, formation of the CuxS com-pound was significantly suppressed. The difference is attributed tothe formation of Cu6Sn5, Cu3Sn and Cu5Zn8 alloys, which wouldsuppress the preferential reaction between metallic Cu and sulfurvapor. Instead, relatively reactive or unstable sulfurized com-pounds were probably formed from Cu6Sn5, Cu3Sn and Cu5Zn8

alloys; conversion of them into the CZTS phase occurred promptly.Significant differences in structures of CZTS films obtained fromthe metallic precursors without and with preheating treatmentshown in Fig. 5 are, therefore, attributed to the differences in thegrowth mechanism of the CZTS phase between them. In addition,there is no experimental evidence to explain the morphologicalinhomogeneity observed on the CZTS film derived from themetallic precursor with preheating for longer (40 min) duration(see Fig. 5c and f). As mentioned above, overgrowth of Cu6Sn5 andCu5Zn8 alloys during the 350 1C preheating accompanied by theformation of appreciable voids is likely to induce reduction of therate of grain growth of the CZTS phase, resulting in appreciablesmall grains remaining in the final CZTS film.

Solar cells with a device structure of Al:ZnO/CdS/CZTS/Mo/glasswere prepared by depositing CdS and Al:ZnO layers on CZTS filmsprepared from CTZ(0), CTZ(20), and CTZ(40). J–V characteristics ofthese cells under simulated AM1.5 irradiation are shown in Fig. 7.It clear that the J–V curve of the cell based on the CZTS filmderived from CTZ(0) shows relatively poor electrical rectification:it gave conversion efficiency (η) of only 0.43%, as shown in Fig. 7a.

Fig. 7. Raman spectra of (a, d) CTZ(0), (b, e) CTZ(20), and (c, f) CTZ(40) films aftersulfurization for (a, b, c) 1 min and (d, e, f) 2 min.

Fig. 8. Light J–V data of CZTS-based solar cells derived from (a) CTZ(0), (b) CTZ(20),and (c) CTZ(40).

Fig. 9. An EQE spectrum of the CZTS-based solar cells derived from CTZ(20).

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎6

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i

As discussed above, appreciable crevices and voids in the CZTS filmshould induce significant cell leakage. The presence of the Cu2SnS3secondary phase (as well as other unidentified impurities) in theCZTS film would also act as a shunt pass. Owing to these structuralfailures, the CZTS film was not useful for photovoltaic application.On the other hand, solar cells based on CZTS films derived fromCTZ(20) and CTZ(40) exhibited appreciable performance (Fig. 7band c). The CZTS-based cell derived from CTZ(20) exhibited betterperformance, with η of 5.6%, short circuit current density (JSC) of19.8 mA cm�2, open circuit potential (VOC) of 0.64 V and fill factor(FF) of 0.45, than that of the cells obtained from the CTZ(40)metallic precursor (η¼2.6%, JSC¼17.5 mA cm�2, VOC¼0.50 V, andFF¼0.30). As can be seen in Fig. 7c, the J–V curve of the CZTS-based cell derived from CTZ(40) shows relatively poor electricalrectification: it had a low shunt resistance (Rsh¼0.04 kΩ cm�2)and a high series resistance (Rs¼15Ω cm�2). The less-packednature of the CZTS film (Fig. 4c) led to the generation of anappreciable shunt pass: the low Rsh significantly reduced VOC.Moreover, the presence of appreciable amounts of voids at theinterface of CZTS and Mo/glass and the relatively thick MoS2 layer(Fig. 4f) should cause such inferior qualities of ohmic contact: theresulting high Rs led to a drop in the FF value as well as a decreasein JSC. On the other hand, the CZTS solar cell obtained from the CTZ(20) metallic precursor shows good rectification, as can be seen inFig. 7b. Significant improvement of VOC compared to that of theCZTS-based cell derived from the CTZ(40) film is due to the highdegree of compactness of the CZTS film, which leads to suppres-sion of shunting, i.e., it has relatively high Rsh (0.15 kΩ cm�2).Large values of FF and JSC are attributed to the good adherencebetween the CZTS film as well as relatively thin MoS2 layer,resulting in efficient majority carrier (hole) transport, thus leadingto relatively low Rs (7.6Ω cm�2).

Figs. 8 and 9 show the external quantum efficiency (EQE)spectrum of the best cell (i.e., the cell derived from CTZ(20)) inwavelengths ranging from 300 nm to 1000 nm. The EQE spectrumincreases gradually from the onset at ca. 350 nm to ca. 450 nm,steeply rises to ca. 520 nm, and then declines gradually to theonset wavelength of ca. 900 nm, corresponding to the band gapenergy of CZTS (1.5 eV) [1,3]. The reduction of EQE response in theshort wavelength region is due to absorption losses in the window(Al:ZnO) and buffer (CdS) layers. On the other hand, the gradualweakening of the EQE response in the long wavelength regionindicates loss of deeply absorbed photons due to poor minoritycarrier diffusion length and/or insufficient penetration of deple-tion width into the absorber [16,31]. Such a failure would becaused by high majority carrier concentration of the CZTS film, i.e.,defect density in the CZTS film is still too high to achieve anoptimum. Thus, further investigation should be carried out toimprove the device performance by efficient production of a well-defined CZTS film with a small amount of defects and majoritycarrier density.

4. Conclusion

In this study, we showed the importance of preheating of theelectrodeposited metallic stack composed of Cu, Sn, and Zn layersto obtain a CZTS film with good compactness, homogeneity andadhesion to the bottom Mo layer. A solar cell with a Al:ZnO/CdS/CZTS/Mo structure derived from the CZTS film was shown to haveefficient device performance with conversion efficiency of morethan 5%. Although the efficiency achieved is significant among theCZTS-based cells reported in the literature, there is still room forfurther improvement in device performance. A limiting factor ofthe best cell is insufficient utilization of photons of relatively longwavelength, as a result of a highly doped state of the CZTS film. We

believe that this failure can be improved by further optimization ofpreheating conditions as well as precise control of metalliccompositions. Since various parameters can be changed in thepreheating process and since electrodeposition can simply controlamounts of deposited metals by changing deposition durationsand/or potentials, further investigations along these lines are nowin progress.

Acknowledgment

This work was carried out as part of a program supported byNew Energy and Industrial Technology Development Organization(NEDO). LY is grateful for financial support by the National NaturalScience Foundation of China (No. 21203069).

References

[1] H. Katagiri, K. Saitoh, T. Washio, H. Shinohara, T. Kurumadani, S. Miyajima,Development of thin film solar cell based on Cu2ZnSnS4 thin films, SolarEnergy Materials and Solar Cells 65 (2001) 141–148.

[2] K. Ito, T. Nakazawa, Electrical and optical properties of stannite-type quatern-ary semiconductor thin films, Japanese Journal of Applied Physics 1 (27)(1988) 2094–2097.

[3] S. Chen, A. Walsh, J.-H. Yang, X. Gong, L. Sun, P.-X. Yang, J.-H. Chu, S.-H. Wei,Compositional dependence of structural and electronic properties of Cu2ZnSn(S,Se)4 alloys for thin film solar cells, Physical Review B 83 (2011) 125201.

[4] W. Shockley, H.J. Queisser, Detailed balance limit of efficiency of p–n junctionsolar cells, Journal of Applied Physics 32 (1961) 510–519.

[5] L.M. Peter, Towards sustainable photovoltaics: the search for new materials,Philosophical Transactions of the Royal Society A: Mathematical Physical andEngineering Sciences 369 (2011) 1840–1856.

[6] C. Wadia, A.P. Alivisatos, D.M. Kammen, Materials availability expands theopportunity for large-scale photovoltaics deployment, Environmental Scienceand Technology 43 (2009) 2072–2077.

[7] B. Shin, O. Gunawan, Y. Zhu, N.A. Bojarczuk, S.J. Chey, S. Guha, Thin film solarcell with 8.4% power conversion efficiency using an earth-abundantCu2ZnSnS4 absorber, Progress in Photovoltaics: Research and Applications 21(2013) 72–76.

[8] H. Katagiri, K. Jimbo, S. Yamada, T. Kamimura, W.S. Maw, T. Fukano, T. Ito,T. Motohiro, Enhanced conversion efficiencies of Cu2ZnSnS4-based thin filmsolar cells by using preferential etching technique, Applied Physics Express 1(2008) 041201.

[9] S.E. Habas, H.A.S. Platt, M.F.A.M. van Hest, D.S. Ginley, Low-cost inorganic solarcells: from ink to printed device, Chemical Reviews 110 (2010) 6571–6594.

[10] Y.B.K. Kumar, G.S. Babu, P.U. Bhaskar, V.S. Raja, Preparation and characteriza-tion of spray-deposited Cu2ZnSnS4 thin films, Solar Energy Materials and SolarCell 93 (2009) 1230–1237.

[11] Y. Miyamoto, K. Tanaka, M. Oonuki, N. Moritake, H. Uchiki, Optical propertiesof Cu2ZnSnS4 thin films prepared by sol–gel and sulfurization method,Japanese Journal of Applied Physics 47 (2008) 596–597.

[12] Z. Zhou, Y. Wang, D. Xu, Y. Zhang, Fabrication of Cu2ZnSnS4 screen printedlayers for solar cells, Solar Energy Materials and Solar Cell 94 (2010)2042–2045.

[13] T.K. Todorov, K.B. Reuter, D.B. Mitzi, High-efficiency solar cell with earth-abundant liquid-processed absorber, Advanced Materials 22 (2010)E156–E159.

[14] T.K. Todorov, J. Tang, S. Bag, O. Gunawan, T. Gokmen, Y. Zhu, D. B. Mitzi, Beyond11% efficiency: characteristics of state-of-the-art Cu2ZnSn(S,Se)4 solar cells,Advanced Energy Materials 3 (2013) 34–38.

[15] A. Ennaoui, M. Lux-Steiner, A. Weber, D. Abou-Ras, I. Kotschau, H.W. Schock,R. Schurr, A. Holzing, S. Jost, R. Hock, T. Voss, J. Schulze, A. Kirbs, Cu2ZnSnS4thin film solar cells from electroplated precursors: novel low-cost perspective,Thin Solid Films, 517, 2511–2514.

[16] J.J. Scragg, D.M. Berg, P.J. Dale, A 3.2% efficient Kesterite device from electro-deposited stacked elemental layers, Journal of Electroanalytical Chemistry 646(2010) 52–59.

[17] S. Ahmed, K.B. Reuter, O. Gunawan, L. Guo, L.T. Romankiw, H. Deligianni, Ahigh efficiency electrodeposited Cu2ZnSnS4 solar cell, Advanced EnergyMaterials 2 (2012) 253–259.

[18] S.M. Lee, S. Ikeda, T. Yagi, T. Harada, A. Ennaoui, M. Matsumura, Effects ofpreheat treatment on their structural, electrochemical and solar cell proper-ties, Physical Chemistry Chemical Physics 13 (2011) 6662–6669.

[19] S.M. Lee, S. Ikeda, Y. Otsuka, T. Harada, M. Matsumura, Electrochim. Acta,Homogeneous electrochemical deposition of In on a Cu-covered Mo substratefor fabrication of efficient solar cells with a CuInS2 photoabsorber, Electro-chimica Acta 79 (2012) 189–196.

[20] E. Guaus, J. Torrent-Burgués, Tin-zinc electrodeposition from sulphate–tartratebaths, Journal of Electroanalytical Chemistry 575 (2005) 301–309.

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i

[21] J. von Appen, R. Dronskowski, K. Hack, A theoretical search for intermetalliccompounds and solution phases in the binary system Sn/Zn, Journal of Alloysand Compounds 379 (2004) 110–116.

[22] D.B. Mitzi, O. Gunawan, T.K. Todorov, K. Wang, S. Guha, The path towards ahigh-performance solution-processed kesterite solar cell, Solar Energy Mate-rials and Solar Cells 95 (2011) 1421–1436.

[23] P.A. Fernandes, P.M.P. Salomé, A.F. da Cunha, Growth and Raman scatteringcharacterization of Cu2ZnSnS4 thin films, Thin Solid Films 517 (2009)2519–2523.

[24] P.A. Fernandes, P.M.P. Salomé, A.F. da Cunha, CuxSnSxþ1 (x¼2, 3) thin filmsgrown by sulfurization of metallic precursors deposited by dc magnetronsputtering, Physica Status Solidi c 7 (2010) 901–904.

[25] P.A. Fernandes, P.M.P. Salomé, A.F. da Cunha, A study of ternary Cu2SnS3 andCu3SnS4 thin films prepared by sulfurizing stacked metal precursors, Journalof Physics D: Applied Physics 43 (2010) 215403.

[26] S. Chen, X. Gong, A. Walsh, S.H. Wei, Defect physics of the kesterite thin-filmsolar cell absorber Cu2ZnSnS4, Applied Physics Letters 96 (2010) 021902.

[27] A. Walsh, S.Y. Chen, S.H. Wei, X.G. Gong, Kesterite thin-film solar cells:advances in materials modelling of Cu2ZnSnS4, Advanced Energy Materials 2(2012) 400–409.

[28] H. Katagiri, K. Jimbo, W.S. Maw, K. Oishi, M. Yamazaki, H. Araki, A. Takeuchi,Development of CZTS-based thin film solar cells, Thin Solid Films 517 (2009)2455–2460.

[29] H. Yoo, J. Kim, L. Zhang, Sulfurization temperature effects on the growth ofCu2ZnSnS4 thin film, Current Applied Physics 12 (2012) 1052–1057.

[30] H. Yoo, J. Kim, Comparative study of Cu2ZnSnS4 film growth, Solar EnergyMaterials and Solar Cells 95 (2011) 239–244.

[31] I. Repins, C. Beall, N. Vora, C. DeHart, D. Kuciauskas, P. Dippo, B. To, J. Mann,W.C. Hsu, A. Goodrich, R. Noufi, Co-evaporated Cu2ZnSnSe4 films and devices,Solar Energy Materials and Solar Cells 101 (2012) 154–159.

Y. Lin et al. / Solar Energy Materials & Solar Cells ∎ (∎∎∎∎) ∎∎∎–∎∎∎8

Please cite this article as: Y. Lin, et al., Mechanistic aspects of preheating effects of electrodeposited metallic precursors on structuraland photovoltaic properties of..., Solar Energy Materials and Solar Cells (2013), http://dx.doi.org/10.1016/j.solmat.2013.09.006i