Making vs. Modeling:
A paradox of progress in nanotechnology

by Eric Drexler on 2009/02/25

Knowledge and know-how often go together. Where technologies are concerned, we tend to understand the things we make, and often can make the things we understand. This is a widespread pattern, but it’s important to recognize the exceptions, and nanofabrication is one of them.

There’s no necessary connection between understanding something and being able to make it. Astrophysicists understand a lot about what ordinary stars are and how they work, but this is no help at all in making one. To take another space example, it’s easy to calculate an orbit for a spacecraft, but an entirely different matter to build and launch it. Advanced materials for nanomachines present a similar case.

Paths Forward

An attractive development pathway toward advanced nanotechnologies leads from self-assembly of materials like protein, through softly guided assembly of nanoscale building blocks, to mechanosynthesis of rigid, fine-grained materials like pyrite, ceria, silicon carbide, and diamond. Along this line of development, ease of design and ease of fabrication shift in opposite directions.

Easy but Hard

Biopolymers are easy to make because cells provide the necessary molecular fabrication machines, and these are readily programmed. The problems are entirely those of design. Complex structures require polymeric strands that fold to make components that self-assemble, and to this requires meeting stringent constraints of stability, solubility, and selective binding. Modeling is difficult because stability, solubility, and binding are sensitive to small differences in free energy, and these are difficult to calculate. Also, in working with materials like protein, design is an indirect process that requires clever computational search. On top of all this, building functional systems tends to require larger structures when the materials are soft and coarse-grained, like biopolymers.

In short, in engineering soft machines today, making is easy today, but designing is hard.

Hard but Easy

Now consider structures made from materials at the far end of the path. We can’t make intricate, atomically precise structures out rigid, fine-grained materials today; that’s what puts them at the far end of the path. But what about designing them? In components made of these materials, shape and stability don’t depend on a balance of weak forces, they result from strong covalent and ionic bonding forces that are easily modeled with ample accuracy by standard methods in computational chemistry. Because the materials are fine-grained and rigid, structures can be simple enough and regular enough for a person to design directly, without computational search.

What this enables us to understand

Designing complex nanomechanical systems in detail would of course be an enormous task (and a waste of time for several reasons), but designing basic mechanical components and characterizing their performance has proved to be straightforward. It turns out that there is enough information for standard engineering methodologies to provide conservative estimates of at least some of what can be done with technologies based on these materials and components. The aim is not, of course, to predict what will be made by hordes of smart people in the future, but to get a rudimentary idea of what can be made with components and materials of this general kind.

In short, for advanced, rigid-material-based nanomechanical systems, making is beyond reach today, but designing is surprisingly easy — much easier than comparable design tasks involving biomolecular structures.

This is what makes it possible to speak with confidence about important (yet very limited) aspects of what will become possible at the far end of a long series of developments. I think it‘s clear why this should seem surprising, but only superficially so.


I’ve explored this and related topics in much more depth in “Toward Integrated Nanosystems: Fundamental Issues in Design and Modeling”, in Volume 1 of the Handbook of Theoretical and Computational Nanotechnology.


I’ve discussed materials, assembly, and development paths in a series of posts. The first gives an overview of a new and highly promising direction for lab research in atomically precise nanotechnologies.

  1. Modular Molecular Composite Nanosystems
       Biomolecular engineering for atomically precise nanosystems
  2. Toward Advanced Nanotechnology: Nanomaterials (1)
       Why I’ve never advocated starting with diamond
  3. Toward Advanced Nanotechnology: Nanomaterials (2)
       Stiffness matters (and protein isn’t remotely like meat)
  4. Self-Assembly for Nanotechnology
       The virtues of self-assembly and the benefits of external guidance
  5. From Self-Assembly to Mechanosynthesis
       Mechanosynthesis begins with soft machines
  6. Toward Advanced Nanotechnology: Nanomaterials (3)
       Mechanical engineering meets thermal fluctuations
  7. Nanostructures, Nanomaterials, and Lattice-Scaled Stiffness
       A figure of merit for machine materials
  8. Nanomachines, Nanomaterials, and Klm
       High Klm materials that can be made in a beaker.

Previous post:

Next post: