This paper has been submitted at Nov 4th, 1996, for publication in a
scientific journal and is now subject to the reviewing process.
You may ask for a hardcopy with the complete manuscript, which contains all figures and all
greek symbols.
Everybody
is invited to comment to this paper. I will consider every argument and
comment back.
This is the 3rd revised version.
Dr. Bernhard Wessling
Zipperling Kessler / Ormecon Chemie
D-22949 Ammersbek
4. Discussing the "solution hypothesis" again
4.1 What indicates that the systems could be true solutions?
I will now try to consider the option, that such polyaniline-solvent systems were real solutions, even though the consequences from 2.1 and 2.2 do not any more allow this. What is the experimental background provided by the scientific community for the conclusion or hypothesis of "solution"? To be open, there are only a few weak arguments.
For most people in the field it is sufficient to state, that a solvent/polyaniline system is "clear" (green). But as there are a lot of absolutely transparent, crystal clear dispersions (beginning with Faraday's gold sol, up to micro-dispersions of oil/water/selected detergents[24] are clear and stable!). This observation is not at all convincing.
The first publication with the message "conducting polymer solutions are feasible" came from J. Frommer[25]. She claimed to dissolve poly-(phenylene sulfide) in AsF3 during "doping" with AsF5. It was mentioned being important that the "doping" is performed in the solvent; first doping and then bringing the product in contact with the solvent only a few seconds later "results in incomplete dissolution". This observation was interpreted by cross-linking.
Later emeraldine base (EB), the base form of polyaniline was treated similarly when liquids capable of both "doping" and "dissolving" were used, H2SO426 and m-cresol[27]. In the first case the authors also presented viscosity curves[28] and stated, that the viscosity dropped over time (when letting the solution stay around) which was interpreted by a decrease in molecular weight. It was not considered as a possibility that - in case they would deal with a dispersion - the particle size could be changed over time, here: increased by aggregation. Such a process would also decrease the viscosity significantly.
We do not yet see any indication of a real dissolution effect of these doping solvents. One should have to look for other experiments which could prove this assumption and disprove my proposal of that the solvent is first "doping" and secondly wetting the particle surface. Isn't it reasonable to assume that a liquid which is able to react with EB in form of an acid-base reaction or also in form of an oxidation (like m-cresol, or AsF4, which could be the active species when having AsF3 in AsF5) is also able to wet the resulting particles?
There are two publications in which some serious arguments supporting the "solution hypothesis" have been developed from extensive experimental studies:
4.1.1. Dynamic light scattering of PAni-CSA
Heeger et al.[29] have published a "solution characterization of surfactant solubilized polyaniline". We might first ask, what the term "surfactant solubilized" might have as a meaning, as "surfactants" are generally active at a surface, reducing their surface tension - but the authors are claiming a "solubilization", hence no surface to be treated! But from the text it becomes clear, that they believe that "PAni-CSA exists as single chains in solution at low concentrations in m-cresol".
The dynamic light scattering study was performed by comparing the results with those of polyelectrolytes. But here, the ions are quasi infinitely separated, which - as we know from section 2.3 - cannot be made with polyaniline. Also the authors admit, that "the previous work on polyelectrolyte-surfactant solutions will give insights into the solution properties of PAni-CSA", but they are nevertheless using equations derived for such systems for interpreting their measurement results.
One aspect which is clearly different between the polyelectrolyte solutions and the PAni systems investigated, is that PAni is insoluble without the addition of, what they call "surfactant", whereby the polyelectrolyte is soluble without "surfactant".
The group reports first about the preparation procedure. It should be asked, why it was necessary to "sonicate" (i.e.: treating the PAni/CSA mixture in m-cresol at very low concentration like 0.01% under ultrasound) for 5 hours! The use of ultrasound is a widely used technique for preparation of colloids[30], where the colloidal particles are hard to disperse. We have described this technique for the dispersion of conductive polymers as early as 1984[31].
Despite this very high energy input, the "solution" had to be filtered through millipore filters. And even then, there was evidence for a certain degree of aggregation, as was detected from the slower of two relaxation effects. This relaxation was attributed to large particles convecting through the laser beam. A particle diameter was not given.
The faster relaxation was attributed to "single chain polymer diffusion". It remains unknown from where the authors concluded, that they were observing "single chains", even more as they had shown, that the particles which were detected did not show a behaviour characteristic of extended (more rigid) rods, which one would expect from solvated PAni chains.
What was clearly observed by the authors was (1) a hydrodynamic radius of between 10 and 50 nm (2) an increase of the radius with decreasing PAni concentration (3) an increase of the radius with increasing amount of water in m-cresol (4) no "chains" or "particles" were observable with a significant difference between length and diameter (hence the particles or chains were more or less globular).
Their interpretation of these findings was: the PAni is completely dissolved (although they mentioned "some degree of aggregation"), and the PAni chain is present as a more random coil, not as a rigid rod. They concluded this from the difference of the hydrodynamic radius in the different concentrations of PAni. This behaviour was analogous to phenomena found for polyelectrolytes. But it remains unclear whether one can deduct anything valuable from this apparent analogy. From the difference in hydrodynamic radius they concluded a varying persistence length of the chain ranging from 0.5 nm (!?) to 9 nm (so over a factor of 20) depending of a concentration difference between 0.01 % and 0.001%. In other words: It was claimed, that the random PAni coil is completely folded at 0.01% and completely defolded (and more stretched) at 0.001%.
It seems at least questionable, if the conclusions are compelling. In contrast, the observations can also be interpreted by assuming dispersed particles. This can be seen when taking into account
- the necessity of sonicating
- the presence of bigger particles even at such low concentrations
- the general observation, that dispersion using ultrasound is more efficient when the medium has a higher viscosity (i.e.: at the higher concentration of 0.01%, the shear rate induced by the sonotrodes is higher, leading to a better dispersion, cf [ nonequi 2]
- this accounts for the dependance of particle size from concentration
- the presence of water (becoming adsorbed on the PAni particles) decreasing the dispersion ability of m-cresol, which leads to bigger particles (cf. section 3.)[32].
It also seems unreasonable, that PAni, dissolved in a single chain, should be present as a random coil, and not as a rigid rod. But even more it must be asked if the persistence length can vary so drastically at so relatively low concentrations.
It must be stressed that a direct confirmation of single solvated chains has not been given in this publication.
4.1.2. Liquid crystalline "solutions"
Another very interesting observation was published by Y. Cao and P. Smith[33]: liquid crystalline "solutions" (as they claimed" of electrically conducting polyaniline. Again they studied PAni-CSA in m-cresol and found liquid crystalline behaviour in contrast to emeraldine base in m-cresol.
First it must be known that birefringeance and other characteristics of liquid crystalline systems are not limited to solutions. In fact, these phenomena are widely known and described for many colloidal systems: everywhere, where a high shear (which was obviously applied by sonicating) is induced in viscous media of colloidal particles, which are anisodimensional, the inner friction of the systems leads to an orientation leading to liquid crystalline behaviour.
In contradiction to the (later) findings of the same group from dynamic light scattering, Cao and Smith claimed: "Apparently, complete protonation by strong acids was essential for transforming the polyaniline chains from flexible coils (in emeraldine base) to more rigid entities and, consequently, for the formation of liquid-crystalline solutions of polyaniline."
But when looking on our model for colloidal systems of PAni, we had deducted a certain anisotropy from other experiments [13] and hypothesized a structure for the adsorbed solvent layer on the PAni particles, leading to an orientation of them in dispersions and gels (fig. 10.2).
In fact, the assumption of real solutions of intrinsically conductive polymers has not well been supported by experimental results yet, and only very few attempts have been made in that direction, probably because the authors didn't realize the importance of this question.
