PerspectiveNanocrystal Structure

The coordination chemistry of nanocrystal surfaces

See allHide authors and affiliations

Science  06 Feb 2015:
Vol. 347, Issue 6222, pp. 615-616
DOI: 10.1126/science.1259924

In the 1990s, when quantum confined colloidal semiconductor nanocrystals (NCs, or quantum dots) were first synthesized with narrow size distributions, there was an explosion of effort to harness their bright and narrow luminescence for optoelectronic devices and fluorescence labeling (1). However, the surfactant ligands that stabilized NCs also influenced their electronic structure and optical properties. Encapsulating the NC cores within an insulating inorganic shell reduced the effect of surface structure on charge recombination (2) and forced the radiative recombination of photoexcited charges. These structures greatly increased the photoluminescence quantum yield (PLQY) and enabled their recent use in liquid crystal displays. However, PLQYs of core-shell nanocrystals remain sensitive to their surfaces and if NCs are to be useful within electrical devices, such as photovoltaic (PV) cells, the complex relation between their surface structure and their frontier orbital structure must be better understood.

Surface atoms of NCs have lower coordination than bulk atoms, which results in weaker bonds that in turn create electronic states within the semiconductor band gap that trap photoexcited charges before they can radiatively recombine. Surfactant ligands coordinate to surface atoms, which strengthens their bonding, and “passivates” these midgap electronic states, and enables luminescence. Thus, tailoring the ligand shell for its interactions with the surrounding medium, be it cellular cytoplasm or the conducting matrix of a light-emitting diode (LED), will influence the surface-derived electronic structure and the optical performance. Similarly, surface trap states define the lowest energy path for charge transport (3). Both effects limit the efficiencies of NC PV cells, LEDs, and photodetectors.

The influence of ligation on electronic structure makes surface coordination chemistry critically important in NC science. Density functional theory (DFT) simulations of surface ligand interactions can be informative, but experimental structures have only recently reached sufficient accuracy to enable realistic DFT studies (4). Untangling the atomic structure of NC facets requires methods beyond the ones used with bulk single-crystal surfaces, many of which work well only in ultrahigh vacuum and with well-ordered, flat surfaces. Thus, two key questions remain: What is the nature of the surface ligand interaction, and how do these interactions influence the frontier orbital structure and the fate of excited charge carriers?

Surface-ligand chemistry.

(A) Examples of several ligand exchange reactions are shown. (B) The coordination of different types of ligands in Green's formulation (4) to metal-chalcogenide nanocrystals (NCs, such as cadmium selenide) are illustrated. R is an alkyl group; Bu is n-butyl.

Early investigations of II-VI NCs—in particular, cadmium selenide synthesized in tri-n-octylphosphine oxide—concluded that the dominant ligand type is a datively bound, neutral donor, a so-called L-type ligand (see the figure) (5). Langmuir-like adsorption should result that could be manipulated according to Le Chatelier's principle: If the NCs are placed into a solution containing a much greater concentration of a new L-type ligand, the original ligands should be displaced, regardless of their relative binding affinity. However, early attempts to exchange the ligand shell—for example, by extended heating of NCs in neutral Lewis basic ligand solutions (such as pyridine or other amines)—were only moderately successful.

Several groups then explored ligand exchange reactivity by systematically measuring the effects of ligands on PLQY (68). These studies revealed both a complex underlying surface coordination chemistry and a complex relation between PL and ligation that remains a difficult and central topic in colloidal crystal science. In recent years, it has been realized that L-type ligands are not the primary mode of surface-ligand stabilization. In a landmark study of lead selenide NCs, Moreels et al. (9) reported a size-dependent, metal-rich composition that increasingly deviated from the bulk stoichiometry as the NC radius decreased. They concluded that the NC surfaces are covered by a monolayer of lead atoms. Solution 1H and 31P nuclear magnetic resonance (NMR) spectroscopy revealed that a shell of anionic or X-type ligands provided the charges needed to balance the cationic charge of a metal-rich NC (10).

The anionic ligand–cationic NC description helped explain the difficulty encountered in early exchange studies and opened the door to the design of new reactions that maintain the charge balance between the NC and its X-type ligand shell. In the past 5 years, successful ligand exchange studies have used a modified Le Chatelier's approach in which excess anionic chalcogenide and halide salts displace the native carboxylate and phosphonate ligands, or an acid-base metathesis approach where an acidic proton, trimethylsilyl group, or alkylating agent remove the native carboxylate or phosphonate anions in exchange for another anion of interest. These strategies have led to dramatic improvements in the charge-transport mobilities of NC field-effect transistors and record efficiency of NC PV cells (11).

More recently, NMR spectroscopy revealed that the surface metal ion layer that enriches the NC formula is labile and can be displaced as a complex along with its associated ligand anions, or Z-type ligands (12). The NC stoichiometry is concentration dependent and controlled by the medium in which the NC is suspended. The same study also reported a dramatic and well-behaved dependence of the PLQY on the surface coverage of metal carboxylate complexes.

Absent better understanding of these issues, practical control over the composition and structure of metal surfactant complexes that bind the NC surface will remain erratic. Batch-to-batch variability is typical in the most common methods used to synthesize NCs and variations in the NC composition arise from methods that afford uncontrolled reactivity or terminate the precursor conversion prematurely to obtain a desired size. These effects are compounded by isolation procedures and standards for sample purity that are largely unstudied. Thus, new synthetic methods are needed that reproducibly prepare and isolate NC with known compositions on a larger scale if the structural origins of NC properties are to approach our understanding of bulk semiconductor crystals. This level of control will be necessary for NC technologies to impact not only lighting and photovoltaic technologies but also the biological sciences.

References

View Abstract

Stay Connected to Science

Navigate This Article