When polymer scientists are asked to select goals for the development of polymer nanostructures and mesoscale engineering, they find self-organization is an important task for scientists and polymers alike.

Polymers with defined size, shape, architecture and chemical functionality, either from biological or non-biological origin, are among the most promising and versatile building blocks for nanostructured materials. Whereas the characteristic length and shape of the formed material structures (mostly by self-assembly and self-organization) is given by the size and geometry of the starting units, specialized functions can be independently encoded in the chemical composition and patterning of the individual polymers. That way, features such as controlled adhesion and bio-compatibility, intelligent and responsive surfaces, the design of nanoscale actuators and pumps and so on, can be engineered in a rational manner.

From this broad range of possibilities, several attractive applications that illustrate the potential of functional polymer materials to a wider community were identified at a polymer scientists meeting in Crete^* last year. It is thought that these could be realistically achieved by collaborative efforts in the next few years. Most of the engineering challenges are at the 'mesoscale', with the aim of closing the gap between so-called top-down and bottom-up approaches to materials assembly. The ultimate goal is the ability to control the arrangement and interactions of nanoscale objects by 'functional interfacing'¹, which means the regulation of both the common operation and the flow of information between a group of objects. For that purpose, it is also useful to identify medium-term goals. What follows is a personal hot-topic list, selected from the issues that emerged during the Crete meeting. These seven topics are clearly not exhaustive and are intended to stimulate further discussion.

The goals are:

To extend the limits of conventional top-down approaches to polymer processing, such as lithography, micro-embossing and microstamping. This can be done by improving the resolution of these techniques by using non-classical tricks, such as modern electrochemistry, or the use of electric field instabilities and flow patterns, as shown by Steiner and colleagues² on page 48 of this issue. Figure 3 on page 50 depicts a model structure created by this technique, indicating the precision and the high aspect ratio that can be achieved.

To break down the three-dimensional (3D) symmetry of the interaction between nanoscale objects into dimension-specific communication. For example, to allow the objects to differentiate between left/right, up/down and forward/backward, thereby enhancing the bottom-up approach to materials assembly. Simple versions of such structures are found in Janus objects³, but the final goal is the controlled 3D assembly of a nanostructure from their encoded building blocks⁴ (Fig. 1). One approach is seen in experiments with nanocrystals⁵ where functional polymers selectively adsorbed on specific surfaces induce such 3D selectivity. The resulting high structural complexity is determined by the symmetry of the primary crystal structure, which together with the adsorbed polymers regulates the number and size of the exposed crystal faces.

To control mutual adhesion of objects by using nanostructured surface patterns. Such surface structures can be created by microstamping, but also by lithographic techniques combined with subsequent chemical reactions. For example, surfaces patterned with hydrophilic and hydrophobic spots will preferentially bind to a matching pattern, but result in non-binding for non-compatible structures. Such pattern recognition is error-tolerant, because the degree of binding is proportional to the 'similarity' of the two patterns, so relative patterns are also recognised. In this way, patterns based on gradients can be used to precisely position two objects with nanometre precision (Fig. 2a).

To make the surfaces of objects 'responsive' to external stimuli, such as heat, light or their chemical environment. This can be done by grafting mixed polymer brushes⁶ or block-copolymer structures on to surfaces (Fig. 2b). Such surfaces can be reversibly triggered, patterned, and switched, for applications such as data storage and microfluidics.

To reduce randomness in polymer sequences built from two or more different monomers. This can be done by generating macromolecules in strong local force-fields (for example, along a surface pattern or at a phase boundary), which help determine the monomer sequence and therefore the polymer function. Such 'field imprinting' was predicted by theoretical and modelling studies⁷ to lead to protein-like polymers, which can recognise surface patterns, particle sizes or polarity gradients, owing to the 'memory' of their formation.

To bridge the gap between synthetic polymers and biological structures. This can be done either by molecular hybrid structures containing elements of both biomatter and synthetic polymers (in which the biological building block is preferentially made by biotechnological approaches), or by biofunctional surfaces. This biomaterials interface is crucial for implants and medical devices⁸, and its potential and importance is still underestimated in polymer research. Conversely, biotechnological separation and purification procedures are still based on commodity polymers⁹, whereas customized hybrid polymers may also significantly improve and simplify these processes.

To create synthetic polymer 'muscles', valves, and motors from nanoscale structures. Some scientists are investigating the idea¹⁰ of using the sequence of the protein elastin (valine-proline-glycine-valine-glycine) as a building block for polymer structures that emulate the natural properties of elastin. But other 'simple' nanostructures, such as corrugated nanosheets made of metal-coated rubbers (Fig. 3), have been identified as highly efficient actuators, and are compatible with current microengineering.

The field of functional polymer materials is currently in the process of defining goals and principles for the rational design of clever molecular structures and techniques. In trying to turn their visions into reality, and therefore serve the future demands of technology, it is my personal opinion that polymer scientists must first break down these dreams into well-defined tasks and tools that need to be developed. It is too soon to tell if the scientists meeting in Crete have come away with a recipe for success, but they have taken some promising steps in the right direction.