In situ particle size estimation for crystallization processes by multivariate image analysis

Crystal size estimation from in situ images has received attention recently as a means to estimate product properties in real-time. In this paper, an automated image analysis strategy that combines classical image analysis techniques with multivariate statistics has been developed for online analysis of in situ images from crystallization process. The strategy introduces a novel image segmentation step based on information extracted from multivariate statistical models. Experimental results for batch cooling crystallization of monosodium glutamate show that the strategy effectively extracts crystal size and shape information from in situ images. The robustness and efficiency of the method has been established by comparing its performance with those obtained by manual analysis of the images. The method yields reasonably good estimates of particle length and is also fast enough to provide online measurements for the purpose of online optimization and control of a typical crystallization process.     

Crystal shape monitoring and supersaturation measurement in cooling crystallization with quartz crystal oscillator

A crystallization monitoring system using a quartz crystal oscillator was utilized to predict different shapes of crystal formation by measuring crystal growth rate and to measure supersaturation. Applying different rates of cooling, crystal formation of different shapes was induced, and the frequency variation of the oscillator and the crystal shape observed with an SEM were compared to determine how the frequency variation can be interpreted for the prediction of produced crystal shape. The experimental results obtained from the crystallization of potassium nitrate and cupric sulfate solutions showed that the proposed frequency measurement technique could be applied in the prediction of crystal shape of cooling crystallization processes. In addition, supersaturation was determined from the measurements of solution and coolant temperatures.     

膜辅助添加晶种的过硫酸铵冷却结晶在线监测与过程调控

冷却结晶是经典的溶液结晶过程,常用于分离溶解度随温度变化较大的物质,制备高品质晶体产品。直接进行降温会导致成核速率不可控,得到的晶体产品质量差。在工业中通常选择在溶液结晶介稳区内投放适量晶种来诱导成核,但晶种制备过程复杂,而且成功的添加晶种过程取决于晶种的粒度分布、数量、投放时机和操作人员的经验等因素,降低了产品质量的批次重复性。本文利用聚四氟乙烯（PTFE）中空纤维膜组件为结晶溶液和冷却液提供换热界面,结晶溶液温度降低,在膜界面处形成较均匀的过冷度梯度,进而在低过饱和度下发生异相成核,实现膜辅助添加晶种的过硫酸铵冷却结晶过程调控。膜组件中产生的晶种进入结晶釜中继续生长,将成核和生长过程进行解耦。在线结晶检测系统捕捉到的照片证实了通过控制膜组件使用温度和时长两个操作参数便可得到具有较好的形貌、较窄的粒度分布的晶种。相比直接冷却结晶,在相近的降温速率下,膜辅助添加晶种过程制备的晶体产品具有更大的平均粒径,且粒度分布更集中,表面更加光滑。因此,膜辅助冷却结晶呈现了良好的成核控制能力,有望实现晶种自动制备和添加功能,为高附加值晶体产品的冷却结晶过程开发提供了新方向。     

无水醋酸钠结晶过程中析晶温度和颗粒粒径在线测量

结晶过程是化工单元操作中一种高效节能的固液分离与提纯技术,广泛应用于食品、医药、染料等生产过程。结合超声法、图像法和光学浊度法研究了无水醋酸钠的结晶过程,通过超声模型结合最优正则化反演算法求解晶体颗粒粒度分布,在线测量了不同降温速率和搅拌速率下的析晶温度、晶体尺寸和形貌变化。结果表明：较快的降温速率可促进晶体生长,搅拌速率较快初期晶体尺寸较大,但后期由于颗粒磨损致平均尺寸减小;相同条件下光学浊度法较超声法较早测得析出晶体,实验中析晶温度偏差低于15%;图像法直观观察到了晶体生长过程并较好表征了晶粒形貌,与超声法具有接近的晶粒增长趋势,当晶体稳定析出后两者体积中位径偏差在15%以内,均较好地反映了实验条件对结晶晶体尺寸的影响。光学方法可更精细体现结晶初期特征,而在结晶后期超声法仍可对粒径进行有效测量。     

On the use of process analytical technologies and population balance equations for the estimation of crystallization kinetics. A case study

The batch cooling solution crystallization of ammonium oxalate was performed in water at various constant cooling rates. Measurements of the solute concentration were obtained using in situ attenuated total reflectance fourier transform infrared (ATR‐FTIR) spectroscopy, and final estimates of the crystal size distribution (CSD) were computed; thanks to in situ image acquisition and off‐line image analysis. The crystallization process was then simulated using population balance equations (PBEs). Estimates of the nucleation and the growth parameters were computed through model/experiments fitting. According to the cooling rate, the PBE model allowed distinguishing between two distinct crystallization regimes, separated by an “intermediate regime.” The respective contributions and shortcomings of solute concentration measurements and granulometric data to the identification of nucleation and growth kinetic parameters are analyzed and discussed. It is shown in particular that no real separate estimation of nucleation and growth parameters can be obtained in the absence of CSD data. © 2011 American Institute of Chemical Engineers AIChE J, 2012     

FOX-7晶体的制备和热性质

为了解FOX-7晶体形状和尺寸对其热性质的影响,通过冷却结晶和阻溶剂结晶的方法从DMF/H2O、NMP/H2O体系中获得FOX-7晶体.SEM结果表明,得到的晶体基本为延长的四面或六面棱柱和斜方菱晶两种类型,不同溶剂体系中得到晶体的尺寸不同和包含的颗粒数量也不同,超声处理得到的晶体具有单一形状,只是其表面较粗糙.DSC测试结果表明,不同结晶方式和不同溶剂体系中获得的FOX-7晶体的热性质有一定的差别.由此得出,不同的结晶方式和溶剂体系可以获得不同形状、尺寸和热性质的FOX-7晶体,而超声处理可以很好地控制晶体的形状.     

硫酸镍间歇结晶过程的工艺优化

文章采用聚焦光束反射测量仪(FBRM)研究了硫酸镍间歇结晶过程晶体的成核和生长规律,考察了搅拌桨形式、降温方式、搅拌速率、晶种添加量和晶种粒度对结晶过程中结晶动力学的影响。结晶工艺条件优化后,得到了粒径大、晶型好、分布均匀的晶体产品,为硫酸镍间歇结晶过程的工艺优化和工业放大提供了依据。     

Particle shape manipulation and optimization in cooling crystallization involving multiple crystal morphological forms

A population balance model for predicting the dynamic evolution of crystal shape distribution is further developed to simulate crystallization processes in which multiple crystal morphological forms co‐exist and transitions between them can take place. The new model is applied to derive the optimal temperature and supersaturation profiles leading to the desired crystal shape distribution in cooling crystallization. Since tracking an optimum temperature or supersaturation trajectory can be easily implemented by manipulating the coolant flowrate in the reactor jacket, the proposed methodology provides a feasible closed‐loop mechanism for crystal shape tailoring and control. The methodology is demonstrated by applying it to a case study of seeded cooling crystallization of potash alum. © 2009 American Institute of Chemical Engineers AIChE J, 2009     

Application of inline imaging for monitoring crystallization process in a continuous oscillatory baffled crystallizer

In this study, an in situ imaging system has been analysed to characterize the crystal size, the shape and the number of particles during a continuous crystallization process in a Continuous Oscillatory Baffled Crystallizer (COBC). Two image analysis approaches were examined for particle characterization in the suspension containing both small nuclei and larger grown crystals (nonspherical and irregular in shape). The pattern matching approach, in which the particles are approximated to be spherical, did result in an overestimation of the size. Alternatively, a segmentation‐based algorithm resulted in reliable crystal size and shape characteristics. The laser diffraction analysis in comparison to the image analysis overestimated the particle sizes due to the agglomeration of particles upon filtration and drying. The trend in the particle counts during the start of crystallization process, including nucleation, determined by the image analysis probe was comparable with the one measured by FBRM, highlighting the potential of in situ imaging for process monitoring. © 2018 American Institute of Chemical Engineers AIChE J, 64: 2450–2461, 2018     

Effects of crystallization conditions on sedimentation in canola oil

The effects of various factors on sediment formation in canola oil were studied. The crystallization temperature of sediment varied with cooling rate, whereas the melting temperature depended on heating rate as well as the cooling rate during sediment formation. The final crystal size depended on cooling rate. The crystal habit of sediment was generally rod-like but could change to a round and leaf-like shape at low cooling rates (<0.5°C/min). Crystal nucleation occurred in the initial stage of crystallization, while crystal growth was observed during the whole crystallization process, decreasing as cooling proceeded. Crystal growth rate of the sediment was proportional to the crystal surface area Lecithin did not affect the phase transition temperatures of sediment, but retarded crystal growth.     

Modelling of crystallization process and optimization of the cooling strategy

To obtain a uniform and large crystal in seeded batch cooling crystallization, the cooling strategy is very important. In this study, an optimal cooling strategy is obtained through simulation and compared to linear and natural cooling strategies. A model for a crystallization process in a batch reactor is constructed by using population balance equation and material balance for solution concentration, and a prediction model for meta-stable limit is formulated by the dynamic meta-stable limit approach. Based on this model, an optimal cooling strategy is obtained using genetic algorithm with the objective function of minimizing the unwanted nucleation and maximizing the crystal growth rate. From the simulation results, the product from the optimal cooling strategy showed uniform and large crystal size distribution while products from the other two strategies contained significant amount of fine particles.     

Multiple recrystallization behavior of poly(1,1-dimethysilacyclobutane): A combined calorimetric and small angle X-ray scattering study

The crystallization and melting behavior of a series of poly(1,1-dimethysilacyclobutane) (PDMSB) samples accessible via living anionic polymerization protocols with different molar masses were studied by differential scanning calorimetry (DSC), small angle X-ray scattering (SAXS) and X-ray diffraction (XRD). It has been found that in the cooling process only one crystallization exotherm was observed by using DSC, but in subsequent heating two clear endotherms appeared. DSC measurements additionally revealed that there will be two melting peaks when the isothermal crystallization temperature, T_c, is low enough, but only one melting peak could be obtained when T_c is high. The value of the lower melting peak T_m1 increased with T_c, showing the increase of lamella thickness due to increase of T_c, which was also evidenced by using SAXS measurements. However, the value of the higher melting peak T_m2 stayed almost constant with T_c. XRD analysis showed that there is only one crystal form independent of T_c. While SAXS results revealed only one long range order, indicating this multiple endotherms phenomenon was caused by the recrystallization process during heating, but not isothermal thickening and thinning process as observed for e.g. poly(ethylene oxide)s (PEOs) with low molar masses. Heating rate dependent DSC experiments showed that there are two recrystallization processes with different time scales. The recrystallization at low temperature always showed up independent of heating rate while the recrystallization at high temperature showed up only when the heating rate was very low. This interesting phenomenon was explained by the different energy barrier for the recrystallization process at low and high temperature as shown by in situ SAXS measurements during heating.     

Optimization of temperature swing strategy for selective cooling crystallization of α-form l-glutamic acid crystals

l-glutamic acid can be crystallized as metastable α-form and stable β-form crystal. The α-form is desired because of its prismatic shape. Production of α-form of l-glutamic acid by cooling crystallization is not well-defined and α-form solid is commercially not available. In this study, an optimal cooling strategy to selectively produce large and narrowly distributed α-crystals is found by modeling and optimizing the crystallization and polymorphic transformation of l-glutamic acid. The optimal temperature profile is found to be cooling-heating-cooling concept where short nucleation period is followed by growth period in metastable zone. The obtained α-form of L-glutamic acid by optimal strategy had improved mean size, distribution, and purity compared to constant cooling.     

Entropic analysis of a batch-crystallization process

The analysis of crystallization processes considering an entropic perspective is the primary purpose of this paper. Although the modeling of crystallization processes is well established in the literature, the entropic representation of these processes still needs to be studied. The modeling considering the second law of thermodynamics was investigated, which resulted in a model that represents the entropy production rate of a batch-crystallization process. The results indicate that the entropy production of the batch cooling crystallization is related to the variability of the crystal size distribution. The model of the entropy production rate could be used as a restriction criterion for multi-objective optimization of the cooling temperature profile of a batch crystallizer to improve the quality of the crystal size distribution of the final product.     

An algorithm for analyzing noisy, in situ images of high-aspect-ratio crystals to monitor particle size distribution

An image analysis algorithm for extracting particle size distribution information from in situ images of suspended, high-aspect-ratio crystals is described. The accuracy of the algorithm is assessed using comparisons with results obtained by manually sizing the imaged crystals. The algorithm can accurately extract crystal size information from in situ images of suspended, high-aspect-ratio crystals for moderate solids concentrations. The algorithm's performance declines somewhat for high solids concentrations and high levels of particle attrition. The speed with which the algorithm analyzes the images is suitable for real-time monitoring and control of particle size distribution mean and variance.     

阿司匹林结晶过程的在线分析

晶体的颗粒尺寸分布和形状是结晶产品的两个关键质量指标,不仅影响结晶产品的性质,还影响下游的过滤、干燥及运输存储等过程。利用超声粒度分析仪、衰减全反射傅里叶红外光谱仪、浊度仪与二维成像系统等分析仪器在线测量了不同搅拌速率和不同降温速率下阿司匹林乙醇溶液结晶过程中温度、浓度、颗粒尺寸分布和形状的变化情况。实验结果表明：较低的降温速率或者较大的搅拌速率条件下得到含有大量细晶的阿司匹林结晶产品;较高的降温速率下得到长宽比较大的阿司匹林结晶产品。调节降温速率和搅拌速率是一种有效控制阿司匹林结晶产品尺寸分布与形状的方法。     

DYNAMIC OPTIMAL CONTROL OF BATCH CRYSTALLIZATION PROCESSES

This work applies an on-line optimal control strategy developed by Zhang (2001) to two cooling batch crystallization processes. The algorithm initially finds the optimal crystallizer temperature and subsequently uses a feedback control system in order to achieve the desired final product quality of the crystals expressed in terms of the final crystal size distribution. In both batch processes, it is shown that the on-line optimal control approach provides better final product quality as compared with a simplified optimal cooling policy. The improvement is especially noticeable in the presence of plant/model mismatch or errors in the initial conditions.     

对二甲苯悬浮熔融结晶动力学

结晶法是工业上生产对二甲苯的主要方法之一。现有对二甲苯结晶动力学参数均单纯由结晶母液的温度和浓度变化通过非线性优化法而获得,未检测对二甲苯的晶体粒度数据,因而其准确性难以得到保证。本文利用超声在线粒度仪(OPUS)检测对二甲苯晶体的粒度分布,通过添加晶种的间歇悬浮熔融结晶实验,应用矩量变换法测定82%（质量）对二甲苯-间二甲苯体系中的对二甲苯结晶动力学。利用最小二乘法对动力学实验数据进行多元线性回归后得到了对二甲苯结晶动力学方程,研究结果表明,在对二甲苯悬浮熔融结晶过程中,溶液相对过饱和度对对二甲苯晶体成核速率的影响大于对晶体生长速率的影响,搅拌速率对成核过程影响明显,而晶浆悬浮密度对成核速率的影响不大。     

Crystallization behaviour of random block copolymers of poly(butylene terephthalate) and poly(tetramethylene ether glycol)

The morphology and crystallization behaviour of random block copolymers of poly(butylene terephthalate) and poly(tetramethylene ether glycol) have been investigated. Single crystals have been grown in thin films crystallized from the melt. Well defined lamellae, exhibiting (hkO) single crystal electron diffraction patterns have been observed in copolymers containing down to 49 wt% (0.83 mole fraction) poly(butylene terephthalate). WAXS and electron diffraction support a model of a relatively pure poly(butylene terephthalate) crystal core with the poly(tetramethylene ether glycol) (soft segment) sequences and short hard segments being rejected to the lamellar surface and the soft segment rich matrix. The lateral dimensions of the lamellae are determined by the number of hard segment sequences long enough to traverse the stable crystal size at the crystallization temperature. This leads to an initial population of crystals formed at T_c and a second set of smaller crystals that grow from the short hard segment sequences upon cooling to room temperature. The result is fractionation by sequence length due to a coupling of the sequence distribution with the stable crystal size at the crystallization temperature.