High-Pressure Crystallization of Glass-Forming Liquids …
37
slows down and moves out of the experimental window as the glass transition is
approached. Since ε
(T ) dependences collected with the cooling rates from 0.1 K/min
up to 5 K/min are almost identical, one can suppose that PC indeed easily reach the
glassy state. On the other hand, it should be also remembered that the presence of
only a small crystalline fraction might be hardly detectable in the dielectric response
of the bulk material.
Subsequent heating starting from the glassy state (Fig. 7b) reveals dramatic
changes in the crystallization behavior of PC. In this case, recrystallization event
is detected as a sudden drop of the dielectric permittivity, with the onset that shifts
towards higher temperatures with increasing the heating rate. We associate such a
sudden drop in the dielectric permittivity with reducing the number of the reorientating dipoles as the liquid volume fraction decrease. In each case, the crystal melts
at the same temperature, T m = 218 K, indicating for the presence of only one polymorphic form. From the results presented above, it becomes evident that preventing
crystallization of PC upon heating from the glassy state is far more complicated than
on cooling a liquid from the melt. This type of behavior is very typical for numerous
molecular liquids and agree with the schematic picture of the nucleation and growth
rates maxima located at different temperature regions, as presented in Fig. 2.
Similarly, we have performed compression/decompression rate-dependent studies
carried out for PC under isothermal conditions, T = 243 K. The results of these
experiments are collected in Fig. 7c, d, respectively. In analogy to cooling rate dependent scans, we have also observed a characteristic dispersion curve in the dielectric
response of the pressurized sample, indicating slowing down of the molecular movements and approaching the glassy state. For the two slowest compression rates (2.6
and 5.8 MPa/min) at least partial crystallization of PC was observed, as both curves
deviate from all the others. This was additionally certified by comparing the values
of the dielectric permittivity for the glassy and crystalline materials obtained on
increased pressure [61]. Upon decompression from the glassy state, just like when
heating the glassy sample, crystallization is far more difficult to avoid and this requires
higher scanning rates (see Fig. 7d). Nevertheless, the values of ε
do not drop down
completely to that characteristic for the crystalline material (~2.55), meaning that
there is still some substantial amount of the liquid volume fraction in the depressurized material. Using decompression rate of 154 MPa/min, we have not observed
any traces of the PC crystallization which indicates that at such high-compression
rates there is not enough time for the nuclei to grow. Therefore, we conclude that
by varying with compression/decompression rate it is possible to affect crystallization/vitrification tendency of the molecular liquids, same as when using the temperature. The results presented in Fig. 7d also reveal only one melting event (p m =
200 MPa at 243 K), which is again consistent with the heating rate dependent studies
carried out at ambient pressure. However, it has been also observed—especially for
ice phases—that different compression/decompression rates might induce elusive
polymorphic transformations [70, 71]. The increased tendency of the depressurized
PC to crystallization conform the schematic picture of the nucleation and growth
rates maxima located at slightly different pressure regions, as introduced in Fig. 6.
37
slows down and moves out of the experimental window as the glass transition is
approached. Since ε
(T ) dependences collected with the cooling rates from 0.1 K/min
up to 5 K/min are almost identical, one can suppose that PC indeed easily reach the
glassy state. On the other hand, it should be also remembered that the presence of
only a small crystalline fraction might be hardly detectable in the dielectric response
of the bulk material.
Subsequent heating starting from the glassy state (Fig. 7b) reveals dramatic
changes in the crystallization behavior of PC. In this case, recrystallization event
is detected as a sudden drop of the dielectric permittivity, with the onset that shifts
towards higher temperatures with increasing the heating rate. We associate such a
sudden drop in the dielectric permittivity with reducing the number of the reorientating dipoles as the liquid volume fraction decrease. In each case, the crystal melts
at the same temperature, T m = 218 K, indicating for the presence of only one polymorphic form. From the results presented above, it becomes evident that preventing
crystallization of PC upon heating from the glassy state is far more complicated than
on cooling a liquid from the melt. This type of behavior is very typical for numerous
molecular liquids and agree with the schematic picture of the nucleation and growth
rates maxima located at different temperature regions, as presented in Fig. 2.
Similarly, we have performed compression/decompression rate-dependent studies
carried out for PC under isothermal conditions, T = 243 K. The results of these
experiments are collected in Fig. 7c, d, respectively. In analogy to cooling rate dependent scans, we have also observed a characteristic dispersion curve in the dielectric
response of the pressurized sample, indicating slowing down of the molecular movements and approaching the glassy state. For the two slowest compression rates (2.6
and 5.8 MPa/min) at least partial crystallization of PC was observed, as both curves
deviate from all the others. This was additionally certified by comparing the values
of the dielectric permittivity for the glassy and crystalline materials obtained on
increased pressure [61]. Upon decompression from the glassy state, just like when
heating the glassy sample, crystallization is far more difficult to avoid and this requires
higher scanning rates (see Fig. 7d). Nevertheless, the values of ε
do not drop down
completely to that characteristic for the crystalline material (~2.55), meaning that
there is still some substantial amount of the liquid volume fraction in the depressurized material. Using decompression rate of 154 MPa/min, we have not observed
any traces of the PC crystallization which indicates that at such high-compression
rates there is not enough time for the nuclei to grow. Therefore, we conclude that
by varying with compression/decompression rate it is possible to affect crystallization/vitrification tendency of the molecular liquids, same as when using the temperature. The results presented in Fig. 7d also reveal only one melting event (p m =
200 MPa at 243 K), which is again consistent with the heating rate dependent studies
carried out at ambient pressure. However, it has been also observed—especially for
ice phases—that different compression/decompression rates might induce elusive
polymorphic transformations [70, 71]. The increased tendency of the depressurized
PC to crystallization conform the schematic picture of the nucleation and growth
rates maxima located at slightly different pressure regions, as introduced in Fig. 6.