January 2001 Vol. 31, No. 1, pp. 34–40.

 

Bernhard Wessling

From conductive polymers to organic metals

Organic metals were a euphoric dream in the 1980s and a disappointment in the 1990s. Now they are moving to the marketplace.

The year was 1977. Hideki Shirakawa, Alan MacDiarmid, and Alan Heeger (winners of the 2000 Nobel Prize in Chemistry) published their discovery that upon reaction with iodine, polyacetylene exhibited electrical conductivity many orders of magnitude higher than the neutral unreacted film (1). Here, it seemed, was a polymer chain on which electrons could move—a molecular conductor. Chemists and physicists who dreamed of lightweight batteries and a replacement for copper power lines, and the billion-dollar market for these products, got their hopes up; their goal seemed just over the horizon.

I, too, was fascinated by these materials, and I began to think about them seriously in 1978, when I had a research position in a small company. In 1981, when I moved to Zipperling Kessler GmbH, I introduced my conductive polymer project there. Shortly thereafter, an announcement in a German automotive association magazine almost killed the project. It cited a major battery manufacturer’s claim: Soon the old lead batteries would be replaced by polyacetylene batteries. My boss came and said, “What are we going to do in this business; why the hell are you doing this research?” I somehow convinced him that this must have been a misunderstanding—I just could not imagine that this polymer could become a battery! Even if this could be true, there would be so many different application possibilities for these wonderful materials that we should not work too short-sightedly, but with a long-term strategic focus. I did not know that long-term meant more than 20 years.

The research hypothesis and strategy
I argued that it was premature to direct our research to specific products, because I suspected that these polymers might have behaviors that were totally different from those described in the literature. It was likely that synthesis, properties, processing, and application could be fundamentally different from anything imaginable at that time.

I was lucky; my intuition was right. The conductive polymers are completely different than we had expected in the late 1970s or early 1980s and 1990s. An important crossroads for our research strategy occurred very early when I thought about polyacetylene’s insolubility and unmoldability: What if we just accept that it is insoluble and unmoldable? Could we process this polymer by dispersing it in a nonconducting medium such as a solvent or a polymer matrix (2)?

In our lab, we went on to research polyacetylene, many polypyrroles, polythiophenes, and poly(perinaphthalenes); we finally concentrated on polyanilines. We designed suitable synthetic procedures that yielded a dispersable powder with an ultrafine primary particle size of 10–15 nm (3). We found that the surface tension was extremely high (>150 mN/m), leading to a terrible tendency to agglomerate, much stronger than in any pigment (the field in which my company had mastered dispersion technology).

Figure 1. TEM of flocculated polyaniline in a polymer matrix. Fine structure is visible in the secondary particles, which are ~100 nm across. Used with permission from Ormecon GmbH.

So we designed dispersion processes capable of overcoming these extreme surface forces (4). In solvents, we can disperse the conductive polyaniline (or other polymers) down to the primary particle size, ~10–15 nm (5). In a polymer matrix, this is not possible because of wetting limitations exhibited by the polymer matrix surface. We are limited to a minimum size of ~50 nm, and we find a typical particle size value is ~100 nm (Figure 1).

Can conductive polymers form true solutions?
Most research groups went the other way: They wanted to introduce solubility and moldability by introducing side chains, synthesizing copolymers, and even designing new monomer types. Although this was very fruitful research that increased our understanding, it did not lead to practical materials. Many new monomers and polymers were made. Many of them were claimed to be soluble and moldable, but none of them made it to the market (6). It was even questionable whether the “solutions” of conductive polymers were “true” solutions because it was never proven that

• single molecules were present in the solvent,
• each monomer unit was completely solvated, and
• any intrachain interaction or interactions among monomer units were replaced by interactions between the chains and the solvent.

On the contrary, in every case we found very fine particles in the “solutions”, so we continued the dispersion research. Now we can be sure that true solutions from conductive polymers not only do not exist, but also that they cannot exist. The reason is found in thermodynamics (7):

• The surface tension for these conductive polymers is far higher than that of any of the available solvents. To form a true solution, the surface tension of the solvent must be similar to that of the solute. Water has a surface tension of 72.9 mN/m, the highest known value for a solvent, but still far below the 150 mN/m observed in polyaniline.
• By definition, the free energy of formation for a true solution must be negative. This is always the case if the solvation energy and enthalpy terms are negative. For crystalline solutes, the enthalpy of melting also must be considered. This enthalpy can either be provided by the solvent, or the heat may be provided by an external source. The conducting polymer mixtures have an extremely high (possibly infinite) melt enthalpy, making it impossible to have a negative free energy of dissolution.
• Because a conductive polymer is a polymeric salt, we can estimate the lattice energy by constructing a Born–Haber cycle. The resulting lattice energy value is far above any potential solvation or hydration energy.

Dispersions: Diverse nonequilibrium systems
How can dispersions of conductive polymers conduct electricity? Aren’t the chains, which are supposed to be the “wire” for the electrons, cut and destroyed by the dispersion? Wouldn’t the particles in the matrix be isolated from each other, surrounded by the matrix polymer?

Fortunately, no. As we understand it, the chains are not stretched out, but folded into the primary particles. No particle has any connecting chain that links it to a neighbor particle to provide a transport “wire”, so we did not risk cutting the wires by dispersing the particles. Moreover, nature did us a favor, allowing self-organization to occur under nonequilibrium conditions (8).

Dispersions are complex and diverse nonequilibrium systems (9). Even today, a complete theoretical understanding of dispersions is lacking. We know that dispersion begins with the removal of adsorbed species to create surfaces; this is an extremely energy-demanding step because 10-nm particles have a specific surface area of 300 m²/g. After the surfaces are created, the matrix surface wets the particle surfaces or replaces formerly present adsorbents.

This arrangement of particles does not conduct electricity. Particles in the fully dispersed stage are in fact isolated from each other. But at a critical concentration, these particles join together and form elongated “pearl chains”. More precisely, the layer of the matrix adsorbed on the particles joins and merges to form a hollow flexible pipe with the dispersed particles inside. Now, the particles can touch each other. From outside, this looks like a snake that has swallowed too many golf balls.

Figure 2. SEM of a break surface in a polymer blend at the flocculation point concentration (6%). Spherical polyaniline particles are encased in the blend matrix to form the characteristic “pearl chain” morphology. Used with permission from Ormecon GmbH.

These “golf ball snakes” or “pearl chains” are branched, and they form elongated 2-D networks (Figure 2). Their layers are bent and form 3-D continuous structures, which interpenetrate to form a complex co-continuous system.

On the way to the organic metal
To our surprise, the conductivity for these polymer dispersions was in the same range as for the pure undispersed polyaniline. We also found the first hints of hidden metallic properties in the thermoelectric power behavior (10). This was confirmed by microwave absorption studies: We concluded that in a primary 10-nm-diam particle, an 8-nm metallic core is surrounded by a ~1-nm nonmetallic (maybe even insulating) shell (11).

The conductivity mechanism, therefore, is a complicated sum of two contributions: a metallic part within the particle, and a tunneling of electrons through the shells from particle to particle through a barrier of ~2 nm thickness. The picture of electrons moving on the chain like a train on the track is gone. Instead, electrons within the particle are metallic in nature, like a “cloud”, or “electron gas”, which does not need the chain as a guiding substrate. The electrons are contained in a lattice and form a metallic conduction band; at the same time, they are quantums capable of tunneling. The thermally activated conductivity (it decreases with decreasing temperature, in contrast to metals) indicated that tunneling was the most prominent transport mechanism.

Years later, we encountered another surprise: In a commercially prepared poly(methyl methacrylate) (PMMA) blend containing 40% polyaniline, we found a conductivity ~10 times higher (!) than for the undispersed raw polyaniline. The dispersion is metallic from 300 to 230 K. Its conductivity is independent of temperature in the region from 230 to 150 K, and it is nonmetallic at temperatures below 150 K. The conductivity at very low temperatures (4 K) was still 20% of the original value, in contrast to a factor of 10^–9 for previously made blends (12)! The spectrum in the near-IR showed a much broader band for the blend than for the undispersed raw material, indicating a higher degree of electron delocalization.

Something fundamentally different had happened. We studied the changes in X-ray diffraction and the reduced activation energy for conductivity at temperatures between 0.5 and 10 K. We found that our polyaniline had crossed the insulator-to-metal-transition during the dispersion process (13). If we can rely on the model unit cell applied for interpreting the X-ray spectra, we can conclude that the chain packing is much denser after dispersion, and the Ph–N–Ph angle decreases from 166° to 134°.

Figure 3. Morphological hierarchy of polyaniline particles and aggregates. Individual molecules cluster to form primary particles (organic metals), which then form primary aggregates. Primary aggregates cluster to form secondary particles.

Figure 4. Simulation of a polyaniline primary particle. Used with permission from Ormecon GmbH.

Not only does dispersion provide us with the only viable tool for processing polyaniline, but also, under appropriate conditions, it allows us to cross the boundary between insulators and metals. Most recently, using small-angle X-ray scattering, we may have revealed the inner structure of the particles during dispersion: One primary particle (10–15 nm) may consist of ~20 individual molecules, folded to a diameter of 3.5 nm, which nevertheless form a coherent metallic core (Figure 3). The primary particles agglomerate to a first hyperstructure of ~45 nm and then form secondary particles of ~100 nm in polymer matrices (Figure 4) (14).

The dispersion technology is the basis of all commercial applications of any conductive polymer or organic metal. For example, poly(3,4-ethlyenedioxythiophene) (PEDOT) is an insoluble polymer made by Bayer that has a totally different chemistry from polyaniline and is only available in dispersion form (15).

Our organic metal is first present in form of a predispersion (e.g., in PMMA). The powder used to make the predispersion is useless if it is not transformed to a predispersed form. The advantage of our powder and predispersion technology is that we can limit ourselves to one, or at most two, chemically different polyanilines and only half a dozen to a dozen predispersions. This is the basis for many useful formulations, including paints, lacquers, and dispersions.

Applications
In these formulations, the organic metal can display a unique set of properties:

• It is more noble than copper and slightly less noble than silver.
• Thanks to its redox chemistry, it can act as a catalyst (see Figure 5).
• It is the first metal that can be stoichiometrically and reversibly reduced.
• Because of its small particle size, it can build very thin, though conductive and reactive, layers, or thin transparent green conductive coatings.

During the early 1980s, when we decided to do basic chemical, physical, and materials research that was not explicitly directed to a special product, we still were thinking about applications. None of these has been realized until now, except for transparent green antistatic coatings, which is a small market niche. Later, we found technologies we couldn’t even imagine then.

Figure 6. Reaction mechanism for the passivation of iron using an organic metal (ES) coating. LE indicates the leucoemeraldine form of polyaniline.

Corrosion protection
Applied as a 20-µm thick primer coating, our organic metal shifts a steel (ST37) surface potential from about –400 mV to +250–400 mV, which is quite far into the noble region (16). Moreover, an Fe₂O₃ layer up to 1 µm thick is formed between the steel surface and the primer. We have discovered an interesting reaction: Polyaniline oxidizes the iron to Fe²⁺, and in return is reduced to the leucoemeraldine form, a hydrogenated form (the hydrogen coming from the protons that had been attached to every second nitrogen; see Figure 6). In the presence of oxygen and water, Fe²⁺ is oxidized to Fe³⁺, the leucoemeraldine form is oxidized to the polyaniline in emeraldine base form, and oxygen is reduced to OH^–.

As soon as the stoichiometrically appropriate amounts of Fe³⁺ and OH^– are generated, they form Fe₂O₃, and the poly aniline base is transformed back to the protonated salt. We call this “passivation of iron”, because Fe₂O₃ is the most stable iron oxide (17).

We have developed technically useful primers and coatings based on fundamental studies that are focused on corrosion processes and the passivation mechanism (18). Scanning Volta potential studies tell us about the velocity at which the corrosive electrolyte can penetrate the iron–primer interface and build an iron–salt–water interface—the prerequisite for corrosion.

In our systems, this velocity is <3 µm/h. The potentially corrosive electrolyte comes into contact with the fully oxidized Fe₂O₃, rather than an iron or less-oxidized iron oxide surface. Electrochemical impedance spectroscopy demonstrates the ability of an electrolyte to penetrate the coating and the ability of ions such as Fe²⁺ to migrate toward the outside. We define the transparency (or nontransparency) of coatings for ions by the dielectric constant . Water has an of 75; vacuum has a value of 1. We want a value close to 1; our target range is 1– 12. This value must be stable and not increase over time or upon the action of corrosive agents. Conventional coatings usually have a good starting value of <12, but increases relatively quickly with exposure to corrosive materials, indicating an increasing transparency for electrolytes.

Our coating systems have proven their outstanding performance in many commercial products under harsh corrosive conditions. These coatings are in use in many countries in waste water treatment plants, salt production operations, on ships and boats, in harbor construction materials, in commercial steel structures, and in many other applications. Experience shows that we can expect a lifetime increase by a factor of 2–10 over conventional coatings (19).

Printed circuit board surface finish
No portable phone, coffee machine, television, computer, printer, elevator, or CD player works without printed circuit boards (PCBs). They connect the transistors, resistors, and chips by an intelligent circuit. The electronic components are assembled and soldered on the proper places in the last step of PCB manufacturing. Hence, the next-to-the-last production step must provide a final surface finish that can be soldered without failure. Here, the organic metal offers a new technology, where previously only liquid solder (tin–lead or gold) could guarantee solderability.

We coat the bare copper pads with ~80 nm of a water dispersion of the organic metal, which acts as a catalyst for the electroless deposition of pure tin to form a tin layer only 800 nm thick, at one-third the cost of gold. This layer is lead-free, and it can finish fine-pitch structures (those especially intricate areas that the liquid solder cannot), providing the same solderability as liquid tin–lead or gold solder (20).

Organic and polymeric LEDs
Organic and polymeric light emitters are an exciting development in the manufacture of flat-panel light-emitting displays. They are very thin, come in all colors, can be applied in a patterned structure, are relatively cost-efficient to manufacture, and produce bright pictures (21).

During the development process, it became clear that the emitter should not be placed directly on the anode (usually indium tin oxide, ITO). A hole-injection or anode buffer layer should be placed in between to provide a tuned anode work function and a more even potential and surface topology.

Figure 7. A flexible yellow light-emitting diode (LED) constructed using the organic metal. Used with permission from Ormecon GmbH.

PEDOT and polyaniline were considered promising candidates. It now seems that water dispersions of our polyaniline offer the best performance. These dispersions have an average particle size of only 35 nm, and conductivities can be tuned to match the requirements in LED displays. Light can be emitted at lower voltage, thus using less power; and devices exhibit a longer lifetime than those using any other anode material. We produce this dispersion on a pilot scale and have licensed the product to Covion Organic Semiconductors GmbH (Frankfurt, Germany), the leading manufacturer of light-emitting polymers (Figure 7).

Visions on the horizon
We are convinced that this is not the end. Sensors have been created; polyaniline is being tested as anode material in organic solar cells; and “smart windows” should still be an interesting target for industrial development.

We are heavily involved in molecular simulation studies in our laboratories and hope to resolve the detailed molecular structure soon. We hope this will lead us to an even better insight into the structure–conductivity relationship and allow us to increase the conductivity by another factor of 30, to ~3000 S/cm. This would allow us to provide thin coatings that display an excellent electromagnetic interference (EMI) shielding efficiency. This would be the first step to replacing conventional metals, one of the original dreams of 20 years ago.

"This is our friendliest bird. She was owned by a chemist and answers to 'Polyaniline'."

References

1. Shirakawa, H.; Louis, E.; MacDiarmid, A.; Chiang, C.; Heeger, A. Chem. Commun. 1977, 578.
2. Wessling, B. Makromol. Chem. 1984, 185, 1265–1275.
3. Wessling, B. Electrical Conductivity in Heterogeneous Polymer Systems; In Electronic Properties of Conjugated Polymers; Kusmany, H., Mebring, M., Eds.; Springer Verlag: Berlin, 1988; Vol. 76, pp 407–412.
4. Wessling, B. In Handbook of Organic Conductive Molecules and Polymers; Nalwa, H. S., Ed.; Wiley & Sons: New York, 1997; Vol. 3, pp 497– 632.
5. Wessling, B. Adv. Mater. 1993, 5, 300–305.
6. Wessling B. In Handbook of Conducting Polymers, 2nd ed.; Skotheim, T. A., Elsenbaumer, R. L., Reynolds, J. R., Eds.; Marcel Dekker: New York, 1998; pp 467–530.
7. Wessling, B. In Hand book of Nanostructured Materials and Nanotechnology; Nalwa, H. S., Ed.; Academic Press: New York, 1999; Vol. 5, pp 501– 576
8. Wessling, B. Polym. Eng. Sci. 1991, 31, 1200–1206.
9. Wessling, B. Z. Phys. Chem. 1995, 191, 119–135.
10. Subramaniam, C. K.; Kaiser, A. B.; Gilberd, P. W.; Wessling, B. J. Polym. Sci., Part B: Polym. Phys. 1993, 31, 1425–1430.
11. Pelster, R.; Nimtz, G.; Wessling, B. Phys. Rev. B 1994, 49, 12718– 12723.
12. Subramaniam, C. K.; Kaiser, A. B.; Gilberd, P. W.; Liu, C.-J.; Wessling, B. Solid-State Commun. 1996, 97, 235–238.
13. Wessling, B.; Srinivasan, D.; Rangarajan, G.; Mietzner, T.; Lennartz, W. Eur. Phys. J. E 2000, 2, 207–210.
14. Lennartz, W.; Mietzner, T.; Nimtz, G.; Wessling, B. Morphological Changes in PAni–PMMA Blends during Dispersion Studied by SAXS. Presented at the International Conference on the Science and Technology of Synthetic Metals (ICSM) 2000, Bad Gastein, Austria, July15–21, 2000.
15. Groenendaal, L.; Jonas F.; Freitag, D.; Pielastzik, H.; Reynolds, J. R. Adv. Mater. 2000, 12, 481–494.
16. Wessling, B. Adv. Mater. 1994, 6, 226–228.
17. Wessling, B. Mater. Corros. (Engl. Transl.) 1996, 47, 439–445.
18. Posdorfer, J.; Wessling, B. Fresenius’s J. Anal. Chem. 2000, 367, 343– 345.
19. Port, O. Business Week, Sept 4, 2000, p 78.
20. Morawska, Z.; Koziol, G. Lead-free solderability preservative coatings of PCBs; Printed Circuit Board, in press.
21. Burroughs, J.; Bradley, D.; Brown, A.; Marks, R.; Mackay, K.; Friend, R.; Burn, P.; Holmes, A. Nature 1990, 347, 539.

---

Bernhard Wessling is president and managing partner of Ormecon Chemie (Ormecon Chemie GmbH & Co. KG, Ferdinand-Harten-Strasse 7, D-22949 Ammersbek, Germany; +49-40-604-106-18; wessling@zipperling.do.uunet.de). He received his Ph.D. in chemistry from the University of Bochum, studying natural compound synthesis and circular dichroism. He has worked in the polymer industry, combining basic and applied research on dispersions in polymer systems. He was managing director, shareholder, and CEO of Zipperling Kessler until 1996, when Zipperling founded Ormecon Chemie as a fully owned subsidiary, and he took his present position. He has filed numerous international patents and published several dozen scientific and technical papers.

 

 © 2001 American Chemical Society.  

---