Deep reinforcement learning for the direct optimization of gradient separations in liquid chromatography.

While Reinforcement Learning (RL) has already proven successful in performing complex tasks, such as controlling large-scale epidemics, mitigating influenza and playing computer games beyond expert level, it is currently largely unexplored in the field of separation sciences. This paper therefore aims to introduce RL, specifically proximal policy optimization (PPO), in liquid chromatography, and evaluate whether it can be trained to optimize separations directly, based solely on the outcome of a single generic separation as input, and a reward signal based on the resolution between peak pairs (taking a value between [-1,1]). More specifically, PPO algorithms or agents were trained to select linear (1-segment) or multi-segment (2-, 3-, or 16-segment) gradients in 1 experiment, based on the outcome of an initial, generic linear gradient (ϕstart=0.3, ϕend=1.0, and tg=20min), to improve separations. The size of the mixtures to be separated varied between 10 and 20 components. Furthermore, two agents, selecting 16-segment gradients, were trained to perform this optimization using either 2 or 3 experiments, in sequence, to investigate whether the agents could improve separations further, based on previous outcomes. Results showed that the PPO agent can improve separations given the outcome of one generic scouting run as input, by selecting ϕ-programs tailored to the mixture under consideration. Allowing agents more freedom in selecting multi-segment gradients increased the reward from 0.891 to 0.908 on average; and allowing the agents to perform an additional experiment increased the reward from 0.908 to 0.918 on average. Finally, the agent outperformed random experiments as well as standard experiments (ϕstart=0.0, ϕend=1.0, and tg=20min) significantly; as random experiments resulted in average rewards between 0.220 and 0.283, and standard experiments resulted in average rewards of 0.840. In conclusion, while there is room for improvement, the results demonstrate the potential of RL in chromatography and present an interesting future direction for the automated optimization of separations.

Comprehensive two-dimensional liquid chromatography with high resolution mass spectrometry to investigate the photoelectrochemical degradation of environmentally relevant pharmaceuticals and their degradation products in water.

The widespread presence of organic micropollutants in the environment reflects the inability of traditional wastewater treatment plants to remove them. In this context, advanced oxidation processes (AOPs) have emerged as promising quaternary wastewater treatment technologies since they efficiently degrade recalcitrant components by generating highly reactive free radicals. Nonetheless, the chemical characterization of potentially harmful byproducts is essential to avoid the contamination of natural water bodies with hazardous substances. Given the complexity of wastewater matrices, the implementation of comprehensive analytical methodologies is required. In this work, the simultaneous photoelectrochemical degradation of seven environmentally relevant pharmaceuticals and one metabolite from the EU Watch List 2020/1161 was examined in ultrapure water and simulated wastewater, achieving excellent removal efficiencies (overall >95%) after 180 min treatment. The reactor unit was linked to an online LC sample manager, allowing for automated sampling every 15 min and near real-time process monitoring. Online comprehensive two-dimensional liquid chromatography (LC × LC) coupled with high resolution mass spectrometry (HRMS) was subsequently used to tentatively identify degradation products after photoelectrochemical degradation. Two reversed-phase liquid chromatography (RPLC) columns were used: an SB-C18 column operated with 5 mM ammonium formate at pH 5.8 (1A) and methanol (1B) as the mobile phases in the first dimension and an SB-Aq column using acidified water at pH 3.1 (2A) and acetonitrile (2B) as the mobile phases in the second dimension. This resulted in a five-fold increase in peak capacity compared to one-dimensional LC while maintaining the same total analysis time of 50 min. The LC x LC method allowed the tentative identification of 12 venlafaxine, 7 trimethoprim and 10 ciprofloxacin intermediates. Subsequent toxicity predictions suggested that some of these byproducts were potentially harmful. This study presents an effective hybrid technology for the simultaneous removal of pharmaceuticals from contaminated wastewater matrices and demonstrates how multidimensional liquid chromatography techniques can be applied to better understand the degradation mechanisms after the treatment of micropollutants with AOPs.

Comprehensive analysis of the effective and intra-particle diffusion of weakly retained compounds in silica hydrophilic interaction liquid chromatography columns.

A detailed analysis of intra-particle volumes and layer thicknesses and their effect on the diffusion of solutes in hydrophilic interaction liquid chromatography (HILIC) was made. Pycnometric measurements and the retention volume of deuterated mobile phase constituents (water and acetonitrile) were used to estimate the void volume inside the column, including not only the volume of the mobile phase but also part of the enriched water solvent acting as the stationary phase in HILIC. The mobile phase (hold-up) volume accessible to non-retained components was estimated using a homologous series approach. The joint analysis of the different approaches indicated the formation of enriched water layers on the hydrophobic silica mesopore walls with a thickness varying significantly with mobile phase composition. The maximal thickness of the enriched water layers, which corresponded to the minimum void volume accessible to unretained solutes, marked a transition in the retention behavior of the studied analytes. Discrepancies between deuterated solvent measurements and pycnometry were explained by the existence of an irreplaceable water layer adsorbed on the silica surface. Regarding the diffusion behavior in HILIC, peak parking experiments were used to interpret the influence of the acetonitrile content on the effective diffusion coefficient Deff. A systematic decrease in Deff and molecular diffusion Dm was observed with decreasing acetonitrile concentration, primarily attributed to variations in mobile phase viscosity. Notably, Deff/Dm remained nearly unaffected by variations in mobile phase composition. Finally, the effective medium theory was used to make a comprehensive analysis of Dpart/Dm to study the contribution to band broadening when the solute resides in the mesopores. The obtained data unveiled a curvature with a minimum corresponding to conditions of maximum water-layer thickness and retention. For the weakly retained compounds (k' < 0.5) the Dpart/Dm-values were found to be relatively high (order of 0.35-0.5), which directly reflects the high γsDs/Dm-values that were observed (order 0.35-7).

Detailed analysis of the effective and intra-particle diffusion coefficient of proteins at elevated pressure in columns packed with wide-pore core-shell particles.

To determine the efficiency that can be obtained in a packed-bed liquid-chromatography column for a particular analyte, a correct determination of the molecular and effective diffusion coefficients (Dm and Deff) of the analyte is required. The latter is usually obtained via peak parking experiments wherein the flow is stopped. As a result, the column pressure rapidly dissipates and the measurement is essentially conducted at ambient pressure. This is problematic for analytes whose retention depends on pressure, such as proteins and potentially other large (dipolar) molecules. In that case, a conventional peak parking experiment is expected to lead to large errors in Deff. To obtain a better estimate ofDeff, the present study reports on the use of a set-up enabling peak parking measurements under pressurized conditions. This approach allowed us to report, for the first time, Deff for proteins at elevated pressure under retained conditions. First, Deff was determined at a (average) pressure of about 105 bar for a set of proteins with varying size, namely: bradykinin, insulin, lysozyme, ß-lactoglobulin, and carbonic anhydrase in a column packed with 400 Å core-shell particles. The obtained data were then compared to those of several small analytes: acetophenone, propiophenone, benzophenone, valerophenone, and hexanophenone. A clear trend between Deff and analyte size was observed. The set-up was then used to determine Deff of bradykinin and lysozyme at variable (average) pressures ranging from 28 bar to 430 bar. These experiments showed a decrease in intra-particle and surface diffusion with pressure, which was larger for lysozyme than bradykinin. The data show that pressurized peak parking experiments are vital to correctly determine Deff when the analyte retention varies significantly with pressure.

Automated method development in high-pressure liquid chromatography.

Method development in liquid chromatography is a crucial step in the optimization of analytical separations for various applications. However, it is often a challenging endeavour due to its time-consuming, resource intensive and costly nature, which is further hampered by its complexity requiring highly skilled and experienced scientists. This review presents an examination of the methods that are required for a completely automated method development procedure in liquid chromatography, aimed at taking the human out of the decision loop. Some of the presented approaches have recently witnessed an important increase in interest as they offer the promise to facilitate, streamline and speed up the method development process. The review first discusses the mathematical description of the separation problem by means of multi-criteria optimization functions. Two different strategies to resolve this optimization are then presented; an experimental and a model-based approach. Additionally, methods for automated peak detection and peak tracking are reviewed, which, upon integration in an instrument, allow for a completely closed-loop method development process. For each of these approaches, various currently applied methods are presented, recent trends and approaches discussed, short-comings pointed out, and future prospects highlighted.

Detailed comparison of in-house developed and commercially available heart-cutting and selective comprehensive two-dimensional liquid chromatography systems.

Recently, two-dimensional liquid chromatography (2D-LC) has become a popular approach to analyze complex samples. This is partly due to the introduction of commercial 2D-LC systems. In the past, 2D-LC was carried out on in-house developed setups, typically consisting of several switching valves and sample loops as the interface between the two dimensions. Commercial systems usually offer different 2D-LC modes in combination with specialized software to operate the instrument and analyze the data. This makes them highly user-friendly, however, at an increased cost compared to in-house developed setups. This study aims to make a comparison between an in-house developed 2D-LC setup and a commercially available 2D-LC instrument. The comparison is made based on experimental differences, in addition to more general differences, including cost price, flexibility, and ease of operation. Special attention is also paid to the different strategies to deal with the mobile phase incompatibility between the highly orthogonal separation mechanisms considered in this work: hydrophilic interaction liquid chromatography (HILIC) and reversed-phase LC (RPLC). For the commercial 2D-LC instrument, this is done using active solvent modulation (ASM), a valve-based approach allowing the on-line dilution of the effluent eluting from the first dimension column before transfer to the second dimension (2D) column. For the in-house developed setup, a combination of restriction capillaries and a trap column is used. Using a sample of 28 compounds with a large polarity range, peak shapes and recoveries of the 2D-chromatograms are compared for both setups. For early eluting compounds, the selective comprehensive approach, currently only possible on the commercial 2D-LC instrument, results in the best peak shapes and recoveries, however, at the cost of an increased analysis time. In general, depending on the analytical goal (single heart-cut versus full-comprehensive 2D-LC), an in-house developed system can be satisfactory for the analysis of specific target compounds/samples. For more complex problems, it can be interesting to use a more specialized commercial 2D-LC instrument. Overall, this comparison study provides advice for analytical scientists, who are considering to use 2D-LC, on the type of equipment to consider, depending on the needs of their particular applications.

A perspective on the use of deep deterministic policy gradient reinforcement learning for retention time modeling in reversed-phase liquid chromatography.

Artificial intelligence and machine learning techniques are increasingly used for different tasks related to method development in liquid chromatography. In this study, the possibilities of a reinforcement learning algorithm, more specifically a deep deterministic policy gradient algorithm, are evaluated for the selection of scouting runs for retention time modeling. As a theoretical exercise, it is investigated whether such an algorithm can be trained to select scouting runs for any compound of interest allowing to retrieve its correct retention parameters for the three-parameter Neue-Kuss retention model. It is observed that three scouting runs are generally sufficient to retrieve the retention parameters with an accuracy (mean relative percentage error MRPE) of 1 % or less. When given the opportunity to select additional scouting runs, this does not lead to a significantly improved accuracy. It is also observed that the agent tends to give preference to isocratic scouting runs for retention time modeling, and is only motivated towards selecting gradient scouting runs when penalized (strongly) for large analysis/gradient times. This seems to reinforce the general power and usefulness of isocratic scouting runs for retention time modeling. Finally, the best results (lowest MRPE) are obtained when the agent manages to retrieve retention time data for % ACN at elution of the compound under consideration that spread the entire relevant range of ACN (5 % ACN to 95 % ACN) as well as possible, i.e., resulting in retention data at a low, intermediate and high % ACN. Based on the obtained results, we believe reinforcement learning holds great potential to automate and rationalize method development in liquid chromatography in the future.

Automated tuning of denoising algorithms for noise removal in chromatograms.

Different algorithms, such as the Savitzky-Golay filter and Whittaker smoother, have been proposed to improve the quality of experimental chromatograms. These approaches avoid excessive noise from hampering data analysis and as such allow an accurate detection and quantification of analytes. These algorithms require fine-tuning of their hyperparameters to regulate their roughness and flexibility. Traditionally, this fine-tuning is done manually until a signal is obtained that removes the noise while conserving valuable peak information. More objective and automated approaches are available, but these are usually method specific and/or require previous knowledge. In this work, the L-and V-curve, k-fold cross-validation, autocorrelation function and residual variance estimation approach are introduced as alternative automated and generally applicable parameter tuning methods. These methods do not require any previous information and are compatible with a multitude of denoising methods. Additionally, for k-fold cross-validation, autocorrelation function and residual variance estimation, a novel implementation based on median estimators is proposed to handle the specific shape of chromatograms, typically composed of alternating flat baselines and sharp peaks. These tuning methods are investigated in combination with four denoising methods; the Savitsky-Golay filter, Whittaker smoother, sparsity assisted signal smoother and baseline estimation and denoising using sparsity approach. It is demonstrated that the median estimators approach significantly improves the denoising and information conservation performance of relevant smoother-tuner combinations up to a factor 4 for simulated datasets and even up to a factor 10 for an experimental chromatogram. Moreover, the parameter tuning methods relying on residual variance estimation, k-fold cross-validation and autocorrelation function lead to similar small root-mean squared errors on the different simulated datasets and experimental chromatograms. The sparsity assisted signal smoother and baseline estimation and denoising using sparsity approach, which both rely on the use of sparsity, systematically outperform the two other methods and are hence most appropriate for chromatograms.

Treatment of antimicrobial azole compounds via photolysis, electrochemical and photoelectrochemical oxidation: Degradation kinetics and transformation products.

The degradation kinetics and transformation products of pharmaceutical azole drugs from Watch List 2020/1161 (fluconazole, FCZ; miconazole, MCZ; clotrimazole, CTZ; and sulfamethoxazole, SMX) are examined individually and as a mixture in Milli-Q water and simulated wastewater (SWW) upon treatment with three different advanced oxidation processes: (i) photolysis (UV), (ii) electrochemical (eAOP), and (iii) photoelectrochemical (eAOP/UV). For individual pollutant degradation, UV was found to be significantly more effective for SMX and CTZ compared to MCZ and FCZ. Whereas when treating the azole drugs mixture, eAOP/UV was determined to be the most effective treatment method. The degradation efficiency was higher in Milli-Q than in SWW because the treatment efficiency depended on the matrix compositions. The degradation products formed under different processes were identified, and the routes of transformation were proposed. The results of this study can assist in the selection of the most suitable treatment technology depending upon the pollutant or matrix.

Electrochemical degradation of 17α-ethinylestradiol: Transformation products, degradation pathways and in vivo assessment of estrogenic activity.

Conventional water treatment methods are not efficient in eliminating endocrine disrupting compounds (EDCs) in wastewater. Electrochemical Advanced Oxidation Processes (eAOPs) offer a promising alternative, as they electro-generate highly reactive species that oxidize EDCs. However, these processes produce a wide spectrum of transformation products (TPs) with unknown chemical and biological properties. Therefore, a comprehensive chemical and biological evaluation of these remediation technologies is necessary before they can be safely applied in real-life situations. In this study, 17α-ethinylestradiol (EE2), a persistent estrogen, was electrochemically degraded using a boron doped diamond anode with sodium sulfate (Na2SO4) and sodium chloride (NaCl) as supporting electrolytes. Ultra-high performance liquid chromatography coupled to quadrupole time-of-flight mass spectrometry was used for the quantification of EE2 and the identification of TPs. Estrogenic activity was assessed using a transgenic medaka fish line. At optimal operating conditions, EE2 removal reached over 99.9% after 120 min and 2 min, using Na2SO4 and NaCl, respectively. The combined EE2 quantification and in vivo estrogenic assessment demonstrated the overall estrogenic activity was consistently reduced with the degradation of EE2, but not completely eradicated. The identification and time monitoring of TPs showed that the radical agents readily oxidized the phenolic A-ring of EE2, leading to the generation of hydroxylated and/or halogenated TPs and ring-opening products. eAOP revealed to be a promising technique for the removal of EE2 from water. However, caution should be exercised with respect to the generation of potentially toxic TPs.

In-Depth Performance Analysis and Comparison of Monolithic and Particulate Zwitterionic Hydrophilic Interaction Liquid Chromatography Polymer Columns.

The kinetic performance of different zwitterionic hydrophilic interaction liquid chromatography polymer columns is evaluated and compared in-depth. For this purpose, two lab-made monolithic columns, synthesized with different crosslinkers, and a commercial particle packed column are considered. It is found that performance evaluation techniques, such as comparing plate height curves or fitted A-, B- and C-terms, obtained by fitting experimental plate height data to a plate height model, are complicated by the determination of a reliable characteristic length. This is due to the very different morphology of these column types, and the heterogeneity of the monolithic columns. The occurrence of a convective flow through the packed particle column further complicates the interpretation of the obtained fitting parameters, as part of the C-term is wrongfully attributed to the A-term. Therefore, the use of the kinetic plot method is suggested for the comparative evaluation of these columns, as kinetic plots do not require the determination of a characteristic length, nor rely on any fitting parameters. With the kinetic plot method, it is demonstrated that the lab-made monolithic columns outperform the packed particle column for plate counts between 10,000 and 800,000. This is attributed to the higher column efficiency of these columns, due to their small domain and skeleton sizes, and their high permeability, resulting from their high external porosity and the occasional occurrence of preferential flow paths.

Sterilisation of biopharmaceuticals: Effect of gamma irradiation, e-beam irradiation and nitrogen dioxide on human insulin.

Biopharmaceuticals are administered parenterally and therefore sterility is required. Sterility can be obtained via different processes including exposure to steam or dry heat. Sterilisation studies on biopharmaceuticals, which are highly sensitive medicinal products, are scarce. This study investigates the effect of different sterilisation processes on recombinant human insulin in solid state (gamma and e-beam irradiation (w/wo dry ice), nitrogen dioxide (NO2)) and in aqueous solution (gamma irradiation (w/wo dry ice, w/wo glycerin)) using ultra-high performance liquid chromatography-diode array detection-mass spectrometry. It is observed that NO2 substantially degrades the solid samples, while gamma and e-beam irradiation result in lower levels of degradation (mean normalized peak areas of 95.2-96.2 % with respect to the non-sterilised samples). Gamma irradiation of insulin solutions with and without dry ice at 2.5 kGy results in mean normalised peak areas of 85 % and <40 % with respect to the non-sterilised samples, respectively. It is concluded that sterilisation using ionising radiation of liquid biopharmaceuticals with insulin and sterilisation of insulin dry powder using NO2 is less suitable with the set-ups used here because of substantial degradation. In contrast, evidence is presented in favour of sterilisation of insulin dry powder using ionising radiation.

Optimizing transfer and dilution processes when using active solvent modulation in on-line two-dimensional liquid chromatography.

Two-dimensional liquid chromatography (2D-LC) is becoming increasingly popular for the analysis of complex samples, which is partly due to the recent introduction of commercial 2D-LC systems. To deal with the mobile phase incompatibility between highly orthogonal retention mechanisms, such as hydrophilic interaction liquid chromatography (HILIC) and reversed-phase LC (RPLC), several strategies have been introduced over the years. One of these strategies is active solvent modulation (ASM), a valve-based approach allowing the on-line dilution of the effluent eluting from the first dimension before transfer to the second dimension. This strategy has gained a lot of attention and holds great potential, however, no clear guidelines are currently in place for its use. Therefore, this study aims to investigate how the ASM process can be optimized when using highly incompatible LC combinations, such as HILIC and RPLC, in a simplified selective comprehensive 2D-LC set-up (sHILIC x RPLC) to suggest guidelines for future users. Using a representative sample, the dilution factor (DF), the duration of the ASM phase, the filling percentage of the sample loops, and their unloading configuration are investigated and optimized. It is observed that a DF of 10 with an optimal ASM phase duration, a sample loop filling of maximum 25%, and an unloading configuration in backflush mode, result in the best peak shapes, intensities, and recoveries for early eluting compounds, while keeping the total analysis time minimal. Based on these results, some general recommendations are made that could also be applied in other 2D-LC modes, such as comprehensive 2D-LC (LC x LC), heart-cutting 2D-LC (LC-LC), and other chromatographic combinations with mobile phase incompatibility issues.

Application of partial ozonation on tank truck cleaning concentrate and the influence on biodegradability and ecotoxicity: a pilot-scale study.

This study investigates the pilot-scale ozone treatment of reverse osmosis concentrate (ROC), originating from variable tank truck cleaning wastewater. The influence of ozonation on short- and long-term biodegradation potential was examined through respirometry and Zahn-Wellens, respectively. Ecotoxicity was also examined for several concentrate batches and ozonation steps. Chemical oxidation through ozone had a beneficial effect on chemical oxygen demand removal, with a removal efficiency up to 56%. Formation of short-term biochemical oxygen demand (BODst) was induced for several, but not all batches, showing the potential of subsequent biological treatment of ozonated ROC. An increase in the inherent biodegradability through Zahn-Wellens was observed for all tested samples after ozonation, rising to a maximum of 68% after 3 hours of ozonation, highlighting the importance of sludge adaptation. Ecotoxicity, tested with Artemia franciscana and the saltwater algae P. tricornutum, showed initial decreases in algae inhibition after short ozonation periods. An increase in algae inhibition was, however, seen after prolonged ozonation for all tested ROC samples, pointing to the formation of ecotoxic by-products. Artemia showed no significant toxicity effects. When applying biological treatment through Zahn-Wellens, a decrease in ecotoxicity was observed for several samples, likely through biological oxidation of the produced degradation products.

On the occurrence of very low intra-particle diffusion rates in zwitterionic hydrophilic interaction liquid chromatography polymer columns.

Intra-particle diffusion is investigated in particle-packed and monolithic zwitterionic hydrophilic interaction liquid chromatography (HILIC) polymer columns. For this purpose, plate heights are first measured over a broad range of velocities for polar compounds with zone retention factors between k″ =  2 and 10 on all column types, multiplied with the corresponding velocity, and extrapolated to zero velocity. The thus obtained B-term coefficients are additionally verified via peak parking experiments and indicate a very low degree of longitudinal diffusion in all columns, that moreover seems to be independent of the zone retention factor. This is in contrast to earlier observations on reversed-phase liquid chromatography columns and bare silica HILIC columns. To obtain more insight into the reasons for these low levels of longitudinal diffusion, the experimentally obtained effective diffusion coefficients are modeled to equations derived from the effective medium theory, such as the models developed by Torquato and Jefrey & Hashin. It is, however, demonstrated that these models fail to adequately describe effective diffusion in the studied zwitterionic HILIC columns, due to the low degree of intraparticle diffusion in these stationary phases. Fitting the experimental effective diffusion coefficients to the two limiting cases of the effective medium theory, i.e., the pure parallel-connection and pure serial-connection case, it is observed that to stay within these limits, intraparticle diffusion must be extremely low, i.e., Dpz/Dm ≤ 0.01, indicating that the diffusion inside the pores is at least 100 times smaller than the bulk diffusion coefficient. There are several possible explanations for these low intraparticle diffusion coefficients, such as a strongly hindered diffusion in the polymer matrix, extremely low surface diffusion rates in the aqueous layer adsorbed onto the stationary phase and/or the occurrence of slow localized adsorption events.

Review of recent insights in the measurement and modelling of the B-term dispersion and related mass transfer properties in liquid chromatography.

In this contribution, we review the recent literature relating to the measurement and modelling of all diffusion-dominated processes contributing to the efficiency of a chromatographic column. In first instance, this involves the measurement and modelling of the overall effective diffusion coefficient Deff (determining the so-called B-term band broadening). The latter manifests itself most clearly during a so-called peak parking experiment. Using effective medium theory modelling, the measured Deff-value can subsequently be decomposed into its constituent contributions, of which the intra-particle or the mesoporous zone and the surface diffusion coefficient are the most important ones. As an accurate estimation of the diffusion processes also allows computing the C-term plate height contribution terms, the review ends with some recent insights obtained when using the established B- and C-term contributions to compute the degree of eddy-dispersion in contemporary packed bed columns.

Convolutional neural network for automated peak detection in reversed-phase liquid chromatography.

Although commercially available software provides options for automatic peak detection, visual inspection and manual corrections are often needed. Peak detection algorithms commonly employed require carefully written rules and thresholds to increase true positive rates and decrease false positive rates. In this study, a deep learning model, specifically, a convolutional neural network (CNN), was implemented to perform automatic peak detection in reversed-phase liquid chromatography (RPLC). The model inputs a whole chromatogram and outputs predicted locations, probabilities, and areas of the peaks. The obtained results on a simulated validation set demonstrated that the model performed well (ROC-AUC of 0.996), and comparably or better than a derivative-based approach using the Savitzky-Golay algorithm for detecting peaks on experimental chromatograms (8.6% increase in true positives). In addition, predicted peak probabilities (typically between 0.5 and 1.0 for true positives) gave an indication of how confident the CNN model was in the peaks detected. The CNN model was trained entirely on simulated chromatograms (a training set of 1,000,000 chromatograms), and thus no effort had to be put into collecting and labeling chromatograms. A potential major drawback of this approach, namely training a CNN model on simulated chromatograms, is the risk of not capturing the actual "chromatogram space" well enough that is needed to perform accurate peak detection in real chromatograms.

Kinetics and mechanisms of the carbamazepine degradation in aqueous media using novel iodate-assisted photochemical and photocatalytic systems.

The present study investigates the kinetics and mechanisms of carbamazepine (CBZ) degradation using a novel UV/iodate (IO3-) system for the first time and explores the influence of process conditions on its degradation. UV/IO3- showed high degradation efficiencies in a wide range of pHs, especially under neutral and acidic conditions, indicating that the system can be considered as a promising method to deal with effluents under various pH conditions. Radical scavenging experiments show that both iodine radicals (IO, IO2 and IO3) and hydroxyl radicals play an important role in CBZ degradation. Furthermore, the combination of UV/IO3- with TiO2 was studied to explore the potential of the addition of IO3- to improve the efficiency of the conventional TiO2 photocatalytic system. Scavenging experiments indicated that iodine radicals (IO, IO2 and IO3) were mainly involved in the degradation of CBZ in the UV/IO3-/TiO2 system, and the reaction mechanism equations were proposed for the first time for the studied UV/IO3-/TiO2 system. Several degradation products and four possible pathways of CBZ degradation were also elucidated using ultra-high-performance liquid chromatography in combination with a quadrupole time-of-flight mass spectrometer (Q-TOF MS). Respirometric tests indicated that the treatment has a positive impact on biomass behavior during subsequent biological purification, highlighting that the developed IO3--assisted AOPs are eco-friendly.