Dr Vasanth Kumar Kannuchamy

[Display omitted] •Deep neural networks are successfully deployed in chemical engineering class rooms.•Deep neural networks accurately predict the adsorption equilibrium data.•Neural networks can be deployed in fast-track mode. The latest industrial revolution, Industry 4.0, is progressing exponentially and targets to integrate artificial intelligence and machine learning algorithms with existing technology to digitalise chemical processes across the industry, especially in the area of online monitoring, predictive analysis and modelling. Machine learning algorithms are being constantly implemented in both academic laboratories and industry to uncover the underlying correlations that exist in the high-dimensional and complex experimental and synthetic data that describes a chemical process. Indeed soon, proficiency in artificial intelligence methodology will become a required skill of a chemical engineer. It is therefore becoming essential to train chemical engineers with these methods to help them to adapt to this new era of digitised industries. Keeping these issues in mind, we introduced deep neural networks to the final-year chemical engineering students through a computer laboratory exercise. The exercise was delivered in fast-track mode: the students were asked to develop deep neural networks to model and predict the equilibrium adsorption of uptake of three different acids by activated carbon at four different temperatures. In this manuscript, we discuss in detail this laboratory exercise from delivery and design to the results obtained and the students’ feedback. In the classroom, the students compared the adsorption equilibrium data obtained using the established theoretical adsorption isotherms and empirical correlations with the neural networks developed in the classroom. The experience obtained from the classroom confirmed that this exercise gave the students the essential knowledge on the AI and awareness on the jargons in the world of machine language and obtained the required level of coding skills to develop a simple neural net with one layer or a sophisticated deep networks to model an important unit operation in chemical engineering and to accurately predict the experimental outcomes.

Characterization of the guest-host interactions and the heterogeneity of porous materials is essential across the physical and biological sciences, for example for gas sorption and separation, pollutant removal from wastewater, biological systems (protein-ligand binding) and molecular recognition materials such as molecularly imprinted polymers. Information about the guest-host interactions can be obtained from calorimetric experiments. Alternatively, more detailed information can be obtained by properly analysing the experimentally acquired adsorption equilibrium data. Adsorption equilibrium is usually interpreted using theoretical adsorption isotherms that correlate with the equilibrium concentration of the adsorbate in the solid phase and in the bulk fluid at a constant temperature. Such theoretical isotherms or expressions can accurately predict the adsorbent efficiency (at equilibrium) as a function of process variables such as the initial adsorbate concentration, adsorbent mass, reactor volume and temperature. Detailed analysis of the adsorption isotherms permits the calculation of the number density of the adsorbent sites, their binding energy for the guest molecules and information about the distribution of adsorption site binding energies. These analyses are discussed in this review. A critical evaluation of the analytical and numerical methods that can characterize the heterogeneity and guest-host interactions involved in terms of discrete or continuous binding site affinity distribution was performed. Critical discussion of the limitations and the advantages of these models is provided. An overview of the experimental techniques that rely on calorimetric and chromatographic principles to experimentally measure the binding energy and characteristic properties of adsorbent surfaces is also included. Finally, the potential use of site energy distribution functions and their potential to provide new information about the binding energy of adsorbents for a specific guest molecule application is discussed.

During the growth of crystals in impure solution, impurities can pin the flow of the elementary steps and decrease the growth rate or even arrest the crystal growth. In this work, for the first time, we showed that curcumin crystals can grow in impure solution that contains two structurally similar impurities, following a non-classical crystallisation pathway that deviates from the pinning mechanism. We showed that, in a highly impure solution that contains 20 wt% of impurities, a high supersaturation can initiate the crystal growth via sympathetic nucleation that involves the formation of new growth surfaces on the seed crystals. These new surfaces formed on the seed crystals at the expense of higher supersaturation act like active growth surfaces and dictate the entire crystal growth kinetics especially at lower supersaturations. We showed that, if we can artificially create new surfaces that look like giant macrosteps at the micron scale on the crystal surface, then these macrosteps can not only speed-up the crystallisation rate but also control the rate of transfer of impurities into the bulk crystals.

Adsorption is a widely used chemical engineering unit operation for the separation and purification of fluid streams. Typical uses of adsorption include the removal of targeted pollutants like antibiotics, dyes, heavy metals, and other small to large molecules from aqueous solutions or wastewater. To date several adsorbents that vary in terms of their physicochemical properties and costs have been tested for their efficacy to remove these pollutants from wastewater. Irrespective of the type of adsorbent, nature of the pollutant, or experimental conditions, the overall cost of adsorption depends directly on the adsorption contact time and the cost of the adsorbent materials. Thus, it is essential to minimize the amount of adsorbent and the contact time required. We carefully reviewed the attempts made by several researchers to minimize these two parameters using theoretical adsorption kinetics and isotherms. We also clearly explained the theoretical methods and the calculation procedures involved during the optimization of the adsorbent mass and the contact time. To complement the theoretical calculation procedures, we also made a detailed review on the theoretical adsorption isotherms that are commonly used to model experimental equilibrium data that can be used to optimize the adsorbent mass.

Impurities are frequently encountered during the crystallisation of active pharmaceutical compounds. Impurities can either adsorb onto active sites or replace atoms of the crystal lattice. Locating the impurities, especially structurally-similar impurities, on the crystal surface is challenging. In this work, we showed that time-of-flight secondary ion mass spectrometry (TOF-SIMS) can be successfully used to simultaneously quantify the composition of the two structurally-similar impurities, demethoxycurcumin (DMC) and bisdemethoxycurcumin (BDMC), on the surface of the curcumin (CUR) crystals obtained from crystal growth and cooling crystallisation experiments. The distribution of the three curcuminoids on the crystal surface was evaluated through imaging of the specific ions that correspond to intact molecules of the curcuminoids. In terms of location, both BDMC and DMC were evenly distributed on the surfaces of the facets of the curcumin crystals. Additionally, we found that BDMC and DMC occupy their own sites or they are located on the same site on the crystal surface providing evidence for lattice replacement of these molecules. Thus our results demonstrate the potential of TOF-SIMS to provide new information that can help to understand crystallisation.

Crystallisation from solution is an important process in pharmaceutical industries and is commonly used to purify active pharmaceutical ingredients. Crystallisation involves phase change and the mechanisms involved are random which makes the process stochastic. This creates a variation in the time required to reach a fixed percentage of yield from batch to batch. It is essential to regulate the batch crystallisation process and make it more predictable for industrial applications for the ease of process chain scheduling of upstream and downstream unit operations. In this work, we propose a new technique called dotted crystallisation, where carbon dots are used to dictate and regulate events associated with nucleation and crystallisation processes. Following the rules of two-step nucleation theory, the carbon dots intentionally added to a supersaturated solution of curcumin anchors the crystallising compound to form prenucleation clusters that evolve into stable nuclei. Using curcumin as a model compound, we showed that the nucleation of this compound in isopropanol can be regulated, and the nucleation rate can be improved via addition of small quantities of carbon dots to the supersaturated solution. Our results confirmed that the nucleation rate of curcumin by dotted crystallisation was roughly four times higher than the nucleation rate by conventional cooling crystallisation and produced smaller sized crystals with a narrow size distribution.

Adsorption is a widely used unit operation for concentrating solutes from large solution volumes onto small quantities of solid adsorbents. Understanding adsorption kinetics is essential for assessing the process efficiency and the impact of operating variables. Kinetic data are typically analyzed using theoretical models that reveal adsorption mechanisms, variations in pore size, and diffusion behavior, which are critical parameters for designing fixed bed adsorption systems and developing in silico simulations. Adsorption kinetic models are generally classified as empirical, semiempirical, or theoretical, each offering distinct insights into adsorption mechanisms. The choice of model depends on whether the aim is to fit experimental data or to interpret the underlying mechanism. This review summarizes the most widely used kinetic models, their mathematical foundations, and physical interpretations. Practical guidance is provided on model selection, identification of rate-limiting steps, and the use of error functions to determine the best-fitting kinetic expression, with examples from wastewater treatment applications.

Crystals growing via nonclassical crystallisation pathways often lose the shape factor and the mechanisms involved are often anomalous. This makes it difficult to predict the crystal growth kinetics using the conventional crystal growth kinetic models. A new pseudo second order growth kinetic expression to predict the anomalous growth behaviour of curcumin crystals is introduced. This model also allows several important parameters relevant to the kinetics of the crystal growth, such as mass crystallised with respect to time, overall crystal growth kinetic constant, the maximum number of active sites involved in the crystal growth process and the initial crystal growth rate, to be obtained. According to the pseudo second order kinetics, the activation energy for the nonclassical crystal growth of curcumin was found to be similar to 69 kJ mol-1, which seems to agree with the activation energy for the classical crystal growth of organic compounds reported in the literature. The total number of active sites involved in the crystal growth process was found to be similar to 1022 per gram of the seeds. Pseudo second order kinetics can accurately predict the crystal growth kinetics.

Biochar (BC) and hydrochar (HC) have attracted considerable attention owing to their versatile characteristics and proven effectiveness in diverse technical fields. Solid BC is generated as a result of the dry carbonisation process of pyrolysis, in contrast to the slurry HC, which is produced during the hydrothermal carbonisation process. In this study, we evaluated the adsorption potential of two hydrochar samples (HCs) and three biochar samples (BCs) produced from sugar cane bagasse. The adsorption capacity of these samples was tested for ammonium, nitrate, and phosphate ions under various conditions. The BCs and HCs were subjected to characterisation using a CHNS/O analyser, the zeta potential, and Fourier transform infrared (FTIR). Elevating the pyrolysis temperature of the biochar resulted in changes in the fixed carbon and ash contents, while the volatile matter and H/C and O/C atomic ratios decreased. As the residence time increased, the H/C ratio and volatile matter content of the hydrochars (HCs) decreased. However, the fixed carbon content, ash content, and O/C and C/N ratios exhibited an increase. Thermodynamics, adsorption isotherms, and pH were also taken into consideration. The FTIR spectra analysis indicated that the carboxyl and ester functional groups present in both the BCs and HCs displayed reduced peak intensities subsequent to the adsorption of the three ions. While the adsorption was exothermic, we noticed that the adsorption capacity increased with temperature. The results indicate that sorption was homogenous across all binding sites, as evidenced by the optimal fit to the Langmuir isotherm. The research findings indicate that the adsorption capacity of various BC and HC adsorbents is significantly influenced by the surface area of the adsorbents in the case of nitrate and phosphate, but in the case of ammonia, adsorption is dictated by the functional polar groups present on the adsorbent surface.

Anisotropic, biomorphic (wood-derived) potassium-based geopolymer (KGP) - biocarbon (CB) composites with net shape were manufactured by infiltrating KGP slurry into monolithic porous biocarbon (CB) structures (~77 vol% porosity) derived from pyrolyzing beech wood. About 70% of the pores in the three-dimensional (3D) CB structures were infiltrated by the KGP slurry. Compared to pure KGP, the energy absorption per unit volume in compression loading of the KGP-CB composites was increased by ten-fold. After heat treatment at 1000 °C for 1 h in N2, the compressive strength of the KGP-CB composites increased from ~7–24 MPa, accompanied by the formation of crystalline leucite (K2O·Al2O3·4SiO2) phase in the KGP. The KGP-CB composite also exhibited three orders higher electrical conductivity than pure KGP. The effect of temperature on the formation of crystalline phases in KGP and KGP-CB composites was investigated. FTIR, TGA and SEM analyses were used to investigate the changes in microstructures and phase formation during thermal treatment.

We present the design and evaluation of a simulation-based learning approach implemented in a second-year chemical engineering module, integrating Aspen HYSYS with a methanol production case study to embed sustainability and develop student awareness of the United Nations Sustainable Development Goals (SDGs). Through process simulation, we enable students to model complete flowsheets, explore scenario-based analyses, and evaluate sustainability-driven process modifications. Within a structured mini-design project involving reaction and separation systems, we require students to analyse mass and energy balances, optimise process performance, and evaluate trade-offs between efficiency and sustainability. We also ask students to identify and justify relevant SDGs within their designs. Our analysis of student outputs shows improved ability to link engineering decisions with sustainability objectives, enhanced energy-focused reasoning, and stronger systems-level thinking. Overall, the results highlight the effectiveness of simulation-based learning in integrating technical competence with sustainability literacy in undergraduate chemical engineering education.

Cooling crystallization experiments of curcumin in isopropanol confirmed that curcumin can crystallize via classical or nonclassical pathways, depending on the levels of supersaturation and supercooling. Light microscopy analysis revealed that classical crystallization produces needle-shaped single crystals with an equilibrium habit, while nonclassical crystallization results in spherulitic mesocrystals. Through a series of experiments under various conditions, we developed a crystal habit phase diagram for curcumin in pure isopropanol. Presented here for the first time, this diagram illustrates the relationship between supersaturation, supercooling, and crystal habit, offering a valuable guide for controlling curcumin crystallization pathways.

Crystallization via nucleation can isolate active pharmaceutical ingredients from their crudes. While chemical engineering textbooks provide theoretical knowledge on crystallization and nucleation theories, they often fall short in providing provide practical insights on the nucleation mechanism. To bridge this gap, we introduced a virtual experiment on nucleation in second-year chemical engineering classrooms. The main goal is to educate students on crystallization procedures in research and process industries, teaching them how to analyse and manage collected data while integrating theoretical knowledge. This includes conveying the kind of information that can be obtained from a crystallisation process and instructing students on how to analyse and manage the data collected in the light of the theories learned. We devised an original chemical engineering problem on nucleation, derived directly from the raw data collected in the classroom from virtual experiments. This method differs from the conventional approach of solving standard textbook problems. The textbook problems, regrettably often lack crucial information on how nucleation rate or surface free energy are directly obtained from raw data. By the conclusion of the virtual experiment, students have acquired a comprehensive understanding encompassing both practical and theoretical aspects of crystallization, with a particular focus on nucleation. The methodologies elucidated in this study can be applied across a spectrum of chemical engineering modules, including process engineering, unit operations in chemical engineering, mass transfer, and can even be integrated into specialized courses dedicated to crystallization.

Developing theory-informed standard operating procedures (SOPs) for the continuous crystallisation of pharmaceuticals still remains a bottleneck. For the continuous manufacturing of pharmaceuticals, the current methods rely on the laborious trial-and-error approach to identify process conditions such as the dilution rate (flow per unit volume of reactor) and initial supersaturation, where the productivity will be at maximum at steady-state conditions. This approach, while proven and considered to be useful, lacks or ignores the information obtained from batch kinetics. Herein for the first time, we propose a theoretical method to develop batch kinetics-informed theoretical procedures for the continuous manufacturing of a model compound curcumin in isopropanol. The theoretical approach uses batch kinetic constants to theoretically identify the optimum dilution rate and the corresponding mass of the model compound curcumin when crystallised, as well as its productivity at steady-state conditions as a function of initial supersaturation. The theory-informed procedures will serve as a valuable guideline to develop operating procedures for the continuous production of the target compound and thus eliminate the trial-and-error approach used to develop the protocols for the continuous manufacturing of pharmaceuticals. We also showed that our methods allow for the estimation of the dilution rate that corresponds to washout conditions (i.e., where all the crystals in the crystalliser will be washed out due to the high flow rate of the input stream) during the continuous manufacturing of crystals.

The Langmuir isotherm is used to determine the properties of a theoretical “holy grail” adsorbent that can meet the US Department of Energy’s methane storage target of 0.5 g/g and 266 v/v. For a storage tank operating between 5 and 65 bar, the adsorbent requires a maximum adsorption capacity of 0.8388 g/g, a binding affinity of 0.05547 bar–1, and a material density of 377 g/L. For a tank operating between 5 and 80 bar, the binding affinity should be 0.05 bar–1, with the same capacity and density. The Langmuir isotherm is also applied to calculate the necessary adsorbent properties, including the number of adsorption sites and binding energies, to achieve the volumetric storage target of 266 v/v based on the material’s density.

Background: Metastable zone width (MSZW) and solubility are crucial for developing crystallization procedures in the purification of active pharmaceutical ingredients (APIs). Traditionally, determining these properties involves labor-intensive methods that can take weeks or even months. With advancements in process analytical technologies (PAT) and the increasing focus on quality by design (QbD) in pharmaceutical manufacturing, more efficient and reliable protocols are needed. In this study, we employ in situ Fourier Transform Infrared (FTIR) spectroscopy and Focused Beam Reflectance Measurement (FBRM) to establish protocols for measuring solubility at different temperatures and MSZW at varying cooling rates. Methods: We experimentally determined MSZW and solubility using FTIR spectroscopy and FBRM. IR spectra were analyzed to obtain solubility concentrations, while FBRM counts were used to extract MSZW and supersolubility concentrations. The collected data were assessed using four theoretical models, including a newly developed model based on classical nucleation theory. By fitting experimental MSZW data to these models, we determined nucleation kinetics and thermodynamic parameters. Results: Our novel model exhibited excellent agreement with experimental MSZW data across different cooling rates, demonstrating its robustness. The nucleation rate constant and nucleation rate ranged between 10(2)(1) and 10(2)(2) molecules/m(3)s. The Gibbs free energy of nucleation was calculated as 3.6 kJ/mol, with surface energy values between 2.6 and 8.8 mJ/m(2). The estimated critical nucleus radius was in the order of 10(-)(3) m. Conclusions: The protocols we developed for predicting MSZW and solubility of paracetamol using PAT can serve as a guideline for other APIs. Our theoretical model enhances the predictive accuracy of nucleation kinetics and thermodynamics, contributing to optimized crystallization processes.

We propose a new mathematical model based on the classical nucleation theory to predict the nucleation rate, kinetic constant, and Gibbs free energy of nucleation using metastable zone width (MSZW) data as a function of solubility temperature. Unlike widely used models by N & yacute;vlt, Kubota, and Sangwal, which are limited in capturing the impact of varying cooling rates, the proposed model allows direct estimation of nucleation rates from MSZW data obtained under different cooling conditions. This is particularly advantageous for continuous or semi-batch crystallisation design, where cooling rate is a critical variable. The model has been successfully validated using experimental data from 22 solute-solvent systems, including 10 APIs, one API intermediate, lysozyme, and glycine, as well as 8 inorganic compounds. Predicted nucleation rates span from 1020 to 1024 molecules per m3 s for APIs, and up to 1034 molecules per m3 s for lysozyme, the largest molecule studied. Gibbs free energy of nucleation varies from 4 to 49 kJ mol-1 for most compounds, reaching 87 kJ mol-1 for lysozyme. The model also enables accurate prediction of induction time and key thermodynamic parameters such as surface free energy, critical nucleus size, and number of unit cells-based solely on MSZW data obtained at different cooling rates.

In this work, we report the new protocols that are developed to determine the shape factors of Form I paracetamol crystals using a combination of techniques that rely on the state-of-the-art X-ray computed tomography and Morphologi G3. The determined shape factors successfully predicted the crystal growth rate of paracetamol in 2-propanol and the length of the crystals growing in the supersaturated solution. X-ray computed tomography and optical microscopy can be effectively combined to predict crystal shape factors with reasonable accuracy.

Impurities play a critical role in crystal growth by adsorbing onto growth steps and reducing step advancement rates. In this study we introduce the Statistical Distribution Model (SDM), which incorporates the random distribution of immobile impurities using Poisson statistics. We define two key parameters: the velocity reduction factor (A), which quantifies how strongly an impurity reduces step velocity once pinned, and the pinning probability scaling constant (B), which reflects how likely it is for impurity pairs to fall within a critical pinning distance. Using these two parameters, we classify impurities into at least five types, ranging from those that weakly reduce velocity and rarely pin to those that strongly inhibit growth through frequent pinning events. Depending on impurity strength and coverage, we show that SDM accurately describes three commonly observed behaviors during the growth of crystals in impure solutions: complete growth arrest, partial inhibition, and linear reduction in velocity. The model shows excellent agreement with experimental data from ADP, KBr, and sucrose systems, offering improved insight into impurity-induced growth kinetics.