The Proceedings of the 1989 NanoCon Northwest regional nanotechnology conference, with K. Eric Drexler as Guest of Honor.

Return to NanoCon Top
Return to NanoCon Proceedings Table of Contents
Return to Jim's Molecular Nanotechology Home Page


C. Bruce Robinson: Conducting Polymers for Electronic Switches

For my part of this presentation, I want to talk about the conducting polymers in the DNA computer: what it takes to make a conducting material and how we can control the structure of those conducting materials.

Given the way nature works, it is very difficult to predict the future of technology in detail. An example from the history of conducting polymers involves the first generation of such materials: polyacetylene. Allied Chemical put a great deal of effort and publicity into the use of polyacetylene to make batteries, but this project came to naught and was abandoned. A second generation of electroactive conducting polymers was based on polyaniline. We heard very little about this, but the Japanese latched onto polyaniline, and they now make and sell several hundred thousand batteries per year based on polyaniline. Clearly technologies evolve and practical applications happen, but predicting just how they will evolve is a problem.

I assume that whatever is going to be made will be made from polymers. I will be talking about polymers that conduct electricity. Another tenet of faith of mine is that nanotechnological objects will have to be self-assembling. The first stage of building the assembler is making something self-assembling.

1. Designing Polymers

The first step in designing a polymer is to ask what kind of properties one wants: So we see that the list of requirements is quite long and rather challenging by itself.

Another example of the unexpected way that things develop is that 20-25 years ago Little proposed a method of making high temperature superconductors and was ridiculed. His suggestions did not necessarily spawn the discovery of high temperature superconductors, but now that they exist, Little's ideas are being re-examined for clues as to how they work.

Another general comment about future technologies is that however we get there, we will have to build on present technologies. The paths that we take can be diverse, but they have to grow out of things that we know how to do now. Thus I would like to talk about things that we do know how to do now, and that might form a basis for what we do in the future.

2. Conducting Polymers

I want to talk about electroactive or conducting polymers (Appendix E: Ref. 1). This type of material takes an organic molecule and puts a metal ligand in the middle. These polymers form spontaneously under the right conditions because metals like to coordinate with a variety of things. If you could regulate this coordination, you could make a linear polymer. A large field is developing on this framework.

Let me consider two examples of self-assembling, metal-containing, linear polymers that have been developed. The porphyrins - planar molecules that form a ring around a central metal atom and include, for example, the business end of hemoglobin - form 2-dimensional networks in solution (Appendix E: Figure 1). They do stack and form solids, but how they stack can not be controlled, and they are not conductors by themselves. You can control the stacking by putting certain metals in the middle. The second figure (Appendix E: Figure 2) illustrates the general approach by which stacked, metal-containing, linear conductors have been constructed. You can put silicon in with oxygen bridges, bake in an oven, and form crystals. This compound has been made, and it is a conductor, and can be doped to be highly conductive, as is typical of a semi-conductor.
AUDIENCE: Would that be conductive solely in that line?

B. ROBINSON: Yes. It is highly anisotropic in its conductivity. It is a conductor in the direction of polymerization and an insulator in the other two directions.
What controls the conductivity? It turns out that the conductivity of this particular material is controlled by the organic part. You can change the metal from one metal to another and you don't change the conductivity.
Some theoretical work (Appendix E: Ref. 1) on a related polymer shown in Figure 3 (Appendix E) said that if you put iron in, it would be a very good conductor. It turns out that it just can't be made. End of story. The silicon analog of this compound, however, can be made. Unfortunately, the silicon analog is air-sensitive, and it is an insulator.

In another example (Appendix E: Figure 4), use of another ligand permitted the spacing of the components so that conductivity could be controlled based on the nature of the metal in the center. A better conductor has been made by substituting biphenyl to separate the porphyrin rings. The spacing of this material will also control the nature of the conductivity. These materials can also be doped to make them semi-conductors.

3. Chemical Switches

Now, I want to tell you about another type of experiment. One could imagine controlling the ability of this central metal to push or pull electrons into this "wire", and thus controlling its conductivity. A prototype experiment has been tried with polyaniline and swelling it with an aqueous solution of iron chloride (Appendix E: Ref. 2). Is the conductivity of the polyaniline changed by changing the oxidation state (valence) of this iron? The answer is "Yes", surprisingly. Changing the oxidation state of the iron changes the conductivity of polyaniline from insulator to semi-conductor (6 orders of magnitude). This is a hybrid of solution and solid-state chemistry demonstrating that an oxidation could change conductivity.

That leads to our idea of a chemically controlled switch that can be built. We are proposing to use conductive, ladder polymers as ligands that connect metal centers. We want things that are themselves conductive and can be made arbitrarily long.

The father of this class is trans-polyacetylene (tPA), the prototype one-dimensional conductor, (Appendix E: Figure 5 and Ref. 3). Its properties were theoretically predicted in the 60's, but it couldn't be made until a Japanese graduate student of Shirakawa got the formula backward and thereby succeeded in making it. Had he been a more careful student, he would have failed. This opened up the field, and Longett-Higgins' predictions (Appendix E: Ref. 4) from the 60's were found to be reasonably accurate.

There has been much controversy as to the nature of the conductivity that has been observed. tPA should just be alternating double and single bonds. If you add an electron, it produces a "defect" called a soliton. It was proposed that this soliton migrates up and down the chain (Appendix E: Ref. 5). This sort of conductivity was not encouraging from my point of view because it required getting nuclei to move to propagate the soliton; i.e., it is phonon limited. But experiments show that this entity does not move along the chain. Furthermore, it behaves like a one-dimensional metal. Yet when you make the material, it acts like an insulator because the metal-like part (the defect) is just a small part of the total material. Although it looks like aluminum foil, it is not conductive. You need to string together the small regions that are conductive and excise the large regions that are not. We propose to make an object of just that length, about 50 carbon atoms. We imagine that ligands of this length would be connected to various transition metals.

How to connect the ligand to a transition metal; how to make the ligand the correct size? Some other materials have been made that have the electronic structure of tPA but are very solid - they look like brick dust and are being used in composites for airplanes. They are called ladder polymers. Two examples POL and PTL are shown in Figure 5 (Appendix E). They terminate in sulfur and nitrogen and will thus complex metal ions and so become self-assembling systems: the transition metal ions connect the ligands.

4. Conducting Polymers & DNA Scaffolds

Hanging such conducting material off of DNA leads to a supra-molecular self-assembling system. We don't even need the metal to make the ligand self-assemble. The self-assembling DNA cruciforms that have been discussed forms the scaffolding that forces these conductive polymers to align in a particular way, and you can then introduce the metal at your leisure. Figure 6 (Appendix E) illustrates the self-assembling nature of the DNA-conductive polymer complex. We don't know yet whether it will be conducting, but the evidence we have suggests that it could work.

Much of the progress in this field has been slowed by the fact that making electroactive polymers requires making a crystal, which may or may not be possible to do. With the DNA scaffold, you now have control over where the polymers will be in space. This should be a very fruitful avenue even for basic research into how these materials work. It was just last month that tPA was finally reported to be a metal - it could actually be made to conduct at the speed of copper. However, the object that carries the charge is a defect in the lattice. With our DNA lattice assembly, there is no lattice to constrict the defect that we require and thus choke off the conductivity. If we can control the redox of the metal, we should be able to control the conductivity.

Now we really hit the realm of science fiction. How do we control this metal's redox state? Crude experiments can be done in which you simply apply voltages (Appendix E: Ref. 2). We also have some suggestions that come from Ratner (Appendix E: Ref. 6) on rectification, or the changing of redox. In one example, a voltage source can be used to change the redox of ruthenium.
AUDIENCE: Could you briefly define "redox"?
B. ROBINSON: Change in the valence which depends on how many electrons are around this ruthenium in orbitals available for bond formation.
Finally, if one can have 3-dimensional arrays of these things, one can also have 3-dimensional addressing through these molecular wires. The atomic bit is thus a point in space at the intersection of wires that allow it to be set and sensed in close analogy to how typical RAM devices work. We have described such a scheme in detail (Appendix E: Ref. 7).

D. General Discussion of Molecular Computers

AUDIENCE: What about the interfacing problem, which always emerges with molecular electronic devices?

B. ROBINSON: We don't know yet how to handle that. As Eric also suggested, we think "fan out" will be useful. Another idea is that etching techniques are progressing to smaller scales so that interfacing should become possible. Further, if you don't like one atom as your bit, we may be able to cluster 20-30 atoms together, which may be a necessary intermediate to get readable voltages in and out, and that should surely be interface-able to conventional devices.

E. DREXLER: What you're talking about here is very much like a transistor so that you would get the kind of fan out that you get in transistor logic, which is very different from solitons, where you are moving around a particle and can not duplicate it to get multiple ones.
Can you say anything about the expected switching speeds for this phenomenon?

B. ROBINSON: We hope it will be electronic, not phonon limited. I'd rather not speculate on numbers.

J. CRAMER: I was wondering about quantum mechanical effects. You're talking about highly periodic structures. You can imagine that the wavelengths of the conduction electrons, described quantum mechanically, might be either commensurate or incommensurate with that regular structure.

B. ROBINSON: What little I know about solid state physics suggests that if I have a regular lattice, the band structure will follow. Another point is that one is going to look at how one tailors the length of the ligand that connects (the metal-atom bits)...

J. CRAMER: Particularly if they connect by more than one path.

B. ROBINSON: Right. It is exactly that length dependence that has to be explored in detail. That was one reason that I suggested these ladder polymers - they can be made in specific lengths and be separated and tested according to length. But it is very important that it be a regular repeating unit.

AUDIENCE: I seem to recall from studying genetics that crossing-over is a dynamic process. If so, or if there are other dynamic processes happening to DNA molecules, how do you hold them still long enough to build structures.

N. SEEMAN: The process of recombination in general is enzymatically controlled at numerous steps. Otherwise it does not occur. Secondly, the branch isomerization that I showed you is one of the dynamic features that is occurring on a molecular scale, and, as I indicated, we eliminate that by destroying the sequence symmetry that flanks the branching site. We control the conditions that the DNA in our reactions is exposed to, and these are not at all conducive to recombination. I only spoke of recombination to indicate how we got interested in the problem.

AUDIENCE: Since you're looking at the DNA molecule from a structural rather than a biological stand-point, is there the possibility of using other, similar molecules?

N. SEEMAN: Yes. That's a very good question. There are many variations on the theme. Among the options available are doing things to eliminate the charge so that we could move into non-aqueous environments. Bruce's polymers would be much happier in a non-aqueous environment. However, there are some problems in doing that. If you eliminate the charge, you may weaken the base-pair specificity. We will have to find out experimentally.
Switching to RNA, or something similar with a functional group that may allow a derivatization is another way of hanging Bruce's polymers off the DNA. One could also imagine using proteins that bind to specific sites on DNA, such as cro, which Eric mentioned last night, as a way to attach the polymer to the DNA. There are also small drug molecules that have a limited specificity for binding DNA.
One of the things that we like about DNA is that it is a big fat molecule. We don't have to worry about whether we hang a single one of Bruce's polymers or a whole bundle of them from a scaffold as large as DNA.

J. LEWIS: One thing that is perhaps worth mentioning is to compare the computational density of this DNA-based electronic computer with Eric's nano-mechanical computer. Your estimates were that it would be a thousand-fold less dense in terms of bits per cubic nanometer and yet you've estimated that you could still store about 3000 times as much information as all the books that have ever been written in a volume the size of a sugar cube.

AUDIENCE: Have you thought about how super-conducting materials would fit into your model.

B. ROBINSON: I've only begun to think about that. No one really knows the physical basis for the working of high temperature super-conductors, so its hard to guess how that would go.
N. SEEMAN: Do you know how thin a wire you could make from a super-conductor?

B. ROBINSON: On the nanometer scale. But super-conductors are not the sort of objects that make nice, well-defined small things. They are made by the "shake-and-bake" approach.

AUDIENCE: One of the problems with super-conductors is the difficulty in making it into a wire. If you could supply a support, like a DNA molecule...

B. ROBINSON: It might be possible in principle, but the scale of what units go together and stick together is important, versus what units go together with a lot less energetics. My contention is that organic conductors are a good choice because, once made, they are stable objects. They don't continue to grow by epitaxy. After the ligands (organic conductors) are made, then weaker forces come into play, leading to self-assembly. High temperature super-conductors, on the other hand, are oxides of various metals and whether they can be assembled in regular ways that would still be super-conductors is not known.

AUDIENCE: If I wanted to design a computer circuit out of your scheme, can I assume that I have both p and n type CMOS transistors and just basically make a 3-dimensional network.

B. ROBINSON: Fundamentally, yes.

AUDIENCE: Is there a reason that you chose to publish a memory scheme instead of a computer scheme?

B. ROBINSON: We thought that it was an easier goal.
AUDIENCE: There was something last week in the popular press about some sort of new development about photographing DNA at the University of Washington?

J. LEWIS: I think you're talking about the visualization of a molecule of DNA electrophoresing through an agarose gel, work done by Stephen Smith.

Appendix E. Conductive Polymer Structures by Dr. Bruce Robinson

REFERENCES (for conference presentation):
1. M. Hanach, A. Datz, R. Fay, K. Fischer, U. Keppeler, J. Koch, J. Metz, M. Mezger. "Synthesis and Properties of Conducting Bridged Macrocyclic Metal Complexes" Chapter 5 (pp 133-204) in Handbook of Conducting Polymers I (T. A. Skatherin, ed.) 1986, Marcel Dekker, New York. This chapter is an excellent review. Figures 1 through 4 are taken from this chapter.
2. E. W.Paul, A. J. Ricco, M. S. Wrighton. (1985) J. Phys. Chem. 89: 1441-1447.
3. B. H. Robinson, J. M. Schurr, A. L. Kweroin, H. Thomonn, H. Kim, A. Morrobel-Sosa, P. Bryson, L. R. Dalton. (1985) J. Phys. Chem. 89: 4994 (see references therein).
4. H. C. Longuet-Higgins and L. Salam. (1959) Proc. Roy. Soc. London, A 251: 172.
5. W. P. Su, J. R. Schrieffer, and A. J. Heeger. (1980) Phys. Rev. B., 22: 2099-2111.
6. A. Aviram and M. A. Ratner. (1974) Chem. Phys. Lett. 29: 277-283.
7. B. H. Robinson and N. C. Seeman. (1987) "The design of a biochip: a self-assembling molecular-scale memory device" Protein Engineering 1: 295-300. Figures 5 and 6 are taken from this paper.

FIGURE 1. The basic structure of tetrazaporphorin [1] and phthalocyanine [2].

The basic structure of tetrazaporphorin [1] and phthalocyanine [2].
This figure was re-drawn from reference [1] (Hanach et al.).

FIGURE 2. General construction scheme for bridged macrocyclic transition metal polymers.
General construction scheme for bridged macrocyclic transition metal polymers

The (:) indicates a lone pair of electrons on N that can coordinate with a metal ion.
This figure was re-drawn from reference [1] (Hanach et al.).

FIGURE 3. Polymeric phthalocyaninatometal complexes.
Polymeric phthalocyaninatometal complexes.
Although it may not be obvious, the atoms in the central portions of the ring structures are nitrogen (N), and the central atom is a metal (M).
This figure was re-drawn from reference [1] (Hanach et al.).

FIGURE 4. Phthalocyaninato-pyrazine-iron (II)
Phthalocyaninato-pyrazine-iron (II)

This figure was re-drawn from reference [1] (Hanach et al.).

FIGURE 5. Conducting polymers. (a) trans-polyactylene (tPA) is shown. A region containing a defect is indicated by the dotted line. (b) Ladder polymer polyphenothiazine (PTL). (c) Ladder polymer polyphenoxazine (POL).

Conducting polymers.

This figure was re-drawn from reference [7] (Robinson and Seeman).

FIGURE 6. Nucleic acid junctions as scaffolds for docking conducting polymers. The nucleic acids are shown as thick lines and the conducting polymers as thin lines. The 5' --> 3' directionality of the nucleic acid chains is indicated by the arrowheads at the end of each chain. Above, two 4-arm nucleic acid junctions are shown. The right arm of the junction on the left is shown with an overhanging cohesive (sticky) end complementary to the overhanging cohesive end of the left arm of the junction to the right. The conducting polymers are attached by the tethers every helix turn. When the two arms have been enzymatically ligated, as shown on the bottom, the conducting polymers line up in a proper fashion to chelate a metal ion which forms a conducting bridge.

This figure was re-drawn from reference [7] (Robinson and Seeman).

Next page of the NanoCon Proceedings
Previous page of the NanoCon Proceedings
Return to NanoCon Proceedings Table of Contents
Return to Jim's Molecular Nanotechology Home Page
Return to Web Site Design and Authoring Services

Comments and correspondence? Send email to Jim Lewis:
NanoCon Proceedings © Copyright 1989, by NANOCON
This page is part of Jim's Molecular Nanotechnology Web, copyright ©1996 James B. Lewis Enterprises. All rights reserved.
Last updated 24June96.
The URL of this document is: