Most large molecules made by conventional synthesis are floppy

The techniques of organic synthesis chiefly work with molecules in liquid solution. Large, precise structures can be built by combining smaller molecules under suitable conditions (temperatures, solvents, catalysts,...), but it is an empirical fact that most large molecules made this way have many parts linked by single covalent bonds. Organic synthesis can make precise structures in which every multi-atom part is linked to every neighboring multi-atom part by multiple bonds, but all such structures made to date have either been small, or are organized in a simple, regular, and repetitive way. The reasons for this involve the details of what can and cannot be made using currently available solution-phase procedures.

Structures that contain many parts linked by single bonds tend to be floppy, that is, to have extensive conformational flexibility. The illustration on this page shows a typical structure of this sort (which forms part of a larger structure that has been called a “molecular machine”). Each bond marked with a blue arrow allows rotation of the parts above relative to the parts below. These motions are not entirely free — for example, many of those shown have three preferred directions, separated by low energy barriers. Some bonds allow no rotation — here, the bonds within the pair of six-membered rings have partial double-bond character, inhibiting rotation, and their embedding in a flat ring blocks rotation entirely. Taken together, however, the many available rotations about various bond-axes enable the molecule to twist, bend, and writhe.

Most sorts of machines cannot be made from parts that are so flexible — they require parts with stable shapes, able to move in controlled ways. In biological molecular machines, one finds these stable shapes. Although their molecules are inherently floppy (considering only their covalent bonds), these floppy molecules fold back on themselves and pack together to form solid, atomically precise structures. In these structures, non-covalent interactions hold parts in place and block rotation. Protein engineers have learned to design such structures, opening a path toward molecular machine systems that can be built with current techniques.

Easier to design and model (but impossible to make with current techniques) are large, atomically precise structures made stiff by dense networks of covalent bonds. These structures have engineering advantages, but their production will require new techniques.