The dihydropyridines, represented by 1,4-DHPs, belong to a class of heterocyclic compounds very important for synthetic and medicinal organic chemistry because they have several pharmacological properties. Their derivatives act, mainly, as drugs for the treatment of cardiovascular diseases, acting as calcium channel blockers. The Hantzsch multicomponent reaction is the classical method for obtaining 1,4-DHPs. This reaction was first described in 1882 by Arthur Rudolf Hantzsch. He performed the cyclocondensation of β-ketoester, aldehyde and ammonia, obtaining, as a product, the symmetric 1,4-dihydropyridine, known as Hantzsch ester. However, the synthesis of 1,4-DHPs by means of the classical Hantzsch reaction presents long reaction times and low yields.

In this work, an alternative methodology for this multicomponent reaction was performed using a heteropolyacid, HPW, as heterogeneous catalyst and the green solvent PEG 400 under microwave heating. The method showed advantages such as: improved yield, short reaction times, selectivity for only one of the generally formed DHP isomers and a greener approach to the synthesis.

Thin Layer Chromatography (TLC) is a technique widely used in criminal forensic institutions due to simplicity, low cost and high analytical frequency. Nevertheless, it presents disadvantages regarding resolution, detectability, reproducibility and subjectivity in the interpretation of the results. To provide objective measurements of the signals, several chemometric tools can be used. In this sense, the present study aimed to evaluate the efficiency of the separation for the identification and quantification of phytocannabinoids contained in seized samples by the Polícia Civil do Distrito Federal (PCDF). Samples of pressed marijuana and cannabis sativa L. products, such as extracts, oils and resins, were packed in plastic microtubes and subjected to chloroform extraction and colorimetric reaction with azo Fast Blue B (FBBS) salt, according to previously defined protocol. Then, applications of the organic phase were made in TLC plates, which were later inserted inside glass vats containing moving phases for elution under the established chromatographic conditions for the TI and TJ systems. For comparative purposes, the coupling reactions of phytocannabinoids with FBBS were made before and after the applications. Images of the TLC plates were first obtained by the VSC8000® and then by a common scanner. The images were imported into Matlab®, where adjustments were made to the resolution, resulting in 210 chromatograms per plate (30 applications x 1 illumination x 7 layers of color). After noise reduction and baseline correction, the results of migration distance, area and peak height were obtained. Finally, the stecheometry of the phytocannabinoid coupling reaction was elucidated by Liquid Chromatography coupled to Time of Flight Mass Spectrometry (LC TripleTOF), which due to the high separation capacity and possibility of quantification after calibration, there is greater reliability in the identification. The results obtained with seized samples were compared with those obtained with analytical patterns of the main phytocannabinoids present in C. sativa: delta-9-tetrahydrocannabinol (THC), cannabinol (CBN), cannabidiol (CBD) and canabichromene (CBC) showed well-defined and reproducible peaks, while in the quantitative assay, significant regressions were obtained from serial dilutions, with R2 of 94.59% for the blue layer and 95.67% for the saturation layer. Furthermore, the characterization tests showed that although FBBS azo salt presents two diazonian groups, the diazoic coupling reaction occurs mostly at the ratio of 1:1 for THC, CBD and CBN and 2:1 for CBC. It was possible to distinguish these four phytocanbiniodes both in TLC and LC TripleTOF analyses

Water is the most important resource for the maintenance and survival of living beings. Access to quality water is also a human right. Water is used in various human activities, and therefore, it is essential that it be treated after use so that it returns to the environment in a suitable quality to prevent any harm to human health and/or cause ecological imbalance. Wastewater discharged into the sewage system is called wastewater. This wastewater goes to treatment plants to be properly cleaned of excess waste and residues that could contaminate and/or upset the environment. The efficiency and control of treatment are measured by laboratory analysis of different quality parameters. These analyses often require a lot of time and resources, both financial and human. For these reasons, research and technological innovations that can assist in the water quality control process, reducing analysis time and resource usage, are important. Among the most current technologies, chemical sensors appear to be very promising, as they are portable devices capable of detecting various chemical species of interest while being implemented in the water treatment process, enabling on-site and real-time monitoring. In this line, the present project proposes the development and evaluation of a multisensory system, an electronic tongue, for the quality control of wastewater treatment. The electronic tongue to be evaluated is composed of an array of nonspecific impedimetric sensors, consisting of interdigitated electrodes modified with ultra-thin films of various nanomaterials, an impedance analyzer, and one or more sample classification algorithms. The water samples were provided by the Environmental Sanitation Company of the Federal District (CAESB), obtained from different sewage treatment plants in the city of Brasília-DF and surrounding areas. A total of 212 samples from 7 different stations were analyzed, collected between the months of November and December 2022, and from January and from September to December 2023. It was found that the developed electronic tongue can distinguish, with fair efficiency, the samples according to the originating treatment station. With the classification algorithm, for four quality parameters: chemical oxygen demand (COD), total suspended solids (TSS), total nitrogen (TN), total phosphorus (TP), accuracy rates with the electronic tongue ranged from 40% to 60%.

Fingerprints are common and widely used evidence to help solve crimes. Estimating how long this trace was left at the crime scene would represent a huge gain for the criminal investigation. In view of the scarcity of a reliable analytical method to estimate the age of a fingerprint, the present study aims to evaluate the possibility of creating a robust, fast, reliable and non-destructive dating method using hyperspectral images in the near infrared (HSI-NIR), capable of estimating how long a latent fingerprint was left as a mark at a crime scene. To this end, two experimental designs were created. The first one, presented in Chapter 1, evaluated the degradation of fingerprints over a period of 18 days of aging and 13 volunteers participated in this study. The second one (Chapter 2) evaluated the effect of 8 weeks on the degradation of fingerprints of 6 volunteers. Time estimation, in both cases, was carried out using the chemometric classification technique Partial Least Squares Discriminant Analysis (PLS-DA). An individual aging pattern was created for each donor, used as a training set, since a universal aging pattern is not possible. In all cases, two test samples for each donor were used as a prediction set, with one time period being more recent and the other corresponding to a longer degradation time. In all cases, the percentage of pixels in the images that were correctly classified into the corresponding class was determined. Different time groupings were evaluated in all cases. In the first experimental design, 4 PLS-DA models were executed - which obtained a correct classification rate of prediction samples of 20.92%, 34.62%, 42.31% and 66.67%, respectively, and in the second one there were 2 models, with average correct percentages of 30.56% and 63.89%, respectively. Additionally, the most important spectral regions in class discrimination were evaluated using VIP Scores, noting the great importance of regions associated with fatty acids, amino acids, water, lipids, carbohydrates and their degradation products. Both approaches noted the great challenge of building an efficient dating model for each and every individual, given the great variability of factors that surround and interfere with the aging process of fingerprints.

Biodegradable polymers synthesized from renewable raw materials have become very attractive, due to the increasing number of inappropriately discarded materials with high degradation time. This way, poly(lactic acid) emerges as one of the most promising materials, being a product based on renewable raw material, biodegradable with properties similar to polymers based on fossil raw material, thus being able to replace them in different types of applications. This work aims at the synthesis of poly(L-lactic acid) from a process of direct polycondensation of lactic acid, evaluating the products obtained in the process with the use of organometallic catalysts. The analyzes of the obtained polymers were made through the acid index, infrared spectroscopy, viscosity, thermogravimetry (TGA) and differential scanning calorimetry (DSC) and the obtained results showed that there was a process of polyesterification of the monomers, obtaining different types of products in each reaction.

Concerns regarding personal health have been taking consumers (especially those with diabetes and other disease comorbidities that depend of dietary) to replace sugar with natural sweeteners. Among this group, stevia along with the polyols xylitol and erythritol stands on for its benefits such as cariostatic properties, satiety prolongation and antioxidant properties. With the increased demand and high market prices, these products are easy targets for adulteration, by illicit addition of low-cost artificial sweeteners in the product. Thus, the aim of this project is to determine natural sweeteners in tabletop commercial samples using near and mid infrared spectroscopy and a DD-SIMCA classification model in order to attest the authenticity of the products. The results showed the model was capable of classify correctly most of the validation samples (natural sweeteners), among these were samples of other classes and samples adulterated with saccharin, sucrose and acesulfame. The efficiency rate of the DD-SIMCA model reached 90% for most of training sets. The use of infrared spectroscopy combined with the DD-SIMCA algorithm provides a promising alternative to verification of authenticity of erythritol, stevia and xylitol.

Paper devices (PAD) are simple, low cost and use small amounts of reagent, the use of digital imaging to perform determinations in this medium is usually done through colorimetry, requiring multiple steps between signal acquisition and results. Aluminium chlorohydrate is an active ingredient of most commercially available antiperspirants, and the determination of aluminum is one way to perform quality control of these products, since the analyte is potentially harmful to human health. For this purpose, a fluorimeter capable of performing fluorescence determinations in an agile manner using PAD and digital images, with control and reading done by smartphone, using an app developed for this purpose, was developed. Aluminium chlorohydrate determinations were performed in commercial antiperspirant samples using the proposed instrument, and the results obtained were then compared with a reference method. The developed method for aluminum determinations was effective in determining the samples with errors less than 8.2% and no difference at the 95% confidence level for eight samples compared to the reference method.

Membrane-active peptides (MAPs) exhibit a broad biotechnological potential, both as alternatives to conventional drugs in combating multidrug-resistant microorganisms (MDR), and as carriers of conjugated drugs. However, their application relies on the selectivity with which these molecules recognize their targets, prokaryotic membranes, in relation to host membranes. Antimicrobial peptides (AMPs) and intragenic antimicrobial peptides (IAPs), the last-mentioned found in the primary structure of various proteins, are often poorly selective, being both antimicrobial and cytotoxic. The peptide Hs02 is an example of an IAP, consisting of 16 amino acid residues, a positive net charge, and an α-helix structure when adsorbed onto membranes. Nonetheless, while Hs02 is a potent and broad-spectrum antimicrobial agent, it exhibits moderate selectivity towards eukaryotic cells. With the purpose of enhancing this IAP’s selectivity, two generations of Hs02 analogs were synthesized, purified, and characterized. In one of these series, three shortened analogs were synthesized, followed by three others with point mutations at critical sites. The evaluation of their structures was carried out through Circular Dichroism spectroscopy, in the presence of large unilamellar vesicles (LUVs) with neutral and negative charges, as well as lipopolysaccharides (LPS). Furthermore, antimicrobial activities were tested against Gram-positive and Gram-negative bacteria, and cytotoxicities were measured using MTT cell viability assays with BV-2 microglial cells as a model. The obtained results demonstrated a significant increase in the therapeutic index (TI) of all second-generation analogs against the tested Gram-negative bacteria, particularly in the case of 16.3. Regarding Gram-positive bacteria, the developed analogs exhibited a decrease in their TI, suggesting that the cell wall composition of these bacteria might hinder the passage of cationic antimicrobial peptides. Despite this reduction, 16.3 showed an almost 5-fold increase in the Therapeutic Index against E. coli and P. aeruginosa, attributed to reduced cytotoxicity and maintained antimicrobial activity compared to the original peptide. The findings of this study indicate that the balance between Hs02's cationicity and minimal hydrophobicity was optimized, significantly altering the TI of this MAP.

In this master’s thesis, the synthesis of novel carba-isoflavonoids is described. They were formed through the reaction between an aryl bromide and a tetralone, enhancing the α-arylation method of tetralones using a palladium catalytic system with a bulky ligand, a strong base, and a solvent to establish a new (sp²―sp³) bond between aryl groups and the α position of tetralones. Microwave-assisted heating was employed to reduce the reaction time to 25 and 30 minutes. This represents a simple and efficient strategy for the formation of α-aryl carbonyl compounds. Through this work, 10 previously unreported compounds were obtained with varying yields (30-84%). A synthesis of carba-pterocarpens from the products of α-arylation of tetralones was achieved using the BBr3-promoted cyclization method. The results were promising as it proved to be a rapid and practical reaction for obtaining the desired products. The substances prepared in this study were spectroscopically characterized using ¹H NMR, ¹³C NMR, 2D-NMR, MS and IV (FTIR).

It is known that the use of plastic bags has made it more practical for moving goods, as well as storage, but polyethylene is obtained from petroleum, fossil and finite source. The use of plastic bags to collect domestic sink garbage does not allow organic matter to decompose correctly, leading to an increase in waste, in addition to the pollution already caused by the improper disposal of plastic. A possibly viable alternative proposal is the use of kraft paper, which is rich in cellulose and resistant, coated with a layer of latex, composed mainly of natural rubber (polyisoprene) and water and stabilized with an anticoagulant, preferably natural, such as tannin, which allows the temporary waterproofing of the container, allowing the bag to receive domestic organic waste, be stored for a limited time, and be disposed of for composting along with the food waste inside. Tests were carried out to determine the most adequate treatment, which would allow the best water repellency with the surface, but which would maintain the temporary characteristic of waterproofing. Then, it was determined the capacity to retain larger amounts of water of the treatments with the best results of water repellency. The temporarily waterproofed kraft paper was tested also for its possibility to develop microorganisms under appropriate conditions in petri dishes. Finally, the latex-tannin protected kraft paper samples were submitted to destructibility in a degrading environment based on ground earth with earthworms. Of the various treatments performed, the best result was kraft paper coated with two layers of 1% latex-tannin.

The current work reports the synthesis and investigation by physical-chemistry and spectroscopic methods of four dithiocarbazates HL¹, HL² H₂L³ and H₂L⁴ and eight novel metal complexes, which four complexes are derived from 2-acetylpyridine [Cu(L¹)Cl] (1), [Cu(L¹)Br] (2), [Cu(L²)Cl] (3) and [Cu(L²)Br] (4), explored in detail in section 1; and four complexes are derived from 1,1,1-trifluoro-2,4-pentanedione, [Ni(L³)PPh₃] (5), [Ni(L³)Py] (6), [Ni(L⁴)PPh₃] (7) and [Ni(L⁴)Py] (8), explored in detail in section 2. The crystal structures of Cu(II) complexes were elucidated by single crystal X-ray diffraction, which showed distorted square planar geometry to the metal centers, which tridentate ligands coordinated by NNS system and an additional halogen (Cl^- or Br^-) complete the coordination sphere. Mass spectrometry data indicated the presence of [Cu(L¹)(DMF)]⁺ and [Cu(L²)(DMF)]⁺ and characteristic fragmentation of the compounds. The Density Functional Theory (DFT) B97-3c method was used to optimize the geometries, and indicated that the results are in agreement with the experimental data. Hydrogen bonds and π-π stacking interactions were observed by the Hirshfeld surface, and the main interactions are attributed to the H---H contacts. In addition, the compounds were characterized by FT-IR and by ultraviolet–visible spectroscopy, which showed good agreement with the proposed structures. The biological activity of the compounds was evaluated against human glioma cells (U251) and the results revealed that free dithiocarbazates present high antitumor activity in vitro, which is increased after complexation with copper. The measurement of the cytotoxicity of the compounds also showed activity in a low concentration range, which indicates the high efficiency of the complexes as potential drugs. In section 2, the crystal structures of the H₂L³ and the (5)-(8) complexes are described after being elucidated by single crystal X-ray diffraction. Additionally, all compounds were characterized by a set of techniques that corroborate the proposed structures. In general, the results showed square planar geometry to the metal centers, which dithiocarbazates tridentate and a triphenylphosphine or pyridine molecule as coligand. The mass spectrometry data indicated the presence of the molecular ions [M+ H]⁺ and characteristic fragmentations of the compounds, which indicated the same behavior of the compounds as a solid and in solution. The ligands and complexes in this section were tested against four cell lines (MDA-MB-231 - breast cancer; U251 - glioma; NALM-6 and 697 - leukemia), furthermore, a molecular docking study was made to evaluate the predominant mode of binding of each compound with proteins of known 3D structures.

The present work concerns pioneer studies about the preparation of carbon-supported Mo-based catalysts and their use in the production of hydrocarbon biofuels through the hydroprocessing of lipidic feedstocks. A starting mesopore-rich activated carbon (AC) was prepared by chemical activation of dried coconut shell with H3PO4. This AC was used as obtained or after being treated with HNO3 or under an H2 atmosphere at elevated temperature, which permitted to increase or reduce the content of surface acidic groups, respectively. Preparation conditions were searched in order to give rise to catalysts with improved activity for the hydodeoxygenation (HDO) of lipidic feedstocks. HDO tests were carried out in a batch system having dodecanoic acid (a model compound) or coconut oil as substrate. Efficient catalysts were obtained, which promoted complete feedstock deoxygenation. The best results were verified for the catalyst prepared using the unmodified AC as support, adding Ni as activating agent, and depositing Mo and Ni successively from individual solutions by wet impregnation. It is valid to highlight that the prepared catalysts presented good stability in successive HDO cycles. Due to the low acidity of the carbon supports, the occurrence of isomerization reactions was not significant, so that the obtained biofuels were basically constituted by n-alkanes. The produced biofuels present a great potential to be used as drop-in substitutes for fossil fuels like diesel and jet fuel.

Cope and Claisen rearrangements are thermally activated [3,3] sigmatropic rearrangements that occur through cyclic six-membered transition states. Despite extensive investigation over the years, the magnetochemistry of these rearrangements still poses questions. In light of this, an investigation was conducted based on chemical-computational models to unravel the mechanistic and magnetochemical aspects of Cope and Claisen rearrangements using Density Functional Theory (DFT), CBS-QB3, Atoms in Molecules (AIM) theory, Multicenter Bond Order, Noncovalent Interaction (NCI), GIAO, and GIMIC methods. It was found that the Cope and Claisen rearrangements proceed via concerted mechanisms and cyclic six-membered aromatic transition states. The molar and anisotropic susceptibility (𝜒), NICS(1), and NICSzz exhibited a good correlation with J(nA.T-1), yet these descriptors proved insufficient to estimate the aromatic character of these transition states. Among the magnetic and topological descriptors employed in this study, magnetically induced current density (J) emerged as an excellent strategy for estimating the aromatic character of the transition states involved in the investigated rearrangements. Comparing the chair conformations of Cope rearrangement (RCO) and Claisen rearrangement (RCL), it was observed that greater aromatic character is associated with less-favored kinetics. For the chair conformations, RCL (TS2) exhibited 𝐽=8.63 nA.T-1, while RCO (TS4) showed 𝐽=10.43 nA.T-1. By utilizing the J(nA.T-1) index, the following aromatic character order was established: TS4 > TS2 > TS3 > TS1 for the analyzed structures, with boat conformations displaying high paramagnetic currents, which decrease the overall current in the ring. This result suggests that the aromatic character and thermodynamics in RCO and RCL cannot be directly correlated since the most stable transition states are not necessarily the most aromatic ones.

The synthesis of new selective fluorescent markers, with specific photophysical and biological properties, is highly desirable, and this area is growing more and more. In this sense, this work presents the synthesis of a fluorescent derivative called BTD-CHO, with the objective of being used in cellular bioimaging. The architecture of the derivative was intuitive to obtain a marker capable of selectively detecting hydrazine. The characterization of the compound controlled that the systemic occurred effectively. Photophysical studies appreciate large Stokes shifts and good stability in the excited state. The addition of hydrazine showed a shift in fluorescence from the green to the blue region, evidencing effective and selective labeling, even in the presence of other ions. Biological studies approved that the use of BTD-CHO can be performed both in living cells and in fixed cells, in addition to having been successfully employed in multicellular models, such as C. elegans and Zebrafish. Bioimaging revealed that the derivative was able to cross the cell membrane and selectively mark lipid bodies, showing strong fluorescence intensity, higher than that observed with the commercial marker Bodipy®, and also specificity for lipids that are not marked by the commercial marker. Furthermore, the derivative was superior to the commercial marker in terms of cytotoxicity and solubility in aqueous media. Such results demonstrate and consolidate the BTD core as an excellent and promising alternative for cellular bioimaging applications.

The presence of polycyclic aromatic hydrocarbons (PAHs) in firefighters' personal protective equipment is a concern. PAHs are compounds that have two or more condensed aromatic rings and are considered carcinogenic, mutagenic, and teratogenic. Several studies have shown that occupational exposure of these professionals to toxic compounds increases the risk of developing lung cancer, trachea, and leukemia, in addition to cardiovascular diseases. Some habits, such as adequate storage, washing method and washing frequency, can reduce this risk. With that in mind, we developed a form to assess attitudes and working practices of firefighters from Distrito Federal and Pernambucorelated to the presence of these pollutants in personal protective equipment. [iW1] The results show good awareness level from firefighters related to health attitudes, but, on the other hand, it was evidenced the lack of structure inside fire departments, making more difficult to adopt good habits that could reduce the exposure to carcinogenic compounds.

Besides that, we also evaluated the presence of PAHs deposited in firefighters' personal protective equipment during training exercise. To the best of our knowledge, this is the first study regarding the theme performed with Brazilian firefighters. Samples were collected after a firefighting training exercise and analysis were done with gas chromatography coupled to mass spectrometry. A set of 16 target compounds were searched and 8 were found in samples that were collected on the sleeve, chest, shoulder, and glove of firefighters’ suits. The highest concentration and variety of pollutants was found at an instructor, very probably due to the higher exposure frequency compared to those firefighters who carry out one-off combats.

Finally, we in we have evaluated some methods to decontaminating firefighters' fire-fighting suits. It is important to quote that it is mandatory that decontamination remove PAHs while maintaining the integrity of the suits. Photolysis with an ultraviolet and white lamp were evaluated, as well as the addition of an oxidizing agent (H₂O₂) to oxidize two probe PAHs. Regarding the fibers, damage was verified by scanning electron microscopy, optical microscopy and attenuated total reflection Fourier transform infrared spectroscopy. In addition, a tear resistance test (ASTM D2261) was performed on the outer layer fabric of the jacket, before and after treatment.

Best result was achieved using white lamp photolysis (WLP), when it was possible to remove more than 83% of the pyrene and 9-methylanthracene previously deposited on the fabric outer layer firefighting suit. The WLP was compared with the most used techniques (washing in a washing machine and brushing with soap and water) and showed better results. The experiment was also carried out on real samples, with the three layers of the fabric. Photolysis presented a superior result to the other techniques, removing 73% of PAHs, while washing and brushing promoted a reduction of 44% and 32%, respectively.

With this, it was possible to develop a simple and inexpensive method of decontamination based on an advanced oxidative process, photolysis, which is effective in the degradation of PAHs, in addition to maintaining the integrity of Kevlar® and Nomex® fibers.

Some proteins may present internal segments with shape and physical-chemical properties compatible with antimicrobial peptides (AMPs), such that, when synthesized as individual entities, present similar biological activities. Normally, the prospection of those is done based on helical AMPs, which results in the identification of protein segments compatible with cationic and amphiphilic α-helices. The present work explores a new form for encrypted antimicrobial peptides in proteins based on cationic and amphiphilic β-sheets and aims to evaluate their activity as antibacterial and anti-inflammatory molecules. For such, 6 putative short intragenic antimicrobial peptides (IAPs) were prospected from the human proteome by the software Kamal, and 4 of these were synthetized and purified with success, named Hs11, Hs12, Hs13, Hs14. One control peptide, the HsCtrl1, also was synthetized. Peptides Hs11 and Hs12 presented wide antimicrobial activity against bacteria, presenting a minimum inhibitory concentration between 8 and 128 µM. Furthermore, it was verified that none of the tested peptides reduced the cellular viability of murine macrophages (BMDM) after 24 hours of incubation in concentrations up to 100 µM. All the peptides reduced, in different ways, the biogenesis of lipid bodies in BMDM treated or not with lipopolysaccharides (LPS). Hs11, Hs12, Hs13 and Hs14 decreased the secretion of TNF-α and IL-6 in BMDM not stimulated with

LPS. The peptide Hs11 reduced the secretion of TNF-α to basal levels in cells that were previously incubated with LPS. At last, nuclear magnetic resonance studies were conducted to investigate the interaction between LPS aggregates and the peptide Hs11. In conclusion, short amphiphilic peptides ranging from 8 to 9 amino acid residues are interesting new small molecules, as they present antibacterial and anti-inflammatory activity that is comparable to previously known AMPs. The methodology described herein has the potential to generate alternatives in the combat of bacterial infections caused by resistant microorganisms, which are expected to victimize more people than cancer per year by 2050.

Micro-pollutants of emerging concern have imposed a major technological challenge: pesticides, drugs and other anthropogenic substances are increasingly found in aquatic and atmospheric environments and even in water supplies, being related to adverse effects on biota and human health. Overcoming this challenge requires understanding the behavior of these species in the environment and the development of technologies that allows for minimizing their dissemination. Viable alternatives applied in this thesis include the use of radical-based oxidation processes using both experimental – via the competition kinetics method – and theoretical protocols – blend of kinetic, quantum chemistry and machine learning calculations. In a first study, the mechanisms, kinetics, and an evaluation of the toxicity of picloram degradation – a pesticide widely used in the world - initiated by OH radicals indicate that: i) two favorable pathways occur by addition to the pyridine ring, ii) picloram and the majority of degradation products are estimated as harmful; however, ii) these compounds can suffer photolysis by sunlight. However, the competition kinetic method and the quantum chemistry description make the degradation analyses a formidable enterprise, considering the costs of ad hoc instrumental equipment’s and dedicated computational efforts. To overcome the demanding conventional procedures, we developed a free and user-friendly web application (www.pysirc.com.br) based on holistic machine learning combined with molecular fingerprints models that permits compilation of kinetic parameters and mechanistic interpretation of radical-based oxidation attacks according to the OECD principles. Machine learning algorithms were implemented, and all models provided high goodness-of-fit for radical-based degradation in aquatic and atmospheric environment. The models were interpreted using the SHAP (SHapley Additive exPlanations) method: the results showed that the model developed made the prediction based on a reasonable understanding of how electron-withdrawing/donating groups interfere in the reactivity of the radicals. We argue that our models and web interface can stimulate and expand the application and interpretation of kinetic research on contaminants in water and air treatment units based on advanced oxidative technologies.

Statistical learning models that use machine learning methods like GSA have been used for decades in chemistry for a variety of tasks. With these techniques, our research groups are mainly dedicated to the kinetic and mechanistic study of reactive and non-reactive processes allied to the development of new models and techniques for carrying out these researches. With the current increase in the processing capacity of computers and the advent of sophisticated methods of machine learning, such as deep learning, it has become increasingly necessary to create new analysis techniques as well as the development of platforms and tools that systematize these techniques. Accordingly, the present work makes three proposals for theoretical models that will be presentend with the development of codes for their use, with graphical and intuitive interfaces: 𝑖) We propose here the transitivity plot: a new perspective under Arrhenius plot that is able to linearize non-Arrhenius curvatures and thus observe regime changes. Its application was able to linearize the curvature of the propylene carbonate relaxation process and revealed a break in the regime at the temperature of 198K, with changes in phenomenological parameters such as the barrier height. In addition, the Transitivity code systematizes tools developed over the last 10 years to deal with the kinetics of non-Arrhenius processes, providing options for phenomenological estimation of the reaction kinetic constant in liquid and gas phases and perform model fitting with experimental kinetic data using GSA. As an example of its application, a publication on the free energy surface topography of the Claisen-Schmidt reaction was carried out and will be shown in this work. 𝑖𝑖) we also propose the generalized model of the temperature coefficient Qd10, used to evaluate the degree of dependence of biological processes as a function of temperature. The model was applied to a series of biological processes present in recent literature and unified the conflicting results obtained by the different researchers: the quality of the fits reinforced the initial hypothesis that the temperature coefficient must be described using deformed exponentials (Qd10) instead of the usual Arrhenius model (Q10). Furthermore, the standard model Q10 is a particular case of the proposed model, as it is retrieved when the deformed parameter is equal to zero, thus validating its general justification. The Qd10-GSA interface was developed and it estimates parameters related to the Arrhenius and AquilantiMundim formulations from experimental data, thus drawing a parallel between the usual model Q10 and the model proposed here Qd10. 𝑖𝑖𝑖) We also propose deep learning models that use artificial neural networks and learn from a spectral database, obtained from NIST, to make inferences about structural properties of molecules and perform spectra reconstructions from molecular featurizers. CNN ResNet34 and the proposed tabular model performed well in the classification of molecular structural properties using infrared and mass spectra, with F1-Score values between 0.83 and 0.88 and accuracies of 96%. The reconstruction models showed a mean square error of 0.005 and 0.007 in the reconstruction of the mass and infrared spectra, respectively. Thus, it is concluded that the development of these models suggests that neural networks are suitable for the search for new spectral analysis techniques and represent a step towards the creation of autonomous analysis techniques. Finally, the codes developed here present very intuitive and user-friendly interfaces in order to disseminate the proposed new concepts and the authors hope that they can also be used as powerful research tools and didactic tools for teaching thermodynamics, process kinetics and related issues.