230
S. Napolitano
Ultrathin films were prepared by spincoating solutions of PET in a mixture (5:2)
of trifluoroacetic acid (TFA) and chloroform. Optimizing the relative concentration
of the two liquids was possible by considering that TFA is a good solvent for PET,
while chloroform increases the volatility of the mixture. Using solutions of PET in
pure TFA for spincoating resulted in films with an undulated surface. This feature
is usually attributed to a nonhomogenous distribution of solvent in the film during
spinning, probably related to the low volatility of TFA. Adding chloroform permits
to decrease the surface roughness of PET. Samples as prepared were kept for 4 h at
353 K in order to remove residual solvent and reduce stresses induced by the spincoating preparation. No amorphization was necessary in this case, measurements via
atomic force microscopy and X-rays proved that the used annealing conditions do
not result in the formation of crystalline structures [54].
In the thickness range considered (7 nm–10 μm), the crystallization rate varies by
more than five orders of magnitude. Investigating this enormous range of crystallization times requires a serious feasibility study before performing the dielectric experiments. Owing at investigating all the samples at the same experimental conditions,
that is, the same annealing/crystallization temperature, the dielectric function was
recorded at 373 K, where the characteristic crystallization time, t cry , is ~7 min in bulk
samples. Performing experiments at lower temperatures might have given a better
control of the measuring conditions for bulk samples, but would have led to tremendously slower kinetics in the thinnest films; in addition to reducing the maximum
number of samples to investigate within a given project time, measuring at lower
temperatures is also prohibitive because of the larger consumptions in exchange gas
used, and the finite volume of LN2 tanks, if employed. Working at higher temperatures, on the contrary is limited by the frequency window where the segmental
relaxation is not affected by artifacts.
At 373 K, the structural relaxation process of PET appears as a strong peak in the
imaginary component of the dielectric function, centered around 10 kHz (for bulk)
(i.e., τ ~ 16 μs). As observed for PHB, ε decreased during isothermal annealing;
see Fig. 5. While for thick films this phenomenon could be directly explained in
terms of crystallization, that is, via a drop in ε following an Avami-like law of the
type [23, 55]:
ε ∼ −exp
−t/t cry
β
(3)
with 1 < β < 4. in the case of thinner films and extra term should be included. Prior to
crystallization, a reduction in ε not following Eq. (3) is observed for films thinner
than ≈500 nm. This extra term is linear in a plot of ε versus logt and increases
in amplitude upon reduction of the thickness; see Fig. 5. Such further reduction
in dielectric strength was associated with interfacial rearrangements increasing the
monomer/surface density without affecting the number of chains adsorbed per unit
surface [56]. Same as in the case of crystal perfection, this process yields an increase
in the number of monomers directly adsorbed onto the substrate, without affecting
the total adsorbed amount, that is. Under such conditions, the number of monomers
S. Napolitano
Ultrathin films were prepared by spincoating solutions of PET in a mixture (5:2)
of trifluoroacetic acid (TFA) and chloroform. Optimizing the relative concentration
of the two liquids was possible by considering that TFA is a good solvent for PET,
while chloroform increases the volatility of the mixture. Using solutions of PET in
pure TFA for spincoating resulted in films with an undulated surface. This feature
is usually attributed to a nonhomogenous distribution of solvent in the film during
spinning, probably related to the low volatility of TFA. Adding chloroform permits
to decrease the surface roughness of PET. Samples as prepared were kept for 4 h at
353 K in order to remove residual solvent and reduce stresses induced by the spincoating preparation. No amorphization was necessary in this case, measurements via
atomic force microscopy and X-rays proved that the used annealing conditions do
not result in the formation of crystalline structures [54].
In the thickness range considered (7 nm–10 μm), the crystallization rate varies by
more than five orders of magnitude. Investigating this enormous range of crystallization times requires a serious feasibility study before performing the dielectric experiments. Owing at investigating all the samples at the same experimental conditions,
that is, the same annealing/crystallization temperature, the dielectric function was
recorded at 373 K, where the characteristic crystallization time, t cry , is ~7 min in bulk
samples. Performing experiments at lower temperatures might have given a better
control of the measuring conditions for bulk samples, but would have led to tremendously slower kinetics in the thinnest films; in addition to reducing the maximum
number of samples to investigate within a given project time, measuring at lower
temperatures is also prohibitive because of the larger consumptions in exchange gas
used, and the finite volume of LN2 tanks, if employed. Working at higher temperatures, on the contrary is limited by the frequency window where the segmental
relaxation is not affected by artifacts.
At 373 K, the structural relaxation process of PET appears as a strong peak in the
imaginary component of the dielectric function, centered around 10 kHz (for bulk)
(i.e., τ ~ 16 μs). As observed for PHB, ε decreased during isothermal annealing;
see Fig. 5. While for thick films this phenomenon could be directly explained in
terms of crystallization, that is, via a drop in ε following an Avami-like law of the
type [23, 55]:
ε ∼ −exp
−t/t cry
β
(3)
with 1 < β < 4. in the case of thinner films and extra term should be included. Prior to
crystallization, a reduction in ε not following Eq. (3) is observed for films thinner
than ≈500 nm. This extra term is linear in a plot of ε versus logt and increases
in amplitude upon reduction of the thickness; see Fig. 5. Such further reduction
in dielectric strength was associated with interfacial rearrangements increasing the
monomer/surface density without affecting the number of chains adsorbed per unit
surface [56]. Same as in the case of crystal perfection, this process yields an increase
in the number of monomers directly adsorbed onto the substrate, without affecting
the total adsorbed amount, that is. Under such conditions, the number of monomers