Common to glass-forming monoalcohols is the extra Debye type relaxations in dielectric spectra, manifested by the slower dynamics than the structural (α-) relaxation, which associates with liquid viscosity and calorimetric glass transitions (Hassion and Cole, 1955; Dannhauser, 1968; Kudlik et al., 1997; Petong et al., 1999; Johari et al., 2001; Wang and Richert, 2004; Fragiadakis et al., 2010; Pawlus et al., 2010; Gainaru et al., 2011; Power et al., 2011). A number of unique dynamic characters have been discovered for the Debye relaxation (Hansen et al., 1997; Wang and Richert, 2005a; Wang et al., 2005; Huth et al., 2007; Power et al., 2007; Jakobsen et al., 2008; Gainaru and Böhmer, 2010). The understanding of the Debye relaxation has been developed in the latest decades, and the chain-like associations imposed by intermolecular hydrogen-bond (H-bond) are considered to be responsible for the Debye relaxation, although the special relaxation might not be the direct signature of the overall reorientation of the associations as a whole (Fragiadakis et al., 2010; Gainaru et al., 2011; Singh and Richert, 2012). However, a complete understanding of the Debye relaxation is still unclear and, in particular, the formation of the associations (or clusters) corresponding to the Debye relaxation in H-bond liquids and the relation between the Debye and α-relaxations have not yet been solved.

The studies of the dielectric dynamics in supercooled monoalcohols with small molecular size are expected to extend the structural knowledge (clusters) associated with the Debye relaxation. The exploration of the possibility that similar chain-like associations are present in methanol would consequently be of particular interest because of the smallest size among monoalcohols. Yet, the rapid crystallization in supercooled methanol near the glass transition region impedes the accurate determination of the dynamics and the glass transition temperature, T_g (Souda, 2004). Recent studies of the dielectric relaxation in supercooled ethanol found that it is not straightforward to confirm the Debye relaxation as the main relaxation becomes broader than pure Debye dynamics (Brand et al., 2002), especially when the plastic crystal is involved in ethanol.

In order to learn the dynamics of methanol, a common practice is to study the dynamics of binary mixtures of methanol (Sun et al., 2011). In the past few decades, aqueous solutions of methanol have been explored extensively; however, much less experimental studies are available for the dynamics in glass-forming mixtures composed of methanol and well-defined Debye liquids (Denney and Cole, 1955). Our recent studies of the mixtures composed of two monoalcohols suggest that the main relaxation in the dielectric spectra of the mixtures is still Debye typed, while the Debye relaxation disappears in monoalcohols diluted by non-Debye liquids at a specific mole fraction (Wang and Richert, 2005b; Wang et al., 2005). In this paper, we examine the dielectric behaviors in the glass-forming solutions of methanol mixed with 2-ethyl-1-hexanol, a typical Debye liquid featured by the separated Debye and α-relaxations and high glass-forming ability (Murthy, 1996; Wang and Richert, 2004; Wang et al., 2005; Jakobsen et al., 2008; Fragiadakis et al., 2010; Bauer et al., 2013). A broad composition range is explored up to 80 mol% methanol, and the main dielectric relaxation with the slowest dynamics in the mixtures is focused. The dynamic features of the Debye relaxation in supercooled liquids are identified.

Broadband dielectric relaxation measurements of the mixtures of methanol (Sigma Aldrich, 99%) and 2-ethyl-1-hexanol (2E1H, Alfa Aesar, 99%) are made in a Novocontrol broadband dielectric spectrometer (Concept 80) equipped with a liquid nitrogen cooling system (Chen et al., 2009). The dielectric relaxation is isothermally measured at each temperature by holding liquids between two separated brass electrodes by Teflon strips of thickness 25 μm. The scanning frequency range spans from 0.01 Hz to 10 MHz. The temperature is controlled by a Novocontrol Quatro controller with temperature accuracy within 0.1 K. The dielectric data are analyzed to obtain the dynamic parameters in terms of Havriliak–Negami (HN) expressions (Havriliak and Negami, 1967), (1)ε*(ω)=ε∞+∑iΔεi(1+(iωτi)αi)γi+σdciε0ω, where ε_∞ is the high-frequency dielectric constant, Δε_i dielectric strength, τ_i dielectric relaxation time, and α_i and γ_i profile shape factors of the relaxation dispersion for the ith dynamics. With α = 1, the HN equation becomes the Cole–Davidson (CD) equation, while γ = 1 leads to the Cole–Cole (CC) equation. σ_dc is the dc conductivity.

The calorimetric glass transition of the mixtures is recorded in a Perkin–Elmer (PE) Diamond differential scanning calorimetry (DSC) at fixed cooling and heating rates of 20 K/min. The instrument is calibrated by indium and cyclohexane for the high and low temperature, following a previous procedure described elsewhere (Wang et al., 2002). The glass transition temperatures, T_g–cal, are determined by the onset temperatures of the heat capacity jumps from glasses to supercooled liquids.

The dielectric loss spectra measured for the methanol – 2E1H mixtures are shown in Figure 1. In Figures 1A,B, two relaxations I and II are easily seen in the isothermal spectra of the mixtures with 10 and 30 mol% methanol (i.e., 10M–90H and 30M–70H), in analogy to the dielectric behaviors of pure 2E1H (Wang and Richert, 2004). The slow relaxation is consequently expected to be Debye typed. Fits to the loss peaks of the 10M–90H mixtures show a CD type profile for the Debye relaxation with γ_I = 1, while the α-relaxation shows typical HN profile, giving typical values of α_II~ 0.75 and γ_II ~ 0.55. As the concentration of methanol increases, the CC type profiles for the slow relaxation are favored based on the comparison of the fitting quality. In Figure 1B, a fitting with α_I = 0.9 and γ_I = 1 for the 163 K curve of the 30M–70H mixture well explains the main relaxation, while the shape parameters of the α-relaxation are comparable with those in the 10M–90H mixture. The dielectric spectra obtained in the mixtures with 60 mol% methanol are presented in Figure 1C. It is difficult to distinguish two dynamics, and apparently one HN equation seems to basically explain the loss spectra. Figure 1D shows the dielectric spectra of the mixtures with 80 mol% methanol, and two relaxations are resolved and the CC equation is also proven to be suitable for the slow relaxation.

Figure 2 presents a close inspection of the relaxation dispersion profiles of the mixtures with the loss spectra normalized by the loss maxima positioned at 1–10 Hz. It is seen that as the fraction of methanol increases, the slope of the low-frequency flank of the slow relaxation reduces continuously from unity. Further analyses of the relaxation dispersion parameters (see Figure 4) suggest that the continuous broadening of the relaxation I cannot be explained by the interference of the relaxation II, although the dynamic separation is reduced between the two relaxations.

The relaxation strength and relaxation dispersion of the dynamics I and II are presented in Figures 3 and 4. It is seen that in the range of the methanol mole fraction below 0.7, the total relaxation strength slightly increases, while for the higher fraction, the large enhancement in the relaxation strength is observed, in particular when the dielectric data of pure methanol (Denney and Cole, 1955) is added. The similar case is also observed for the relaxation strength of the relaxation I. The numerical comparison finds that the relaxation I basically explains the total relaxation strength for all the mixtures, suggesting that the relaxation II is relatively weak. A close inspection of the relaxation II also finds a continuous increase of the relaxation strength with increased concentration of methanol. Figure 4 shows the relaxation dispersion parameters of the two relaxations, and the CC type character is revealed for the relaxation I of the mixtures as the methanol mole fraction exceeds 0.3. α_I is observed to continuously reduce, indicative of increased broadening of the relaxation I. In contrast, the relaxation dispersion parameters, α_II and γ_II, of the relaxation II basically keep constant, consistent with the temperature dependence of the α-relaxation dispersion in generic liquids, where the relaxation width either keeps constant or decreases in the supercooled regions as temperature increases (Dixon et al., 1990; Olsen et al., 2001; Wang and Richert, 2007).

Figure 5 presents the reciprocal temperature dependence of the relaxation times for the relaxations I and II of the mixtures. The Vogel–Fulcher–Tammann (VFT) equation, log₁₀τ = A + B/(T – T₀), where τ is the relaxation time and A, B, and T₀ are constants, is used to analyze the data. The kinetic glass transition temperature, T_g–kin is usually defined by a temperature where the relaxation time approaches 100 s (Angell et al., 2000). And thus, two T_g–kin values can be obtained for the methanol-deficit mixtures, denoted by TτD=100 s for the relaxation I and Tτα=100 s for the relaxation II. T_g can also be determined from calorimetric measurements, and Figure 6 presents the calorimetric results of the mixtures with T_g–cal defined by the onset transition temperature in the heatflow curves from glassy to supercooled liquid states. Only one glass transition is clearly identified, and this suggests that no phase separation occurs in the mixtures.

Two relaxations are observed in the dielectric spectra of the methanol–2E1H mixtures at the methanol-rich and 2E1H-rich compositions. For the 2E1H-rich compositions, it can be easily concluded that the relaxation I is of the same nature as the Debye relaxation detected in pure 2E1H, since the Debye relaxation has been proven not to disappear in the diluted monoalcohols until a certain concentration (Wang et al., 2005; El Goresy and Böhmer, 2008). In contrast, the identification of the slow relaxation (I) in the methanol-rich compositions might not be straightforward. Fortunately, the unique dynamic features in supercooled states found in earlier publications can be used to serve as the basic criteria to distinguish the Debye relaxation from α-relaxation in pure liquids or liquid mixtures (Wang and Richert, 2005a). Naturally, two prominent dynamic features of the Debye relaxation will be examined for the methanol-rich mixtures, one being the difference between the kinetic T_g in terms of the Debye relaxation time and the calorimetrically measured T_g (Wang and Richert, 2005a; Wang et al., 2008) and the other the ideal mixing law of the Debye relaxation time in the binary mixtures of two primary monoalcohols (Wang and Richert, 2005b).

The dynamic analyses together with the calculation of the Kirkwood factors of the methanol–2E1H mixtures suggest that the Debye relaxation is present in the dielectric spectra of pure methanol. The conclusion is also supported by a recent study of the comparison of the dielectric and low-frequency Raman spectra of methanol, finding that the slow dielectric dynamics, which is the Debye relaxation, does not correspond to the Raman spectrum, and the faster processes reflect common microscopic dynamics (Fukasawa et al., 2005). The validity of the ideal mixing law for the Debye relaxation time of the mixtures advances the construction of the temperature dependence of the Debye relaxation time of the pure methanol, which will facilitate the evaluation of the dynamics in deeply supercooled methanol. In addition, broadened Debye relaxation is expected in the dielectric spectra of the methanol, and this is quite different from earlier observation of the Debye liquids.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.