Method for generating a characteristic representation of a substance

Abstract

A method for generating a characteristic representation of a substance is described. The method comprises: obtaining one or more portions of nuclear magnetic resonance (NMR) data of the substance, wherein each NMR data portion comprises one or more NMR data sets each comprising intensity data and relaxation data, processing each NMR data portion(s) into a multidimensional distribution of fitted data sets by performing multiple fittings to a function, wherein the function comprises at least 3 fitting parameter(s), and generating the characteristic representation of the substance from the fitted data sets. The processing comprises applying at least one start guess. Each fitted data set comprises data embedding at least one fitted data subset comprising an apparent longitudinal relaxation time (T1*), an associated apparent transverse relaxation time (T2*), and an associated intensity (M0), preferably the function is a function which represents a model of the obtained NMR data portion(s).

Description

[0001] METHOD FOR GENERATING A CHARACTERISTIC REPRESENTATION OF A SUBSTANCE

[0002] TECHNICAL FIELD

[0003] The invention relates to a method and a system for generating a characteristic representation, for example a multidimensional map, such as a T1-T2 map, of a substance, for example a complex substance, such as a food product, a substance containing chemical pollutants, or a manure slurry.

[0004] BACKGROUND ART

[0005] Substances, in particular complex substances comprising a plurality of elements, such as fat and / or water and / or protein and / or nucleic acids and / or carbohydrates and / or vitamins and / or minerals and / or chemical pollutants and / or nutrients, in which chemical species of each element interacts with other chemical species in the substance are generally characterized in a laboratory settings using expensive devices and time consuming methods, such as chromatography.

[0006] Substance characterizations are used and often required in many applications. One example is the food industry where food companies are required to label the nutrients in the product, and thus must characterize their products frequently in terms of constituents or in a continuous flow to monitor variations. Another example is conversion of biomass to recycle energy, where it is desired to characterize a slurry during the conversion from biomass to energy to identify the distribution of elements, in order to control the process. A third example is the chemical industry where companies transform organic and inorganic raw materials using chemical processes to chemicals, and where characterization of the chemicals and intermediate products are essential to control the process and to ensure quality products.

[0007] In general, nuclear magnetic resonance (NMR) has shown to be an excellent non- invasive technique for studying microscopic molecular interactions, wherein Laplace transformations usually are applied for converting NMR echo data into a distribution of relaxation times.

[0008] WO 2018 / 163188A1 discloses a method of characterizing chemical and / or morphological features of a material, comprising acquiring energy relaxation data from 1H low field nuclear magnetic resonance (1H LF-NMR) measurements of the material, converting the relaxation signals into a multidimensional distribution of longitudinal and transverse relaxation times by solving an inverse problem under both L1 and L2 regularizations and further imposing a non-negativity constraint, and identifying one or more characteristics of the material with the aid of the multidimensional T1-T2 distribution.

[0009] There is a substantial need for alternative or improved methods for characterizing a substance and in particular a method that is very efficient and / or a method which is suitable for generating valuable information of a substance even for a complex substance having a complex composition.

[0010] DISCLOSURE OF THE INVENTION

[0011] An objective of the invention is to provide a method for characterizing a substance to provide valuable information about chemical, physical, biological, and / or morphological features of the substance.

[0012] In an embodiment it is an objective to provide a method for producing a characteristic representation of a substance wherein the characteristic representation describes one or more characteristics features of the substance, preferably correlated to the content of the substance and / or components thereof and / or chemical and / or physical and / or biological impacts and / or interactions between components or moieties of components of the substance.

[0013] In an embodiment it is an objective to provide a method for producing a characteristic representation of a substance wherein the characteristic representation describes the substance in the form of a depiction or data representing a depiction of the substance.

[0014] In an embodiment it is an objective to provide a method for producing a characteristic representation of a substance wherein the characteristic representation describes at least one quality feature of the substance, preferably wherein the characteristic representation comprises a depiction or data representing a depiction of a quality parameter of the substance or a signature of a change in a process involving the substance.

[0015] In an embodiment it is an objective to provide a method for producing characteristic representation of a substance, which characteristic representation.

[0016] In an embodiment it is an objective to provide an NMR system for performing a method for characterizing a substance to provide valuable information about chemical, physical, biological, and / or morphological features of the substance. In an embodiment it is an objective to provide a method for generating a trained computer for performing a processing for producing or generating a characteristic representation of a substance of a method of the invention.

[0017] These and other objects have been solved by the invention or embodiments thereof as defined in the claims and / or as described herein below.

[0018] It has been found that the invention or embodiments thereof have a number of additional advantages, which will be clear to the skilled person from the following description.

[0019] The method of the invention comprises a method for generating a characteristic representation of a substance comprising:

[0020] • recording nuclear magnetic resonance (NMR) data of a least one sample of the substance,

[0021] • producing one or more portions of nuclear magnetic resonance (NMR) data of the recorded NMR data, wherein each NMR data portion is produced to comprise one or more NMR data sets ,

[0022] • processing each NMR data portion(s) into a multidimensional distribution of fitted data sets comprising selecting a function and performing multiple fittings to the selected function, and

[0023] • producing the characteristic representation of the substance from the fitted data sets,

[0024] Each data set advantageously, comprises associated intensity data and relaxation data. The intensity data and relaxation data may be associated in that these data is from a common isotope.

[0025] The selected function comprises at least 3 fitting parameter(s).

[0026] The processing of each of the NMR data portion(s) comprises selecting and applying at least one start guess.

[0027] In the following the NMR data portion(s) and the NMR data set(s) are also referred to as input NMR data. The function comprises at least 3 fitting parameter(s) and is preferably a function which represents a model of the obtained NMR data portion(s).

[0028] The function is also referred to as a model function or a model equation, such as y = the function.

[0029] The processing comprises applying at least one start guess. Each fitted data set comprises data embedding at least one fitted data subset comprising fitted data set parameters comprising an apparent longitudinal relaxation time (T1*), an associated apparent transverse relaxation time (T2*), and an associated intensity (MO).

[0030] The one or more data sets of each NMR data portion advantageously comprises NMR data for a common NMR readable isotope. In an embodiment, the one or more NMR data portions comprises data for the same or for different isotopes. In an embodiment, the one or more NMR data portions comprises data of one or more readable isotopes interacting with other isotopes.

[0031] The phrases “NMR readable” and “NMR active” are herein used interchangeable.

[0032] The inventors have found that the method of the invention provides a very effective tool for accomplishing a chemical, physical, biological, and / or morphological analyses of the substance even where the substance is a complex substance comprising a plurality of components that may interact, have any impact on each other or even reacting with each other. Thus, the method may provide a very detailed analysis of the substance which may be demonstrated, embedded in or even directly revealed from the characteristic representation. Thus, in an embodiment the characteristic representation may even provide a unique visual characteristic, such as a biometric equivalent or analogous characteristic of the substance, such as a fingerprint e.g. a composition fingerprint, such as a quality fingerprint.

[0033] The fingerprint may advantageously encompass characteristics of the combination of elements and / or compounds of the substance that contribute to overall identification and / or quality of the substance.

[0034] The phrase "quality fingerprint" is herein used to mean a unique set of characteristics, parameters and / or attributes that define the quality of the substance. The method of the invention thereby provides a highly valuable tool for examining substances, e.g. for comparing with reference substances with known chemical, physical, biological, and / or morphological features, e.g. for determining if a quality parameter is fulfilled, for estimating compatibility, for estimating reactivity or for estimating a response factor to or of the substance.

[0035] Thus, the method of the invention has been found to provide a valuable tool for use in productions for example for controlling quality, such as at least one quality parameter and / or a quality fingerprint of a raw material, of a precursor material, and / or of a produced material, such as a food substance for example as part of process control.

[0036] The method of the invention has been found to be a useful and effective tool for fast and effectively detection of a potential presence of one or more impurities for example in a raw material or in a final product.

[0037] In an embodiment, the method may be used in forensic e.g. for analyzing a complex substance found at a crime location or the method may be applied for analyzing and generating a characteristic representation of a fluid sample of a biological organism, e.g. to compare with a characteristic representation reference substance found at a crime location.

[0038] It should be emphasized that the term "comprises / comprising" when used herein is to be interpreted as an open term, i.e. it should be taken to specify the presence of specifically stated feature(s), such as element(s), unit(s), integer(s), step(s), component(s), and combination(s) thereof, but does not preclude the presence or addition of one or more other features.

[0039] Throughout the description or claims, the singular encompasses the plural, and the plural encompasses the singular unless otherwise specified or required by the context.

[0040] The method of the invention may be referred to as a computer implemented invention, in that a computer system may advantageously be applied in at least the performing of the multiple fittings to the selected function. The term "wt%" means herein percent by weight.

[0041] The "an embodiment" should be interpreted to include examples of the invention comprising the feature(s) of the mentioned embodiment.

[0042] The term "substantially" should herein be taken to mean that ordinary product variances and tolerances are comprised. All features of the invention and embodiments of the invention as described herein, including ranges and preferred ranges, may be combined in various ways within the scope of the invention, unless there are specific reasons not to combine such features.

[0043] Unless other is specified, any properties, ranges of properties and / or determination and / or assay condition is given or provided at 1 atmosphere and 37 °C.

[0044] All features of the invention and embodiments of the invention as described herein including ranges and preferred ranges may be combined in various ways within the scope of the invention, unless there are specific reasons not to combine such features.

[0045] Nuclear magnetic resonance (NMR) spectroscopy is a well-known analytical technique for structure elucidation of small and macro molecules for both quantitative and qualitative analysis. The present invention provides an analysis tool that may result in a characteristic representation for a substance which may comprise several different molecules and components and where the method may provide a characteristic representation for an entire substance and not merely an individual molecule of the substance.

[0046] By the performing of the multiple fittings to the function, at least one fitted data set may be obtained, wherein each of the fitted data sets, preferably comprises embedded a plurality of fitted data subsets which fitted data subsets are collected to provide the characteristic representation of the substance, e.g. comprising data describing the characteristic representation of the substance and / or comprising a plot of data of the characteristic representation of the substance.

[0047] The performing of the multiple fittings to the function, may advantageously comprise performing multiple regression fittings to the function. Advantageously, the performing of the multiple fittings to the function comprises performing multiple fittings, such as at least two, such as at least 4, such as from 5 to 50 fittings to the function for each of the NMR data sets, wherein the multiple fittings advantageously comprise multiple regression fittings to the function.

[0048] The skilled person will understand that each individual fitting may have certain uncertainty. However, is has been found that when performing multiple fittings the statistical uncertainty of individual components may become very small and even practically irrelevant. Thereby the characteristic representation may be produced to have a very high accuracy, even providing that the produced characteristic representation constitutes an accurate and reproducible identification characteristic representation, such as a substance fingerprint.

[0049] Methods of generating NMR data such as NMR intensity data, NMR frequency data and NMR relaxation data are known in the art and may for example be as described in WO2022 / 117170.

[0050] The method comprises recording nuclear magnetic resonance (NMR) data of a least one sample of the substance. The recording of NMR data may comprise performing a plurality of readings wherein the plurality of readings may be identical or may differ from each other as further described below.

[0051] The producing of the one or more portions of nuclear magnetic resonance (NMR) data of the recorded NMR data, may comprise deriving data sets from the recorded NMR data and producing one or more NMR data portions from the derived datasets, wherein each data set is produced to comprise associated intensity data and relaxation data.

[0052] In an embodiment, each of the NMR data sets comprises frequency data, such as at least one of resonance frequency, chemical shift, quadrupole coupling, J coupling, or dipole-dipole coupling.

[0053] The selection of the desired resonance frequency data for the NMR data sets may preferably be performed in dependence of the isotope that the NMR data sets apply for. The frequency data may in particular be advantageous where the isotope that is applied is included in many different chemical or physical constellations of the substance.

[0054] Advantageously, each of the NMR data sets comprises at least one of inversion recovery data or echo NMR data. Preferably each of the NMR data sets comprises data representing at least one inversion recovery delay (I RD) and / or data representing at least one echo delay (ED). It has been found that inversion recovery delay data and / or echo delay data and / or any combination of these may increase discrimination between types of components of the substance such as fat components and non-fat components of the substance.

[0055] In an embodiment, the echo data comprises at least one of spin echo data, fast spin echo data, stimulated echo data, or gradient echo data. Such echo data may for example be obtained as described by Hahn in “Spin Echo”. Each of the NMR data sets may conveniently comprise data recorded with a pulse sequence comprising at least one period where the recorded data is affected by T1 (longitudinal) relaxation and / or at least one period influenced by T2 (transverse) relaxation.

[0056] In an embodiment, each of the NMR data sets comprises NMR data recorded with a pulse sequence comprising at least one period where the recorded data is affected by chemical exchange. The NMR data sets comprising NMR data recorded with a pulse sequence comprising at least one period where the recorded data is affected by chemical exchange may conveniently comprise frequency data.

[0057] In an embodiment, each of the NMR data sets comprises NMR data recorded by at least one reading comprising a pulse sequence with at least one period where the recorded data is affected by diffusion, such as diffusion of molecules, such as diffusion of ions, such as spin diffusion.

[0058] The analytical potential where the data sets of the input NMR data comprise multicomponent relaxation time or diffusion data may provide valuable information characterizing the substance.

[0059] Based on the evolution of spin systems, while applying data sets comprising two or more relaxation times and / or correlated diffusion data, such as their diffusion coefficients (D), may provide valuable characteristic of molecular species and specific dynamics that depend on the molecular environment and dipolar interactions of the substance and components of the substance. In addition to time resolution, information related to chemical exchanges and molecular transfers from one site to another of the components of the substance may be enhanced, making the resulting characteristic representation more detailed and providing additional information on the structural environment of the components of the substance.

[0060] Advantageously, the one or more of the NMR data portions comprises at least 2 NMR data sets, such as at least 3 NMR data sets, such as at least 4 NMR data sets, such as at least 8 NMR data sets, such as 16 or more NMR data sets. On the one hand, the more data sets the more accurate or detailed characteristic representation may be obtained. On the other hand, the more data sets the more data processing may be required for generating the characteristic representation of the substance.

[0061] The data of at least one of the NMR data sets may advantageously comprise data from a common or identical pulse sequence. In an embodiment, one or more of the NMR data sets may - even all of the NMR data sets may - comprise intensity data from a common or identical pulse sequence and relaxation data from a different pulse sequence.

[0062] In an embodiment, wherein the data of at least one of the data sets are or comprises data from a common or identical set of pulse sequences comprising one or more of the sequences an inversion-recovery pulse sequence, a Carr-Purcell-Meiboom-Gill (CPMG) sequence, and / or a spin-echo sequence with one or more echoes.

[0063] Optionally the data set(s) of at least one of the NMR data portion(s) comprises data from a common or identical set of pulse sequence.

[0064] The input NMR data may comprise NMR data determined on any NMR active isotope comprising NMR active nuclei, preferably on one or more isotopes that each has a relatively high abundance and / or that comprises an NMR active nucleus having a sensitivity of at least 0.01 %, such as at least 1 % compared to the sensitivity of1H.

[0065] Advantageously, the data of at least one of the data sets are based on data from a common NMR readable isotope, preferably the data set(s) of at least one of the NMR data portion(s) are based on data from a common NMR readable isotope.

[0066] It should be noted that different NMR data portions beneficially may comprise data from different isotopes.

[0067] In an embodiment, the processing of each NMR data portion of two or more NMR data portions, comprises performing the multiple fittings of each of the NMR data portion individually to the formula.

[0068] In an embodiment, the processing of each NMR data portion of two or more NMR data portions, comprises performing the multiple fittings of more than one or even all of the NMR data portions simultaneously and in combination.

[0069] In an embodiment, each NMR data portion is processed separately from the processing of other NMR data portions. The respective processing’s of the respective NMR data portion may preferably comprise fitting to the same formula. In an embodiment, the respective processing’s of the respective NMR data portion may comprise fitting to different formulas. In an embodiment, the processing of each of the NMR data portions, comprises performing the multiple fittings of the one or more data sets of the NMR data portion in a common fitting procedure to the formula.

[0070] The performing of the multiple fittings of the NMR data portion may advantageously comprise applying the at least one start guess for at least one of the fittings of the multiple fittings. Preferably, the performing of the multiple fittings of the NMR data portion comprises applying the at least one start guess for a plurality of the multiple fittings, such as all the fittings of the multiple fittings, wherein the at least one start may be equal or different.

[0071] The respective one or more start guess may comprise a specific value or it may comprise an interval of values. The specific value(s) and / or the interval(s) of values may, e.g., be selected as specific value(s) and / or interval(s) of values with a high probability optionally obtained from previous fittings to the formula.

[0072] In an embodiment, the performing of the multiple fittings of the NMR data portion comprises applying the a first start guess for at least one fitting parameter of a first of the plurality of the multiple fittings and applying subsequent start guesses for the at least one fitting parameter of the subsequent fittings, wherein the subsequent start guesses for the at least one fitting parameter are in the form of an average of fitted results of the at least one fitting parameter of previous fittings of the multiple fittings or in the form of one fitted result of the at least one fitting parameter of a previous fitting of the multiple fittings.

[0073] Advantageously, the start guess comprises a start guess for at least one of the at least 3 fitting parameters for at least one fitting of the multiple fittings. Preferably the start guess comprises a start guess for at least one of the at least 3 fitting parameters for a plurality of the multiple fittings, such as for all of the multiple fittings, wherein the start guess may be identical or differ from a fitting to a subsequent fitting according to a selected pattern.

[0074] Optionally the start guess comprises a start guess for two or more of the at least 3 fitting parameters for at least one fitting of the multiple fittings. This is especially relevant where the start guesses for at least one of the fitting parameters comprises an interval.

[0075] In an embodiment, the at least one start guess comprises at least one of a random start guess or a start guess generated by applying bootstrapping. The Bootstrapping may for example comprise creating simulated data based on an original dataset. The original dataset may, e.g., be a number of previous determined fitting parameters or fitting parameters derived from an estimated interval of probable or possible fitting parameters. Examples of bootstrapping techniques comprises Bayesian bootstrapping and parametric bootstrapping.

[0076] The at least one start guess may be applied to a fitting parameter in a plurality of the fittings, such as to each of the fittings of the multiple fittings to the function. In an embodiment, the start guess applied to the at least one fitting parameter is the same start guess in at least two or more of the multiple fittings to the function, such as the same start guess in each of the multiple fittings to the function. This embodiment may be preferred where there are many fitting parameters, e.g., 4 or more, such as 5 or more fitting parameters to optimize information content / uniqueness relative to required computing resources.

[0077] In an embodiment, where the input NMR data is produced to comprise two or more NMR data portions, the processing of the respective NMR data portions of the plurality of NMR data portions may advantageously comprise applying a randomly selected start guess for at least one of the fitting parameters for each of the NMR data portions.

[0078] In an embodiment, the at least one start guess is characterized by being of Monte Carlo type random start guess.

[0079] In an embodiment, the at least one start guess may be selected to be a start range. For example, the start guess may be within a start range represented by a span of selected values, wherein the selected span of values comprises or consists of typical expected values and / or estimated values. The at least one start guess may comprise a start guess for each of one or more fitting parameters.

[0080] The at least one start guess may conveniently be selected to be a start guess for a fitting parameter which, is correlated to at least one expected fitted parameter or to at least one fitted parameter of the fitted data set from a previous fitting. In an embodiment, the at least one start guess may be selected to be a start guess for a fitting parameter which is correlated to an expected fitted data subset to fitted data subset from a previous fitting.

[0081] Advantageously, the start guess is based on or comprises a start range within a selected span of values, wherein the selected span of values comprises or consist of typical expected values, wherein the typical expected values is determined from typical values of the at least one expected or previous fitted parameter.

[0082] In an embodiment, the at least one start guess is selected to be a start guess for a fitting parameter which is correlated to at least one parameter of the fitted data set subset, wherein the at least one fitted data set parameter is selected from apparent intensity (MO), T1* and T2*, chemical exchange, frequency or diffusion parameters for the species in the sample.

[0083] In an embodiment the at least one start guess is selected to be a start guess for a fitting parameter which is correlated to a derivable parameter which is derivable from at least one parameter of one or more fitted data sets.

[0084] In an embodiment the at least one start guess is correlated to T1 or T2 for the substance.

[0085] The function may be selected in dependence of the complexity of the substance and / or in dependence of how detailed the characteristic representation is desired to be. The complexity of the substance may be determined with respect to number of different components and / or interactions and / or impacts between the components or portions thereof, such as moieties or molecules. Advantageously, the function is a non-linear function.

[0086] In an embodiment, the function comprises one or more exponential terms. In one embodiment the function is a multilinear function comprising two or more exponential terms, such as up to 5 exponential terms, such as 2 or 3 exponential terms. In an embodiment, the function comprises one or more hyperbolic terms.

[0087] In an embodiment, the function comprises one or more trigonometric terms.

[0088] In an embodiment, the function comprises one or more terms selected from a logarithm term, a sinusoidal term, a polynomial term, or a differential term.

[0089] In an embodiment, the function comprises at least one exponential term combined with one or more other terms, such as one or more terms selected from a logarithm term, a sinusoidal term, a polynomial term or a differential term.

[0090] The function may in principle have any number of fitting parameters. To obtain a characteristic representation with a desired degree of detail, the number of filling parameter should advantageously not be too low. To ensure that the required computing resources is not too high and / or that the processing time is not too long, the number of filling parameter should advantageously not be too high. Advantageously, function comprises 3-10 fitting parameters for each spin type, such as 4 fitting parameters, such as 5 fitting parameters, such as 6 fitting parameters, such as 7 fitting parameters, such as 8 fitting parameters or more.

[0091] In an embodiment, the function comprises a function Fl wherein AbBbCbDLand Etare fitting parameters and IRDj, TEk lare input data from the respective NMR data sets;

[0092] / is a type of spin among n different spin types;

[0093] IRDj is an inverse recovery delay applied when obtaining respective j NMR data sets among m different inversion recovery delays;

[0094] TEk.i is an echo time with k representing the ’th length of each echo in a train of echoes from 1 to p and I is the number of echo(es) from 1 to q in the respective echo trains; wherein n, m, p, and q, independently of each is an integer from 1 and up, such as up to 100, such as from 1 to 50, such as from 4 to 25.

[0095] The term “type of spin” or a spin type is herein used to mean a spin, that differs from other spin in at least one parameter, for example a spin, that differs from other spin in at least one of Nuclear Spin Angular Momentum, Spin Angular Momentum Magnitude, chemical shift, spin - spin coupling, spin - spin splitting, spin - lattice relaxation, Spin - spin relaxation, such as spins that differs in at least one of T1*, T2*, intensity M0, frequency (such as resonance frequency, chemical shift, quadrupole coupling, J coupling, or dipole-dipole coupling), chemical exchange, diffusion or other spin parameter(s), such as spin parameters which may be determined by the method of one or more embodiments of the invention.

[0096] In an embodiment, the function is a function Fl:

[0097] As it can be seen this function Fl has 2 exponential terms.

[0098] In an embodiment, the function is a function Fll: wherein AbBbCbDLand Etare fitting parameters and IRDj, TEk land RDhare input data from the respective NMR data sets;

[0099] / is a type of spin among n different spin types;

[0100] I RD; is an inverse recovery delay applied when obtaining respective j NMR data sets among m different inversion recovery delays;

[0101] TE / <, / is an echo time with k representing the ’th length of each echo in a train of echoes from 1 to p and I is the number of echo(es) from 1 to q in the respective echo trains;

[0102] RD / , is a recycle delay during which the magnetization recovers or partially recovers from running one to running subsequent NMR recordings of the h’te recycle delays among 1 to r recycle delays; wherein n, m, p, q, and r, independently of each is an integer from 1 and up, such as up to 100, such as from 1 to 50, such as from 4 to 25.

[0103] The phrase “partially recovers” means advantageously that the magnetization recovers at least 5 %, such as at least 20 %, such as at least 50 %.

[0104] A desired formula, which has shown to be suitable to generate a characteristic representation of a desired high quality is the function Fll as specified in the following:

[0105] - exp (Ci * ?Dh))

[0106] As it can be seen this function Fl I has 3 exponential terms.

[0107] The intensity data of the respective NMR data sets advantageously comprises an echo time and wherein the relaxation data of the respective NMR data sets comprises an inversion recovery delay, preferably the respective NMR data sets further comprise a recycle delay.

[0108] In an embodiment, the parameter C, is correlated to a T1 or T1* of a fitted data subset correlated to echo data representing a spin type / among n spin types.

[0109] The parameter D, may be correlated to a T2 or T2* of a fitted data subset correlated to echo data representing a spin type / among n spin types. One of or both of the parameters A and B, may be correlated to an intensity MO and / or a chemical exchange value of a fitted data subset correlated to echo data representing a spin type / among n spin types.

[0110] Advantageously, each fitted data subset comprises a factor Beta;, wherein Betai=-1 / A factor for spin type / among n spin types of each fitted data set representing chemical exchange.

[0111] In an embodiment, each fitted data subset comprises an MO factor, wherein MOj=- A*Bi factors for spin type / among n spin types of each fitted data set representing an intensity of the spin type.

[0112] The MO factor may be an intensity.

[0113] In an embodiment, T1* is determined according to T1*=-1 / C / factors for spin type / among n spin types of each fitted data set, wherein T1* is representing a T1* of the spin type / .

[0114] In an embodiment, T2* is determined according to T2*i=-1 / Dzfactors for spin type / among n spin types of each fitted data set, wherein T2* is representing a T2* of the spin type / . The Ei factors for spin type / among n spin types of each fitted data set may advantageously represent a diffusion parameter of a molecule, a moiety or an ion comprising the spin type / .

[0115] The number n of spin types is advantageously 2 or larger, such as up to 1000. Advantageously, the number n of spin types is from 3 to 100.

[0116] Advantageously, each fitted data subset represents a characteristic of a spin type, preferably comprising at least a T1*, an associated T2* and an associated intensity (M0). Preferably each fitted data subset represents a characteristic of a spin type further comprising one or more of a diffusion parameter, a frequency shift, a chemical exchange, a J coupling, a quadrupole coupling, or a dipole-dipole coupling.

[0117] In an embodiment, the production of the characteristic representation of the substance from the fitted data sets comprises deriving from each fitted data set, a representative fraction comprising the at least one fitted data subset comprising the T1*, the associated T2* and the associated intensity (M0) to obtain a plurality of representation fractions and assembling the plurality of representation fractions to produce the characteristic representation of the substance.

[0118] The characteristic representation may be continuously improved by adding additional representation fractions to the characteristic representation.

[0119] Thereby the characteristic representation may be produced by being build up over time as the processing of respective NMR data portions or NMR data sets are completed.

[0120] Thereby, when a user is aiming to identify a certain sub-characteristic of a substance, and this sub-characteristic is achieved during a buildup of the characteristic representation, the building up of the characteristic representation may be terminated and the production of the characteristic representation may be deemed to be completed. Thereby, time may be saved which may increase the effectivity of the method.

[0121] In an embodiment, the T1*, the associated T2* and the associated intensity (M0) of the at least one fitted data subset may comprise data representing a characteristic of spin diffusion of molecular or ionic species in inhomogeneous static or radiofrequency fields. In an embodiment, the T1*, the associated T2*, and the associated intensity (MO) of the at least one fitted data subset may comprise data representing characteristic of a chemical exchange involving one, two or preferably multiple spin types, such as 20 or more, such as 50 or more, such as 5 to 100 or even more spin types in the material.

[0122] The characteristic representation may advantageously comprise a probability map of T1* and T2* of the respective spin types. Such probability map may be provided in the form of a visual probability map or in the form of data representing such probability map.

[0123] In an embodiment, the characteristic representation is in the form of or comprises a depiction, preferably illustrating the derived characteristics of some of or all of the n spin types. The depiction may be selected in dependence of the desired use of the method. For example, where the method is used for determine if the substance fulfill a one or more quality criterions, the depiction is selected to show or even argument characteristics related to this or these quality criterion(s).

[0124] It has been found that where the solved Beta; factors (-1 / Aj) of one or more of the fitted data subset differs from an expected Beta factor of x2 (i.e. (expected Beta factor)2), it may be an indicator for that chemical exchange or diffusion is going on similar time scale range as at least one I RDj.

[0125] The expected Beta factor may advantageously be determined empirically, optionally it may be selected to be a value between 1 and 2.

[0126] In an embodiment, wherein the solved Beta; factor(s) (-1 / Aj) of one or more of the fitted data subset differs from an expected Beta factor of x2, it is an indicator for that chemical exchange or diffusion on a time scale within a range of from 25% to + 200% of at least one I RDj.

[0127] The T1* and the associated T2* of a fitted data subset may in an embodiment not represent true T 1 and T2 values for the spin in question, for example because the respective spin of the respective nuclei of components in the substance is under influence of temporarily or permanently adjacent and / or neighboring element and / or moieties of components. However, in an embodiment, the true T1 and T2 of a spin in question may be derived from T1* and T2* values of the spin upon consideration of chemical exchange and / or spin diffusion, e.g., in the process of fitting. In an embodiment, at least one of the fitted data sets comprises data embedding two or more fitted data subsets each comprising a T1*, an associated T2*, and an associated intensity (MO). Advantageously, the production of the characteristic representation of the substance from the at least one fitted data set comprising the two or more fitted data subsets, comprises deriving a representation fraction comprising the two or more fitted data subsets of the T1*, the associated T2* and the associated intensity (MO).

[0128] Advantageously, at least one of the fitted data subsets, each comprising a characteristic of a pair or a group of spin types comprising at least two spin types, and wherein the method comprises deriving a chemical exchange characteristic from the characteristic of the pair or the group of spin types and determining T 1 and T2 (the real T 1 and T2) from the T1* and the T2* and the characteristics of the associated chemical exchange process or processes. Thereby additional information about the substance may be determined.

[0129] Advantageously, each of the fitted data subsets comprises at least one Beta factor and / or at least one diffusion parameter, wherein the least one Beta factor and / or the at least one diffusion parameter comprises a characteristic of one or a pair or a group of spin types. It has shown that a chemical exchange characteristic may be determined from these parameters and advantageously the method comprises deriving such chemical exchange characteristic from the characteristic of the pair or the group of spin types and / or diffusion characteristics for at least one spin type.

[0130] The input NMR data may comprise NMR data from any NMR active isotope.

[0131] In an embodiment, the NMR data portion(s) comprises data from at least one NMR readable isotope. Preferably, the NMR data from each of the one or more of the data portions is single isotope NMR data comprising NMR data from one isotope.

[0132] In an embodiment, the one or more NMR data portions comprises data from at least two isotopes. Preferably the method comprises obtaining at least a plurality of NMR data portions, comprising at least a first data portion comprises NMR data from a first isotope and a second NMR data portion comprising data from a second isotope. The plurality of NMR data portions advantageously comprises at least a number of NMR data portions comprising NMR data from respective isotope. The number of NMR data portions may conveniently be up to 50, such as up to 10, such as from 2 to 5. To obtain a high quality characteristic representation, it is desired that the respective NMR data portions, each comprising NMR data from one isotope only are individually fitted to the function.

[0133] Examples of suitable isotopes for the at least one isotope is selected from the group of isotopes1H,2H,6Li,7Li,10B,11B,13C,14N,15N,17O,19F23Na,27AI,29Si31P,33S,35CI,37CI , and39K,41K,43Ca,47Ti,49Ti,50V,51V,53Cr,55Mn,57Fe,59Co,61Ni,63Cu,65Cu,67Zn,69Ga,71Ga,75As,77Se,79Br,81Br,83Kr,85Rb,87Rb,87Sr,89Y,91Zr,93Nb,95Mo,97Mo,105Pd,107Ag,109Ag,111Cd,113Cd,117Sn,119Sn,115Sn,121Sb,135Ba,137Ba177Pb,199Hg,201Hg, and207Pb.

[0134] The first and second or the respective isotopes may conveniently comprise at least two isotopes selected from the group of isotopes1H,2H,6Li,7Li,10B,11B,13C,14N,15N,17O,19F23Na,27AI,29Si31P,33S,35CI,37CI , and39K,41K,43Ca,47Ti,49Ti,50V,51V,53Cr,55Mn,57Fe,59Co,61Ni,63Cu,65Cu,67Zn,69Ga,71Ga,75As,77Se,79Br,81Br,83Kr,85Rb,87Rb,87Sr,89Y,91Zr,93Nb,95Mo,97Mo,105Pd,107Ag,109Ag,111Cd,113Cd,117Sn,119Sn,115Sn,121Sb,135Ba,137Ba177Pb,199Hg,201Hg, and207Pb.

[0135] The one or more isotopes applied are preferably selected to comprise one or more isotopes of expected to be present in the substance, preferably comprising one or more isotopes with a high presence in the substance. Example of preferred isotope comprises1H,13C,14N,15N,17O,19F23Na,39K,41K,35CI,37CI, and43Ca.

[0136] In an embodiment, the processing of each NMR data portion(s) comprises performing multiple fittings to a two or more different functions, where the two or more different functions may be as described above.

[0137] In an embodiment, the method further comprises fitting two or more of the determined fitted data sets to a Bloch McConnell equation, such as a Bloch McConnell equation for two-site exchange or a Bloch McConnell equation for three- site exchange. The fitted data sets applied in the fitting to the Bloch McConnell equation comprises each T1 , T2, MO.

[0138] The fitted data sets are advantageously fitted to the Bloch McConnell equation in one or more groups of two or more, thereby additional information about chemical exchange and parametrize this in terms of kinetic and / or exchange constants or rates. In an embodiment, the fitted data sets are fitted to the Bloch McConnell equation in groups of two fitted data sets representing respective spins, and at the method comprises determining least two kinetic constants representing the chemical exchange between the respective spins.

[0139] Fitting two or more of the determined fitted data sets to a Bloch McConnell equation is in particular beneficial where the solved Beta; factor(s) (-1 ZAj) of one or more of the fitted data subsets differ from an expected Beta factor of x2.

[0140] Bloch McConnell equations are known in the art and was developed by Harden M McConnell based on the Bloch equations as described in “Reaction Rates by Nuclear Magnetic Resonance” by Harden M. McConnell, The Journal of Chemical Physics 28, 430 (1958); doi: 10.1063 / 1.1744152.

[0141] The recording of the NMR data may be performed by any methods, such as the methods known in the art. Advantageously, the step of obtaining the one or more portions of NMR data comprises performing at least one NMR reading using an NMR spectrometer, such as a low field NMR spectrometer.

[0142] The low-field NMR spectrometer may advantageously be configured for generating a static magnetic field, such as a low field of from about 0.1 to about 5 tesla. It has been found to be very beneficial using a low-field NMR spectrometer with a maximal magnetic field about 5 Tesla, preferably about 3 Tesla or less, such as about 3 tesla or less.

[0143] The recording of the NMR data advantageously comprises performing at least one reading, such as preferably a plurality of readings.

[0144] Where there are several readings, the readings may be performed on a common sample of the substance or on different samples of the substance.

[0145] In an embodiment, the at least one NMR reading comprises subjecting a sample of the substance to a sequence of RF exiting pulses and reading relaxation / dephasing signals of at least one spin of at least one isotope.

[0146] The RF exiting pulses may comprise any RF irradiation exiting the nuclei of the isotope to be read. Advantageously, the at least one NMR reading comprises subjecting a sample of the substance to one or more sequences of RF irradiation comprising frequencies and / or periods suitable for determining T 1 and T2 relaxation and / or J coupling evolution and / or dipole-dipole coupling evolution and / or quadrupolar coupling evolution and / or evolution depending on chemical exchange and / or diffusion.

[0147] In an embodiment, the at least one NMR reading comprises subjecting a sample of the substance to one or more sequences of RF exciting pulses comprising at least one of the pulse sequences DEPT (Distortionless Enhancement by Polarization Transfer), DEPTQ (DEPT with retention of quaternary carbons), HSQC (Heteronuclear Single Quantum Coherence), INEPT (Insensitive Nuclei Enhanced by Polarization Transfer), BIRD (Bilinear Rotation Decoupling pulses), TANGO (Testing for Adjacent Nuclei with a Gyration Operator) or NOE (Nuclear Overhauser Effect).

[0148] The Nuclear Overhauser Effect (NOE) is the transfer of nuclear spin polarization from one nuclear spin population to another via cross-relaxation. It is a common phenomenon observed by nuclear magnetic resonance (NMR) spectroscopy. Nuclear Overhauser Effect can for example be used to determine intra- (and even inter-) molecular distances. The NOE effect is the change in population of one proton (or other nucleus) when another magnetic nucleus close in space is saturated by decoupling or by a selective 180 degree pulse. Preferably the pulses used are soft pulses.

[0149] Further information about DEPT and NOE and how to perform them can e.g. be found in “New Method for NMR Signal Enhancement by Polarization Transfer, and Attached Nucleus Testing” by John Homer et al. J. Chem. Soc.,Chem. Commun., 1994 and “Fundamentals of NMR” Chapter 1, by Thomas L. James, Department of Pharmaceutical Chemistry, University of California, 1998 (http: / / qudev.ethz.ch / content / courses / phys4 / studentspresentations / n r / James Fund a entals of .. .R.pdf)

[0150] The at least one NMR reading may conveniently be subjecting a sample of the substance to a sequence of RF exiting pulses comprising a plurality of pulses comprising at least one 180° RF pulse and at least one 90° RF pulse. In an embodiment, the at least one NMR reading comprises subjecting a sample of the substance to a sequence of RF exiting pulses comprising a Hahn echo sequence or a CPMG sequence or a combination comprising at least one of these.

[0151] Advantageously, the at least one NMR reading comprises subjecting a sample of the substance to a sequence of RF exiting pulses comprising a spin echo sequence, a fast spin echo sequence, a stimulated echo sequence, a gradient echo sequence or any combinations comprising at least one of these e.g. as described above

[0152] In an embodiment, the at least one NMR reading comprises subjecting a sample of the substance to one or more sequences of RF exiting pulses comprising one or more of the sequences an inversion-recovery pulse sequence, a Carr-Purcell- Meiboom-Gill (CPMG) sequence and / or a spin-echo sequence or any combinations comprising at least one of these.

[0153] Preferably, the at least one NMR reading comprises subjecting a sample of the substance to a sequence of RF pulses and delays comprising a sub sequence SE1

[0154] • applying a first 180° RF pulse

[0155] • applying a first inversion recovery delay (I RD)

[0156] • applying a first 90° RF pulse and

[0157] • determine a resonance intensity, directly or throughout an echo sequence, such as CPMG.

[0158] The SE1 subsequence may be repeated optionally applying different IRD values. In addition, it may comprise determining a null IRD value where the determined resonance intensity passes through a point of zero intensity.

[0159] This delay time Tnuii is related to T1 in that Tnuii is 0.71 times T 1.

[0160] In an embodiment, the at least one NMR reading comprises subjecting a sample of the substance to a sequence of RF pulses and delays comprising a sub sequence SE1 i. applying a 90° RF ii. applying an echo delay iii. applying a 180° RF pulse iv. repeating steps ii and iii N times v. applying a recycle delay,

[0161] The SE2 subsequence may be repeated e.g. applying different recycle delays. Advantageously, the at least one NMR reading comprises subjecting the sample of the substance to a repeated sequence of RF exiting signals comprising

[0162] _n_riNV_ ^ / 2 _ ^E / 2 _ 7T _rE _71_ T’w _ invert IRD excite 1stED 1strefocus ED refocus wait RD repeat x (N-l) wherein and / 2 designates radio-frequency pulses with pulse duration and radiofrequency pulse field strengths adjust to provide pulse flip angles of (such as 180°) and K / 2 (such as 90°), respectively, Tim is the inversion recovery delay (IRD), TE is the echo delay (ED), Tn / the recycle delay (RD), and N the number of echoes.

[0163] The SE2 pulse sequence may in an embodiment involve two TE - K elements repeated A / -1 and A / ’-1 times and separated by a delay period.

[0164] Thereby, highly useful input NMR data may be obtained suitable for generating a high quality characteristic representation.

[0165] The pulse sequence may conveniently be repeated a plurality of time with different composition of one more of the parameters IRD, ED, N, RD, radio-frequency pulse field strengths, or radio-frequency pulse durations to improve the NMR data and thereby the portion(s) of NMR data forming part of the input NMR data.

[0166] The NMR data for the production the NMR portion(s) comprising different isotopes may for example be obtained by recording signals from one isotope during one or more of the delay periods of a recording from another isotope. This is herein referred to as interleaved NMR recording.

[0167] In an embodiment, the pulse sequence is repeated a plurality of time with different composition and number of elements of inversion, waiting, and echo periods.

[0168] Advantageously, the NMR recording comprises reading at least according to any of the preceding claims, wherein the data sets are produced to comprise data from at least two different NMR readable isotopes. Preferably one of the NMR readable isotope being1H. The at least two different NMR readable isotopes may preferably be read using interleaved NMR recording. In an embodiment, the NMR reading comprises reading at least two different NMR readable isotopes, wherein one of the NMR readable isotope being13C.

[0169] In an embodiment, the NMR reading comprises reading multiple NMR readable isotopes in parallel, comprising initiating a further pulse sequence for another NMR readable isotope in an RD, IRD, or ED period of a previous NMR readable isotope pulse sequence, i.e. , applying interleaved NMR recording.

[0170] The NMR recording may advantageously comprise performing one or more additional readings over time, wherein an additional NMR data portion from each additional reading may be obtained. The method may preferably comprise processing the additional NMR data portion to obtain an additional representation fraction and comparing and / or adding the additional representation fraction to the characteristic representation.

[0171] Thereby valuable information of potential changes in the sample may be obtained.

[0172] This may for example be applied to determine the stability of a substance, a development of a substance, a change of a substance in a production line, a reactivity of a substance, or similar.

[0173] The information of changes in the substance may for example comprise a change in physical condition, such as temperature, hydration, pressure, salt content, pH, solvent type, phase, or concentration of components / elements in the substance.

[0174] In an embodiment, the fitting of a plurality of portions of data for the sample, may improve the measuring accuracy in terms of standard deviation for one or more of the parameters individually and / or reduces a required data measurement time for a selected measurement accuracy.

[0175] The substance may in principle be any kind of substance comprising active NMR isotope(s). The substance may be solid, liquid, gaseous, or any mixture thereof. Advantageously, the substance comprises at least two components, preferably each component comprises a unique NMR readable isotope. The method of the invention and embodiments thereof have shown to be very beneficial for analyzing of complex substances, such as substances comprising several different components, e.g., as described above.

[0176] Advantageously, the substance is a substance comprising two or more components and preferably, such as a food product, such as a sample in any stage of wastewater treatment, such as an animal slurry or manure, such as a livestock feed, such as a biogas digester slurry, such as one or more samples at any stage of biorefinery, such as one or more samples in any stage of mining or mineral refining processes.

[0177] In an embodiment, the substance comprises at least an organic and / or an inorganic component(s). The substance may for example comprises monomers that is capable of polymerize and the method may comprise performing one or more additional recordings over time and for each additional recording producing an additional NMR data portion, processing the additional NMR data portion to obtain an additional representation fraction and determining the stage of and / or quality of polymerization.

[0178] In an embodiment, the substance comprises at least an organic and / or an inorganic component(s). The substance may for example comprise chemical or biological reactants that are capable of interacting in chemical or biological synthesis and the method may comprise performing one or more additional recordings over time and for each additional recording producing an additional NMR data portion, processing the additional NMR data portion to obtain an additional representation fraction and determining the stage of and / or quality of the chemical or biological reaction.

[0179] In an embodiment, the method comprises preparing a plot of the characteristic representation of the substance. The plot of the characteristic representation of the substance may for example comprise a probability map of T1* and T2* of the respective spin types as described above and / or a depiction as described above.

[0180] Advantageously, the method comprises determine a quality parameter of the substance comprising deriving data from the characteristic representation and determining the quality parameter and optionally determine if the quality parameter fulfil a quality criterion, e.g., a quality criterion comprising a selected threshold.

[0181] The quality parameter of the substance may for example be associated with an amount of a component, such as the amount of an organic component, such as the amount of lipid, such as an amount of fat, such as the amount of carbohydrate, such as the amount of protein, such as an amount of nucleic acid, such as an amount of chemical reagent, such as an amount of an isotope, such as an amount of mobile sample components, such as an amount of rigid / immobile sample components, such as an amount of species undergoing chemical exchange, such as an amount of species undergoing molecular ionic, or spin diffusion.

[0182] In an embodiment, the quality parameter of the substance is associated with a chemical property of the substance, such as a pH value.

[0183] The characteristic representation may comprise a plurality of identifiers, wherein each identifier representing a statistical distribution of the sample components.

[0184] The method may conveniently comprise comparing the characteristic representation of the substance with at least one historical characteristic representation, to identify a difference, such as a chemical and / or biological and / or a compositional difference(s) between the substance and the substantially similar substance.

[0185] The method may for example comprise determining a quality parameter based on the identified difference.

[0186] The at least one historical characteristic representation may for example comprise at a historical characteristic representation of a reference substance or a historical characteristic representation synthesized from two or more historical characteristic representation of a reference substance.

[0187] The characteristic representation of the reference substance may for example be applied to define a quality criterion, e.g., comprising a threshold. A generated characteristic representation of a substance may be compared with such characteristic representation of the reference substance, e.g., to determine is a quality criterion is fulfilled.

[0188] The least on reference substance may for example correspond to the substance, in that it comprises at least 25 % of the components present in the substance, such as at least 50 % of the components present in the substance, such as at least 90 % of the components present in the substance.

[0189] In an embodiment, the characteristic representation is a visual characteristic representation, such as a visual probability map or in the form of data representing such probability map, such as a depiction, the method advantageously comprises determine at least one quality parameter of the substance by comparing the visual characteristic representation with a visual characteristic representation of a reference substance or a historical visual characteristic representation, wherein the at least one quality parameter comprises a similarity parameter, such as a similarity parameter comprising a Bhattacharyya Distance. In an embodiment the method comprises classifying the produced visual characteristic representation according to the determined similarity parameter.

[0190] In an embodiment, the processing may be performed by a trained computer wherein the trained computer is trained to determine one or more quality parameters and one or more sample identifiers, for example to determine quantitative parameters and / or to monitoring differences in historical data such as in chemical, physical or biological processes for example in a process monitoring setup. Preferably the trained computer is trained to classify a substance based on a characteristic representation generated by an embodiment of the method on that substance according to one or more quality parameters and / or according to one or more sample identifiers.

[0191] The invention also comprises a method for generating a trained computer. The method comprises subjecting the computer to a machine learning process comprising feeding data to the computer, wherein the data comprises data packets of previous obtained NMR data correlated to at least one fitted data subset comprising a T1*, an associated T2*, and an associated intensity (MO). The method for generating the trained may comprise feeding data packets comprising data representing a characteristic representation of a substance and the NMR data portions applied for generating the characteristic representation, e.g., as described above.

[0192] In an embodiment, the processing of the NMR data portion(s) comprise(s) applying artificial intelligence, machine learning, reinforcement learning, or similar in supervised, unsupervised or any other form.

[0193] The invention also comprises a trained computer obtainable by the method as described above.

[0194] The invention further comprises an NMR system for performing the method as escribed above, The NMR system comprises a computer system which is configured for receiving one or more input NMR data, wherein the input NMR data comprises a plurality of NMR data sets each comprising intensity data and relaxation data, and wherein the computer system is configured for

[0195] • processing the NMR data portion(s) into a multidimensional distribution of fitted data sets by performing multiple fittings to a function, wherein the function comprises at least 2 fitting parameter(s) and is a function which represents a model of the obtained NMR data portion(s), and

[0196] • generating the characteristic representation of the substance from the fitted data sets.

[0197] The NMR system may advantageously be configured for performing the data processing as described above.

[0198] The computer system may be in the form of a single computer or it may comprise separate computer portions adapted for communication by wire or wireless data communication, preferably via the Internet and / or via Bluetooth, or similar means. The computer system may comprise a persona compute a tablet, a smart prone and / or any integrated or external memory devices. Advantageously, the NMR system comprises an NMR spectrometer, preferably the NMR spectrometer is in data communication with the computer system and optionally comprises a computer forming part of the computer system.

[0199] In an embodiment, the computer system comprises a super computer system, such as a computer capable of performing at least 1014floating-points operations per second (FLOPS), such as a computer capable of performing at least 1015FLOPS, such as a computer capable of performing at least 1016FLOPS.

[0200] In an embodiment, the computer system comprises a cluster of computers, computer systems with multiple CPU’s, computer systems with multiple GPU’s, such as 1000 or more GPUs, optionally located at a remotely location relative to the NMR spectrometer.

[0201] Advantageously, the computer system comprises the trained computer as described above.

[0202] BRIEF DESCRIPTIONS OF EMBODIMENTS AND EXAMPLES

[0203] In the following the invention will be further illustrated by the description of a number of illustrative and non-limiting embodiments and examples of the present invention, with reference to the appended drawings.

[0204] The figures are schematic and are not drawn to scale and may be simplified for clarity. Throughout, the same reference numerals are used for identical or corresponding parts.

[0205] Figure la, figure lb, and figure lc illustrate visual characteristic representations obtained in example 1.

[0206] Figure 2 is a process diagram illustrating an embodiment of the method of the invention comprising generating a characteristic representation of at least two NMR data portions.

[0207] Figure 3 is a process diagram illustrating an embodiment of the method of the invention comprising determining if a change has occurred in a substance. Figure 4 is a process diagram illustrates an embodiment of the method of the invention comprising determining if a sample substance fulfills a quality criterion.

[0208] Figures 5a, 5a', 5b and 5b' are visual characteristic representations associated to example 2.

[0209] Figures 6a, 6a', 6b and 6b' are visual characteristic representations associated to example 3.

[0210] Figures 7a, 7b, 7c, 7d, 7e are visual characteristic representations associated to example 4.

[0211] Figures 8a, 8a', 8b and 8b' are visual characteristic representation associated to example 5.

[0212] The process comprising generating a characteristic representation of at least two NMR data portions of a substance.

[0213] The substance may be as described above.

[0214] In step 20a a first NMR data portion is obtained from a first isotope, for example one of the above described active isotopes. The first NMR data portion comprises a plurality of NMR data sets each comprising intensity data and relaxation data. In step 20b the first NMR data portion is processed into a multidimensional distribution of fitted data sets by performing multiple fittings to a function X1 to obtain fitted data sets. In step 20c, first fitted data subsets are derived from the fitted data sets. In this embodiment the first fitted data subsets comprise associated T1*, T2* M0, and Beta factor. In variations thereof the Bata factor may be omitted and / or each fitted data subset may comprises additional fittings with functions and associated parameters that provide more detailed information about characteristics such as chemical exchange and / or diffusion as described above. In step 20d a first representation fraction is generated from the fitted data subsets.

[0215] In step 21a a second NMR data portion is obtained from a second isotope, for example one of the above described active isotopes. The second NMR data portion comprises a plurality of NMR data sets each comprising intensity data and relaxation data. In step 21b the second NMR data portion is processed into a multidimensional distribution of fitted data sets by performing multiple fittings to a function X2 to obtain fitted data sets. In step 21c, second fitted data subsets are derived from the fitted data sets. In this embodiment the first fitted data subsets comprise associated T1*, T2*, MO, and Beta factor, for spin types with Beta factors differing from a factor x2 potentially supplemented with fittings to other functions and parameters, e.g. of the Bloch McDonnell type, to characterize in more details characteristics of chemical exchange and / or spin diffusion as described above. In step 20d a second representation fraction is generated from the fitted data subsets.

[0216] In step 22, the characteristic representation is generated by combining the first representation fraction and the second representation fraction. It should be observed that each of the first representation fraction and the second representation fraction also represents a characteristic representation. In an embodiment one or more further representation fractions are generated in a similar way as the production of the first representation fraction and the second representation fraction and the one or more further representation fraction may be added to the characteristic representation.

[0217] The formulas X1 and X2 may be equal, or they may differ from each other. The formulas X1 and X2 may independently of each other be as described above.

[0218] In step 23, the characteristic representation is generated to comprise a probability map of T 1* and T2*, e.g., as described above.

[0219] In step 24, the characteristic representation may be generated to comprise a depiction illustrating at least a portion of the first and / or second fitted data subsets, e.g., as described above or as shown in figure 1c.

[0220] The process diagram of figure 3 illustrates an embodiment comprising determining if a change has occurred in a substance.

[0221] In step 30a one or more NMR data portions are obtained at a first point of time of the substance. The NMR data portions may comprise NMR data from one or more active isotopes. In step 30b, each NMR data portion is processed into a multidimensional distribution of fitted data sets by performing multiple fittings to a function X3 to obtain fitted data sets. In step 30c, fitted data subsets are derived from the fitted data sets. In this embodiment the first fitted data subsets comprises associated T1*, T2*, MO, and Beta factor, for spin types with Beta factors differing from a factor x2 potentially supplemented with fittings to other functions and parameters to characterize in more details characteristics of chemical exchange and / or spin diffusion as described above. In step 30d a characteristic representation is generated from the fitted data subsets.

[0222] At as second point of time the process is repeated.

[0223] In step 31a one or more NMR data portions are obtained at the second point of time of the substance. The NMR data portions may preferably comprise NMR data from the same one or more active isotopes as the NMR data portions obtained at the first point of time. In step 31b, each NMR data portion is processed into a multidimensional distribution of fitted data sets by performing multiple fittings to the function X3 to obtain fitted data sets. In step 31c, fitted data subsets are derived from the fitted data sets.

[0224] In step 32 the characteristic representation of step 31 d is compared with the characteristic representation of step 30d and it may be determined if a change has occurred in the period from the first point of time to the second point of time.

[0225] This embodiment illustrated in the process diagram of figure 3, may be highly suitable for determining the stability of a substance and / or a reactivity of a substance. The embodiment may for example be applied for determining a drugs activity when mixed with a biological fluid and / or to detect a potential allergic reaction, where blood samples taken at first, second and optionally subsequent points of time are acquired and analyzed by generating a characteristic representation. The second point of time may be from a second to months or even years after the first point of time, depending on the potential change that is analyzed for.

[0226] The process diagram of figure 4 illustrates an embodiment of the method comprising determining if a sample substance fulfills a quality criterion.

[0227] The process comprises generating a reference characteristic representation based on one or more reference substances and thereafter generating a characteristic representation of a sample substance and determining if the sample substance fulfills a quality criterion.

[0228] In step 40a one or more NMR data portions are obtained of the reference substance. The NMR data portions may comprise NMR data from one or more active isotopes. In step 40b, each NMR data portion is processed into a multidimensional distribution of fitted data sets by performing multiple fittings to a function X4 to obtain fitted data sets. In step 40c, fitted data subsets are derived from the fitted data sets. In this embodiment the first fitted data subsets comprise associated T1*, T2*, M0, and Beta factor. This may be supplemented with additional fittings to obtain more detailed characteristics of chemical exchange or diffusion phenomena. In step 40d a reference characteristic representation is generated from the fitted data subsets.

[0229] In step 41a one or more NMR data portions are obtained of a sample substance. The NMR data portions preferably comprise NMR data from the same one or more active isotopes as applied for the reference substance. In step 41b, each NMR data portion is processed into a multidimensional distribution of fitted data sets by performing multiple fittings to a function X4 to obtain fitted data sets. In step 41c, fitted data subsets are derived from the fitted data sets. The first fitted data subsets comprise same or similar parameters as the fitted data subsets obtained for the reference substance. In step 41 d a characteristic representation is generated from the fitted data subsets.

[0230] In step 42, the characteristic representation of step 41 is compared with the reference characteristic representation of step 40d. Based on this it may be determined if the sample substance fulfill a quality criterion.

[0231] This embodiment illustrated in the process diagram of figure 4, may be highly suitable for determining if a substance fulfill a quality criterion, for example a quality criterion associated to a content of one or more components, to a quality criterion associated to an interaction between two or more components, to a quality criterion associated to chemical exchange between two or more components, to a quality criterion associated to diffusion of components, to a quality criterion associated to a homogeneity of the substance, to a quality criterion associated to potential presence of impurities or any other chemical, physical, or biological criterions. In the below examples the NMR instrument Tveskaeg™, marketed by NanoNord A / S was used.

[0232] EXAMPLE 1

[0233] In this example the substance was a commercial salted butter based on cow milk.

[0234] A package of the butter was provided.

[0235] A portion of the sample was subject to an 1H NMR analysis obtaining a portion of NMR data of an isotope.

[0236] The NMR data was recorded using a series of CPMG and inversion recovery CPMG experiment with 18 (m) inversion recovery delays distributed logarithmically between 0 and 3 s (i.e., in addition to a standard CPMG experiment, inversion recovery CPMG experiments were performed with 18 different inversion recovery delays (IRD) of 0, 0.001, 0.00164938, 0.00272045, 0.00448704, 0.00740083, 0.0122068, 0.0201336, 0.0332078, 0.0547723, 0.0903401, 0.149005, 0.245765, 0.40536, 0.668592, 1.10276, 1.81887, and 3 s) and the free induction decay recorded in-between a series of 1000 (N) echoes in a CPMG sequence separated by 200 |_is (TE); in each 200 p.s period 101 NMR data points were sampled equidistantly. The experiment used 9 p.s 90° (K / 2) RF pulses and 18 |_is 180° (K) RF refocus pulses, a recycle delay of 4 s (TW), and 2 scans (standard phase cycle to reduce instrumental imperfections), a static magnetic field about 0.9 T, and a sample temperature of about 39 °C.

[0237] Prior to recording the NMR data the NMR spectrometer was in a standard calibrated state with calibrated radiofrequency pulses and the intensity of a CPMG experiment calibrated using a calibrated state with calibrated radiofrequency pulses and the intensity of a standard calibration sample. This is referred to as one-time factory calibration.

[0238] An NMR data portion was produced from the recorded data. The recorded NMR data portion was organized in NMR data sets, each comprising intensity data and relaxation data (T1 and T2). The NMR data portion were processed into a multidimensional distribution of fitted data sets by performing multiple fittings to the function FI as described above: wherein AbBbCbDLand Etare fitting parameters and IRDj, TEk lare input data from the respective NMR data sets;

[0239] / is a type of spin among 4 different spin types;

[0240] IRDj is the inverse recovery delay applied when obtaining respective j NMR data sets among the m different inversion recovery delay of 0-3 s.;

[0241] TEk.i is the echo time of with k representing the ’th length of each echo in a train of echoes from 1 to p and I is the number of echo(es) from 1 to q in the respective echo trains.

[0242] In this example, n = 4 spin types and 1000 random starting guesses on MOi, Tl*i, and T2*i were applied in the fitting process.

[0243] The fitted datasets, were further analyzed and fitted data subset were derived, wherein each fitted data subset comprised associated values of T1*, T2, an intensity (M0), and a Beta factor, wherein the values of the respective data subsets are associated by representing a common spin type and being derived from a common data subset.

[0244] The Tl*, and T2* values were mapped to obtain a characteristic representation in the form of a T1-T2 map as illustrated in figure la. The map in figure la illustrates all possible solutions obtained by the fitted datasets, wherein each point (or circle) in the T1-T2 map represents a set of Tl* and T2* values of a data subset for a spin type; the axis are shown on a linear scale in seconds. In this case the Beta factors does not vary appreciably from the x2 factor, implying that the Tl* and T2* values measured are not influenced noticeably by chemical exchange, while effects from field inhomogeneity may influence values Figure lb shows a probability map obtained by adding intensities from the solutions in figure la, thereby obtaining the statistical probability of distinguishable components in the Tl*-T2* map. In this probability map the axes are on a logarithmic scale to reveal further characteristic information. From the probability map of figure lb, two major regions W and F as encircled by the dotted lines are identified. The region W and the region F are distinguished by their difference in Tl* and T2* values. It was found that the region W represents water and the region F represents fat.

[0245] From the probability map of figure lb and the regions W and F is was determined that the butter sample had a water content of 19.2 wt%, whereas the package from where the butter was taken declared a water content of 15 wt%.

[0246] Figure lc illustrated a characteristic representation in the form of a depiction. The depiction may be provided e.g. as described above. The shown depiction of figure lc is based on the probability map of figure lb to provide a clear illustration of the water versus fat content. By generating the characteristic representation as a depiction it may be simpler for a user, such as an operator or such as a computer in a machine learning setup to relatively fast determine if the substance on which the characteristic representation is generated fulfill a selected criterion.

[0247] EXAMPLE 2

[0248] In this example, a number of food samples in the form of samples of two different creams with different fat content were analyzed.

[0249] Samples of creams with respectively a declared fat content of 8 wt% (low-fat cream) and 37 wt% (high-fat cream) were investigated. Portions of the respective samples were subject to NMR analysis. 1H NMR data were recorded using a series of CPMG and inversion recovery CPMG experiments with 18 (m) inversion recovery delays distributed logarithmically between 0 and 3 s (i.e., in addition to a standard CPMG experiment, inversion recovery CPMG experiments with inversion recovery delays (Ti / v0 of 0.001, 0.00164938, 0.00272045, 0.00448704, 0.00740083, 0.0122068, 0.0201336, 0.0332078, 0.0547723, 0.0903401, 0.149005, 0.245765, 0.40536, 0.668592, 1.10276, 1.81887, 3 s) and the free induction decay recorded in-between a series of 1000 ( / V) echoes in a CPMG sequence separated by 200 |_is (TF); in each 200 p.s period 101 points were sampled equidistantly. The experiment used 9 |_is 90° (K / 2) and 18 |_is 180° (K) pulses, a repeat delay of 4 s (Tw), and 2 scans, a static magnetic field about 1 T, and a sample temperature of about 39 °C.

[0250] Prior to recording the NMR data the NMR instrument were in a standard calibrated state with calibrated radiofrequency pulses and the intensity of a CPMG experiment calibrated using an standard calibration sample. This is referred to as one-time factory calibration.

[0251] A NMR data portion was produced from the recorded NMR data, wherein the NMR data portion comprised a plurality of data sets, each comprising inverse recovery delay (I RD) and associated echo time (ED).

[0252] The NMR data portion were analyzed using the same function as used in example 1 using n = 4 spin types and 1000 random starting guesses on M0i, Tl*j, and T2*i. This lead to the visual characteristic representation in the form of Tl*-T2* maps illustrated in Figure 5a for the low-fat cream (a) and Figure 5b for the high-fat cream.

[0253] For each of respectively the sample of low-fat cream (Figure 5a) and the sample of high-fat cream (Figure 5b), the left map of respectively figure 5a and figure 5b shows all possible solutions with each point representing Tl* and T2* values for a spin type; the axes are shown on a linear scale in seconds. The map to the right of respectively figure 5a and Figure 5b shows a probability map with the axes on a logarithmic scale. For the two cream samples, we have marked the fat region by a box marked F and the water region with a box marked W, clearly showing that the fat / water ratio is substantially higher for the low-fat sample relative to the high-fat sample. The map to the left of respectively figure 5a and figure 5b indicates that some, but not all, of the water is experiencing exchange through beta values larger than 2 (marked in grey). From the data we deduce 9.3 wt% and 36.2 wt% fat conforming well with the known values 8 wt% and 37 wt%, respectively.

[0254] Figures 5a' and 5b' show the maps corresponding to Figures 5a and 5b in color.

[0255] EXAMPLE 3 In this example, a number of food samples in the form of samples of two different vegetable oils were analyzed.

[0256] A sample of coconut oil and a sample of rapeseed oil were subject to NMR analysis. 1H NMR data were recorded using a series of CPMG and inversion recovery CPMG experiments with 18 (m) inversion recovery delays distributed logarithmically between 0 and 3 s (i.e., in addition to a standard CPMG experiment, inversion recovery CPMG experiments with inversion recovery delays (Tm / v) of 0.001, 0.00164938, 0.00272045, 0.00448704, 0.00740083, 0.0122068, 0.0201336, 0.0332078, 0.0547723, 0.0903401, 0.149005, 0.245765, 0.40536, 0.668592, 1.10276, 1.81887, 3 s) and the free induction decay recorded in-between a series of 1000 ( / V) echoes in a CPMG sequence separated by 200 |_is (TE); in each 200 |_is period 101 points were sampled equidistantly. The experiment used 9 |_is 90° (K / 2) and 18 |_is 180° (K) pulses, a repeat delay of 4 s (Tw), and 2 scans, a static magnetic field about 1 T, and a sample temperature of about 39 °C.

[0257] Prior to recording data the NMR instrument were in a standard calibrated state with calibrated radiofrequency pulses and the intensity of an CPMG experiment calibrated using an standard calibration sample. This is referred to as one-time factory calibration.

[0258] A NMR data portion was produced from the recorded NMR data, wherein the NMR data portion comprised a plurality of data sets, each comprising inverse recovery delay (I RD) and associated echo time (ED).

[0259] The NMR data portion were analyzed using the same function as used in example 1 using n = 4 spin types and 1000 random starting guesses on M0i, Tl*j, and T2*i. This lead to the visual characteristic representation in the form of the Tl*-T2* maps illustrated in figure 6a for coconut oil, and figure 6b rapeseed oil.

[0260] For each or the respective samples of coconut oil (Figure 6a) and rapeseed oil (Figure 6b), The map to the right of respectively figure 6a and Figure 6b shows all possible solutions with each point representing Tl* and T2* values for a spin type; the axes are shown on a linear scale in seconds. The map to the right of respectively figure 6a and Figure 6b shows a probability map with the axes on a logarithmic scale. For the two vegetable oils the Tl*-T2* maps reflects the lipid composition (chain length) and mobility of the samples. Thereby the maps provide detailed information which may be used for identification and monitoring of and variations in a production line which may e.g. be used in a quality control. Both samples contain a variety of different lipids giving rise to the ridge like patterns observed in the visual characteristic representations of figure 6a and 6b. The coconut oil is characterized by having shorter lipid chain lengths than rapeseed oil.

[0261] Figures 6a' and 6b' show the maps corresponding to the left map of respectively figures 6a and 6b in color.

[0262] EXAMPLE 4

[0263] In this example substances comprising various flour-containing compounds, including Teflon® and environmentally challenging PFAS / PFOS molecules were analyzed.

[0264] Samples of the various substances were provided. A portion of each sample was subject to an NMR analysis. 19F NMR data were recorded using a series of CPMG and inversion recovery CPMG experiments with 18 (m) inversion recovery delays distributed logarithmically between 0 and 9 s (i.e., in addition to a standard CPMG experiment, inversion recovery CPMG experiments with inversion recovery delays (Ti / v0 of 0.001, 0.00176661, 0.0031209, 0.00551341, 0.00974004, 0.0172068, 0.0303977, 0.0537008, 0.0948683, 0.167595, 0.296075, 0.523048, 0.92402, 1.63238, 2.88378, 5.09451, 9 s) and the free induction decay recorded in-between a series of 1000 ( / V) echoes in a CPMG sequence separated by 200 |_is (TE); in each 200 p.s period 101 points were sampled equidistantly. The experiment used 9 |_is 90° (K / 2) and 18 |_is 180° (K) pulses, a repeat delay of 12 s (Tw), and 2 scans, a static magnetic field about 1 T, and a sample temperature of about 39 °C.

[0265] Prior to recording data the NMR instrument was in a standard calibrated state with calibrated radiofrequency pulses and the intensity of a CPMG experiment calibrated using a standard calibration sample. This is referred to as one-time factory calibration. A NMR data portion was produced from the recorded NMR data, wherein the NMR data portion comprised a plurality of data sets, each comprising inverse recovery delay (I RD) and associated echo time (ED).

[0266] The NMR data portion were analyzed using the same function as used in example 1 using n = 4 spin types and 1000 random starting guesses on M0i, Tl*j, and T2*i.

[0267] This lead to the visual characteristic representation in the form of Tl*-T2* maps Figure for a representative sample containing the PFAS compounds PFDA, PFOS, and PFTeA. Figure 7a shows the Tl* and T2* probability map with Tl* and T2* on axes represented on a logarithmic scale in seconds for a sample containing 30 ppm teflon (left), 2500 ppm of two PFAS molecules (middle), and 20 ppm of a PFAS molecule (right). Figure 7b shows a Tl* and T2* probability map for a sample containing three different PFAS molecules: PFOA, PFOS, PFTeA. The quantity of 19F (mg / L or ppm) measured by the proposed NMR is 469 ppm, which agrees favorably with laboratory measurement of 381 ppm considering uncertainty also for the laboratory data. Figure 7c propose typical 19F Tl*-T2* domains for different types of fluorine as investigated here. Figure 7d shows laboratory and NMR measurements of various floor containing samples. Figure 7e shows a plot of the laboratory and NMR measurements data of figure 7e, which illustrates a correlation there between.

[0268] EXAMPLE 5

[0269] In this example of two samples of animal slurries containing different amounts of total phosphorous (TP), total nitrogen (TN), and ammonium NHx-N were provided and analyzed.

[0270] Each sample was subjected to an NMR analysis, in which NMR data were recorded using a series of CPMG and inversion recovery CPMG experiments with 18 (m) inversion recovery delays distributed logarithmically between 0 and 3 s (i.e., in addition to a standard CPMG experiment, inversion recovery CPMG experiments with inversion recovery delays (7jw) of 0.001, 0.00164938, 0.00272045, 0.00448704, 0.00740083, 0.0122068, 0.0201336, 0.0332078, 0.0547723, 0.0903401, 0.149005, 0.245765, 0.40536, 0.668592, 1.10276, 1.81887, 3 s) and the free induction decay recorded in-between a series of 1000 ( / V) echoes in a CPMG sequence separated by 200 p.s (TE, in each 200 |_is period 101 points were sampled equidistantly. The experiment used 9 |_is 90° (K / 2) and 18 |_is 180° (K) pulses, a repeat delay of 4 s (Tw), and 2 scans, a static magnetic field about 1 T, and a sample temperature of about 39 °C.

[0271] Prior to recording data the NMR instrument was in a standard calibrated state with calibrated radiofrequency pulses and the intensity of a CPMG experiment calibrated using a standard calibration sample. This is referred to as one-time factory calibration.

[0272] An NMR data portion was produced from the recorded NMR data, wherein the NMR data portion comprised a plurality of data sets, each comprising inverse recovery delay (I RD) and associated echo time (ED).

[0273] The NMR data portion were analyzed using the same function as used in example 1 using n = 4 spin types and 1000 random starting guesses on M0i, Tl*j, and T2*i.

[0274] This lead to the visual characteristic representation in the form of Tl*-T2* maps for samples 1 (TP: 508 ppm, TN: 3389 ppm, NHx-N: 2006 ppm) and for sample 2 (TP: 771 ppm, TN: 5213 ppm, NHx-N: 3270 ppm) shown to the left in Figures 8a and 8b, respectively. The right-hand side of Figures 8a and 8b shows the corresponding Tl* and T2* probability map with Tl* and T2* on axes represented on a logarithmic scale in seconds.

[0275] Figures 8a' and 8b' show the maps corresponding to the left map of respectively figures 8a and 8b in color.

Claims

PATENT CLAIMS1. A method for producing a characteristic representation of a substance comprising:• recording nuclear magnetic resonance (NMR) data of a least one sample of the substance,• producing one or more portions of nuclear magnetic resonance (NMR) data of the recorded NMR data, wherein each NMR data portion is produced to comprise one or more NMR data sets ,• processing each NMR data portion(s) into a multidimensional distribution of fitted data sets comprising selecting a function and performing multiple fittings to the selected function, , and• producing the characteristic representation of the substance from the fitted data sets, wherein each data set comprises associated intensity data and relaxation data, wherein the selected function comprises at least 3 fitting parameter(s), wherein the processing of each of said NMR data portion(s) comprises selecting and applying at least one start guess, and wherein each fitted data set comprises data embedding at least one fitted data subset comprising an apparent longitudinal relaxation time (T1*), an associated apparent transverse relaxation time (T2*), and an associated intensity (MO), preferably the function is a function which represents a model of the obtained NMR data portion(s).

2. The method of claim 1, wherein each of the NMR data sets is produced to comprise frequency data, such as at least one of resonance frequency, chemical shift, quadrupole coupling, J coupling, or dipole-dipole coupling.

3. The method of claim 1 or claim 2, wherein each of the NMR data sets is produced to comprise at least one of inversion recovery data or echo NMR data, preferably each of the NMR data sets comprises data representing at least one inversion recovery delay (I RD) and / or data representing at least one echo delay (ED).

4. The method of claim 3, wherein the echo data comprises at least one of spin echo data, fast spin echo data, stimulated echo data, or gradient echo data.

5. The method of any one of the preceding claims, wherein each of the NMR data sets comprises data from the recorded NMR data, wherein the NMR data was recorded with a pulse sequence comprising at least one period where the NMR data is affected by T1 (longitudinal) relaxation and / or at least one period influenced by T2 (transverse) relaxation, such as wherein each of the NMR data sets comprises data of the recorded NMR data recorded with a pulse sequence comprising at least two periods influenced by T1 relaxation or at least two periods influenced by T2 relaxation.

6. The method of any one of the preceding claims, wherein each of the NMR data sets comprises data of the recorded NMR data recorded with a pulse sequence comprising at least one period where the recorded data is affected by chemical exchange.

7. The method of any one of the preceding claims, wherein each of the NMR data sets comprises data of the recorded NMR data recorded comprising a pulse sequence with at least one period where the recorded data is affected by diffusion, such as diffusion of molecules, such as diffusion of ions, such as spin diffusion.

8. The method of any one of the preceding claims, wherein one or more of the NMR data portions comprises at least 2 NMR data sets, such as at least 3 NMR data sets, such as at least 4 NMR data sets, such as at least 8 NMR data sets, such as 16 or more NMR data sets.

9. The method of any one of the preceding claims, wherein the data of at least one of the NMR data sets are data from a common or identical pulse sequence of the recorded NMR data, optionally the NMR data set(s) of at least one of the NMR data portions comprises data from a common or identical pulse sequence of the recorded NMR data.

10. The method of any one of the preceding claims, wherein the data of at least one of the NMR data sets are data from a common or identical set of pulse sequences of the recorded NMR data comprising one or more of the sequences an inversion recovery pulse sequence, a Carr-Purcell-Meiboom-Gill (CPMG) sequence and / or a spin-echo sequence with one or more echoes, optionally the NMR data set(s) of at least one of the NMR data portion(s) comprises data from a common or identical set of pulse sequence.

11. The method of any one of the preceding claims, wherein the data of at least one of the NMR data sets are based on data of the recorded NMR data from a common isotope, preferably the NMR data set(s) of at least one of the NMR data portion(s) are based on data of the recorded NMR data from a common NMR readable isotope.

12. The method of any one of the preceding claims, wherein the processing of said each NMR data portion, comprises performing said multiple fittings of said NMR data portion individually to said formula.

13. The method of any one of the preceding claims, wherein the processing of said each NMR data portion, comprises performing said multiple fittings of said one or more NMR data sets of said NMR data portion in a common fitting procedure to said formula.

14. The method of any one of the preceding claims, wherein said performing of said multiple fittings of said NMR data portion comprises applying said at least one start guess for at least one of said fittings of said multiple fittings, preferably said performing of said multiple fittings of said NMR data portion comprises applying said at least one start guess for a plurality of said multiplefittings, such as all of the fittings of the multiple fittings, wherein said at least one start may be equal or different.

15. The method of claim 14, wherein the start guess comprises a start guess for at least one of the at least 3 fitting parameters for at least one fitting of the multiple fittings, preferably the start guess comprises a start guess for at least one of the at least 3 fitting parameters for a plurality of the multiple fittings, such as for all of the multiple fittings, wherein the start guess may be identical or differ from a fitting to a subsequent fitting according to a selected pattern.

16. The method of any one of the preceding claims, wherein the at least one start guess comprises at least one of a random start guess or a start guess generated by applying bootstrapping.

17. The method of any one of the preceding claims, wherein the at least one start guess is applied to at least one of the fitting parameters in a plurality of the fittings, such as to each of the fittings of the multiple fittings to the function, such as comprising that the start guess applied to the at least one fitting parameter is the same start guess in at least two or more of the multiple fittings to the function, such as the same start guess in each of the multiple fittings to the function.

18. The method of any one of the preceding claims, wherein the processing of respective NMR data portions of a plurality of NMR data portions comprises applying a randomly selected start guess for at least one of the fitting parameters for each of said NMR data portions.

19. The method of any one of the preceding claims, wherein said at least one start guess is characterized by being of Monte Carlo type random start guess.

20. The method of any one of the preceding claims, wherein said at least one start guess is selected to be within a start range, preferably said at least one start guess is a start guess for a fitting parameter and the start guess is within a start range represented as a span of selected values, wherein the selected span of values preferably comprises or consist of one or more typical expected values.

21. The method of any one of the preceding claims, wherein said at least one start guess is selected to be a start guess for a fitting parameter, which is correlated to at least one fitted parameter of the fitted data set and / or to the fitted data subset and wherein the start guess preferably comprises parameters within a start range represented by a selected span of values, wherein the selected span of values comprises or consist of typical expected values, wherein the typical expected values is determined from typical values of said at least one fitted parameter, such as previously determined values of said at least one fitted parameter in a fitting procedure of a processing of another NMR data portion.

22. The method of any one of the preceding claims, wherein said at least one start guess is selected to be a start guess for a fitting parameter which is correlated to at least one parameter of the fitted data subset, wherein the at least one parameter of the fitted data subset is selected from apparent intensity (MO), T1* and T2*, frequency, chemical exchange, or diffusion of said species in said at least one sample.

23. The method of any one of the preceding claims, wherein the function comprises one or more exponential terms.

24. The method of any one of the preceding claims, wherein the function comprises one or more hyperbolic terms.

25. The method of any one of the preceding claims, wherein the function comprises one or more trigonometric terms.

26. The method of any one of the preceding claims, wherein the function comprises one or more terms selected from a logarithm term, an exponential term, a sinusoidal term, a polynomial term, or a differential term.

27. The method of any one of the preceding claims, wherein the function comprises 4 fitting parameters, such as 5 fitting parameters, such as 6 fitting parameters, such as 7 fitting parameters, such as 8 fitting parameters or more.

28. The method of any one of the preceding claims, wherein the function comprises a function Flwherein AbBbCbDLand Etare fitting parameters and IRDj, TEk lare input data from the respective NMR data sets; / is a type of spin among n different spin types;IRDj is an inversion recovery delay applied when obtaining respective j NMR data sets among m different inversion recovery delays;TEk.i is an echo time with k representing the ’th length of each echo in a train of echoes from 1 to p and I is the number of echo(es) from 1 to q in the respective echo trains; and wherein n, m, p, and q, independently of each is an integer from 1 and up, such as up to 100, such as from 1 to 50, such as from 4-25.

29. The method of claim 28, wherein the function Fl is30. The method of any one of the preceding claims, wherein the function is a function Fl Iwherein AbBbCbDLand Etare fitting parameters and IRDj, TEk land RDhare input NMR data from the respective NMR data sets; / is a type of spin among n different spin types;IRD; is an inversion recovery delay applied when obtaining respective j NMR data sets among m different inversion recovery delays;TE / <, / is an echo time with k representing the ’th length of each echo in a train of echoes from 1 to p and I is the number of echo(es) from 1 to q in the respective echo trains;RD / , is a recycle delay during which the magnetization recovers or partially recovers from running one to running subsequent NMR recordings of the / ?’th recycle delay among 1 to r recycle delays; and wherein n, m, p, q, and r, independently of each is an integer from 1 and up, such as up to 100, such as from 1 to 50, such as from 4-25.

31. The method of claim 30, wherein the function Fl I is- exp ( * «Dh))32. The method of any one of the preceding claims, wherein the intensity data of the respective NMR data sets comprises an echo time and wherein the relaxation data of the respective NMR data sets comprises an inversion recovery delay, preferably the respective NMR data sets further comprise a recycle delay.

33. The method of any one of claims 28-32, wherein the parameter C> is correlated to a T1* of a fitted data subset correlated to echo data representing a spin type / among n spin types; the parameter D, is correlated to a T2* of a fitted data subset correlated to echo data representing a spin type / among n spin types and / or at least one of the parameters the A and B, is correlated to an intensity MO and / or a chemical exchange value of a fitted data subset correlated to echo data representing a spin type / among n spin types.

34. The method of anyone of claims 28-33, wherein each fitted data subset comprise a factor Beta;, wherein Betai=-1 / A factor for spin type / among n spin types of each fitted data set representing a signature of chemical exchange.

35. The method of anyone of claims 28-34, wherein each fitted data subset comprises an MO factor, wherein MOj=- A*B, factors for spin type / among n spin types of each fitted data set representing an intensity of the spin type.

36. The method of anyone of claims 28-35, wherein the T1*=-1 / C / factor for spin type / among n spin types of each fitted data set, wherein T1* is representing a T1* of the spin type / .

37. The method of anyone of claims 28-36, wherein the T2*i=-1 / Dzfactor for spin type / among n spin types of each fitted data set, wherein T2* is representing a T2* of the spin type / .

38. The method of anyone of claims 28-37, wherein the E, factor for spin type / among n spin types of each fitted data set, wherein the E, factor is representing diffusion of a molecule, a moiety or an ion comprising the spin type / .

39. The method of anyone of claims 28-38, wherein n is at least 2, such as up to 1000, such as from 4 to 100.

40. The method of any one of the preceding claims, wherein each fitted data subset represents a characteristic of a spin type, preferably comprising at least a T1*, an associated T2*, and an associated intensity (M0), more preferably each fitted data subset represents a characteristic of a spin type further comprising one ormore of a diffusion parameter, a frequency shift, a chemical exchange, a J coupling, a quadrupole coupling, or a dipole-dipole coupling.

41. The method of any one of the preceding claims, wherein the production of the characteristic representation of the substance from the fitted data sets comprises deriving from each fitted data set, a representative fraction comprising said at least one fitted data subset comprising the T1*, the associated T2*, and the associated intensity (MO) to obtain a plurality of representation fractions and assembling said plurality of representation fractions to produce the characteristic representation of the substance.

42. The method of claim 41, wherein said T1*, said associated T2*, and said associated intensity (MO) of said at least one fitted data subset comprises data representing a characteristic of spin diffusion of molecular or ionic species in inhomogeneous static or radio-frequency fields.

43. The method of claim 41 or claim 42, wherein said T1*, said associated T2*, and said associated intensity (MO) of said at least one fitted data subset comprises data representing characteristic of a chemical exchange involving two or multiple spin types in said material.

44. The method of any one of the preceding claims, wherein the characteristic representation comprises a probability map of T1* and T2* of the respective spin types.

45. The method of any one of the preceding claims, wherein if the solved Beta; factors (-1 / A) of one or more of the fitted data subset is beyond or at least different from an expected value of a factor x2, it is an indicator for that chemical exchange or diffusion occurring on similar time scale range as at least one I RDj.

46. The method of any one of the preceding claims, wherein the observed T1*-T2* correlation do not represent the T1-T2 correlation of the spin type, and wherein the method comprises extracting T1-T2 correlation as part of the fitting procedure.

47. The method of any one of the preceding claims, wherein at least one of the fitted data sets comprises data embedding two or more fitted data subsets comprising a T1*, an associated T2,* and an associated intensity (MO), and wherein the generation of the characteristic representation of the substance from the at least one fitted data set comprising said two or more fitted data subsets, comprises deriving from the at least one fitted data set, a representation fraction comprising said two or more fitted data subsets of the T1*, the associated T2*, and the associated intensity (MO).

48. The method of any one of claims 41-47, wherein at least one of the fitted data subsets, each comprising a characteristic of a pair or a group of spin types comprising at least two spin types, and wherein the method comprises deriving a chemical exchange characteristic from said characteristic of the pair or the group of spin types and determining T 1 and T2 from the T1* and the T2* and said characteristics of the associated chemical exchange process or processes.

49. The method of any one of claims 41-48, wherein each of said fitted data subsets comprises at least one Beta; factor and / or at least one diffusion parameter, wherein the least one Beta; factor and / or the at least one diffusion parameter comprises a characteristic of one or a pair or a group of spin types, and wherein the method comprises deriving a chemical exchange characteristic from said characteristic of the pair or the group of spin types and / or diffusion characteristics for at least one spin type, preferably the production of the characteristic representation of the substance comprises producing the characteristic representation of the substance to comprise said chemical exchange characteristic from said characteristic of the pair or the group of spin types and / or diffusion characteristics for at least one spin type.

50. The method of any one of the preceding claims, wherein the NMR data portion(s) comprises data from at least one NMR readable isotope, preferably the NMR data from one of the one or more of the data portions is / are single isotope data comprising NMR data from one isotope.

51. The method of any one of the preceding claims, wherein the NMR data portion(s) comprises data from at least two isotopes, preferably the method comprises producing at least a plurality of NMR data portions, comprising producingat least a first data portion comprises NMR data from a first isotope and a second NMR data portion comprising data from a second isotope, preferably the plurality of NMR data portions comprises at least a number of NMR data portions comprising NMR data from respective isotopes, wherein the number of NMR data portions is up to 50, such as up to 10, such as from 2 to 5.

52. The method of claim 50 or 51 , wherein said at least one isotope is selected from the group of isotopes1H,2H,6Li,7Li,10B,11B,13C,14N,15N,17O,19F23Na,27AI,29Si31P,33S,35CI,37CI , and39K,41K,43Ca,47Ti,49Ti,50V,51V,53Cr,55Mn,57Fe,59Co,61Ni,63Cu,65Cu,67Zn,69Ga,71Ga,75As,77Se,79Br,81Br,83Kr,85Rb,87Rb,87Sr,89Y,91Zr,93Nb,95Mo,97Mo,105Pd,107Ag,109Ag,111Cd,113Cd,117Sn,119Sn,115Sn,121Sb,135Ba,137Ba177Pb,199Hg,201Hg, and207Pb.

53. The method of claim 50 or 51 , wherein said first and second or said respective isotopes comprises at least two isotopes selected from the group of isotopes1H,2H,6Li,7Li,10B,11B,13C,14N,15N,17O,19F23Na,27AI,29Si31P,33S,35CI,37CI , and39K,41K,43Ca,47Ti,49Ti,50V,51V,53Cr,55Mn,57Fe,59Co,61Ni,63Cu,65Cu,67Zn,69Ga,71Ga,75As,77Se,79Br,81Br,83Kr,85Rb,87Rb,87Sr,89Y,91Zr,93Nb,95Mo,97Mo,105Pd,107Ag,109Ag,111Cd,113Cd,117Sn,119Sn,115Sn,121Sb,135Ba,137Ba177Pb,199Hg,201Hg, and207Pb.

54. The method of any one of the preceding claims, wherein the step of producing the one or more portions of NMR data comprises recording said NMR data comprising performing at least one NMR reading using an NMR spectrometer, such as a low-field NMR spectrometer.

55. The method of claim 54, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a sequence of RF exciting pulses and reading relaxation / dephasing signals of at least one spin of at least one isotope.

56. The method of claim 54 or claim 55, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a sequence of RF exciting pulses comprising a plurality of pulses comprising at least one 180° RF pulse and at least one 90° RF pulse.

57. The method of any one of claims 54-56, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a sequence of RF exciting pulses comprising a Hahn echo sequence or a CPMG sequence or a combination comprising at least one of these.

58. The method of any one of claims 54-57, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a sequence of RF exciting pulses comprising a spin echo sequence, a fast spin echo sequence, a stimulated echo sequence, a gradient echo sequence or any combinations comprising at least one of these.

59. The method of any one of claims 54-58, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to one or more sequences of RF exciting pulses comprising one or more of the sequences an inversion-recovery pulse sequence, a Carr-Purcell-Meiboom-Gill (CPMG) sequence and / or a spin-echo sequence or any combinations comprising at least one of these.

60. The method of any one of claims 54-59, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a sequence of RF pulses and delays comprising a sub sequence SE1• applying a first 180° RF pulse• applying a first inversion recovery delay (I RD)• applying a first 90° RF pulse and• determine a resonance intensity, directly or throughout an echo sequence, such as CPMG, preferably the method comprises repeating the SE1 subsequence applying different I RD values, and optionally determining a null I RD value where the determined resonance intensity passes through a point of zero intensity.

61. The method of any one of claims 54-59, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a sequence of RF pulses and delays comprising a sub sequence SE1vi. applying a 90° RF vii. applying an echo delay viii. applying a 180° RF pulse ix. repeating steps ii and iii N times x. applying a recycle delay, preferably the method comprises repeating the SE2 subsequence optionally applying different recycle delays.

62. The method of any one of claims 54-60, wherein the at least one NMR reading comprises subjecting the at least one sample of the substance to a repeated sequence of RF exiting signals comprisinginvert IRD excite 1stED 1strefocus ED refocus wait RD repeat x (N-l) wherein and / 2 designates radio-frequency pulses with pulse duration and radiofrequency pulse field strengths adjust to provide pulse flip angles of (such as 180°) and K / 2 (such as 90°), respectively, Tim is the inversion recovery delay (IRD), TE is the echo delay (ED), Tw the recycle delay (RD), and N the number of echoes.

63. The method of claim 62, wherein the NMR data are recorded during one or more of the delay periods.

64. The method of any one of claims 54-63, wherein the pulse sequence is repeated a plurality of time with different composition of one more of the parameters IRD, ED, N, RD, radio-frequency pulse field strengths, or radio-frequency pulse durations and / or flip angles.

65. The method of any one of claims 54-64, wherein the pulse sequence is repeated a plurality of time with different composition and number of elements of inversion, waiting, and echo periods.

66. The method of any one of claims 54-65, wherein the at least one NMR reading comprises reading at least according to any of the preceding claims, wherein the NMR data sets comprises data from at least two different NMR readable isotopes, one of the two NMR readable isotope being1H.

67. The method of any one of claims 54-66, wherein the at least one NMR reading comprises reading at least two different NMR readable isotopes, one of the two NMR readable isotope being13C.

68. The method of any one of claims 54-67, wherein the at least one NMR reading comprises reading multiple NMR readable isotopes in parallel, comprising initiating a further pulse sequence for another NMR readable isotope in an RD, IRD, or ED period of a previous NMR readable isotope pulse sequence.

69. The method of any one of claims 54-68, wherein the at least one NMR reading comprises performing additional readings over time comprising obtaining an additional NMR data portion from each reading, processing the additional NMR data portion, obtaining additional representation fraction and comparing the additional representation fraction to the characteristic representation to thereby obtain information of changes in said sample.

70. The method of claim 69, wherein the information of changes in said substance comprises change in physical condition, such as temperature, hydration, pressure, salt content, pH, solvent type, or concentration of components / elements in said substance.

71. The method of any one of the preceding claims, wherein the fitting of a plurality of NMR data portions of said sample improves the measuring accuracy in terms of standard deviation for one or more of the parameters individually and / or reduces a required data measurement time for a selected measurement accuracy.

72. The method of any one of the preceding claims, wherein the substance comprises at least two components, preferably each component comprises a unique NMR readable isotope.

73. The method of any one of the preceding claims, wherein the substance is a complex substance, such as a food product, such as one or more samples in any stage of wastewater treatment, such as an animal slurry or manure, such as a biogas digester slurry, such as livestock feed, such as one or more samples at any stage of biorefinery, such as one or more samples in any stage of mining or mineral refining processes.

74. The method of any one of the preceding claims, wherein the substance comprises at least an organic and / or an inorganic component(s).

75. The method of any one of the preceding claims, wherein the method comprises preparing a plot of the characteristic representation of the substance.

76. The method of any one of the preceding claims, wherein the method comprises determining a quality parameter of the substance comprising deriving data from said characteristic representation and determining said quality parameter.

77. The method of claim 76, wherein the quality parameter of the substance is associated with an amount of a component, such as the amount of an organic component, such as the amount of lipid, such as an amount of fat, such as the amount of carbohydrate, such as the amount of protein, such as an amount of nucleic acid, such as an amount of chemical reagent, such as an amount of an isotope, such as an amount of mobile sample components, such as an amount of rigid / immobile sample components, such as an amount of species undergoing chemical exchange, such as an amount of species undergoing molecular, ionic, or spin diffusion.

78. The method of claim 76 or claim 77, wherein the quality parameter of the substance is associated with a chemical property of the substance, such as a pH value.

79. The method of any one of the preceding claims, wherein the characteristic representation comprises a plurality of identifiers, each identifier representing a statistical distribution of said sample components.

80. The method of any one of the preceding claims, wherein the method comprises comparing the characteristic representation of the substance with at least one historical characteristic representation, to identify a difference, such as a chemical and / or biological and / or a compositional difference(s) between the substance and the substantially similar substance.

81. The method of claim 80, wherein the method comprises determining a quality parameter based on the identified difference.

82. The method of claim 80 or claim 81 , wherein the at least one historical characteristic representation comprises at least one historical characteristic representation of at least on reference substance or at least one historical characteristic representation synthesized from two or more historical characteristic representation of at least on reference substance.

83. The method of claim 82, wherein at least on reference substance corresponds to the substance, in that it comprises at least 25 % of the components present in the substance, such as at least 50 % of the components present in the substance, such as at least 90 % of the components present in the substance.

84. The method of any one of the preceding claims, wherein the processing is performed by a trained computer, wherein the trained computer is trained to determine one or more quality parameters and one or more sample identifiers, preferably to determine quantitative parameters and / or to monitoring differences in historical data, such as in chemical, physical, or biological processes, preferably the trained computer is trained to classify according to one or more quality parameters and / or according to one or more sample identifiers.

85. The method of any one of the preceding claims, wherein the processing is involving artificial intelligence, machine learning, reinforcement learning, or similar in supervised, unsupervised or any other form.

86. A method for generating a trained computer comprising subjecting the computer to a machine learning process comprising feeding data to the computer, wherein the data comprises data packets of previous obtained NMR data correlatedto at least one fitted data subset comprising a T1*, an associated T2*, and an associated intensity (MO).

87. The method for generating the trained computer of claim 86, wherein the data packets comprises data representing a characteristic representation of a substance and the NMR data portions applied for generating the characteristic representation.

88. A trained computer obtainable by the method according to any one of claims 86 or claim 87.

89. An NMR system for performing the method of any one of the preceding claims, wherein the NMR system comprises a computer system, wherein the computer system is configured for receiving one or more portions of nuclear magnetic resonance (NMR) data of the substance, wherein the NMR data portion(s) comprises a plurality of NMR data sets, each comprising intensity data and relaxation data, and wherein the computer system is configured for• processing the NMR data portion(s) into a multidimensional distribution of fitted data sets by performing multiple fittings to at least one function, wherein the function(s) comprise(s) at least 2 fitting parameter(s) and is a function(s) which represent(s) a model of the obtained NMR data portion(s), and• generating the characteristic representation of the substance from the fitted data sets.

90. The NMR system of claim 89, wherein the NMR system is configured for carrying out the method of any one of claim 1-87.

91. The NMR system of claim 89 or claim 90, wherein the computer system comprises separate computer portions adapted for wireless data communication, preferably via the Internet.

92. The NMR system of any one of the preceding claims 89-91 , wherein the NMR system comprises an NMR spectrometer, preferably the NMR spectrometer is in data communication with the computer system.

93. The NMR system of any one of the preceding claims 89-92, wherein the computer system comprises a super computer system, a cluster of computers, computer systems with multiple CPU’s, computer systems with multiple GPU’s, preferably located at a remotely location relative to the NMR spectrometer.

94. The NMR system of any one of the preceding claims 89-93, wherein the computer system comprises the trained computer according to any one of claims 84-88.

Images