The spherulitic crystal growth velocity in selenium supercooled liquid has been measured by infrared microscopy in isothermal conditions for rapidly heated samples of a-Se from temperatures well below T g . These data are compared and analyzed along with previously published data obtained on samples quenched from a temperature well above the melting point. The spherulites grew linearly over a course of time that corresponds to crystal growth controlled by crystal−liquid interface kinetics. The crystal growth velocity data obtained for these two different thermal histories can be described by the normal growth model for moderate supercoolings (ΔT <60 K). The screw dislocation growth provides a better description for larger supercoolings (ΔT > 90 K) that is also consistent with morphological observations. However, the prediction based on this model still significantly deviates for intermediate supercoolings. It is shown that all experimental data in the whole temperature range (T g < T < T m ) can be described by a combined approach including both these models, taking into account actual viscosity scaling of crystal growth η −1 . The kinetic information captured in the DSC curve is analyzed, providing evidence that the non-isothermal crystallization process can be described by the Johnson-Mehl-Avrami model. Two distinct regions characterized by different values of apparent activation energy were found. The transition between these regions coincides with morphological changes of spherulitic crystals and other properties. These regions are characterized by distinct apparent activation energies whose values are consistent with those obtained from microscopic measurement of crystal growth velocity.

The isothermal crystal growth in Se 100−x Te x bulk glasses (x = 10, 20, and 30) was studied directly using infrared microscopy. The crystals grew spherulitically and linearly in the course of time, which is typical for crystal growth controlled by liquid-crystal interface kinetics. An operative growth model was found using a combination of growth and viscosity data, and using two different approaches for calculations of the Gibbs free energy change between the undercooled melt and crystalline phase. The study shows that the exact knowledge of the Gibbs free energy change calculated from both, heat capacities, and the simple approximation proposed by Turnbull, can provide comparable results regarding determination of an operative crystal growth model. A detailed discussion about the relationship between the kinetic coefficient of crystal growth rate and viscosity (u kin ∝ η −ξ ) is presented. Moreover, the activation energies of crystal growth were found to be higher than the activation energies of the overall crystallization process obtained by differential scanning calorimetry. The relation between these two quantities is considered under the experimental conditions. ■ INTRODUCTIONChalcogenide glasses and thin films are very interesting materials that exhibit unique physical properties. Because of their diverse active properties, chalcogenide materials can be used in various optical and optoelectronical devices or in various electronic thresholds and switches. 1−3 Interesting applications of chalcogenide materials also are found in modern high-tech memory devices. 4−6 Thermal stability and crystallization play a key role in processing and usage of the materials. In particular, the crystallization process needs to be considered from two points of viewin order to obtain an ideal glass the crystallization has to be prevented, and, on the other hand, the controlled amorphous-to-crystalline transformation is a fundamental process of considered technology (modern phase change materials, PCM). Thus, knowledge and understanding of crystal growth kinetics and nucleation in such materials are essential for their future applications. The crystallization studies 7−11 focus on evaluation of mechanism of crystal growth and nucleation, and on prediction of crystallization behavior in a wide temperature range. Such a description can be useful for tailoring and optimization of new high-tech materials as they provide a possibility to predict the crystallization behavior in similar materials by revealing the basic mechanisms and properties of the material.Selenium-based materials have been commercially applied in photoreceptors, photovoltaic materials, etc. 12 The properties of pure selenium can be significantly improved by alloying with other elements such as tellurium, germanium, antimony, arsenic, etc. Se−Te glasses and thin films are very attractive materials that exhibit an intermediate behavior between pure selenium and tellurium. The Se−Te mixtures exhibit numerous advantages in comparison with pure Se, for example, great...

Crystal growth in the surface of selenium bulk samples and in selenium thin films of different thicknesses has been studied under isothermal conditions using different microscopy techniques (optical, infrared, and scanning electron microscopy). The structure of the formed crystals is described with respect to previous publications focused on crystal growth in selenium thin films and bulk samples. Crystal growth rates were obtained from the linear dependence of crystal sizes on annealing time. Such behavior assumes that crystal growth is driven by liquid–crystal interface kinetics. The crystal growth rates found in the surface of bulk samples are comparable with those found in thin films and a few orders of magnitude higher than previously published growth rates of volume crystals formed in selenium undercooled melts. The crystal growth rates were scaled with the viscosities to analyze the Stokes–Einstein relation. A relatively high decoupling between the crystal growth rate and viscosity occurs in the studied samples of amorphous selenium. Therefore, the standard screw dislocation growth model is corrected for the decoupling, which provides a satisfactory description of the crystal growth rate over a wide temperature range.

Crystal growth, viscosity, and melting were studied in Ge2Sb2Se5 bulk samples. The crystals formed a compact layer on the surface of the sample and then continued to grow from the surface to the central part of the sample. The formed crystalline layer grew linearly with time, which suggests that the crystal growth is controlled by liquid-crystal interface kinetics. Combining the growth data with the measured viscosities and melting data, crystal growth could be described on the basis of standard crystal growth models. The screw dislocation growth model seems to be operative in describing the temperature dependence of the crystal growth rate in the studied material in a wide temperature range. A detailed discussion on the relation between the kinetic coefficient of crystal growth and viscosity (ukin ∝ η(-ξ)) is presented. The activation energy of crystal growth was found to be higher than the activation energy of crystallization obtained from differential scanning calorimetry, which covers the whole nucleation-growth process. This difference is considered and explained under the experimental conditions.

No abstract

No abstract

Crystal growth rates in GeSbSe bulk glass and thin film were measured using optical and scanning electron microscopy under isothermal conditions. The studied temperature region was 255-346 °C and 254-286 °C for bulk glass and thin film, respectively. The compact crystalline layer growing from the surface into the amorphous core was formed in bulk glasses and no bulk crystallization was observed. In the case of thin films, needle-shape crystals were formed. The crystalline layer and needle-shape crystals grew linearly with time that corresponds to a crystal growth controlled by the crystal-liquid interface kinetics. In the narrow temperature range, crystal growth rates exhibit simple exponential behavior, so the activation energies of crystal growth for the studied temperature regions were estimated (E = 294 ± 6 kJ/mol for bulk glass and E = 224 ± 12 kJ/mol for thin film). Viscosity of GeSbSe material was measured in the region of the undercooled melt and glass. The extrapolation of viscosity data into the immeasurable, but important, temperature range is discussed. The experimental growth data were combined with melting and viscosity data and the appropriate growth models were proposed to describe crystal growth in a wide temperature region. The standard crystal growth models are based on a simple proportionality of the crystal growth rate to the viscosity (u ∝ η). This simple proportionality holds for the bulk material. Nevertheless, in the thin films the decoupling of the crystal growth rate from the inverse viscosity occurs, and the standard kinetic growth models need to be corrected. Such corrections provide better description of experimental data and more realistic value of the parameter describing the mean interatomic distance in the crystal-liquid interface layer, where the crystal growth takes place.

Non-isothermal measurements of dc conductivity were used to study crystallization in bulk glasses from the TexSe100−x system (x = 10, 20, and 30). In combination with infrared microscopy, it provides a very good qualitative picture of the complex crystallization processes, as the formation of conductive paths proceeds more or less separately for each involved process. In addition, due to the pronounced signal corresponding to the surface crystallization, the dc conductivity measurements seem to bear great potential for crystallization kinetic studies. Based on identified characteristic states corresponding to the particular crystallization mechanisms, the activation energies were calculated for surface and bulk processes and compared to the differential scanning calorimetry (DSC) results reported earlier. Furthermore, the conversion rate α was calculated from the conductivity data by using the Odolevsky equation, the obtained dependence is in a good agreement with the results from residual enthalpies measurements performed by DSC. In a supplemental study a thorough kinetic analysis was applied to the DSC measurements of residual enthalpies. The observed decrease of the Johnson-Mehl-Avrami kinetic exponent m was explained by decrease in dimensionality of the bulk crystallization process.