Elsevier

Acta Materialia

Volume 132, 15 June 2017, Pages 276-284
Acta Materialia

Full length article
Formation of nanometer-sized Cu-Sn-Se particles in Cu2ZnSnSe4 thin-films and their effect on solar cell efficiency

https://doi.org/10.1016/j.actamat.2017.04.056Get rights and content

Abstract

Atom probe tomography and transmission electron microscopy are used to study the formation of nano-sized Cu-Sn-Se particles in Cu2ZnSnSe4 thin-films. For a Cu-rich precursor, which was deposited at 320 °C under Cu- and Zn-rich growth conditions, Cu2-xSe grains at the surface are detected. During annealing the precursor at 500 °C in a SnSe + Se atmosphere most of the Cu2-xSe is transformed to Cu2ZnSnSe4 via the consumption of excessive ZnSe and incorporation of Sn. However, atom probe tomography studies also reveal the formation of various nanometer-sized Cu-Sn-Se particles close to the CdS/Cu2ZnSnSe4 interface. One of those particles has a composition close to the Cu2SnSe3 compound. This phase has a smaller band gap than Cu2ZnSnSe4 and is proposed to lead to a significant drop in the open-circuit voltage and could be the main cause for a detrimental p-n junction and the zero efficiency of the final device. Possible effects of the other phases on solar cell performance and formation mechanisms are discussed as well.

Introduction

The compound semiconductors Cu2ZnSn(S,Se)4 (CZTS(e)) have recently attracted a great deal of attention as absorber materials for thin-film solar cells [1], [2], [3], [4]. As a result of intensive research world-wide, CZTS(e) based solar cells have already achieved 12.6% efficiency [2].

The growth of CZTS(e) films is often hampered by the formation of secondary phases due to the narrow existence regions of CZTS(e) compounds as confirmed by theoretical [5], [6] and experimental studies [7], [8]. Secondary phases that were observed are predominantly ZnS(e) [9], [10], [11], [12], CuxS(e) [13], [14], Cu2SnS(e)3 [15], [16], and SnS(e)x [9], [14]. In general, they are considered to be detrimental to the cell's performance. Cu2SnS(e)3, Cu2S(e) and SnS(e) have smaller band gaps than CZTS(e) and should be avoided by any means as their presence will decrease the open circuit voltage [3]. In agreement with this presumption, CZTS(e) based solar cells with the highest efficiencies are found to be Cu-poor ([Cu]/([Zn]+[Sn])<1) and Zn-rich ([Zn]/[Sn]>1) [3], [17]. Under such growth conditions ZnS(e) can be easily formed. However, ZnS(e) has a larger band gap than CZTS(e) and its detrimental effect on cell efficiency is moderate [3].

Although a Cu-poor and Zn-rich CZTS(e) composition is preferred, an intermediate Cu-rich ([Cu]/([Zn]+[Sn])>1) growth step can be included. For Cu(In,Ga)Se2 it was shown that a Cu-rich ([Cu]/([Ga]+[In])>1) composition for a limited time during the growth process is beneficial for the cell performance due to an increase in grain size and a reduction in recombination activity [18], [19]. In a previous work we could demonstrate that the Cu-rich growth step for CZTSe can be easily implemented by using a Cu-rich precursor [20]. The Cu excess leads to the formation of Cu2-xSe at the surface of the Cu-rich precursor, which can be removed by KCN etching. Subsequently, the KCN-treated Cu-rich precursor was annealed in a SnSe + Se atmosphere and led to working cell devices. However, when the Cu2-xSe phase was not removed by KCN etching, the annealed precursor yielded zero efficiency. It was speculated that this observation is due to the formation of a detrimental Cu-Sn-Se compound during annealing.

Despite these interesting observations, a detailed formation mechanism of the Cu-Sn-Se compound could not be given. In general, there is a lack of knowledge about the formation of Cu-Sn-S(e) compounds during CZTS(e) growth. Direct evidence of the formation of Cu2SnS3 was given by Cheng et al. [15]. Using electron dispersive X-ray spectroscopy (EDX) and confocal Raman imaging formation of Cu2SnS3 on the surface upon sulfurization of a Zn/Cu/Sn metal stack was detected. This observation was attributed to an incomplete formation of CZTS. Other groups [9], [21] used a similar growth procedure and found indications of Cu-Sn-S formation at the CZTS/Mo interface by EDX and X-ray photoelectron spectroscopy. However, it has not yet been clarified whether Cu-Sn-S formation is due to an incomplete sulfurization of the metal stacks or due to decomposition of the CZTS compound at the interface with the Mo back contact [22], [23].

Mousel et al. [20] and Wang et al. [24] prepared precursors by a low temperature (320 °C and 110 °C, respectively) co-evaporation process and applied a post-deposition annealing process in a SnSe + Se atmosphere and in a S atmosphere, respectively. Using secondary ion mass spectrometry (SIMS), Mousel et al. [20] detected a Sn-enrichment at the surface of the annealed precursor, which suggested the formation of a Cu-Sn-Se compound, whereas Wang et al. [24] detected Cu-Sn-S compounds at the CZTS/Mo interface by EDX.

However, in most studies direct detection and/or quantification of possibly nano-sized Cu-Sn-Se compounds was too challenging or even impossible due to insufficient spatial resolution of the applied techniques and/or due to the similar structure of e.g. Cu2SnS(e)3 with CZTS(e) and ZnS(e). To overcome these challenges, we perform complementary atom probe tomography (APT) and scanning transmission electron microscopy (STEM)-EDX measurements in this work. We provided insights into the formation of Cu-Sn-Se particles and the influence of these particles on the overall cell performance.

Section snippets

Thin film preparation and device fabrication

The CZTSe films were grown by a sequential process. Two precursors were fabricated in a molecular beam epitaxy system by co-evaporation of Cu, Zn, Sn, and Se onto a Mo-coated soda-lime glass (SLG) substrate at a temperature of 320 °C. Both precursors were grown under Zn-rich ([Zn]/[Sn]>1) conditions. One precursor was Cu-rich ([Cu]/([Zn]+[Sn])>1) and the other one was Cu-poor ([Cu]/([Zn]+[Sn])<1). The absorber films were fabricated by annealing the precursor at 500 °C for 30 min in a SnSe + Se

Precursor films

In a previous work [20] the present authors detected a Cu-rich surface on the Cu-rich precursor by SIMS depth profiling and found indications for the formation of a Cu-Se phase. However, the distribution of this phase on the sub-micrometer scale was not studied in detail.

STEM-EDX maps shown in Fig. 1a)–d) reveal wedge-shaped regions at the film surface as Cu2-xSe grains as it is also evident from the EDX intensity profile shown in Fig. 1e). The compositions are: 57.6 ± 0.4 at.% Cu and

Formation of secondary phases

In this part we focus on the development of a growth model, which we suggest for the formation of secondary phases in the Cu-rich precursor and the annealed Cu-rich precursor without prior KCN etching. The other samples only contain ZnSe as secondary phase and its formation mechanism is similar to the ZnSe formation in the Cu-rich precursor.

Besides CZTSe, several secondary phases are formed in the Cu-rich precursor film due to the Cu-rich and Zn-rich growth. After an initial growth of small

Conclusions

We conducted APT and STEM-EDX measurements to study the formation of Cu-Sn-Se compounds and their effect on cell performance. We detected Cu2-xSe grains mainly at the surface of the co-evaporated Cu-rich precursor due to Cu-rich growth conditions. Some of them are doped with Zn and Sn. The Sn incorporation can be as high that Cu-Sn-Se compounds are already formed in the precursor. The incorporation of Sn from the SnSe + Se atmosphere during annealing without prior KCN etching of the Cu-rich

Acknowledgements

The authors thank Susanne Siebentritt from the University of Luxembourg as well as Baptiste Gault from Max-Planck-Institut für Eisenforschung GmbH for fruitful discussions. Furthermore, the authors are grateful to Uwe Tezins and Andreas Sturm for their support to the APT and FIB facilities at Max-Planck-Institut für Eisenforschung GmbH. This work was funded by the German Research Foundation (DFG) (Contract CH 943/2-1 and GA 2450/1-1), and by the Luxembourgish Fonds National de la Recherche.

References (39)

  • S. Fiechter et al.

    Phase relations and optical properties of semiconducting ternary sulfides in the system Cu–Sn–S

    J. Phys. Chem. Solids

    (2003)
  • S. Siebentritt et al.

    Kesterites - a challenging material for solar cells

    Prog. Photovoltaics

    (2012)
  • W. Wang et al.

    Device characteristics of CZTSSe thin-film solar cells with 12.6% efficiency

    Adv. Energy Mater.

    (2014)
  • G. Brammertz et al.

    Characterization of defects in 9.7% efficient Cu2ZnSnSe4-CdS-ZnO solar cells

    Appl. Phys. Lett.

    (2013)
  • A. Walsh et al.

    Kesterite Thin-film solar cells: advances in materials modelling of Cu2ZnSnS4

    Adv. Energy Mater.

    (2012)
  • A. Redinger et al.

    Detection of a ZnSe secondary phase in coevaporated Cu2ZnSnSe4 thin films

    Appl. Phys. Lett.

    (2011)
  • T. Schwarz et al.

    Atom probe study of Cu2ZnSnSe4 thin-films prepared by co-evaporation and post-deposition annealing

    Appl. Phys. Lett.

    (2013)
  • J.T. Wätjen et al.

    Direct evidence of current blocking by ZnSe in Cu2ZnSnSe4 solar cells

    Appl. Phys. Lett.

    (2012)
  • G.S. Babu et al.

    Effect of Cu/(Zn plus Sn) ratio on the properties of co-evaporated Cu2ZnSnSe4 thin films

    Sol. Energ Mat. Sol. C

    (2010)
  • View full text