nanoparticle necklace

Assembling nanoparticles in a controlled manner could lead to new nanodevices and materials. But how do you control where the linkages go?

Nanoparticles can potentially be used as giant artificial atoms that may be linked together to form new kinds of material. But whereas atoms tend to join together in particular geometric arrangements and with specific valences, depending on the nature of their electron shells, spherical nanoparticles on their own are more like simple balls that will stack like a greengrocer's oranges. Now researchers at the Massachusetts Institute of Technology have found a way to give metal nanocrystals a specific 'valency' and directional bonding preference.

Francesco Stellacci and his co-workers have made gold nanoparticles with molecular linker groups situated precisely at opposite poles. These can act as the 'monomer' units in linear, polymerized chains of nanoparticles. Previously it has been very difficult to exert any control over precisely where on the particle surface such linker groups might become attached, or how many of them will bind.

Earlier attempts to give nanoparticles a valency have been restricted to supplying them with a single binding site: a valency of 1. For example, Joseph Jacobson and co-workers, also at MIT, attached a single reactive ligand to gold nanoparticles by bonding them via amide linkages to the beads of a resin used for solid-phase peptide synthesis; the single linkage is enforced by the limited area accessible on each resin bead. But for nanoparticles that can be 'polymerized', at least two reactive groups per particle are needed.

Stellacci and colleagues have achieved this by taking advantage of a topological constraint on molecules bound to the spherical nanoparticle surfaces. Long-chain alkylthiol molecules, which terminate in an –SH group, will bind to a gold surface via the sulphur atom to form self-assembled monolayers in which the organic molecules are arrayed in an orderly fashion like the threads of a carpet, all with the same tilt angle. On a curved surface, however, it is simply not possible for all the molecules to have the same tilt: there are inevitably 'point defects', where a single molecule has an anomalous tilt. These are analogous to the whorls of hair on the back of one's head.

Specifically, for a spherical surface there must be precisely two such defects. Stellacci and colleagues found that they could ensure the defects form at diametrically opposed poles on a thiol-coated gold nanoparticle if they used a mixture of two different thiols. Previously, Stellacci and his co-workers observed that such mixed monolayers will segregate into alternating rings of the two thiols, visible in a scanning tunnelling microscope as ripples3. The researchers reasoned that the rings would force the defects to become situated at the two poles, which creates minimal disruption of the ripple pattern.

Because they have a different tilt, the lone molecules at the point defects are less stabilized by interactions with their neighbours than are the other molecules in the monolayer that coats a nanoparticle. This means that they will be the most susceptible to displacement in a ligand-exchange process. To see if their hypothesis was valid, the MIT team covered gold nanoparticles with a mixed thiol monolayer and then exposed them to sulphur-terminated ligands that had carboxylic acid groups at the other end (11-mercaptoundecanoic acid, MUA). Exchange of thiols for MUA introduced a reactive group, which could be used as a linker between nanoparticles, for example via diamine molecules, to which carboxylic acids stick at each end in a reaction equivalent to that used in nylon synthesis.

If precisely two MUA molecules were appended to each nanoparticle at opposite poles, as planned, then this reaction would give linear chains of particles. That was just what the researchers observed. If, on the other hand, the MUA molecules had inserted themselves at random into the monolayer coating, the reaction would have linked the nanoparticles into dense three-dimensional aggregates. They found further confirmation that the particles were being joined into chains by the molecular linkers by observing how the average separation between particles varied with the length of the linkers. Stellacci and colleagues estimated that only one MUA molecule per 100 nanoparticles was likely to attach itself somewhere other than at the ‘poles’ defined by point defects.

They found that the nanoparticle chains aggregated at the interface of an aqueous and an organic solvent to form robust films, presumably by entanglement. Unlike individual chains, these films are rather insoluble, and can be considered comparable to the way ordinary polymers will often precipitate from solution when the molecular chains get entwined. Thus, the divalent nanoparticles are already showing promise as the fabric of new types of material.

Source : 25 January 2007 nanozone news nature publication.