Merging of α and slow β relaxation in supercooled liquids

Abstract

Dielectric relaxation spectroscopy (1 Hz -20 GHz) has been performed on supercooled glassformers from the temperature of glass transition (Tg) up to that of melting. Precise measurements particularly in the frequencies of MHz-order have revealed that the temperature dependences of secondary β-relaxation times deviate from the Arrhenius relation in well above Tg. Consequently, our results indicate that the β-process merges into the primary α-mode around the melting temperature, and not at the dynamical transition point T ≈ 1.2Tg.PACS numbers: 64.70. Pf, 77.22.Gm, In recent years, much of the focus on glassy dynamics has been shifting to a considerably higher temperature than T g , the glass-transition one [1,2,3,4]. The topical temperature is located around T D ≡ 1.2T g , where the dynamics of supercooled liquids has been found to change fairly. So far, there have been observed the following phenomena:(i) Rössler scaling reveals that, when cooled, the Debye-Stokes-Einstein relation becomes invalid around T D [5,6]. This indicates a change of diffusion mechanism there.(ii) Stickel analysis [7] clarifies that temperature dependence of viscosity changes around T D . Therefore, in order to fit the primary α-relaxation time τ α using the Vogel-Fulcher-Tamman (VFT) relation,the coefficients (τ 0 , C, T 0 ) have to vary at T B ≈ T D . This suggests that the mechanism of slow structural-relaxation makes some alternation there.(iii) Johari-Goldstein type β-process (secondary process in the context of dielectric relaxation) [8,9,10,11,12,13] merges into the α-relaxation around T D , extrapolating the Arrhenius-type temperature dependence below T g of the β-relaxation times [6,14,15,16,17,18,19].Theoretically, the characteristic temperature T D is thought to be comparable to T C where the idealized Mode Coupling Theory (MCT) predicts a dynamical phase transition [20]. Indeed, the above first two phenomena (i, ii) can be regarded as indicators of the dynamical transition at T C . However, the MCT is irrelevant to the third one (iii), the bifurcation of α, β-modes; even the existence of the Johari-Goldstein type β-process cannot be derived.Furthermore, from experimental aspects, while the dynamical transition phenomena (i, ii) have been confirmed from either Rössler or Stickel plot definitely, the bifurcation of (iii) is inferred from the extrapolation. Actually, however, it remains an open problem as to whether the Arrhenius behavior of the β-process persists in higher temperatures near T D : the T D -decoupling of α, β-relaxations is not conclusive.This letter thus aims to investigate the secondary β-mode in high temperatures well above T g , by carrying out precisely the broad-band measurements of dielectric relaxation. Our main result is the following: as will be seen in Figures 2 and 5, the β-relaxation times deviate from the Arrhenius relation, indicating that the β-relaxation merges into the α-mode not at T D but around the Arrhenius-VFT crossover temperature T A (generally close to the melting one) where the temperatu...

Orientation polarization from faster motions in the ultraviscous and glassy diethyl phthalate and its entropy

“…Therefore, we can reliably deduce only that the activation energy of the ␤ relaxation is higher for the ultraviscous liquid than for the glassy state, but not exactly how much higher. Fujima et al 34 have observed a similar feature in the plot for D-sorbitol. They have shown it as a plot of the symmetric broadening parameter for the ␤ relaxation against 1 / T in Although a detailed resolution of the spectra for D-sorbitol has supported their findings, 34 we investigate if the features observed here could be seen as an artifact of local temperature change in the sample as a result of structural relaxation during the course of measurements.…”

“…Fujima et al 34 have observed a similar feature in the plot for D-sorbitol. They have shown it as a plot of the symmetric broadening parameter for the ␤ relaxation against 1 / T in Although a detailed resolution of the spectra for D-sorbitol has supported their findings, 34 we investigate if the features observed here could be seen as an artifact of local temperature change in the sample as a result of structural relaxation during the course of measurements. A possible local increase in the temperature as a result of structural relaxation at T Ͻ T g , would raise f m and would therefore increase the slope of the log͑f m,␤ ͒ against 1 / T plot at T near T g , but here a decrease is observed in Fig.…”

“…4 increases above T g . Therefore, we conclude that the features observed in the log͑f m,␤ ͒ against 1 / T plot and of the ␣ ␤ against T plots for diethyl phthalate and other liquids 34 are either a consequence of the multiparameter fits in which the Cole-Cole distribution of times has been used for the ␤ relaxation or else they indicate a structural change in this region. It is also conceivable that these features are an indication of the effect of structural unfreezing on heating at T near T g .…”

The new insight into dynamic crossover in glass forming liquids from the apparent enthalpy analysis

“…In subsequent years, it became a key tool for estimating the so-called dynamic crossover temperature T B between two dynamic domains. 7,[18][19][20][21][22][23][24] It was shown that at least two VFT equations are desired for describing pre-vitrificational slowing down in a broader range of temperatures. The temperature T B has been also recognized as a milestone in the way on approaching T g since a lot of exceptional phenomena disclose there, for instance: [18][19][20][21][22][23][24][25][26][27][28][29][30][31] (1) the loss of ergodicity as predicted by mode-coupling theory (MCT), (2) increasing broadening of the structural relaxation time distribution, (3) a marked change in temperature dependence of the nanopore unoccupied volume radius, (4) splitting of the high temperature relaxation into the primary and secondary relaxation times, (5) orientational-translational decoupling, (6) for a broad set of supercooled systems τ (T B ) = 10 −7±1 s or η(T B ) ≈ 10 3 P-i.e., it is near-universal, (7) τ (T B , P B ), η(T B , P B ) = const for a given glass former, (8) it is believed that the dynamical crossover is closely related to the onset of caging and appearing of dynamical heterogeneities, (9) the coefficient D T in the VFT equation almost always increases on crossing to the dynamical domain in the immediate vicinity of T g .…”

“…22 These facts are worth stressing because the dynamic crossover is sometimes linked to the "FS" crossover. 32 However, there are fundamental weak points in the discussion employing the "Stickel analysis" carried out so far, namely: (1) an impressive number of papers discussing the dynamical crossover explores only the "formal" plot ϕ T vs. 1/T, or its pressure counterpart ϕ P = (dlog 10 τ /dP) −1/2 vs. P, without an attempt of finding a possible physical meaning hidden behind [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33] (2) the general validity of the VFT equation, underlying ϕ T (T) "Stickel" function, has been strongly questioned in recent years.…”