Secondary Phase Formation Mechanism in the Mo-Back Contact

Research Article Cite This: ACS Appl. Mater. Interfaces 2019, 11, 23160−23167

www.acsami.org

Secondary Phase Formation Mechanism in the Mo-Back Contact Region during Sulfo-Selenization Using a Metal Precursor: Effect of Wettability between a Liquid Metal and Substrate on Secondary Phase Formation Se-Yun Kim,†,§ Seung-Hyun Kim,†,§ Sanghun Hong,‡ Dae-Ho Son,† Young-Ill Kim,† Sammi Kim,† Kwangseok Ahn,† Kee-Jeong Yang,† Dae-Hwan Kim,*,† and Jin-Kyu Kang*,† Downloaded via BUFFALO STATE on July 19, 2019 at 01:41:05 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

†

Convergence Research Center for Solar Energy, Daegu-Gyeongbuk Institute of Science and Technology (DGIST), Daegu 42988, Republic of Korea ‡ School of Materials Science and Engineering, Kyungpook National University, Daegu 41566, Republic of Korea S Supporting Information *

ABSTRACT: Recently, highly efficient CZTS solar cells using pure metal precursors have been reported, and our group created a cell with 12.6% efficiency, which is equivalent to the long-lasting world record of IBM. In this study, we report a new secondary phase formation mechanism in the back contact interface. Previously, CZTSSe decomposition with Mo has been proposed to explain the secondary phase and void formation in the Mo-back contact region. In our sulfo-selenization system, the formation of voids and secondary phases is well explained by the unique wetting properties of Mo and the liquid metal above the peritectic reaction (η-Cu6Sn5 → ε-Cu3Sn + liquid Sn) temperature. Good wetting between the liquid Sn and the Mo substrate was observed because of strong metallic bonding between the liquid metal and Mo layer. Thus, some ε-Cu3Sn and liquid Sn likely remained on the Mo layer during the sulfo-selenization process, and Cu−SSe and Cu−Sn−SSe phases formed on the Mo side. When bare soda lime glass (SLG) was used as a substrate, nonwetting adhesion was observed because of weak van der Walls interactions between the liquid metal and substrate. The Cu−Sn alloy did not remain on the SLG surface, and Cu−SSe and Cu−Sn−SSe phases were not observed after the final sulfo-selenization process. Additionally, Mo/SLG substrates coated with a thin Al2O3 layer (1−5 nm) were used to control secondary phase formation by changing the wetting properties between Mo and the liquid metal. A 1 nm Al2O3 layer was enough to control secondary phase formation at the CZTSSe/Mo and void/Mo interfaces, and a 2 nm Al2O3 layer was enough to perfectly control secondary phase formation at the Mo interface and Mo−SSe formation. KEYWORDS: CZTSSe, metal precursor, Mo back contact, secondary phase formation mechanism, wettability

■

INTRODUCTION

maintenance. Thus, methods using a compound or metal precursor prepared by a sputtering process have been used to replace difficult hydrazine solution processes.5−10 In particular, excellent results have been reported for the use of metal precursors.11−13 In our group, PCEs of 12.3%5 and, very recently, 12.6%, which tie with the world record efficiency certified by Newport, were achieved by using metal precursors

Among emerging photovoltaic technologies, kesterite materials (Cu2SnZnS4, CZTS; Cu2SnZnSe4, CZTSe; and Cu2ZnSn(S1−xSex)4, CZTSSe) with high adsorption coefficients are considered promising candidates for large-area module production using earth-abundant, nontoxic elements.1−3 Thus far, the CZTSSe cell, with the highest power conversion efficiency (PCE, 12.6%), was obtained by a two-step process using a hydrazine solution.4 In terms of explosion risk and wastewater treatment, hydrazine solution processes using various organic solvents necessitate much attention and © 2019 American Chemical Society

Received: March 4, 2019 Accepted: June 6, 2019 Published: June 6, 2019 23160

DOI: 10.1021/acsami.9b03969 ACS Appl. Mater. Interfaces 2019, 11, 23160−23167

Research Article

ACS Applied Materials & Interfaces and a sulfo-selenization process;2 more details about the latter will be reported in the near future. Interestingly, when using metal precursors, many secondary phases and large voids are commonly observed in the Mo-back contact side.6,9,10,14−17 These secondary phases and voids near the Mo-back contact reduce the fill factor and output current of solar cells.18−21 For example, it was reported that a highly conductive Cu−S secondary phase near the Mo-back contact can decrease RSh22,23 and provide a recombination site. Additionally, ZnS and MoS2 secondary phases have been reported to increase RS,24−26 and the Cu2SnS3 phase can also act as a recombination center because of the formation of a type I heterointerface.27 Thus far, the representative formation mechanism of secondary phases and voids in the Mo-back contact interface has been the decomposition of CZTS by Mo at 550 °C;28,29 it was reported that CZTSe can be decomposed by Mo at 400 °C. 30−33 Mo-assisted decomposition of CZTS(e) was proposed based on a thermodynamic calculation, as shown in eq 1; the facile reduction of Sn(IV) and the favorability of MoS(e)2 formation can induce phase separation by removing S(e) from CZTS on the Mo-back contact side, and volatile Sn(II)S(e) is formed by the reduction of Sn(IV)S(e)2.28,29

controlling the wettability between a substrate and liquid metal. Experimental Details. The metal precursors for the CZTSSe absorber layer were deposited onto the Mo layer using 99.99% pure Sn, Cu, and Zn sputtering targets with a stacking order of Sn (275 nm)/Cu (160 nm)/Zn (188 nm)/ Mo. The sample was heated from room temperature to 300 °C for 560 s and then maintained at 300 °C for 900 s. Subsequently, the sample was heated from 300 to 480 °C for 1800 s and then maintained at 480 °C for 600 s. The experimental details for synthesizing CZTSSe have already been described in detail in previous research.36 To compare the wettability of liquid Sn on different substrates, Mo/SLG and SLG substrates were used. The Sn reflow test was conducted at 400 °C for 10 min under an Ar flow. The Al2O3coated Mo/SLG substrate was used to investigate the effect of an intermediate layer on secondary phase formation at the Moside. Al2O3 (from 1 to 5 nm thickness) was deposited by plasma enhanced atomic layer deposition (PE-ALD, NCD Co., model LUCIDA M200 PL) using trimethylaluminum (TMA) and oxygen as the precursor and reactant materials, respectively. Prior to deposition, the Mo/SLG substrate was cleaned using acetone, methanol, and buffered oxide etchant (BOE) to remove particles and any oxide layer, sequentially. Argon was also used as the carrier gas for the TMA precursor and as the purge gas after each cycle. The TMA precursor was kept at 10 °C throughout the deposition process, and Al2O3 layers were deposited at a substrate temperature of 100 °C. Each cycle of the Al2O3 PE-ALD process consisted of a TMA injection pulse (0.2 s), an Ar purge (10.0 s), an oxygen pulse (2.0 s), a radio frequency (RF) plasma pulse (3.0 s), and an Ar purge (10.0 s). A RF plasma pulse that was capacitively coupled with a low frequency plasma source and was applied for 3.0 s only during the RF plasma pulse in order to produce oxygen radicals. The RF plasma power and electrode− substrate distance were, respectively, fixed at 100 W and 40 mm. Deposition rate per cycle was 1.47 Å/cycle. Surface and cross-sectional images of the absorber layers were obtained by field emission scanning electron microscopy (FESEM, Hitachi Co., model S-4800). Scanning transmission electron microscopy−energy dispersive X-ray spectroscopy (STEM−EDS) measurements were performed using a QUANTAX-200 instrument (Bruker Co.) to analyze the composition of the absorber layers.

2Cu 2ZnSnS(e)4 + Mo → 2Cu 2S(e) + 2ZnS(e) + 2SnS(e) + MoS(e)2

(1)

It has been suggested that the SnS(e) phase can easily evaporate, and Cu migrates into the absorber and MoS(e)2; thus, ZnS(e) with voids mainly remains in the CZTS(e)/Mo interfacial region. Based on these formation mechanisms, different techniques, such as inserting an intermediate layer between CZTSSe and Mo or changing the metal electrodes, have been utilized to prevent decomposition of CZTSSe by Mo.34,35 In our previous study involving processes that yielded PCEs as high as 12.5%, the formation mechanisms of the CZTSSe double layer and voids in the CZTSSe were proposed.36 Based on our previous experimental results, there are two types of void formation mechanisms. First, large voids are produced due to persistent dezincification and the preferential chalcogenide reaction between Zn and S and/or Se at temperatures below approximately 400 °C; almost all Zn is consumed to form a ZnSSe layer, leaving a Cu−Sn alloy with a Cu/Sn ratio of 2:1 under the ZnSSe layer. Second, additional small voids form due to the mass transfer of Cu and Sn from below the ZnSSe layer to above the layer. In other words, the Mo decomposition model does not well explain the void formation mechanisms in our process. Thus, we should consider a different formation mechanism for the secondary phases that form on the Mo interface to understand the cause of the secondary phase formation and to find a way to reduce the formation of secondary phases. In this study, we proposed a new model that may explain the formation mechanism of secondary phases without Mo decomposition. The formation of a secondary phase at the Mo interface can be expected to be determined by the wetting behavior between the substrate and liquid metal. To compare the bonding strengths between substrates and nonreactive liquid metals, a Mo/soda lime glass (SLG) substrate, a bare SLG substrate, and an Al2O3-coated Mo/SLG substrate were used. The results confirmed that the formation of a secondary phase in the Mo-back contact region can be well controlled by

■

RESULTS AND DISCUSSION As mentioned above, we previously reported the detailed CZTSSe phase formation schemes of our 2-step process.36 Voids are formed by the migration or dezincification of Zn to form ZnS(e) and the mass transfer of Cn and Sn through the ZnS(e) grain boundary; in our previous report, the formation of these voids was not explained by the model based on Moassisted CZTSSe decomposition and SnSSe evaporation near the back contact. Interestingly, secondary phases, such as Cu− Sn−SSe and Cu−SSe, were observed not only at the CZTSSSe/Mo interface but also at the void/Mo interface of the 12.5% CZTSSe absorber layer. However, the cause of the secondary phase distribution in the Mo-back contact side, such as the CZTSSe/Mo and void/Mo interfaces, was not addressed in depth in the previous study.36 To examine the detailed phase evolution of the secondary phase during the sulfo-selenization process, the samples were removed after being cooled down to specific temperatures. The 23161


Research Article

ACS Applied Materials & Interfaces

Figure 1. (a,b) Cross-sectional STEM−EDS mapping images at different positions on the sample cooled at 400 °C. (c) Cross-sectional STEM− EDS mapping images of the CZTSSe absorber layer after sulfo-selenization at 480 °C for 10 min. (The details are in the results of a previous report.34) (d−h) Schematic diagram of the secondary formation mechanism; (d) good wetting behavior between a liquid metal and metal substrate, and (e−g) secondary phase formation at approximately 400 °C, and (h) distribution of the secondary phase at the void/Mo interface after the final process.

sample cooled down to 400 °C showed a slightly nonuniform graded surface, which indicates a rapid change at approximately 400 °C and a slightly higher edge temperature. Thus, if we scan from the center of the 400 °C sample to the edge, we should easily observe the region where an abrupt change occurs; the FIB samples for STEM−EDS mapping were prepared from different sites on one sample, and the temperature of the sample in Figure 1a is actually lower than that of the sample in Figure 1b. It was observed that the ZnSSe layer is preferentially formed on the surface, and a Cu−Sn alloy exists under the ZnSSe layer at around 400 °C, as shown in Figure 1a. Interestingly, the secondary phases, such as Cu−SSe and Cu−Sn−SSe, in the CZTSe/Mo and void/Mo interface regions had already formed at around 400 °C, as shown in Figure 1b. X-ray diffraction (XRD) analysis according to the temperature tracking experiment of our sulfo-selenization process was added in Figure S1, however, it was difficult to define each secondary phase, clearly. Based on the results of STEM−EDS and XRD data, it was confirmed that the Cu−Sn alloy and liquid Sn remained in the CZTSSe/Mo and void/Mo interface as residues at around 400 °C, and these residues became a secondary phase, such as Cu−SSe or Cu−Sn−SSe. As shown in Figure 1c, the distribution of the secondary phases observed in the CZTSSe absorber layer after sulfo-selenization at 480 °C for 10 min is similar to that of the secondary phase distribution at 400 °C. That is, the secondary phase observed in the final sample can be expected to form at 400 °C and did not significantly change until the final stage. So far, some representative results have been reported in which the void and secondary phase are successfully controlled by inhibiting CZTS decomposition by Mo. Scragg et al. proposed a method to prevent CZTS decomposition by inserting a TiN diffusion barrier in the process which annealed at 560−570 °C for 10 min using a Cu−Sn−Zn−S precursor.29 Liu et al. also reported that voids and ZnS phases can be controlled by inserting an Al2O3 diffusion barrier in the process, which annealed at 560 °C for 5 min using a Cu/ZnS/

SnS precursor.37 In the two cases reported above, it is expected that the reaction path might be different from our case using a pure metal precursor. Also, because the heat treatment was conducted above a temperature (550 °C) capable of thermodynamic decomposition, it can be expected that the decomposition may be successfully suppressed by inserting the diffusion barrier. On the other hand, Liu et al. reported that the void was reduced when the ZnO diffusion barrier was inserted in the process which annealed at 570 °C for 30 min using the Cu/Sn/Zn stacked metal precursor.38 In this case, it is expected that ZnO has a diffusion barrier role because the heat treatment had proceeded for a comparatively long time above the decomposing temperature. However, in our previous study, it was clearly reported that void formation began at 300 °C, very large voids were formed at 400 °C, and small voids were additionally formed at over 400 °C;36 that is, it was confirmed that the secondary phase was distributed on the rear contact surface even though the voids were not formed by decomposition.36 Thus, we propose a new secondary phase formation mechanism, as shown in Figure 1d,f. When a liquid metal droplet falls on a metal substrate, the wettability between the substrate and metal should be good, as shown in Figure 1d; when a nonreactive liquid metal is on a metal substrate a wetting angle of less than 30° has been reported.39 When a metal precursor was sulfo-selenized at approximately 400 °C, the total ratio of Cu to Sn of the residual Cu−Sn alloys was expected to be approximately 2:1 because all the Zn from the Cu−Zn−Sn metal precursors was consumed to form the ZnSSe layer.36 Then, the Cu−Sn alloys (η-Cu6Sn5 and εCu3Sn) with this ratio pass through the peritectic reaction (ηCu6Sn5 → ε-Cu3Sn + liquid Sn) point; liquid Sn containing approximately 14% Cu is produced.40 Thus, it was expected that the wetting characteristics of liquid Sn and Mo can be important factors. Good wetting behavior was expected due to the metal bonding between the Mo layer and the Cu−Sn alloys (ε-Cu3Sn and liquid Sn). Previous studies have reported that 23162


Research Article


der Waals interactions, resulting in a wetting angle greater than approximately 120°. Additionally, when BOE-treated Mo/SLG substrates were used, the wetting behavior was expected to be better than that with Mo/SLG, as shown in Figure S2, because of the increase in the surface energy caused by damage or natural MoOx etching. Figure 3a shows STEM−EDS mapping images of CZTSSe on the SLG substrate after annealing at 400 °C, and Figure 3b shows the images obtained after the final process. Unlike Figure 1b, most of the metal alloy is adhered to the SLG substrate below the ZnSSe layer, not on the surface. In other words, the metal alloy has better wetting behavior on the ZnSSe layer than the SLG surface. As a result, a relatively clean interface between the SLG surface and the CZTSSe layer can be obtained during the sulfo-selenization process, as shown in Figure 3b; XRD patterns of CZTSSe on the Mo/SLG and SLG substrates were added in Figure S3. Additionally, there is no secondary phase at the void/SLG substrate interface. Based on these results, we can explain the absence of bottom secondary phases, as shown in the schematic diagrams in Figure 3c−g. When a liquid metal droplet falls on an SLG substrate, poor wettability between the SLG substrate and liquid metal is expected, as shown in Figures 2b and 3c. In this case, as shown in Figure 3d−f, most of the Cu and Sn is expected to migrate to the upper side of the ZnSSe layer through the ZnSSe grain boundary without leaving any residue on the Mo substrate due to poor wettability. This tendency is the result of the poor wetting behavior of the solid and liquid forms.41 Based on these results, we tried to control the secondary phase formation in the Mo-back contract region by changing the surface conditions of the Mo layer to cause metallic bonding. Thus, a 5 nm-thick Al2O3 layer was coated on the substrate by ALD. Cross-sectional FESEM images of the CZTSSe product obtained from a sulfo-selenization process using 5 nm-Al2O3-coated Mo/SLG substrates are shown in Figure 4a−c. The MoSSe phase was not observed at the CZTSSe/Mo and void/Mo interfaces, probably due to the blocking effect of Al2O3, and no secondary phases, such as Cu−SSe and Cu−Sn−SSe, were observed. The bottom of the CZTSSe layer and the surface of Al2O3/Mo were exposed by exfoliation, and their surface morphologies are shown in Figure 4d,e, respectively. Because the bottom morphology of CZTSSe is similar to that of the Al2O3-coated Mo/SLG morphology, there is likely no secondary phase between CZTSSe and Al2O3coated Mo. The cross-sectional STEM−EDS mapping images

Cu and Sn below the ZnSSe layer migrate to the top of the ZnSSe layer;36 see previous work for more details. Accordingly, Cu and Sn movement is very important, and residual Cu and Sn are likely depending on the wetting characteristics of the substrate. When Cu and Sn migrate through the ZnSSe layer to the upper side, the Cu−Sn alloys remain on the Mo substrate due to the good wettability with the Mo layer, as shown in Figure 1e, which shows the void/Mo interface. The metal residues are expected to react with chalcogens to form secondary phases, such as Cu−SSe and Cu−Sn−SSe, as shown in Figure 1f. Thus, the wettability behavior between the Cu−Sn alloy and the Mo/SLG substrate is a key factor for controlling secondary phase formation in the Mo-back contact region. Based on this concept, the wettability behavior between the liquid Sn metal and Mo/SLG or SLG substrates was investigated, as shown in Figure 2. Sn was confirmed to be

Figure 2. Surface FESEM images of Sn films on (a) Mo/SLG and (b) SLG substrates annealed at 400 °C for 10 min under Ar.

relatively well dispersed on the Mo/SLG substrate, while Sn was relatively aggregated on the SLG substrate, as shown in Figure 2a,b, respectively. As mentioned above, good wetting behavior between a nonreactive liquid metal and a solid metal substrate can be observed due to their strong metallic interfacial bonds.39 However, nonreactive liquid metals do not exhibit wetting behavior on carbon materials (C) and ionocovalent ceramic (Al2O3, SiO2, and BN) substrates;39 in these systems, nonwetting adhesion is provided by weak van

Figure 3. Cross-sectional STEM−EDS mapping images of the sample (a) annealed at 400 °C and (b) after the final process. (c−g) Schematic diagram of the secondary formation mechanism. (e) Poor wetting behavior between a liquid metal and the SLG substrate; (d−f) secondary phase formation at approximately 400 °C and (g) distribution of the secondary phase at the void/Mo interface after the final process. 23163


Research Article


Figure 4. (a) Cross-sectional FESEM images of the CZTSSe layer on the 5 nm Al2O3 layer deposited on the Mo/SLG substrate. (b,c) Highmagnification FESEM images of the CZTSSe/Al2O3/Mo interface. Surface FESEM images of the (d) CZTS backside and the (e) Al2O3 side after exfoliating the CZTSSe layer. (f) Cross-sectional STEM−EDS mapping images. (g) High-magnification STEM−EDS mapping images of the CZTSSe/Al2O3/Mo interface. (h) High-magnification TEM image of the CZTSSe/Al2O3/Mo interface.

Figure 5. (a) Surface FESEM images of exfoliated CZTSSe using Mo/SLG substrates and (b) high-magnification images of (a). (c) Schematic diagram of the distribution of voids and the bottom CZTSSe of the CZTSSe layer using a Mo/SLG substrate. (d) Surface FESEM images of exfoliated CZTSSe using an Al2O3-coated Mo/SLG substrate and (e) high-magnification images of (d). (f) Schematic diagram of the distributions of void and the bottom CZTSSe of the CZTSSe layer using an Al2O3-coated Mo/SLG substrate.

This difference in wetting behavior affects the distribution of the lower part of the CZTSSe double layer and the formation of secondary phases at the Mo interface. The exfoliated CZTSSe surface was observed to confirm the distribution of the lower part of the CZTSSe double layer, as shown in Figure 5. Figure 5a,b show the back surface FESEM images of exfoliated CZTSSe synthesized using a Mo/SLG substrate. Based on the results of Figure 5a, it can be expected that the voids were isolated when the Mo/SLG substrate was used; similar results were confirmed by the etched surface of CZTSSe in our previous work.36 A schematic diagram of the distribution of the lower part of CZTSSe is shown in Figure 5c.

of CZTSSe/Al2O3-coated Mo are shown in Figure 4f. The high-magnification STEM−EDS image in Figure 4g shows a uniform Al2O3 layer between Mo and CZTSSe. Figure 4h shows the Al2O3 layer with a thickness of approximately 5 nm. Additionally, the Al2O3 layer with a thickness of 5 nm was confirmed to well tolerate a temperature increase to 480 °C, and a heat treatment was performed for 10 min. The results show that the formation of Cu−SSe and Cu−Sn−SSe phases in the Mo-back contact region can be suppressed by controlling the wettability between the liquid metal and Al2O3-coated Mo layer. 23164


Research Article


of secondary phase formation, and the impurity of Al or/and O. When a 5 nm Al2O3-coated Mo/SLG substrate was used, we observed atypical grains as the liquid phase at the bottom of the exfoliated CZTSSe layer, as shown in Figure S7. For Al2O3 layers with a thickness of more than 2 nm, peeling of the CZTSSe layer occurred during the CdS CBD process, as shown in Figure S8. The origin of the secondary phase formation in the Mo-back contact region was suggested as the wetting behavior, and some methods for improving efficiency by suppressing secondary phase formation were suggested, as mentioned below. First, based on our experimental conditions, because the secondary phase present in the Mo-back contact region is expected to migrate to the CZTSSe layer, it is necessary to tune the composition for optimization by reducing the Cu and Sn contents. Second, another intermediate layer with poor wettability with liquid Sn, such as BN or carbon intermediate layers, which can be applied as electrodes to control wetting, can also be used.39 Third, Na should be added to induce sufficient defect passivation. Fourth, fine dot patterning is expected to be able to control the wetting characteristics, so optimizing the dot pattern can be a method that does not interfere with hole transport while controlling wettability.

Figure 5d,e are the back surface FESEM images of exfoliated CZTSSe synthesized using a 5 nm Al2O3-coated Mo/SLG substrate; the same distribution of the lower part of CZTSSe was observed when using the SLG substrate, as shown in Figure S4. Interestingly, the voids were connected, while the lower part of the CZTSSe double layer was isolated. A schematic diagram of the distribution of the lower part of the CZTSSe double layer with a 5 nm Al2O3-coated Mo/SLG substrate is shown in Figure 5f. As a result, when the wetting behavior is good, the liquid Sn does not actively migrate. Therefore, the lower part of the CZTSSe double layer forms where the liquid Sn is. On the other hand, when the wetting behavior is poor, liquid Sn easily migrates and agglomerates; thus, the lower part of the CZTSSe double layer appears to be isolated as a large island. The above results confirmed that the intermediate layer greatly influences the formation of the secondary phases and voids. Therefore, we investigated the differences that developed with different intermediate layer thicknesses. The cross-sectional FESEM images of a sample without the Al2O3 intermediate layer and samples with 1−5 nm Al2O3 layers are shown in Figure 6a−f. Secondary phases can be observed in

■

CONCLUSIONS In this study, we explained that Cu−SSe and Cu−Sn−SSe secondary phases form at the CZTSSe/Mo and void/Mo interfaces because of the wettability behavior between the substrate and metal precursor when using a metal precursor. It was suggested that good wetting properties between a liquid metal and Mo substrate induce residual metal formation on the Mo-back contact side, and the residual metal transforms into the Cu−SSe and Cu−Sn−SSe phases. The effect of wettability changes on secondary phase formation in the Mo-back contact region was confirmed using SLG- and Al2O3-coated Mo/SLG substrates. The formation of secondary phases in the Mo-back contact region can be effectively controlled by changing the wettability between the liquid metal and the back-contact electrode. In future work, we will attempt to optimize the composition of Cu and Sn, add a Na source, use an intermediate layer other than Al2O3, and control the secondary phase through fine patterning of the intermediate layer.

Figure 6. Cross-sectional FESEM images of CZTSSe (a) without an Al2O3 layer and with (b) 1, (c) 2, (d) 3, (e) 4, and (f) 5 nm Al2O3 layers.

the CZTSSe/Mo or void/Mo interfaces in Figure 6a, while relatively clean CZTSSe/Al2O3 or void/Al2O3 interfaces are observed in Figure 6c−f. Interestingly, a secondary phase-free interface between CZTSSe/Al2O3 was observed when the Al2O3 thickness was 2 nm, and MoSSe was not observed under these conditions. A 2 nm layer of Al2O3 deposited by ALD can well withstand these process conditions (480 °C for 10 min); in contrast to this result, it has been reported that when an Al2O3 layer of a few nanometers is deposited by sputtering, the Al2O3 layer decomposes during heat treatment.37 Interestingly, the formation of Cu−SSe and Cu−Sn−SSe secondary phases in the Mo-back contact region can be suppressed by Al2O3 coating with a thickness of 1 nm, as shown in Figures 6b and S5, although the role of the diffusion barrier inhibition of MoSSe formation disappeared. An Al2O3 layer with a thickness of 1 nm likely changes the wettability at the wetting stage (approximately 400 °C), and then, the Al2O3 layer decomposes due to the increase in temperature, 480 °C. Thus, the MoSSe layer can grow. When an Al2O3 layer with a thickness of 1 nm was applied, the PCE decreased from 12.5 to 2.7%, as shown in Figure S6. The lower efficiency is expected to be due to the effect of Na diffusion interruption because of Al 2 O 3 passivation, the compositional change due to the inhibition

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.9b03969.

■

FESEM images of the Sn layer on the different substrates, FESEM images of the bottom side of exfoliated CZTSSe films which were grown on SLG-, Mo/SLG-, and 5 nm Al2O3-coated Mo/SLG substrates, photoimages of the cell devices, I−V curve of the CZTSSe device with a 1 nm Al2O3-coated Mo/SLG substrate, and cross-sectional STEM−EDS images of the CZTSSe layer on the 1 nm Al2O3-coated Mo/SLG substrate (PDF)

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (D.-H.K.). *E-mail: [email protected] (J.-K.K.). 23165


Research Article

ACS Applied Materials & Interfaces ORCID

Efficiency Enabled by Heterojunction Heat Treatment. Nat. Energy 2018, 3, 764−772. (12) Sun, R.; Zhuang, D.; Zhao, M.; Gong, Q.; Scarpulla, M.; Wei, Y.; Ren, G.; Wu, Y. Beyond 11% Efficient Cu2ZnSn(Se,S)4 Thin Film Solar Cells by Cadmium Alloying. Sol. Energy Mater. Sol. Cells 2018, 174, 494−498. (13) Yan, C.; Sun, K.; Huang, J.; Johnston, S.; Liu, F.; Veettil, B. P.; Sun, K.; Pu, A.; Zhou, F.; Stride, J. A.; Green, M. A.; Hao, X. Beyond 11% Efficient Sulfide Kesterite Cu2ZnxCd1−xSnS4 Solar Cell: Effects of Cadmium Alloying. ACS Energy Lett. 2017, 2, 930−936. (14) Fairbrother, A.; Fontané, X.; Izquierdo-Roca, V.; EspíndolaRodríguez, M.; López-Marino, S.; Placidi, M.; Calvo-Barrio, L.; PérezRodríguez, A.; Saucedo, E. On the Formation Mechanisms of Zn-rich Cu2ZnSnS4 Films Prepared by Sulfurization of Metallic Stacks. Sol. Energy Mater. Sol. Cells 2013, 112, 97−105. (15) Jung, H. R.; Shin, S. W.; Suryawanshi, M. P.; Yeo, S. J.; Yun, J. H.; Moon, J. H.; Kim, J. H. Phase evolution pathways of kesterite Cu 2 ZnSnS 4 and Cu 2 ZnSnSe 4 thin films during the annealing of sputtered Cu-Sn-Zn metallic precursors. Sol. Energy 2017, 145, 2−12. (16) Li, W.; Chen, J.; Yan, C.; Liu, F.; Hao, X. Transmission electron microscopy analysis for the process of crystallization of $\text{C}{{\text{u}}_{2}}\text{ZnSn}{{\text{S}}_{4}}$ film from sputtered Zn/CuSn precursor. Nanotechnology 2014, 25, 195701. (17) Yoo, H.; Kim, J. Growth of Cu2ZnSnS4 Thin Films Using Sulfurization of Stacked Metallic Films. Thin Solid Films 2010, 518, 6567−6572. (18) Zhuk, S.; Kushwaha, A.; Wong, T. K. S.; Masudy-Panah, S.; Smirnov, A.; Dalapati, G. K. Critical Review on Sputter-Deposited Cu2ZnSnS4 (CZTS) Based Thin Film Photovoltaic Technology Focusing on Device Architecture and Absorber Quality on the Solar Cells Performance. Sol. Energy Mater. Sol. Cells 2017, 171, 239−252. (19) Li, W.; Chen, J.; Cui, H.; Liu, F.; Hao, X. Inhibiting MoS2 Formation by Introducing a ZnO Intermediate Layer for Cu2ZnSnS4 Solar Cells. Mater. Lett. 2014, 130, 87−90. (20) Hiroi, H.; Sakai, N.; Kato, T.; Sugimoto, H. High Voltage Cu 2ZnSnS 4 Submodules by Hybrid Buffer Layer. Proceedings of the 39th IEEE Photovoltaic Specialists Conference (PVSC ’13); IEEE: Tampa, Fla, USA, 2013; pp 863−866. (21) Chalapathy, R. B. V.; Jung, G. S.; Ahn, B. T. Fabrication of Cu2ZnSnS4 Films by Sulfurization of Cu/ZnSn/Cu Precursor Layers in Sulfur Atmosphere for Solar Cells. Sol. Energy Mater. Sol. Cells 2011, 95, 3216−3221. (22) Marchionna, S.; Garattini, P.; Le Donne, A.; Acciarri, M.; Tombolato, S.; Binetti, S. Cu2ZnSnS4 Solar Cells Grown by Sulphurisation of Sputtered Metal Precursors. Thin Solid Films 2013, 542, 114−118. (23) Abusnina, M.; Matin, M.; Moutinho, H. R.; Blackburn, J. L.; Alleman, J.; DeHart, C.; To, B.; Al-Jassim, M. Suppression of the Cu2−xS Secondary Phases in CZTS Films Through Controlling the Film Elemental Composition. IEEE J. photovolt. 2015, 5, 1470−1475. (24) Li, W.; Chen, J.; Yan, C.; Hao, X. The Effect of ZnS Segregation on Zn-Rich CZTS Thin Film Solar Cells. J. Alloys Compd. 2015, 632, 178−184. (25) Nam, D.; Cho, S.; Sim, J.-H.; Yang, K.-J.; Son, D.-H.; Kim, D.H.; Kang, J.-K.; Kwon, M.-S.; Jeon, C.-W.; Cheong, H. Solar Conversion Efficiency and Distribution of ZnS Secondary Phase in Cu2ZnSnS4 Solar Cells. Sol. Energy Mater. Sol. Cells 2016, 149, 226− 231. (26) Yang, K.-J.; Sim, J.-H.; Jeon, B.; Son, D.-H.; Kim, D.-H.; Sung, S.-J.; Hwang, D.-K.; Song, S.; Khadka, D. B.; Kim, J.; Kang, J.-K. Effects of Na and MoS2 on Cu2ZnSnS4 Thin-film Solar Cell. Prog. Photovoltaics Res. Appl. 2015, 23, 862−873. (27) Bao, W.; Masaya, I. Influence of Secondary Phases in KesteriteCu 2 ZnSnS 4 Absorber Material Based on the First Principles Calculation. J. Photoenergy 2015, 2015, 592079. (28) Scragg, J. J.; Wätjen, J. T.; Edoff, M.; Ericson, T.; Kubart, T.; Platzer-Björkman, C. A Detrimental Reaction at the Molybdenum Back Contact in Cu2ZnSn(S,Se)4 Thin-Film Solar Cells. J. Am. Chem. Soc. 2012, 134, 19330−19333.

Se-Yun Kim: 0000-0003-1465-477X Author Contributions §

S.-Y.K. and S.-H.K. contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE), Republic of Korea (no. 20173010012980), the Technology Development Program to Solve Climate Change of the National Research Foundation (NRF) funded by the Ministry of Science and ICT, Republic of Korea (2016M1A2A2936781), and the DGIST R & D Programs of the Ministry of Science and ICT, Republic of Korea (19-BD-05). We thank Cheon, Eun, and Sung in the Center for Core Research Facilities (CCRF) of DGIST for STEM measurements and Al2O3 coating.

■

REFERENCES

(1) Best Research-Cell Efficiencies chart in the US Department of Energy’s National Renewable Energy Laboratory (accessed June 14, 2019); www.nrel.gov/pv/cell-efficiency.html. (2) Green, M. A.; Hishikawa, Y.; Dunlop, E. D.; Levi, D. H.; HohlEbinger, J.; Yoshita, M.; Ho-Baillie, A. W. Y. Solar Cell Efficiency Tables (Version 53). Prog. Photovoltaics Res. Appl. 2019, 27, 3−12. (3) Mitzi, D. B.; Gunawan, O.; Todorov, T. K.; Wang, K.; Guha, S. The Path Towards a High-Performance Solution-Processed Kesterite Solar Cell. Sol. Energy Mater. Sol. Cells 2011, 95, 1421−1436. (4) Wang, W.; Winkler, M. T.; Gunawan, O.; Gokmen, T.; Todorov, T. K.; Zhu, Y.; Mitzi, D. B. Device Characteristics of CZTSSe ThinFilm Solar Cells with 12.6% Efficiency. Adv. Energy Mater. 2014, 4, 1301465. (5) Yang, K.-J.; Son, D.-H.; Sung, S.-J.; Sim, J.-H.; Kim, Y.-I.; Park, S.-N.; Jeon, D.-H.; Kim, J.; Hwang, D.-K.; Jeon, C.-W.; Nam, D.; Cheong, H.; Kang, J.-K.; Kim, D.-H. A Band-Gap-Graded CZTSSe Solar Cell with 12.3% Efficiency. J. Mater. Chem. A 2016, 4, 10151− 10158. (6) Brammertz, G.; Buffière, M.; Oueslati, S.; ElAnzeery, H.; Messaoud, K. B.; Sahayaraj, S.; Köble, C.; Meuris, M.; Poortmans, J. Characterization of Defects in 9.7% Efficient Cu2ZnSnSe4-CdS-ZnO Solar Cells. Appl. Phys. Lett. 2013, 103, 163904. (7) Kim, G. Y.; Jeong, A. R.; Kim, J. R.; Jo, W.; Son, D.-H.; Kim, D.H.; Kang, J.-K. Surface Potential on Grain Boundaries and Intragrains of Highly Efficient Cu2ZnSn(S,Se)4 Thin-films Grown by Two-Step Sputtering Process. Sol. Energy Mater. Sol. Cells 2014, 127, 129−135. (8) Neuschitzer, M.; Sanchez, Y.; López-Marino, S.; Xie, H.; Fairbrother, A.; Placidi, M.; Haass, S.; Izquierdo-Roca, V.; PerezRodriguez, A.; Saucedo, E. Optimization of CdS buffer layer for highperformance Cu2ZnSnSe4solar cells and the effects of light soaking: elimination of crossover and red kink. Prog. Photovoltaics 2015, 23, 1660−1667. (9) Gang, M. G.; Shin, S. W.; Hong, C. W.; Gurav, K. V.; Gwak, J.; Yun, J. H.; Lee, J. Y.; Kim, J. H. Sputtering Processed Highly Efficient Cu2ZnSn(S,Se)4 Solar Cells by a Low-Cost, Simple, Environmentally Friendly, and Up-Scalable Strategy. Green Chem. 2016, 18, 700−711. (10) Li, J.; Kim, S.; Nam, D.; Liu, X.; Kim, J.; Cheong, H.; Liu, W.; Li, H.; Sun, Y.; Zhang, Y. Tailoring the Defects and Carrier Density for Beyond 10% Efficient CZTSe Thin Film Solar Cells. Sol. Energy Mater. Sol. Cells 2017, 159, 447−455. (11) Yan, C.; Huang, J.; Sun, K.; Johnston, S.; Zhang, Y.; Sun, H.; Pu, A.; He, M.; Liu, F.; Eder, K.; Yang, L.; Cairney, J. M.; EkinsDaukes, N. J.; Hameiri, Z.; Stride, J. A.; Chen, S.; Green, M. A.; Hao, X. Cu2ZnSnS4 Solar Cells with Over 10% Power Conversion 23166


Research Article

ACS Applied Materials & Interfaces (29) Scragg, J. J.; Kubart, T.; Wätjen, J. T.; Ericson, T.; Linnarsson, M. K.; Platzer-Björkman, C. Effects of Back Contact Instability on Cu2ZnSnS4 Devices and Processes. Chem. Mater. 2013, 25, 3162− 3171. (30) Redinger, A.; Berg, D. M.; Dale, P. J.; Valle, N.; Siebentritt, S. Rout Touard High-Efficiency Single-Phase Cu2ZnSn(S,Se)4 ThinFilm Solar Cells:Model Experiments and Literature Review. IEEE J. photovolt. 2011, 1, 2156−3381. (31) Weber, A.; Mainz, R.; Schock, H. W. On the Sn Loss from Thin Films of the Material System Cu−Zn−Sn−S in High Vacuum. J. Appl. Phys. 2010, 107, 013516. (32) Redinger, A.; Siebentritt, S. Coevaporation of Cu2ZnSnSe4 Thin Films. Appl. Phys. Lett. 2010, 97, 092111. (33) Redinger, A.; Berg, D. M.; Dale, P. J.; Siebentritt, S. The Consequences of Kesterite Equilibria for Efficient Solar Cells. J. Am. Chem. Soc. 2011, 133, 3320−3323. (34) Kaur, K.; Kumar, N.; Kumar, M. Strategic Review of Interface Carrier Recombination in Earth Abundant Cu−Zn−Sn−S−Se Solar Cells: Current Challenges and Future Prospects. J. Mater. Chem. A 2017, 5, 3069−3090. (35) Zhang, Z.; Yao, L.; Zhang, Y.; Ao, J.; Bi, J.; Gao, S.; Gao, Q.; Jeng, M.-J.; Sun, G.; Zhou, Z.; He, Q.; Sun, Y. Modified Back Contact Interface of CZTSSe Thin Film Solar Cells: Elimination of Double Layer Distribution in Absorber Layer. Adv. Sci. 2018, 5, 1700645. (36) Kim, S.-Y.; Son, D.-H.; Kim, Y.-I.; Kim, S.-H.; Kim, S.; Ahn, K.; Sung, S.-J.; Hwang, D.-K.; Yang, K.-J.; Kang, J.-K.; Kim, D.-H. Void and Secondary Phase Formation Mechanisms of CZTSSe Using Sn/ Cu/Zn/Mo Stacked Elemental Precursors. Nano Energy 2019, 59, 399−411. (37) Liu, F.; Huang, J.; Sun, K.; Yan, C.; Shen, Y.; Park, J.; Pu, A.; Zhou, F.; Liu, X.; Stride, J. A.; Green, M. A.; Hao, X. Beyond 8% Ultrathin Kesterite Cu2ZnSnS4 Solar Cells by Interface Reaction Route Controlling and Self-Organized Nanopattern at the Back Contact. NPG Asia Mater. 2017, 9, No. e401. (38) Liu, X.; Cui, H.; Li, W.; Song, N.; Liu, F.; Conibeer, G.; Hao, X. Improving Cu2ZnSnS (CZTS) solar cell performance by an ultrathin ZnO intermediate layer between CZTS absorber and Mo back contact. Phys. Status Solidi RRL 2014, 8, 966−970. (39) Eustathopoulos, N. Wetting by Liquid MetalsApplication in Materials Processing: The Contribution of the Grenoble Group. Metals 2015, 5, 350−370. (40) Fürtauer, S.; Li, D.; Cupid, D.; Flandorfer, H. The Cu−Sn Phase Diagram, Part I: New Experimental Results. Intermetallics 2013, 34, 142−147. (41) Lafuma, A.; Quéré, D. Superhydrophobic States. Nat. Mater. 2003, 2, 457−460.

23167


Secondary Phase Formation Mechanism in the Mo-Back Contact

Recommend Documents