Ph.D. Thesis
Title The origin of the penultimate unit effect
in free-radical copolymerization
Adviser Professor Thomas P. Davis
Thesis Committee T. Fukuda, Institute for Chemical Research,
Kyoto University, Kyoto, Japan; K. Matyjaszewski, Dept. of Chemistry,
Carnegie Mellon University, Pittsburgh PA, USA; G. Moad, CSIRO Molecular
Science, Clayton South, Victoria, Australia.
Essay
Free radicals are fundamentally important in a wide
range of chemical and biological processes, including the degradation
of synthetic and biological polymers, stratospheric and atmospheric
processes, several important enzymatic processes such as photosynthesis
and the biosynthesis of DNA, and even the growth of complex chemicals
in interstellar clouds. The objective of my PhD research was to contribute
to a deeper understanding of the fundamental influences on radical
reactivity and selectivity, with a specific focus on remote substituent
effects.
My research was motivated by the specific problem of
kinetic modeling in free-radical copolymerization. Free-radical polymerization
is one of the most important bulk chemical industries and copolymerization
is often utilized to impart the properties of more than one type of
monomer to a product. The properties of the polymer depend heavily
on such things as its molecular weight distribution, chemical composition,
and sequence distribution -and these are governed by the rates of
the various reactions that occur during the polymerization process.
Modeling copolymerization is very complicated as, not only are there
different classes of reaction to consider (such as initiation, propagation,
termination, and chain-transfer), but the radicals undergoing these
reactions can also have countless different chemical compositions
(being composed of various combinations of the different monomer units).
To develop useful models it is necessary to make simplifying assumptions
about the fundamental influences on radical reactivity, so that the
large set of chemically different reactions can be grouped together
into a smaller number of sets of kinetically equivalent reactions.
For nearly 50 years it was assumed that remote substituent
effects were unimportant in the propagation step of free-radical polymerization,
and a simplified model based on this assumption -the terminal model-
was adopted for modeling the composition, sequence distribution and
overall propagation rate. By treating the reactivity ratios of the
monomers as adjustable parameters, this model could be made to fit
the composition data for almost all systems tested. However, although
widely used, the model was not critically tested until the seminal
work of Prof. Fukuda in 1985. Unlike previous workers, Fukuda actually
used the parameters obtained from a terminal model fit to composition
data, to predict the propagation rate coefficients for the same system,
and compared them against independently measured values. When subjected
to this critical testing, the model failed comprehensively. This failure
has since been demonstrated for a number of other systems, and it
is now generally accepted that the terminal model cannot be made to
describe simultaneously the composition and propagation rate coefficients
of ordinary copolymerization systems.
While it is now clear that the terminal model is incorrect,
there has been a reluctance to call into question the large number
of terminal model reactivity ratios that have been measured over the
previous 50 years, as well as the numerous (terminal-model based)
empirical schemes for explaining reactivity in free-radical polymerization
that have been proposed. In order to retain this earlier work, Fukuda
proposed a compromise model in which remote substituent effects were
considered to be unimportant in models for copolymer composition but
important in models for propagation kinetics. In chemical terms this
amounted to the assumption that the penultimate unit affects radical
reactivity but not selectivity. This 'restricted penultimate' model
has now been widely adopted, but the physical validity of its assumptions
had not, until my work, been tested.
My research was an extensive theoretical and experimental
examination of the nature of the penultimate unit effect, with the
aim of establishing whether or not such effects exist, and whether
they are more important in radical selectivity or reactivity. This
would facilitate a critical test of Fukuda's restricted model and
would thus establish whether or not the previous 50 years of work
based upon the terminal model is valid.
In a series of high-level ab initio molecular orbital
calculations of the reaction barrier (and other related quantities)
in a series of model propagation reactions, I provided direct evidence
for the existence of significant penultimate unit effects. Furthermore,
the calculated effects were strongly dependent on the nature of the
reacting monomer (ie. there were effects on selectivity), and were
likely to be polar in origin. Although my calculations revealed significant
penultimate unit effects on radical stability, such effects largely
canceled from the reaction barrier owing to the early transition state.
I thus showed that penultimate unit effects on selectivity are much
more significant than those on reactivity, which in turn shows that
Fukuda's restricted model -and thus the terminal model composition
equation- is not physically valid.
In order to provide experimental support for this result,
I also performed a series of kinetic studies. I initially attempted
to discriminate between alternative versions of the penultimate model
purely on their capacity to describe simultaneously propagation and
composition data for typical co- and ter-polymerization systems. However,
despite compiling the most extensive data-sets to date, using the
most accurate available experimental methods, I found that, owing
to their adjustable parameters, any number of models could be made
to fit the data. Nonetheless, by designing model copolymerization
systems, such as pairs of sterically similar but electronically different
para-substituted styrene systems, I was able to provide direct evidence
against the restricted model. In an additional study, I showed that
the penultimate unit effect is temperature dependent, and thus has
a significant enthalpic component, which (coupled with previous studies
of solvent effects) is further evidence for polar penultimate unit
effects.
In conclusion, my research provided evidence for the
existence of penultimate unit effects, and indicated that these effects
will be more important in selectivity than reactivity. These results
imply that the failure of the terminal model to describe propagation
rate coefficients must be taken as evidence of its failure to describe
composition data. This result not only entails that current kinetic
modeling of free-radical polymerization needs to revised, but it also
suggests that many of our fundamental ideas about radical reactivity
in general -in particular, about the importance of remote substituent
effects- need to be re-examined .