Aqueous Synthesis of High-Quality Cu2ZnSnS4 ... - ACS Publications

Jan 2, 2018 - ARC Centre of Excellence in Exciton Science, School of Chemistry and Bio21 Institute, University of Melbourne, 30 Flemington. Road, Park...
0 downloads 3 Views 2MB Size
Subscriber access provided by UNIV OF DURHAM

Article

Aqueous Synthesis of High-Quality Cu2ZnSnS4 Nanocrystals and their Thermal Annealing Characteristics Cameron Ritchie, Anthony Chesman, Mark J Styles, Jacek Jasieniak, and Paul Mulvaney Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03885 • Publication Date (Web): 02 Jan 2018 Downloaded from http://pubs.acs.org on January 3, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Aqueous Synthesis of High-Quality Cu2ZnSnS4 Nanocrystals and their Thermal Annealing Characteristics Cameron Ritchiea,b, Anthony S.R. Chesmanc, Mark Stylesc, Jacek J. Jasieniak*,b, and Paul Mulvaney*,a a

ARC Centre of Excellence in Exciton Science, School of Chemistry and Bio21 Institute, University of Melbourne, 30 Flemington Road, Parkville, Victoria 3010, Australia, b ARC Centre of Excellence in Exciton Science, Materials Science and Engineering, and Energy Materials & Systems Institute, Monash University, 20 Research Way, Clayton, Victoria 3800, Australia, c CSIRO Manufacturing, Ian Wark Laboratories, Bayview Avenue, Clayton, Victoria 3168, Australia. ABSTRACT: Copper zinc tin sulfide (CZTS) nanocrystal inks are promising candidates for the development of cheap, efficient, scalable and non-toxic photovoltaic (PV) devices. However, optimisation of the synthetic chemistry to achieve these goals remains a key challenge. Herein we describe a single-step, aqueous-based synthesis that yields high quality CZTS nanocrystal inks, while also minimizing residual organic impurities. By exploiting simultaneous redox and crystal formation reactions, square platelet-like CZTS nanocrystals stabilized by Sn2S64− and thiourea are produced. The CZTS synthesis is optimised by using a combination of inductively coupled plasma analysis, Raman spectroscopy, Fourier transform infrared spectroscopy and synchrotron powder x-ray diffraction to assess the versatility of the synthesis and identify suitable composition ranges for achieving phase-pure CZTS. It is found that mild heat treatment between 185-220 °C is most suitable to achieve this, because this temperature range is sufficiently high to thermalize existing ligands and ink additives, while minimising tin loss, which is problematic at higher temperatures. The low temperatures required to process these nanocrystal inks to give CZTS thin-films are readily amenable to production scale processes.

1.0. INTRODUCTION The quaternary I2-II-IV-VI4 compound copper zinc tin sulfide (CZTS) is regarded as a promising photovoltaic material due to its excellent optoelectronic properties, tuneable bandgap, and earth abundant elemental composition.1 Various methods have been reported for the deposition of CZTS thin films, including decomposition of molten salts,2-3 reactive sputtering,4-5 electroplating,6-7 vapour deposition,8-9 precursor solution deposition,10-13 and nanoparticle ink sintering.14-40 Of these methods, the approach using CZTS nanoparticle inks is one of the most promising, commonly achieving efficiencies over 8.5%,14, 20-25 with the highest efficiency nanoparticlebased CZTS solar cells reported to date (11.1%) being fabricated using a hydrazine-based synthetic route.21 This success is largely due to the better phase and compositional control achieved through solution processing, compared to the alternative synthetic approaches. However, synthetic routes for the production of high quality CZTS nanoparticles typically involve the use of toxic solvents and extensive post-synthesis processing to remove organic impurities.16-18, 25, 34-37, 40 These are clearly undesirable when considering that reproducibility and scale-up are essential for commercial production, which requires the minimization of hazards and additional processing steps. While existing “greener” CZTS nanoparticle syntheses based on water and ethanol dispersions overcome the risks associated with using toxic hydrazine,14, 24, 31-33, 38-39 eliminating organic contaminants, which are usually present as an excess of aliphatic ligands used to achieve homogeneous dispersions of nanocrystals, remains a challenge. Furthermore, incomplete combustion

of these organic ligands also forms insulating layers at interfaces following thermal annealing,15 generating trap states that enhance carrier recombination.1 Both of these effects reduce exciton diffusion lengths and the concentration of charge carriers,41 resulting in lower device performance. As ligands are required to stabilize colloidal nanoparticles, removing them or eliminating them entirely while maintaining stability in solution is impossible. A viable alternative is to replace these non-volatile, long-chained aliphatic ligands with short-chained ligands that will readily vaporise during annealing.42 Another option is to remove carbon containing stabilizing agents completely, and use highly charged, molecular, metal chalcogenide ligands.19, 43 By employing the latter approach, Tang et al.19 developed a method for fabricating CZTS thin-films from aqueous solutions containing alloyed CuS/ZnS nanoparticles coated with a tin metal chalcogenide. This organic ligand-free approach yielded overall PV efficiencies above 5%, despite not using phase pure CZTS nanoparticles. Careful control of the processing parameters during the deposition of a CZTS ink is critical to successful device fabrication. Ink composition and the effects of thermal annealing on elemental composition, grain growth, and phase purity are of great importance for the production of high-quality thin films.1 Copper sulfide is often formed as a by-product during CZTS nanoparticle synthesis as it has the lowest solubility product of all the phases,  = 2.7 × 10   at 25 °C, more than 20 orders of magnitude less than zinc sulfide.44-45 However, due to the relatively high atomic diffusivity of lattice Cu+ ions, Cu+ quickly diffuses into the CZTS phase during annealing.46 Tin is the least stable element in CZTS during

1 Environment ACS Paragon Plus

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

annealing, with the formation of SnS above 300 °C and the significant vapour pressure of tin sulfide species above 400 °C detrimentally affecting the interfacial properties and overall performance of CZTS PV devices.47 SnS loss can be reduced if high temperature annealing is performed in a sulfur or selenium containing atmosphere, and can be eliminated completely if controlled amounts of SnS vapour are also present.48-51 A sulfur, selenium, tin sulfide, or tin selenide containing atmosphere also improves the growth kinetics of crystal grains in CZTS, as the chalcogenide vapour assists in the dissolution and reformation of the anion-based lattice.42, 47-50, 52-56 Understanding how to enhance crystallite growth while maintaining the phase purity of CZTS is essential to minimizing point defects, dislocations, and stacking faults in the CZTS lattice, as these defects have been shown to shorten charge carrier lifetimes.57-58 At present there is no CZTS nanoparticle synthesis that simultaneously provides stoichiometric compositional control and phase purity using a low toxicity solvent, while also yielding negligible residual carbon impurities following thermal annealing. Herein, we demonstrate a versatile, stoichiometric, one-pot synthetic method that produces aqueous, high quality CZTS nanocrystals without the use of long-chained organic ligands. These nanocrystals are found to exhibit a unique, square platelet morphology. Our synthetic strategy is based on simultaneous reduction and crystal formation reactions and is eminently scalable. In preparation of their use in the fabrication of high-efficiency PV devices, we also investigate the annealing properties, phase purity and grain growth characteristics for our CZTS nanocrystals under inert atmosphere.

Page 2 of 12

container in air before resealing to prevent the loss of ammonia during the reaction. The original mustard-coloured, turbid mixture was allowed to react with stirring until no more unreacted tin was present (~ 45 minutes), with the temperature maintained at 40 °C using a water bath. This yielded a 0.36 M Sn2S64− transparent solution (2.5 mmol in 13% ammoniumsulfide water, 6.94 mL) with a deep-orange colour. Dissociative route: Tin (IV) sulfide powder (5 mmol, 0.914 g) was added to ammonium-sulfide (4.44 mL) and Milli-Q water (2.5 mL), then the reaction container was sealed to prevent the loss of ammonia during the reaction. The original mustard-coloured, turbid mixture was allowed to react with stirring until no more solid was present (~ 4 minutes), with the temperature maintained at 40 °C using a water bath. This yielded a 0.36 M Sn2S64− solution comparable to that obtained via the redox route. The aqueous Sn-MCC solutions cannot be stored for long periods due to a slow degradation reaction that converts the soluble Sn-MCC to a black precipitate of tin oxide. The rate of degradation is significantly slower for Sn−MCC prepared via the dissociation route (6 days for 1 mg of precipitate) compared to Sn−MCC synthesised via the redox route (3 hours for 1 mg of precipitate). No other differences during the synthesis or in the resulting nanocrystals were observed.

2.0. EXPERIMENTAL SECTION 2.1. Materials Tin powder (99%), copper nitrate hemi-pentahydrate (≥98%), ammonia (25% in water), thiourea (≥99%), and ammonium sulfide (20% in water) were purchased from Sigma Aldrich. All of the above reagents and solvents were used as received. The tin powder was stored in a glove box under a nitrogen atmosphere to prevent oxidation. Deionized water was obtained from a Milli-Q system (18.2 MΩ.cm resistivity). Reaction containers were polypropylene specimen containers with polypropylene screw caps purchased from Techno Plas. Note that CZTS nanocrystals were found to adsorb strongly onto glass reagent vessels. Tin (IV) sulfide was synthesized via reaction of sodium sulfide nonahydrate (99.99%, Sigma Aldrich) and tin (IV) chloride pentahydrate (98%, Sigma Aldrich) in water. Tin (IV) sulfide was purified by vacuum filtration and washing with Milli-Q water.

2.2. Preparation of Tin Metal Chalcogenide There are two possible routes to form identical tin metal chalcogenide (Sn-MCC) solutions: (i) a redox route using pure tin and sulfur powders and (ii) a dissociative route using tin (IV) sulfide. These are shown schematically in Figure 1Figure 1 a & b. Both were trialled in this study. Redox route: Tin powder (5 mmol, 0.594 g) and sulfur powder (10 mmol, 0.321 g) were added to a reaction container and sealed under a nitrogen atmosphere to prevent the premature oxidation of tin. Ammonium-sulfide (4.44 mL) and then Milli-Q water (2.5 mL) were quickly added to the reaction

Figure 1. a) Metallic tin and sulfur redox reaction route to SnMCC. b) Tin (IV) sulfide dissociative route to Sn-MCC. c) A simplified process diagram for aqueous CZTS nanoink preparation.

2.3. Synthesis of CZTS Nanocrystals A typical synthesis of CZTS nanocrystals with an elemental ratio of 2:1:1 Cu:Zn:Sn and total metal concentration of 40 g/L is as follows: a 1 M thiourea solution (12 mmol in 12 mL of Milli-Q water) and a 0.36 M Sn-MCC solution (2.5 mmol in 6.94 mL of 13% ammonium-sulfide water) were added to Milli-Q water (34 mL, to adjust the final ink concentration to 40 g/L) at 40 °C with stirring. A 0.76 M Zn(NO3)2 solution (5 mmol in 6.54 mL of 18% ammonia solution) was then added, causing the formation of a pale yellow precipitate, which redissolved within 2 minutes to return the solution to its original transparent orange appearance. Once the solution was completely transparent, a 3.02 M Cu(NO3)2 solution (10 mmol in 3.31 mL of Milli-Q water) was added quickly with stirring, causing the immediate formation of a dark red/black CZTS nanocrystal ink. The yield by mass was ≥ 95% (2.4 g of dried

2 Environment ACS Paragon Plus

Page 3 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

CZTS nanocrystals). This synthetic method was successfully scaled-up to produce 50 g of dried CZTS nanocrystal powder. CZTS nanocrystals of different elemental ratios and concentrations can be synthesised stoichiometrically by the addition of different amounts of copper, zinc, and tin precursor solutions while keeping a ~1.2-fold excess of thiourea to copper ions in solution. This synthesis can be varied to make any of the nanocrystal inks discussed later in this work at concentrations of up to ~80 g/L.

2.4. Characterisation As-synthesised CZTS ink powders were obtained by vacuum drying at 40 °C. Purified CZTS nanocrystal powders were obtained by adding ethanol or isopropanol as an antisolvent, centrifuging at > 2100 rcf (~5 mins), disposing of the supernatant and redispersing the precipitant in Milli-Q water. This process was repeated 2-3 times and finally the precipitant was vacuum dried at 40 °C. Transmission electron microscopy (TEM) images were recorded on a FEI TF 20 TEM operated with an acceleration voltage of 200 kV. Purified 2:1:1 CZTS nanocrystals were prepared as described above, and redispersed with Milli-Q water (0.1 g/L) for TEM analysis. The dilute nanocrystal dispersions (10 µL) were drop cast onto parafilm. A carboncoated copper grid (400 mesh) pretreated with a dilute poly(dimethyldiallyl ammonium chloride) (PDMAC) solution to create a monolayer of cationic polymer was placed on top for two minutes then removed and allowed to dry. Scanning electron microscopy (SEM) images were recorded on a FEI Nova Dual Beam FIB SEM operated with an acceleration voltage of 20 kV, current of 2.4 nA, and magnification of 100000×. 2:1:1 CZTS nanocrystals films were prepared from purified CZTS nanocrystals redispersed in Milli-Q water (20 g/L), and drop cast on plasma cleaned silicon wafers. These films were annealed at 400 °C, 500 °C, 600 °C and 700 °C for 5 minutes in a tube furnace under a nitrogen atmosphere. Samples for AFM were prepared by spin-coating purified 2:1:1 CZTS nanocrystals redispersed in Milli-Q water (0.5 g/L) onto a silicon wafer. AFM measurements were performed on an Asylum Research MFP-3D-Bio. The sample was scanned in tapping mode with a Tap300Al−G cantilever with a tip radius of ~9 nm. Raman spectroscopy was performed using a Renishaw InVia-Raman spectrometer using a λ = 514 nm laser source and Renishaw RM 1000 Raman spectrometer using a λ = 325 nm laser in a range from 100 to 720 cm−1 and the peaks were fitted using Origin commercial software. Fourier transform infrared spectra were collected with a Perkin Elmer Spectrum One FTIR spectrometer in a range from 650 to 3600 cm−1. Tandem thermogravimetric analysis combined with Fourier transform infrared spectroscopy (TGA-FTIR) was conducted on a Netzsch TG 209 F1 Libra instrument in tandem with a STA 449 F1 Jupiter-FT-IR Coupling system. The CZTS ink powder used for this analysis was synthesized at a concentration of 20 g/L with a Cu:Zn:Sn ratio of 2:1:1. The heating rate was 2.5 °C/minute with mass measured every 0.6 seconds and FTIR spectra acquired every 22 seconds. The original CZTS mass was 86 mg and a nitrogen flow rate of 20 mL/min was used to minimize oxygen in the system. FTIR peaks were analysed based on reference spectra from the National Institute of

Standards and Technology (NIST) database where no publications are cited. Temperature dependent powder x-ray diffraction (PXRD) analysis was carried out on the PD Beamline at the Australian Synchrotron. A variable temperature ramp was applied using a heated blower with temperature sensor. The PXRD measurements were collected at a wavelength of λ = 0.8235 Å. The CZTS nanocrystal powders were prepared as described above and finally loaded into a capillary, which was placed under an over-pressure nitrogen atmosphere to ensure no oxygen entered the system. Each XRD pattern was obtained using an acquisition time of ~85 seconds and a heating rate of 20 °C/minute. Extra patterns were collected at constant temperatures of 26 °C and 600 °C to confirm that the samples were fully equilibrated throughout the temperature ramping process. The mean temperature during the acquisition time of each XRD pattern is quoted as the temperature value in Figure 7. For analysis, individual P1 and P2 data sets were merged using CONVAS259 to remove detector gaps and the merged data sets were analysed using Rietveld refinement-based quantitative phase analysis60-61 implemented in the launch mode of TOPAS (Bruker AXS). Crystal phase wt% values were calculated via the Hill and Howard algorithm.62 An anisotropic crystallite size broadening function, adapted by Coelho63 from March-Dollase,64 was used to improve the quality of the fit by taking into account the crystallite size and micro-strain. Inductively coupled plasma elemental analysis (ICP−MS) was performed using a GBC Optimass 9500 ICP−TOF−MS. For ICP−MS a portion of CZTS nanocrystal powder was digested in HCl and diluted to 2% HCl in Milli-Q water to achieve ~250 ppb levels for analysis. A single calibration standard solution was purchased from Choice Analytical of 100 ppm of Au, Cu, Zn, Sn, S, and Se was diluted to multiple concentrations between 10 and 200 ppb in order to produce a calibration curve for each of the element. Hartree-Fock geometry optimizations were performed using a basis set of 6−31G in the software package ChemBio3D 13.0 utilizing the GAMESS interface.

3.0. RESULTS AND DISCUSSION 3.1. Synthesis, Morphology and Surface Chemistry of CZTS Nanocrystals We build on the work of Tang et al.19 and follow four design principles in our synthesis of high quality CZTS nanocrystals: (1) the materials and processing must be of low cost; (2) green and non-toxic materials and techniques must be used; (3) all techniques employed must be scalable; and (4) impurities, such as carbon, nitrogen and oxygen, must be minimized. The first three principles are evident in the selection of precursors, the use of low temperature and simple processing techniques, and the ability to scale up to ≥ 50 g batches of dried nanocrystal powder. In order to minimize carbon, nitrogen, and oxygen impurities in the CZTS crystals, while creating high quality nanocrystals, careful consideration of the precursors, their reaction mechanisms, and their relative concentrations is required. Furthermore, a low-carbon colloidal stabilizer that works at high particle concentrations and high ionic strength is needed. A tin metal chalcogenide was chosen as an organic-free alternative in order to colloidally stabilize the CZTS nanocrystals in aqueous solutions via electrostatic repulsion.43 Formation of metal hydroxides and metal oxides, which would

3 Environment ACS Paragon Plus

Langmuir 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

usually form in aqueous solution at high pH, were suppressed by the addition of excess HS−. Thiourea acts as both a mild reducing agent to convert Cu(II) to Cu(I)65 and a potential source of excess HS− in the solution. Although thiourea may also act as a source of carbon and nitrogen impurities, in aqueous solutions thiourea decomposes above ~150 °C to cyanamide and isothiocyanic acid, which further decomposes into ammonia and carbonyl sulfide that are immediately released as gases.66 The use of the nitrate salts of copper and zinc was adopted to obviate the contamination of CZTS by chloride ions and organic species, which are present in most CZTS syntheses and are difficult to remove during post-processing. Nitrate ions can be decomposed into nitric dioxide gas and water in the presence of an excess of ammonia above ~180 °C.67 The addition of ammonia fulfils the requirement of a basic solution to decompose nitrates, while also being easily removed at mild temperatures. In order to form CZTS nanocrystals and avoid premature formation of CuS, the simultaneous reaction of thiourea and sulfur with all metal ions is essential. The simultaneous reduction reaction of Cu2+ with thiourea to form Cu+, and crystal formation reaction which occurs while copper ions are in the presence of HS− and the metal sources Zn2+ and Sn2S64−, results in the formation of singlephase sphalerite CZTS nanocrystals, rather than alloys consisting of multiple secondary phases covered by tin metal chalcogenides. This approach overcomes the typical difficulties of scalability and commercialisation by using cheap, non-toxic precursors, while eliminating the need for post processing to remove impurities and organic ligands.

Page 4 of 12

hypothesis, AFM was used to measure the nanocrystal thickness as seen in Figure 3.

Figure 3. a) 5 µm × 5 µm low resolution tapping-mode AFM image of CZTS nanocrystals with an elemental ratio of 2:1:1 Cu:Zn:Sn spin coated onto a silicon wafer. b) High resolution AFM image of the green square area of a) showing a topographic image of a single CZTS nanocrystal; c) AFM cross section of the single nanocrystal in b).

Figure 3 a) shows a low-resolution, tapping-mode AFM measurement of discrete CZTS nanocrystals on a silicon wafer. These revealed uniform nanocrystals with little variation in width or height. A high resolution AFM image of a single CZTS nanocrystal is presented in Figure 3 b). 30 nanocrystals in total were imaged, with a typical cross-section shown in Figure 3 c). All 30 nanocrystals measured were consistent with the hypothesis that the CZTS nanocrystals are square platelets with a mean height of 2.5 nm ± 0.8 nm. It should be noted that due to AFM tip convolution effects, the lateral dimensions are overestimated.68

Figure 2. a) TEM image of CZTS nanocrystals with an elemental ratio of 2:1:1 Cu:Zn:Sn produced at an original concentration of 20 g/L showing relatively monodisperse square nanocrystals. Left Inset: histogram indicating a mean size of 13.4 nm b) High resolution TEM image showing the presence of crystal lattice spacings of 0.31 nm and 0.54 nm (green), and an inter-nanocrystal spacing of ~0.8 nm (red). Right Inset: The corresponding FFT image.

TEM was carried out in order to identify the size distribution and morphology of the nanocrystals produced. Figure 2 a) shows monodisperse cubic or square CZTS nanocrystals with an average width of 13.4 nm and a standard deviation of 1.7 nm. The high resolution TEM (HR-TEM) image shown in Figure 2 b) clearly reveals the presence of crystal lattice spacings of ~ 0.31 nm and ~ 0.54 nm, which are consistent with the inter-planar d112 spacing of sphalerite and kesterite CZTS phases, and the lattice constant a, respectively.26 The HR-TEM image also shows two seemingly overlapping CZTS nanocrystals in the bottom left corner, which suggests that the nanocrystals may be platelets rather than cube shaped. If the nanocrystals are indeed very thin, we would expect all of the nanocrystals to be oriented flat on the surface of the TEM grid so the shorter axis would not be observable. In order to test this

Figure 4. a) FTIR spectra showing the functional groups present in the dried CZTS nanocrystal ink (black), twice washed and dried CZTS nanocrystals (blue), and twice washed and annealed at 250 °C CZTS nanocrystals (red). b) and c) Optimized geometries of Sn2S64− anions and thiourea respectively obtained from HartreeFock simulations with a 6–31G basis set.

4 Environment ACS Paragon Plus

Page 5 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

A constant intercrystal spacing of ~0.8 nm is observed in Figure 2 b). This may be attributed to the steric repulsions between surface bound molecules on the nanocrystals. In order to confirm this, we identified the surface bound molecules by FTIR in Figure 4 a). In the unwashed ink, all of the expected functional groups stemming from the original reaction chemistry were present, except the C=S vibration from thiourea at 1140 cm−1. A C–S peak at 700 cm−1 is still observed, which implies that the thiourea could be bound to the surface of the nanocrystals by a sulfur bond as C−S−CZTS. After purifying the nanocrystals, only the surface bound groups S−Sn, N−H and C−S were present, corresponding to Sn2S64− and thiourea, respectively. Computational geometry optimization was performed to calculate whether these surface molecules could be responsible for the observed 0.8 nm spacing between nanocrystals in Figure 4 b) & c). The calculations predict that both molecules have mean lengths ~0.4 nm, which is consistent with a 0.8 nm intercrystal spacing arising from one layer of either ligand on the surface of each adjacent nanocrystal. The final CZTS surface chemistry probably constitutes both Sn−MCC anions and thiourea molecules. The former confers some negative charge on the nanocrystals, while the thiourea probably provides colloidal stability via hydrogen bonding to the solvent.

3.2. Crystal Phase Quality of CZTS Nanocrystals To determine the CZTS crystal phase we utilized synchrotron PXRD and Raman spectroscopy. CZTS nanocrystals were prepared as described in the experimental section, washed twice with ethanol, and dried at 40 °C under vacuum to give a dry nanocrystal powder. Raman spectra of the crystals were collected (i) as-prepared and (ii) after annealing at 300 °C for 20 minutes in a nitrogen filled tube furnace (Figure 5 a) & b), respectively). The experimental data for 2:1:1 CZTS nanocrystals are overlaid by the fits to the characteristic CZTS Raman peaks1. The results suggest the formation of a pure CZTS phase at both temperatures, however ternary phases which overlap the CZTS raman scattering peaks could also be present.

indicated by the green arrow highlights a portion of the PXRD patterns, which is not fit well by a model of CZTS with 2 nm crystal size. It exhibits a much sharper {220}/{204} peak corresponding to a larger crystallite size (~12 nm) and a broad peak underneath.

The two temperatures, 25 °C and 300 °C, were selected from temperature dependent XRD measurements to match the Raman spectra, and also to give an example of the quality of the XRD phase fitting analysis. Further details for the PXRD analysis can be found in the Supporting Information. The experimental PXRD patterns and their corresponding phase fits are shown in Figure 5 c) & d) in black and red, respectively. The 25 °C sample PXRD pattern in Figure 5 c) shows broad diffraction peaks, which could be identified as any of a number of phases. If we assume pure CZTS, as suggested by the corresponding Raman spectra, fitting of the diffraction pattern predicts crystallites of ~2 nm size due to the reduced periodicity of individual crystal domains. The insert in Figure 5 c) shows a region where the model fit does not accurately follow the experimental patterns. A sharp peak corresponding to the {220} or {204} peak superimposed over two, merged, broad peaks is highlighted by the green circle and arrow. The sharp peak is well fitted by a model with a crystal size of 12 nm, while the broad peak underneath is fit well by a model with a 1.5 nm crystal size. The model for the fit in each case assumes equal dimensions in all directions; however, we have confirmed from AFM that this is not the case for these nanocrystals. It is likely that other peaks are also significantly sharper than expected, but due the intensity of the {220}/{204} peak, this is the easiest to identify. In contrast, the diffraction pattern of the sample treated at 300 °C in Figure 5 d) exhibits welldefined peaks, with good agreement to a fit that assumes a 100% sphalerite CZTS phase with a crystallite size of 12 nm. These results suggest that the sphalerite CZTS nanocrystals grow from 2 nm to 12 nm between 25 °C and 300 °C without any intervening phase transitions.

3.3. Phase Map of Annealed CompositionallyVariant CZTS Nanocrystals In order to create a compositionally-dependent crystal phase map, 18 different compositions of CZTS nanocrystals were analysed by both synchrotron PXRD and Raman spectroscopy. Samples were prepared with theoretical elemental ratios as shown in Figure 6, at a concentration of 20 g/L, washed twice with ethanol, and dried at 40 °C under vacuum to give a dry nanocrystal powder. In order to confirm that the CZTS nanocrystal synthesis is stoichiometric, we performed ICP−MS on the precipitates to compare the theoretical and experimental elemental ratios and the results are presented in Figure 6. The values were found to be within measurement and pipetting errors (