Abstract Heating inorganic glasses above their respective glass transition temperature leads to surface crystallization in many cases. Analyzing the immediate surface of the resulting glass-ceramics using electron backscatter diffraction (EBSD) has shown that the topmost layer of crystals already shows orientation preferences, aka textures. As the information depth of EBSD is less than 100 nm, these textures can only result from oriented nucleation. This is noteworthy, considering that random nucleation is assumed in the classic nucleation theory for glasses. Since the effect was first described in 2010, oriented nucleation has been proven for 13 crystal phases including cases from all seven crystal symmetries ranging from cubic to triclinic. Of the 13 confirmed cases, six phases show a single texture while seven show up to three coexisting textures at the immediate surface. The texture intensities range from weak to extreme. It has been shown that modifying the crystallization temperature can affect oriented nucleation as well as modifying the chemical composition of the glass by e.g. systematically increasing the amount of the network forming component. Of the described textures, at least 20 indicate a low indexed lattice plane being aligned parallel to the sample surface while eight are more complex. In all cases, the orientation alignment is not strict: while most show tolerances of less than ±15°, some cases also feature a preference of much wider orientation domains. A fundamental explanation for the occurrence of oriented nucleation has not been proposed so far. While varying surface energies for different lattice planes of a phase could explain the effect, surface energies are usually unknown and hard to measure. Hence they cannot be used to predict the orientation preference at the surface at the current time. Considering the crystal structure of the phases, however, is an alternative and plausible approach. A detailed analysis of the cases analyzed so far shows that the smallest low indexed lattice plane of the crystal unit cell is often aligned parallel to the surface. This approach, however, can only explain the orientation preference, if an a-, b- or c-axis is perpendicular to the surface because higher indexed lattice planes usually have larger areas. Considering the smallest area per atom of the least mobile species being oriented parallel to the surface leads to an even better correlation between the approach and the experimental observations. This might be explained as follows: During nucleation, in a first step, an arrangement of atoms is formed at the immediate surface of the glass, which is similar to a certain plane of the unit cell of the formed crystal. For this purpose, the diffusion of atoms is necessary. Hence it seems plausible that first a layer rich in less mobile atoms should be formed and subsequently the more mobile atoms are added. As the least mobile atoms in the glass structure are the network formers, the first step of nucleation should then be that the lattice plane with the highest number of network formers per area is formed at the immediate glass surface. This lattice plane should initially remain parallel to the surface as further atoms are added. The orientation of the crystals below the immediate surface may change during crystal growth into the bulk so that the fastest growing crystallographic direction becomes oriented perpendicular to the surface. Here we review the work concerning oriented nucleation in glasses and present two models addressing the possible cause of the effect. Both a critically discussed and compared to the experimentally observed cases of oriented nucleation.

中文翻译：

---

无机玻璃中的定向表面成核 - 综述

摘要 在许多情况下，将无机玻璃加热到其各自的玻璃化转变温度以上会导致表面结晶。使用电子背散射衍射 (EBSD) 分析所得玻璃陶瓷的直接表面表明，晶体的最顶层已经显示出取向偏好，也就是纹理。由于 EBSD 的信息深度小于 100 nm，这些纹理只能由定向成核产生。这是值得注意的，因为在经典的玻璃成核理论中假设了随机成核。自 2010 年首次描述该效应以来，定向成核已被证明适用于 13 个晶相，包括从立方到三斜晶的所有七种晶体对称性的情况。在13例确诊病例中，六个阶段显示单个纹理，而七个阶段在直接表面显示多达三个共存的纹理。纹理强度范围从弱到极端。已经表明，改变结晶温度可以通过例如系统地增加网络形成组分的量来影响定向成核以及改变玻璃的化学组成。在所描述的纹理中，至少有 20 个表示低索引晶格平面与样品表面平行排列，而 8 个更复杂。在所有情况下，方向对齐并不严格：虽然大多数情况下的公差小于 ±15°，但某些情况下还倾向于更宽的方向域。迄今为止，尚未提出对定向成核发生的基本解释。虽然相的不同晶面的不同表面能可以解释这种效应，但表面能通常是未知的并且难以测量。因此，它们不能用于预测当前时间表面的方向偏好。然而，考虑相的晶体结构是另一种可行的方法。对迄今为止所分析案例的详细分析表明，晶体晶胞的最小低折射率晶格平面通常与表面平行排列。然而，这种方法只能解释取向偏好，如果 a 轴、b 轴或 c 轴垂直于表面，因为更高索引的晶格平面通常具有更大的面积。考虑到与表面平行取向的最低移动物种的每个原子的最小面积导致该方法与实验观察之间的更好的相关性。这可以解释如下：在成核过程中，第一步，在玻璃的直接表面形成原子排列，它类似于形成的晶体的晶胞的某个平面。为此，原子的扩散是必要的。因此，首先应该形成富含较少流动原子的层，然后添加较多流动原子的层似乎是合理的。由于玻璃结构中流动性最小的原子是网络形成体，因此成核的第一步应该是在紧邻玻璃表面形成每面积具有最高数量网络形成体的晶格平面。随着更多原子的加入，该晶格平面最初应保持平行于表面。在晶体生长到块体中期间，直接表面下方的晶体的取向可能会发生变化，从而使生长最快的晶体学方向垂直于表面取向。在这里，我们回顾了有关玻璃中定向成核的工作，并提出了两种模型来解决该效应的可能原因。对定向成核的实验观察情况进行了批判性讨论和比较。在这里，我们回顾了有关玻璃中定向成核的工作，并提出了两种模型来解决该效应的可能原因。对定向成核的实验观察情况进行了批判性讨论和比较。在这里，我们回顾了有关玻璃中定向成核的工作，并提出了两种模型来解决该效应的可能原因。对定向成核的实验观察情况进行了批判性讨论和比较。