Controllable Multinary Alloy Electrodeposition for Thin-Film Solar Cell Fabrication: A Case Study of Kesterite Cu2ZnSnS4

摘要

class="head no_bottom_margin" id="sec1title">IntroductionElectrodeposition (ED) is a non-vacuum and commercial technology with the demonstrated capability of depositing functional coatings on the base materials by using electric current to reduce dissolved salts in electrolyte baths. The feasibility of ED has been evaluated for the majority of the elements in the periodic table, including conventional metal elements (e.g., Cu, Zn, Sn, Au, Pt), alkaline earth metal elements (e.g., Mg [], Sr [], Eu []), transition group metal elements (e.g., Nb [], Ta [], Bi []), and non-metal elements (e.g., Se [], Te []). Based on these elements, a variety of multifunctional coatings can be electrodeposited for applications of finishing, microelectronics, nanobiosystems, solar cell productions, etc. (; , ). In particular, ED has met with success in preparing a wide range of photoelectrochemical and photovoltaic solar absorber films, including Sb2Se3 (), CdTe (), Cu2O (), CuSbS2 (, href="#bib67" rid="bib67" class=" bibr popnode">Septina et al., 2014), Cu(In, Ga)Se2 (CIGS) (href="#bib20" rid="bib20" class=" bibr popnode">Duchatelet et al., 2013, href="#bib52" rid="bib52" class=" bibr popnode">Lincot et al., 2004), Cu2ZnSnS(e)4 [CZTS(e)] (href="#bib15" rid="bib15" class=" bibr popnode">Colombara et al., 2015, href="#bib58" rid="bib58" class=" bibr popnode">Peter, 2015), and CH3NH3PbI3 (href="#bib13" rid="bib13" class=" bibr popnode">Chen et al., 2015, href="#bib41" rid="bib41" class=" bibr popnode">Huang et al., 2015). Aside from the absorbers, functional ZnO window layers (href="#bib73" rid="bib73" class=" bibr popnode">Tsin et al., 2015, href="#bib74" rid="bib74" class=" bibr popnode">Tsin et al., 2016a), Zn-based finger grids (href="#bib75" rid="bib75" class=" bibr popnode">Tsin et al., 2016b), and CuSCN (href="#bib80" rid="bib80" class=" bibr popnode">Ye et al., 2015) charge transport layers in the thin-film solar cells can be as well prepared by ED. Besides, ED has already demonstrated great success in the roll-to-roll CIGS solar panel manufacture, which is based on the precursor type of an electrodeposited Cu-In-Ga alloy covered by an electrodeposited In-Se or Ga-Se single layer (href="#bib2" rid="bib2" class=" bibr popnode">Aksu et al., 2012, href="#bib7" rid="bib7" class=" bibr popnode">Başol et al., 2009).Kesterite CZTS(e) solar cells have attracted increasing attention in the past decade because of their non-toxicity and earth-abundant nature. ED emerging as a technoeconomic and large-scale deposition technology has been widely employed to fabricate the kesterite absorber precursors. Post-chalcogenization of the ED precursors is additionally needed to form the final kesterites. Promisingly, high-performance (7–9%) CZTS(e) solar cell devices have been frequently reported based on this two-stage approach (href="#bib38" rid="bib38" class=" bibr popnode">Guo et al., 2014, href="#bib43" rid="bib43" class=" bibr popnode">Jeon et al., 2014, href="#bib77" rid="bib77" class=" bibr popnode">Vauche et al., 2016). Specifically, electrodeposited precursors can be classified into two types: stack layers (i.e., Cu/Sn/Zn) and alloy layers (i.e., Cu-Zn-Sn), both of which have successfully attained the highly efficient kesterite devices. The device performances closely hinge on the quality of the kesterite absorber, which usually requires a Cu-poor and Zn-rich composition (Cu/Sn = 1.6–1.7, Cu/Zn ≤ 1.7, Zn/Sn ≥ 1.1), absence of secondary phases, and a large-area layer uniformity (href="#bib22" rid="bib22" class=" bibr popnode">Fairbrother et al., 2015, href="#bib71" rid="bib71" class=" bibr popnode">Tai et al., 2016, href="#bib77" rid="bib77" class=" bibr popnode">Vauche et al., 2016). Thus, it calls for ED with the ability to tune the precursor film composition and control its morphology sophisticatedly.For the stacking ED precursor, the composition can be manipulated by varying the thickness of each individual layer. But the nucleation and uniformity of the overlayers largely depend on those of the underlying layers, and as a result, the stacking ED precursors usually fail to exhibit a smooth, uniform, and compact morphology (href="#bib66" rid="bib66" class=" bibr popnode">Scragg et al., 2010, href="#bib77" rid="bib77" class=" bibr popnode">Vauche et al., 2016). Meanwhile, the morphological inhomogeneity always comes with the compositional fluctuation. The direct chalcogenization annealing of stacked precursors usually results in poor film morphologies even with the formation of secondary phases, ultimately leading to poor device efficiencies (href="#bib50" rid="bib50" class=" bibr popnode">Lin et al., 2014, href="#bib66" rid="bib66" class=" bibr popnode">Scragg et al., 2010). The pre-mixing of the stacking ED precursors at moderately lower temperatures is known as one expedient to improve the compactness, morphology roughness, and atomic homogeneity, leading to an 8.2% efficiency using a kesterite selenide CZTSe absorber (href="#bib38" rid="bib38" class=" bibr popnode">Guo et al., 2014, href="#bib77" rid="bib77" class=" bibr popnode">Vauche et al., 2016). In contrast, the alloy ED precursor offers a tantalizing advantage over the stacking ED, making the pre-annealing of the precursors no longer needed before the chalcogenization, and an 8% efficiency has been achieved based on an alloy ED kesterite selenide CZTSe absorber (href="#bib43" rid="bib43" class=" bibr popnode">Jeon et al., 2014). Nonetheless, alloy ED seems more challenging in the practical aspects on how to control the alloy composition, largely because of the big difference between the reduction potentials of metal ions. Although the use of complexing reagents, such as low-cost citrate and tartrate salts (href="#bib1" rid="bib1" class=" bibr popnode">Abd El Rehim and El Ayashy, 1978, href="#bib35" rid="bib35" class=" bibr popnode">Gougaud et al., 2013, href="#bib37" rid="bib37" class=" bibr popnode">Guaus and Torrent-Burgués, 2005, href="#bib45" rid="bib45" class=" bibr popnode">Kazimierczak and Ozga, 2013, href="#bib49" rid="bib49" class=" bibr popnode">Lee et al., 2013, href="#bib54" rid="bib54" class=" bibr popnode">Mkawi et al., 2014, href="#bib69" rid="bib69" class=" bibr popnode">Shin et al., 2016, href="#bib70" rid="bib70" class=" bibr popnode">Slupska and Ozga, 2014), may minimize this potential difference, the reduction of Zn ion is rather negative and competes with the hydrogen evolution. To make the deposition of Zn more efficient, one expedient is to increase the deposition potential (or current density) (href="#bib1" rid="bib1" class=" bibr popnode">Abd El Rehim and El Ayashy, 1978, href="#bib12" rid="bib12" class=" bibr popnode">Chen et al., 2000, href="#bib35" rid="bib35" class=" bibr popnode">Gougaud et al., 2013, href="#bib45" rid="bib45" class=" bibr popnode">Kazimierczak and Ozga, 2013, href="#bib49" rid="bib49" class=" bibr popnode">Lee et al., 2013); worse, its side effect leads to hydrogen evolution parallel to the film deposition, particularly for acidic baths (pH < 5), and even leads to precursor films with a rough and dendritic morphology (href="#bib35" rid="bib35" class=" bibr popnode">Gougaud et al., 2013, href="#bib49" rid="bib49" class=" bibr popnode">Lee et al., 2013, href="#bib69" rid="bib69" class=" bibr popnode">Shin et al., 2016, href="#bib70" rid="bib70" class=" bibr popnode">Slupska and Ozga, 2014). As a result, few groups are able to obtain well-performing kesterite solar cells using alloy ED. Hitherto, none of the kesterite publications could succeed in refining electrodeposit compositions without sacrificing the deposit morphology. This may be due, largely, to the increased complexity of the ternary alloy ED that needs to handle with three different component metal elements, compared with the industrialized single or binary ED techniques.The control of the alloy composition along with the deposit morphology needs the comprehensive and insightful knowledge of the co-plating process and the plating parameters, including bath compositions, plating potentials/current density, plating time, use of additives, and bath stability. This work addresses in detail these complicated operation variables of multinary alloy ED using kesterite CZTS as a case study. It is shown that the alloy composition and morphology can be exactly controlled through manipulation of these co-plating variables. Under the optimized conditions of alloy ED, the layer uniformity of precursors can rival those that were fabricated by the vacuum-based methods in the aspects of morphology and composition, manifesting the promise of alloy ED for the low-cost industrial fabrication of solar devices.