8.6% Efficient CZTSSe Solar Cells Sprayed from Water–Ethanol CZTS

Oct 12, 2014 - Evolution of Morphology and Composition during Annealing and Selenization in Solution-Processed Cu2ZnSn(S,Se)4. James A. Clark ...
0 downloads 0 Views 3MB Size
Letter pubs.acs.org/JPCL

8.6% Efficient CZTSSe Solar Cells Sprayed from Water−Ethanol CZTS Colloidal Solutions Gerardo Larramona,* Stéphane Bourdais, Alain Jacob, Christophe Choné, Takuma Muto, Yan Cuccaro, Bruno Delatouche, Camille Moisan, Daniel Péré, and Gilles Dennler IMRA Europe S.A.S., 220 rue Albert Caquot, BP 213, 06904 Sophia Antipolis Cedex, France S Supporting Information *

ABSTRACT: Copper zinc tin sulfide-selenide, Cu2ZnSn(S1−xSex)4 (CZTSSe), thin film photovoltaic devices were fabricated using a fast and environmentally friendly preparation method, consisting of the following steps: An instantaneous synthesis of a Cu−Zn−Sn−S (no Se) colloid, a nonpyrolytic spray of a dispersion of this colloid in a water−ethanol mixture, and a sequential annealing first in a N2 atmosphere and second in a Se atmosphere. The achievement of cell efficiencies up to 8.6% under AM1.5G (cell area 0.25 cm2) and without antireflecting coating indicates that this method can compete with other vacuumbased or more complex wet deposition methods.

SECTION: Energy Conversion and Storage; Energy and Charge Transport uring the last five years (2009−2014), the number of scientific articles dealing with copper zinc tin sulfideselenide Cu2ZnSn(S1−xSex)4 (CZTSSe)1 has been multiplied by a factor of 10, confirming that this material is generally acknowledged as a potential alternative to Cu(In,Ga)Se2 (CIGS). Recently, a hero value of 12.6% has been reported as the result of a thorough light management optimization.2 However, and alike previous record values, these performances were obtained on devices fabricated by a wet deposition method based upon a very toxic and flammable solvent, namely hydrazine.3 A large part of the scientific efforts within the CZTSSe community is currently focused on the development of more environment friendly and industry compatible processes.4−10 In a recent previous paper,11 we reported a swift and simple fabrication method of Cu2ZnSnS4 (CZTS) films that allowed us to achieve promising photovoltaic performances of 5.0% with pure-sulfur CZTS films (i.e., not containing Selenium and having a bandgap of 1.5 eV). This novel process is based on an ultrafast (actually instantaneous) synthesis of a Cu−Zn−Sn−S aqueous colloid of around 10 nm nanoparticles. This colloid is then formulated into an additive-free water−ethanol (90−10% vol) ink, which is sprayed on 300 °C heated 50 × 50 mm2 Mo coated soda-lime glass substrates (typically precut in 25 × 25 mm2 pieces) within a N2 glovebox. The simplicity and straightforwardness of this procedure makes it very attractive for large scale industrial applications and presents many competitive advantages compared to other reported wet routes, namely (i) a very fast synthesis of the colloidal solution as compared to the traditional syntheses of well-crystallized

D

© XXXX American Chemical Society

nanoparticles,4,5,9 (ii) the absence of toxic solvents and/or organic ligands during the synthesis, and (iii) the absence of organic additives (i.e., like ligands, surfactants, thickeners) in the intermediate solution or in the final ink used for the deposition of the photoactive layer.4,9 As a direct consequence of the last two points, we naturally avoid the formation of a carbon-rich fine grained layer between the active layer and the Mo substrate9 and favor a low level of impurities known to be detrimental for minority carriers12 in photovoltaic devices. In spite of the fact that the 5.0% efficiency we could initially achieve with pure-sulfur CZTS films was respectable, it was clearly lagging behind the typical 8−10% efficiency level reported with Se-containing CZTSSe active layers processed with much less environmental friendly and up-scalable approaches.4−9 In this communication, we report the introduction of Se in our unique process; this allowed us to reduce the bandgap of our active layers to the optimum value around 1.15 eV,13 and thus led us to cell performances reaching 8.6% when combined with a CdS buffer layer. The experimental details regarding the CZTS (without Se) colloid synthesis and the ink formulation are similar to what we have reported previously.11 In the new study presented herein, we investigated several approaches to fabricate Se-containing CZTSSe films which were compact and with large grains. The first approach we tried consisted in synthesizing Se-containing colloids, Cu−Zn−Sn−S−Se, in a similar way as the one Received: September 3, 2014 Accepted: October 12, 2014

3763

dx.doi.org/10.1021/jz501864a | J. Phys. Chem. Lett. 2014, 5, 3763−3767

The Journal of Physical Chemistry Letters

Letter

Figure 1. SEM images of top (1) and cross-section (2) views of CZTS or CZTSSe films on commercial Mo glass substrates: As-deposited CZTS film (a), CZTS film after N2 annealing (b), and CZTSSe film after Se annealing (c).

think that the secondary phases formed by the interaction between the Cu−Zn−Sn−S layer and the Mo substrate play an important role, also in the case of the annealing in Se atmosphere. The third and last approach allowed us to significantly improve both the film morphology and the cell performances: It consisted in depositing a pure-sulfur CZTS thin film, and submitting this latter to a sequential annealing comprised of (i) a first annealing under nitrogen followed by (ii) a second annealing under Se atmosphere. The first annealing was performed in a N2 glovebox and on a simple hot plate closed with a lid, with the following temperature profile: A 10 min ramp from room temperature (RT) to 200 °C, a 10 min dwell at 200 °C, a 30 min ramp from 200 to 525 °C, a 15 min dwell at 525 °C, and finally a natural cooling down to RT. After this first annealing step, the samples were placed in a graphite box covered with a lid and containing 80 mg of Se and 5 mg of SnS. This box was then transferred to a three-zone Carbolite HZS 12/600 split tubular oven with a flow of Argon at atmospheric pressure. The thermal cycle of this second annealing was designed to follow approximately a 20 min ramp from RT to 550 °C, a 30 min dwell at 550 °C, and a forced cooling (by opening the tubular oven) below 50 °C in approximately 1 h. All these conditions were chosen after some optimization of the cycle in our selenization oven. The Se mass in the graphite box was tuned to reach our current optimum of Se content in the film. The SnS mass was less than 5% of the Se mass, as a small amount is sufficient to reach its very low saturated vapor pressure: The presence of SnS in the gas form is believed to diminish the possible tin-related loss, which can take place during the selenization annealing process. The structure of the photovoltaic devices was similar to the standards found in the literature, namely soda-lime glass/Mo/ CZTSSe/CdS (60 nm)/ZnO (60 nm)/indium tin oxide (ITO, 250 nm). Manual mechanical scribing was used to define cell areas of 5 × 5 mm2 (cell area ∼0.25 cm2). The exact area for the best cells was measured accurately from digitalized images. For the standard cell fabrication, a simple silver paste dot was

employed for pure-sulfur Cu−Zn−Sn−S (see Supporting Information II). We used aqueous solutions of NaHS and NaHSe as S and Se precursors respectively (the latter being synthesized in house according to a literature procedure, as detailed in Supporting Information I), and an acetonitrile solution of Cu, Zn and Sn chlorides as the metal precursors. Although we could obtain thin films which were compact and made of large grains (upon certain annealing conditions), the best performances did not reach 5% (with short-circuit currents