Guest Editor: J. Bachmann
Beilstein J. Nanotechnol. 2019, 10, 198–210. doi:10.3762/bjnano.10.18
Received 22 Sep 2018, Accepted 04 Dec 2018, Published 15 Jan 2019
Antimony sulfide (Sb2S3), an environmentally benign material, has been prepared by various deposition methods for use as a solar absorber due to its direct band gap of ≈1.7 eV and high absorption coefficient in the visible light spectrum (1.8 × 105 cm−1 at 450 nm). Rapid, scalable, economically viable and controllable in-air growth of continuous, uniform, polycrystalline Sb2S3 absorber layers has not yet been accomplished. This could be achieved with chemical spray pyrolysis, a robust chemical method for deposition of thin films. We applied a two-stage process to produce continuous Sb2S3 optical coatings with uniform thickness. First, amorphous Sb2S3 layers, likely forming by 3D Volmer–Weber island growth through a molten phase reaction between SbCl3 and SC(NH2)2, were deposited in air on a glass/ITO/TiO2 substrate by ultrasonic spraying of methanolic Sb/S 1:3 molar ratio solution at 200–210 °C. Second, we produced polycrystalline uniform films of Sb2S3 (Eg 1.8 eV) with a post-deposition thermal treatment of amorphous Sb2S3 layers in vacuum at 170 °C, <4 × 10−6 Torr for 5 minutes. The effects of the deposition temperature, the precursor molar ratio and the thermal treatment temperature on the Sb2S3 layers were investigated using Raman spectroscopy, X-ray diffraction, scanning electron microscopy, energy dispersive X-ray spectroscopy and UV–vis–NIR spectroscopy. We demonstrated that Sb2S3 optical coatings with controllable structure, morphology and optical properties can be deposited by ultrasonic spray pyrolysis in air by tuning of the deposition temperature, the Sb/S precursor molar ratio in the spray solution, and the post-deposition treatment temperature.
Keywords: antimony sulfide; thin films; ultrasonic spray; vacuum annealing; Volmer–Weber growth
Antimony sulfide (Sb2S3) is an environmentally benign material. As Sb and S are abundant elements in the Earth’s crust, enough raw materials can be supplied to manufacture large quantities of Sb2S3 in the long term. Sb2S3 can be applied as the inorganic absorber in solar cells due to its direct band gap of ≈1.7 eV [1,2].
Sb2S3, prepared by a chemical bath deposition (CBD) [3,4], spin coating , atomic layer deposition (ALD)  or chemical spray pyrolysis (CSP)  method, has been applied in extremely thin absorber (ETA) solar cells due to its excellent absorption coefficient in the visible light spectrum (1.8 × 105 cm−1 at 450 nm) [1,2]. Improvements in photocurrent density have been sought by utilizing a transparent, nanostructured window layer instead of planar window layers with the ETA Sb2S3 absorber layer [4,7]. Previous studies show that achieving sufficient repeatability alongside optimization of the component layers, i.e., transparent (structured) window layer, Sb2S3 absorber layer, and hole transport material layer, and their respective interfaces, is a tremendous undertaking .
Attention has surged toward planar heterojunction Sb2S3 solar cells due to their simpler structure, less intricate production, and enhanced repeatability vs structured solar cells . Planar ≈1.7 eV absorber layers can be applied in semitransparent solar cells as well as in tandem solar cells.
Chemical spray pyrolysis (CSP) is a robust and industrially scalable chemical method for rapid deposition of thin films . Our research group first investigated spray-deposited Sb2S3 by pneumatically spraying aqueous solutions (tartaric acid added as complexing agent to prevent hydrolysis , akin to studies by Rajpure et al. ) or methanolic solutions of SbCl3. Following, we studied the effect of the Sb/S precursor molar ratio in solution on ultrasonically sprayed Sb2S3 layers and presented the first planar TiO2/Sb2S3/P3HT solar cells comprising ultrasonically sprayed Sb2S3 (power conversion efficiency η ≤ 1.9%) .
SbCl3 and thiourea (SC(NH2)2) are often used in the field to deposit Sb2S3 thin films. Spraying the SbCl3/SC(NH2)2 (henceforth Sb/S) 1:6 molar ratio solution at 250 °C in air yielded separate Sb2S3 grains, which did not cover the TiO2 substrate entirely, whereas spraying the Sb/S 1:3 solution yielded an inhomogeneous mix of amorphous and polycrystalline Sb2S3 . We learned to produce continuous uniform layers of polycrystalline Sb2S3 by a two-step process on ZnO nanorod/TiO2 substrates . In this study, we applied this two-step process, i.e., depositing amorphous Sb2S3 layers on planar substrates, followed by post-deposition crystallization.
The aim of this study was to produce crystalline, continuous, Sb2S3 optical coatings with uniform thickness to be applied as a photovoltaic absorber by ultrasonic spraying on planar glass/ITO/TiO2 substrates, followed by a post-deposition treatment. To this end, we studied the effect of the deposition temperature (TD), the molar ratio of precursors SbCl3 and thiourea (SC(NH2)2) in the spray solution, and the post-deposition treatment temperature on the structure, morphology and optical properties of ultrasonically sprayed Sb2S3 thin films.
Two sequential operations were used to obtain homogeneous Sb2S3 optical coatings with uniform thickness on planar TiO2 substrates. First, we tuned the deposition temperature and molar ratio of Sb/S precursors in spray solution to deposit continuous amorphous Sb2S3 layers. An intimate contact, which is a prerequisite for high power conversion efficiency in solar cells , is formed at the interface between TiO2 and Sb2S3 during deposition of amorphous Sb2S3 layers. Second, all layers were thermally treated in an inert environment (vacuum, <4 × 10−6 Torr) to induce crystallization, without oxidation.
Preliminary experiments at deposition temperatures lower than 182 °C (decomposition of SC(NH2)2 [14,15]) yielded inhomogeneous red-brown layers. Furthermore, in our previous paper, 250 °C was found to be too high a deposition temperature to obtain sufficient coverage of TiO2 substrate by polycrystalline Sb2S3 thin films, despite the suitable band gap of 1.6 eV and high phase purity . Restricted to deposition temperatures in the range 182–250 °C, we sprayed Sb/S 1:3 and 1:6 molar ratio precursor solutions at TD = 200, 210, and 220 °C. We varied the aforementioned parameters to attain the conditions to deposit dense and homogeneous layers of amorphous Sb2S3, which we then crystallized by a post-deposition thermal treatment.
Based on the scanning electron microscopy (SEM) images, preliminary experiments revealed that spraying Sb/S 1:6 solutions consistently yielded twice thinner layers compared to layers deposited from Sb/S 1:3 solutions. Sb2S3 layers of comparable thickness were deposited by spraying Sb/S 1:6 solutions for 40 minutes and Sb/S 1:3 solutions for 20 minutes.
The samples are named in the text as follows: A-B-C, where A is the S/Sb molar ratio in solution, B is the deposition temperature, and C is the specification of the treatment. [Sb/S molar ratio in solution: “3” for Sb/S 1:3 or “6” for Sb/S 1:6]-[deposition temperature: “200”, “210” or “220” (°C)]-[treatment: “As-dep.” for as-deposited and “170”, “200” or “250” (°C) for samples thermally treated in vacuum].
The samples in which Sb2S3 layers were deposited from either Sb/S 1:3 or 1:6 solution at TD = 200 °C, followed by thermal treatment in vacuum at 200 °C (3-200-200, 6-200-200), contain no Sb2S3, as it likely volatilized completely during the vacuum thermal treatment. Likewise, treating the Sb2S3 layers at temperatures higher than 200 °C caused Sb2S3 to completely volatilize during treatment. Photographs of the samples (Figure S1) and the description of the vapor pressure calculations (Comment S1) are provided in the Supporting Information File 1. Consequently, only as-deposited samples and samples thermally treated in vacuum at 170 °C and 200 °C are eligible for discussion.
Raman spectroscopy provides quantitative and qualitative information on the vibrational modes in solids. The wide Raman band centered at 290 cm−1 [12,16] associated with metastibnite, i.e., amorphous Sb2S3, is characteristic of as-deposited orange colored (photograph in Supporting Information File 1, Figure S1) samples (3-200-As-dep., 3-210-As-dep., Figure 1A; 6-200-As-dep., Figure 1B). The band centered at 145 cm−1 is a low frequency Eg vibrational mode of anatase-TiO2 , which is observed due to the laser beam penetrating to the substrate [12,16] through the discontinuous Sb2S3 layers. The TiO2 vibrational band is absent in spectra of Sb2S3 layers containing less pinholes, as the signal is captured only from Sb2S3.
The narrower bands, attributed to orthorhombic Sb2S3 [16,18-20], are present in the spectra of as-deposited and thermally treated lustrous gray (photograph in Supporting Information File 1, Figure S1) samples (3-200-170, 3-210-170, 3-210-200, 3-220-As-dep., 3-220-170, 3-220-200, Figure 1A; 6-200-170, 6-210-As-dep., 6-210-170, 6-210-200, 6-220-As-dep., 6-220-170, 6-220-200, Figure 1B; photograph in Supporting Information File 1, Figure S1). According to group theory, orthorhombic Sb2S3 has 30 predicted Raman active modes: ΓRaman = 10Ag + 5B1g + 10B2g + 5B3g [18,20]. The Raman spectra were deconvoluted using Lorentzian fitting into vibrational bands of Sb2S3 based on the literature [12,16,21,22]. The centers of the bands of Sb2S3 in the deconvoluted Raman spectra (Table 1, symmetries taken from [20,21]) are similar to values reported in our previous studies [7,12]. Band centers, relative single peak intensities and full widths at half maximum (FWHM) of the narrow bands centered at 281, 301 and 310 cm−1 can be respectively found in Tables S1, S2, and S3 of Supporting Information File 1.
Table 1: Raman band centers and assigned active modes for the studied Sb2S3 layers.
|Center of Raman band, cm−1||Symmetry||Vibrational mode, [21-23]|
|This study||Ref. ||Ref. ||Ref. ||Ref. |
|188||189||186||B1g||B1g||antisym. S–Sb–S bending|
|237||237||239||B1g||B1g/B3g||symmetric S–Sb–S bending|
|281||281||282||Ag||Ag/B2g||antisym. S–Sb–S stretching|
|301||300||299||Ag||Ag/B2g||antisym. S–Sb–S stretching|
|310||310||312||Ag||Ag/B2g||symmetric S–Sb–S stretching|
The FWHM of the vibrational band centered at 281 cm−1 narrows from ≈24 cm−1 to 21–23 cm−1 after vacuum thermal treatment of the samples deposited at 210–220 °C from both Sb/S 1:3 and Sb/S 1:6 solutions (3-210-170, 3-220-170, 6-210-170 and 6-220-170) at 170 °C (3-210-170, 3-220-170, 6-210-170 and 6-220-170) and narrows by 5 cm−1 at most after vacuum thermal treatment at 200 °C (3-210-200). The narrowing of the Raman bands due to thermal treatment leads us to suppose that crystallization continues during the vacuum thermal treatment and proceeds further at higher thermal treatment temperatures . The vibrational bands corresponding to Sb2O3 were not detected by Raman spectroscopy in any of the studied glass/ITO/TiO2/Sb2S3 samples.
X-ray diffraction (XRD) provides qualitative information on the phase composition and crystal structure. XRD patterns of reference glass/ITO/TiO2 samples and samples containing XRD-amorphous Sb2S3 (3-200-As-dep., 3-210-As-dep., Figure 2A; 6-200-As-dep., Figure 2B) show only diffraction peaks corresponding to cubic In2O3 (2θ = 21.3°, 30.4°, 35.3°, 37.4°, 41.4°, 45.3°, ICDD PDF 03-065-3170) and anatase-TiO2 (25.3°, 48.2°, ICDD PDF 00-016-0617). The diffraction peaks of orthorhombic Sb2S3 (ICDD PDF 01-075-4012), space group Pnma (D2h16) [20,24,25], appear in XRD patterns of lustrous gray as-deposited and thermally treated Sb2S3 samples (3-200-170, 3-210-170, 3-210-200, 3-220-As-dep., 3-220-170, 3-220-200, Figure 2A; 6-200-170, 6-210-As-dep., 6-210-170, 6-210-200, 6-220-As-dep., 6-220-170, 6-220-200, Figure 2B). The 2θ angles of observed Sb2S3 diffraction peaks and corresponding crystal plane indices are presented in Supporting Information File 1, Table S4. Experimentally determined mean lattice constants a, b and c of Sb2S3 are 11.25 ± 0.07 Å, 3.810 ± 0.025 Å and 11.16 ± 0.07 Å, respectively. Our experimentally determined mean unit cell volume (479 ± 4 Å3) lies between the experimentally determined volume (486.7 Å3) and the theoretically determined volume (470.5 Å3) calculated from orthorhombic Sb2S3 powder (>99.99 wt %) data presented by Ibáñez et al. .
Sb2S3 layers deposited from Sb/S 1:6 solution at 210 °C (6-210-As-dep., Figure 2B) are polycrystalline, whereas layers deposited from Sb/S 1:3 solution (3-210-As-dep., Figure 2A) are XRD-amorphous. Sb2S3 layers deposited at 220 °C from both Sb/S 1:3 (3-220-As-dep., Figure 2A) and 1:6 (6-220-As-dep., Figure 2B) solution are polycrystalline. Several diffraction peaks corresponding to orthorhombic Sb2S3 were detected in these samples. No additional phases were detected by XRD in any studied samples. The presence or absence of amorphous Sb2O3 as a minor phase in the Sb2S3 layers, as it is difficult to ascertain by Raman or XRD analyses, has not been conclusively demonstrated.
The diffraction peak of the (2 0 0)/(0 0 2) plane of Sb2S3 is absent in most samples deposited from Sb/S 1:6 solution. Conversely, the diffraction peak of the (1 0 1) plane of Sb2S3 is absent in all samples deposited from Sb/S 1:3 solution. Sb2S3 crystallites in most of our samples have no preferred orientation. Only crystallites in as-deposited and vacuum treated (170 °C) samples deposited from Sb/S 1:6 solution (6-220-As-dep., 6-220-170, Figure 2B) show a preferred orientation parallel to the substrate surface along the (0 2 0) plane normal of Sb2S3. Interestingly, this preferred orientation of crystallites does not extend to the sample with Sb2S3 deposited in the same conditions, but thermally treated in vacuum at 200 °C (6-220-200, Figure 2B).
The larger crystallite size is a boon to the power conversion efficiency of all solar absorber materials because decreasing the amount of grain boundaries likely increases charge carrier mobility . The crystallite sizes of as-deposited and thermally treated Sb2S3 layers are presented in Table 2. The effect of the deposition temperature is observed in Sb/S 1:3 Sb2S3 layers, as the crystallite size increases after vacuum annealing at 170 °C from 19 ± 8 nm to 100 ± 23 nm by raising TD from 200 to 220 °C. The crystallite size in Sb/S 1:6 Sb2S3 layers (42 ± 15 nm) does not change significantly with TD or vacuum treatment. Furthermore, vacuum treatment at 200 °C vs 170 °C does not substantially affect the crystallite size of Sb2S3 layers.
Table 2: Crystallite size (D) of as-deposited and vacuum treated Sb2S3 thin films. The crystallite size was calculated by the Scherrer equation from the (2 0 2) diffraction peak of as-deposited and vacuum treated (170 °C, 200 °C, 5 minutes) Sb2S3 thin films deposited on glass/ITO/TiO2 substrates from Sb/S 1:3 and 1:6 precursor solution at TD = 200, 210, 220 °C.
|Sb/S in sol.||1:3||1:6|
|as-dep.||amorph.||amorph.||33 ± 10||amorph.||39 ± 4||47 ± 1|
|vac. 170 °C||19 ± 8||38 ± 6||100 ± 23||37 ± 8||35 ± 4||49 ± 3|
|vac. 200 °C||no layera||32 ± 8||67 ± 12||no layera||45 ± 6||52 ± 3|
aNo Sb2S3 was detected by XRD or Raman.
In comparison, the largest crystallites in Sb2S3 layers grown on TiO2 substrates via CBD and annealed at 270 °C in N2 for 30 min oriented along the (2 0 0) plane parallel to the substrate were 74 nm in size . The crystallites oriented along the (2 0 1) plane were 24 nm in size in Sb2S3 layers grown on SnO2/F (FTO) coated glass substrates via thermal evaporation . The crystallite size was 52 nm along the (3 0 1) plane in Sb2S3 layers grown on glass substrates at 250 °C via spray pyrolysis , similar to the crystallite size in some of our samples. We conclude that the mean crystallite size in our Sb2S3 layers is in the general range of values obtained in the literature using both chemical and physical methods.
The aim of this study was to obtain uniform Sb2S3 layers, which continuously coat the TiO2 substrate. According to SEM surface studies, layers deposited from both Sb/S 1:3 and Sb/S 1:6 solutions at 200 and 210 °C (3-200-As-dep., 3-210-As-dep., Figure 3G,H, Supporting Information File 1, Figure S2A,B, Figure S3A,B; 6-200-As-dep., Figure 3A,B; 6-210-As-dep., Figure 3C,D) cover the substrate almost entirely. Grain boundaries and larger clusters of grains have formed in layers deposited from Sb/S 1:6 solutions for 40 minutes at 210 °C (6-210-As-dep., Figure 3C,D, Figure S5C,D). Cap-shaped islands (Ø 70 nm) in Sb2S3 layers deposited from Sb/S 1:6 solution at TD = 210 °C for 20 minutes (Figure S4A,B), have grown (Ø 100 nm) and coalesced further after 40 minutes of deposition at 200–210 °C (6-200-As-dep., Figure 3A,B, Figure S5A,B; 6-210-As-dep., Figure 3C,D, Figure S5C,D, Figure S6A,B), thereby covering the TiO2 substrate to a greater extent. The layers deposited from Sb/S 1:6 solution at 220 °C for 40 minutes (6-220-As-dep., Figure 3E,F, Figure S5E,F) consist of various agglomerates, separated by pinholes, and grains flowing randomly along the partially exposed TiO2 substrate (lower left, Figure 3E).
Increasing the deposition temperature from 210 to 220 °C significantly transforms the surface morphology in Sb/S 1:3 layers, as instead of the planar grains (3-210-As-dep., Figure 3G,H) domains of elongated rod-shaped grains (length ≈ 100 nm) appear either upright or sideways on the substrate (3-220-As-dep., Figure 3I,J, Figure S3C,D). Rod-shaped Sb2S3 grains were able to grow due to the nature of the material as well as due to complex interactions between the substrate and the turbulence of the spray during deposition .
Increasing the sulfur precursor concentration in the spray solution from Sb/S 1:3 to 1:6 (and deposition time from 20 to 40 minutes) yields Sb2S3 layers consisting of agglomerated grains (6-220-As-dep., Figure 3E,F). As the deposition time was simultaneously increased from 20 to 40 minutes, it is uncertain whether the morphology of the Sb2S3 layers is affected more by the Sb/S molar ratio in solution or by the deposition time. Sb2S3 tends to yield different morphologies in similar deposition conditions, possibly due to liquid phase reactions between molten-boiling SbCl3 (mp 73.4 °C, bp 223.5 °C ) and molten thiourea (TU, mp 182 °C [14,15]) catalyzed by the highly active surface of the TiO2 substrate .
We have consistently observed twice slower growth of Sb2S3 by spraying solutions with Sb/S 1:6 (Supporting Information File 1, Figure S4A,B) vs Sb/S 1:3 (Figure 3G,H) molar ratio at 200–220 °C. We speculate that doubling the concentration of TU could sterically inhibit the formation of solid Sb2S3 nuclei on the surface of the active TiO2 substrate due to more intense bubbling of volatile TU decomposition products (CS2, NH3, HCN, COS, SO2, HCl, HNCS at 200–220 °C in air based on decomposition studies of pure TU , Cu(TU)3Cl , Zn(TU)2Cl2 , and Sn(TU)2Cl2 ) in the surrounding liquid phase.
In summary, the most uniform and continuous Sb2S3 thin films were deposited from Sb/S 1:3 solution at 200–210 °C.
The thermal treatment of X-ray amorphous Sb2S3 layers (6-200-As-dep., Figure 3A,B; 3-200-As-dep.; 3-210-As-dep., Figure 3G,H, Supporting Information File 1, Figure S2A,B) in vacuum at 170 °C for 5 minutes yields enhanced substrate coverage at the expense of decreased layer thickness due to coalescence of grains and film formation (6-200-170, Figure 4A,B; 3-200-170, Figure 4G,H; 3-210-170, Figure 4I,J). Complete substrate coverage is observed in the Sb2S3 layers deposited at 210 °C from Sb/S 1:3 solution as coalescence is facilitated during treatment in vacuum at 170 °C due to the near-continuous coverage of the TiO2 substrate in the as-deposited layers (3-210-170, Figure 4G,H, Figure S2C,D, Figure S7A,B).
Planar grain agglomerates in thermally treated Sb2S3 layers (3-210-170, Figure 4G,H, Supporting Information File 1, Figure S7A,B; 6-200-170, Figure 4A,B, Figure S9A,B; 6-210-170, Figure 4C,D, Figure S9C,D) range from 100 nm to over 10 µm in size. These agglomerates, consisting of smaller grains separated by ridges, resemble the surface morphology of 300 nm thick polycrystalline Sb2S3 films grown via thermal evaporation and annealed for 10 min at 300 °C in N2 , and that of metal halide perovskites obtained by Volmer–Weber growth via hot casting . The layers deposited at 220 °C from both Sb/S 1:3 and Sb/S 1:6 solutions, and thermally treated at 170 °C, consist of numerous grains and pinholes (3-220-170, Figure 4I,J; 6-220-170, Figure 4E,F).
Sb2S3 layers deposited at 210 °C from both Sb/S 1:3 and Sb/S 1:6 solutions, and thermally treated in vacuum at 200 °C (3-210-200, Figure 5A,B, Supporting Information File 1, Figure S8A,C,E; 6-210-200, Figure 5C,D, Figure S8B,D,F), are porous, inhomogeneous and ≈20 nm thinner (Table 3) vs the uniform in thickness layers after treatment at 170 °C (3-210-170, Figure 4I,J; 6-210-170, Figure 4C,D).
Table 3: Thicknesses of Sb2S3 layers estimated from SEM images.
|Sb2S3 layer thickness, nm|
|Sb/S 1:3 in sol., 20 min dep.||Sb/S 1:6 in sol., 40 min dep.|
|vac., 170 °C||70–90||70–90||80/150a||30–40||60/400a||40/400a|
|vac., 200 °C||no layerb||60–70||N/A||no layerb||60–70||N/A|
aThickness of formations shown in the Supporting Information File 1 in Figures S5, S7, S8 and S9. bNo Sb2S3 was detected by XRD or Raman.
The decreasing layer thickness indicates that approximately a quarter of Sb2S3 by volume has either evaporated or sublimated, i.e., volatilized. Incongruent evaporation, i.e., depletion of sulfur in Sb2S3 during evaporation, may cause the change in Sb2S3 layer morphology, as volatilization of the planar regions around the nucleating islands has been reported during thermal treatment of both Sb2Se3 layers grown via thermal evaporation  and oxide containing Sb2S3 layers grown via CBD .
The calculated vapor pressure of Sb2S3 is ≈2 × 10−10 Torr at 170 °C, 7 × 10−9 Torr at 200 °C and 9 × 10−7 Torr at 250 °C , whereas the dynamic system pressure is ≈4 × 10−6 Torr. The calculated partial pressure of Sb2S3 is ≈0.0050% at 170 °C, 0.18% at 200 °C and 23% at 250 °C (Comment S1 in Supporting Information File 1). The loss of a quarter of the Sb2S3 layer thickness in samples that were vacuum annealed at 200 vs 170 °C (Table 3) correlates with the exponential increase in Sb2S3 vapor pressure in the 170–250 °C range.
In conclusion, the most uniform and continuous Sb2S3 thin films were produced by vacuum treatment at 170 °C for 5 min of Sb2S3 layers deposited from Sb/S 1:3 solution at 200–210 °C.
The elemental composition of Sb2S3 in as-deposited and thermally treated glass/ITO/TiO2/Sb2S3 samples was determined using energy dispersive X-ray spectroscopy (EDX). The EDX results of studied Sb2S3 layers in terms of S to Sb atomic ratio (S/Sb) are presented in Table 4. S/Sb in both as-deposited and vacuum annealed polycrystalline Sb2S3 layers deposited at TD = 220 °C is close to the stoichiometric value of 1.5 of Sb2S3, whereas the S/Sb ratio of as-deposited and thermally treated Sb2S3 layers (Sb/S 1:3 in solution, TD 200–210 °C, 3-200-As-dep., 3-210-As-dep., 3-200-170, 3-210-170) is ≈1.3. S/Sb is ≈1.5–1.6 in layers deposited from Sb/S 1:6 solution at 200–220 °C.
Table 4: S/Sb atomic ratio of as-deposited and thermally treated Sb2S3 layers calculated from EDX data.
|S/Sb in layer|
|Sb/S in sol.||1:3||1:6|
|vac., 170 °C||1.3||1.3||1.5||1.6||1.6||1.5|
|vac., 200 °C||N/A||1.4||N/A||N/A||1.5||N/A|
We note that interpretation of EDX spectra of very thin layers is difficult. Most of our Sb2S3 layers are thinner than 100 nm, which could explain the divergence in the elemental composition of our Sb2S3 layers. Therefore, future studies by more surface sensitive methods are required. Overall, S/Sb in most studied samples approximates the stoichiometric value of 1.5 of Sb2S3.
Oxygen could not be quantified by EDX due to the thin layers and high concentration of O in the glass/ITO/TiO2 substrate. In addition, C and Cl levels were below the detection limit of the used EDX setup in all studied Sb2S3 layers, meaning most C and Cl species exit the growing Sb2S3 layer during deposition in open environment (Supporting Information File 1, Figure S11). We believe that this reinforces our claim that formation of Sb2S3 proceeds through a molten phase reaction between SbCl3 and TU, where the denser (4562 kg/m3 ) Sb2S3 precipitates and nucleates, while the remainder of the volatile compounds (SbCl3, and various decomposition products of TU) exit the system [14,15,38,40].
The three most common growth mechanisms of solids can be described by the following equations :
Where σSG is the surface free energy of the substrate–gas interface (TiO2–air), σLG is the surface free energy of the layer–gas interface (Sb2S3–air) and σSL is the surface free energy of the substrate–layer interface (TiO2–Sb2S3). The surface free energy (σ) is the driving force of fluids and solids to seek a condition of minimum energy by contracting interfacial surface area . Separate 3D islands grow if Equation 1 is valid, a.k.a. Volmer–Weber growth; 2D layer-by-layer growth occurs if Equation 2 is valid, a.k.a. Frank–Van der Merwe growth; combined 2D layer-by layer and 3D island growth occurs if Equation 3 is valid, a.k.a. Stranski–Krastanov growth [36,41-43].
Furthermore, SEM surface studies show cap-shaped islands indicative of Volmer–Weber growth in Sb2S3 layers deposited on Si/SiO2 alternative substrates by ultrasonic spraying (Supporting Information File 1, Figure S10A,B). Metastibnite-Sb2S3 forms when formation of stibnite-Sb2S3 is halted by insufficient reaction time and energy [44-46]. Volmer–Weber island growth of amorphous Sb2S3 (and in some cases leaf-like grains of polycrystalline Sb2S3) have been observed in Sb2S3 layers grown by chemical bath deposition on glass [47,48], In2O3/Sn (ITO) , planar TiO2  and TiO2 nanotube arrays , by sequential deposition  and spin coating [8,52] on planar TiO2, by photochemical deposition on mesoporous TiO2 , by thermal evaporation on planar CdS  and planar TiO2 . Supported by these numerous observations, we consider the Volmer–Weber growth characteristic of Sb2S3, given that the substrate and deposition conditions are met. Indeed, metastibnite, the naturally occurring mineral form of amorphous Sb2S3, has the botryoidal characteristic, preferentially forming globular clusters . We have also observed 3D growth of extremely thin TiO2 layers by spray pyrolysis . Therefore, 3D island growth may partially be imposed by the use of the spray pyrolysis method as well.
Based on the above observations, the morphology and crystallinity of as-deposited layers seems to determine the nature of Sb2S3 layer morphology as formed during vacuum thermal treatment. Our proposed growth mechanism of Sb2S3 by ultrasonic spraying in air is illustrated in Figure 6.
The absorption coefficient (α) and band gap (Eg) values of Sb2S3 in both as-deposited and thermally treated glass/ITO/TiO2/Sb2S3 samples were determined using an approximated Sb2S3 layer thickness of 100 nm derived from SEM images (Table 3). The absorption coefficient α was determined as
where d is the layer thickness, R is the total reflectance, included to compensate for thin film interference, and T is the total transmittance.
The band gap of Sb2S3 layers was determined by plotting (αhν)1/r vs hν, where h is the Planck constant, ν is the frequency and r = 1/2 is the exponent corresponding to the assumed direct optical transition . Extrapolating the linear region of this curve to the hν-axis yields the optical band gap. Thin film interference could not be completely removed by accounting for reflectance in α calculations. Thus, the absolute values of α may deviate from the expected values with the uncertainty introduced by using a constant layer thickness in calculations.
The α vs wavelength plots of samples, which contain as-deposited or vacuum-treated Sb2S3 layers deposited from Sb/S 1:3 solution, are shown in Figure 7A. Likewise, α vs wavelength plots of Sb/S 1:6 samples are shown in Figure 7B. The α in samples containing amorphous Sb2S3 increases steadily from 103–104 cm−1 at 600–800 nm to 105 cm−1 at around 400 nm. The α increases significantly faster in samples containing as-grown crystalline Sb2S3 or vacuum crystallized Sb2S3. The value of α surges by an order of magnitude from around 104 cm−1 to 105 cm−1 as the wavelength decreases from 750 nm to 650 nm due to the onset of absorption in crystalline Sb2S3. At shorter wavelengths beyond the absorption edge, α increases at a slower rate, from around 105 cm−1 at 650 nm to more than 5 × 105 cm−1 at 300 nm. The optical absorption results are in agreement with XRD, which shows that these samples (3-220-As-dep., 3-210-170, 6-210-As-dep. and 6-200-170) contain orthorhombic Sb2S3 (Figure 2A,B). Comparing the α spectra of samples containing amorphous and crystalline Sb2S3 further confirms that the Sb2S3 layers deposited from Sb/S 1:3 solution at 200–210 °C, and from Sb/S 1:6 solution at 200 °C, are indeed amorphous. Namely, α is an order of magnitude smaller at around 600 nm in samples containing amorphous Sb2S3 layers (3-210-As-dep. and 6-200-As-dep.).
The experimentally determined Eg are ≈2.7 and 1.8 eV for amorphous and polycrystalline Sb2S3, respectively (Table 5, Tauc plots in Supporting Information File 1, Figure S12). In comparison, Eg of amorphous CBD-Sb2S3 on glass substrates is ≈2.5 eV  and Eg of polycrystalline Sb2S3 prepared by physical and chemical methods is commonly reported as 1.6–1.8 eV [1,22,58-60]. As such, we find the Eg of our polycrystalline Sb2S3 layers lies satisfactorily in the range of published values.
Table 5: Band gap (Eg) of as-deposited and thermally treated Sb2S3 layers, as estimated assuming direct optical transition and Tauc plotsa of optical transmittance spectra of glass/ITO/TiO2/Sb2S3 samples.
|Sb/S in sol.||1:3||1:6|
|vac. 170 °C||1.8||1.8||1.8||1.8||1.8||1.8|
|vac. 200 °C||no layerb||1.8||1.8||no layerb||1.8||1.8|
aSupporting Information File 1, Figure S12A,B. bNo Sb2S3 was detected by XRD or Raman.
The structure, the morphology, and the optical properties of Sb2S3 layers could be controlled by varying the spray deposition temperature and the molar ratio of precursors in spray solution. Nonuniform, discontinuous layers of polycrystalline Sb2S3 (Eg 1.8 eV) were deposited by ultrasonic spray pyrolysis of SbCl3/SC(NH2)2 1:3 solution at TD ≥ 220 °C or 1:6 solution at TD ≥ 210 °C on glass/ITO/TiO2 substrates in air. Increasing the concentration of the sulfur precursor in spray solution from Sb/S 1:3 to 1:6 reduced the crystallization temperature of Sb2S3 layers by ≈10 °C. Uniform layers of amorphous Sb2S3 (Eg ≈ 2.7 eV, S/Sb 1:3) were deposited on glass/ITO/TiO2 substrates in air by ultrasonic spray pyrolysis of Sb/S 1:3 solution at TD = 200–210 °C. High quality, uniform, pinhole-free coatings of polycrystalline orthorhombic Sb2S3 (Eg 1.8 eV, S/Sb 1.3) with lateral grain size as large as 10 μm were produced by crystallization of amorphous Sb2S3 layers in vacuum at 170 °C for 5 minutes. Such Sb2S3 optical coatings are very attractive for future application as low-cost absorber layers in solar cells.
Commercial 1.1 mm thick soda-lime glass coated with 150 nm 25 Ω∙sq−1 tin doped indium oxide (ITO) from ZSW was used as a substrate. The substrates were rinsed with deionized water, methanol (99.9 vol %), deionized water, dipped in aqueous room temperature H2SO4 (1 vol %), rinsed again with deionized water, and dried at 105 °C in air.
TiO2 was prepared by methods used in our previous papers [7,12]. The TiO2 film thickness was ≈80 nm based on SEM images. The Sb2S3 layers were deposited from 30 mM SbCl3 (99 wt %) and SC(NH2)2 (99 wt %) methanolic (99.9 vol %) solutions at molar ratios of Sb/S 1:3 and Sb/S 1:6. All chemicals were purchased from Sigma-Aldrich and used without any additional processing. The precursor solutions were prepared inside a glovebox with controlled humidity (<14 ppm).
The solutions were ultrasonically nebulized and guided by compressed air at a flow rate of 5 L·min−1 onto glass/ITO/TiO2 substrates at deposition temperatures of 200, 210, and 220 °C for 20 min (Sb/S 1:3) or 40 min (Sb/S 1:6). After deposition, some of the samples were thermally treated in dynamic vacuum (<4 × 10−6 Torr) at 170, 200 or 250 °C for 5 min. The average heating and cooling rate was ≈8 °C·min−1.
The elemental composition of the films was determined by energy dispersive X-ray spectroscopy (EDX) using a Bruker spectrometer with ESPRIT 1.8 system at the Zeiss HR FESEM Ultra 55 scanning electron microscope (SEM) operating at an accelerating voltage of 7 kV. The surface and cross-sectional morphologies of the layers were recorded by the same SEM system at an electron beam accelerating voltage of 4 kV.
Unpolarized micro-Raman measurements were conducted at room temperature using a Horiba Jobin Yvon Labram HR 800 spectrometer in backscattering geometry. The laser intensity was attenuated to ca. 143 µW·µm−2 over a focal area of Ø 5 µm to prevent oxidation of the Sb2S3 layers, a common oversight according to Kharbish et al. . Deconvoluted band centers in Raman shift, band intensities and full widths at half maximum (FWHM) were fitted using a Lorentzian function .
X-ray diffraction (XRD) patterns were recorded on a Rigaku Ultima IV powder diffractometer in θ-2θ mode (Cu Kα1 λ = 1.5406 Å, 40 kV, 40 mA, step 0.02°, 5°/min, silicon strip detector D/teX Ultra). The crystal structure and phase composition were analyzed using Rigaku PDXL 2 software.
Optical total transmittance and total reflectance spectra of glass/ITO/TiO2 reference and glass/ITO/TiO2/Sb2S3 samples were measured in the 250–1600 nm range vs air as a reference using a Jasco V-670 UV-VIS-NIR spectrophotometer equipped with a 40 mm integrating sphere and Spectra Manager II software.
|Supporting Information File 1: Additional XRD, EDX data, SEM images, Lorentzian fitting of Sb2S3 Raman vibrational bands, and Tauc plots.|
|Format: PDF||Size: 3.0 MB||Download|
We acknowledge Dr. Valdek Mikli from the Laboratory of Optoelectronic Materials Physics at Tallinn University of Technology for recording SEM images and EDX measurements, Estonian Research Council project IUT19-4 “Thin films and nanomaterials by wet-chemical methods for next-generation photovoltaics” and European Regional Development Fund project TK141 “Advanced materials and high-technology devices for sustainable energetics, sensorics and nanoelectronics” for funding.