Subsuming the Metal Seed to Transform Binary Metal Chalcogenide Nanocrystals into Multinary Compositions

Direct colloidal synthesis of multinary metal chalcogenide nanocrystals typically develops dynamically from the binary metal chalcogenide nanocrystals with the subsequent incorporation of additional metal cations from solution during the growth process. Metal seeding of binary and multinary chalcogenides is also established, although the seed is solely a catalyst for nanocrystal nucleation and the metal from the seed has never been exploited as active alloying nuclei. Here we form colloidal Cu–Bi–Zn–S nanorods (NRs) from Bi-seeded Cu2–xS heterostructures. The evolution of these homogeneously alloyed NRs is driven by the dissolution of the Bi-rich seed and recrystallization of the Cu-rich stem into a transitional segment, followed by the incorporation of Zn2+ to form the quaternary Cu–Bi–Zn–S composition. The present study also reveals that the variation of Zn concentration in the NRs modulates the aspect ratio and affects the nature of the majority charge carriers. The NRs exhibit promising thermoelectric properties with very low thermal conductivity values of 0.45 and 0.65 W/mK at 775 and 605 K, respectively, for Zn-poor and Zn-rich NRs. This study highlights the potential of metal seed alloying as a direct growth route to achieving homogeneously alloyed NRs compositions that are not possible by conventional direct methods or by postsynthetic transformations.

. TEM of the NCs derived from aliquot solutions withdrawn at different temperatures displaying the Bi seed formation, Bi-Cu 2-x S heterostructures formation, and structural evolution of the Bi-Cu 2-x S heterostructures into Cu-Bi-Zn-S based NRs.   (PDF No. 01-071-2426), and Cu 2.94 Bi 4.8 S 9 (PDF No. 03-065-5469) obtained from the mentioned PDF cards marked as   , and  respectively (b)Percentage of the phases calculated from Rietveld refinement of the XRD patterns obtained from aliquot thin films on silicon zero background disc. From XRD analysis of aliquot 150 and 170 °C, beside Bi, presence of BiOCl is ascertained. It can be suggested at 150 and 170 °C unreacted BiCl3 is still present. Exposure of unreacted BiCl3 to air and moisture produced BiOCl (BiCl3+ H2O 2HCl+ BiOCl).           Figure S17. (a) Rietveld refinement and lattice parameters of the experimental XRD pattern of NR1 in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra; (b) Rietveld refinement and lattice parameters of the experimental XRD pattern of NR2 in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. For NR1 PDF No. 03-065-5469 and for NR2 PDF No. 03-065-4965 were used as reference.

Effect of TOPO:
Trioctylphosphine oxide (TOPO) was used because of its ability to increase the solubility of metal precursors as a co-ordinating solvent in the colloidal systemFurthermore, in our case, changing the TOPO concentration has no effect on the morphology of the nanorods as shown in Figure S21. However, the nanorods synthesized without TOPO display poor colloidal stability with wide aggregation compared to the optimized conditions with the use of TOPO. Figure S23. Scanning electron micrographs of Cu-Bi-Zn-S NRs using different concentrations of Trioctylphosphine oxide (TOPO). Highlighted concentration has been used in the main synthesis. Effect of thiol mixture: Alkanethiols are soft Lewis base with a strong affinity for Cu + cations in colloidal reaction media. Compared to bulky tert-dodecanethiol (t-DDT), linear 1dodecanethiol (1-DDT) shows stronger affinity for Cu + which slows the nucleation kinetics. Hence, mixing sterically hindered t-DDT helps to regulate the reactivity of Cu + during nucleation and growth. It should be noted the decomposition temperature of 1-DDT is ~ 280 °C compared to ~230 °C for t-DDT. Hence, a thiol mixture containing more t-DDT will release more active S for nucleation. Additionally, by modulating the concentration of 1-DDT in the thiol mixture, we observed significant changes in the shape of the NCs. By using only t-DDT, the NRs are formed with few NRs displaying higher width ( Figure R2a, b). While increasing the amount of 1-DDT to more than 50 % in the thiol mixture, the morphology of the NCs changes to nail shape (50% 1-DDT, Figure R2c) and multipods (100 %, 1-DDT Figure R2d). Hence, it can be said insufficient passivation of facets by t-DDT resulted in wider NRs when zero concentration of 1-DDT is used. In comparison, the higher decomposition temperature of 1-DDT, higher Cu binding affinity, and high passivation resulted in irregular shape of the NCs when a high amount of 1-DDT is used. Figure S24. Scanning electron micrographs of Cu-Bi-Zn-S NCs using different concentration of 1-DDT (a) 0%, (b) highlighted wide NRs, (c) 50% and (d) 100%, in t-DDT and 1-DDT mixture keeping the total amount 1 ml as used in optimized synthesis. Figure S25. Rietveld refinement of the experimental XRD pattern of the NCs collected from 150 °C aliquot in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. For Bi, denoted as  (PDF No. 04-003-1496), and BiOCl, denoted as  (PDF No.00-006-0249) reference were used. Figure S26. Rietveld refinement and lattice parameters of the experimental XRD pattern of of the NCs collected from 170 °C aliquot in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. For Bi, denoted as  (PDF No. 04-003-1496), and for BiOCl, denoted as  (PDF No.00-006-0249) reference were used. Figure S27. Rietveld refinement and lattice parameters of the experimental XRD pattern of of the NCs collected from 190 °C aliquot in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. For Bi, denoted as  (PDF No. 04-003-1496), for BiOCl, denoted as  (PDF No.00-006-0249) reference were used. Figure S28. Rietveld refinement and lattice parameters of the experimental XRD pattern of of the NCs collected from 210 °C aliquot in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. Figure S29. Rietveld refinement and lattice parameters of the experimental XRD pattern of of the NCs collected from 230 °C aliquot in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. Figure S30. Rietveld refinement and lattice parameters of the experimental XRD pattern of of the NCs collected from 250 °C aliquot in black with red spectra depicting the calculated pattern and blue curve presents the difference spectra. Figure S31. XPS analysis of the solid products derived from the aliquots collected at 150, 210, 230 °C. (a) Bi4f and S2p spectra, the peaks at ~ 156.5 and 161.8 eV correspond to metallic bismuth of the seed and the peaks at ~161 and ~162 eV corresponds to sulphide from metal sulphide bond (Cu 2-x S and Bi x Cu y S z ) for all temperatures. The peaks near 163.8 and 158.9 eV belong to Bi 3+ of Bi-S bond for 210, and 230 °C aliquot, which is absent in 150 °C suggesting the absence of any Bi x Cu y S z phase. For the 150 °C aliquot, the peaks of 164.5 and 159.2 eV belong to Bi 3+ from BiCl 3 or BiOCl, which might have formed from the exposure of unreacted BiCl 3 to air and moisture. Figure S32. XPS survey of the solid products derived from the aliquots collected at 150, 210, and 230 °C. The XPS survey exhibited a peak for Zn 2p around 1045 eV suggesting the presence of Zn 2+ in NCs collected at 230 °C; however, no Zn 2+ signal was found from the aliquot collected at 210 °C further corroborating our observation that Zn 2+ will come into effect only after 230 °C.

Liquid phase analysis:
The supernatant of the aliquot collected at 150, 210, and 230 °C were characterised using FTIR.
The -C=C-H (~3050 cm -1 ), C-H alkyl stretching (~2920 cm -1 symmetric and 2850 cm -1 asymmetric) belongs to oleylamine and 1-ocatdecene present in the liquid. It was further supported by the presence of -C-N stretching mode near ~1120 cm -1 and C=C stretching mode at 1645 cm -1 . The C-S stretching mode is present near ~700 cm -1 . However, S-H stretching mode is absent near ~2500 cm -1 suggesting the removal of S-H due to the interaction of alkane thiols with cationic sources and thermolysis of alkanethiol to produce active sulfide species.
Similarly, the stretching mode of P=O displayed a broad peak ~1150 cm -1, suggesting the presence of TOPO.