The Molecular Machine Path
to Molecular Manufacturing (2):
Exploiting Improved Methods and Building Blocks

by Eric Drexler on 2009/12/27

Part 4 of a series on the history and prospects of advanced nanotechnology
concepts, prompted by the upcoming 50th anniversary of Feynman’s historic talk,
“There’s Plenty of Room at the Bottom”.

Schafmeister bis-peptides
Rigid, structurally diverse
bis-peptide oligomers
C. Schafmeister, JACS, 2006

In this post, I’d like to outline the promise of fabrication technologies that are within reach of chemical and ribosome-class polymer synthesis (to make nanoscale components) in combination with the familiar mechanism of spontaneous, Brownian assembly (to put them together). These technologies can enable many applications, including productive nanosystems that are beyond ribosome-class machines, further along the path to advanced molecular manufacturing.

Where we are today

In the previous post in this sequence, on “Foldamers and Brownian Assembly”, I outlined current capabilities for designing and making molecular components that spontaneously assemble to form complex systems. Some of these components (polypeptide foldamers) have mechanical and functional properties that compare favorably with high-performance engineering polymers, and these components are compatible with others (polynucleotides) that enable facile design and fabrication methodologies that can produce new million-atom, 100-nm class structures in less than a day.

A major bottleneck today is the weakness of design software and the scarcity of integrative, cross-disciplinary work. Experimental demonstrations of foldamer design and applications show the potential for far-reaching engineering developments, but without well-targeted support for research and research-enabling software, progress will continue to lag far behind its potential. I know this both because it’s obvious from the scientific literature, and because I know several of the leading researchers in the fields I’ve mentioned, and we talk about these issues.

Toward improved components and design methods

These emerging capabilities have potential applications to a range of useful products, including ribosome-class machines that would broaden the range of accessible and economical materials, molecular components, and composite, multi-component products.

Since it is often useful to describe criteria and metrics in technology development, here is an outline of a challenging objective that I am persuaded can be achieved by creative, collaborative work:

Objectives for a next-generation molecular engineering technology, taking as a benchmark the current capabilities of structural DNA nanotechnology:

  • Exploits components that are competitive with DNA oligomers in ease of design and production, and are markedly superior in the diversity of their available shapes, properties, and binding capabilities for dissimilar components.
  • Yields products that are competitive with DNA origami in scale, atomic precision, and ease of assembly, and are markedly superior in their robustness and functionality (including potential applications in nanostructured materials, nanophotovoltaics, hybrid digital systems, productive nanosystems, etc.)


With sufficient progress in design methodologies, existing kinds of polymers (or combinations of them) could achieve these objectives; I’m thinking here of polypeptides and polypeptoids, together with polynucleotides and their synthetic analogs.

New foldamer materials, however, could offer greater inherent rigidity and stronger intermolecular binding, with potential advantages that include low entropy of unfolding (which tends to correlate with greater predictability of folding), and greatly improved mechanical and environmental stability. Regarding greater inherent rigidity I’m thinking here of rigid-rod polymers and Chris Schafmeister’s bispeptide oligomers; regarding stronger intermolecular binding, I’m inspired by (among others) the effectiveness of extended hydrogen bond arrays [pdf] in forming strong non-covalent links [pdf], and the burgeoning field of systematic design of metal-organic frameworks [Chemical Society Reviews special issue]. All these are within the scope of current chemical synthesis methods, and are also potential target products for an early-generation of non-biological productive nanosystems.

The characteristics of advanced foldamers have the potential to facilitate design (especially if they are accompanied by fine-grained structural regularity), and adding new components to the existing toolkit is almost guaranteed to expand the scope of potential systems and system performance.

Progress will enable accelerated progress

When components become substantially easier to design, make, and assemble, and enable the engineering of more robust, highly functional products, technological progress in molecular systems engineering can be expected to accelerate.

  • More effective design techniques will facilitate system architecting and more ambitious objectives.
  • Quicker fabrication and assembly techniques will speed the design-build-test cycle that develops workable concepts into working systems.
  • Greater product robustness and functionality will expand applications, motivating growth in research and research funding.
  • Growth in research and research funding will yield both more applications and further advances in design, fabrication, and capabilities.

Thus, we can expect a virtuous circle of faster development, improved performance, broader applications, and investment in further improvement of the technology base.

For both scientific and practical reasons, there will be good reasons to pursue improvements in productive nanosystems, extending ribosome-like, 1D control of mechanosynthesis to a wider range of 1D polymers (and perhaps to row-by-row synthesis of 2D covalent-network structures). Productive nanosystems of this class have the potential to increase yields and drop costs for already-accessible synthetic structures, and to extend atomically precise control to structures that are beyond reach of unguided solution-phase chemistry.

Ribosomes are needlessly complex

Ribosomes are dauntingly complex, but mostly for irrelevant reasons. In ribosomes, the basic polymer-building mechanism has become encrusted with structural and functional features that stem from of billions of years of evolutionary tinkering, responsive to specifically biological needs. There is no need to imitate this complexity —or even the kinetic-proofreading tricks that compensate for the poor binding specificity inherent in 3-base codons.

Toward increasingly complex and capable systems

The potential advantages of implementing more complex systems are obvious in the case of hybrid digital systems (think of nanolithographic structures as providing circuit boards and sockets for increasingly complex self-assembled plug-ins), but there are also potential advantages in implementing more complex productive nanosystems.

The discussion above has considered existing fabrication methods (chemical synthesis and biological productive nanosystems), while noting the potential for ribosome-class polymer builders. Although the latter devices are deceptively dissimilar from advanced molecular manufacturing systems, this gap will shrink with increasing with complexity and performance — moving from 1D to 2D positional control of reaction locations, moving along the spectrum between self-assembly and mechanosynthesis, and expanding the range of molecular building blocks and products that can be made. Eventually, productive nanosystems can be organized in scalable arrays that feed hierarchical assembly systems. In a mature form, systems of this kind fit the criteria for high-throughput atomically precise manufacturing.

[Edited 28 Dec.]

See also:

Studies of advanced atomically precise fabrication:

{ 4 comments… read them below or add one }

Valkyrie Ice December 29, 2009 at 5:56 am UTC

I have been spending a lot of time recently following many of the trends in research which you have written about here, and thinking very much the same things you have pointed out. You did so in a far more precise and informative way than I would have been able too. Thank you.

I have been following the development of nanotechnology for twenty years now, and seeing the beginnings of real progress is extremely exciting. Hopefully the mental bug you described will be overcome quickly. I have often been asked why I believe that nanofacturing will become a real technology, and tell them to look in a mirror. As an example of the most sophisticated nanomechanical system on the planet, I’ve never doubted that we would overcome all the inherent difficulties to true mechanosynthesis. It’s always simply been a matter of time. With these new breakthroughs, I think that time might have just gotten significantly shorter.

Eric Drexler December 30, 2009 at 4:06 am UTC

@ Valkyrie Ice — Yes, and I am persuaded that the necessary science is in surprisingly good shape. There is no new and necessary [kind of] component or phenomenon that hasn’t been demonstrated. [I say more about what I mean by this in a comment below.]

The great and fundamental problem is institutional and cultural. Although there are research institutions that collectively have ample technical resources, these institutions are configured to do lab-level research, and have no way to undertake the design, task breakdown, coordinated problem-solving, component testing, system assembly, testing, and iterative design and process revision that are necessary to do engineering.

In other words, without architects, plans, and masons, all the bricks and mortar in the world won’t make a house.

This is related to “The Antiparallel Structures of Science and Engineering”.

Chris Phoenix December 31, 2009 at 10:08 am UTC

1) What about molecular actuators? They’d vastly increase the design space, but there don’t seem to be a lot of them ready to be integrated. Recent work from IBM on self-assembling DNA staple structures to lithographed surfaces in particular positions and orientations would seem to enable addressing individual actuators electrically.

2) Some of the complexity of the ribosome is for efficiency. Early artificial “ribosome analogs” would not have to be nearly as efficient.


Eric Drexler January 1, 2010 at 1:11 am UTC

@ Chris Phoenix — I should clarify what I had in mind in my comment above.

There’s obviously much more to do than integrate existing components, in part because the components in question weren’t made for the purpose. There are, however, demonstrations of all the necessary kinds of devices, in the sense of basic functions; these include the actuators that you mention.

Aside from demos, the important point is that commonplace nanoscale materials and phenomena would, if properly configured, provide sets of components that can fulfill the engineering requirements for implementing, among many other things, extraordinarily capable productive nanosystems.

I’m tempted to say that no scientific breakthroughs are necessary, but science reporting comes close to reporting new omelets as breakthroughs in breakfast science, so I’m sure there will be thousands of them (and some even useful).

{ 6 trackbacks }

Leave a Comment

Previous post:

Next post: