Kilns consume about half of the energy necessary to operate lithium refineries and their decarbonisation requires accurate modelling of the calcination of spodumene concentrates fed to the process. This contribution applies the isoconversional methodology to investigate the kinetic parameters of the transition of spodumene to its high-temperature polymorph of spodumene, using the heat flux measurements from the differential scanning calorimetry (DSC). The normalised energy demand (), presented as a function of temperature, characterises these measurements. The activation energy and the product of the reaction model and the frequency factor , (), depend on . As the process involves multi-step reactions, we deploy the Friedman differential method and the accurate flexible integral method of Vyazovkin to obtain the kinetic parameters. We also modify the method of Ortega to acquire additional estimates of and () and apply the rigid integral method of Starink for comparison. The Friedman, Vyazovkin and modified methods deliver the same estimates of the kinetic parameters within their error bands. The Starink method works surprisingly well for predicting the conversion time despite the inaccuracies in the derived values of and (). This comes to pass because of the compensation effect between these parameters. The activation energy declines rapidly from around 1000 kJ mol at the commencement of the heat treatment to 668 kJ mol at = 0.22, then, decreases gently to 577 kJ mol at = 0.98, during successive recrystallisation events. Average uncertainties in these results amount to 13 kJ mol. The frequency factors fall between 58.5 (±1.0) min and 51.0 (±3.2) min, as computed at = 0.23 and 0.98, respectively. The so-called false compensation analysis reveals that the first-order reaction model (in ) governs the energy demand for calcination for ≥ 0.23, but, initially, the transformation proceeds through the dissociation-diffusion regime that is not part of the established reaction models. This regime must not be ignored in modelling the calcination of spodumene, as it consumes around 20 % of energy required for the transformation reactions. The results reveal significant differences in the predictions of the treatment time, by more than two orders of magnitude, from the existing kinetic models, and explain the differences by the experimental conditions to collect the data for the models. The dissociation of Si-O bonds and diffusion of Si ions out of their tetrahedral cages govern the onset of the thermal treatment of -spodumene and account for the elevated values of the activation energy in the dissociation-diffusion regime. The two recrystallisation events are limited by the multicomponent diffusion, especially Si, in partly crystallised structure. The recrystallisation of - to -spodumene defines the required retention time of concentrate particles in the kiln, to maximise the effectiveness of the subsequent recovery of lithium from the treated material. Fitting and ln() to polynomials allows a convenient integration of the model equation for making predictions under any heating program. The model forecasts well the transformation of particles of spodumene characterised by = 315 µm, studied in additional experiments, using the X-ray powder diffraction to quantify the conversion of -spodumene. The predictions from the isoconversional model also concur well with other conversion measurements available in the literature, within the expected variability of different spodumene concentrates.

中文翻译：

---

锂辉石煅烧动力学（α-LiAlSi2O6）


窑炉消耗了锂精炼厂运营所需能源的大约一半，其脱碳需要对输送到该工艺的锂辉石精矿的煅烧进行精确建模。该贡献应用等转化方法，利用差示扫描量热法 (DSC) 的热通量测量结果，研究锂辉石转变为锂辉石高温多晶型物的动力学参数。标准化能量需求 () 作为温度的函数表示，表征了这些测量结果。反应模型的活化能以及乘积和频率因子 () 取决于 。由于该过程涉及多步反应，我们采用Friedman微分法和Vyazovkin精确柔性积分法来获得动力学参数。我们还修改了 Ortega 的方法以获得 和 () 的额外估计，并应用 Starink 的刚性积分方法进行比较。 Friedman、Vyazovkin 和修改后的方法在其误差带内提供了相同的动力学参数估计值。尽管 和 () 的派生值不准确，但 Starink 方法在预测转换时间方面效果出人意料地好。这是因为这些参数之间的补偿效应。在连续的重结晶过程中，活化能从热处理开始时的约 1000 kJ mol 快速下降至 = 0.22 时的 668 kJ mol，然后缓慢下降至 = 0.98 时的 577 kJ mol。这些结果的平均不确定性达到 13 kJ mol。频率因子介于 58.5 (±1.0) 分钟和 51.0 (±3.2) 分钟之间，计算结果分别为 = 0.23 和 0.98。 所谓的错误补偿分析表明，一级反应模型 (in ) 控制着 ≥ 0.23 的煅烧能量需求，但最初，转化是通过解离-扩散机制进行的，该机制不属于已建立的反应模型的一部分。在对锂辉石的煅烧进行建模时，一定不能忽视这种情况，因为它消耗了转化反应所需能量的 20% 左右。结果表明，处理时间的预测与现有动力学模型存在两个以上数量级的显着差异，并通过收集模型数据的实验条件解释了差异。 Si-O 键的解离和 Si 离子从四面体笼中的扩散控制了锂辉石热处理的开始，并解释了解离-扩散区域中活化能的升高。这两个再结晶事件受到多组分扩散的限制，尤其是部分结晶结构中的 Si。锂辉石的重结晶决定了精矿颗粒在窑中所需的保留时间，以最大限度地提高随后从处理过的材料中回收锂的效率。将 ln() 拟合到多项式可以方便地积分模型方程，以便在任何加热程序下进行预测。该模型很好地预测了以 = 315 µm 为特征的锂辉石颗粒的转化，在其他实验中进行了研究，使用 X 射线粉末衍射来量化 β 锂辉石的转化。 在不同锂辉石精矿的预期变化范围内，等转化模型的预测也与文献中可用的其他转化测量结果非常一致。