A liquid chromatography gradient automatic optimization method based on real-time feedback of chromatographic peaks

Abstract

This invention discloses an automated gradient optimization method for liquid chromatography based on real-time peak feedback, applicable to liquid chromatography systems including detectors, chromatographic pumps, and data acquisition and control units. The method includes: executing an initial gradient; acquiring chromatograms and extracting the retention time and half-peak width of each peak; calculating the resolution of adjacent peaks; if there are substandard peak pairs, reducing the rate of change of the mobile phase gradient within the corresponding elution window to improve resolution; if all peaks meet the standard, identifying and compressing the inter-peak blank time period to shorten analysis time; and outputting the optimal gradient through closed-loop iteration when both resolution and blank periods are simultaneously satisfied. This invention is fully automated, using deterministic feedback based on measured peak parameters. It uniquely incorporates resolution correction and time compression into the same closed loop and automatically handles their contradictions, ultimately converging to the optimal balance between resolution and analysis time, significantly reducing solvent consumption and manual intervention.

Claims

1. A method for automatic optimization of liquid chromatography gradients based on real-time feedback of chromatographic peaks, applied to a liquid chromatography system including a detector, a chromatographic pump, and a data acquisition and control unit, characterized in that, The data acquisition and control unit performs the following steps: Step S100: Control the chromatographic pump to execute the preset initial linear gradient program B0, perform gradient elution on the sample, and acquire the chromatogram through the detector; Step S200: Perform baseline correction and noise filtering on the chromatogram, then identify all chromatographic peaks and extract the retention time t of each peak. Ri and half-peak width W i , where i is the peak number sequentially by time; Step S300: For each pair of adjacent peaks i and i+1, calculate the separation degree R according to the following formula. s(i) : R s(i) =2×(t R(i+1) -t Ri ) / (W i +W (i+1) ) If any R exists s(i) Less than the preset separation threshold R std If all R s(i) All are greater than or equal to R std If so, proceed to step S500; Step S400: For each pair of adjacent peaks with unsatisfactory resolution, determine its elution window on the chromatographic time axis, and within this window, reduce the gradient rate of change of the organic phase ratio in the mobile phase as time increases, generating a correction gradient program B. new And write to the chromatography pump control unit, then return to step S100; Step S500: Establish a time axis occupancy map based on the retention time and half-peak width of each peak, and identify compressible blank time periods between adjacent peak occupancy intervals whose duration is greater than a preset blank threshold. Step S600: Increase the gradient change rate of the organic phase ratio in the mobile phase during the compressible blank time period to compress the analysis time and generate an optimized gradient program B. opt And write to the chromatography pump control unit, then return to step S100; Step S700: Terminate the iteration loop and output the final gradient program and corresponding chromatogram when both of the following conditions are met: Condition 1: The resolution of all adjacent peaks is greater than or equal to R. std Condition 2: There is no compressible blank time period in the chromatogram.

2. The method according to claim 1, characterized in that, The method for determining the elution window in step S400 is as follows: the retention time of the peak before the peak that does not meet the standard is pushed forward by twice its half-peak width as the starting point of the window, and the retention time of the subsequent peak is pushed backward by twice its half-peak width as the ending point of the window, and the window boundary does not exceed the analysis time range.

3. The method according to claim 1, characterized in that, The method for reducing the gradient change rate in step S400 is to replace the organic phase ratio-time function in the elution window with a linear function and a monotonically increasing upward convex function with a gradually decreasing rate of change.

4. The method according to claim 3, characterized in that, The convex monotonically increasing function is in the form of a power function φ(t) = φ(t). start )+a·(tt start ) b Where φ represents the organic phase ratio, a is the amplitude coefficient, and b is the curvature exponent and satisfies 0 <b<1。 5. The method according to claim 1, characterized in that, The method for identifying the compressible blank time period in step S500 is as follows: for adjacent peaks i and i+1, calculate the blank duration Δt. blank(i) = (t R(i+1) - W (i+1) ) - (t Ri + W i If Δt blank(i) >K × (W i + W (i+1) If K = 2, it is determined to be a compressible blank time period, where K is the preset blank tolerance coefficient, and the value range is 2~5.

6. The method according to claim 1, characterized in that, The method to increase the gradient change rate in step S600 is as follows: keep the gradient program of the peak elution section outside the compressible blank time period unchanged, and only increase the slope of the organic phase ratio-time function corresponding to the compressible blank time period to m times the original slope, where m ranges from 1.5 to 3.

0.

7. The method according to claim 1, characterized in that, In steps S400 and S600, a smoothing process is performed at the connection boundary between the modified gradient segment and the unmodified segment. The smoothing process adopts a weighted cosine transition or a linear interpolation transition to ensure that the entire gradient program is continuous in time and the first derivative is continuous.

8. The method according to claim 1, characterized in that, The method also includes an anomaly protection step: presetting a maximum allowed number of iterations N. max When the number of iterations reaches N max The loop is forcibly terminated, the current best gradient is output and the user is prompted; and when the resolution of the same adjacent peak pair continues to decrease in multiple consecutive iterations, and the current resolution decreases by more than 20% relative to the highest resolution reached by the peak pair during the optimization process, the loop is automatically paused and the user is prompted to check the column status.

9. The method according to claim 5, characterized in that, The blank tolerance coefficient K is not a fixed value, but is based on the minimum resolution R of all adjacent peaks in the current chromatogram. s_min Dynamic setting: When R s_min When K is ≤ 1.8, it takes values ​​from 4.0 to 5.0; when it is 1.8... <R s_min When R ≤ 2.5, K takes values ​​from 3.0 to 4.0; when R s_min When K > 2.5, K takes values ​​from 2.0 to 2.

5.

10. The method according to claim 1, characterized in that, In the iterative cycle of step S700, when the chromatogram after step S600 is determined by step S300 to have peak pairs with unsatisfactory resolution, the method first executes step S400 to restore the resolution, and after the resolution is restored to the standard, steps S500 and S600 are executed again for blank compression. This collaborative self-repair mechanism achieves the optimal balance between resolution and analysis time.

Images