Dan Allen and Dr. William Evanson, Physics and Astronomy
In recent years much research has focused on the characteristics of metal and semiconductor “quantum dots” or “nanocrystals”. Cadmium selenide (CdSe) semiconductor nanocrystals of diameter less than ~10nm exhibit unique quantum mechanical properties not present in bulk materials. CdSe nanocrystals of 2-5nm in diameter photoluminesce brightly in the visible spectrum, making them attractive for applications such as photodiodes and light-emitting displays. These nanocrystals are also being researched for applications such as single electron transistors, transistors with multiple logic states, and quantum computing. Present commercial applications include single electron memory devices and luminescent tags for biological molecules.
Photoluminescence efficiency is highly dependent on effective passivation of unpaired bonds at the surface of the nanocrystal. For a 3nm spherical crystal containing ~103 atoms, nearly half of the atoms are within an atomic layer of the surface. Yet, the somewhat irregular surface of CdSe nanocrystals is not well understood. Hess, et al. report a photo-reversible, temperature-induced darkening of nanocrystal luminescence.i This is attributed to a change in the surface bonding arrangements. Solid state NMR, XPS, and high resolution TEM have been used to study the surface of CdSe nanocrystals. Results indicate slight faceting with some surface disorder; however, much remains to be investigated if useful models for surface structure and interactions are to be produced.
A nuclear hyperfine technique called perturbed angular correlation (PAC) may yield new information about the lattice and electronic structure at the surface of the nanocrystals. PAC is highly suitable for the study of defects and bonding arrangements in crystalline structures. Information obtained by PAC spectroscopy about temperature-induced changes of nanocrystal surfaces could be useful in developing robust nanocrystals for commercial applications.
The PAC technique correlates the angle between two gamma rays emitted from a radioactive probe atom after a decay to the strength of the electric field gradient of varying lattice arrangements. 111In is the radionuclide used most often for PAC experiments. It is also ideal for PAC studies of CdSe nanocrystals, since 111In decays directly to 111Cd.
In order to study the structure and surface of CdSe nanocrystals by PAC, I developed a procedure for synthesizing In-doped CdSe nanocrystals. Because In is an electron donor in CdSe, In-doped nanocrystals (an N-type quantum dot) may also exhibit new properties useful in device applications.
Ideally, the location of In atoms within a nanocrystal would be random, and include both surface and interior sites. Diffusion of In into CdSe thin films at moderate temperaturesii points to a favorable incorporation of In into the bulk CdSe matrix. However recent theoretical calculations show that for small CdSe molecular clusters of 17 and 35 atoms, surface sites are significantly energetically favorable for In dopant atoms.iii The probable location of dopant atoms within the larger 4-5nm (1000 atom) nanocrystals remains an important question.
Using the method of Dabbousi, et al.iv, I prepared a solution of the CdSe nanocrystal precursors dimethylcadmium (CdMe2) and trioctylphosphine selenide (TOP-Se) in trioctylphospine (TOP). To this I added indium chloride (InCl3) dissolved in TOP in molar ratios of 1/10, 1/100, and 1/200 (In to Cd).
The mixture was injected into degassed trioctylphosphine oxide stirring at 350° C under N2. The mixture immediately changed from clear to yellow/orange indicating the formation of nanocrystals. The absorbtion spectra showed a peak at 530nm (green) characteristic of ~4nm diameter nanocrystals. Aliquots of In-doped nanocrystals were removed, and the remaining solution was capped with ZnS according to the published method. Nanocrystals were isolated by precipitation with methanol and redispersal in hexanes.
Doped nanocrystals of all concentrations (1/10, 1/100, 1/200) were highly photoluminescent with a narrow emission peak at 540nm, typical of 4nm diameter particles.
Inductively coupled plasma spectroscopy (ICP) of (digested) 1/10 and 1/100 doped nanocrystals show relative In to Cd concentrations of 1/9 and 1/80, respectively.
X-ray photoelectron spectroscopy (XPS) verified the presence of a significant quantity of In before and after ligand exchanges with thioglycolic acid. (Ligand exchange can remove any dopants loosely bound to the surface since S binds to the CdSe surface and ZnS capping layer very strongly.)
In qualitative comparison with undoped nanocrystals, 1/10 In-doped (uncapped) nanocrystals exhibited somewhat higher photoluminescence yield. This may be due to partial surface passivation by the In dopants. Capped particles did not appear to be significantly different, however calibrated quantum yield experiments need to be performed.
The question of location of dopants is important for PAC spectroscopy. Future experiments will include successive etchings of the nanocrystal surface with tripyrridinophosphine oxide, followed by XPS or ICP to depth profile the concentration of dopant atoms. Also, electron spin resonance (ESR) experiments will be performed at cryogenic temperatures to look for unpaired electrons indicative of successful N-type doping. In addition, the symmetry of the lattice positions of such electrons may be examined by ESR.
In-doped quantum dots may also have unique properties, useful in opto-electronic devices. For instance, mobile conduction band electrons from In donors may partially screen electronic defects in larger nanocrystals, and reduce trapping of holes and electrons at the surface, leading to higher luminescence yeild. In-doped CdSe particles may also provide better photoconductivity in photodiodes made of nanocrystalline filmsv, due to the increased density of conduction electrons.
In follow-up research, I intend to compare the photoconductivity of In-doped nanocrystal films to undoped films. I will also compare the current-voltage (I-V) characteristics of single N-type In-doped nanoparticles with conductive atomic force microscopy data I recently obtained for undoped CdSe nanocrystals.
In conclusion, I have demonstrated a simple method for doping CdSe nanocrystals with In. This procedure can be used for PAC spectroscopy studies of nanocrystal surfaces. Future investigation of In-doped nanocrystals may reveal new conduction properties associated with N-type dopants.
I would like to thank ORCA for funding this project, Dr. William Evenson, my advisor, Dr. Roger Harrison for suggesting indium chloride as a precursor and performing ICP experiments, Yit-Yian Lua for XPS data, and Dr. Robert Davis for useful consultations.
References
i B. C. Hess, et al., Phys. Rev. Lett. 86, 3132-3135 (2001).
ii G. J. Scilla, F. C. Luo, Appl. Phys. Lett 42, 538 (1983).
iii K. E. Anderson, C. Y. Fong, W. E. Pickett (University of California-Davis), APS Meeting, March 2001, session L33: Properties of Quantum Dots.
iv B. O. Dabbousi, et al., J. Phys. Chem. B 101, 9463-9475 (1999).
v C. A. Leatherdale, et al., Phys. Rev. B 62, 2669-2680 (2000).
Note to sponsors: This results of this ORCA project will be used in a National Science Foundation proposal for research funding in collaboration with the University of Maryland, Baltimore County. I will also be examining In-doped CdSe nanocrystal photodiodes as potential commercial products as part of my senior thesis research.