Chemistry Reference
In-Depth Information
sample is deformed slowly enough to allow such movements to take place at the
experimental temperature. Such deformation will not be recovered when the stress
is released if the experiment has been performed above
T
g
. If the rate at which
the specimen is deformed in a particular experiment is too fast to allow the mac-
romolecular segments to respond by flowing, the polymer will be observed to be
glassy. It will either break before the test is completed or recover its original
dimensions when the stress is removed. In either event, the experimental tempera-
ture will have been indicated to be below
T
g
. As a consequence, observed glass
transition temperatures vary directly with the rates of the experiments in which
they are measured.
The
T
g
values quoted in
Table 4.2
are either measured by very slow rate meth-
ods or are obtained by extrapolating the data from faster, nonequilibrium tech-
niques to zero rates. This is a fairly common practice, in order that the glass
transition temperature can be considered as characteristic only of the polymer and
not of the measuring method.
Many relatively slow or static methods have been used to measure
T
g
. These
include techniques for determining the density or specific volume of the polymer
as a function of temperature (cf.
Fig. 4.1
) as well as measurements of refractive
index, elastic modulus, and other properties. Differential thermal analysis and dif-
ferential scanning calorimetry are widely used for this purpose at present, with
simple extrapolative corrections for the effects of heating or cooling rates on the
observed values of
T
g
. These two methods reflect the changes in specific heat of
the polymer at the glass-to-rubber transition. Dynamic mechanical measurements,
which are described in
Sections 4.7.1 and 4.8
, are also widely employed for locat-
ing
T
g
.
In addition, there are many related industrial measurements based on softening
point, hardness, stiffness, or deflection under load while the temperature is being
varied at a stipulated rate. No attempt is usually made to compensate for heating
rate in these methods, which yield transition temperatures about 10
20
higher
than those from the other procedures mentioned. Some technical literature that is
used for design with plastics quotes brittleness temperatures rather than
T
g
. The
former is usually that temperature at which half the specimens tested break in a
specified impact test. It depends on the polymer and also on the nature of the
impact, sample thickness, presence or absence of notches, and so on. Since the
measured brittleness temperature is influenced very strongly by experimental con-
ditions, it cannot be expected to correlate closely with
T
g
or even with the impact
behavior of polymeric articles under service conditions that may differ widely
from those of the brittleness test method.
Heat distortion temperatures (HDTs) are widely used as design criteria for
polymeric articles. These are temperatures at which specimens with particular
dimensions distort a given amount under specified loads and deformations.
Various test methods, such as ASTM D648, are described in standards compila-
tions. Because of the stress applied during the test, the HDT of a polymer is
invariably higher than its
T
g
.
