Coreduction Colloidal Synthesis of III–V Nanocrystals: The Case of InP

Abstract

Colloidal Group 13–Group 15 semiconductor nanocrystals (III–V NCs) have been the subject of intensive studies during the last two decades because of rich phenomena associated with quantum confinement. However, studies of III–V NCs are restricted owing to synthetic difficulties. As a consequence of the attractive applications as luminescence probes for bioimaging and as photovoltaic devices, InP NCs have become the most extensively studied III–V system. The synthetic studies of InP NCs are therefore more advanced compared to other III–V systems. By adaptating Wells0 dehalosilylation reaction, some feasible synthetic methods for colloidal InP NCs have been established by several groups. These methods usually involve the reaction of an indium salt with tris(trimethylsilyl)phosphine, P(TMS)3, in a high-boiling-point solvent at high temperatures. At first, Micic et al. chose coordinating solvents (e.g., trioctylphosphine oxide (TOPO) and trioctylphosphine (TOP)) as the reaction medium for the synthesis. To obtain crystalline NCs, growth must be carried out over a long period of time (up to seven days). Even then, the assynthesized NCs showed a broad size distribution, and a further size-selective post-treatment was required to achieve monodispersity. Later, Battaglia and Peng further developed this synthesis by replacing the phosphorous-based coordinating solvent with a noncoordinating one, such as octadecene, and using fatty acids as capping ligands. This synthetic variation not only greatly shortened the reaction time to a few hours, but also generated monodisperse InP NCs without any size sorting. Recently, Xu et al. showed that InP NCs of similar quality could also be rapidly produced in some weak coordinating solvents such as fatty-acid esters. However, all of these methods are dependent on the use of expensive P(TMS)3 precursor as the phosphorus source. This results in a high cost, hampering scale-up of the synthesis. To decrease the cost of the synthesis, a few phosphorous compounds, such as In(tBu2P)3 and Na3P, as well as white phosphorus (P4) have been explored as alternative phosphorus sources. However, the replacement of P(TMS)3 by In(tBu2P)3 or Na3P leads to much less control over the particle size. Moreover, the use of these phosphorus sources brings about some new problems. For example, the synthesis of the organometallic precursor is complex and also highcost, and the preparation of Na3P requires a handling of hazardous and pyrophoric sodium metal and P4. [17] When using the simplest phosphorus source, that is, P4, the syntheses are usually conducted under hydrothermal or solvothermal conditions; the as-synthesized InP NCs are polydisperse and aggregated, giving rise to very low quality. Recently, our group reported that colloidal InP NCs could be prepared using a wet-chemical reduction approach in which P4 and LiBH(C2H5)3 (superhydride) were involved. [22] However, the as-prepared NCs showed a broad size distribution. Therefore, it is still a big challenge to explore an alternative economic phosphorus source for the synthesis of InP NCs of an acceptable quality. As phosphorus halides, such as PCl3, can be reduced to elemental phosphorus with extremely high activity, these compounds may be superior phosphorus sources to P4 for the synthesis of InP NCs. It is anticipated that chemical reaction between freshly reduced indium and phosphorus results in a more rapid nucleation burst, which favors the formation of higher-quality NCs. Herein, we report the first example of using PCl3 to synthesize high-quality InP NCs through a coreduction colloidal approach. The synthesis was carried out in octadecene in the presence of stearic acid as capping ligand; the reactions involve simultaneous reduction of In(OAc)3 and PCl3 with superhydride. The hot-injection method, which has been extensively adopted for the colloidal synthesis of monodisperse NCs, is however inapplicable to our system owing to the low boiling point of PCl3 (76 8C). We thus have to carry out the redox reactions at a low temperature (ca. 40 8C) to retain PCl3 in the starting solution, and then elevate the temperature for the NC growth. Figure 1 depicts an X-ray diffraction (XRD) pattern and a transmission electron microscopy (TEM) image of a typical sample grown at 250 8C for 4 h. All detectable diffraction peaks in Figure 1a can be indexed to those of the zinc blende structure InP (ICDD PDF Card No. 73-1983). The broad nature of these peaks indicates the extremely small size of the particles. Figure 1b shows that the InP NCs are dot-shaped and quasi-monodisperse. The average particle size is about (3.5 0.5) nm based on statistical sampling of this image. These results clearly demonstrate that the present synthetic method is effective to generate InP NCs of relatively high quality. The growth process of InP NCs was investigated by monitoring the UV/Vis absorption spectra of samples grown [*] Dr. Z. Liu, D. Xu, J. Zhang, Dr. Z. Sun, Prof. J. Fang Department of Chemistry State University of New York at Binghamton Binghamton, NY 13902 (USA) Fax: (+1)607-777-4478 E-mail: jfang@binghamton.edu