 |
| painting by Baptiste Ibar (2010) "Dual-Split" http://www.baptisteibar.com |
My first attempts to influence the local structure of
molded polymers with vibrational means were done in the basement of a MIT lab,
way back in 1973, when I and a fellow student vibrated loud speakers (his) attached
to a thin aluminum plate positioned on top of a heated plaque of polystyrene,
while cooling those with dry ice. We blasted Pink-Floyd music as loud as we
could hoping to modify the free volume content and distribution ! The plaque did look very different under
polarized light. Yet, the idea was found ludicrous by my Ph-D supervisor who
requested to stop such experiments.
But I was intrigued by the potential of the use of vibration
in molding processing and refined my original idea a couple of years after I
left MIT. I convinced investors to build
me a lab where I could design, build and test a vibratory compression molding
machine I had in mind (I called it a “bb-machine” because it was designed to
modify the amount of bb-grains in the glassy state-see later). I worked for several years on small round
specimens of polystyrene and polypropylene, 2 to 3 mm thick, submitting them to
a pressure vibration history while cooling from a high molten temperature until
they had solidified. The vibro-molded specimens were then cut into ASTM tensile
bars and tested mechanically. I observed the important influence of vibration
on mechanical properties, in particular a spectacular improvement in the
stiffness and strength of the vibrated samples. These first results and
concepts were published in a 1979 French patent letter, the first in a series
of patents dedicated to the use of vibrational means during molding to modify
properties of molded products. I called his new technology Rheomolding to emphasize the
particular importance of rheology on molding.
The main idea behind Rheomolding was the ability to raise at will the
value of phase transitions by combining pressure and vibration, hence to
provide means to “ Rheo-cool”, i.e. go beyond the limitations of cooling by
conduction to produce super fast quench rates, regardless of sample thickness
(see the Blog of September 14, 2012). I assumed that the large improvements in
strength and stiffness observed for vibro-molded specimens were due to the fast
Rheo-cooling rates, only achievable by combining true cooling rate and the
effect of a change of vibration
frequency on the Tg or Tm of the polymeric melt.
In a second phase of my research on the influence of
vibrational means on molding, I designed and built several prototype machines
capable of applying the vibrational pressure history to classical molding
operations, such as injection molding and extrusion. This resulted in a new
series of patent disclosures. The technology required that processors buy
special equipment mounted on their existing extruders and injection molding
machines, and be trained to know how to implement the resin specific
vibrational parameters. One draw back of this new technology became apparent:
any attempt to use vibration during molding without specific guidance to the
correct frequency, amplitude of vibration and timing sequence, could
potentially result in the degradation of the mechanical performance, not its
improvement. Besides, on top of being rather difficult and expensive to
install, the new technology did not raise the interest of resin manufacturers
to the level expected, because, as one resin manufacturer put it:” once a customer has bought the proper
equipment to be able to Rheomold my resin, what will prevent him from buying
from my competitors?”.
Rheomoldingtm technology has become public property in Europe and throughout the world since September
2001.
In 1994, I left the Company I had created to
concentrate on the theoretical side of polymers. I worked for two years on perfecting
the theoretical model I had developed at MIT in my Ph-D thesis and carried out simulation
work of melt processing by computer, using C-Mold and working with the team of
Prof. K.K. Wang at Cornell. During that
time, I also published a review of past work regarding manipulation of melt
rheology by vibration, and a controversial paper questioning the validity of
existing theories of polymer physics, especially with regard to spectroscopic and
dielectric data (Do we Need a New Theory
in Polymer Physics?).
Meanwhile, I continued to search for alternative ways
to apply melt vibration to a melt without the inherent drawbacks attached to
Rheomolding. In 1997, I was granted a new patent on vibrating gas inside (and
outside) the mold cavity to produce a means to vibrate a melt during the
filling stage of an injection molding process. This made vibration benefits
available to the small processors, because of the small costs associated with
the Vibration Gas Assisted Molding equipment (VIBROGAIMtm).
About 10 years
ago, in 1999 and 2001, I was granted two new patents related to an innovative
technology of “disentanglement” of macromolecules to produce an ultimate
control of viscosity of molten plastics prior to and/or during a molding operation.
The technology (which I called EZ-Flow) was based on a new concept of entanglements
derived from a generalization of the statistical model of conformers in collective
interactions I had introduced in the lectures at Cornell. I had made a true
breakthrough in inventing a new way to introduce the vibration of the melt
without the need to have an external vibrator involved (although this was still
a complementary option). This was done
through the insertion and the disposition of i-ribs (‘intelligent ribs) on the
surfaces touching the melt as it passed through. I operated those i-ribs
machines for several years, proving that I was right: it could decrease the
melt viscosity of polymers and their blends by several orders of magnitude even
preserve some of the viscosity reduction in pellets made up at the end of the
treatment line.
The technology resulted in major benefits to plastic
processors , especially to compounders who battled with viscosity issues as
they increased the amount of additives (fibers, nanoparticles) in their
masterbatches. The technology was very successful, yet the machines were big
and expensive and limited to rather modest throughput rates, because the
treatment efficiency depended on the residence time in the disentangling
processor, a major drawback to address the high throughput rates of the
commercial compounding lines (several tons per hour). Another breakthrough had to
be made.
In 2007, I started an academic career, dedicating my
time understanding what I thought were extraordinary results which the current
theoretical models could not predict nor explain. My first task was to
understand in depth the existing rheological models and question the meaning of
the experiments I had conducted. This is why I studied with a lab dynamic
rheometer the effect of strain and how
it differed or coupled with the effect of strain-rate (frequency). By focusing
more and more on the way entanglements were mathematically described by the
current models, I realized the limitations and shortcomings of their definition,
especially under deformation conditions which brought the melt into the “non-linear”
visco-elastic region, i.e. at higher strain.
By the end of 2009 I had read and studied all
the theses of the University of Pau which dealt with polymer flow and rheology.
I also studied the relevant references. I was now totally convinced that the
established models of entanglements were too simplistic and missed the
fundamental understanding of the concept of chain components interactions. I started to believe and suggest that a new
theory of polymer physics had to be presented not only to account for the
spectroscopy results discussed in my paper of 1997 (Brillouin scattering, RMA
results, Low frequency Raman), but also to account for rheology in the
non-linear region.
I generalized the equations of the Dual-Phase statistics,
which applied to a macro-coil system of conformers of size M < Me and discovered that it gave birth to another level of split,
the two dual-phase statistics, which I coined the Cross-Dual-Phase Statistical model.
The interesting result was that the cross-dual-phases would only hold stable if
the interpenetrating macro-coils were of a size superior to a critical value Me.
This provided a definition of Me as the critical system size for a split of the
dual-phase to occur and remain stable, an entirely new concept. The network of
entanglement resulted from the interpenetration of one of the two dual-phases
into the other, defining a network of channeling boundaries perpetually in
motion. Using that definition of the entanglement network, I was able, in 2011, to develop the rheological
equations responsible for linear and non-linear effects such as shear-thinning,
normal force and strain softening. The width of the rubbery plateau with molecular
weight, as well as the very existence of the rubbery plateau itself, could also
be understood and derived from the same
concepts.
Finally, I showed in 2012 that the sacred-saint exponent 3.4 found empirically for the molecular weight
dependence of the Newtonian viscosity, which had driven so much attention from the
de Gennes school of reptation, was in fact ill-defined since it did not represent the sole influence
of the molecular weight alone, incorporating also a mix of free volume contribution.
The real molecular weight dependence was characterized by an exponent 5.3, not
3.4, at constant free volume contribution.
All in all, it appeared that the new physics which
gave birth to the new concept of entanglement network through the stability of
the crossed-dual-phase, was better adapted to the description of the
experimental results I had obtained combining the effect of shear-thinning and
strain softening, which I had called “Rheo-Fluidification”.
The innovation really came from the new statistics of
the conformers I created to describe the interplay between the inter and intra
molecular interactions between the conformers. Initially, in my thesis, I had
described the conformational state of the bonds by a series of circularly
bounded chained kinetic equations, the
way it is used in chemistry in the case of multi-stage chemical reactions. But,
a few years later, after dropping the assumption that local order could
stabilize thermodynamically the systems of interactions, I introduced a
dissipative term between two “splitted” kinetic equations that provided the
rate dependence of the population of the conformers. The presence of the dissipative
term coupled the (b/F) statistics with the conformational statistics (transßà gaucheßàcis). This became the Dual-Phase statistics (the two
phases being the b and F “local” phases), also called the Dual-Split Kinetics,
which described the properties of single macro-coils and of interpenetrated
macro-coils for M < Me.
The solution of the dual nature of the interactions
between the conformers (which were covalently intra-molecularly bonded along
the chain and inter-molecularly interacting with neighboring conformers) was
therefore represented by the formation and the fluctuation of bb-grains
surrounded by F-conformers. The grain structure of the b-conformers came about
intuitively from the notion that they locally interacted, forming, at least
temporarily, a cooperative unit. The cohesion of the melt resulted from the local
fluctuation of the grain formation and dissolution, giving the appearance of the
delocalization of the grain structure in space (this was not true, of course,
below Tg). At first, I associated the fluctuation with the thermal motion,
but I recently understood, with the
introduction of the Grain-Field Statistics, that this was not a trivial issue at all that it was, in fact, the most
important theoretical problem I had to address and understand (with
ramifications in all branches of physics).
Consider a “normal statistics”: the energy level
between two states is constant, say DEo , which defines an homogeneous field. I define a Granular-Field a
fluctuating field, for instance a series of DEo values and 0 values over dt time intervals. The Grain-Field parameters would be associated with
the amplitude DEo and the frequency
corresponding to dt. This field is not homogeneous, it is granular. The
granular aspect, in time, of the field can be described via a Fourier transform
into a series of sine waves, so the problem of studying the response generated
by a granular field is, in fact, equivalent to studying a variable field such
as DE= DEoi +DEa sin(wt+Q). This is a fascinating research on its own, which
I will publish separately.
What interests me, in this communication, is the
impact of introducing a grain-field into the dual-split kinetic equations,
especially to describe the interactions between two interpenetrating macro-coil
systems, leading to their self-diffusion. As I said above, the stability of the
Cross-Dual-Phase solution is the consequence of the existence of the Granular
Field. The very existence of
entanglements would be due to the granular aspect of the field defining the conformational
statistics!
Also, to make a long story short, it can be shown that
the local grain structure, giving rise to the localization or delocalization of
the bb-grains, derives from the fluctuation of the conformational field. The frequency of the local fluctuation of
density (due to the (b/F) <---> (c,g,F) statistics) is not simply due to thermal
fluctuations, it is correlated to the parameters of the Grain-Field.
The Pink-Flow technology originated from this new
understanding of the interactions between the conformers (summarized as “entanglements”)
which came about from the theoretical work I conducted , making granular the
conformation field of the dual-split equations. This was 2010-2011. It now appears
possible to determine a specific set of processing conditions which can modify “plastically”
the network of phase-entanglement (versus “visco-elastically” in the 1st
generation of disentanglement processors) so that the major drawbacks of the first
generation of machines could be overcome. The 2nd generation Rheo-Fluidification
technology is compact and adapted to the high throughput rates of the industry.
Its efficiency no longer depends on the
residence time in the Rheo-Fluidizer. Validation tests are the next phase.
The GOOD VIBRATIONS are still on.
