© 2002 Robert
A. Freitas Jr. and Ralph C. Merkle
28 October 2002
Literature citation: Robert A. Freitas Jr. and Ralph C. Merkle, “Molecularly Precise Fabrication and Massively Parallel Assembly: The Two Keys to 21st Century Manufacturing,” 28 October 2002; http://www.MolecularAssembler.com/Nanofactory/TwoKeys.htm
Permanent URL of this document: http://www.MolecularAssembler.com/Nanofactory/TwoKeys.htm
This essay was written as a tutorial to present a general overview of molecular manufacturing to an audience of potential policymakers and funding sources. A modified version was later published by Zyvex Corp. as a corporate "white paper" that differs in some particulars from this original document.
Abstract. Two technological capabilities are key to economically viable molecular manufacturing, and are essential to reap the tremendous benefits and the full potential of manufacturing in the 21st century. The first key is the ability to fabricate physical structures with molecular precision. The second key is the ability to fabricate massive quantities of molecularly precise structures, or to assemble larger objects from vast numbers of molecularly precise smaller objects – that is, massively parallel assembly. The end result of the research program proposed here will be molecular manufacturing systems capable of producing macroscale quantities of molecularly precise diamondoid structures. We will need a theoretical and experimental program to develop molecularly precise fabrication of diamondoid structures using machine-phase nanotechnology. We will also need a theoretical and experimental program to develop methods for massive parallelization of these newly developed machine-phase techniques for molecularly precise fabrication. This combined effort will result in the full realization of the tremendous potential for 21st century manufacturing, yielding inexpensive and transformative products in the computer, medical, aerospace, defense, environmental, energy, and household sectors of the economy.
and Molecular Manufacturing
II. Molecularly Precise Fabrication
III. Pathway to Molecularly Precise Fabrication
IV. Massively Parallel Assembly
V. Pathway to Massively Parallel Assembly
VI. Commercial Applications of Nanofactories and Molecular Assemblers
VII. Financial Benefits of Nanofactories and Molecular Assemblers
I. Nanotechnology and Molecular Manufacturing
Nanotechnology is the premier economic driver for manufacturing in the 21st century.
The National Nanotechnology Initiative (NNI; http://www.nano.gov) defines “nanotechnology” as research and technology development in the length scale of approximately 1 to 100 nanometers, with the overall objective of providing a fundamental understanding of phenomena and materials at the nanoscale and to create and use structures, devices and systems that have novel properties and functions because of their small or intermediate size. According to the NNI: “Nanotechnology research and development includes manipulation under control of the nanoscale structures and their integration into larger material components, systems and architectures. Within these larger scale assemblies, the control and construction of their structures and components remains at the nanometer scale. In some particular cases, the critical length scale for novel properties and phenomena may be under 1 nm (e.g., manipulation of atoms at ~0.1 nm) or be larger than 100 nm (e.g., nanoparticle reinforced polymers [with] unique features at ~ 200-300 nm as a function of the local bridges or bonds between the nanoparticles and the polymer).”
In a fundamental sense, the NNI is charged with expanding our technological capabilities for molecular manufacturing, and addressing the question of which arrangements of atoms are possible and desirable to make. Thus the NNI accelerates trends in manufacturing dating back centuries or more. The ultimate limits are dictated by basic physics and economics, and may be defined in terms of what is possible:
· Flexibility: arranging
atoms in most of the ways permitted by physical law.
· Quality: getting almost every atom in the right place.
· Cost: manufacturing at modest increments above the expense of raw materials.
If the U.S. could manufacture large-scale products with high flexibility, high quality, and extremely low cost, it would possess an economic driver much larger than the whole of computing technology in the last quarter century. This is not an exaggeration, nor is it a description of a free lunch. It is the recognition of an economic opportunity that will accrue to any country that develops molecular manufacturing first.
However, developing molecular manufacturing will be difficult. It requires a time perspective beyond the usual commercial and industrial R&D horizons. It will take long-term R&D funding that is focused on specific and accountable performance objectives. This fits the high-risk, high-reward profile for federally funded R&D. But unlike projects that are essentially “science experiments” with unknown or unaccountable outcomes or performance objectives, we already know that molecular manufacturing of large objects at little more than the cost of raw feedstock is possible because it happens in nature. For example, trees are solar-powered molecular manufacturing systems that convert the raw feedstock of water, soil and atmospheric carbon dioxide (CO2) into tons of wood.
Two technological capabilities are key to economically viable molecular manufacturing, and are essential to reap the tremendous benefits and the full potential of manufacturing in the 21st century. The first key is the ability to fabricate physical structures with molecular precision (Sections II and III, below). The second key is the ability to fabricate massive quantities of molecularly precise structures, or to assemble larger objects from vast numbers of molecularly precise smaller objects – that is, massively parallel assembly (Sections IV and V, below).
II. Molecularly Precise Fabrication
The first key to economically viable molecular manufacturing is the ability to fabricate physical structures with molecular precision. There are two principal techniques for achieving molecularly precise fabrication of physical structures. These two techniques are known as self-assembly and positional assembly.
Molecular self-assembly involves stirring molecular parts together and, by clever design of those parts, making parts that spontaneously assemble into the desired result. Self-assembly is a probabilistic process in which only the final state (the desired end configuration), and not the pathway taken to it, is specified. Molecular self-assembly is a strategy for nanofabrication that involves designing molecules and supramolecular entities so that shape- and charge-complementarity causes them to spontaneously aggregate into desired structures. Self-assembly thus refers to a process in which molecularly precise fabrication occurs spontaneously and stochastically “by itself.” Self-assembly is widely found in natural biological systems (e.g., cell membrane lipid bilayers, protein folding, etc.), and is extensively used in traditional bulk chemistry synthesis. Mechanical self-assembly via physical agitation of painstakingly-designed macroscale wood block parts was demonstrated by Roger Penrose in the 1950s.
Molecular self-assembly is experimentally accessible today. Indeed, virtually all of the structures fabricated in contemporary nanotechnology research (e.g., nanotubes, dendrimers, functionalized nanoparticles) are produced via self-assembly. The most complex artificially designed self-assembled molecularly precise structures have been created by Nadrian Seeman at New York University (Figure 1). Seeman uses complementary self-assembling artificial DNA strands to fabricate nanoscale wire-frame objects in almost any regular geometric shape, an ion-activated DNA-based mechanical actuator, and most recently a DNA-based rotary motor. Because the self-assembly process is inherently statistical and because DNA self-assembly permits many mismatches to occur, the yield of perfect product structures is low. Redesign of the artificial DNA strands to make a slightly different product structure demands great effort and time, and the materials available for the core structures are limited to DNA.
Figure 1. DNA-based 3-D molecular objects and an actuator, synthesized by Nadrian Seeman
“Cube”; http://seemanlab4.chem.nyu.edu/nano-cube.html; “Truncated Octahedron”; http://seemanlab4.chem.nyu.edu/nano-oct.html; “A DNA Nanomechanical Device Predicated on the B-Z Transition of DNA”; http://seemanlab4.chem.nyu.edu/BZ.Device.html
Molecular positional assembly involves holding the molecular parts in the proper position and orientation so that when they touch they will join together the way we want them to. Positional assembly is a deterministic process in which the components used in a construction are held in known positions and are constrained to follow desired intermediate physical pathways during the entire construction sequence. Molecular positional assembly is a strategy for nanofabrication that allows a far broader range of starting components to be used – in particular, highly reactive structures that might have prematurely reacted elsewhere if not positionally constrained. The use of highly reactive compounds permits a simple and direct synthetic process that should provide greater flexibility and make possible the synthesis of a vastly wider range of structures. (Attempts to self-assemble reactive components, even if chemically distinct, must fail when those components collide with each other in undesired but still reactive conformations.) Positional assembly is also found in natural biological systems (e.g., ribosomes, DNA polymerase, etc.) and is extensively employed in traditional macroscale industrial manufacturing processes such as robotic assembly lines for automobiles and electronic circuit boards.
Molecular positional assembly – also known as mechanosynthesis – is experimentally accessible today. However, far less progress has been achieved. In part, this is because much of the work to date has employed more specialized laboratory equipment such as scanning probe microscopes (SPMs), high-vacuum systems, and ultra-low-temperature chambers. It also reflects its more recent origins – dating from the first SPMs in the early 1980s. The most complex positionally assembled molecularly precise structures have been created by Wilson Ho at the University of California at Irvine (Figure 2). Ho uses a scanning probe tip at very low temperature to locate two carbon monoxide (CO) molecules and one iron (Fe) atom adsorbed on a silver surface in vacuum. Using the probe tip, Ho picks up one CO molecule, moves it to the Fe atom, and covalently bonds it to the Fe atom using an electric pulse, then repeats the procedure at the same site with the second CO molecule. The result is a positionally-assembled molecule of Fe(CO)2, a simple rabbit-ear-shaped molecularly precise structure that has been fabricated at a specific site on the silver surface.
In the future, and after a great deal of further development, the deterministic positional assembly process could achieve remarkably high reliability, permitting the direct fabrication of relatively large structures. With suitable tool tips, redesign of new product structures should be relatively simple and quick, allowing rapid prototyping of new designs, and a far greater range of atoms and materials should become accessible for molecularly precise fabrication. Robotic assembly of nanostructures using the atomic force microscope is being investigated experimentally at Ari Requicha’s Molecular Robotics Lab at USC, and elsewhere.
Figure 2. Positionally assembly of Fe(CO)2 molecule by Wilson Ho
H.J. Lee, W. Ho, “Single Bond Formation and Characterization with a Scanning Tunneling Microscope,” Science 286(1999):1719-1722; http://www.physics.uci.edu/%7Ewilsonho/sci_112699_fig1.gif and http://www.physics.uci.edu/%7Ewilsonho/sci_112699_fig4.gif
Programmable positional assembly at the molecular scale is the central mechanism for achieving both great flexibility and the ultimate in precision and quality in manufacturing. While ubiquitous at the scale of centimeters and meters, positional assembly at the molecular scale is still rudimentary – but its promise is immense.
III. Pathway to Molecularly Precise Fabrication
Molecularly precise fabrication or assembly involves holding feedstock molecules or nanoscale components in the proper position and orientation so that when they touch they will join together in the desired manner. Old claims that thermal noise or quantum uncertainty must make molecular machines impossible are disproved by the ubiquity of biological molecular machines (e.g., molecular rotary motors) in nature; more recent claims that scanning probe microscope tips cannot arrange atoms, let go of atoms, or induce positionally-precise chemical reactions on surfaces are contradicted by published experimental results. Currently proposed molecular-scale positional devices crudely resemble normal-sized robotic devices, but about one ten-millionth as big. A nanoscale robotic arm could sweep systematically back and forth, adding and removing atoms and molecules from a surface to build any structure that the computer controller instructed it to make. Such an arm, composed of a few million atoms, might be 100 nanometers long and 30 nanometers in diameter. It would have fewer than 100 moving parts but would use no lubricants – at this scale, a lubricant molecule would act like a piece of grit. Such molecular manipulators should be able to position their tips to within a small fraction of an atomic diameter. Trillions of such devices would occupy little more than a few cubic millimeters (a speck the size of a pinhead).
Unlike macroscale robots, nanoscale robotic manipulators and nanoscale products at intermediate stages of assembly will be buffeted by thermal noise. Atoms and molecules are in a constant state of wiggle and jiggle. The higher the temperature, the more vigorous the motion. To maintain its position, a nanoscale robotic arm (and the workpiece upon which it labors) must be extremely stiff. The strength and lightness of a material depends on the number and strength of the bonds that hold its atoms together, and on the lightness of the atoms. The element that best fits these criteria is carbon, which is both lightweight and forms stronger bonds than other elements. The carbon-carbon bond is especially strong: each carbon atom can bond to four neighboring atoms. And carbon atoms can make the stiffest material available: diamond. In diamond, a dense network of strong bonds creates a strong, light, and stiff material. Just as historians have named the Stone Age, the Bronze Age, and the Steel Age after the materials that humans could make, some have suggested that the new epoch of machine-phase nanotechnology that we are entering should be called the Diamond Age.
How can a diamond device of this scale be produced? One answer comes from looking at how we grow diamond today. In a process somewhat reminiscent of spray painting, we build up layer after layer of diamond by holding a surface in a cloud of reactive hydrogen atoms and hydrocarbon molecules. When these molecules bump into the surface they change it, either by adding, removing, or rearranging atoms. By carefully controlling the pressure, temperature, and the exact composition of the gas in this process, called chemical vapor deposition or CVD, we can create conditions that favor the growth of diamond on the surface.
But randomly bombarding a surface with reactive molecules does not offer fine control over the growth process and is more like building a wristwatch with a sand blaster. To achieve molecular precise fabrication, all chemical reactions must occur at precisely specified places on the surface. A second problem is how to make the diamond surface reactive at the particular spots where we want to add another atom or molecule. A diamond surface is normally covered with a layer of hydrogen atoms. Without this layer, the raw diamond surface would be highly reactive because it would be studded with unused (or “dangling”) bonds from the carbon atoms. While hydrogenation prevents unwanted reactions, it also renders the entire surface inert, making it difficult to add carbon (or anything else) to it.
To overcome this problem, we could use a set of molecular-scale tools that would, in a series of well-defined steps, prepare the surface and create hydrocarbon structures on a layer of diamond, atom by atom and molecule by molecule. The first step in the process might be to remove a hydrogen atom from each of two specific adjacent spots on the diamond surface, leaving behind two reactive dangling bonds. This could be done using a hydrogen abstraction tool – an as yet theoretical molecular structure that has a high chemical affinity for hydrogen at one end but is elsewhere inert. The tool’s unreactive region serves as a handle. The tool would be held by a molecular positional device, initially perhaps a scanning probe microscope tip but ultimately a molecular robotic arm, and moved directly over particular hydrogen atoms on the surface. One suitable molecule for a hydrogen abstraction tool is the acetylene radical – two carbon atoms triple bonded together. One carbon would be the handle, and would bond to a nanoscale positioning tool. The other carbon has a dangling bond where a hydrogen atom would be in ordinary acetylene. The environment around the tool would be inert (e.g., vacuum or a noble gas such as neon).
Once the abstraction tool has created adjacent reactive spots by selectively removing hydrogen atoms from the diamond surface, the second step is to deposit carbon atoms at the desired sites. In this way a diamond structure is built, molecule by molecule, according to plan. One proposal for this deposition function is the dimer placement tool. A dimer is a molecule consisting of two of the same atoms or molecules stuck together. In this case, the dimer would be C2 – two carbon atoms connected by a triple bond, with each carbon in the dimer connected to a larger unreactive handle structure. The dimer placement tool, also held by a molecular positional device, is brought close to the reactive spots along a particular trajectory, causing the two dangling surface bonds to react with the ends of the carbon dimer. The dimer placement tool would then withdraw, breaking the bonds between it and the CC dimer and transferring the carbon dimer from the tool to the surface (Figure 3). Since the energy released during the reaction is much larger than thermal noise, the dimer will “snap” onto the surface and stay there. A positionally controlled dimer could be attached almost anywhere on a growing molecular workpiece, leading to the construction of virtually any desired shape.
Figure 3. Dimer placement tool tip (tool handle structure not shown) and the carbon deposition sequence (below)
Ralph C. Merkle, Robert A. Freitas Jr., “Theoretical analysis of a carbon-carbon dimer placement tool for diamond mechanosynthesis,” paper presented at the 10th Foresight Conference, October 2002; http://www.foresight.org/Conferences/MNT10/Abstracts/Merkle/index.html
After a molecularly precise structure has been fabricated by a succession of hydrogen abstractions and carbon dimer depositions, the fabricated structure must be passivated to prevent additional unplanned reactions. While the hydrogen abstraction tool is intended to make an inert structure reactive by creating a dangling bond, the hydrogen donation tool does the opposite: makes a reactive structure inert by terminating a dangling bond. Such a tool would be used to stabilize reactive surfaces and prevent the surface atoms from rearranging in unexpected and undesired ways. The key requirement for a hydrogen donation tool is that it include a weakly attached hydrogen atom. Many molecules fit that description, but the bond between hydrogen and tin is especially weak. A tin-based hydrogen donation tool should be effective.
These three molecular tools, plus a few others, should enable us to make a wide range of molecularly precise stiff structures – but only those that are composed of hydrogen and carbon. This is a far less ambitious goal than attempting to use all 90+ natural chemical elements in the periodic table. In exchange for narrowing our focus to this more limited class of structures, we make it much easier to analyze in detail those structures that can be fabricated and the synthetic reactions needed to make them. Diamond and its shatterproof variants fall within this category, as do the fullerenes (sheets of carbon atoms rolled into spheres, tubes, and other shapes). These materials can compose all of the parts needed for basic nanomechanical devices such as struts, bearings, gears, and robotic arms. Later on, a handful of additional elements can be added, such as dopant atoms to fabricate diamond electronic devices. These and related structures, composed primarily of carbon and hydrogen in combination with a few atoms of nitrogen, oxygen, fluorine, silicon, phosphorus, sulfur, chlorine, or other elements, constitute what we call the class of “diamondoid” materials.
While a molecular robotic arm could build another molecular robotic arm, we must initially build a molecular robotic arm with something other than a molecular robotic arm. We could, for example, use a macroscopic positional device – such as an improved version of an existing atomic-force microscope – to build the first molecular robotic arm. Alternatively, we could self-assemble a simplified molecular positional device or its components, and there are other possibilities. Once the first crude positional device exists, it can be used to fabricate better ones. The end result of this development process would be a basic nanofactory or molecular assembler that employs machine-phase nanotechnology (e.g., nanoscale gears, struts, springs, motors, casings) to fabricate molecularly precise diamondoid structures, following a set of instructions to build a desired specific design.
The magnitude of this challenge should not be underestimated. Present proposals for a nanofactory or molecular assembler able to fabricate diamondoid structures involve at least hundreds of millions of atoms – with no atom out of place. Even a simple robotic manipulator arm, which might be composed of only a few million atoms, would have to be accompanied by other components. The robotic arms might work in a vacuum, dictating the need for a shell around the arm to maintain that vacuum. Essential ancillary subsystems might include acoustic receivers, computers, pressure-actuated ratchets, and binding sites. If each operation, such as hydrogen atom abstraction or carbon dimer placement, typically handles one or a few atoms, then the error rate must be less than one in a billion.
Although such perfection is theoretically attainable, existing technology is not up to the task. For example, a process of chemical synthesis that contemporary chemists would view as very good converts 99% of the reactants to the desired product. Yet that 99% yield represents an error rate of one in 100, which is ten million times less perfect than required. Biology achieves somewhat lower error rates. The synthesis of proteins from amino acids by ribosomes has an error rate of roughly one residue per 2,000; and DNA, by relying on extensive error detection and correction along with built-in redundancy (the molecule has two complementary strands), achieves a net error rate of only one base per billion when replicating itself.
One technique that can position individual atoms is the scanning probe microscope (SPM), in which a sharp tip is brought down to the surface of a sample, generating a signal that allows the probed surface to be mapped, crudely analogous to a blind person tapping with a cane to sense the path ahead. Some SPMs literally push on the atomic surface and record how hard the surface pushes back, or connect the probe and the surface to a voltage source and measure the current flow when the probe gets close to the surface. A host of other probe-surface interactions can be measured and are used to make different types of SPMs. Besides mapping, the SPM can also change a surface – for instance, by depositing individual atoms and molecules in a desired pattern. In a well-publicized case, scientists arranged 35 xenon atoms on a nickel surface to form the letters identifying their employer as “IBM” (Figure 4). But this SPM manipulation required cooling to 4 degrees above absolute zero – hardly ideal conditions for large-scale manufacturing. SPMs also have error rates high enough to require relatively sophisticated error detection and correction methods. While these systems can move around a few atoms or molecules, they can't manufacture large amounts of precisely structured diamond of the kind that might be used to build a molecular robotic arm. Today’s SPMs also are much too slow. In nature, bacterial ribosomes take at least 25 milliseconds to add a single amino acid to a growing protein. If a nanofactory production line or molecular assembler is to manufacture a copy of itself (or its own mass) in about a day, and if this requires about a hundred million atom-placement operations, then each such operation must be completed in ~1 millisecond, somewhat faster than the ribosome. Today’s SPMs, by contrast, may take up to an hour to arrange a single atom or molecule.
Figure 4. In 1989, a scanning probe microscope spelled out “IBM” using xenon atoms
IV. Massively Parallel Assembly
Complex objects assembled from simpler components may be manufactured either serially or in parallel. In serial assembly, objects are manufactured one at a time by a stepwise manufacturing process. Examples include handcrafted unique items such as an antique pocket watch, classical industrial “mass production” items such as automobiles which emerge only one by one at the end of an assembly line, or the traditional serial digital computer which executes instructions one by one in a linear sequence. In parallel assembly, objects are manufactured along many pathways simultaneously or at many different sites, such as polysomes in living cells (multiple ribosomes translating a single mRNA strand simultaneously), mask lithography deposition of multiple circuits simultaneously on a single semiconductor wafer, or the modern parallel computer which at any moment is executing different instructions on hundreds or even tens of thousands of independent processors in a highly parallel manner.
Biology provides perhaps the best example of the power of massive parallelism in assembly. A single ribosome, able to make a single protein as directed by a single molecule of messenger RNA, is a marvelous manufacturing system. Yet by itself, it would have little economic impact. But billions and billions of ribosomes, operating together in each living cell, can make all the proteins in a tree or – still more quickly – all the proteins in a rapidly growing kelp plant, which can literally grow six inches per day.
The difference between serial and parallel processing is equally crucial in molecular manufacturing, where the basic parts are very small. If a typical molecularly precise simple component is 1 nm3 in volume, then to manufacture a single macroscale but molecularly precise 1 cm3 product requires the assembly of 1000 billion billion (1021) individual simple molecular components. With serial manufacturing, just one molecular component is handled at a time – even at a 1 GHz operating frequency it would take many thousands of years, clearly not economically viable. But with parallel manufacturing, vast numbers of molecular components can be processed simultaneously, reducing batch processing times to days, hours, or even less. Massively parallel assembly is the key to the economic viability of molecular manufacturing.
There are two principal pathways for achieving massively parallel assembly of molecularly precise physical structures: self-assembly and positional assembly. In commercial chemical synthesis, self-assembly usually takes place in fluid phase among mole (~1023) quantities of reactant molecules, which interact to produce mole quantities of product molecules. In Seeman’s experiments producing DNA-based structures and in other related experiments involving supramolecular or biomolecular self-assembly, the number of product objects produced per batch is vastly less than mole quantities but is still very large by conventional standards in macroscale manufacturing. The inherent parallelism of self-assembly is the main advantage of this pathway over positional assembly in manufacturing.
To overcome this advantage and reap the full benefits of flexibility, precision and quality in 21st century molecular manufacturing using positional assembly – also known as machine-phase nanotechnology – new techniques for massively parallel positional assembly must be developed. At least two such techniques have already been clearly identified: (1) Massively parallel manipulator arrays and (2) self-replicating systems.
Massively parallel manipulator arrays would use a very large array of independently actuated manipulation devices (e.g., scanning probe tips, robot arms, etc.) to process a very large number of molecular precise components simultaneously to build a larger product object. Using conventional microlithography, researchers at IBM’s Zurich Research Laboratory have fabricated scanning probe tip arrays of up to 1024 individual tips (Figure 5). Simpler mechanical ciliary arrays consisting of 10,000 independent microactuators on a 1 cm2 chip have been fabricated at the Cornell National Nanofabrication Laboratory for microscale parts transport applications. An alternative is Zyvex’s RotapodTM exponential assembly design concept (Figure 6), in which a single robotic arm on a wafer makes a second robotic arm on a facing surface by picking up micron-size lithographically-produced parts – carefully laid out in advance in exactly the right locations so the tiny robotic arm can find them – and assembling them. The two robotic arms then make two more robotic arms, one on each of the two facing surfaces. These four robotic arms, two on each surface, then make four more robotic arms. This process continues with the number of robotic arms steadily increasing in the pattern 1, 2, 4, 8, 16, 32, 64, etc., until some manufacturing limit is reached (e.g., both surfaces are completely covered with tiny robotic arms). Thus a single manipulator uses supplied parts to build a large manipulator array which can subsequently undertake the desired massively parallel manufacturing operations. However, the present RotapodTM manipulator design also lacks sufficient precision to achieve flexible and molecularly precise fabrication.
Figure 5. Millipede concept of IBM Zurich Research Laboratory
“1000 Tips for Ultrahigh-Density Data Storage,” IBM News, 11 October 1999; http://www.zurich.ibm.com/news/99/millipede.html
Figure 6. Zyvex’s RotapodTM exponential assembly design concept
George D. Skidmore et al, “Exponential assembly,” November 2000; http://www.zyvex.com/Publications/papers/exponentialGS.html; “Exponential Assembly”, 25 September 2001; http://www.zyvex.com/Research/exponential.html
A self-replicating system achieves massively parallel assembly first by fabricating copies of itself, and allowing those copies to fabricate further copies, resulting in a rapid increase in the total number of systems. Once the population of replicated manipulator systems is deemed large enough, the manipulator population is redirected to produce useful product objects, rather than more copies of itself. Self-replicating systems are widely found in natural biological systems but have not been pursued explicitly in macroscale manufacturing for at least two reasons: (1) the widespread but erroneous perception of great technical difficulty, and (2) the correct perception that such massive parallelism is unnecessary for traditional macroscale manufacturing. Nevertheless, ever since John von Neumann’s theoretical studies of replicating systems in the 1940s and 1950s, and the well-known 1980 NASA engineering study of self-replicating lunar factories (Figure 7), manufacturing automation has been slowly progressing toward the goal of the fully self-replicating factory – including most notably Fujitsu Fanuc’s nearly “unmanned” robot factory in Yamanashi Prefecture that uses robot arms to make robot arms (Figure 8). It is worth noting that self-replicating systems can be fully remote-controlled, fully autonomous, or any combination in between.
Figure 7. Self-replicating lunar factories from 1980 NASA study
Robert A. Freitas Jr., William P. Gilbreath, eds., Advanced Automation for Space Missions, NASA Conference Publication CP-2255, 1982; http://www.islandone.org/MMSG/aasm/
Figure 8. “Unmanned” robot-making factory of Fujitsu Fanuc Ltd.
FANUC, “New Robot Factory started the full Operation,” News Track, April 1998; http://www.fanuc.co.jp/en/news/h10/h10_04.htm
In the last few years there has been renewed academic research interest in the challenge of mechanical self-replicating systems, in part due to the realization that replication can be a fundamentally simple process. Today there are several ongoing university research programs, both theoretical and experimental, on mechanical (nonbiological) self-replicating machines. Most notably, Gregory Chirikjian at Johns Hopkins University has an active research program on macroscale kinematic self-replicating systems. Chirikjian’s group has built and operated the world’s first Lego-based self-replicating machine (Figure 9). The first robots were remote-controlled (teleoperated) rather than autonomous to focus on the mechanical issues involved in the design of self-replicating systems. “The study of self-replicating robots is an interesting research area which has not been extensively pursued in recent years,” writes Chirikjian. “Our future work will be to develop truly autonomous (rather than remote-controlled) self-replicating robots.” The biotechnology and molecular engineering communities are just beginning to seriously study mechanical replicators operating in the nanoscale size domain.
Figure 9. Chirikjian’s Lego-based teleoperated kinematic replicator made of 5 parts
Gregory S. Chirikjian, Jackrit Suthakorn, “Toward self-replicating robots,” 8th Intl. Symp. on Experimental Robotics (ISER ’02), 8-11 July 2002; http://www.iser02.unisa.it/papers/67.pdf
Current methods of self-assembly, while allowing massively parallel assembly, lack the flexibility, precision and quality that are needed for 21st century molecular manufacturing. Current methods of positional assembly, including massively parallel manipulator arrays and self-replicating systems, should allow molecularly precise massively parallel assembly, although further theoretical and experimental work will be required to fully realize this capability.
V. Pathway to Massively Parallel Assembly
Economically viable molecular manufacturing systems based on molecularly precise massively parallel assembly are likely to take longer to develop than the usual five to ten year time horizon of the private sector. The private venture capital sector has shown considerable enthusiasm for funding nanoscale science and engineering projects that focus on novel electrical or physical properties of nanoscale materials. But they are not focusing on the high-risk, high-payoff opportunity of developing molecular manufacturing machine components and systems with moving parts. There are some European and Japanese initiatives to develop molecular manufacturing components and systems. The key rationale for U.S. government funding is that molecular manufacturing might not happen first in the U.S., or will happen much more slowly in the U.S., if we rely on the private sector for initial R&D stage funding. The question of who develops this technology first has profound economic, security, military, and environmental significance.
A successful molecular manufacturing system that might be deployed in the 2010s or 2020s could be described and analyzed today. Such a system would almost certainly be composed mostly of systems and subsystems that are not experimentally accessible at present, for the simple reason that we cannot yet build the relevant components. But if we are to think about and analyze systems that we cannot build today, and if we are to do so with any certitude, then we must initiate a carefully conceived theoretical and computational R&D program expressly for this purpose. Existing tools in computational chemistry can be harnessed to analyze molecular structures, regardless of whether or not those structures are immediately buildable. Computational modeling of known experimentally accessible structures gives us confidence about the capabilities (and limits) of the modeling software, and permits us to evaluate structures that have not yet been made – and perhaps cannot directly be made – using our current 20th century technology base.
The value of such theoretical and computational work, particularly when used to assess systems that exceed our immediate experimental capabilities, is sometimes debated. But the alternative is to abandon active investigation of systems and structures that cannot be built today. Inability to think systematically about what cannot yet be built is very likely to delay our ability to build it. If we are to build machine-phase molecular manufacturing systems in the next two decades – systems that are experimentally inaccessible today – then methodical design work on such systems is both necessary and urgent.
It is important to reiterate the need to develop and analyze systems. The existing evaluation of scientific research is effective in considering specific issues, but is much less effective in generating (possibly complex) systems proposals for engineering assessment and analysis. The story of the scientist who discovers some new and useful property of matter after accidentally leaving the Bunsen burner turned on while away at lunch is well known. But the story of the engineer who accidentally develops a computer or Saturn V booster is not only unknown, but seems remarkably unlikely.
If – as we believe – the successful development of machine-phase molecular manufacturing systems requires the design of massively parallel systems (with some proposals calling for the design of self-replicating systems) then we need to explicitly create programs that solicit systems proposals – proposals that can reasonably be expected to fulfill the goals of molecular manufacturing as outlined above. Systems proposals can be analyzed by theoretical and computational tools that examine the systems as a whole, together with the subsystems and components from which they might be composed.
VI. Commercial Applications of Nanofactories and Molecular Assemblers
The end result of the research program proposed here will be molecular manufacturing systems capable of producing macroscale quantities of molecularly precise diamondoid structures. Nanofactories or molecular assemblers will make possible the manufacture of fundamentally novel products having the intricate complexity currently found only in biological systems, but operating with greater speed, power, reliability, and, most importantly, entirely under human control. Here are a few examples:
Molecular Electronics and Molecular Computers. The manufacture of computer chips for data storage and computation will undergo a profound change. There are fundamental limits to further improvements in conventional lithography, the current process for chip manufacture. If improvements to computer hardware are to continue at the present pace, in a decade or so lithography must be replaced by some new manufacturing technology. Designs for computer logic elements composed of fewer than 1,000 atoms have already been suggested – but each atom in such a small device has to be in exactly the right place. Nanofactories or molecular assemblers could build these devices. A 1 cm3 block of diamond contains more than 1023 carbon atoms, enough to build and connect more than 1018 computer logic elements. With machine-phase nanotechnology, we should be able to build mass storage devices that can store about 100 billion billion bits of information (~1 million copies of the Library of Congress) in a volume the size of a sugar cube – a billion times more than today’s desktop computers. Massively parallel computers with a 1 cm3 core (if properly cooled) should likewise deliver about 1000 billion billion instructions per second – a trillion times more than today’s desktop computers. Both achievements would represent a phenomenal acceleration of the standard Moore’s Law curve by many orders of magnitude.
Diamond is an excellent electronic material. It outperforms silicon in several key respects. For one thing, electrons move faster in diamond than in silicon. More importantly, as chips get faster and faster, their performance is limited by the need to dissipate heat generated in the circuitry – and diamond can work better than silicon at high temperatures. Diamond has this advantage for two reasons. First, diamond has greater thermal conductivity than silicon, allowing heat to exit a diamond transistor more quickly. Second, the “bandgap” of diamond (5.5 electron volts) is larger than that of silicon (1.1 electron volts). The bandgap is the minimum amount of energy required to boost an electron from its relatively immobile state into the semiconductor’s conduction band, where the electron moves freely under the influence of a voltage. As temperature increases, more electrons gain the energy needed to jump into the conduction band. When too many electrons do this, the device changes from a semiconductor into a conductor; the transistor shorts out and stops working. Diamond’s higher bandgap means that it shorts out at a higher temperature.
Quantum computing is another area where molecular precision in the manufacturing process is likely to be essential. Quantum computers, if they prove feasible, will let us solve certain types of problems in a few days that conventional computers couldn’t solve in billions of years. Of greatest interest is factoring, a problem at the heart of modern cryptographic systems that has great commercial and national security implications. While it is still too early to say which of the many competing proposals for a quantum computer will prove most successful, it seems likely that the winning approach will benefit from greater precision in the manufacturing process. There is already one proposal which requires individual phosphorus atoms to be placed at specific lattice sites in a silicon or diamond crystal – something that today’s technology can’t do. Machine-phase nanotechnology is the obvious way to inexpensively build such remarkably precise structures.
Nanomedicine. Once we learn how to design and construct complete artificial nanorobots using strong diamondoid materials, nanometer-scale parts, and onboard subsystems including sensors, motors, manipulators, power plants, and molecular computers, the practice of medicine will be forever changed. One example is the artificial mechanical red cell called a respirocyte (Figure 10). Still entirely theoretical, the respirocyte is a micron-wide spherical nanorobot made of 18 billion atoms precisely arranged in a diamondoid structure to form a tiny tank for compressed gas, that can be safely pressurized up to 1,000 atmospheres. Several billion molecules of oxygen and carbon dioxide can be absorbed or released from the tank in a controlled manner using computer-controlled molecular pumps powered by glucose and oxygen. External gas concentration sensors would allow respirocytes to mimic the action of the natural hemoglobin-filled red blood cells, with oxygen released and carbon dioxide absorbed in the tissues, and vice versa in the lungs. Each respirocyte can hold 236 times more gas per unit volume than a natural red cell, so a few cubic centimeters injected into the human bloodstream would exactly replace the gas carrying capacity of the patient’s entire 5.4 liters of blood. A half-liter dose could keep a patient’s tissues safely oxygenated for up to 4 hours in the event that a heart attack caused the heart to stop beating. Or this large dose would enable a healthy person to sit quietly at the bottom of a swimming pool for four hours, holding his breath, or to sprint at top speed for at least 15 minutes without breathing.
Figure 10. Artificial red cell (respirocyte, left) and artificial white cell (microbivore, right)
http://www.foresight.org/Nanomedicine/Respirocytes.html (Designer Robert Freitas, Artist PhleschBubble; © 2001 PhleschBubble)
Other proposed medical nanorobots potentially offer equally astonishing performance improvements over nature. For instance, nanorobotic phagocytes (artificial white cells) called microbivores (Figure 10) could patrol the human bloodstream, seeking out and digesting unwanted pathogens including bacteria, viruses, or fungi. Each one of these nanorobots could completely destroy one pathogen in just 30 seconds – about 100 times faster than natural leukocytes or macrophages – releasing a harmless effluent of amino acids, mononucleotides, fatty acids and sugars. It will not matter that a bacterium has acquired multiple drug resistance to antibiotics or to any other traditional treatment. The microbivore will remove it anyway, achieving complete clearance of even the most severe septicemic infections in minutes to hours, as compared to weeks or even months for antibiotic-assisted natural white cell defenses – and without increasing the risk of sepsis or septic shock. Related nanorobots could be programmed to recognize and digest cancer cells, or to mechanically clear circulatory obstructions in a time scale on the order of minutes, thus quickly rescuing the stroke patient from ischemic damage. Nanorobotic clottocytes (artificial mechanical platelets) could make possible complete hemostasis in just 1 second, even for moderately large wounds, a response time 100-1000 times faster than the natural system. Clottocytes may perform a clotting function that is equivalent in its essentials to that performed by biological platelets, but at only 0.01% of the bloodstream concentration of those cells or about 20 nanorobots per cubic millimeter of serum. Hence clottocytes appear to be about 10,000 times more effective as clotting agents than an equal volume of natural platelets.
Medical nanorobots will also be able to intervene at the cellular level, performing in vivo cytosurgery. The most likely site of pathological function in the cell is the nucleus – more specifically, the chromosomes. In one simple cytosurgical procedure, a nanorobot controlled by a physician would extract existing chromosomes from a diseased cell and insert new ones in their place. This is called chromosome replacement therapy. The replacement chromosomes will be manufactured to order, outside of the patient’s body in a clinical benchtop production device that includes a molecular assembly line, using the patient’s original individual genome as the blueprint. The replacement chromosomes would be appropriately methylated, thus expressing only the appropriate exons that are active in the cell type to which the nanorobot has been targeted. If the patient chooses, inherited defective genes could be replaced with nondefective base-pair sequences, permanently curing a genetic disease. Because aging is believed to be the result of a number of interrelated molecular processes and malfunctions in cells, and because cellular and genetic malfunctions will be largely reversible, middle-aged and older people who gain access to an advanced nanomedicine can expect to have most of their youthful health and appearance restored. In the first half of the 21st century, nanomedicine should eliminate virtually all common diseases of the 20th century, and virtually all medical pain and suffering as well.
The idea of placing autonomous self-powered nanorobots inside of us might seem a bit odd, but the human body already teems with such nanodevices. For instance, more than 40 trillion single-celled microbes swim through our colon, outnumbering our tissue cells almost ten to one. Our bodies also maintain a population of more than a trillion motile biological nanodevices called fibroblasts and white cells such as neutrophils and lymphocytes, each measuring perhaps 10 microns in size. These beneficial natural nanorobots are constantly crawling around inside of us, repairing damaged tissues, attacking invading microbes, and gathering up foreign particles and transporting them to various organs for disposal from the body.
Aerospace and Defense. Natural diamond is expensive, can’t be formed into arbitrary desired shapes, and readily shatters. Nanofactories or molecular assemblers will allow the inexpensive fabrication of shatterproof diamond (e.g., diamond fibers) in exactly the shapes we want. This would allow the design and construction of a space shuttle with a structural mass just 1/50th of today’s version without any sacrifice of strength. Made-to-order diamondoid assemblies would benefit all load-bearing structures, and most of the mass in most of our products is load bearing. Beams, struts, buildings, cars, planes, boats and almost all other products would benefit – for example, automobiles could weigh less than the weight of their passengers, and 1-kilogram bicycles would delight racers. The most dramatic improvements are possible for rockets, where the strength-to-weight ratio and the cost of components are critical. Personal ground-to-orbit vehicles the size of a station wagon with launch costs of ~$1/kg to orbit should be feasible, allowing easy personal access to space.
Arrays of molecular sensors on airborne nanorobots and other surveillance devices should allow immediate detection and response to chemical and biological warfare attacks, as well as a defense against a variety of biological and nuclear terrorism threats. Once detected, malicious chemical, biological, and nuclear agents can be rapidly analyzed and neutralized. Machine-phase nanotechnology has a wide range of both defensive and offensive applications. Admiral David E. Jeremiah, USN (Ret), former Vice Chairman of the Joint Chiefs of Staff, said: “Military applications of molecular manufacturing have even greater potential than nuclear weapons to radically change the balance of power.”
Environment. Nanofactories and molecular assemblers are the ultimate “green” technology. Clean assembler-based factories using molecularly-precise feedstocks would eliminate the pollution typically produced by traditional bulk manufacturing and would allow complete recycling of previously manufactured molecular structures. Products could be manufactured and used without generating wastes or noxious effluents, with high process efficiency and lighter materials thus reducing both material and energy consumption. Forced flow of polluted water through nanopore filters, possibly using simple mechanical power supplied by people or animals, could provide a cheap source of fresh potable water for impoverished third-world populations. Molecular nanorobots also could enable comprehensive environmental remediation, quickly and cheaply cleaning up existing pollution and toxic wastes left behind from the 20th century.
Energy. Low cost solar cells and batteries produced by molecular manufacturing could replace coal, oil and nuclear fuels with clean, cheap and abundant solar power. Humanity currently consumes about 10 terawatts of power, but over 100,000 terawatts of solar energy continuously fall on the Earth, most of it unused directly by man. Nanofactories or molecular assemblers could radically alter the economics of energy production. We already know how to make efficient solar cells. Nanotechnology could cut costs, finally making solar power economical. Here we need not make new or technically superior devices – just by making inexpensively what we already know how to make expensively, we would move solar power into the mainstream.New fabrics and building materials can be embedded with billions of tiny motors, sensors, and even computers, allowing our clothing, furniture, and houses to react instantly and intelligently to external conditions and to our ever-changing needs.
Household. A special class of nanofactory or molecular assembler could also build edible foods. Nano-manufactured food would be structurally and nutritionally perfect – no contaminants and precisely specified levels of vitamins, minerals, fats, bones, fibers, sugars, flavors, colors, textures and intoxicants.
VII. Financial Benefits of Nanofactories and Molecular Assemblers
Nanofactories and molecular assemblers will dramatically reduce the manufacturing costs of most manufactured products.
Since molecular manufacturing systems can be used to make more molecular manufacturing systems, it is believed that the capital costs of production can be quite low, possibly in the range familiar in agriculture and in the production of industrial chemicals – on the order of $1/kilogram, or less. After amortization of possibly quite high initial development costs, the price of nanofactories or molecular assemblers (and of the objects they build) should be no higher than the price of other complex structures made by self-replicating systems. For example, potatoes – which have a staggering design complexity involving tens of thousands of different genes and different proteins directed by many megabits of genetic information – can be purchased for less than a dollar per kilogram.
Molecular manufacturing should eventually allow inexpensive production of large batches of designed nanorobots. As a very crude analogy, even in 1995 a typical mail-order bioscience vendor could manufacture 15 micromoles (~9 x 1018 product items) of customized oligonucleotides for $18/base up to at least 110 bases with a convenient 48-hour turnaround on orders; other online vendors offer similar or slightly higher prices up to 200-mer sequences with additional charges for necessary purifications. If a billion-base structure could be induced to self-assemble (Section II) from 10 million different sets of 100-mer custom-built oligonucleotides (total ~20 billion atoms), forming a desired ~20-billion-atom nanomachine, then, optimistically assuming a ~100% yield, the cost would be ~$2 x 10-9 per nanomachine or ~$2000 per ~1 cm3 of product volume which would include 1012 finished nanomachines. At the end of the 20th century, many specialty biotechnology-related drug treatments were comparably expensive. For example, treatments for multiple sclerosis using two closely related ~22.5 kilodalton interferons were administered in dosages of ~1 milligram/week at a treatment cost of ~$10,000/yr, equivalent to a treatment cost of ~$2 x 10-9 (2 nanodollars per giga-atom). The famous HeLa (Henrietta Lacks) cell line is an example of a once tiny population of unique cancerous cells that were purposely replicated in vitro (~24-hour replication time) over many decades and are now widely available worldwide at very low cost (e.g. ~$200 per one-cm3 ampule). Swallowable therapeutic pills containing bacteria (i.e., natural biological nanorobots) are widely available over the counter for gastrointestinal refloration, as for example SalivarexTM which contains a minimum of ~2.9 billion beneficial bacteria per capsule and AlkadophilusTM which contains ~1.5 billion organisms per capsule, both at a 2003 price of ~$0.2 x 10-9 per “nanodevice”. Recycling medical nanorobots might reduce their costs by another factor of 10-100, and nonmedical devices for energy or transportation applications could be cheaper still. The cost of human-engineered molecularly-precise macroscale products should approach the cost of bulk commodities in most applications.
Another way to estimate the manufacturing cost of goods produced by nanofactories or molecular assemblers is to consider the expense of energy inputs, probably the single largest materials input cost for machine-phase nanoscale fabrication. It has been theorized that energy dissipation during molecularly precise fabrication may be as low as ~ 0.1 MJ/kg or less (corresponding roughly to thermal noise at room temperature, e.g., kT ~ 4 zJ/atom at 298 K) if one assumes the development of a set of mechanochemical processes capable of transforming feedstock molecules into complex product structures using only reliable, nearly reversible steps. However, the energy content of various organic feedstock materials is much higher (e.g., 16 MJ/kg for glucose, 17 MJ/kg for vegetable protein, 18 MJ/kg for animal protein, 19 MJ/kg for wood, 39 MJ/kg for fats, and 33 MJ/kg for the thermodynamic heat of formation of diamond from CO2), and it is estimated that the typical energy dissipation caused by chemical transformations involving carbon-rich materials will be ~100 MJ/kg of final product using readily-envisioned irreversible methods in mature mechanosynthetic systems where low energy dissipation is not a primary design objective. Even using today’s high pre-assembler-era cost of energy – $0.028/MJ (at $0.10/Kw-hr for electricity) or $0.014/MJ (at $1.50/gallon for gasoline) – a product that requires 100 MJ/kg to manufacture has an energy input cost of only ~$2/kg. By comparison, the total manufacturing cost of the thin lithographically-patterned multilayer active surface of a semiconductor computer chip is approximately $10,000,000/kg, including about $100,000/kg for the energy costs alone. Molecular manufacturing may yield phenomenal cost savings, possibly up to 5-7 orders of magnitude in computer component fabrication.
It seems likely that the development of machine-phase nanotechnology will require time, focus, and resources. The creation of nuclear weapons took billions of dollars and a very focused development project. The Apollo program likewise took a focused effort over many years, along with billions of dollars and vast amounts of creative talent. The unfolding of the computer industry, while following a very different pattern (private versus governmental, incremental “pay as you go” versus large up-front funding), also involved major funding and many years of focused effort. It is too early to know exactly what pattern the development of machine-phase nanotechnology will follow, but it is not too early to observe that it is likely to require major resources. Whoever makes the decision to commit these resources is unlikely to do so unless there is a clear picture of both the goal and how to achieve it.
How can we obtain this clear picture? As a first step, we must explicitly pursue research into the question: “What would a molecular manufacturing system look like?” The possible answers to this question can be explored using today’s theoretical and computational methods to explore and analyze different components and approaches.
This achieved, we will need a theoretical and experimental program to develop molecularly precise fabrication of diamondoid structures using machine-phase nanotechnology. We will also need a theoretical and experimental program to develop methods for massive parallelization of these newly developed machine-phase techniques for molecularly precise fabrication. This combined effort will result in the full realization of the tremendous potential for 21st century manufacturing, yielding inexpensive and transformative products in the computer, medical, aerospace, defense, environmental, energy, and household sectors of the economy.
Last modified on 22 June 2006
since 14 June 2006