Carbohydrates, vital components of biological systems, are well-known for their structural diversity. Nuclear Magnetic Resonance (NMR) spectroscopy plays a crucial role in understanding their intricate molecular arrangements and is essential in assessing and verifying the molecular structure of organic molecules. An important part of this process is to predict the NMR chemical shift from the molecular structure. This work introduces a novel approach that leverages E(3) equivariant graph neural networks to predict carbohydrate NMR spectral data. Notably, our model achieves a substantial reduction in mean absolute error, up to threefold, compared to traditional models that rely solely on two-dimensional molecular structure. Even with limited data, the model excels, highlighting its robustness and generalization capabilities. The model is dubbed GeqShift (geometric equivariant shift) and uses equivariant graph self-attention layers to learn about NMR chemical shifts, in particular since stereochemical arrangements in carbohydrate molecules are characteristics of their structures. 

Small-angle X-ray scattering (SAXS) is a useful technique for nanoscale structural characterization of materials. In SAXS, structural and spatial information is indirectly obtained from the scattering intensity in the spectral domain, known as the reciprocal space. Therefore, characterizing the structure requires solving the inverse problem of finding a plausible structure model that corresponds to the measured scattering intensity. Both the choice of structure model and the computational workload of parameter estimation are bottlenecks in this process. In this work, we develop a framework for analysis of SAXS data from disordered materials. The materials are modeled using Gaussian Random Fields (GRFs). We study the case of two phases, pore and solid, and three phases, where a third phase is added at the interface between the two other phases. Further, we develop very fast GPU-accelerated, Fourier transform-based numerical methods for both structure generation and SAXS simulation. We demonstrate that length scales and volume fractions can be predicted with good accuracy using our machine learning-based framework. The parameter prediction executes virtually instantaneously and hence the computational burden of conventional model fitting can be avoided. Copyright © 2022 Röding, Tomaszewski, Yu, Borg and Rönnols.

This study describes development of methods for identification and quantification of a set of mono- and diaryl compounds in lignin. The diaryls consist of a set of stilbenes and resinols, while the monoaryls consist of guaiacol, vanillin, acetovanillone and three syringyl counterparts. The methods rely on HPLC-MS and GC-MS. These were quantified in a set of technical lignins and were found to comprise 1 – 11 % of the total mass of the samples.

Most of our knowledge on kraft pulping comes from studies on dissolved lignin in the freely drainable black liquor and isolated residual lignin in pulp. However, entrapped liquor in the delignified chips has been shown to differ significantly from the free liquor. The present study has compared three liquor fractions: free, lumen and fiber wall liquor. The free liquor was obtained by draining the delignified chips, the lumen liquor was separated by centrifugation and the fiber wall liquor by subsequent leaching. The liquor in the fiber wall had the lowest concentration of lignin and hydrosulfide ions and the highest concentration of monovalent cations. The dissolved lignin in the fiber wall liquor had the highest molar mass and the highest content of xylan. The highest concentration of dissolved lignin was in the liquor filling the lumen cavities. The lignin in the free liquor had the lowest molar mass and the lowest content of lignin structures containing β-O-4 linkages and aliphatic hydroxyl groups. The lowest mass transfer rate of dissolved lignin was from the lumen liquor to the free liquor probably restricted by the tortuosity of the chip. 

Small angle X-ray scattering (SAXS) is extensively used in materials science as a way of examining nanostructures. The analysis of experimental SAXS data involves mapping a rather simple data format to a vast amount of structural models. Despite various scientific computing tools to assist the model selection, the activity heavily relies on the SAXS analysts’ experience, which is recognized as an efficiency bottleneck by the community. To cope with this decision-making problem, we develop and evaluate the open-source, Machine Learning-based tool SCAN (SCattering Ai aNalysis) to provide recommendations on model selection. SCAN exploits multiple machine learning algorithms and uses models and a simulation tool implemented in the SasView package for generating a well defined set of datasets. Our evaluation shows that SCAN delivers an overall accuracy of 95%-97%. The XGBoost Classifier has been identified as the most accurate method with a good balance between accuracy and training time. With eleven predefined structural models for common nanostructures and an easy draw-drop function to expand the number and types training models, SCAN can accelerate the SAXS data analysis workflow.

Benchtop nuclear magnetic resonance (NMR) spectroscopy is an emerging field with an appealing profile for industrial applications. The instrumentation offers the possibility to measure NMR spectra in situations where high-field NMR spectroscopy is considered too expensive or complicated. In this study, we investigated the scope and limitations of 1H NMR measurements on kraft lignins and black liquors at low magnetic field strengths (1.0 and 1.5 T). The ability to quantify different classes of compounds was investigated and found to be promising. NMR-based diffusion measurements were performed, with the aim of gaining insight into the molar mass of the lignins at hand. These measurements were fast, repeatable and in good agreement with established methods.

Size Exclusion Chromatography (SEC) of lignin poses many challenges. In numerous studies chromatograms of lignin show a bimodal molar mass distribution. Is this a true characteristic of lignin, is it caused by molecular associations or aggregations, or could it could be an artefact from using column combinations with porosities that do not match properly? To improve resolution and enable separation over a larger molar mass range, multiple columns with different porosities are often connected in series. If the porosities do not match properly, the result appears as a shoulder or bimodality in the chromatogram. To understand whether the bimodal distribution is a sample characteristic or an analyze artefact, we have used different columns, column combination and samples to see when the results is a bimodal distribution and when only one peak is formed. Results show that the bimodality of lignin can be an artifact originating from column mismatch. Using single porosity columns with a low molar mass cut-off should be avoided since it can cause false bimodality.

This paper describes improvements in the methodology for NMR spectroscopic analysis of dissolved kraft lignins and black liquors, for structural analysis as well as for reaction monitoring. The described methodologies are variations and applications of non-uniform sampling (NUS) in 2D spectra and diffusion ordered spectroscopy (DOSY), with reduced experiment time and/or increased resolution and novel ways to track reactions through monitoring of diffusion of a reactant mixture.

The use of NMR spectroscopy at high magnetic fields is a common tool in the analysis of lignin samples. In the presented study, NMR measurements on a group of softwood and hardwood kraft lignins at low field (1.0 T) with a benchtop NMR spectrometer, containing a permanent magnet, are investigated and evaluated. NMR based diffusion measurements were performed, for which the results were found to be fast, repeatable, and in good agreement numbers to measurements at high field. Measurements were also performed on a sample in alkaline solution, as a model for black liquor analysis, with promising initial results.