Chromatographic Separation Model Calibration Using Sparse Multi-Mode Runs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current chromatographic separation methods face challenges in determining molecule-specific parameters for isotherm equations, particularly when binding behavior does not follow simple stoichiometric displacement models, leading to time-consuming and high-dimensional curve fitting processes.
Innovation Solution
A method involving a series of chromatography runs with non-identical operation modes, recording retention behavior data, generating a computational grid, selecting an adsorption model, calculating partitioning coefficients, and minimizing an aggregated distance measure to quickly derive model parameters, applicable to mechanistic modeling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional curve fitting methods are used to determine adsorption model parameters, then measurement precision can be achieved, but the method requires time-consuming computation and high-dimensional parameter optimization
Solution Approach 1:
The patent segments the parameter determination process into two distinct phases: (1) determining linear region parameters from low-concentration experimental data, and (2) determining non-linear region parameters from high-load chromatography experiments. This segmentation avoids the need for simultaneous optimization of all parameters, dramatically reducing computation time while maintaining accuracy.
Solution Approach 2:
The patent performs preliminary determination of linear region parameters before tackling the non-linear parameters. By first establishing the linear region characteristics using simple experiments, the foundation is laid for subsequently determining non-linear parameters more efficiently, avoiding the need to optimize all parameters simultaneously from scratch.
2Device complexity
If the Yamamoto method is used for parameter determination, then the dimensionality of the parameter problem is reduced, but the method is limited to simple ion-exchange chromatography and requires the binding behavior to follow stoichiometric displacement models
Solution Approach 1:
The patent introduces dynamic adaptation by providing multiple adsorption model options (Langmuir, Freundlich, Hill, etc.) that can be selected based on the specific binding behavior of the target molecule. The method dynamically adjusts to different binding mechanisms rather than being constrained to a single model, thereby maintaining versatility while managing complexity through the segmented approach.
Solution Approach 2:
The patent changes the approach to parameter determination by separating linear and non-linear parameter optimization into distinct stages. This parameter change strategy allows the method to handle complex binding models without requiring simultaneous optimization of all parameters, thus maintaining adaptability to different binding mechanisms while controlling computational complexity.
3Reliability
If high-dimensional curve fitting is performed to determine all adsorption parameters simultaneously, then comprehensive model calibration is achieved, but the computational expenditure and time required increase significantly
Solution Approach 1:
The patent segments the comprehensive parameter determination into two reliable phases: linear region parameters from low-concentration data and non-linear parameters from high-load data. This segmentation maintains model calibration completeness by ensuring both regions are properly characterized, while simultaneously improving productivity by avoiding the computational burden of simultaneous high-dimensional optimization.
Solution Approach 2:
The patent uses the linear region parameters as an intermediary step that facilitates the subsequent determination of non-linear parameters. By first establishing the linear region characteristics, the method creates a foundation that simplifies the subsequent non-linear parameter optimization, thereby achieving comprehensive calibration with reduced computational expenditure.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for rapid derivation of model parameters from few experiments at low load density, reducing complexity and time required for chromatography model calibration and setting determination.
Implementation Method 1
The adsorption and desorption of solutes is described by so-called isotherm equations
Implementation Method 2
chromatographic separation of at least one target molecule
Data Source
AI summary
The present invention relates to a method (500), the method comprising performing (510) an experiment as a series of chromatography runs, wherein the series of chromatography runs comprises at least two chromatography runs using non-identical operation modes, recording (520) retention behavior data of at least one target molecule for each of the chromatography runs, wherein the retention behavior data indicates at least measured points of elution (M-POE) and properties of the measured points of elution (M-POEJ), generating (530) at least one computational grid (MTX) in at least one dimension, where the extent of the grid is indicative of maximum ranges between a starting point of the experiment (SP) and an end point of the experiment (EP), selecting (540) an adsorption model from a set of predetermined adsorption models, the adsorption model having initial candidate values of adsorption model parameters, calculating partitioning coefficient values (550) for all points in the computational grid using the selected adsorption model, calculating integrals (560) of a function of the calculated partitioning coefficient values over at least two paths in the computational grid (MTX) to obtain properties of calculated points of elution (C-POEJ), calculating (570) an aggregated distance measure (D) from the properties of calculated points of elution (C-POEJ) to respective experimental reference values of properties of the measured points of elution (M-POEJ), selecting (580) new candidate values of adsorption model parameters, repeating (590) the steps of calculating (550) partitioning, calculating integrals (560), calculating the distance measure (570), and selecting (580) new candidate values until the aggregated distance measure (D) is minimized, and determining parameters 595 of the adsorption model using the selected values of adsorption model parameters that minimizes the aggregated distance measure (D).