ACS PRF | ACS
All e-Annual Reports
42364-AC7
Glass Transition and Dynamics of Hydrogen-Bonded Liquids
The structure and dynamics of hydrogen bonded liquids, e.g. alcohols, water, etc., is addressed commonly by dielectric relaxation experiments. The strong dielectric signals of these materials have been assumed to identify the relaxation times associated with molecular motion, while the Kirkwood-Fröhlich correlation factors g have been used to derive structures resulting from hydrogen bonding. This project investigates the relation between structural and dielectric relaxation, casting serious doubts on the previous interpretation of the dielectric results. The focus is on glass-forming alcohols, where the features of interest are particularly obvious and where a single relaxation time or Debye peak is absent in any simple molecular liquid.
Past reporting period results:
For the large and narrow dielectric peak found in many hydrogen-bonding systems, the following features have been observed: This dielectric signal occurs in monohydroxy alcohols, not in diols and polyols, and it is also seen in several amides, in mixtures with other liquids only at very high alcohol content, and in mixtures where both pure systems show a Debye peak across a broader range. In binary systems, the logarithmic relaxation time of a Debye peak follows the ideal mixing law (unlike structural relaxation). The factor by which the Debye peak is slower than the true structural relaxation peak varies at Tg, but approaches 100 for all systems at higher temperatures. Additionally, the glass transition temperature Tg derived from the dielectric Debye signal is not consistent with the Tg derived from the heat capacity step seen in differential scanning calorimetry experiments. However, there is always another dielectric signature in terms of a much smaller dielectric feature that displays all typical properties of a structural relaxation peak.
Current reporting period results:
Regarding the comparison of dielectric and calorimetric signals, the decisive answer for identifying the time scale relations comes from a recent collaboration with Chr. Schick in Rostock, Germany, who performed isothermal scans of the frequency resolved dynamic heat capacity, Cp' and Cp'' versus frequency at different temperatures. These Cp'' data can be compared directly to the dielectric loss (eps''), which has been done for 2-ethyl-1-hexanol. The result of this study is very clear: the calorimetric time scale matches that of the smaller and faster dielectric peak, while there is no detectable calorimetric mode that corresponds to the Debye peak. This verifies our earlier conclusions that have been based on more indirect comparisons, but the relation between energy fluctuation and dielectric polarization fluctuations has not been this clearly and unambiguously.
Dynamic and thermodynamic behavior that does not involve the dielectric Debye polarization has also been studied in order to observe possible anomalies of alcohols other than the occurrence of the dielectric Debye process. In this respect, even monohydroxy alcohols display the typical behavior of other molecular liquids. We have tested a well established link between thermodynamic and dynamic fragility that reads m = 40 Cp(Tg)/Sm, where m is the dynamic fragility define as slope of dlog(tau)/d(Tg/T) evaluated at Tg, Cp(Tg) is the heat capacity step at the glass transition, and the Sm is the melting entropy. Alcohols and other hydrogen bonded systems are no exception in this strong quantitative correlation between dynamic and thermodynamic variables. A similar result is obtained by studying the relation between the boiling temperature Tb (a thermodynamic phase transition) and the glass transition temperature Tg (a purely kinetic phenomenon). For many liquids, Tb and Tg are positively correlated, as expected. However, within a series of isomers, Tb and Tg are negatively correlated, and this novel feature is observed equally for alkanes and alcohols. These studies emphasize that the prominent dielectric Debye peak of monohydroxy alcohols is the only unusual feature of these liquids, while true measures of structural relaxation such as calorimetry and mechanical relaxation do not reveal any anomalies.
In stead of setting up a 3-omega type calorimeter which would fail to detect the Debye process, we decided to apply a more promising technique which relies on supplying energy directly to the slow degrees of freedom. In the case of the monohydroxy alcohols, the Debye process would absorb energy from a high external electric field and the time scale of heat flow is assesses by dielectric experiments. These experiments are work-in-progress and at this point show the feasibility of the method.
A microscopic picture of the origin of the Debye peak has not been found yet, but the detailed observations compiled so far provide stringent tests of a future picture of the structural and dynamic properties of alcohols. The main challenge to be resolved is that the strong dipoles (1.68 D) of these molecules fail to generate an accordingly strong dielectric signal on the time scale of the calorimetric and mechanical modes, and that the prominent dielectric signal that could match the dipole strength is so much slower.
