Calibration transfer is commonly used for spectra obtained on different spectrometers or other conditions. This paper reports a multivariate transfer approach for 2D diffusion‐ordered spectroscopy (DOSY) measurements among high‐resolution nuclear magnetic resonance (NMR) spectrometers on the basis of partial least squares (PLS) regression. As the test system previously published quantitative model to predict molecular weight of heparin, low molecular weight heparin (LMWH) was used.
For multivariate modeling, piecewise direct standardization (PDS) and direct standardization (DS) were employed. PDS showed the best performance for estimating heparin molecular weight resulting in a significant decrease in root mean square error of prediction (RMSEP) from 647 and 393 Da without standardization to 513 and 135 Da when PDS was applied for heparin and LMWH, respectively. DS is not recommended for high‐resolution NMR data.
The study showed the success of standardization procedure in cases when multivariate models have to be applied to the 2D NMR spectra recorded on different NMR spectrometers. The proposed approach appears to be a valuable achievement towards application of transfer models also to 2D NMR data.