Abstract

Uniform, alloyed Cu-In-Zn-S nanoflowers with sizes of  nm and  nm composed of aggregated 4.1 nm and 5.6 nm primary crystallites, respectively, were obtained in a one-pot, heat-up reaction between copper, indium, and zinc acetate with tert-dodecanethiol in the presence of trioctylphosphine oxide. Larger aggregates were obtained by diluting tert-dodecanethiol with oleylamine, which lowered the reactivity of the indium and zinc precursors and led to the formation of copper rich particles. The thermal decomposition of tert-dodecanethiol stabilizing the primary crystallites induced their agglomeration, while the presence of trioctylphosphine oxide on the surface of the nanoflowers provided them with colloidal stability and prevented them from further aggregation.

1. Introduction

“Green” methods of synthesis and environmentally friendly, nontoxic materials gain increasing scientific attention during the last years [110]. One of these materials is the Cu-In-Zn-S (CIZS) alloy [36, 1116]. Copper indium disulfide (CIS) and ZnS have similar crystallographic structure and the lattice mismatch between both compounds is relatively small (2.2%); therefore, they can form alloys within the whole composition range [17]. Nanocrystalline CIZS alloys can be synthesized with methods similar to that developed for pure CIS [3]. The synthesis of quasispherical and elongated particles as well as nanocubes of this material has been reported, so far [11, 13, 15, 1825]. Depending on the composition, CIZS alloys can have bandgaps with values between those of pure ZnS (3.78 eV) and CIS (1.45 eV) [26]. Furthermore, the presence of Zn ions within the CIS lattice can passivate some of the defects, which are usually easily formed in pure CIS, which generally leads to higher emission quantum yields in CIZS alloys, compared to pure CIS [19]. In contrast to pure semiconductor compounds, where the optical properties can be adjusted by the size and the shape of the nanocrystals, due to the quantum size effect [27, 28], the tunability of the bandgap of an alloy is independent from the morphological parameters of the crystallites. This extends the range of possible combinations of optical and electronic properties with sizes and shapes and, thus, the application potential of these nanomaterials. However, for many applications both the properties of isolated nanocrystals in colloidal solution and a possibility to generate larger structures composed of self-assembled nanoparticles are important [2937]. One of the strategies leading to such larger nanostructures is the limited ligand protection strategy [31], resulting in a controlled aggregation of primary particles and the formation of nanoflower-like geometries. This tactic has been successfully applied to synthesize nanoflowers of several different materials, for example, ZnTe, In2O3, CuInS2, or CuInSe2 [3033, 38, 39].

Here, we describe a one-pot, heat-up synthesis of alloyed CIZS nanocrystals with zinc blende structure, forming uniform, nanoflower shaped aggregates. We attribute the aggregation of the particles to a partial thermal decomposition of the ligand shell. The size of these aggregates could be controlled by changing the composition of the reaction solution, that is, by diluting tert-dodecanethiol, which serves as the solvent, one of the ligands and the source of sulfur, with oleylamine.

2. Materials and Methods

2.1. Materials

Indium(III)acetate (99.99%—In) (InAc3), copper(I)acetate (99.99%—Cu) (CuAc), zinc(II)acetate (99.99%—Zn) (ZnAc2), oleylamine (OLAM), and tert-dodecanethiol (t-DDT) were purchased from Sigma Aldrich. All the chemicals were used without further purification.

2.2. Synthesis

In a typical synthesis 1 mmol copper acetate, 1 mmol indium acetate, 1.7 mmol zinc acetate, and 4.2 mmol TOPO were dissolved in 10 mL t-DDT under vacuum at room temperature. After 10 min the reaction mixture was rapidly heated to 220°C. Aliquots were taken from the reaction solution at different reaction time, between 5 min and 1 h. The resulting nanocrystals were precipitated with ethanol and redissolved in hexane. This cleaning procedure was conducted three times.

2.3. Characterization

Transmission electron microscopy (TEM) images were obtained on a Zeiss EM902A electron microscope with an accelerating voltage of 80 kV. High resolution TEM (HRTEM) micrographs were taken with a JEOL JEM2100F electron microscope. Samples were prepared on a carbon coated copper grid and dried under air for 24 hours. UV/Vis-absorption spectra were taken on a Varian Carry 100 Scan Spectrophotometer and samples were prepared by diluting the solution of the nanocrystal solution with n-hexane. Powder X-ray diffraction (XRD) measurements were performed with a PANalytical X’Pert PRO MPD diffractometer using Cu Kα radiation (1.54 Å), standard Bragg-Brentano geometry, and variable slits. The samples were measured on low background silicon sample holders and prepared by dropping colloidal solution on the holder and letting the solvent evaporate (by heating to 70°C for 30 min). The integral stoichiometry was obtained by the EDAX detector integrated into a FEI Quanta 200 3D scanning electron microscope.

3. Results and Discussion

Alloyed copper indium zinc sulfide nanocrystals were synthesized using a heat-up procedure, starting with copper, indium, and zinc acetate as the sources of the cations. The reaction took place in t-DDT. This ternary thiol can be easily decomposed thermally; therefore, it does not only play the role of the solvent and ligand in this synthesis; but it also serves as the source of sulfur for the growing nanoparticles [11, 18, 4044]. The thiol is a soft Lewis base, especially capable of stabilizing copper (soft Lewis acid) rich particles, or regulating the activity of the copper monomers in the reaction solution. Because of the presence of In3+ ions (hard Lewis acid), TOPO (hard Lewis base) was used as an additional stabilizer.

Particles obtained in this reaction have nanoflower morphology: they are composed of several particles attached together. These aggregates have a uniform size of  nm (Figure 1(a)). The size of the nanoflowers increases, when the composition of the reaction solution is changed. Diluting t-DDT with 50% oleylamine results in the formation of larger aggregates of CIZS nanoparticles, with a diameter of  nm (Figure 1(b)). A closer look at the structure of such aggregates reveals that they are composed of several crystalline domains (Figure 2(a)). Their random orientation is reflected in the random distribution of the spots corresponding to the (111) lattice planes in the fast Fourier transform (FFT) pattern (Figure 2(b)) obtained from the micrograph in Figure 2(a).

The crystallinity of the nanoflowers was further studied by powder X-ray diffraction (see Figure 3). The size of the crystalline domains was calculated from the broadening of the reflections of the XRD pattern using the Debye-Scherrer formula (with for undefined geometry). The crystallites within the smaller and larger nanoflowers have a diameter of 4.1 nm and 5.6 nm, respectively; thus, the nanoflowers are not single crystalline but consist of several nanoparticles grown together in random orientation, which confirms the observations from HRTEM. These primary particles have zinc blende structure (see Figures 3(a) and 3(b)) and their lattice parameters have values between those for pure CIS and ZnS; this indicates the formation of the CIZS alloy. Assuming a linear dependence between the composition of the alloy and the lattice parameters (Vegard’s law), the particles with a lattice constant  Å (smaller nanoflowers, Figure 3(a)) and  Å (larger nanoflowers, Figure 2(b)) contain 45% and 39% ZnS, respectively.

In order to obtain more precise information about the composition of the particles, the samples were also studied by EDX measurements. The stoichiometry of the smaller nanoflowers is Cu0.71In0.76Zn1.0S2.8; thus, the particles are slightly copper deficient (Cu : In ratio is 0.92 : 1) and sulfur rich. Because of the presence of a high concentration of thiols in the reaction solution, we can assume that the particles are, at least partly, stabilized by thiols. Therefore, the excess sulfur atoms are most likely located on the surface of the particles. The ZnS content obtained from EDX is 56%, which is in agreement with the value calculated using Vegard’s law.

Not only the size, but also the composition of the nanocrystals is affected by the change of the composition of the solvent. The stoichiometry of the larger nanoflowers is Cu1.31In0.78Zn1.0S3.14; thus in contrast to the synthesis without OLAM, the nanocrystals are copper rich. Apparently, the addition of OLAM reduces the activity of indium monomers in the reaction solution, compared with the activity of copper monomers. The In : Zn ratio is not influenced by the composition of the reaction solution; thus, also the reactivity of Zn monomers is affected by the presence of OLAM.

Taking into account the composition of the samples obtained from XRD and EDX, the optical bandgap of the particles should lie in the range between 1.82 and 1.84 eV (values for 39% and 56% ZnS). The bandgap of the particles obtained from the absorption spectrum in Figure 4(a) (smaller nanoflowers) is 2.0 eV [17]. This bandgap widening is most likely a result of the quantum size effect, because the size of the crystallites equals the Bohr radius for CIS (4.0 nm). Another possible reason for the observed larger bandgap value can be the copper deficiency of the particles. A blue shift of the absorption spectra has been observed before for copper poor CIS nanocrystals, which was due to the lowering of the valence band, determined by the hybridization of the Cu d and S p orbitals [45]. The absorption spectrum of the larger nanoflowers is slightly red shifted, compared to that of the smaller particles. The reason for the smaller bandgap can be both the larger size and different composition of the particles, especially the higher copper to indium ratio.

The formation of nanoflowers from small nanocrystals indicates changes in the stabilization of the particles during the reaction. In the beginning of the synthesis the surface of the emerging nanocrystals can be passivated in an efficient way, preventing their further growth and, consequently, leading to the formation of relatively small particles (~4–6 nm). The subsequent aggregation of the primary particles is most likely due to the thermal decomposition of t-DDT, which originally passivates the surface copper atoms. Because of the presence of indium ions on the surface of the particles, their ligand shell should also contain TOPO, the other stabilizer molecule. However, in copper rich particles, the fraction of TOPO on the surface should be smaller, and, consequently, the colloidal stability of copper rich primary particles should be lower, compared to particles containing more indium. Therefore, the formation of larger aggregates in the reaction conducted in a mixture of t-DDT and OLAM is most likely due to the larger fraction of t-DDT on the surface of the primary particles. After the thermal decomposition of t-DDT, the colloidal stabilization of the aggregates is due to the presence of thermally stable TOPO molecules; therefore, we expect no further changes in their size during heating of the reaction solution. Indeed, the size of the aggregates remains constant, when they are kept at 220°C for one hour (data not shown here for brevity). Thus, we can conclude that t-DDT is the main ligand stabilizing the primary particles, while the presence of TOPO keeps the nanoflowers in solution and prevents their further aggregation.

4. Conclusions

The use of t-DDT and TOPO as stabilizers is a suitable strategy to synthesize uniform, alloyed CIZS nanoflowers. The low thermal stability of t-DDT is responsible for the aggregation of the primary nanocrystallites, comprising the nanoflowers, while the presence of TOPO provides the agglomerates with sufficient colloidal stability. This strategy, relying mainly on the properties of the organic stabilizing molecules, could be applied to synthesize nanoflowers of other ternary or quaternary copper sulfide based materials, such as copper tin sulfide or copper zinc tin sulfide.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

The authors gratefully acknowledge funding of the EWE Research Group “Thin Film Photovoltaics” by the EWE AG, Oldenburg. The authors acknowledge also the financial support by the MWN (Material World Network) Program between the National Science Foundation of the US and the Deutsche Forschungsgemeinschaft (Grant no. Li 580/8-1).