Highly efficient solution-processed CZTSSe solar cells based on a convenient sodium-incorporated post-treatment method

Highly efficient solution-processed CZTSSe solar cells based on a convenient sodium-incorporated post-treatment method

Journal of Energy Chemistry 40 (2020) 196–203 Contents lists available at ScienceDirect Journal of Energy Chemistry journal homepage: www.elsevier.c...

2MB Sizes 14 Downloads 43 Views

Journal of Energy Chemistry 40 (2020) 196–203

Contents lists available at ScienceDirect

Journal of Energy Chemistry journal homepage: www.elsevier.com/locate/jechem

Highly efficient solution-processed CZTSSe solar cells based on a convenient sodium-incorporated post-treatment method Biwen Duan a,c, Linbao Guo a,c, Qing Yu a,c, Jiangjian Shi a,c, Huijue Wu a, Yanhong Luo a,c,d, Dongmei Li a,c,d,∗, Sixin Wu e, Zhi Zheng f, Qingbo Meng a,b,c,d,∗ a

Key Laboratory for Renewable Energy (CAS), Beijing Key Laboratory for New Energy Materials and Devices, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China Center of Materials Science and Optoelectronics Engineering, University of Chinese Academy of Sciences, Beijing 100049, China c School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China d Songshan Lake Materials Laboratory, Dongguan 523808, Guangdong, China e Key Laboratory for Special Functional Materials of MOE, Henan University, Kaifeng 475004, Henan, China f Key Laboratory for Micro-Nano Energy Storage and Conversion Materials of Henan Province, College of Advanced Materials and Energy, Institute of Surface Micro and Nanomaterials, Xuchang University, Xuchang 461000, Henan, China b

a r t i c l e

i n f o

Article history: Received 15 March 2019 Revised 22 March 2019 Accepted 26 March 2019 Available online 18 April 2019 Keywords: Thin film solar cell CZTSSe Sodium doping Post-treatment Spin-coating method

a b s t r a c t In CZTSSe solar cells, a simple sodium-incorporation post-treatment method toward solution-processed Cu2 ZnSnS4 precursor films is presented in this work. An ultrathin NaCl film is deposited on Cu2 ZnSnS4 precursor films by spin-coating NaCl solution. In subsequent selenization process, the introduction of NaCl is found to be benefacial for the formation of Cu2- x Se, which can further facilitate the element transportation, leading to dense and smooth CZTSSe films with large grains and less impurity Cu2 Sn(S,Se)3 phase. SIMS depth profiles confirm the gradient distribution of the sodium element in Na-doped absorbers. Photoluminescence spectra show that the introduction of appropriate sodium into the absorber can inhibit the band tail states. As high as 11.18% of power conversion efficiency (PCE) is achieved for the device treated with 5 mg mL−1 NaCl solution, and an average efficiency of Na-doped devices is 10.71%, 13% higher than that of the control groups (9.45%). Besides, the depletion width and the charge recombination lifetime can also have regular variation with sodium treatment. This work offers an easy modification method for high-quality Na-doped CZTSSe films and high-performance devices, in the meantime, it can also help to further understand the effects of sodium in CZTSSe solar cells. © 2019 Science Press and Dalian Institute of Chemical Physics, Chinese Academy of Sciences. Published by Elsevier B.V. and Science Press. All rights reserved.

1. Introduction Kesterite semiconductors including Cu2 ZnSnS4 (CZTS), Cu2 ZnSnSe4 (CZTSe), Cu2 ZnSn(S,Se)4 (CZTSSe), have been attractive absorber materials for photovoltaics due to their outstanding advantages, such as appropriate band width (1.0–1.5 eV), high light absorption coefficients (>104 cm−1 ), non-toxic and earth-abundant component elements [1–4]. CZTSSe solar cells are widely considered to be a potential alternative to Cu(In,Ga)Se2 (CIGS) thin film photovoltaics. Currently, the PCE of CZTSSe solar cells is 12.6%, still far lower than that of the CIGS [5,6]. Since these two devices have

∗ Corresponding authors at: Key Laboratory for Renewable Energy (CAS), Beijing Key Laboratory for New Energy Materials and Devices, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China. E-mail addresses: [email protected] (D. Li), [email protected] (Q. Meng).

analogous device structures, CZTSSe solar cells indeed borrowed some strategies of CIGS to promote device performance [7–10]. Sodium or alkali incorporation has been proved to be a very fruitful methodology to enhance the PCEs of copper-based thin film solar cells [11,12]. The benefit of Na and other alkali elements (K, Rb, Cs) to CIGS devices is extraordinarily notable in achieving its world record-settings from passive Na-doping from soda lime glass (SLG) substrates to intentionally using NaF or KF post-deposition treatment (PDT) towards the CIGS absorber [12,13]. At present, the PCE of CIGS photovoltaic device has reached 22.6%, which is based on the heavy alkali elements Rb and Cs incorporation [14]. It has been generally suggested that alkali metal doping can (1) increase the hole concentration [15–17]; (2) reduce point defects [17,18]; (3) promote the crystal growth of the absorber [19] and (4) passivate grain boundaries [20].

https://doi.org/10.1016/j.jechem.2019.03.029 2095-4956/© 2019 Science Press and Dalian Institute of Chemical Physics, Chinese Academy of Sciences. Published by Elsevier B.V. and Science Press. All rights reserved.

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

Accordingly, some attempts have also been carried out in CZTSSe solar cells. Prabhakar et al. compared the polycrystalline CZTS films grown on SLG and borosilicate substrates, and found that both CZTS grain sizes and hole carrier concentration increased due to sodium diffusion from SLG [21]. Hlaing Oo et al. also observed larger grains of CZTS samples grown on SLG or dipped in Na2 S aqueous solution [22]. Sun et al. sputtered Na-doped Mo (Mo-Na) back contact for sodium incorporation, thus achieved CZTS solar cell with the efficiency of 6.2% on flexible stainless steel substrate [23]. Zhou et al. reported an effective defect passivation method by using CZTS:Na nanocrystals, which improved the device PCE from 3.89% to 6.14% [24]. Gershon et al. deposited NaF thin film on the top of Mo layer to obtain CZTSe devices with over 10% PCE, and revealed that Na-CZTSe films have shallower band tail states, and larger hole mobility [25]. Tiwari et al. simultaneously introduced NaCl and Sb(OAc)3 into the precursor solution to improve crystallization and reduce structural disorder of the CZTS absorber, and this Na:Sb co-doping could present better cell performance than only Na or Sb doping [26]. Neuschitzer et al. evaporated NaF+Ge onto the precursors before selenization, and demonstrated the interaction of Na and Ge could avoid the formation of deep defects [27]. In general, alkali metals (especially sodium) incorporation into CZTSSe absorbers can be realized by pre-deposition [23,25,28], post-deposition [27,29,30] or co-existence in the precursor solution [24,26], in contrast to fabricating CZTSSe precursor layers. Unlike CIGS, although some experimental and theoretical efforts have been done to confirm positive roles of the sodium incorporation in the CZTSSe solar cells, fundamental understanding toward the nature of the interaction between Na and other elements or layers in the device and the underlying mechanism have not been fully clear [11]. Besides, the doping process is still needed to be optimized as well. In this work, an easy and eco-friendly sodium incorporation post-treatment method has been developed for high quality CZTSSe absorbers, that is, NaCl ethanol/water solution is directly spin-coated onto Cu2 ZnSnS4 precursor films. The introduction of NaCl is found to enable to improve morphology and homogeneity of CZTSSe films. It is found that, the gradient distribution of the sodium ions in the absorber can reduce the localized defects of CZTSSe. Up to 11.18% PCE has been achieved by optimizing the concentration of NaCl solution, and the average PCE increase is more than 13% compared with the untreated devices. Further investigation reveals that the depletion width and the free carrier recombination lifetime can also be improved by sodium treatment. This work offers a simple method to prepare high-quality Na-doped CZTSSe films for high-performance devices, meanwhile, it can also help us to further understand the effects of sodium incorporation in CZTSSe solar cells.

2. Experimental 2.1. Reagents and materials Thioglycolic acid (HSCH2 COOH, 98%), 2-methoxyethanol (HOCH2 CH2 OCH3 , 99.5%), ethanolamine (HOCH2 CH2 NH2 , 99%), thiourea (TU, 99.0%), and cadmium sulfate (CdSO4 ·8/3H2 O, 99.99%) were purchased from Aladdin. Se particles (99.999%) were purchased from Zhongnuo Advanced Material (Beijing) Technology Co., Ltd. Ammonium hydroxide (AR) was from Beijing Chemical Works, NaCl (AR) was from Xilong Chemical Co., Ltd. CuO (99.9%, Zhongnuo), ZnO (99.99%, Sigma-Aldrich) and SnO (99.9%, Aladdin) were used as metal sources. All the chemicals were directly used without further purification.

197

2.2. Fabrication of CZTSSe solar cells The precursor solution was prepared by following our previous work [31–33]. In brief, CuO (0.1400 g, 1.76 mmol), ZnO (0.0977 g, 1.20 mmol) and SnO (0.1378 g, 1.02 mmol) were dissolved into a mixed solvent composed of 2-methoxyethanol (4 mL), thioglycolic acid (1.2 mL) and ethanolamine (2 mL). A clear yellow solution was obtained after 2 h stirring at room temperature. The fabrication process of precursor films was carried out in the nitrogenfilled glove box. The CZTS precursor solution was spin-coated at 50 0 0 rpm for 20 s on a 2 × 2 cm2 Mo-sputtered soda lime glass (SLG) substrate with a SiO2 layer as alkali diffusion barrier (SLG/SiO2 /Mo), which was annealed on 410 °C hot plate for 2 min. The same process was repeated six times to obtain desired thickness. For sodium doping post-treatment, a series of NaCl solutions with 1, 5, 10 mg mL−1 concentration in ethanol and ultra-pure water mixed solvent (volume ratio= 1:1) were prepared. NaCl solutions were spin-coated onto the above precursor films, followed by an annealing process on 60 °C hot plate for 10 min. Finally, asprepared CZTS precursor films with and without NaCl treatment were selenized at 540 °C for 15 min under Se atmosphere and nitrogen flow (105 Pa) in a rapid thermal processing (RTP) furnace. CZTSSe solar cells were fabricated with the structure of SLG/SiO2 /Mo/CZTSSe/CdS/ZnO/ITO/Ag. A 50 nm-thickness CdS buffer layer was deposited on the top of the CZTSSe absorber by chemical bath deposition (CBD) method, and then followed by 50 nm ZnO and 200 nm ITO as the window layer derived from the radio frequency (RF) magnetron sputtering. Finally, a 100 nm Ag grid was deposited by thermal evaporation, yielding an active area of 0.18 cm2 for each cell.

2.3. Characterization Compositions of CZTSSe absorbers were determined by an energy dispersive X-ray fluorescence (XRF) spectrometer (EDX-70 0 0, Shimadzu). X-ray diffraction (XRD) patterns of the absorber samples were collected on an X-ray diffractometer with Cu Kα as the radiation source (Empyrean, PANaltical). The film morphologies were obtained by scanning electron microscopy (S4800-SEM, Hitachi) and atomic force microscope (Multimode 8, Bruker). Steady-state photoluminescence (PL) spectrum was obtained on PL spectrometer (FLS 900, Edinburgh Instruments), using a pulsed diode laser (EPL-640, ∼20 nJ cm−2 ) as the excitation source and equipped with a low temperature controller (LakeShore, 335). The composition depth profiles were obtained by secondary-ion mass spectroscopy (Hiden SIMS). For analysis throughout the films, a primary Ar+ beam was used with an impact energy of 3 keV and a beam current of 200 nA. Current density-voltage (J-V) characteristics of the cells were collected on Keithley 2400 Source Meter under AM1.5 illumination (10 0 0 mW cm−2 ) from Zolix SS150A. Light intensity of the solar simulator was calibrated by a standard monocrystalline Si reference solar cell. The external quantum efficiency (EQE) was measured with a lab-made IPCE setup under 0.3–0.9 mW cm−2 monochromic light illumination without bias illumination, where both certified Si and Ge diodes were used as the reference detectors [34]. The capacitance-voltage (C-V) study was conducted on an electrochemical workstation (Versa STAT3, Princeton), which was performed at different DC bias voltages ranging from −1.0 to 0.5 V with a perturbation AC voltage of 50 mV at 100 kHz. Time-resolved photovoltage (TPV) of the cells was measured by a lab-made setup, in which the cells were excited by 532 nm pulse laser (20 Hz, 4 ns, Brio) with an ultralow light intensity of about 10 nJ cm−2 and the photovoltage decay process was recorded by a digital oscilloscope (DPO 7104, Tektronix) with an input impedance of 1 MΩ [35].

198

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

Fig. 1. Schematic illustration of the preparation of sodium incorporation posttreatment toward CZTS precursor films.

3. Results and discussion 3.1. Microstructure of CZTSSe films In general, three post-treatment methods have been developed to incorporate sodium into CZTSSe solar cells, including vacuum evaporation, dipping and spin-coating methods [29,36,37]. Spincoating method is easy and controllable, however, few work on this spin-coating post-treatment method has been reported. In this work, CZTS precursor films were treated by spin-coating NaCl ethanol/water solution before selenization, as shown in Fig. 1. For experiment details, please see Experimental Section. For clarity, NaCl solutions with 0, 1, 5 and 10 mg mL−1 concentration are labeled as Na0, Na1, Na5 and Na10, respectively. First, X-ray diffraction (XRD) was used to explore the influence of sodium introduction on the phase of the same batch of Na0∼Na10-CZTSSe films on quartz glass substrates. As can be seen in Fig. 2(a), no significant difference in XRD peak positions is found for all the four CZTSSe films, no matter sodium is doped or not, suggesting that the sodium incorporation does not change the kesterite phase, maybe due to a small amount of sodium incorporation cannot induce obvious lattice distortion in the bulk [38]. And the peaks are centered at 15.7°, 17.5°, 23.4°, 27.3°, 29.6°, 36.4°, 43.7°, 45.4° and 55.8° corresponding to (002), (101), (110), (112), (103), (202), (105), (220) and (312) planes of kesterite, respectively (PDF# 26-0575 and PDF# 52-0806). In the meantime, a small peak located at 24.6° is also found for Na0∼Na5-CZTSSe samples, as shown in Fig. 2(b). Besides, with the NaCl concentration increasing, this peak gradually decreases and finally disappears for Na10-CZTSSe sample. Further investigation reveals that this peak is assigned to 1¯ 1¯ 5 plane of Cu2 Sn(S,Se)3 (CTSSe) phase (according PDF# 27-0198). In fact, CZTSSe and CTSSe phases have similar XRD patterns with main peaks overlapping, however, this small peak at 24.6° belongs to CTSSe phase, not CZTSSe phase. Appearance of CTSSe phase may be caused by the local Zn deficiency during selenization. Obviously, sodium incorporation could restrain the formation of the CTSSe phase. Typically, the introduction of sodium is suggested to facilitate the mass transfer in the process of selenization and avoid local composition heterogeneity. Top-view SEM images of Na0 and Na5-CZTSSe films are presented in Fig. 2(c)–(f). As can be seen from Fig. 2(c) and (d), smaller grain sizes and nonuniformity are observed for the Na0CZTSSe film, in comparison to those for Na5-CZTSSe film. Specifically, the Na0-CZTSSe film is basically composed of 1–1.5 μm size grains with large grain boundaries, whereas Na5-CZTSSe grains with about 2 μm size connect tightly to each other to give a more uniform film. Obviously, sodium incorporation indeed increases the crystal size and improves the CZTSSe crystallinity, in line with the GIGS. Generally, the selenization process follows a liquid-assisted grain growth mechanism, which includes five steps below: (1) Se(l) condensed from the vapor phase penetrates into the precursor films; (2) Cu is dissolved into Se(l); (3) Cu2- x Se nucleation occurs at the surface of the films; (4) Sn and Zn are incorporated into

the growing grains as these matter are dissolved/transported in the liquid phase/solid-state diffusion; (5) finally lager-grain CZTSSe forms [39]. According to this grain growth mechanism, Cu2- x Se as the flux agent could assist the matter transport and grain growth. However, the formation of Cu2- x Se is strongly dependent on the existence of alkali metals (i.e. sodium). According to previous report, if no extra sodium is involved, the Cu2- x Se phase will not form until the low-sodium borosilicate glass substrate is heated up to 275 °C. When an extra sodium is involved, the Cu2- x Se appears just at 225 °C [40]. Obviously, in the sodium-assisted selenization process, the Cu2- x Se formation temperature can be effectively reduced [12]. Consequently, Na5-CZTSSe films have higher quality crystalline grains than Na0 films. Besides, according to cross-sectional images in Fig. S1(a) and (b) of the supplementary information, similar thicknesses of larger CZTSSe grain layer and fine grain layer are also observed in both Na0 and Na5-CZTSSe absorbers, suggesting that sodium incorporation cannot thoroughly eliminate the fine grain layer or improve the crystallization of the fine grain layer either. In fact, according to previous reports, the formation of small grain sublayer is highly related to the carbon residue caused by the unvolatilized organic solvent [41]. The surface morphology of CZTSSe absorber layers is further investigated by atomic force microscope (AFM). As shown in Fig. 3(a) and (b), Na0-CZTSSe film has a clear grain structure and surface relief, whereas the surface roughness of Na5-CZTSSe film significantly decreases. Here, height profiles across grain boundaries are derived from AFM morphologies by selecting three typical positions. As for Na0-CZTSSe films, the height difference between the grain center and the boundary ranges from 250 to 300 nm, whereas for Na5-CZTSSe films, the height difference is reduced to 150–200 nm, as shown in Fig. 3(c) and (d). This further verifies that the CZTSSe absorber derived from the precursor CZTS film treated by 5 mg mL−1 NaCl solution is much smoother and more compact than untreated one. And this smooth surface is supposed to be beneficial for decreasing both the dark current and interfacial states in the device when a p-n junction is formed between the CZTSSe and CdS buffer layer [42]. As we know, appropriate elemental constituent (i.e. Cu/(Zn+Sn), Zn/Sn ratios) is critical for high quality CZTSSe films and highly efficient devices. Theoretical calculation and extensive experiments have shown that Cu-poor and Zn-rich growth condition with Cu/(Zn+Sn) ≈ 0.8 and Zn/Sn ≈ 1.2, favor a high quality CZTSSe phase with fewer defects, and the S/Se ratio will influence the bandgap as well [2,43,44]. Here, an energy dispersive X-ray fluorescence (XRF) spectrum has been employed to see if Na-doping amounts will influence on the composition and bandgaps of Na0– Na10-CZTSSe films. To all the three Na-doping samples, no significant difference in Cu/(Zn+Sn), Zn/Sn and S/(S+Se) ratios is found, which are basically kept at about 0.8, 1.2, and 0.12, respectively, comparable to Na0-CZTSSe sample. Meanwhile, their bandgaps of these absorbers are estimated to be 1.09 eV from external quantum efficiency (EQE) curves by fitting a plot of [hν · ln (1 − EQE)]2 vs. photon energy, as shown in Fig. S2 and Table S1 in the supplementary information. Obviously, their bandgaps are independent of Na-doping and doping amounts (Here, a negligible difference in bandgaps estimated from Fig. S2 of the supplementary information is basically due to the measurement error). This is in good agreement with XRD patterns of their corresponding CZTSSe films, indicating that sodium incorporation does not influence the kesterite phase of CZTSSe absorbers. 3.2. Investigation on sodium distribution in Na-doped CZTSSe absorbers In order to further understand the dynamic distribution of sodium ions during the device preparation, SIMS depth composi-

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

199

Fig. 2. (a) XRD patterns of Na0–Na10-CZTSSe films onto quartz glass; (b) Zoomed XRD patterns in the range of 24°–25.5°; Top-view SEM images of Na0-CZTSSe film (c, e) and Na5-CZTSSe films (d, f).

Fig. 3. AFM topography images and corresponding height profiles across gain boundaries of Na0-CZTSSe (a, c) and Na5-CZTSSe films (b, d).

tional profiles and relevant schematic illustrations of three samples, Na5-CZTS precursor film, Na5-CZTSSe film (selenized) and CdS/Na5-CZTSSe sample, are shown in Fig. 4. We can see that, the elemental composition for all samples can be well identified in SIMS profiles. For Na5-CZTS precursor film, Na+ is found to accumulate on the surface of the CZTS precursor film after being spin-coated by the NaCl solution, and some sodium diffuses into the whole precursor film. Then, selenization of the Na5-CZTS precursor film produced bilayer-structure Na5-CZTSSe film. The largegrained layer is the CZTSSe phase, and the fine-grained layer is a

mixture of CZTSSe and other chalcogenide nanocrystals embedded in a matix of carbon residue. Thus in the small grain sublayer, the SIMS signals of metallic element are relatively low, and the carbon signal is relatively high [45]. For Na5-CZTSSe film, more sodium is found in fine grain layer than that in large CZTSSe crystal layer, meanwhile, an ultrathin Na layer still exists on the top of the absorber. When CdS was further deposited onto as-prepared Na5CZTSSe by CBD method, the Na layer on the top of CZTSSe disappeared, however, no obvious change in CZTSSe bulk. According to previous study, cation substitutions such as NaCu and NaZn may

200

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

Fig. 4. SIMS depth compositional profiles and corresponding schematic illustrations of Na5-CZTS precursor film (a, b), Na5-CZTSSe selenized film (c, d) and CdS/Na5-CZTSSe sample (e, f).

occur due to relatively low formation energy [46]. When CdS is deposited on the top of CZTSSe layer by CBD method, some Na+ from the surface of CZTSSe will dissolve into water, leaving some copper vacancy (VCu ) on the CZTSSe surface. And these VCu will be occupied by Cd2+ , beneficial for the formation of a homogenous p-n junction [16]. Accordingly, the gradient distribution of sodium in the CZTSSe absorber is suggested, which impact on CZTSSe solar cells will be discussed below. 3.3. Photovoltaic performance of CZTSSe solar cells As-prepared CZTSSe absorbers were fabricated into CZTSSe solar cells with the structure of SLG/SiO2 /Mo/CZTSSe/CdS/ZnO/ITO/Ag, as shown in Fig. 5(a). Here, 18 devices for each condition were selected in order to investigate the influence of sodium incorporation on the cell performance. The statistical cell parameters (JSC , VOC , FF and PCE) are shown in Fig. 5(c)–(f) and detailed values are listed in Table 1. We can see that, when no excess Na is introduced into the absorber, the CZTSSe solar cell can give 9.95% PCE with 34.9 mA cm−2 of JSC , 443 mV of VOC and 0.645 of FF. When an excess Na+ is introduced into the absorbers, the averaged JSC , VOC , FF and PCE continuously rise at first, and then decline.

A champion CZTSSe solar cell is obtained when the concentration of NaCl solution is 5 mg mL−1 , which PCE can reach 11.18% with 37.1 mA cm−2 of JSC , 461 mV of VOC and 0.654 of FF, respectively. And the average efficiency increased by more than 13% from 9.45% (control) to 10.71%. However, when the concentration of NaCl solution further increases to 10 mg mL−1 , the device presents relatively poor cell performance with an average PCE of 8.11%. Here, one thing has to be mentioned, the variation tendency of JSC is in good agreement with that of the PCEs, however, this increase in JSC cannot be assigned to the change in light absorption range, due to almost the same bandgap of Na0–Na10-CZTSSe absorbers (Table S1 in the supplementary information). External quantum efficiency (EQE) curves (Fig. 6a) show that the photoelectric response intensity in medium- and long-wavelength ranges gradually increases from Na0 to Na5-CZTSSe. This improvement suggests more efficient photo-generated carrier collection capability, which may be benefited from fewer defects in the CZTSSe bulk and wider depletion region of CZTSSe/CdS interface. In order to evaluate the CZTSSe absorbers and the interfaces in these devices, C-V measurement was performed at different DC bias voltages ranging from −1.0 to 0.5 V with a perturbation AC voltage of 50 mV at 100 kHz. The C − 2 vs. V curves are given as

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

201

Fig. 5. (a) CZTSSe device structure; (b) J-V curves of Na0 (black line) and Na5 (orange line) solar cells; Statistical box charts of Na0–Na10 solar cells: (c) JSC , (d) VOC , (e) FF and (f) PCE. 18 devices for each condition were selected. Table 1. Photovoltaic parameters, electrical characteristics and TPV lifetimes of Na0–Na10 solar cellsa .

τ TPV

Solar cells

JSC (mA cm−2 )

VOC (mV)

FF

PCE (%)

WD (nm)

NCV (cm−3 )

Na0

34.0 (34.9) 35.5 (36.4) 36.8 (37.1) 36.1 (37.5)

441 (443) 442 (446) 452 (461) 404 (404)

0.63 (0.64) 0.64 (0.64) 0.64 (0.65) 0.56 (0.60)

9.45 (9.95) 10.12 (10.35) 10.71 (11.18) 8.11 (9.07)

246

1.47 × 1016

112

257

9.28 × 10

15

172

265

9.92 × 1015

239

221

1.27 × 1016

62

Na1 Na5 Na10 a

(μs)

Values in parentheses are the average parameters of 18 devices.

Fig. S3 in the supplementary information. Electrical properties of Na0–Na10 solar cells, including charge density (NCV ) and depletion width (WD ), are given in Table 1. The profiling position (W) is derived from the following equation [47,48]:

W = ε0 εr A/C

(1)

where C is measured capacitance for each bias, A is the device area and ε r is dielectric constant for CZTSSe absorber. Here, ε r

is fixed to be 8.1 based on the assumption, that is, ε r is more affected by anion [49]. The WD is thus derived at zero bias. As shown in Fig. 6(b) and Table 1, the charge densities about 1016 cm−3 for all the samples are almost not influenced by sodium incorporation amounts. However, the depletion width is gradually extended from 246 to 265 nm for Na0 to Na5-CZTSSe, then drops to 221 nm for Na10-CZTSSe. As discussed above, VCu is generated from NaCu when some Na+ ions are dissolved into water during

202

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

Fig. 6. (a) EQE curves; (b) C-V profiles and (c) transient photovoltage decay (TPV) spectra of Na0–Na10 solar cells; (d) Normalized steady-state PL emission spectra of Na0 (black line) and Na5-(orange line) CZTSSe films on the quartz glass at 40 K.

CBD, on the other hand, some Cd2+ ions occupy VCu acceptors to give CdCu donors on the surface of CZTSSe [13]. It is supposed that the concentration of CdCu is higher than that of undoped sample, and the electronic inversion of Na5-CZTSSe surface is more intense. It thus strengthens the buried homogenous p-n junction, which could be the reason for wider depletion region. This relatively wider depletion width can promote the separation of photogenerated electron-hole pairs and be beneficial for carrier transportation, in good agreement with the variation of corresponding cell performance. For a practical device, the transient photovoltage (TPV) method is a useful technique to directly investigate the process in time scale ranging from picoseconds to milliseconds [50,51]. Here, TPV spectra are obtained under open-circuit condition in the dark, as presented in Fig. 6(c). According to fitted results by single exponential mode, it is found that, the TPV lifetime (τ TPV ) increases from 112, 172 to 239 μs for Na0, Na1 and Na5 solar cells, as listed in Table 1. Amongst, the Na5 solar cell exhibits the longest τ TPV (τ TPV = 239 μs), twice as that of the device without sodium treatment (τ TPV = 112 μs). But the Na10 solar cell presents the shortest τ TPV (τ TPV = 62 μs). As we know, relatively longer decay time in the complete device implies less recombination. It is thus supposed that appropriate amount of Na+ introduction can reduce defects of CZTSSe itself or inhibit the band tail states to reduce free carrier recombination, thus beneficial for the carrier collection and cell performance. However, too much sodium involved in the absorber will increase recombination possibility, leading to unsatisfied cell performance. To further understand the influence of sodium incorporation on photogenerated carrier recombination in the absorber, low-temperature steady-state photoluminescence (PL) measurement was carried out for Na0 and Na5-CZTSSe films

on the quartz glass, as shown in Fig. 6(d). The PL emission peak position of Na5-CZTSSe film is slightly blue-shifted, in comparison to that of Na0 film. Obviously, for Na5-CZTSSe film, the difference between the Eg and the PL peak is relatively smaller than that for Na0-CZTSSe film, suggesting that some band tail states are inhibited by sodium incorporation. That is, the shallow defect states of Na5-CZTSSe film are reduced [52]. As mentioned above, appropriate sodium incorporation could improve both the morphology and the electrical properties of CZTSSe films, however, more sodium incorporation (i.e. Na10 devices) shows unsatisfactory performance (low FF, narrow depletion region, short TPV lifetime). An unsatisfied p-n junction is supposed to be the main reason to this phenomenon. As shown in Fig. S1(c) of the supplementary information, there are many bright spots on the top-view SEM image of Na10 film, which are probably residual sodium-rich particles (i.e. NaSex ) after selenization. As we know, some Na+ may dissolve in water when CdS is deposited on the surface of Na10 film during CBD, however, the existence of relatively larger sodium-rich particles will strongly influence the surface activity of selenized films even the CdS deposition, see Fig. 4(c) and (f). On one hand, these sodium-rich particles on the surface could cause poor contact of p-n junction, on the other hand, the band structure of CdS may also be changed, thus leading to the poor performance of Na10 devices. Further investigation on how to eliminate these residual sodium-rich particles, is also needed in the future. 4. Conclusions In this work, we report a simple, nontoxic, sodium incorporation post-treatment method for high quality CZTSSe films and

B. Duan, L. Guo and Q. Yu et al. / Journal of Energy Chemistry 40 (2020) 196–203

high efficiency CZTSSe solar cells. That is, NaCl water/ethanol mixed solutions are directly spin-coated onto the precursor films. Introduction of Na ions is suggested to facilitate the formation of Cu2- x Se as an intermediator, which is further beneficial for the element transfer during selenization. In the meantime, dense and smooth CZTSSe films with large grains and less impurity Cu2 Sn(S,Se)3 phase are also obtained. Sodium incorporation is proposed to improve the CZTSSe crystal quality during selenization. Further investigation reveals that, Na ions exhibit gradient distribution in CZTSSe absorber by the aid of SIMS profiles. Besides, it is suggested that appropriate Na+ amount and its gradient distribution could decrease the band-tail states density and reduce carrier recombination, leading to better cell performance. Besides, appropriate sodium-doping can broaden the depletion width of p-n junction and promote back interface performance, leading to higher JSC . As high as 11.18% of power conversion efficiency has been achieved for the device treated with 5 mg mL−1 NaCl solution, and the average efficiency has increased by more than 13% from 9.45% (untreated) to 10.71%. Our work offers an easy method to prepare high-quality Na-doped CZTSSe films and highperformance devices as well; in the meantime, it can also help us to further understand the effects of sodium in CZTSSe solar cells. Author contributions The manuscript was written through contributions of all authors. All authors have approval to the final version of the manuscript. Notes The authors declare no competing financial interest. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (Nos. 51421002, 51627803, 91733301, 51761145042, 21501183, 51402348, 53872321, and 11874402), the Knowledge Innovation Program and the Strategic Priority Research Program (Grant XDB 12010400) of the Chinese Academy of Sciences. The authors thank Jieru Xu and Hua Zhang for technical assistance and use of facilities at secondary-ion mass spectroscopy (SIMS) at Institute of Physics. The authors also would like to thank Jionghua Wu and Yiming Li for TPV measurement and discussion. Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jechem.2019.03.029. References [1] [2] [3] [4]

A. Walsh, S. Chen, S.H. Wei, X.G. Gong, Adv. Energy Mater. 2 (2012) 400–409. S. Chen, A. Walsh, X.G. Gong, S.H. Wei, Adv. Mater. 25 (2013) 1522–1539. Y. Fan, H. Qin, B. Mi, Z. Gao, W. Huang, Acta Chim. Sin. 72 (2014) 643–652. X. Zhang, Y. Han, S.-Z. Chai, N.-T. Hu, Z. Yang, H.-J. Geng, H. Wei, Acta Phys. Chim. Sin. 32 (2016) 1330–1346. [5] W. Wang, M.T. Winkler, O. Gunawan, T. Gokmen, T.K. Todorov, Y. Zhu, D.B. Mitzi, Adv. Energy Mater. 4 (2014) 1301465. [6] NREL, Photovoltaic research. https://www.nrel.gov/pv/assets/pdfs/pvefficiency-chart.pdf, 2019 (accessed 3 January 2019). [7] I. Repins, C. Beall, N. Vora, C. DeHart, D. Kuciauskas, P. Dippo, B. To, J. Mann, W.-C. Hsu, A. Goodrich, R. Noufi, Sol. Energy Mater. Sol. Cells 101 (2012) 154–159.

203

[8] H. Xin, S.M. Vorpahl, A.D. Collord, I.L. Braly, A.R. Uhl, B.W. Krueger, D.S. Ginger, H.W. Hillhouse, Phys. Chem. Chem. Phys. 17 (2015) 23859–23866. [9] K.-J. Yang, D.-H. Son, S.-J. Sung, J.-H. Sim, Y.-I. Kim, S.-N. Park, D.-H. Jeon, J. Kim, D.-K. Hwang, C.-W. Jeon, D. Nam, H. Cheong, J.-K. Kang, D.-H. Kim, J. Mater. Chem. A 4 (2016) 10151–10158. [10] F. Liu, C. Yan, J. Huang, K. Sun, F. Zhou, J.A. Stride, M.A. Green, X. Hao, Adv. Energy Mater. 6 (2016) 1600706. [11] P.M.P. Salomé, H. Rodriguez-Alvarez, S. Sadewasser, Sol. Energy Mater. Sol. Cells 143 (2015) 9–20. [12] Y. Sun, S. Lin, W. Li, S. Cheng, Y. Zhang, Y. Liu, W. Liu, Engineering 3 (2017) 452–459. [13] A. Chirila, P. Reinhard, F. Pianezzi, P. Bloesch, A.R. Uhl, C. Fella, L. Kranz, D. Keller, C. Gretener, H. Hagendorfer, D. Jaeger, R. Erni, S. Nishiwaki, S. Buecheler, A.N. Tiwari, Nat. Mater. 12 (2013) 1107–1111. [14] P. Jackson, R. Wuerz, D. Hariskos, E. Lotter, W. Witte, M. Powalla, Phy. Status Solidi RRL 10 (2016) 583–586. [15] T. Nakada, D. Iga, H. Ohbo, A. Kunioka, Jpn. J. Appl. Phys. 36 (1997) 732–737. [16] F. Pianezzi, P. Reinhard, A. Chirila, B. Bissig, S. Nishiwaki, S. Buecheler, A.N. Tiwari, Phys. Chem. Chem. Phys. 16 (2014) 8843–8851. [17] S.-H. Wei, S.B. Zhang, A. Zunger, J. Appl. Phys. 85 (1999) 7214–7218. [18] P.T. Erslev, J.W. Lee, W.N. Shafarman, J.D. Cohen, Thin Solid Films 517 (2009) 2277–2281. [19] D. Braunger, D. Hariskos, G. Bilger, U. Rau, H.W. Shock, Thin Solid Films 361-362 (20 0 0) 161–166. [20] L. Kronik, D. Cahen, H.W. Schock, Adv. Mater. 10 (1998) 31–36. [21] T. Prabhakar, N. Jampana, Sol. Energy Mater. Sol. Cells 95 (2011) 1001–1004. [22] W.M. Hlaing Oo, J.L. Johnson, A. Bhatia, E.A. Lund, M.M. Nowell, M.A. Scarpulla, J. Electr. Mater. 40 (2011) 2214–2221. [23] K. Sun, F. Liu, J. Huang, C. Yan, N. Song, H. Sun, C. Xue, Y. Zhang, A. Pu, Y. Shen, J.A. Stride, M. Green, X. Hao, Sol. Energy Mater. Sol. Cells 182 (2018) 14–20. [24] H. Zhou, T.B. Song, W.C. Hsu, S. Luo, S. Ye, H.S. Duan, C.J. Hsu, W. Yang, Y. Yang, J. Am. Chem. Soc. 135 (2013) 15998–16001. [25] B.T. Gershon, Y.S. Lee, R. Mankad, O. Gunawan, T. Gokmen, D. Bishop, B. McCandless, S. Guha, Appl. Phys. Lett. 106 (2015) 123905. [26] D. Tiwari, T. Koehler, X. Lin, R. Harniman, I. Griffiths, L. Wang, D. Cherns, R. Klenk, D.J. Fermin, Chem. Mater. 28 (2016) 4991–4997. [27] M. Neuschitzer, M.E. Rodriguez, M. Guc, J.A. Marquez, S. Giraldo, I. Forbes, A. Perez-Rodriguez, E. Saucedo, J. Mater. Chem. A 6 (2018) 11759–11772. [28] T. Mise, S. Tajima, T. Fukano, K. Higuchi, T. Washio, K. Jimbo, H. Katagiri, Prog. Photovolt Res. Appl. 24 (2016) 1009–1015. [29] C.M. Sutter-Fella, J.A. Stückelberger, H. Hagendorfer, F. La Mattina, L. Kranz, S. Nishiwaki, A.R. Uhl, Y.E. Romanyuk, A.N. Tiwari, Chem. Mater. 26 (2014) 1420–1425. [30] Y.-R. Lin, V. Tunuguntla, S.-Y. Wei, W.-C. Chen, D. Wong, C.-H. Lai, L.-K. Liu, L.-C. Chen, K.-H. Chen, Nano Energy 16 (2015) 438–445. [31] Q. Tian, G. Wang, W. Zhao, Y. Chen, Y. Yang, L. Huang, D. Pan, Chem. Mater. 26 (2014) 3098–3103. [32] X. Min, J. Shi, L. Guo, Q. Yu, P. Zhang, Q. Tian, D. Li, Y. Luo, H. Wu, Q. Meng, S. Wu, Chin. Phys. B 27 (2018) 016402. [33] X. Min, L. Guo, Q. Yu, B. Duan, J. Shi, H. Wu, Y. Luo, D. Li, Q. Meng, Sci. Chin. Mater. (2018) 1–6. [34] X.Z. Guo, Y.H. Luo, Y.D. Zhang, X.C. Huang, D.M. Li, Q.B. Meng, Rev. Sci. Instrum. 81 (2010) 103106. [35] J. Shi, D. Li, Y. Luo, H. Wu, Q. Meng, Rev. Sci. Instrum. 87 (2016) 123107. [36] G. Altamura, M. Wang, K.L. Choy, Sci. Rep. 6 (2016) 22109. [37] A. Mule, B. Vermang, M. Sylvester, G. Brammertz, S. Ranjbar, T. Schnabel, N. Gampa, M. Meuris, J. Poortmans, Thin Solid Films 633 (2017) 156–161. [38] Z.Y. Zhao, X. Zhao, Inorg. Chem. 53 (2014) 9235–9241. [39] C.J. Hages, M.J. Koeper, C.K. Miskin, K.W. Brew, R. Agrawal, Chem. Mater. 28 (2016) 7703–7714. [40] A. Brummer, V. Honkimäki, P. Berwian, V. Probst, J. Palm, R. Hock, Thin Solid Films 437 (2003) 297–307. [41] V.T. Tiong, Y. Zhang, J. Bell, H. Wang, RSC Adv. 5 (2015) 20178–20185. [42] A.M. Gabor, J.R. Tuttle, D.S. Albin, M.A. Contreras, R. Noufi, A.M. Hermann, Appl. Phys. Lett. 65 (1994) 198–200. [43] D. Shin, B. Saparov, D.B. Mitzi, Adv. Energy Mater. 7 (2017) 1602366. [44] W. Xiao, J.N. Wang, X.S. Zhao, J.W. Wang, G.J. Huang, L. Cheng, L.J. Jiang, L.G. Wang, Sol. Energy 116 (2015) 125–132. [45] H. Zhou, W.-C. Hsu, H.-S. Duan, B. Bob, W. Yang, T.-B. Song, C.-J. Hsu, Y. Yang, Energy Environ. Sci. 6 (2013) 2822–2838. [46] T. Maeda, A. Kawabata, T. Wada, Phys. Status Solidi C 12 (2015) 631–637. [47] P.A. Barenes, Characterization of Materials, second ed., John Wiley & Sons, 2012. [48] S.S. Hegedus, W.N. Shafarman, Prog. Photovolt: Res. Appl. 12 (2004) 155–176. [49] C. Persson, J. Appl. Phys. 107 (2010) 053710. [50] J. Shi, H. Wei, S. Lv, X. Xu, H. Wu, Y. Luo, D. Li, Q. Meng, ChemPhysChem. 16 (2015) 842–847. [51] Y. Li, Y. Li, J. Shi, H. Li, H. Zhang, J. Wu, D. Li, Y. Luo, H. Wu, Q. Meng, Appl. Phys. Lett. 112 (2018) 053904. [52] M.J. Romero, H. Du, G. Teeter, Y. Yan, M.M. Al-Jassim, Phys. Rev. B 84 (2011) 165324.