Single Component Mixtures Research Articles

Using AI/ML to predict blending performance and process sensitivity for Continuous Direct Compression (CDC)

Utilising three artificial intelligence (AI)/machine learning (ML) tools, this study explores the prediction of fill level in inclined linear blenders at steady state by mapping a wide range of bulk powder characteristics to processing parameters. Predicting fill levels enables the calculation of blade passes (strain), known from existing literature to enhance content uniformity. We present and train three AI/ML models, each demonstrating unique predictive capabilities for fill level. These models collectively identify the following rank order of feature importance: RPM, Mixing Blade Region (MB) size, Wall Friction Angle (WFA), and Feed Rate (FR). Random Forest Regression, a machine learning algorithm that constructs a multitude of decision trees at training time and outputs the mode of the classes (classification) or mean prediction (regression) of the individual trees, develops a series of individually useful decision trees. but also allows the extraction of logic and breakpoints within the data. A novel tool which utilises smart optimisation and symbolic regression to model complex systems into simple, closed-form equations, is used to build an accurate reduced-order model. Finally, an Artificial Neural Network (ANN), though less interrogable emerges as the most accurate fill level predictor, with an r2 value of 0.97. Following training on single-component mixtures, the models are tested with a four-component powdered paracetamol formulation, mimicking an existing commercial drug product. The ANN predicts the fill level of this formulation at three RPMs (250, 350 and 450) with a mean absolute error of 1.4%. Ultimately, the modelling tools showcase a framework to better understand the interaction between process and formulation. The result of this allows for a first-time-right approach for formulation development whilst gaining process understanding from fewer experiments. Resulting in the ability to approach risk during product development whilst gaining a greater holistic understanding of the processing environment of the desired formulation.

Automated component analysis in DOSY NMR using information criteria

This study introduces a model selection technique based on Bayesian information criteria for estimating the number of components in a mixture during Diffusion-Ordered Spectroscopy (DOSY) Nuclear Magnetic Resonance (NMR) data analysis. As the accuracy of this technique is dependent on the efficiency of parameter estimators, we further investigate the performance of the Weighted Least Squares (WLS) and Maximum a Posteriori (MAP) estimators. The WLS method, enhanced with meticulously tuned L2-regularization, effectively detects components when the difference in self-diffusion coefficients is more than two-fold, especially when the component with the smaller coefficient has a larger weight ratio. The MAP method, strengthened by a substantial database of prior information, exhibits outstanding precision, decreasing this threshold to 1.5 times. Both estimators provide weight ratio estimates with standard deviations of approximately around 1 percentage point, although the MAP method tends to overestimate the component with a larger self-diffusion coefficient. Deviations from the expected values can exceed 10 percentage points, often due to inaccuracies in component detection. The error estimates are determined using data resampling techniques derived from a large-scale 1000-point experiment and an additional five measurements from a single-component mixture. This approach allowed us to thoroughly examine data distribution characteristics, thereby laying a robust groundwork for future refinement efforts.

A Model for Lithium-Ion Battery Electrolyte Mixture Densities

Lithium-ion battery (LIB) electrolytes are generally composed of an organic hydrocarbon and lithium salt (e.g., LiPF6). The effect of salt addition to the electrolyte solvent increases its density. Experimental density data for basic and realistic, multicomponent mixtures are scarce in the literature, and a predictive model for electrolyte solution density is not yet available. This work resolves to create a predictive method that can be used to estimate the density of these mixtures. A predictive method which accounts for intramolecular forces in the electrolyte mixture was developed and fit to the sparse density data available in the literature. The model exhibited high accuracy for single component electrolyte mixtures with most predictions falling within 1% of measured values and all predictions falling within 5%. The model was further extended to more realistic, multi-solvent electrolyte mixtures which exhibited similar accuracy, seen in Fig. 1. It was noted that the multi-component predictions were moderately worse than for the single component mixtures. The source of this disparity is not clear but may be derived from uncertainty in the data reported by electrolyte vendors and/or the presence of solvent-to-solvent interactions in the solution which are not accounted for in the current modeling framework. In addition to the density model, an accurate method for computing absolute molar and mass fractions of multi-solvent mixtures with specified volumetric concentrations (e.g., 1.2 M LiPF6 in 1:1:1 %vol. EC/ DEC/DMC) is also described for when that is the only known relation for a solution. Computed mass concentration values yielded an average error of 1.4% with a maximum value of 9.6%. The molar and mass approaches were combined to yield a modeling framework capable of computing molar concentrations and densities of all LIB electrolyte solutions based on LiPF6 loaded in any combination of ethylene carbonate (EC), ethyl methyl carbonate (EMC), diethyl carbonate (DEC), dimethyl carbonate (DMC), and propylene carbonate (PC). The largest fault with the final modeling system is that it suffers from a lack of sufficient, high-quality experimental data for validation of its accuracy over a wide range of electrolyte formulations and conditions. This model’s accuracy is dependent on the experimental data utilized for correlation. Accordingly, more data from differing sources will help increase the accuracy and confidence of the model developed herein and allow it for expansive use in LIB fields. Future work will be required to extend the current approach to any novel electrolyte formulations developed in the research and development sector. Figure 1

Enhanced heavy metal ions adsorption by 4‑aminobenzoic acid grafted on chitosan/epichlorohydrin composite: Kinetics, isotherms, thermodynamics and desorption studies

The multi-functional chitosan is important for the development of novel adsorbents which is needed for the effective removal of scavenging metal ions from water and wastewater. In this study the functionalized materials were characterize by FTIR, SEM/EDX, TGA/DTA, BET and XRD. Physicochemical parameters including pH, contact time, ionic strength and temperature were investigated in batch mode to optimize the conditions to be applied on a commercial scale for the maximum removal of metal ions using batch applications. The results showed that adsorption performance of grafted cross-linked chitosan beads can be modeled efficiently by Langmuir isotherm and pseudo-second order kinetic model. Thermodynamic parameters such as Gibb's free energy change (∆G°), enthalpy change (∆H°) and entropy change (∆S°) were calculated and the results showed the adsorption of Pb(II), Cu(II), Ni(II), Zn(II) and Cd(II) ions onto grafted crosslinked chitosan beads (G/ECH-CS) are spontaneous and endothermic in nature. Batch regeneration of the spent adsorbent from the single component mixture was effective. Multiple adsorption/desorption experiments were also carried out, and it was found that 0.0% mass loss of the G/ECH-CS was observed after the third cycle of adsorption/desorption studies.

A stochastic optimization approach for the design of organic fluid mixtures for low-temperature heat recovery

Over 50% of the heat generated in industry is in the form of low-grade heat (with operating temperatures below 370°C). Recovering heat from these sources with standard Rankine cycles (using water as working fluid) is inefficient and expensive. Organic working fluids have become an attractive alternative to mitigate these inefficiencies. In this work, we address the problem of designing flexible multi-component organic fluids capable of withstanding variability in heat source temperatures and efficiencies of individual cycle equipment units. The design problem is cast as a nonlinear stochastic optimization problem and we incorporate risk metrics to handle extreme variability. We show that a stochastic optimization framework allows us to systematically trade-off performance of the working fluid under a variety of scenarios (e.g., inlet source temperatures and equipment efficiencies). With this, it is possible to design working fluids that remain robust in a wide range of operational conditions. We also find that significant flexibility of the working fluid can be obtained by using optimal concentrations as opposed to using single component mixtures. We also find that state-of-the-art nonlinear optimization solvers can handle highly complex stochastic optimization problems that incorporate detailed physical representations of the system.

Spectrophotometric Multicomponent Analysis of Irbesartan, Hydrochlorothiazide and Ramipril in Pharmaceutical Formulations by Chemometric Techniques

Multivariate spectrophotometric methods offer an extensive application for quantification of multicomponent mixtures that exhibit severe spectral overlapping. Partial least Squares (PLS) and Principal componenet Regression (PCR) methods were proposed for the spectrophotometric multicomponent analysis of a ternary mixture consisting of Telmisartan(TEL), Hydrochlorothiazide (HCZ) and Ramipril (RAM) without prior separation. In these chemometric techniques, the measurement of the absorbance values were made in the spectral range from 200-350 nm in the intervals of  = 1 nm at 151 wavelengths in the zero order, first and second derivative spectra of the ternary mixtures of these active ingredients in 0.1 M NaOH. The prepared calibrations of both techniques using absorbance data and concentration matrix data sets were used to predict the concentration of the active ingredients in their ternary, binary and single component mixtures. The optimized models were tested on an external validation set containing synthetic mixtures of the mentioned components. The models were finally used to assay the studied drugs in the commercial formulations. The methods are selective for the determination of telmisartan, hydrochlorothiazide and ramipril and overcome the spectral interference. The prediction values for TEL, HCZ and RAM in both models were found satisfactory.

Absorption spectroscopy for the quantitative prediction of lanthanide concentrations in the 3LiCl–2CsCl eutectic at 723 K

The absorption spectra of single-component mixtures of six lanthanide chloride salts were obtained in the molten salt eutectic 3LiCl–2CsCl at 723 K, and used to build multivariate regression models to predict concentrations of multi-component mixtures of these metals in solution.

PH-Dependent Selective Protein Adsorption into Mesoporous Silica

The adsorption of lysozyme, cytochrome c and myoglobin, similar-sized globular proteins of approximately 1.5 nm radius, into the mesoporous silica material Santa Barbara Amorphous-15 (SBA-15) with 3.3 nm mean pore radius has been studied photometrically for aqueous solutions containing a single protein type and for binary protein mixtures. Distinct variations in the absolute and relative adsorption behavior are observed as a function of the solution's pH-value, and thus pore wall and protein charge. The proteins exhibit the strongest binding below their isoelectric points pI, which indicates the dominance of electrostatic interactions between charged amino acid residues and the -OH groups of the silica surface in the mesopore adsorption process. Moreover, we find for competitive adsorption in the restricted, tubular pore geometry that the protein type which shows the favoured binding to the pore wall can entirely suppress the adsorption of the species with lower binding affinity, even though the latter would adsorb quite well from a single component mixture devoid of the strongly binding protein. We suggest that this different physicochemical behavior along with the large specific surface and thus adsorption capability of mesoporous glasses can be exploited for separation of binary mixtures of proteins with distinct pI by adjusting the aqueous solution's pH.

An overview of environmental forensics

Environmental forensics has emerged as an important area of environmental studies over the past two decades. There are two basic aspects to any environmental investigation. The first being a conventional approach where the standard EPA (Environmental Protection Agency) methods are used to determine concentrations of selected compounds released into the environment. These methods are extremely well documented and widely used, but only provide information on specific target compounds. Whilst this information may be useful for monitoring purposes it is of little use when trying to determine the source of a spill or contaminants in the environment. If the purpose of an investigation is to determine the source of a contaminant, or point of release, then it is necessary to use a wide variety of analytical techniques and integrate all of the resulting data into one comprehensive data set. It may not always be possible to obtain a unique answer, particularly in the case of groundwater contaminants where there might only be one compound, for example MTBE or PCE. In that case if there are multiple possible sources in the area it may be difficult to narrow it down to a specific source. Furthermore fingerprinting tools that may be useful with complex mixtures may not be directly applicable to single component mixtures. The purpose of this paper will be to provide a brief overview, along with some recent examples of the type of information that is typically obtained in an environmental forensic investigation and how this information may be interpreted. It should be noted that all these examples are related to organic contaminants in the environment since that is the major area of focus at this time. Examples will involve hydrocarbons, chlorinated solvents, and MTBE and BTEX compounds. Techniques will include gas chromatography (GC), gas chromatography-mass spectrometry (GCMS), stable isotopes both bulk and gas chromatography-isotope ratio mass spectrometry (GCIRMS). There are other techniques being used but in a paper of this length we will limit ourselves to these widely used techniques and those applications mentioned above since space does not permit a comprehensive review of all of them.

A Component Slope Linear Model for Mixture Experiments

The Scheffé linear model and the Cox linear model are two first-order model forms commonly used for analyzing mixture experiment data. In a fitted Scheffé linear model, the coefficient for a mixture component is the predicted response at that pure component (i.e., single-component mixture). In a fitted Cox linear model, the coefficient for a mixture component is the predicted difference in response at that pure component and at a pre-specified reference mixture. An alternative first-order model for mixture experiments, the component slope linear mixture (CSLM) model, is proposed in this article. In a fitted CSLM model, the coefficient for a mixture component is the predicted slope of the linear response surface along the Cox effect direction for that component (where the direction is determined by a pre-specified reference mixture). The CSLM model and methods for fitting it are presented. The advantages and disadvantages of the CSLM model, Scheffé linear model, and Cox linear model are discussed. The three models are illustrated and interpreted for an example involving the liquidus temperature of spinel crystals in simulated nuclear waste glass.

A note comparing component-slope, Scheffé and Cox parameterizations of the linear mixture experiment model

A mixture experiment involves combining two or more components in various proportions and collecting data on one or more responses. A linear mixture model may adequately represent the relationship between a response and mixture component proportions and be useful in screening the mixture components. The Scheffé and Cox parameterizations of the linear mixture model are commonly used for analyzing mixture experiment data. With the Scheffé parameterization, the fitted coefficient for a component is the predicted response at that pure component (i.e. single-component mixture). With the Cox parameterization, the fitted coefficient for a mixture component is the predicted difference in response at that pure component and at a pre-specified reference composition. This article presents a new component-slope parameterization, in which the fitted coefficient for a mixture component is the predicted slope of the linear response surface along the direction determined by that pure component and at a pre-specified reference composition. The component-slope, Scheffé, and Cox parameterizations of the linear mixture model are compared and their advantages and disadvantages are discussed.

Multivariate calibration techniques applied to the spectrophotometric analysis of one-to-four component systems

The UV spectrophotometric analysis of a multicomponent mixture containing paracetamol, caffeine, tripelenamine and salicylamide by using multivariate calibration methods, such as principal component regression (PCR) and partial least-squares regression (PLS), was described. The calibration set was based on 47 reference samples, consisting of quaternary, ternary, binary and single-component mixtures, with the aim to develop models able to predict the concentrations of unknown samples containing as many as one-to-four components. The calibration models were optimized by an appropriate selection of the number of factors as well as wavelength ranges to be used for building up the data matrix and excluding any information about the interfering excipients included in pharmaceutics. The PCR and PLS models were compared and their predictive performance was inferred by a successful application to the assays of synthetic mixtures and pharmaceutical formulations.

Relaxation mechanism in some dipolar mixtures

The dielectric constant and loss factor of the following mixtures have been studied near room temperature as a function of frequency in the microwave region: (1) nitrobenzene and o-nitrotoluene in n-dodecane in concentrations of 0.1, 0.2 and 0.3 mole fraction of each of the polar components; (2) binary mixtures of chlorobenzene and bromobenzene in various relative proportions. The static dielectric constants and the refractive indices of the above mixtures have also been measured. In the case of the ternary mixture, the value of the distribution parameter, α, has been found to increase somewhat in direct proportion to the concentration of the polar solutes. The loss-factor-frequency curves for this mixture seem to be a simple sum of the experimental curves for the corresponding single component mixtures. The Cole-Cole plots for the binary mixtures of halobenzenes show normal behavior. The dielectric data for both the mixtures have also been interpreted in terms of two partly superimposed Debye regions. The results for the mixtures studied are consistent with the assumption that the relaxation processes are those of molecules rather than larger liquid regions as proposed by some earlier workers for systems which they studied.