This document is the Accepted Manuscript version of a Published Work that appeared in final form in Chemistry of Materials, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see: https://doi.org/10.1021/acs.chemmater.5b03417 Page 1 of 8 Chemistry of Materials 1 2 3 4 5 6 7 Scalable heating-up synthesis of monodisperse Cu2ZnSnS4 8 9 nanocrystals 10 11 Alexey Shavel,†,ћ,* Maria Ibáñez,† Zhishan Luo,† Jonathan De Roo,¥ Alex Carrete,† Mirjana 12 Dimitrievska,† Aziz Genç,¶ Michaela Meyns,† Alejandro Perez-Rodriguez,†,‡ Maksym V. 13 Kovalenko,┴,ǁ Jordi. Arbiol,¶,§ and Andreu Cabot,†, § 14 † 15 Catalonia Energy Research Institute - IREC, Sant Adrià del Besòs, Barcelona, 08930, Spain. 16 ¶ Institut Català de Nanociència i Nanotecnologia, ICN2, Campus de la UAB, 08193 Bellaterra, Spain. 17 ‡ Departament d’Electrònica, Universitat de Barcelona, 08028 Barcelona, Spain. 18 § 19 Institució Catalana de Recerca i Estudis Avançats - ICREA, 08010 Barcelona, Spain. 20 ¥ Ghent University, Department of Inorganic and Physical Chemistry, Krijgslaan 281 S3, 9000 Gent, Belgium. 21 ┴ Institute of Inorganic Chemistry, Department of Chemistry and Applied Biosciences, ETH Zürich, CH-8093, 22 23 Switzerland. ǁ 24 Empa-Swiss Federal Laboratories for Materials Science and Technology, Dübendorf, CH-8600, Switzerland. 25 26 27 28 ABSTRACT: Monodisperse Cu2ZnSnS4 (CZTS) nanocrystals, with quasi-spherical shape, were prepared by a facile, high 29 yield, scalable and high concentration heating-up procedure. The key parameters to minimize the nanocrystal size 30 distribution were efficient mixing and heat transfer in the reaction mixture through intensive argon bubbling and 31 improved control of the heating ramp stability. Optimized synthetic conditions allowed the production of several grams 32 of highly monodisperse CZTS nanocrystals per batch, with up to 5wt% concentration in a crude solution and a yield above 33 90 %. 34 35 36 37 1. INTRODUCTION spaces, because it competes with simultaneous formation 38 Ternary and quaternary semiconductors have recently of binary and ternary phases, all having similar energies of 39 attracted immense interest in photovoltaics, formation and similar crystal structures. Of course, 40 photocatalysis and thermoelectrics.1-11 Among numerous diversity of phases and compositions in multicomponent 41 chalcogenides, Cu2ZnSnS4 (CZTS), comprising abundant systems offers also important and useful advantages. For 42 and non-toxic elements, has particularly appealing instance, Cu-poor and Zn-rich CZTS materials have 43 characteristics such as high absorption coefficient in the demonstrated better performances as absorbers in solar 44 visible region (>104 cm-1) and a suitable band gap (1.45-1.50 cells than nearly stoichiometric materials.2 On contrary, 45 eV) for solar energy conversion. These properties make it Cu-rich CZTS have shown improved photocatalytic 46 an excellent absorber material not only for the next properties.6 Furthermore, the compositional control 47 generation of solar cells,1 but also for photocatalysts.6-9 allows tuning of the semiconductor charge carrier 48 concentration, which is key to optimize performance in The availability of these quaternary semiconductors in 49 applications such as thermoelectrics.15 the form of nanocrystal-based inks allows low-cost 50 solution processing by common techniques such as spray- Control and analysis of compositional and phase 51 coating or inkjet printing. Inexpensive and scalable homogeneities are experimentally challenging and 52 bottom-up approaches are increasingly needed for require, for instance, scanning transmission electron 53 transferring the compositional versatility and microscopy (STEM) combined with high spatial 54 homogeneity of individual nanocrystals to the device- resolution chemical analysis with energy-dispersive X-ray 55 level macroscale (large area thin films or bulk–like spectroscopy (EDX), electron energy loss spectroscopy 56 nanogranular materials).12-14 (EELS) or Raman spectroscopy to unravel the complexity 57 within the individual NC. To obtain NC-to-NC variation The preparation of multinary nanocrystals has proven 58 in atomistic uniformity, one would need to map to be much more difficult than the synthesis of elemental 59 thousands of NCs, which is difficult experimentally; while and binary nanocrystals. In particular, formation of 60 other characterization techniques provide only average quaternary chalcogenides requires narrow parametric information. In this regard, very useful is observation that ACS Paragon Plus Environment Chemistry of Materials Page 2 of 8 shape and size-uniformity of multinary nanocrystals are hour to ensure the removal of traces of low boiling point 1 indicative of their intrinsic chemical homogeneity.16-19 In impurities and water. Three issues to be noted at this 2 other words, monodisperse nanocrystals are more likely point: i) the purification time may be significantly 3 to be also chemically identical/similar. Since photovoltaic, optimized in a large scale production scenario; ii) the 4 catalytic and nearly all other applications require large main reason to use THF as co-solvent is to facilitate the 5 quantities of nanocrystals with a minimal particle-to- metal salts dissolution, but it also helped to remove water 6 particle variation in composition, the aim of this work from the reaction mixture; iii) ZnO is soluble only with 7 was to develop a high yield and easily scalable synthesis the presence of SnCl4·5H2O in a solution. 8 protocol to produce CZTS nanocrystals, with high After purging, the mixture was cooled to 100 °C and 50 9 compositional and morphological uniformities. mmol (10.12 g) of tert-dodecanethiol and 5 mmol (1.012 g) 10 The ‘heating-up’ synthesis – when all reagents and of dodecanethiol were added through a syringe. Then the 11 solvents are pre-mixed and then simply heated - is solution was heated to 250 °C during 10 minutes and 12 generally easier to upscale than the injection-based maintained at this temperature for 1 hour. A constant 13 procedures, primarily due to highly homogenous heating rate and a high argon flow during the heating 14 concentrations and temperatures throughout the were key be obtain monodisperse nanocrystals. 15 process.20 When the injection of a reagent is involved, Formation of nanocrystals was noticed by a gradual 16 upscaling is limited by difficulties to achieve fast mixing darkening of the reaction mixture at about 135-145ºC. 17 of large volumes of reagents. Few gram-scale synthesis Final CZTS nanocrystals were thoroughly purified by 18 protocols to produce CZTS nanocrystals have been multiple precipitation and re-dispersion steps using 2- 19 already proposed. Among them, the solvent-free, propanol and chloroform. Finally, CZTS nanocrystals 20 mechanochemical synthesis approach is a remarkably were dissolved in THF and the solution was centrifuged at 21 simple pathway,21 albeit with limited control over the 9000 rpm during 5-15 minutes to precipitate poorly 22 composition and morphology of CZTS nanocrystals. soluble unreacted metal complexes and large (> 100 nm) 23 Zn-rich particles (note the initial non-stoichiometric Here we report a 90%-yield synthesis of monodisperse 24 precursor ratio). The supernatant was recovered and CZTS nanocrystals using a heating-up procedure and high 25 stored for its posterior use. No additional size selective initial concentration of precursors, up to 0.6 mmol per 26 precipitation was performed. The proposed synthetic gram of a reaction mixture, that is a factor of 8-10 higher 27 route allowed to prepare about 670 mg of CZTS than commonly used in nanocrystal synthesis. We also 28 nanocrystals in 30 g of reaction mixture, and with a present a detailed nuclear magnetic resonance (NMR) 29 precursor metal concentration of 0.3 mmol/g. It is worth study of the surface capping ligands and an efficient 30 mentioning that total precursors concentration might be organic ligand displacement procedure, with final carbon 31 doubled simply by reducing the amount of solvent. content of just 0.1-0.2 wt%. 32 However, by doing so the viscosity of the reaction mixture 33 2. EXPERIMENTAL SECTION increased significantly, reducing the product quality and 34 Chemicals and solvents: Tert-dodecanethiol (tDDT, rendering the synthetic procedure less convenient. 35 98.5%), 1-dodecanethiol (DDT, ≥98%), tin (IV) chloride Instead, we propose to proportionally scale up the 36 pentahydrate (SnCl4·5H2O, 98%), zinc oxide (ZnO, synthesis a factor 10 by conducting the synthesis in a 1 37 ≥99%), copper (II) chloride dihydrate (CuCl2·2H2O, liter flask. In fact, we regularly performed this synthesis 38 ≥99%), 1-octadecene (ODE, 90%), and oleylamine (OLA, protocol at 5-7 fold scale, producing 5-6 g of CZTS 39 70%) were purchased from Aldrich. Chloroform, nanocrystals per batch with essentially the same results. 40 isopropanol, dimethyl sulfoxide (DMSO) and The typical yield of the synthesis was about 90%. 41 tetrahydrofuran (THF) were of analytical grade and Characterization techniques: The morphological, 42 obtained from various sources. OLA was vacuum distilled chemical and structural characterizations of the 43 before its use. The use of ‘as received’ OLA results in poor nanocrystals were carried out by transmission electron 44 nanocrystal size distributions and irregular nanocrystal microscopy (TEM) and high-resolution TEM (HRTEM). 45 shapes. Others precursors and solvents were used without Carbon-coated TEM grids from Ted-Pella were used as 46 further purification. Cole-Parmer ‘Digi-Sense R/S’ substrates. A solution of the CZTS nanocrystals in THF 47 temperature controllers with enabled ‘auto-tuning’ mode was directly dropped on the grids and dried at the 48 were used for the accurate control of the heating rate ambient condition. HRTEM images were obtained using a 49 during the synthesis. FEI Tecnai F20 field-emission gun microscope operated at 50 CZTS nanocrystals: Quasi-spherical CZTS 200 keV with an embedded Gatan image filter for EELS 51 nanocrystals were prepared by the reaction of copper and analyses. Images were analyzed by means of Quantum 52 tin chlorides and zinc oxide with a mixture of tert- Gatan micrograph software. Powder X-ray diffraction 53 dodecanethiol and dodecanethiol in the presence of OLA. (XRD) patterns were obtained with Cu Kα1 (λ = 1.5406 Å) 54 In a typical synthesis, 5.4 mmol (920 mg) of CuCl2·2H2O, radiation in a reflection geometry on a Bruker D8 55 4.8 mmol (391 mg) of ZnO and 1.8 mmol (630 mg) of operating at 40 kV and 40 mA. Raman scattering 56 SnCl4·5H2O were dissolved in a minimal amount of THF. measurements were performed in backscattering 57 Afterward, 24 mmol (6.43 g) of distilled OLA and 20 g of configuration with a LabRam HR800-UV Horiba-Jobin 58 ODE were added to the reaction mixture. The solution Yvon spectrometer, using the 532 nm line from a Nd:YAG 59 was heated under argon flow (ca. 30 mL/min) to 175 °C solid state laser and the 325.0 nm line from a HeNe laser 60 and maintained at this temperature for approximately 1 as excitation lights. UV-vis optical absorption spectra ACS Paragon Plus Environment Page 3 of 8 Chemistry of Materials were recorded on a LAMBDA 950 UV-Vis software of Bruker. The diffusion coefficients were 1 spectrophotometer from PerkinElmer. Elemental analyses obtained by fitting the appropriate Stejskal-Tanner 2 were performed using an elemental organic analyzer equation to the signal intensity decay.25 3 Thermo EA 1108, working under 120 ml/min helium flow, 3. RESULTS AND DISCUSSION 4 the combustion furnace at 1000 ºC, the chromatographic 5 Representative TEM and HRTEM micrographs of the column oven at 60 ºC, and 10 ml oxygen loop at 100 kPa. 6 CZTS nanocrystals prepared following the above detailed Nuclear Magnetic Resonance (NMR) measurements heating-up procedure are shown on figures 1a and 1b 7 were recorded on a Bruker Avance III HD Spectrometer respectively. CZTS nanocrystals have slightly faceted, 8 operating at a 1H frequency of 500.26 MHz and equipped nearly isotropic spherical shapes. Importantly, the size 9 with a BBFO-Z probe. The sample temperature was set to dispersion was below 10 % (estimated as full-width at half 10 298.2 K. One dimensional (1D) 1H and 2D NOESY 11 maximum of the size histogram collected from several (Nuclear Overhauser Effect Spectroscopy) spectra were hundred NCs). The power spectrum, shown in figure 1b, 12 acquired using standard pulse sequences from the Bruker 13 unambiguously revealed CZTS nanocrystals with a library. For the quantitative 1D 1H measurements, 64k wurtzite phase and lattice parameters; a = b = 0.3839 nm 14 data points were sampled with the spectral width set to 15 and c = 0.6339 nm (P63mc space group). XRD analysis 20 ppm and a relaxation delay of 30 sec. The NOESY confirmed the wurtzite crystal phase of the CZTS 16 mixing time was set to 300 ms and 4096 data points in the 17 nanocrystals (figure 1c). The presence of alkylthiols in the direct dimension for 512 data points in the indirect reaction mixture was found to be crucial to obtain CZTS 18 dimension were sampled, with the spectral width set to 12 19 nanocrystals of this crystal structure. Contrarily, sulfur as ppm. Diffusion measurements (2D DOSY) were a precursor causes the formation of nanoparticles with 20 performed using a double stimulated echo sequence for 21 kesterite structure.30 Raman spectra of the produced convection compensation and with monopolar gradient materials are shown in figure 1d. Various excitation 22 pulses.24 Smoothed rectangle gradient pulse shapes were 23 wavelengths were used to enhance the signal of the used throughout. The gradient strength was varied possible secondary and ternary impurity phases. The 24 quadratically from 2-95 % of the probe’s maximum value Raman spectrum obtained with a 532 nm excitation shows 25 in 64 increments, with the gradient pulse duration and two broad bands at 333 cm-1 and 288 cm-1, which agree 26 diffusion delay optimized to ensure a final attenuation of well with the reported spectra of wurtzite CZTS. The 27 the signal in the final increment of less than 10 % relative presence of CuS, SnS, SnS2, SnS3 or Cu2SnS3 would have 28 to the first increment. For 2D processing, the spectra were associated bands at 475 cm-1, 190-220 cm-1, 315 cm-1, 310 29 zero filled until a 4096 - 2048 real data matrix. Before cm-1 or 350 cm-1,19 respectively, which were not observed. 30 Fourier transformation, the 2D spectra were multiplied Raman spectra with 325 nm excitation were also obtained 31 with a squared cosine bell function in both dimensions, to verify the possible presence of ZnS, which is 32 the 1D spectra were multiplied with an exponential characterized by the presence of two peaks at 350 cm-1 33 window function. Concentrations were obtained using the and 700 cm-1. None of these peaks was observed, ruling 34 Digital ERETIC method, as provided in the standard out the presence of this phase in the final material. 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 Figure 1. a) Representative TEM micrograph and a typical size-histogram of CZTS nanocrystals; b) HRTEM micrograph 60 and corresponding indexed power spectrum on the selected red squared CZTS nanocrystal; c) XRD pattern and d) Raman ACS Paragon Plus Environment Chemistry of Materials Page 4 of 8 spectra with 532 nm and 325 nm excitation of CZTS NCs. The resonant Raman spectrum of ZnS obtained under 325 nm 1 excitation is also shown as a reference. 2 Figure 2 shows the programed temperature profiles and second, ca. 30 ml/min) was found to be a key to obtain 3 actual read-outs from the thermocouples during the high quality CZTS nanocrystals. Argon bubbling 4 synthesis. The initial formation of nanocrystals was facilitates both mixing and heat transfer during the 5 observed as a clear change of color of the reaction mixture heating-up process, as previously indicated by Lynch et al. 6 from yellow to dark-brown at 135-145 ºC. At this point, for the heating-up synthesis of iron oxide nanocrsytals.31 7 copper sulfide nanocrystals nucleated first due to high Besides the scalability of the heating based syntheses, 8 chalcophilicity of this metal. These nanocrystals then highly important is also the ability to synthesize at high 9 grew by incorporation of all other metals from the concentrations and with high reaction yields. Hence these 10 reaction medium (Cu, Zn and Sn ions) as the temperature 11 considerations had also received a great attention in this increases. A similar mechanism of the formation of study. Figure 4a shows a ternary diagram with data points 12 quaternary phases, based on the progressive 13 corresponding to the actual composition of CZTS incorporation of the different cations in an initially nanocrystals produced at gradually higher precursor 14 formed copper chalcogenide lattice, was observed for 15 concentrations, obtained by reducing the amount of several Cu-based quaternary chalcogenide solvent (ODE) and keeping same quantities of all 16 semiconductors such as Cu2CdSnSe4,26 and CZTS.27-29 The 17 reagents. Higher precursor concentrations resulted in a formation of the CZTS nanocrystals was completed close drop of the zinc content and slight increase in Sn content 18 to the reaching of the peak temperate of the temperature 19 in the final CZTS nanocrystals. These changes were then ramp(250 ºC). The final composition and size remained eliminated by re-adjusting the ratios of reagents - 20 unchanged after further annealing at 250 ºC for 30 21 increasing ZnO and decreasing SnCl4·5H2O amounts. As a minutes. side effect of additional ZnO, an impurity of large 22 23 polydisperse ZnS-rich nanocrystals had formed, but can 24 a) b) be easily removed by centrifugation. On the other hand, 30 250 the batch-to-batch compositional dispersion was 25 25 Cu relatively low as shown in figure 4b, showing the 26 200 Me [at%] 20 composition of 16 CZTS nanocrystal batches obtained T [ºC] 27 150 Sn 15 using the same total metal concentration.. 28 100 10 Zn 29 30 5 50 31 0 130 140 150 160 170 0 25 50 75 100 125 150 175 32 Reaction Time [min] Time [min] 33 Figure 2. a) Programed (red) and real (black) temperature 34 profiles; b) evolution of the composition of the CZTS 35 nanocrystals during the synthesis:  Cu;  Zn;  Sn. 36 37 Surprisingly, the heating rate played insignificant role 38 on the size and composition of the formed CZTS 39 nanocrystals. For example, constant-rate 10 or 30 minutes 40 ramps from 100 ºC to 250 ºC resulted in essentially the 41 same final material. However, the stability of the heating 42 rate had much more profound effect on the size 43 distribution of the obtained CZTS nanocrystals, as 44 consistently observed in over 20 syntheses shown in 45 figure 3. We associate this result to a better separation of 46 the nucleation and growth stages and to a higher 47 temperature homogeneity of the solution with a steady heating rate. Practically, this result points to the need for Figure 3. Representative TEM images of CZTS nanocrystals 48 obtained from a floating (left) and a steady (right) heating 49 accurate control and knowledge of the temperature rate. CZTS nanocrystals with 12 % standard size-dispersion 50 profile of the nanocrystal synthesis. The second main were obtained using a relatively irregular heating rate. 51 factor determining the size distribution was mixing of the Nanocrystals produced with a more stable heating rate were 52 reagents. Intense argon bubbling (few bubbles per characterized with size distributions of 7.6 %. 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment Page 5 of 8 Chemistry of Materials 0.1 1 a) 0.1 0.5 b) 0.5 2 3 0.2 0.2 4 0.4 0.4 5 at% 6 at% Zn Zn 0.3 0.3 2 0.3 S 7 Sa 0.3 Sa Sn 2 S 0.078 t% 0.217 Sn 8 t% 0.303 9 0.303 0.078 0.4 0.4 0.634 0.2 10 0.2 11 0.184 12 0.5 0.193 0.301 0.5 0.1 0.1 13 0.236 14 15 0.4 0.5 0.6 0.7 0.8 0.4 0.5 0.6 0.7 0.8 16 Cu2S at% Cu2S at% 17 18 Figure 4. Ternary diagrams of the composition of CZTS nanocrystals. a) Composition for samples obtained using different total 19 metal concentrations in the initial solution, as detailed within the graph in mmol of the metals (Σ(Cu+Zn+Sn)) per gram of a 20 reaction mixture. The symbol color and shape display the initial ratio of metals. The [Cu, Zn, Sn] initial concentration ratios 21 were  = [2, 1, 1], =[2, 1.35, 0.84] and =[2, 1.78, 0.67]. The solid black square, red triangle and blue circle mark the initial 22 precursor concentrations used for the materials represented with empty black squares, red triangles and blue circles, respectively. B) Composition of 16 CZTS samples produced with a metal concentration of 0.303 mmol/g and an initial metal ratio 23 [Cu, Zn, Sn]=[2, 1.78, 0.67]. 24 25 The size, shape and monodispersity of CZTS Figure 5. TEM image of CZTS nanocrystals prepared using 26 nanocrystals did not significantly change with the total different Cu:Zn:Sn ratios and concentration; 3.50:2.67:1.00 27 precursor concentration when maintaining the metal and 0.303 mmol/g (left) and 2.40:1.60:1.00 and 0.219 mmol/g 28 ratios constant. On the other hand, the Cu:Zn:Sn ratio (right). TEM image of the rod-like CZTS nanocrystals 29 was a main factor governing the shape of the CZTS prepared with stoichiometric precursor concentrations is 30 nanoparticles. It is well known that the presence of tert- presented on figure S2. The presence of tDDT in the reaction 31 dodecanethiol in the reaction mixture is a key factor for mixture did not influence the shape of CZTS nanocrystals at 32 the production of rod-shaped CZTS and CTS high precursor concentrations. 33 nanoparticles.32 It is considered that while DDT binds 34 equally to all facets, tDDT preferentially binds to specific To study the surface chemistry of the synthesized CZTS 35 facets, leaving the [001] direction for crystal growth. This nanocrystals, NMR measurements were performed on a 36 difference allows the preparation of CZTS or CTS purified dispersion of CZTS nanocrystals (Figure 6). In the 1 37 nanocrystals with different aspect ratio.32 It is assumed H NMR spectrum, overlapping sharp and broad 38 that DDT binds to the surface of CZTS nanocrystals resonances were detected in the chemical shift range 0-3 39 stronger, allowing the preparation of smaller but more ppm (Figure 6a), rendering it difficult to directly interpret 40 homogeneous CZTS nanorods (See figure S3 for them. However, the isolated alkene resonance at 5.5 ppm 41 representative TEM micrographs of CZTS nanocrystals was clearly broadened, an effect of slower molecular 42 synthesized with and without DDT). Surprisingly, we tumbling, which is attributed to the binding of ligands to 43 found that at high precursor concentrations, the presence NC surfaces.33 As OLA is the only molecule in the reaction 44 of tDDT did not significantly influence the shape of the mixture with an internal alkene, and we thus identify 45 CZTS NCs. Only highly Cu-rich precursor solutions in OLA as a ligand. Confirmation of the bound nature of 46 combination with high initial precursor concentration OLA is provided by the negative (black) cross peaks 47 resulted in the formation of quasi-spherical CZTS between its resonances in the 2D NOESY spectrum nanocrystals, as shown in figure 5. (Figure 6b). In addition, we find that the sharp 48 resonances feature positive (red) cross peaks which 49 means that the corresponding molecule does not interact 50 with the surface, and is a desorbed or unbound ligand. 51 52 The overlapping resonances at 0-3 ppm can be 53 separated along the diffusion dimension (Figure 6c) as the 54 broad resonances correspond to bound OLA, diffusing 55 slowly while the sharp resonances belong to a free, yet 56 unidentified molecule, diffusing quickly. The diffusion 57 coefficient of OLA is 41 µm²/s and can be converted to a 58 solvo-dynamic diameter of 20 nm using the Stokes- 59 Einstein equation. Since this value is in the same order of 60 magnitude as the nanocrystal diameter, we conclude that OLA is tightly bound to the nanocrystal.34 Amines ACS Paragon Plus Environment Chemistry of Materials Page 6 of 8 typically bind in a NC(L) binding motif (a stoichiometric Potential applications of CZTS nanocrystals requires 1 nanocrystal with an L-type ligand coordinating to surface displacement of the bulky ligands from the surface of 2 metal ions) and although they are generally considered as nanocrystals. A number of potentially useful short 3 weakly binding ligands, they were found to bind tightly to surfactants were tested as stabilizers of the CZTS 4 copper indium sulfide nanocrystals.35 Here, we find the nanoparticles in polar media (Table S1). 5 same tight binding to CZTS nanocrystals and we infer Thermogravimetric analysis (SI) showed that weight loss 6 that this is due to the strong affinity of amines for copper. upon heat treatment was only a few % and mainly 7 depended on the drying condition (See supporting Interestingly, neither the reference spectrum of DDT, 8 information for the details). Note that DMF and water are nor the reference spectrum of tDDT match the sharp 9 tightly bound to the surface of CZTS nanocrystals and an resonances of the unbound compound. However, when 10 DDT is added to the sample, its resonances disappear intensive drying procedure needs to be applied in order to 11 after two weeks and the signals of the unidentified species remove them completely. CZTS nanocrystals typically 12 have increased in intensity, suggesting that this loose solubility after such an excessive drying. Exchanged 13 compound is a degradation product of DDT. samples showed exceptionally low carbon content (<1 w% 14 with a record value of 0.1-0.2 w% which is the practical 15 Finally, the OLA concentration was quantified to 5.34 limit of the used analytical instrument) after heat 16 mM and with the nanocrystal diameter (13 nm) and the treatment in a nitrogen atmosphere. We believe that the 17 nanocrystal concentration (13 mM) we calculated a ligand residual carbon content depends more on the practical 18 density of 0.8 nm-2. This is a rather low value since full accuracy during cleaning procedure after ligand exchange 19 coverage typically corresponds to 4-5 ligands nm-2 and it than on the nature of the used stabilizer/solvent couple. 20 can be explained by assuming that the original ligand shell comprised both thiol and amine and that the thiol In summary, we detailed an over 90% yield synthesis to 21 prepare highly monodisperse wurtzite CZTS nanocrystals 22 degraded over time, possibly even catalyzed by the CZTS nanocrystals. However, to confirm these speculations, a with the size distributions below 10%. The synthesis 23 protocol is easy to upscale in order to prepare at least 24 thorough surface chemistry study during all the synthetic steps would be needed, which is outside the scope of this several grams of CZTS nanocrystals per batch. To produce 25 stoichiometric nanocrystals with highly concentrated contribution. The solvo-dynamic diameter from the 26 precursor solutions, initial metal ratios had to be DOSY experiment (20 nm) was also higher than the sum 27 adjusted. Finally, effective mixing and heat exchange of the nanocrystal core diameter (13 nm) and the organic 28 using intense argon bubbling and a careful control of the ligand shell (3 nm). This leads us to believe that the 29 heating ramp stability were found to be key parameters to nanocrystals agglomerate during the drying step – prior 30 produce highly monodisperse CZTS nanocrystals. to dissolution in CDCl3– due to the low ligand coverage. 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 Figure 6: a) 1H NMR spectrum of 14 nm CZTS nanocrystals in CDCl3, concentration is 8.4 µM particles. b) NOESY 47 spectrum of the same sample. c) DOSY spectrum of the same sample. 48 49 50 ASSOCIATED CONTENT Present Addresses 51 Supporting Information. Ligand exchange procedure, TEM ћ Medcom Advance SA, Edifici N5, Campus de Sescelades, 52 images of CZTS nanoparticles prepared with and without Carrer de Marcel·lí Domingo s/n, 43007 Tarragona, Spain. 53 tert-dodecanethiol; TG, zeta potentials and IR spectra of the Author Contributions 54 CZTS nanoparticles; additional HRTEM images of CZTS nanoparticles. This material is available free of charge via the The manuscript was written through contributions of all 55 Internet at http://pubs.asc.org. authors. / All authors have given approval to the final version 56 of the manuscript. 57 58 AUTHOR INFORMATION ACKNOWLEDGMENT 59 Corresponding Author This work was supported by the European Regional 60 * ashavel@medcomadvance.com Development Funds and the Framework 7 program under ACS Paragon Plus Environment Page 7 of 8 Chemistry of Materials projects SCALENANO (FP7-NMP-ENERGY-2011-284486) and of Functional Nanocomposites: PbTe–PbS Thermoelectric 1 KESTCELLS (FP7/2007-2013/316488). MI thanks AGAUR for Properties. ACS Nano 2013, 7, 2573-2586. 2 their Beatriu i Pinós postdoctoral grant (2013 BP-A00344). 14. Ibáñez, M.; Korkosz, R. J.; Luo, Z.; Riba, P.; Cadavid, D.; 3 JDR thanks the FWO (Research Foundation Flanders) for Ortega, S.; Cabot, A. Kanatzidis, M. G. Electron Doping in 4 financial support. M. V. K. acknowledges partial financial Bottom-Up Engineered Thermoelectric Nanomaterials through support by the European Union (EU) via FP7 ERC Starting HCl-Mediated Ligand Displacement. J. Am. Chem. Soc. 2015, 137, 5 Grant 2012 (Project NANOSOLID, GA No. 306733). Authors 4046-4049. 6 15. Ibáñez, M.; Zamani, R.; Lalonde, A.; Cadavid, D.; Li, W.; 7 also acknowledge the funding from Generalitat de Catalunya Shavel, A.; Arbiol, J.; Morante, J. R.; Gorsse, S.; Snyder, G. J. 8 2014 SGR 1638. Cabot, A. Cu2ZnGeSe4 Nanocrystals: Synthesis and 9 Thermoelectric Properties. J. Am. Chem. Soc. 2012, 134, 4060- ABBREVIATIONS 4063. 10 11 CZTS, (Cu2ZnSnS4); ODE, 1-octadene; OLA, oleylamine; THF, 16. Yang, W.-C.; Miskin, C. K.; Carter, N. J.; Agrawal, R., Stach, 12 tetrahydrofuran; tDDT, tert-dodecanethiol; DDT, E. A. Compositional Inhomogeneity of Multinary Semiconductor dodecanethiol; Nanoparticles: A Case Study of Cu2ZnSnS4. Chem. Mater. 2014, 13 26, 6955-6962. 14 REFERENCES 17. Haas, W.; Rath, T.; Pein, A.; Rattenberger, J.; Trimmel, G., 15 Hofer, F. The Stoichiometry of Single Nanoparticles of Copper 1. Todorov, T. K.; Tang, J.; Bag, S.; Gunawan, O.; Gokmen, T.; 16 Zinc Tin Selenide. Chem. Comm, 2011, 47, 2050-2052. Zhu, Y.; Mitzi, D. Beyond 11% Efficiency: Characteristics of State- 17 of-the-Art Cu2ZnSn(S,Se)4 Solar Cells. Adv. Energy. Mater. 2013, 18. Shavel, A.; Cadavid, D.; Ibáñez, M.; Carrete, A., Cabot, A. 18 3, 34-38. Continuous Production of Cu2ZnSnS4 Nanocrystals in a Flow 19 2. Chen, S.; Walsh, A.; Gong, X.-G.; Wei, S.-H. Classification of Reactor. J. Am. Chem. Soc. 2012, 134, 1438-1441. 20 Lattice Defects in the Kesterite Cu2ZnSnS4 and Cu2ZnSnSe4 19. Singh, A.; Geaney, H.; Laffir, F.; Ryan, K. M. Colloidal Earth-Abundant Solar Cell Absorbers. Adv. Mater. 2013, 25, 1522- Synthesis of Wurtzite Cu2ZnSnS4 Nanorods and Their 21 1539. Perpendicular Assembly. J. Am. Chem. Soc. 2012, 134, 2910-2913. 22 20. van Embden, J.; Chesman, A. S. R.; Jasieniak, J. J. The Heat- 23 3. Carrete, A.; Shavel, A.; Fontané, X.; Montserrat, J.; Fan, J.; Ibáñez, M.; Saucedo, E.; Pérez-Rodrguez, A.; Cabot, A. Up Synthesis of Colloidal Nanocrystals Chem. Mater. 2015, 27, 24 2246-2285. Antimony-Based Ligand Exchange To Promote Crystallization in 25 Spray-Deposited Cu2ZnSnSe4 Solar Cells. J. Am. Chem. Soc. 2013, 21. Park, B.-I.; Hwang, Y.; Lee, S. Y.; Lee, J.-S.; Park, J.-K.; 26 135, 15982-15985. Jeong, J.; Kim, J. Y.; Kim, B.; Cho, S.-H.; Lee, D.-K. Solvent-Free 27 4. Wang, P.; Minegishi, T.; Ma, G.; Takanabe, K.; Satou, Y.; Synthesis of Cu2ZnSnS4 Nanocrystals: a Facile, Green, Up- 28 Maekawa, S.; Kobori, Y.; Kubota, J.; Domen, K. Scalable Route for Low Cost Photovoltaic Cells. Nanoscale 2014, Photoelectrochemical Conversion of Toluene to 6, 11703-11711. 29 Methylcyclohexane as an Organic Hydride by Cu2ZnSnS4-Based 22. Chesman, A. S. R.; van Embden, J.; Duffy, N. W.; Webster, 30 N. A. S.; Jasieniak, J. J. In Situ Formation of Reactive Sulfide 31 Photoelectrode Assemblies. J. Am. Chem. Soc. 2012, 134, 2469- 2472. Precursors in the One-Pot, Multigram Synthesis of Cu2ZnSnS4 32 Nanocrystals. Cryst. Growth Des. 2013, 13, 1712-1720. 5. Patra, B. K.; Shit, A.; Guria, A. K.; Sarkar, S.; Prusty, G.; 33 Pradhan, N. Coincident Site Epitaxy at the Junction of Au– 23 10c) Chesman, A. S. R.; Duffy, N. W.; Peacock, S.; 34 Cu2ZnSnS4 Heteronanostructures. Chem. Mater. 2015, 27, 650- Waddington, L.; Webster, N. A. S. Jasieniak, J. J. Non-injection 35 657. synthesis of Cu2ZnSnS4 nanocrystals using a binary precursor 36 6 .Yu, X.; Shavel, A.; An, X.; Luo, Z.; Ibáñez, M.; Cabot, A. and ligand approach. RSC Adv., 2013, 3, 1017-1020. Cu2ZnSnS4-Pt and Cu2ZnSnS4-Au Heterostructured 24. Connell, M. A.; Bowyer, P. J.; Bone, P. A.; Davis, A. L.; 37 Nanoparticles for Photocatalytic Water Splitting and Pollutant Swanson, A. G.; Nilsson, M.; Morris, G. A. Improving the 38 Accuracy of Pulsed Field Gradient NMR Diffusion Experiments: 39 Degradation. J. Am. Chem. Soc. 2014, 136, 9236-9239. 7. Yu, X; An, X.; Shavel, A.; Ibáñez M.; Cabot, A. The Effect of Correction for Gradient Non-Uniformity. J. Magn. Reson. 2009, 40 198, 121-131. the Ga Content on the Photocatalytic Hydrogen Evolution of 41 CuIn1-xGaxS2 nanocrystals. J. Mater. Chem. A 2014, 2, 12317-12322. 25. Sinnaeve, D. The Stejskal - Tanner Equation Generalized 42 8. Yu, X.; Liu, J; Genç, A.; Ibáñez, M.; Luo, Z.; Shavel, A.; for Any Gradient Shape - an Overview of Most Pulse Sequences 43 Arbiol, J.; Zhang, G.; Zhang, Y.; Cabot, A. Cu2ZnSnS4–Ag2S Measuring Free Diffusion. Concepts Magn. Reson. Part A 2012, 44 Nanoscale p–n Heterostructures as Sensitizers for 40, 39-65. Photoelectrochemical Water Splitting. Langmuir 2015, 31, 10555- 26. Ibáñez, M.; Zamani, R.; Li, W.; Shavel, A.; Arbiol, J.; 45 10561. Morante, J. R. Cabot, A. Extending the Nanocrystal Synthesis 46 Control to Quaternary Composition. Cryst. Growth Des. 2012, 12, 47 9. Yu, X.; An, X.; Genç, A.; Ibáñez, M.; Arbiol, J.; Zhang, Y.; Cabot, A. Cu2ZnSnS4–PtM (M= Co, Ni) Nanoheterostructures for 1085-109. 48 27. Coughlan, C.; Ryan, K. M. Complete Study of the Photocatalytic Hydrogen Evolution. J. Phys. Chem. C 2015, 119, 49 21882-21888. Composition and Shape Evolution in the Synthesis of Cu2ZnSnS4 50 10. Liu, M.-L.; Chen, I.-W.; Huang, F.-Q.; Chen, L.-D. (CZTS) Semiconductor Nanocrystals. CrystEngComm, 2015, 17, 51 Improved Thermoelectric Properties of Cu-Doped Quaternary 6914-6922. 52 Chalcogenides of Cu2CdSnSe4. Adv. Mater. 2009, 21, 3808-3812. 28. Liao, H.-C.; Jao, M.-H.; Shyue, J.-J.; Chen, Y.-F.; Su, W.-F. 11. Xiao, C.; Li, K.; Zhang, J.; Tong, W.; Liu, Y.; Li, Z.; Huang, P.; Facile Synthesis of Wurtzite Copper-Zinc-Tin Sulfide 53 Pan, B.; Su, H.; Xie, Y. Magnetic Ions in Wide Band Gap Nanocrystals from Plasmonic Djurleite Nuclei. J. Mater. Chem. A 54 2013, 1, 337-341. 55 Semiconductor Nanocrystals for Optimized Thermoelectric Properties. Mater. Horiz. 2014, 1, 81-86. 39. Collord, A. D.; Hillhouse, H. W. Composition Control and 56 Formation Pathway of CZTS and CZTGS Nanocrystal Inks for 12. Guo, Q.; Hillhouse, H. W., Agrawal, R. Synthesis of 57 Cu2ZnSnS4 Nanocrystal Ink and its Use for Solar Cells. J. Am. Kesterite Solar Cells. Chem. Mater. 2015, 27, 1855-1862. 58 Chem. Soc. 2009, 131, 11672-11673. 30. Gabka, G.; Bujak, P.; Gryszel, M.; Ostrowski, A.; 59 13. Ibáñez, M.; Zamani, R.; Gorsse, S.; Fan, J.; Ortega, S.; Malinowska, K.; Zukowska, G. Z.; Agnese, F.; Pron, A.; Reiss, P. 60 Cadavid, D.; Morante, J. R.; Arbiol, J. Cabot, A. Core–Shell Synthesis and Surface Chemistry of High Quality Wurtzite and Nanoparticles As Building Blocks for the Bottom-Up Production Kesterite Cu2ZnSnS4 Nanocrystals Using Tin(ii) 2- ACS Paragon Plus Environment Chemistry of Materials Page 8 of 8 Ethylhexanoate as a New Tin Source. Chem. Comm. 2015, 51, 1 12985-12988. 2 31. Lynch, J.; Zhuang, J.; Wang, T.; LaMontagne, D.; Wu, H.; 3 Cao, Y. C. Gas-Bubble Effects on the Formation of Colloidal Iron 4 Oxide Nanocrystals. J. Am. Chem. Soc. 2011, 133, 12664-12674. 32. Wang, J.-J.; Liu, P.; Ryan, K. M. A Facile Phosphine-Free 5 Colloidal Synthesis of Cu2SnS3 and Cu2ZnSnS4 Nanorods with a 6 Controllable Aspect Ratio. Chem. Comm. 2015, 51, 13810-13813. 7 33. Hens, Z., Martins, J. C. A Solution NMR Toolbox for 8 Characterizing the Surface Chemistry of Colloidal Nanocrystals. 9 Chem. Mater. 2013, 25, 1211-1221. 10 34. De Roo, J.; Van den Broeck, F.; De Keukeleere, K.; Martins, 11 J. C.; Van Driessche, I.; Hens, Z. Unravelling the Surface 12 Chemistry of Metal Oxide Nanocrystals, the Role of Acids and Bases. J. Am. Chem. Soc. 2014, 136, 9650-9657. 13 35. Dierick, R.; Van den Broeck, F.; De Nolf, K.; Zhao, Q.; 14 Vantomme, A.; Martins, J. C.; Hens, Z. Surface Chemistry of 15 CuInS2 Colloidal Nanocrystals, Tight Binding of L-Type Ligands. 16 Chem. Mater. 2014, 26, 5950-5957. 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 ACS Paragon Plus Environment