Determining Phase Separation Dynamics with an Automated Image Processing Algorithm

Abstract

The problems of extracting products efficiently from reaction workups are often overlooked. Issues such as emulsions and rag layer formation can cause long separation times and slow production, thus resulting in manufacturing inefficiencies. To better understand science within this area and to support process development, an image processing methodology has been developed that can automatically track the interface between liquid–liquid phases and provide a quantitative measure of the separation rate of two immiscible liquids. The algorithm is automated and has been successfully applied to 29 cases. Its robustness has been demonstrated with a variety of different liquid mixtures that exhibit a wide range of separation behaviormaking such an algorithm suited to high-throughput experimentation. The information gathered from applying the algorithm shows how issues resulting from poor separations can be detected early in process development.

Automated Image-Based Color Analysis as an Accessible and Widely Applicable PAT Tool

“…Camera-assisted interphase detection has in fact already been described in the late 1990s, but there was no easily accessible software available . There has also been a method described using a colored float, and recently, examples of dedicated level-detection algorithms were published. ,, …”

“…20 There has also been a method described using a colored float, and recently, examples of dedicated level-detection algorithms were published. 5,21,22 The following example demonstrates the use of image analysis for endpoint determination in a liquid−liquid (toluene−water) separation (Figure 6). The L*-value (luminance or lightness) is compared to the turbidity trend from the microscopy probe (using Mettler Toledo's IC Vision software in combination with a MT EasyViewer 100) and used to determine the endpoint.…”

“…This trend has also been observed in the pharmaceutical industry, with examples ranging from process development to packaging. − While ML algorithms are improving, they often still rely on large sets of historical data, which are used to train the ML algorithm, and can typically only be applied for one dedicated task. In stark contrast with the complex image patterns that ML algorithms analyze, many inherently intricate pharmaceutical reactions and unit operations do involve simple appearance changes in color, turbidity, phase, etc. , Usually, these change points mark the start or end of a process. Even more, the rate of these changes often represents how fast a process proceeds, which can be compared between processes ran under different conditions; therefore, a simple time-dependent color analysis can already provide a wealth of (kinetic) information.…”

“…In stark contrast with the complex image patterns that ML algorithms analyze, many inherently intricate pharmaceutical reactions and unit operations do involve simple appearance changes in color, turbidity, phase, etc. 5,6 Usually, these change points mark the start or end of a process. Even more, the rate of these changes often represents how fast a process proceeds, which can be compared between processes ran under different conditions; therefore, a simple time-dependent color analysis can already provide a wealth of (kinetic) information.…”

Vision-Guided Automation Platform for Liquid–Liquid Extraction and Workup Development

“…To our knowledge, this was the first example in the process development space to integrate commercially available components for hands-free extraction processing with fully automated phase detection to drive appropriate aliquoting and sample preparation for distribution coefficient determination. More recently, researchers at the University of Leeds and AstraZeneca have also designed an image processing algorithm to quantitatively track the settling velocity of biphasic LLE samples …”