- Email: [email protected]
Antimony sulfide-selenide thin film solar cells produced from stibnite mineral
Accepted Manuscript Antimony sulfide-selenide thin film solar cells produced from stibnite mineral
P.K. Nair, Geovanni Vázquez García, Eira Anais Zamudio Medina, Laura Guerrero Martínez, Oscar Leyva Castrejón, Jacqueline Moctezuma Ortiz, M.T.S. Nair PII: DOI: Reference:
S0040-6090(17)30840-4 doi:10.1016/j.tsf.2017.11.004 TSF 36336
To appear in:
Thin Solid Films
Received date: Revised date: Accepted date:
4 May 2017 7 October 2017 3 November 2017
Please cite this article as: P.K. Nair, Geovanni Vázquez García, Eira Anais Zamudio Medina, Laura Guerrero Martínez, Oscar Leyva Castrejón, Jacqueline Moctezuma Ortiz, M.T.S. Nair , Antimony sulfide-selenide thin film solar cells produced from stibnite mineral. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Tsf(2017), doi:10.1016/j.tsf.2017.11.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Antimony sulfide-selenide thin film solar cells produced from stibnite mineral P. K. Nair, Geovanni Vázquez García, Eira Anais Zamudio Medina, Laura Guerrero Martínez, Oscar Leyva Castrejón, Jacqueline Moctezuma Ortiz, M. T. S Nair Instituto de Energías Renovables, Universidad Nacional Autónoma de México, Temixco-62580, Morelos, México. [email protected]
T
Abstract
IP
Stibnite (Sb2S3) is a major mineral of antimony, which occurs as large crystals often with
CR
minor encrustations of other minerals. Locally sourced powdered stibnite containing some quartzite (SiO2) and ferrosilite (FeSiO3) has been used in this work as evaporation source for vacuum deposition of thin film solar cells. The added minerals were left as residue in the crucible
US
and did not incorporate into the Sb2S3 thin film. To stibnite powder was added Sb2Se3 powder prepared in our laboratory as chemical precipitate. Source mixtures of different weight – by –
AN
weight (w/w) ratios gave thin films of chemical composition Sb2SxSe3-x and optical band gap (Eg) within 1.38 eV (Sb2Se3) – 1.88 eV (Sb2S3). Solar cells of SnO2:F/CdS (100 nm)/Sb2S3(250 nm)/C-Ag
M
prepared by using stibnite as evaporation source gave under standard conditions, open circuit voltage (Voc), 0.668 V; short circuit current density (Jsc), 6.95 mA/cm2; and conversion efficiency
ED
(η), 1.62 %. For a 2:1 (w/w) mixture of stibnite:Sb2Se3, solar cell of Sb2S2.14Se0.86 (Eg 1.61 eV) was obtained with Voc, 0.562 V; Jsc, 13.53 mA/cm2 and η, 4.03 %. For 1:5 (w/w) mixture, solar cell of
PT
Sb2S0.5Se2.5 (Eg 1.44 eV) gave Voc, 0.443 V; Jsc, 22.31 mA/cm2 and η, 4.24 %. Electrical conductivity of the Sb2SxSe3-x absorber films in the dark increased from 10-8 to 10-6 Ω-1 cm-1 and their
CE
photoconductivity, from 10-6 to 10-5 Ω-1 cm-1 as the composition changed from Sb2S3 to Sb2SxSe3-x and Sb2Se3. Direct use of an abundant mineral as the evaporation source in thin film solar cell
AC
technology is a novelty.
Key words: antimony chalcogenide, antimony sulfide selenide, Sb-S-Se, Sb2SxSe3-x; antimony chalcogenide solar cells, thin film solar cells; solar energy, renewable energy; semiconductor thin films Highlights: Direct use of stibnite ore-mineral of Sb2S3 to produce solar cells of 1.6 % efficiency Stibnite mineral and Sb2Se3 precipitate source-mix used for solar cells of 4.2 % efficiency Variable electrical conductivity and Eg 1.38 – 1.61 eV in Sb2SxSe3-x from stibnite – Sb2Se3
1
ACCEPTED MANUSCRIPT
1. Introduction Production of solar cells directly from an abundant mineral is a very attractive proposition. In this work we consider stibnite mineral with chemical composition Sb2S3 as a candidate. Stibnite is the predominant ore mineral of antimony, distributed over the earth’s crust at 0.2-0.5 parts per million (ppm) [1]. Global production of antimony (metal content) in the year 2015 was nearly
T
150,000 metric tons. Antimony metal price was close to US$8 per kg (commodity index) during
IP
2015, and shows a declining-price trend over time [1]. Stibnite occurs as large crystals and crystal
CR
clusters weighing over many kg in many parts of the world. While these crystals are generally chemically pure (99% Sb2S3) often some encrustations of calcite, quartzite, etc., are found on them
US
[2]. In a locally sourced stibnite powder used in this work to produce thin film solar cells, quartzite and ferrosilite were evidenced, but these were left as residue in thermal evaporation.
AN
Antimonselite (Sb2Se3) as well as its partially S-substituted mineral Sb2S0.423Se2.577 [3] are also among antimony minerals [1, 2].
M
Both Sb2S3 and Sb2Se3 are of orthorhombic crystalline structure; and they possess direct optical band gap (Eg) of 1.88 and 1.1 eV, respectively [4]. Thin films of solid solutions of antimony
ED
sulfide-selenide (Sb2SxSe3-x) with Eg continuously adjustable within 1.1 – 1.88 eV have been prepared through heating chemically deposited Se-Sb2S3 layers [5]. Such materials would satisfy
PT
many of the criteria prescribed in a 2014 work for thin film solar cell materials with respect to material properties, conversion efficiency (η), stability, and price for commercial photovoltaic
CE
module manufacturing [6]. Photovoltaic absorber materials with Eg 1.3 – 1.7 eV of Sb2SxSe3-x solid solutions fulfil criteria set for thin film solar cell materials so as to combine large-enough open
AC
circuit voltage (Voc) and fill factor (FF) of solar cells with appreciable short circuit current density (Jsc) to achieve high η [7]. These guidelines led to a growing interest in antimony sulfide-selenide solar cells in recent years [8]. Among the best conversion efficiency of antimony sulfide-selenide solar cells is 5.79 %, reported in 2016 [9]. This solar cell used tin-doped In2O3 (ITO):
ITO/CdS/Sb2SxSe3-x/Au. The
absorber film was made by rapid thermal evaporation followed by reaction in sulfur-vapor. Following parameters were reported for it: Voc, 0.50 V; Jsc, 22.33 mA/cm2 and FF, 0.521. The best η to date for a Sb2SxS3-x solar cell appears to be 6.3 % reported in 2014 [10]. Compact and mesoporous (mp) layers of TiO2 on fluorine doped SnO2 (FTO) were used along with hole-transport 2
ACCEPTED MANUSCRIPT organic media (HTM/HTL): FTO/bulk-TiO2/mp-TiO2/Sb2SxSe3-x/HTM/HTL/Au. The absorber layer was prepared by spin coating (Sb2Se3) and chemical deposition (Sb2S3) followed by heating. This solar cell had Voc, 0.475 V; Jsc, 24.9 mA/cm2 and FF, 0.556. Solar cells of Sb2Se3 absorbers consistently lagged in Voc, while η and Jsc were comparable. In 2015 [11] η of 5.6 % was reported for a CdS/Sb2Se3 solar cell with Voc, 0.40 V; Jsc, 25.1 mA/cm2, and FF, 0.557. An η of 4.8 % was reported in ITO/CdS/Sb2Se3/Au with Voc, 0.36 V; Jsc , 25.3 mA/cm2; and FF , 0.52 [12]. Use of Sb2S3
T
alone as the absorber film elevates Voc to 0.6 V, but it also lowers η toward 1%, as reported in 2014
IP
[13] in FTO/CdS/Sb2S3/C-Ag, with Voc, 0.60 V; Jsc, 6.12 mA/cm2; FF, 0.35 and η, 1.27%. By steadily
CR
decreasing x in Sb2SxSe3-x solar cells prepared by thermal evaporation of Sb-S-Se chemical precipitate, η was improved but Voc reduced, as reported in 2016 for FTO/CdS/Sb2S1.7Se1.3/C-Ag with Voc, 0.545 V; Jsc, 13.5 mA/cm2; FF, 0.43; η, 3.12% and for Sb2S1.5Se1.5 with Voc, 0.528 V; Jsc, 15.7
US
mA /cm2; FF, 0.44; and η, 3.63 % [14].
AN
In this work we used a mixture of locally sourced powdered stibnite mineral and Sb2Se3 powder prepared as reported in [14], as thermal evaporation source to produce thin film solar
M
cells of Sb2SxSe3-x absorbers. Conversion efficiency of up to 4.24% has been reached in these solar cells. The basic properties of the stibnite mineral and of the chemical precipitate are presented.
ED
Characteristics of thin film Sb2SxSe3-x and of the solar cells show that direct use of stibnite mineral for solar cells is feasible. This option is attractive because the additive minerals which are often
PT
present in stibnite powder do not seem to affect the quality of the thin films or of solar cells. This approach eliminates the laborious and energy-intensive conversion of hand-sorted stibnite crystals
CE
first into powdered stibnite mineral, then into Sb2O3 or Sb-metal [2, 15] before reconverting it to laboratory reagent Sb2S3.
AC
2. Experimental details
Substrates- Glass slides (Corning) measuring 7.5 mm x 2.5 mm of 1 mm thickness or fluorinedoped SnO2 (FTO) of sheet resistance 15 Ω (TEC 15, NSG Pilkington) cut to the same size. CdS thin film deposition: Chemical deposition method was reported in 1994 [16] and mentioned in [13, 14]. The solution mixture for the deposition contained cadmium nitrate, sodium citrate, ammonium hydroxide, thiourea (Aldrich reagents) and distilled water. Up to six clean substrates were immersed in 80 ml of the deposition bath at 80 oC. At 60 min, a specularly reflective CdS thin
3
ACCEPTED MANUSCRIPT film of 80-100 nm in thickness was deposited. This film is of hexagonal crystalline structure with band gap 2.5 eV; with a photoconductivity under illumination of 0.01-0.1 Ω-1 cm-1 [16]. Sb2Se3 chemical precipitate: The method to prepare thin film or precipitate of Sb2SxSe3-x in the laboratory was described in [14]. A solution containing potassium antimony tartrate, triethanolamine, ammonia (aq.), silicotungstic acid, thioacetamide and/or selenosulfate solution,
T
and distilled water was maintained at 80 oC for 4-5 h. It was then cooled down overnight to room
CR
reagents of 98-99% assay were used for this chemical precipitation.
IP
temperature; filtered, rinsed with deionized water, dried and stored. Aldrich or Baker Analyzed
Stibnite (Sb2S3) powder: Was obtained from a local supplier for industrial grade chemicals. The
US
assay is normally stated at 95-98 % purity (Reasol Reactivos Analiticos, Mexico). For comparison of results, we also used laboratory grade reagent of Sb2S3 of 99% assay (Aldrich).
AN
Vacuum thermal evaporation of stibnite-Sb2Se3 powder mix: Thin films and solar cells were prepared from 600 mg of the evaporation source of five different stibnite:Sb2Se3 weight-by-weight
M
ratio (w/w). The films and solar cells would be labelled accordingly: A (600:0); B (400:200); C (200:400); D (100:500); E (0:600). For solar cells labelled F and G, laboratory grade Sb2S3:Sb2Se3
ED
(300:300) and (400:200) were used. The powder source mix was painted on tungsten crucible by using a few drops of propylene glycol and then dried at 80 oC. A Torr International diffusion pump
PT
system was used for thermal evaporation of this source to produce thin films. The chamber reached a base pressure of 2x10-6 Torr (2.7x10-4 Pa). Up to six substrates (2 glass slides, 4 FTO/CdS
CE
substrates) were mounted on a rotating sample holder located 40 cm from the source. The temperature of the substrate was kept at 400-450 oC using four linear tungsten-halogen lamps
AC
mounted at the back of the sample holder. The temperature control was achieved through intermittent lighting of these lamps linked to a temperature controller circuit. The source mixture starts melting in the crucible at a temperature close to the melting point of Sb2S3 (550 oC) and of Sb2Se3 (610 oC). Temperature of the crucible during the thermal evaporation of the source mixture was nearly 850 oC. The duration of thermal evaporation was 15 min. The film thickness showed variation depending on the composition of the source mixture. We noted that when the source contained only stibnite mineral (sample A), the substrate temperature had to be lowered to 400 oC in order to avoid sublimation loss of the thin film from the substrate. This was the case also with laboratory reagent Sb2S3, as reported previously [13]. Addition of even small quantity of Sb2Se3 to 4
ACCEPTED MANUSCRIPT the source mixture seemed to inhibit such loss. The film thicknesses were: A, 232 nm; B, 306 nm; C, 235 nm; D, 370 nm; E, 320 nm; F, 250 nm; G, 300 nm. No effort was made at this stage of work to secure
exactly the same thickness of films in all cases. Solar cell structures: Figure 1 a) shows the cross-sectional schematics of FTO/CdS/Sb2SxSe3-x/C-Ag solar cells; and 1 b) illustrates the development of such cells through the following steps: CdS thin film (100 nm) was chemically deposited on FTO cut to 25 mm x 75 mm pieces;
(ii)
Deposition via vacuum thermal evaporation of a Sb2SxSe3-x thin film (230-370 nm) was
(iii)
CR
done from stibnite-Sb2Se3 sources (total mass, 600 mg);
IP
T
(i)
Hand-painted colloidal graphite electrodes (SPI Chem) of 0.8-1 cm2 in area were
US
applied and the cells were heated at 300 oC for 30 min in nitrogen at a pressure 20 Torr (2.7 kPa); (iv)
Cell periphery was etched out; colloidal silver-paint (DuPont) was applied on the
AN
graphite-paint and multi-strand copper wire was attached on it for measurements. (The streak of light shows the specular reflection from a tubular fluorescent ceiling
M
lamp on the electrode-free area. This part of the layer-structure was used for grazing-
ED
incidence X-ray diffraction (GIXRD) and optical measurements.) Measurements: Film thickness (d) was measured using Alpha-Step 100. Optical transmittance (T)
PT
against air and near-normal (5o incidence) specular reflectance (R) against standard front mirror were measured for film-side incidence of glass/Sb2SxSe3-x films. Shimadzu 1800 UV
CE
spectrophotometer was used. Optical absorption coefficient (α) of the films was estimated using the basic equation considering multiple reflection [17]: T = [(1 – R)2e-αd]/[1 – R2e-2αd]. X-ray
AC
diffraction (XRD) measurement on the source-powder and grazing incidence – GIXRD (incidence, 0.5o) measurement on Sb2SxSe3-x film (glass/Sb2SxSe3-x) and on FTO/CdS/Sb2SxSe3-x were done on a Rigaku-D-MAX 2000 diffractometer with Cu-Kα (1.5406 Å) radiation. The value of x for the solid solution Sb2SxSe3-x was determined by assuming a proportional variation of the inter-planar distance d(hkl) in the orthorhombic cell of the sample between that of stibnite-Sb2S3 (powder diffraction file, PDF 42-1393) and antimonselite-Sb2Se3 (PDF 15-0861) orthorhombic cells. Electrical conductivity of the films in the dark and under illumination (850 W/m 2 tungstenhalogen light) were determined by applying a bias of 10 V and recording the current during 10 s under the dark, 10 s under illumination, and 10 s in the dark. The J-V characteristics of the solar 5
ACCEPTED MANUSCRIPT cells and external quantum efficiency (EQE) of the solar cells were obtained using a SCIENCETECH (Canada) Photovoltaic Testing System (PTS) under simulated solar radiation of air-mass 1.5 Global spectrum of intensity 1000 W m–2 from a xenon arc-lamp and with a monochromator calibrated using a standard Si – photodiode. Prior to these measurements all solar cells were stabilized under the sun (870 – 980 W/m2) for 6 h. The solar cell parameters were also confirmed from measurements made directly under the sun and from J – V characteristics recorded in a locally
T
fabricated set-up. The stability of the solar cell was ascertained during repeated measurements on
IP
the cells during a period of over six months.
CR
3. Results and discussion
US
3.1 Crystalline structure
In Figure 2 (bottom panel) is seen that the reagent grade Sb2S3 shows an XRD pattern
AN
perfectly matching that of Sb2S3 – stibnite, but that of the locally sourced stibnite mineral shows in addition the major peaks of two added minerals, (101) of quartz (SiO2) and (610) of ferrosilite (FeSiO3). Both these minerals are stable at high temperatures, and are expected to be left as
M
residue in the crucible at temperature of evaporation of Sb2S3 of nearly 850 oC. The XRD pattern of
ED
the as-prepared precipitate of Sb2Se3 in the top panel does not show well-defined peaks, but upon heating it at 300 oC the pattern matches that of Sb2Se3 – antimonselite.
PT
Figure 3 shows that the GIXRD pattern of thin film A produced from stibnite mineral powder matches in peak positions (2θ) with those of the standard pattern for Sb2S3, but the
CE
relative intensities vary significantly from the standard, suggesting preferential orientation of crystalline planes on heated (400 oC) glass substrates. However, the film deposited for the cell
AC
structures A (as well as for B-G cells) is on a CdS substrate layer. In this case the GIXRD pattern matches the standard stibnite pattern in 2θ and relative intensity. Additive minerals present in the powdered stibnite mineral (Figure 2) do not lead to added crystalline matter in the deposited thin film. Variation in relative intensity is a behavior repeated in the case of thin film E deposited on glass substrate (Figure 3 top panel). The thin film constituting the solar cell E matches in 2θ and relative intensity with those of antimonselite – Sb2Se3. In solar cell C, produced from a source-mix of stibnite:Sb2Se3 (200:400), the peak positions and relative intensities are intermediate to those of Sb2S3 and Sb2Se3, as indicated by the dotted lines. The relative position of the peaks along 2θ helps estimate the composition of antimony sulfide selenide solid solution in film/Cell C as 6
ACCEPTED MANUSCRIPT Sb2S1.35Se1.65. The method assumes proportional variation of inter-planar distance for any (hkl) crystallographic planes as the composition changes in Sb2SxSe3-x from Sb2S3 (x = 3) to Sb2Se3 (x = 0), as discussed in ref [5] and illustrated in ref [14]. In a same way the composition of films B, D, F and G were determined as: B, Sb2S2.14Se0.86; D, Sb2S0.5Se2.5, F: Sb2S1.5Se1.5 and G, Sb2S2.3Se0.7. Estimate for crystallite diameter (D) for the materials was obtained from non-overlapping peaks (120) by the XRD software through instrumental-corrected full-width (Δ2θ) at half maximum (FWHM) of
IP
T
peak intensity. The values were in the 16-19 nm interval.
CR
3.2 Optical and electrical characteristics
Optical band gap Eg of absorber materials decides the maximum light-generated current
US
density (JL), which also is the maximum value for Jsc in solar cells. The value of α, and the nature of electronic transitions (direct or indirect) decide the thickness of film (3/α) or (4/α) required to achieve 95 % or 98 % of JL [18]. In the present study a thickness interval (230-370 nm) of Sb2SxSe3-x
AN
films was used (given in the experimental section), which might have to be further optimized for solar cells in future work. Data on R and T on the films A-E in Figure 4 a) and b) help evaluate the
M
plots of α and (αhν)⅔ versus photon energy (hν) of Figure 4 c) and d). The optical interference in thin films is not entirely compensated in T and R spectra because R is measured at an angle of
ED
incidence 5o, and T at 0o (normal incidence), entailing larger optical path in the R measurement of the films. This difference gives rise to some waviness in the plots in 4 c) and d). Further, for highly
PT
absorbing thicker films, the transmittance value is not accurately measured for T < 0.01% due to
flatten-off.
CE
equipment limitation. Hence at high hν and very low T, α tend to be sub-estimated and are seen to
AC
For all the films, α sharply increases and surpasses 105 cm-1 upon the onset of optical absorption. Thus, for effective optical absorption, film thickness used here (3/α, of nearly 300 nm) would suffice. However, to inhibit electrical instability or to increase parallel resistance, higher film thicknesses may have to be sought in a later work. Analysis of Eg in Figure 4 d) for antimony sulfide (A), antimony selenide (E) and Sb2SxSe3-x (B – D) thin films tend to suggest ‘direct gaps with forbidden transitions’ [17] for them, in accordance with previous results on these materials [13, 14]. It is seen that the value of Eg may be continuously varied in the films through varying the stibnite:Sb2Se3 ratio. The high optical reflectance of 45 – 50 % of the films seen in Figure 4 b) originates from a high refractive index (n) of nearly 5, and thus entails high static relative
7
ACCEPTED MANUSCRIPT permittivity ( εrs > n2 , hence > 25). This leads to large exciton Bohr radii (rB) in these materials, bringing them comparable with the value of the crystallite grain diameter D (20 nm). Thus, the Eg becomes susceptible to increase due to quantum confinement of the excitons. The Eg of film E, which is Sb2Se3 in composition (Figure 2) is therefore 1.38 eV, which is greater than for the bulk crystalline value of 1.1 eV [4]. Due to this reason, the proportional variation of Eg for Sb2SxSe3-x films within the boundary values of 1.1 – 1.88 eV for Sb2Se3 – Sb2S3 is not strictly followed. What is
T
important is that the Eg sought for promising solar cell absorber materials of 1.4 – 1.6 eV [6, 7] is
IP
available in films B – D.
-8
-6
-1
-1
CR
Figure 5 shows the electrical conductivity (σ) in the dark (10 to 10 Ω cm ) and under
illumination (10-6 to 10-5 Ω-1 cm-1) of the thin films A – E. Both values increase steadily with
US
decrease in x of Sb2SxSe3-x of the thin films. In Sb2Se3 with Eg of 1.38 eV, photons of a wider spectrum is absorbed compared with that in Sb2S3 with 1.89 eV. This leads to higher
AN
photoconductivity in the film E, and hence it is also expected to result in higher Jsc in the solar cell E.
M
3.3 Solar cell characteristics
ED
Systematic variation in the optical and electrical properties of Sb2SxSe3-x thin films discussed above results in a decreasing trend in Voc from 0.668 V to 0.386 V as x drops from 3 to 0,
PT
seen in Figure 6 a). Under similar cell fabrication method, a decrease in Eg from 1.89 (A) to 1.38 eV (E) entails an increase in the intrinsic carrier concentration and hence in an increase in the reverse
CE
saturation current density (Jo). Hence [ln (JL/Jo)] decreases with decrease in x, and so too would Voc, which is related to [ln (JL/Jo)] through the thermal voltage (26 mV at 300 K). An increase in Jsc with
AC
decrease in x takes place, from 6.95 to 23.06 mA/cm2, in accordance with the increasing trend in photoconductivity of Figure 5. Table 1 provides a summary of the solar cell parameters of cells A – E, which illustrates that systemic trend in variation of parameters, with cell efficiency of 4 % seen accessible. It is pertinent to mention that the source and substrate temperatures for vacuum thermal evaporation of thin films and solar cells need to be optimized for each source-mix. In this feasibility study on Sb2SxSe3-x films produced from stibnite mineral, nearly the same condition of cell preparation was used in all cases. For Cell B, the source contained 66.7 % (w/w) (or 74 mol %) of stibnite mineral and produced thin films of Sb2S2.14Se0.86 with Eg, 1.61 eV. This cell gave Voc, 8
ACCEPTED MANUSCRIPT 0.562 V; Jsc, 13.53 mA/cm2 and η, 4.03 %. The maximum η of 4.24% was obtained for Cell D for which the use of stibnite mineral is only ⅙ (600) =16.67 % (w/w). The lack of charge carrier collection at the graphite electrode is suggested by the continued increase of the photo-generated current density toward reverse bias of – 0.4 V. Thus, better cell design would lead to increase in Jsc in all the cases – for example 16 mA/cm2 for Cell B, 25 mA/cm2 for Cell D (Eg, 1.44 eV), etc., might
T
be feasible, with consequent increases in cell efficiency.
IP
EQE plots in Figure 6 b) obtained under Jsc condition help identify the same limitation – a lack of collection of photo-generated charge carriers at the graphite electrode. A welcome
CR
disparity exists between the wavelength λg (nm) = 1240/Eg(eV) indicated on the plots for the onset for EQE, as expected from the optical absorption plot of Figure 4 c) and d), and the actual onset,
US
which occurs at still longer wavelengths. This could arise from “tail states,” which in this case may be due to some larger crystallites relatively exempt from quantum confinement, and hence with
AN
smaller bandgap, which are not evident in the optical absorption analyzed from T and R data. If some population of such larger crystallites exists in the absorber film, the optical absorption and
M
hence EQE onset would occur closer to 1130 nm (1240/1.1 eV (bulk)) and not at 900 nm (1240/1.38 eV of Fig. 4 d) for Cell E, as observed. The EQE onset at longer wavelength may be also
ED
due to compositional variation among the crystallites. More detailed study is required to discern these mechanisms. Evidenced in the EQE data in Figure 6 b) is also that the values drop drastically
PT
at wavelength shorter than 500 nm (2.48 eV). This is directly related to optical absorption in the non-participative part of the CdS window film, outside of the junction at the FTO/CdS interface.
CE
This behavior in EQE calls for a reduction in the thickness of CdS film from 80-100 nm or the addition of a more transparent-resistive thin film along with such reduction [6]. In the n+-CdS/p-
AC
Sb2SxSe3-x junction, the depletion width in the n+ - layer would be a few nm. However, the reduction in CdS film thickness toward such thickness might leave short-circuit paths which would cause reduction in Voc and instability of the solar cell during prolonged operation. For comparison, solar cells were made under the same conditions as of cells B – E, but using a laboratory reagent grade Sb2S3 without the additive minerals in the powdered stibnite (see Fig. 1 bottom panel). J-V characteristics and EQE of cell F and G prepared by using Sb2S3 (Lab. reagent):Sb2Se3 in 300:300 and 400:200 (w/w) ratio, respectively, are given in Figure 7 a) and b). The composition of the absorber film is Sb2S1.75Se1.25 for solar cell F and Sb2S2.3Se0.7 for G. Data for cells F and G are included in Table 1. The compositional variation generally holds true irrespective 9
ACCEPTED MANUSCRIPT of the Sb2S3 source. The deposition condition needs to be optimized for each source-mix, depending on the mesh size of the mineral powder and the laboratory reagent, which was not done in this study. Thus, the cell parameters including η of 2.71 and 2.0 % for solar cells F and G are only meant to show that powdered stibnite mineral can be used directly in making solar cells of conversion efficiency, 4%. Figure 7 b) shows that EQE maximum is 70 %, and not 80% as in Figure 6 b). Again, this would be due to a lack of optimization of the deposition parameters when
IP
T
the lab reagent is used.
We found in these results that the direct use of stibnite mineral with inclusion of
CR
encrustations of minerals, which are stable at the temperature of vacuum thermal evaporation of Sb2S3 (of nearly 850 oC), is not detrimental to solar cell parameters. However, in the calculation of
US
compositional variation while preparing such mineral source-mixes, one should recognize that Sb2S3 content tend to be overestimated, because the quartzite, ferrosilite, etc., additions in the
AN
powder are left as residue. In the case of cells F and G using laboratory reagent for Sb2S3, at the end of the deposition, the crucible was empty, with no residue found.
M
4. Conclusions
ED
It is possible to use directly powdered stibnite ore-mineral of antimony sulfide as an evaporation source to produce photoconductive thin films and solar cells of Sb2S3 with conversion
PT
efficiency of 1.62 %. If antimony selenide is added to stibnite as the thermal evaporation source, antimony sulfide selenide films are produced, with Eg continuously adjustable in the interval 1.38
CE
eV to 1.88 eV. In this study, solar cell of Sb2S2.14Se0.85 with bandgap 1.61 eV has shown conversion efficiency of 4 %; the best value of 4.24% was obtained for solar cells made from Sb2S0.5Se2.5. Based
AC
on the observations made from the J-V and EQE characteristics of the cells, we consider that the performance of these solar cells may be improved in future work. In such a case, the proposition to use an abundant mineral directly for the production of solar cells without going through oredressing, energy-intensive pyrolytic conversions/chemical reactions, purification steps, appears attractive. Acknowledgments We are grateful to María Luisa Ramón García for XRD measurements, Yareli Colín García for optical measurements; José Campos for measurements on solar cells and Oscar Gomez Daza 10
ACCEPTED MANUSCRIPT for experimental support. Financial support from CeMIESol – 35 and PAPIIT-UNAM IT100917 and
AC
CE
PT
ED
M
AN
US
CR
IP
T
infrastructure support from CONACYT-LIFYCS – 123122 are acknowledged.
11
ACCEPTED MANUSCRIPT References [1] US Geological Survey web sites https://minerals.usgs.gov/minerals/pubs/commodity/antimony/; https://minerals.usgs.gov/minerals/pubs/mcs/2016/mcs2016.pdf, accessed on 10 April 2017. [2] C. G. Anderson, The metallurgy of antimony, Chemie der Erde, 72 (2012) S4 3-8.
IP
T
[3] P. Škácha, J. Plášil, J. Sejkora, V. Goliáš, Sulfur-rich antimonselite, Sb2(Se,S)3 in the Se-bearing mineral association from the Příbram uranium and base metal ore district, Czech Republic, Journal of Geosciences, 60 (2015) 23-29.
CR
[4] O. Madelung (Ed.) Data in Science and Technology, Semiconductors Other than Group IV Elements and III-V Compounds, Springer-Verlag, Berlin (1992).
US
[5] D.Y. Suárez-Sandoval, M.T.S. Nair, P.K Nair, Photoconductive antimony sulfide-selenide thin films produced by heating a chemically deposited Se- Sb2S3 layer, Journal of the Electrochemical Society, 153 (2006) C91-C96.
AN
[6] I. L. Repins, M. J. Romero, J. V. Li, Su-Huai Wei, D. Kuciauskas, C.-S. Jiang, C. Beall, C. De Hart, J. Mann, W.-C. Hsu, G. Teeter, A. Goodrich, and R. Noufi, Kesterite successes, ongoing work, and challenges: a perspective from vacuum deposition, IEEE J. Photovolt. 3 (2013) 439-445.
M
[7] J.R. Sites, Calculation of impact ionization enhanced photovoltaic efficiency, Solar Cells 25 (1988) 163-168.
ED
[8] Kai Zeng, Ding-Jiang Xue and Jiang Tang, Antimony selenide thin-film solar cells, Semiconductor Science and Technology, 31 (2016) 063001.
PT
[9] B. Yang, S. Qin, D. Xue, Chao Chen, Yi-su He, D. Niu, H. Huang, J. Tang, In situ sulfurization to generate Sb2(Se1-xSx)3 alloyed films and their application for photovoltaics, Prog. Photovolt: Res. Appl. 25 (2016) 113-122.
CE
[10] Y. C. Choi, Y. H. Lee, S. H. Im, J. H. Noh, T. N. Mandal, W. Seok Yang, S. Il Seok, Efficient Inorganic-Organic Heterojunction Solar Cells Employing Sb2(Sx/Se1-x)3 Graded-Composition Sensitizers, Adv. Energy Mater. 4 (2014) 1301680.
AC
[11] Y. Zhou, L. Wang, S. Chen, S. Qin, X. Liu, J. Chen, D. Xue, M. Luo, Y. Cao, Y. Cheng, E. H. Sargent, J. Tang, Thin-film Sb2Se3 photovoltaics with oriented one-dimensional ribbons and benign grain boundaries, Nature Photonics, 9 (2015) 409–415. [12] Xinsheng Liu, Chao Chen, Liang Wang, Jie Zhong, Miao Luo, Jie Chen, Ding-Jiang Xue, Dengbing Li, Dengbing Li, Ying Zhou, Jiang Tang, Improving the performance of Sb2Se3 thin film solar cells over 4% by controlled addition of oxygen during film deposition, Prog. Photovoltaics: Res. Appl. 23 (2015) 1828. [13] J. Escorcia-García, D. Becerra, M.T.S. Nair, P.K. Nair, Heterojunction CdS/Sb2S3 solar cells using antimony sulfide thin films prepared by thermal evaporation, Thin Solid Films, 569 (2014) 28-34.
12
ACCEPTED MANUSCRIPT [14] D. Pérez-Martínez, J. D. Gonzaga-Sánchez, F. De Bray-Sánchez, G. Vázquez-García, J EscorciaGarcía, M. T. S. Nair, P. K. Nair, Simple solar cells of 3.5% efficiency with antimony sulfide-selenide thin films, Physica Status Solidi RRL, 10 (2016) 388–396. [15] United States Antimony Corporation USAC Stibnite Processing Outline, accessed (26 April 2017) at: http://usantimony.com/metallurgy.htm
T
[16] M. T. S. Nair, P. K. Nair, R. A. Zingaro, E. A. Meyers, Conversion of chemically deposited photosensitive CdS thin films to n-type by air annealing and ion exchange reaction, Journal of Applied Physics, 75 (1994) 1557-1564.
IP
[17] J. I. Pankove, Optical Processes in Semiconductors, Dover, New York, 1975, p. 36, 94.
AC
CE
PT
ED
M
AN
US
CR
[18] J. Nelson, The Physics of Solar Cells, first ed., Imperial College Press, London, 2003, p. 214.
13
ACCEPTED MANUSCRIPT Table 1: Characteristics of solar cells A – E of Figure 6 with cell structure, SnO2:F/CdS(100 nm)/Sb2SxSe3-x(230-370 nm)/C-Ag, produced by vacuum thermal evaporation of a mixture of stibnite (Sb2S3) mineral and Sb2Se3 precipitate as source. The substrate temperature was 400-450 o C; cell area, 0.8 – 1 cm2. Solar cells F and G of Figure 7 were produced using laboratory reagent Sb2S3 (300 mg, 400 mg) and Sb2Se3 precipitate (300 mg, 200 mg).
0.668 0.562 0.497 0.443 0.386 Voc (V)
M ED PT CE 14
Jsc 2 (mA/cm ) 6.95 13.53 18.86 22.31 23.06 Jsc 2 (mA/cm ) 12.15 11.0
T
Voc (V)
IP
Eg (eV) 1.89 1.61 1.48 1.44 1.38 Eg (eV) -
CR
Film thickness (nm), Composition 232, Sb2S3 306, Sb2S2.14Se0.86 235, Sb2S1.35Se1.65 370, Sb2S0.5Se2.5 320, Sb2Se3 Film thickness (nm, Composition 250, Sb2S1.75Se1.25 300, Sb2S2.3Se0.7
0.465 0.573
US
Sb2Se3 (mg) 200 400 500 600 Sb2Se3 (mg) 300 200
AN
Stibnite (mg) 600 400 200 100 Sb2S3 (mg) 300 400
AC
Solar cell A B C D E Solar cell F G
FF
η (%)
0.35 0.53 0.42 0.43 0.35 FF
1.62 4.03 3.93 4.24 3.11 η (%)
0.48 0.32
2.71 2.0
ACCEPTED MANUSCRIPT Figure captions Figure 1. a) Solar cell structure: SnO2:F/CdS/Sb2SxSe3-x/C-Ag; b) Development of solar cell following the sequence: CdS thin film (100 nm) chemically deposited on TEC-15 SnO2:F cut to 25 mm x 75 mm pieces; Sb2SxSe3-x thin film (250-300 nm); hand-painted colloidal graphite electrodes of 0.8-1 cm2; cell demarcation, painting of colloidal silver-paint (DuPont) electrodes for the measurement. The streak of light shows the specular reflection from a tubular fluorescent ceiling lamp on the
IP
T
electrode-free area, which is used for XRD and optical measurements on the solar cells. Figure 2. XRD patterns of the powder sources used in vacuum thermal evaporation: Bottom panel,
CR
locally sourced stibnite mineral with quartzite and ferrosilite inclusions and a comparison with Sb2S3 laboratory reagent (Aldrich).. Top panel, Sb2Se3 chemical precipitate – as prepared and
US
after heating. Miller indices (hkl) matching those of the powder diffraction files PDF of the respective minerals antimonselite (15-0861) and stibnite (42-1393) are indicated
AN
Figure 3. Grazing Incidence (0.5o) XRD patterns of thin films and solar cells A (stibnite) and E (Sb2Se3). In thin film A on Corning glass substrate, the intensity of diffraction from (020)
M
dominates, whereas in cell A (on CdS) the intensities tend to follow that of PDF 42-1393. In thin film E, (020) and (120) diffractions are prominent, but in cell E, the intensities tend to follow that
ED
of PDF-15-0861. In Sb2SxSe3-x solar cell C, the diffraction peaks (211) and (221) are placed in between those in A (Sb2S3) and E (Sb2Se3).
PT
Figure 4. a) Specular reflectance (R) at 5o (near-normal incidence) for films A-E; b) Normal incidence optical transmittance (T); c) optical absorption coefficient (α) illustrating the steady shift
CE
of the photon energy for the onset of optical absorption from 1.89 eV in A (Sb2S3) through B – D
AC
toward 1.38 eV in E (Sb2Se3); and d) (αhν)⅔ versus hν plots for thin film samples A – E. Figure 5. Photoconductivity response curves of films A (Sb2S3) – E (Sb2Se3) under tungsten halogen radiation (800 W/m2) during 20s-40s interval in each case. Fig. 6. a) J-V Characteristics of the solar cells SnO2:F/CdS/Sb2SxSe3-x/C-Ag. A (stibnite), E (Sb2Se3) and B – D of Sb2SxSe3-x (x = 2.14, 1.35, 0.5) absorbers measured for simulated air mass 1.5 Global (1000 W/m2) solar radiation incident on cell area 0.8-1 cm2. Dark J – V is for cell B; b) External Quantum Efficiency of these solar cells. Figure 7. a) J – V and b) EQE of solar cell F and G prepared by using an evaporation source of Sb2S3 (laboratory reagent, F-300 and G-400 mg) and Sb2Se3 (F-300 and G-200 mg). 15
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1. a) Solar cell structure: SnO2:F/CdS/Sb2SxSe3-x/C-Ag; b) Development of solar cell following the sequence: CdS thin film (100 nm) chemically deposited on TEC-15 SnO2:F cut to 25 mm x 75 mm pieces; Sb2SxSe3-x thin film (250-300 nm); hand-painted colloidal graphite electrodes of 0.8-1 cm2; cell demarcation, painting of colloidal silver-paint (DuPont) electrodes for the measurement. The streak of light shows the specular reflection from a tubular fluorescent ceiling lamp on the electrode-free area, which is used for XRD and optical measurements on the solar cells.
16
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
Figure 2. XRD patterns of the powder sources used in vacuum thermal evaporation: Bottom panel, locally sourced stibnite mineral with quartzite and ferrosilite inclusions and a comparison with Sb2S3 laboratory reagent (Aldrich).. Top panel, Sb2Se3 chemical precipitate – as prepared and after heating. Miller indices (hkl) matching those of the powder diffraction files PDF of the respective minerals antimonselite (15-0861) and stibnite (42-1393) are indicated.
17
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
Figure 3. Grazing Incidence (0.5o) XRD patterns of thin films and solar cells A (stibnite) and E (Sb2Se3). In thin film A on Corning glass substrate, the intensity of diffraction from (020) dominates, whereas in cell A (on CdS) the intensities tend to follow that of PDF 42-1393. In thin film E, (020) and (120) diffractions are prominent, but in cell E, the intensities tend to follow that of PDF-15-0861. In Sb2SxSe3-x solar cell C, the diffraction peaks (211) and (221) are placed in between those in A (Sb2S3) and E (Sb2Se3).
18
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Figure 4. a) Specular reflectance (R) at 5o (near-normal incidence) for films A-E; b) Normal incidence optical transmittance (T); c) optical absorption coefficient (α) illustrating the steady shift of the photon energy for the onset of optical absorption from 1.89 eV in A (Sb2S3) through B – D toward 1.38 eV in E (Sb2Se3); and d) (αhν)⅔ versus hν plots for thin film samples A – E. 19
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 5. Photoconductivity response curves of films A (Sb2S3) – E (Sb2Se3) under tungsten halogen radiation (800 W/m2) during 20s-40s interval in each case.
20
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Fig. 6. a) J-V Characteristics of the solar cells SnO2:F/CdS/Sb2SxSe3-x/C-Ag. A (stibnite), E (Sb2Se3) and B – D of Sb2SxSe3-x (x = 2.14, 1.35, 0.5) absorbers measured for simulated air mass 1.5 Global (1000 W/m2) solar radiation incident on cell area 0.8-1 cm2. Dark J – V is for cell B; b) External Quantum Efficiency of these solar cells.
21
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
Figure 7. a) J – V and b) EQE of solar cell F and G prepared by using an evaporation source of Sb2S3 (laboratory reagent, F-300 and G-400 mg) and Sb2Se3 (F-300 and G-200 mg).
22
P.K. Nair, Geovanni Vázquez García, Eira Anais Zamudio Medina, Laura Guerrero Martínez, Oscar Leyva Castrejón, Jacqueline Moctezuma Ortiz, M.T.S. Nair PII: DOI: Reference:
S0040-6090(17)30840-4 doi:10.1016/j.tsf.2017.11.004 TSF 36336
To appear in:
Thin Solid Films
Received date: Revised date: Accepted date:
4 May 2017 7 October 2017 3 November 2017
Please cite this article as: P.K. Nair, Geovanni Vázquez García, Eira Anais Zamudio Medina, Laura Guerrero Martínez, Oscar Leyva Castrejón, Jacqueline Moctezuma Ortiz, M.T.S. Nair , Antimony sulfide-selenide thin film solar cells produced from stibnite mineral. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Tsf(2017), doi:10.1016/j.tsf.2017.11.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Antimony sulfide-selenide thin film solar cells produced from stibnite mineral P. K. Nair, Geovanni Vázquez García, Eira Anais Zamudio Medina, Laura Guerrero Martínez, Oscar Leyva Castrejón, Jacqueline Moctezuma Ortiz, M. T. S Nair Instituto de Energías Renovables, Universidad Nacional Autónoma de México, Temixco-62580, Morelos, México. [email protected]
T
Abstract
IP
Stibnite (Sb2S3) is a major mineral of antimony, which occurs as large crystals often with
CR
minor encrustations of other minerals. Locally sourced powdered stibnite containing some quartzite (SiO2) and ferrosilite (FeSiO3) has been used in this work as evaporation source for vacuum deposition of thin film solar cells. The added minerals were left as residue in the crucible
US
and did not incorporate into the Sb2S3 thin film. To stibnite powder was added Sb2Se3 powder prepared in our laboratory as chemical precipitate. Source mixtures of different weight – by –
AN
weight (w/w) ratios gave thin films of chemical composition Sb2SxSe3-x and optical band gap (Eg) within 1.38 eV (Sb2Se3) – 1.88 eV (Sb2S3). Solar cells of SnO2:F/CdS (100 nm)/Sb2S3(250 nm)/C-Ag
M
prepared by using stibnite as evaporation source gave under standard conditions, open circuit voltage (Voc), 0.668 V; short circuit current density (Jsc), 6.95 mA/cm2; and conversion efficiency
ED
(η), 1.62 %. For a 2:1 (w/w) mixture of stibnite:Sb2Se3, solar cell of Sb2S2.14Se0.86 (Eg 1.61 eV) was obtained with Voc, 0.562 V; Jsc, 13.53 mA/cm2 and η, 4.03 %. For 1:5 (w/w) mixture, solar cell of
PT
Sb2S0.5Se2.5 (Eg 1.44 eV) gave Voc, 0.443 V; Jsc, 22.31 mA/cm2 and η, 4.24 %. Electrical conductivity of the Sb2SxSe3-x absorber films in the dark increased from 10-8 to 10-6 Ω-1 cm-1 and their
CE
photoconductivity, from 10-6 to 10-5 Ω-1 cm-1 as the composition changed from Sb2S3 to Sb2SxSe3-x and Sb2Se3. Direct use of an abundant mineral as the evaporation source in thin film solar cell
AC
technology is a novelty.
Key words: antimony chalcogenide, antimony sulfide selenide, Sb-S-Se, Sb2SxSe3-x; antimony chalcogenide solar cells, thin film solar cells; solar energy, renewable energy; semiconductor thin films Highlights: Direct use of stibnite ore-mineral of Sb2S3 to produce solar cells of 1.6 % efficiency Stibnite mineral and Sb2Se3 precipitate source-mix used for solar cells of 4.2 % efficiency Variable electrical conductivity and Eg 1.38 – 1.61 eV in Sb2SxSe3-x from stibnite – Sb2Se3
1
ACCEPTED MANUSCRIPT
1. Introduction Production of solar cells directly from an abundant mineral is a very attractive proposition. In this work we consider stibnite mineral with chemical composition Sb2S3 as a candidate. Stibnite is the predominant ore mineral of antimony, distributed over the earth’s crust at 0.2-0.5 parts per million (ppm) [1]. Global production of antimony (metal content) in the year 2015 was nearly
T
150,000 metric tons. Antimony metal price was close to US$8 per kg (commodity index) during
IP
2015, and shows a declining-price trend over time [1]. Stibnite occurs as large crystals and crystal
CR
clusters weighing over many kg in many parts of the world. While these crystals are generally chemically pure (99% Sb2S3) often some encrustations of calcite, quartzite, etc., are found on them
US
[2]. In a locally sourced stibnite powder used in this work to produce thin film solar cells, quartzite and ferrosilite were evidenced, but these were left as residue in thermal evaporation.
AN
Antimonselite (Sb2Se3) as well as its partially S-substituted mineral Sb2S0.423Se2.577 [3] are also among antimony minerals [1, 2].
M
Both Sb2S3 and Sb2Se3 are of orthorhombic crystalline structure; and they possess direct optical band gap (Eg) of 1.88 and 1.1 eV, respectively [4]. Thin films of solid solutions of antimony
ED
sulfide-selenide (Sb2SxSe3-x) with Eg continuously adjustable within 1.1 – 1.88 eV have been prepared through heating chemically deposited Se-Sb2S3 layers [5]. Such materials would satisfy
PT
many of the criteria prescribed in a 2014 work for thin film solar cell materials with respect to material properties, conversion efficiency (η), stability, and price for commercial photovoltaic
CE
module manufacturing [6]. Photovoltaic absorber materials with Eg 1.3 – 1.7 eV of Sb2SxSe3-x solid solutions fulfil criteria set for thin film solar cell materials so as to combine large-enough open
AC
circuit voltage (Voc) and fill factor (FF) of solar cells with appreciable short circuit current density (Jsc) to achieve high η [7]. These guidelines led to a growing interest in antimony sulfide-selenide solar cells in recent years [8]. Among the best conversion efficiency of antimony sulfide-selenide solar cells is 5.79 %, reported in 2016 [9]. This solar cell used tin-doped In2O3 (ITO):
ITO/CdS/Sb2SxSe3-x/Au. The
absorber film was made by rapid thermal evaporation followed by reaction in sulfur-vapor. Following parameters were reported for it: Voc, 0.50 V; Jsc, 22.33 mA/cm2 and FF, 0.521. The best η to date for a Sb2SxS3-x solar cell appears to be 6.3 % reported in 2014 [10]. Compact and mesoporous (mp) layers of TiO2 on fluorine doped SnO2 (FTO) were used along with hole-transport 2
ACCEPTED MANUSCRIPT organic media (HTM/HTL): FTO/bulk-TiO2/mp-TiO2/Sb2SxSe3-x/HTM/HTL/Au. The absorber layer was prepared by spin coating (Sb2Se3) and chemical deposition (Sb2S3) followed by heating. This solar cell had Voc, 0.475 V; Jsc, 24.9 mA/cm2 and FF, 0.556. Solar cells of Sb2Se3 absorbers consistently lagged in Voc, while η and Jsc were comparable. In 2015 [11] η of 5.6 % was reported for a CdS/Sb2Se3 solar cell with Voc, 0.40 V; Jsc, 25.1 mA/cm2, and FF, 0.557. An η of 4.8 % was reported in ITO/CdS/Sb2Se3/Au with Voc, 0.36 V; Jsc , 25.3 mA/cm2; and FF , 0.52 [12]. Use of Sb2S3
T
alone as the absorber film elevates Voc to 0.6 V, but it also lowers η toward 1%, as reported in 2014
IP
[13] in FTO/CdS/Sb2S3/C-Ag, with Voc, 0.60 V; Jsc, 6.12 mA/cm2; FF, 0.35 and η, 1.27%. By steadily
CR
decreasing x in Sb2SxSe3-x solar cells prepared by thermal evaporation of Sb-S-Se chemical precipitate, η was improved but Voc reduced, as reported in 2016 for FTO/CdS/Sb2S1.7Se1.3/C-Ag with Voc, 0.545 V; Jsc, 13.5 mA/cm2; FF, 0.43; η, 3.12% and for Sb2S1.5Se1.5 with Voc, 0.528 V; Jsc, 15.7
US
mA /cm2; FF, 0.44; and η, 3.63 % [14].
AN
In this work we used a mixture of locally sourced powdered stibnite mineral and Sb2Se3 powder prepared as reported in [14], as thermal evaporation source to produce thin film solar
M
cells of Sb2SxSe3-x absorbers. Conversion efficiency of up to 4.24% has been reached in these solar cells. The basic properties of the stibnite mineral and of the chemical precipitate are presented.
ED
Characteristics of thin film Sb2SxSe3-x and of the solar cells show that direct use of stibnite mineral for solar cells is feasible. This option is attractive because the additive minerals which are often
PT
present in stibnite powder do not seem to affect the quality of the thin films or of solar cells. This approach eliminates the laborious and energy-intensive conversion of hand-sorted stibnite crystals
CE
first into powdered stibnite mineral, then into Sb2O3 or Sb-metal [2, 15] before reconverting it to laboratory reagent Sb2S3.
AC
2. Experimental details
Substrates- Glass slides (Corning) measuring 7.5 mm x 2.5 mm of 1 mm thickness or fluorinedoped SnO2 (FTO) of sheet resistance 15 Ω (TEC 15, NSG Pilkington) cut to the same size. CdS thin film deposition: Chemical deposition method was reported in 1994 [16] and mentioned in [13, 14]. The solution mixture for the deposition contained cadmium nitrate, sodium citrate, ammonium hydroxide, thiourea (Aldrich reagents) and distilled water. Up to six clean substrates were immersed in 80 ml of the deposition bath at 80 oC. At 60 min, a specularly reflective CdS thin
3
ACCEPTED MANUSCRIPT film of 80-100 nm in thickness was deposited. This film is of hexagonal crystalline structure with band gap 2.5 eV; with a photoconductivity under illumination of 0.01-0.1 Ω-1 cm-1 [16]. Sb2Se3 chemical precipitate: The method to prepare thin film or precipitate of Sb2SxSe3-x in the laboratory was described in [14]. A solution containing potassium antimony tartrate, triethanolamine, ammonia (aq.), silicotungstic acid, thioacetamide and/or selenosulfate solution,
T
and distilled water was maintained at 80 oC for 4-5 h. It was then cooled down overnight to room
CR
reagents of 98-99% assay were used for this chemical precipitation.
IP
temperature; filtered, rinsed with deionized water, dried and stored. Aldrich or Baker Analyzed
Stibnite (Sb2S3) powder: Was obtained from a local supplier for industrial grade chemicals. The
US
assay is normally stated at 95-98 % purity (Reasol Reactivos Analiticos, Mexico). For comparison of results, we also used laboratory grade reagent of Sb2S3 of 99% assay (Aldrich).
AN
Vacuum thermal evaporation of stibnite-Sb2Se3 powder mix: Thin films and solar cells were prepared from 600 mg of the evaporation source of five different stibnite:Sb2Se3 weight-by-weight
M
ratio (w/w). The films and solar cells would be labelled accordingly: A (600:0); B (400:200); C (200:400); D (100:500); E (0:600). For solar cells labelled F and G, laboratory grade Sb2S3:Sb2Se3
ED
(300:300) and (400:200) were used. The powder source mix was painted on tungsten crucible by using a few drops of propylene glycol and then dried at 80 oC. A Torr International diffusion pump
PT
system was used for thermal evaporation of this source to produce thin films. The chamber reached a base pressure of 2x10-6 Torr (2.7x10-4 Pa). Up to six substrates (2 glass slides, 4 FTO/CdS
CE
substrates) were mounted on a rotating sample holder located 40 cm from the source. The temperature of the substrate was kept at 400-450 oC using four linear tungsten-halogen lamps
AC
mounted at the back of the sample holder. The temperature control was achieved through intermittent lighting of these lamps linked to a temperature controller circuit. The source mixture starts melting in the crucible at a temperature close to the melting point of Sb2S3 (550 oC) and of Sb2Se3 (610 oC). Temperature of the crucible during the thermal evaporation of the source mixture was nearly 850 oC. The duration of thermal evaporation was 15 min. The film thickness showed variation depending on the composition of the source mixture. We noted that when the source contained only stibnite mineral (sample A), the substrate temperature had to be lowered to 400 oC in order to avoid sublimation loss of the thin film from the substrate. This was the case also with laboratory reagent Sb2S3, as reported previously [13]. Addition of even small quantity of Sb2Se3 to 4
ACCEPTED MANUSCRIPT the source mixture seemed to inhibit such loss. The film thicknesses were: A, 232 nm; B, 306 nm; C, 235 nm; D, 370 nm; E, 320 nm; F, 250 nm; G, 300 nm. No effort was made at this stage of work to secure
exactly the same thickness of films in all cases. Solar cell structures: Figure 1 a) shows the cross-sectional schematics of FTO/CdS/Sb2SxSe3-x/C-Ag solar cells; and 1 b) illustrates the development of such cells through the following steps: CdS thin film (100 nm) was chemically deposited on FTO cut to 25 mm x 75 mm pieces;
(ii)
Deposition via vacuum thermal evaporation of a Sb2SxSe3-x thin film (230-370 nm) was
(iii)
CR
done from stibnite-Sb2Se3 sources (total mass, 600 mg);
IP
T
(i)
Hand-painted colloidal graphite electrodes (SPI Chem) of 0.8-1 cm2 in area were
US
applied and the cells were heated at 300 oC for 30 min in nitrogen at a pressure 20 Torr (2.7 kPa); (iv)
Cell periphery was etched out; colloidal silver-paint (DuPont) was applied on the
AN
graphite-paint and multi-strand copper wire was attached on it for measurements. (The streak of light shows the specular reflection from a tubular fluorescent ceiling
M
lamp on the electrode-free area. This part of the layer-structure was used for grazing-
ED
incidence X-ray diffraction (GIXRD) and optical measurements.) Measurements: Film thickness (d) was measured using Alpha-Step 100. Optical transmittance (T)
PT
against air and near-normal (5o incidence) specular reflectance (R) against standard front mirror were measured for film-side incidence of glass/Sb2SxSe3-x films. Shimadzu 1800 UV
CE
spectrophotometer was used. Optical absorption coefficient (α) of the films was estimated using the basic equation considering multiple reflection [17]: T = [(1 – R)2e-αd]/[1 – R2e-2αd]. X-ray
AC
diffraction (XRD) measurement on the source-powder and grazing incidence – GIXRD (incidence, 0.5o) measurement on Sb2SxSe3-x film (glass/Sb2SxSe3-x) and on FTO/CdS/Sb2SxSe3-x were done on a Rigaku-D-MAX 2000 diffractometer with Cu-Kα (1.5406 Å) radiation. The value of x for the solid solution Sb2SxSe3-x was determined by assuming a proportional variation of the inter-planar distance d(hkl) in the orthorhombic cell of the sample between that of stibnite-Sb2S3 (powder diffraction file, PDF 42-1393) and antimonselite-Sb2Se3 (PDF 15-0861) orthorhombic cells. Electrical conductivity of the films in the dark and under illumination (850 W/m 2 tungstenhalogen light) were determined by applying a bias of 10 V and recording the current during 10 s under the dark, 10 s under illumination, and 10 s in the dark. The J-V characteristics of the solar 5
ACCEPTED MANUSCRIPT cells and external quantum efficiency (EQE) of the solar cells were obtained using a SCIENCETECH (Canada) Photovoltaic Testing System (PTS) under simulated solar radiation of air-mass 1.5 Global spectrum of intensity 1000 W m–2 from a xenon arc-lamp and with a monochromator calibrated using a standard Si – photodiode. Prior to these measurements all solar cells were stabilized under the sun (870 – 980 W/m2) for 6 h. The solar cell parameters were also confirmed from measurements made directly under the sun and from J – V characteristics recorded in a locally
T
fabricated set-up. The stability of the solar cell was ascertained during repeated measurements on
IP
the cells during a period of over six months.
CR
3. Results and discussion
US
3.1 Crystalline structure
In Figure 2 (bottom panel) is seen that the reagent grade Sb2S3 shows an XRD pattern
AN
perfectly matching that of Sb2S3 – stibnite, but that of the locally sourced stibnite mineral shows in addition the major peaks of two added minerals, (101) of quartz (SiO2) and (610) of ferrosilite (FeSiO3). Both these minerals are stable at high temperatures, and are expected to be left as
M
residue in the crucible at temperature of evaporation of Sb2S3 of nearly 850 oC. The XRD pattern of
ED
the as-prepared precipitate of Sb2Se3 in the top panel does not show well-defined peaks, but upon heating it at 300 oC the pattern matches that of Sb2Se3 – antimonselite.
PT
Figure 3 shows that the GIXRD pattern of thin film A produced from stibnite mineral powder matches in peak positions (2θ) with those of the standard pattern for Sb2S3, but the
CE
relative intensities vary significantly from the standard, suggesting preferential orientation of crystalline planes on heated (400 oC) glass substrates. However, the film deposited for the cell
AC
structures A (as well as for B-G cells) is on a CdS substrate layer. In this case the GIXRD pattern matches the standard stibnite pattern in 2θ and relative intensity. Additive minerals present in the powdered stibnite mineral (Figure 2) do not lead to added crystalline matter in the deposited thin film. Variation in relative intensity is a behavior repeated in the case of thin film E deposited on glass substrate (Figure 3 top panel). The thin film constituting the solar cell E matches in 2θ and relative intensity with those of antimonselite – Sb2Se3. In solar cell C, produced from a source-mix of stibnite:Sb2Se3 (200:400), the peak positions and relative intensities are intermediate to those of Sb2S3 and Sb2Se3, as indicated by the dotted lines. The relative position of the peaks along 2θ helps estimate the composition of antimony sulfide selenide solid solution in film/Cell C as 6
ACCEPTED MANUSCRIPT Sb2S1.35Se1.65. The method assumes proportional variation of inter-planar distance for any (hkl) crystallographic planes as the composition changes in Sb2SxSe3-x from Sb2S3 (x = 3) to Sb2Se3 (x = 0), as discussed in ref [5] and illustrated in ref [14]. In a same way the composition of films B, D, F and G were determined as: B, Sb2S2.14Se0.86; D, Sb2S0.5Se2.5, F: Sb2S1.5Se1.5 and G, Sb2S2.3Se0.7. Estimate for crystallite diameter (D) for the materials was obtained from non-overlapping peaks (120) by the XRD software through instrumental-corrected full-width (Δ2θ) at half maximum (FWHM) of
IP
T
peak intensity. The values were in the 16-19 nm interval.
CR
3.2 Optical and electrical characteristics
Optical band gap Eg of absorber materials decides the maximum light-generated current
US
density (JL), which also is the maximum value for Jsc in solar cells. The value of α, and the nature of electronic transitions (direct or indirect) decide the thickness of film (3/α) or (4/α) required to achieve 95 % or 98 % of JL [18]. In the present study a thickness interval (230-370 nm) of Sb2SxSe3-x
AN
films was used (given in the experimental section), which might have to be further optimized for solar cells in future work. Data on R and T on the films A-E in Figure 4 a) and b) help evaluate the
M
plots of α and (αhν)⅔ versus photon energy (hν) of Figure 4 c) and d). The optical interference in thin films is not entirely compensated in T and R spectra because R is measured at an angle of
ED
incidence 5o, and T at 0o (normal incidence), entailing larger optical path in the R measurement of the films. This difference gives rise to some waviness in the plots in 4 c) and d). Further, for highly
PT
absorbing thicker films, the transmittance value is not accurately measured for T < 0.01% due to
flatten-off.
CE
equipment limitation. Hence at high hν and very low T, α tend to be sub-estimated and are seen to
AC
For all the films, α sharply increases and surpasses 105 cm-1 upon the onset of optical absorption. Thus, for effective optical absorption, film thickness used here (3/α, of nearly 300 nm) would suffice. However, to inhibit electrical instability or to increase parallel resistance, higher film thicknesses may have to be sought in a later work. Analysis of Eg in Figure 4 d) for antimony sulfide (A), antimony selenide (E) and Sb2SxSe3-x (B – D) thin films tend to suggest ‘direct gaps with forbidden transitions’ [17] for them, in accordance with previous results on these materials [13, 14]. It is seen that the value of Eg may be continuously varied in the films through varying the stibnite:Sb2Se3 ratio. The high optical reflectance of 45 – 50 % of the films seen in Figure 4 b) originates from a high refractive index (n) of nearly 5, and thus entails high static relative
7
ACCEPTED MANUSCRIPT permittivity ( εrs > n2 , hence > 25). This leads to large exciton Bohr radii (rB) in these materials, bringing them comparable with the value of the crystallite grain diameter D (20 nm). Thus, the Eg becomes susceptible to increase due to quantum confinement of the excitons. The Eg of film E, which is Sb2Se3 in composition (Figure 2) is therefore 1.38 eV, which is greater than for the bulk crystalline value of 1.1 eV [4]. Due to this reason, the proportional variation of Eg for Sb2SxSe3-x films within the boundary values of 1.1 – 1.88 eV for Sb2Se3 – Sb2S3 is not strictly followed. What is
T
important is that the Eg sought for promising solar cell absorber materials of 1.4 – 1.6 eV [6, 7] is
IP
available in films B – D.
-8
-6
-1
-1
CR
Figure 5 shows the electrical conductivity (σ) in the dark (10 to 10 Ω cm ) and under
illumination (10-6 to 10-5 Ω-1 cm-1) of the thin films A – E. Both values increase steadily with
US
decrease in x of Sb2SxSe3-x of the thin films. In Sb2Se3 with Eg of 1.38 eV, photons of a wider spectrum is absorbed compared with that in Sb2S3 with 1.89 eV. This leads to higher
AN
photoconductivity in the film E, and hence it is also expected to result in higher Jsc in the solar cell E.
M
3.3 Solar cell characteristics
ED
Systematic variation in the optical and electrical properties of Sb2SxSe3-x thin films discussed above results in a decreasing trend in Voc from 0.668 V to 0.386 V as x drops from 3 to 0,
PT
seen in Figure 6 a). Under similar cell fabrication method, a decrease in Eg from 1.89 (A) to 1.38 eV (E) entails an increase in the intrinsic carrier concentration and hence in an increase in the reverse
CE
saturation current density (Jo). Hence [ln (JL/Jo)] decreases with decrease in x, and so too would Voc, which is related to [ln (JL/Jo)] through the thermal voltage (26 mV at 300 K). An increase in Jsc with
AC
decrease in x takes place, from 6.95 to 23.06 mA/cm2, in accordance with the increasing trend in photoconductivity of Figure 5. Table 1 provides a summary of the solar cell parameters of cells A – E, which illustrates that systemic trend in variation of parameters, with cell efficiency of 4 % seen accessible. It is pertinent to mention that the source and substrate temperatures for vacuum thermal evaporation of thin films and solar cells need to be optimized for each source-mix. In this feasibility study on Sb2SxSe3-x films produced from stibnite mineral, nearly the same condition of cell preparation was used in all cases. For Cell B, the source contained 66.7 % (w/w) (or 74 mol %) of stibnite mineral and produced thin films of Sb2S2.14Se0.86 with Eg, 1.61 eV. This cell gave Voc, 8
ACCEPTED MANUSCRIPT 0.562 V; Jsc, 13.53 mA/cm2 and η, 4.03 %. The maximum η of 4.24% was obtained for Cell D for which the use of stibnite mineral is only ⅙ (600) =16.67 % (w/w). The lack of charge carrier collection at the graphite electrode is suggested by the continued increase of the photo-generated current density toward reverse bias of – 0.4 V. Thus, better cell design would lead to increase in Jsc in all the cases – for example 16 mA/cm2 for Cell B, 25 mA/cm2 for Cell D (Eg, 1.44 eV), etc., might
T
be feasible, with consequent increases in cell efficiency.
IP
EQE plots in Figure 6 b) obtained under Jsc condition help identify the same limitation – a lack of collection of photo-generated charge carriers at the graphite electrode. A welcome
CR
disparity exists between the wavelength λg (nm) = 1240/Eg(eV) indicated on the plots for the onset for EQE, as expected from the optical absorption plot of Figure 4 c) and d), and the actual onset,
US
which occurs at still longer wavelengths. This could arise from “tail states,” which in this case may be due to some larger crystallites relatively exempt from quantum confinement, and hence with
AN
smaller bandgap, which are not evident in the optical absorption analyzed from T and R data. If some population of such larger crystallites exists in the absorber film, the optical absorption and
M
hence EQE onset would occur closer to 1130 nm (1240/1.1 eV (bulk)) and not at 900 nm (1240/1.38 eV of Fig. 4 d) for Cell E, as observed. The EQE onset at longer wavelength may be also
ED
due to compositional variation among the crystallites. More detailed study is required to discern these mechanisms. Evidenced in the EQE data in Figure 6 b) is also that the values drop drastically
PT
at wavelength shorter than 500 nm (2.48 eV). This is directly related to optical absorption in the non-participative part of the CdS window film, outside of the junction at the FTO/CdS interface.
CE
This behavior in EQE calls for a reduction in the thickness of CdS film from 80-100 nm or the addition of a more transparent-resistive thin film along with such reduction [6]. In the n+-CdS/p-
AC
Sb2SxSe3-x junction, the depletion width in the n+ - layer would be a few nm. However, the reduction in CdS film thickness toward such thickness might leave short-circuit paths which would cause reduction in Voc and instability of the solar cell during prolonged operation. For comparison, solar cells were made under the same conditions as of cells B – E, but using a laboratory reagent grade Sb2S3 without the additive minerals in the powdered stibnite (see Fig. 1 bottom panel). J-V characteristics and EQE of cell F and G prepared by using Sb2S3 (Lab. reagent):Sb2Se3 in 300:300 and 400:200 (w/w) ratio, respectively, are given in Figure 7 a) and b). The composition of the absorber film is Sb2S1.75Se1.25 for solar cell F and Sb2S2.3Se0.7 for G. Data for cells F and G are included in Table 1. The compositional variation generally holds true irrespective 9
ACCEPTED MANUSCRIPT of the Sb2S3 source. The deposition condition needs to be optimized for each source-mix, depending on the mesh size of the mineral powder and the laboratory reagent, which was not done in this study. Thus, the cell parameters including η of 2.71 and 2.0 % for solar cells F and G are only meant to show that powdered stibnite mineral can be used directly in making solar cells of conversion efficiency, 4%. Figure 7 b) shows that EQE maximum is 70 %, and not 80% as in Figure 6 b). Again, this would be due to a lack of optimization of the deposition parameters when
IP
T
the lab reagent is used.
We found in these results that the direct use of stibnite mineral with inclusion of
CR
encrustations of minerals, which are stable at the temperature of vacuum thermal evaporation of Sb2S3 (of nearly 850 oC), is not detrimental to solar cell parameters. However, in the calculation of
US
compositional variation while preparing such mineral source-mixes, one should recognize that Sb2S3 content tend to be overestimated, because the quartzite, ferrosilite, etc., additions in the
AN
powder are left as residue. In the case of cells F and G using laboratory reagent for Sb2S3, at the end of the deposition, the crucible was empty, with no residue found.
M
4. Conclusions
ED
It is possible to use directly powdered stibnite ore-mineral of antimony sulfide as an evaporation source to produce photoconductive thin films and solar cells of Sb2S3 with conversion
PT
efficiency of 1.62 %. If antimony selenide is added to stibnite as the thermal evaporation source, antimony sulfide selenide films are produced, with Eg continuously adjustable in the interval 1.38
CE
eV to 1.88 eV. In this study, solar cell of Sb2S2.14Se0.85 with bandgap 1.61 eV has shown conversion efficiency of 4 %; the best value of 4.24% was obtained for solar cells made from Sb2S0.5Se2.5. Based
AC
on the observations made from the J-V and EQE characteristics of the cells, we consider that the performance of these solar cells may be improved in future work. In such a case, the proposition to use an abundant mineral directly for the production of solar cells without going through oredressing, energy-intensive pyrolytic conversions/chemical reactions, purification steps, appears attractive. Acknowledgments We are grateful to María Luisa Ramón García for XRD measurements, Yareli Colín García for optical measurements; José Campos for measurements on solar cells and Oscar Gomez Daza 10
ACCEPTED MANUSCRIPT for experimental support. Financial support from CeMIESol – 35 and PAPIIT-UNAM IT100917 and
AC
CE
PT
ED
M
AN
US
CR
IP
T
infrastructure support from CONACYT-LIFYCS – 123122 are acknowledged.
11
ACCEPTED MANUSCRIPT References [1] US Geological Survey web sites https://minerals.usgs.gov/minerals/pubs/commodity/antimony/; https://minerals.usgs.gov/minerals/pubs/mcs/2016/mcs2016.pdf, accessed on 10 April 2017. [2] C. G. Anderson, The metallurgy of antimony, Chemie der Erde, 72 (2012) S4 3-8.
IP
T
[3] P. Škácha, J. Plášil, J. Sejkora, V. Goliáš, Sulfur-rich antimonselite, Sb2(Se,S)3 in the Se-bearing mineral association from the Příbram uranium and base metal ore district, Czech Republic, Journal of Geosciences, 60 (2015) 23-29.
CR
[4] O. Madelung (Ed.) Data in Science and Technology, Semiconductors Other than Group IV Elements and III-V Compounds, Springer-Verlag, Berlin (1992).
US
[5] D.Y. Suárez-Sandoval, M.T.S. Nair, P.K Nair, Photoconductive antimony sulfide-selenide thin films produced by heating a chemically deposited Se- Sb2S3 layer, Journal of the Electrochemical Society, 153 (2006) C91-C96.
AN
[6] I. L. Repins, M. J. Romero, J. V. Li, Su-Huai Wei, D. Kuciauskas, C.-S. Jiang, C. Beall, C. De Hart, J. Mann, W.-C. Hsu, G. Teeter, A. Goodrich, and R. Noufi, Kesterite successes, ongoing work, and challenges: a perspective from vacuum deposition, IEEE J. Photovolt. 3 (2013) 439-445.
M
[7] J.R. Sites, Calculation of impact ionization enhanced photovoltaic efficiency, Solar Cells 25 (1988) 163-168.
ED
[8] Kai Zeng, Ding-Jiang Xue and Jiang Tang, Antimony selenide thin-film solar cells, Semiconductor Science and Technology, 31 (2016) 063001.
PT
[9] B. Yang, S. Qin, D. Xue, Chao Chen, Yi-su He, D. Niu, H. Huang, J. Tang, In situ sulfurization to generate Sb2(Se1-xSx)3 alloyed films and their application for photovoltaics, Prog. Photovolt: Res. Appl. 25 (2016) 113-122.
CE
[10] Y. C. Choi, Y. H. Lee, S. H. Im, J. H. Noh, T. N. Mandal, W. Seok Yang, S. Il Seok, Efficient Inorganic-Organic Heterojunction Solar Cells Employing Sb2(Sx/Se1-x)3 Graded-Composition Sensitizers, Adv. Energy Mater. 4 (2014) 1301680.
AC
[11] Y. Zhou, L. Wang, S. Chen, S. Qin, X. Liu, J. Chen, D. Xue, M. Luo, Y. Cao, Y. Cheng, E. H. Sargent, J. Tang, Thin-film Sb2Se3 photovoltaics with oriented one-dimensional ribbons and benign grain boundaries, Nature Photonics, 9 (2015) 409–415. [12] Xinsheng Liu, Chao Chen, Liang Wang, Jie Zhong, Miao Luo, Jie Chen, Ding-Jiang Xue, Dengbing Li, Dengbing Li, Ying Zhou, Jiang Tang, Improving the performance of Sb2Se3 thin film solar cells over 4% by controlled addition of oxygen during film deposition, Prog. Photovoltaics: Res. Appl. 23 (2015) 1828. [13] J. Escorcia-García, D. Becerra, M.T.S. Nair, P.K. Nair, Heterojunction CdS/Sb2S3 solar cells using antimony sulfide thin films prepared by thermal evaporation, Thin Solid Films, 569 (2014) 28-34.
12
ACCEPTED MANUSCRIPT [14] D. Pérez-Martínez, J. D. Gonzaga-Sánchez, F. De Bray-Sánchez, G. Vázquez-García, J EscorciaGarcía, M. T. S. Nair, P. K. Nair, Simple solar cells of 3.5% efficiency with antimony sulfide-selenide thin films, Physica Status Solidi RRL, 10 (2016) 388–396. [15] United States Antimony Corporation USAC Stibnite Processing Outline, accessed (26 April 2017) at: http://usantimony.com/metallurgy.htm
T
[16] M. T. S. Nair, P. K. Nair, R. A. Zingaro, E. A. Meyers, Conversion of chemically deposited photosensitive CdS thin films to n-type by air annealing and ion exchange reaction, Journal of Applied Physics, 75 (1994) 1557-1564.
IP
[17] J. I. Pankove, Optical Processes in Semiconductors, Dover, New York, 1975, p. 36, 94.
AC
CE
PT
ED
M
AN
US
CR
[18] J. Nelson, The Physics of Solar Cells, first ed., Imperial College Press, London, 2003, p. 214.
13
ACCEPTED MANUSCRIPT Table 1: Characteristics of solar cells A – E of Figure 6 with cell structure, SnO2:F/CdS(100 nm)/Sb2SxSe3-x(230-370 nm)/C-Ag, produced by vacuum thermal evaporation of a mixture of stibnite (Sb2S3) mineral and Sb2Se3 precipitate as source. The substrate temperature was 400-450 o C; cell area, 0.8 – 1 cm2. Solar cells F and G of Figure 7 were produced using laboratory reagent Sb2S3 (300 mg, 400 mg) and Sb2Se3 precipitate (300 mg, 200 mg).
0.668 0.562 0.497 0.443 0.386 Voc (V)
M ED PT CE 14
Jsc 2 (mA/cm ) 6.95 13.53 18.86 22.31 23.06 Jsc 2 (mA/cm ) 12.15 11.0
T
Voc (V)
IP
Eg (eV) 1.89 1.61 1.48 1.44 1.38 Eg (eV) -
CR
Film thickness (nm), Composition 232, Sb2S3 306, Sb2S2.14Se0.86 235, Sb2S1.35Se1.65 370, Sb2S0.5Se2.5 320, Sb2Se3 Film thickness (nm, Composition 250, Sb2S1.75Se1.25 300, Sb2S2.3Se0.7
0.465 0.573
US
Sb2Se3 (mg) 200 400 500 600 Sb2Se3 (mg) 300 200
AN
Stibnite (mg) 600 400 200 100 Sb2S3 (mg) 300 400
AC
Solar cell A B C D E Solar cell F G
FF
η (%)
0.35 0.53 0.42 0.43 0.35 FF
1.62 4.03 3.93 4.24 3.11 η (%)
0.48 0.32
2.71 2.0
ACCEPTED MANUSCRIPT Figure captions Figure 1. a) Solar cell structure: SnO2:F/CdS/Sb2SxSe3-x/C-Ag; b) Development of solar cell following the sequence: CdS thin film (100 nm) chemically deposited on TEC-15 SnO2:F cut to 25 mm x 75 mm pieces; Sb2SxSe3-x thin film (250-300 nm); hand-painted colloidal graphite electrodes of 0.8-1 cm2; cell demarcation, painting of colloidal silver-paint (DuPont) electrodes for the measurement. The streak of light shows the specular reflection from a tubular fluorescent ceiling lamp on the
IP
T
electrode-free area, which is used for XRD and optical measurements on the solar cells. Figure 2. XRD patterns of the powder sources used in vacuum thermal evaporation: Bottom panel,
CR
locally sourced stibnite mineral with quartzite and ferrosilite inclusions and a comparison with Sb2S3 laboratory reagent (Aldrich).. Top panel, Sb2Se3 chemical precipitate – as prepared and
US
after heating. Miller indices (hkl) matching those of the powder diffraction files PDF of the respective minerals antimonselite (15-0861) and stibnite (42-1393) are indicated
AN
Figure 3. Grazing Incidence (0.5o) XRD patterns of thin films and solar cells A (stibnite) and E (Sb2Se3). In thin film A on Corning glass substrate, the intensity of diffraction from (020)
M
dominates, whereas in cell A (on CdS) the intensities tend to follow that of PDF 42-1393. In thin film E, (020) and (120) diffractions are prominent, but in cell E, the intensities tend to follow that
ED
of PDF-15-0861. In Sb2SxSe3-x solar cell C, the diffraction peaks (211) and (221) are placed in between those in A (Sb2S3) and E (Sb2Se3).
PT
Figure 4. a) Specular reflectance (R) at 5o (near-normal incidence) for films A-E; b) Normal incidence optical transmittance (T); c) optical absorption coefficient (α) illustrating the steady shift
CE
of the photon energy for the onset of optical absorption from 1.89 eV in A (Sb2S3) through B – D
AC
toward 1.38 eV in E (Sb2Se3); and d) (αhν)⅔ versus hν plots for thin film samples A – E. Figure 5. Photoconductivity response curves of films A (Sb2S3) – E (Sb2Se3) under tungsten halogen radiation (800 W/m2) during 20s-40s interval in each case. Fig. 6. a) J-V Characteristics of the solar cells SnO2:F/CdS/Sb2SxSe3-x/C-Ag. A (stibnite), E (Sb2Se3) and B – D of Sb2SxSe3-x (x = 2.14, 1.35, 0.5) absorbers measured for simulated air mass 1.5 Global (1000 W/m2) solar radiation incident on cell area 0.8-1 cm2. Dark J – V is for cell B; b) External Quantum Efficiency of these solar cells. Figure 7. a) J – V and b) EQE of solar cell F and G prepared by using an evaporation source of Sb2S3 (laboratory reagent, F-300 and G-400 mg) and Sb2Se3 (F-300 and G-200 mg). 15
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
Figure 1. a) Solar cell structure: SnO2:F/CdS/Sb2SxSe3-x/C-Ag; b) Development of solar cell following the sequence: CdS thin film (100 nm) chemically deposited on TEC-15 SnO2:F cut to 25 mm x 75 mm pieces; Sb2SxSe3-x thin film (250-300 nm); hand-painted colloidal graphite electrodes of 0.8-1 cm2; cell demarcation, painting of colloidal silver-paint (DuPont) electrodes for the measurement. The streak of light shows the specular reflection from a tubular fluorescent ceiling lamp on the electrode-free area, which is used for XRD and optical measurements on the solar cells.
16
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
Figure 2. XRD patterns of the powder sources used in vacuum thermal evaporation: Bottom panel, locally sourced stibnite mineral with quartzite and ferrosilite inclusions and a comparison with Sb2S3 laboratory reagent (Aldrich).. Top panel, Sb2Se3 chemical precipitate – as prepared and after heating. Miller indices (hkl) matching those of the powder diffraction files PDF of the respective minerals antimonselite (15-0861) and stibnite (42-1393) are indicated.
17
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
Figure 3. Grazing Incidence (0.5o) XRD patterns of thin films and solar cells A (stibnite) and E (Sb2Se3). In thin film A on Corning glass substrate, the intensity of diffraction from (020) dominates, whereas in cell A (on CdS) the intensities tend to follow that of PDF 42-1393. In thin film E, (020) and (120) diffractions are prominent, but in cell E, the intensities tend to follow that of PDF-15-0861. In Sb2SxSe3-x solar cell C, the diffraction peaks (211) and (221) are placed in between those in A (Sb2S3) and E (Sb2Se3).
18
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Figure 4. a) Specular reflectance (R) at 5o (near-normal incidence) for films A-E; b) Normal incidence optical transmittance (T); c) optical absorption coefficient (α) illustrating the steady shift of the photon energy for the onset of optical absorption from 1.89 eV in A (Sb2S3) through B – D toward 1.38 eV in E (Sb2Se3); and d) (αhν)⅔ versus hν plots for thin film samples A – E. 19
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Figure 5. Photoconductivity response curves of films A (Sb2S3) – E (Sb2Se3) under tungsten halogen radiation (800 W/m2) during 20s-40s interval in each case.
20
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Fig. 6. a) J-V Characteristics of the solar cells SnO2:F/CdS/Sb2SxSe3-x/C-Ag. A (stibnite), E (Sb2Se3) and B – D of Sb2SxSe3-x (x = 2.14, 1.35, 0.5) absorbers measured for simulated air mass 1.5 Global (1000 W/m2) solar radiation incident on cell area 0.8-1 cm2. Dark J – V is for cell B; b) External Quantum Efficiency of these solar cells.
21
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
Figure 7. a) J – V and b) EQE of solar cell F and G prepared by using an evaporation source of Sb2S3 (laboratory reagent, F-300 and G-400 mg) and Sb2Se3 (F-300 and G-200 mg).
22