Search

Due to the complexity and diversity of polyolefinic plastic waste streams and the inherent non-selective nature of the pyrolysis chemistry, the chemical decomposition of plastic waste is still not fully understood. Accurate data of feedstock and products that also consider impurities is, in this context, quite scarce. Therefore this work focuses on the thermochemical recycling via pyrolysis of different virgin and contaminated waste-derived polyolefin feedstocks (i.e., low-density polyethylene (LDPE), polypropylene (PP) as main components), along with an investigation of the decomposition mechanisms based on the detailed composition of the pyrolysis oils. Crucial in this work is the detailed chemical analysis of the resulting pyrolysis oils by comprehensive two-dimensional gas chromatography (GC × GC) and ICP-OES, among others. Different feedstocks were pyrolyzed at a temperature range of 430–490 °C and at pressures between 0.1 and 2 bar in a continuous pilot-scale pyrolysis unit. At the lowest pressure, the pyrolysis oil yield of the studied polyolefins reached up to 95 wt%. The pyrolysis oil consists of primarily α-olefins (37–42 %) and n-paraffins (32–35 %) for LDPE pyrolysis, while isoolefins (mostly C9 and C15) and diolefins accounted for 84–91 % of the PP-based pyrolysis oils. The post-consumer waste feedstocks led to significantly less pyrolysis oil yields and more char formation compared to their virgin equivalents. It was found that plastic aging, polyvinyl chloride (PVC) (3 wt%), and metal contamination were the main causes of char formation during the pyrolysis of polyolefin waste (4.9 wt%).

Despite the increasing number of electrified vehicles the transportation system still largely depends on the use of fossil fuels. One way to more rapidly reduce the dependency on fossil fuels in transport is to replace them with biofuels. Evaluating the potential of different biofuels in different applications requires knowledge of their physicochemical properties. In chemistry, message passing neural networks (MPNNs) correlating the atoms and bonds of a molecule to properties have shown promising results in predicting the properties of individual chemical components. In this article a machine learning approach, developed from the message passing neural network called Chemprop, is evaluated for the prediction of multiple properties of organic molecules (containing carbon, nitrogen, oxygen and hydrogen). A novel approach using transfer learning based on estimated property values from theoretical estimation methods is applied. Moreover, the effect of multi-task learning (MTL) on the predictions of fuel properties is evaluated. The result show that both transfer learning and multi-task learning are good strategies to improve the accuracy of the predicted values, and that accurate predictions for multiple fuel properties can be obtained using this approach.

We present a group contribution method (SoluteGC) and a machine learning model (SoluteML) to predict the Abraham solute parameters, as well as a machine learning model (DirectML) to predict solvation free energy and enthalpy at 298 K. The proposed group contribution method uses atom-centered functional groups with corrections for ring and polycyclic strain while the machine learning models adopt a directed message passing neural network. The solute parameters predicted from SoluteGC and SoluteML are used to calculate solvation energy and enthalpy via linear free energy relationships. Extensive data sets containing 8366 solute parameters, 20,253 solvation free energies, and 6322 solvation enthalpies are compiled in this work to train the models. The three models are each evaluated on the same test sets using both random and substructure-based solute splits for solvation energy and enthalpy predictions. The results show that the DirectML model is superior to the SoluteML and SoluteGC models for both predictions and can provide accuracy comparable to that of advanced quantum chemistry methods. Yet, even though the DirectML model performs better in general, all three models are useful for various purposes. Uncertain predicted values can be identified by comparing the three models, and when the 3 models are combined together, they can provide even more accurate predictions than any one of them individually. Finally, we present our compiled solute parameter, solvation energy, and solvation enthalpy databases (SoluteDB, dGsolvDBx, dHsolvDB) and provide public access to our final prediction models through a simple web-based tool, software packages, and source code. ispartof: JOURNAL OF CHEMICAL INFORMATION AND MODELING vol:62 issue:3 pages:433-446 ispartof: location:United States status: published

The solubility of organic molecules is crucial in organic synthesis and industrial chemistry; it is important in the design of many phase separation and purification units, and it controls the migration of many species into the environment. To decide which solvents and temperatures can be used in the design of new processes, trial and error is often used, as the choice is restricted by unknown solid solubility limits. Here, we present a fast and convenient computational method for estimating the solubility of solid neutral organic molecules in water and many organic solvents for a broad range of temperatures. The model is developed by combining fundamental thermodynamic equations with machine learning models for solvation free energy, solvation enthalpy, Abraham solute parameters, and aqueous solid solubility at 298 K. We provide free open-source and online tools for the prediction of solid solubility limits and a curated data collection (SolProp) that includes more than 5000 experimental solid solubility values for validation of the model. The model predictions are accurate for aqueous systems and for a huge range of organic solvents up to 550 K or higher. Methods to further improve solid solubility predictions by providing experimental data on the solute of interest in another solvent, or on the solute's sublimation enthalpy, are also presented. Pre-print: 10.26434/chemrxiv-2022-92hl1-v2 ispartof: JOURNAL OF THE AMERICAN CHEMICAL SOCIETY vol:144 issue:24 pages:10785-10797 ispartof: location:United States status: published

Furans are an important class of compounds that can be thermochemical or enzymatically produced from biomass. Despite of their importance little is known about the thermal decomposition of furans with oxygenated substituents. In this work, the influence of the -CH3, -CH2OH and -CHO functional groups on the molecular and radical decomposition chemistry is studied with a combined quantum chemical and experimental approach using 2-furfuryl alcohol and 5-methyl furfural as model components. The quantum chemistry calculations show that both reactants can decompose by a ring-opening isomerization reaction and through carbene intermediates. The latter are formed by the shift of a hydrogen atom or a -CHO functional group within the furan ring structure. The -CHO functional group on the furan ring structure accelerates the molecular ring-opening isomerization reaction, while it suppresses carbene formation channels compared to other functional groups. The weaker C H and C O bonds in 2-furfuryl alcohol and 5-methyl furfural compared to furan and furfural respectively result in a higher importance of radical chemistry that cannot be neglected. This is confirmed experimentally by analyzing the product spectrum with molecular beam synchrotron VUV photoionization mass spectrometry at a pressure of 0.04 bar and for temperatures between 923 K to 1223 K for 2-furfuryl alcohol and 973 K to 1273 K for 5-methyl furfural. For both reactants several radical intermediates are observed starting from 923 K for 2-furfuryl alcohol and from 973 K for 5-methyl furfural. Examples of measured radicals are those initial formed from the reactant by a C H homolytic bond scission and methyl, allyl, propargyl, 1,2-butadiene-4-yl, 2-furanyl-methyl, 2,5-dihydrofuran-2-yl and 1‑hydroxyl-2-furanyl-methyl radicals.

A prerequisite for the generation of detailed fundamental kinetic models is the availability of accurate thermodynamic properties. To address the scarcity of accurate experimental data, theoretical calculations can be used. The accuracy of these quantum chemistry methods for determination of thermodynamic properties can be improved by making use of empirical correction methods, such as the isodesmic bond additivity correction (BAC) method. In this work, ab initio calculations for a set of 371 molecules have been performed to determine a new set of BACs for the CBS-QB3 level of theory. For each of these molecules also accurate experimental data, that is with an experimental uncertainty less than 3 kJ mol(-1), is available. This broad dataset of hydrocarbons and heteroatomic compounds contains (non)cyclic molecules with a wide range of functional groups consisting of hydrogen, carbon, oxygen, nitrogen, and sulfur. The new set of 26 BAC parameters is obtained via linear regression analysis, minimizing the differences between experimental and corrected CBS-QB3 values. The CBS-QB3 method combined with BACs succeeds in approximating the experimental standard enthalpy of formation at 298 K with an accuracy of 4 kJ mol(-1) for almost all species. The BACs reduce the mean absolute deviation for the complete dataset from 5.65 to 2.37 kJ mol(-1), corresponding to a decrease of the root mean square deviation from 6.95 to 3.00 kJ mol(-1).

International audience; 3-Carene is an important potential biofuel with properties similar to the jet-propellant JP-10. Its thermal decomposition and combustion behavior is to date unknown, which is essential to assess its quality as a fuel. A combined experimental and kinetic modeling study has been conducted to understand the initial decomposition of 3-carene. The pyrolysis of 3-carene was investigated in a jet-stirred quartz reactor at atmospheric pressure, at temperatures varying from 650 to 1050 K, covering the complete conversion range. The decomposition of 3-carene was observed to start around 800 K, and it is almost complete at 970 K. Online gas chromatography shows that primarily aromatics are generated which suggests that 3-carene is not a good fuel candidate. The potential energy surface for the initial decomposition pathways determined by KinBot shows that a hydrogen elimination reaction dominates, giving primarily cara-2,4-diene. Next to this molecular pathway, radical pathways lead to aromatics via ring opening. The kinetic model was automatically generated with Genesys and consists of 2565 species and 9331 reactions. New quantum chemical calculations at the CBS-QB3 level of theory were needed to calculate rate coefficients and thermodynamic properties relevant for the primary decomposition of 3-carene. Both the conversion of 3-carene and the yields of the primary products (ie, benzene and hydrogen gas) are well predicted with this kinetic model. Rate of production analyses shows that the dominant pathways to convert 3-carene are hydrogen elimination reaction and radical chemistry.

ispartof: Journal Of Co2 Utilization vol:70 status: published

Industrial-scale application of end-of-life plastic pyrolysis still faces significant problems, such as a lack of detailed knowledge on degradation mechanisms, parameters affecting the degradation, and formation pathways of the primary pyrolysis products. Today, the degradation mechanisms based on radical chain scission are insufficiently understood to explain the pyrolysis chemistry from feedstock-to-product comprehensively. In this study, the impact of operating conditions on the degradation mechanism is evaluated by pyrolyzing end-of-life and virgin high-density polyethylene (HDPE) in a continuous pilot-scale unit at temperature and pressure ranges of 450–504 ºC and 0.1–2 bara. The pyrolysis products were analyzed based on the detailed product composition obtained using comprehensive two-dimensional gas chromatography (GC×GC). A simplified kinetic mechanism was proposed to describe the main production pathways of the various components by considering the weakest points along the polymer chain. The results showed that the chain-end scission mechanism is the main mechanism in the HDPE pyrolysis process, even at low temperatures and pressures in the studied ranges. The pyrolysis of virgin HDPE under sub-atmospheric pressure, 0.1 bara, at 464 ºC reactor temperature, yields the highest concentration of linear hydrocarbons in the pyrolysis oil (93.2 wt%). At higher pressure and temperature, the cyclic and branched hydrocarbons had a higher share of up to 17.4 wt% compared to 6.8 wt% at vacuum pressure and lower temperature. Interestingly, the pyrolysis of end-of-life HDPE at atmospheric pressure and 450 ºC led to more cyclic and branched hydrocarbons (sum: 22.1 wt%), as opposed to that of virgin HDPE which is more prone to the production of linear hydrocarbons at the studied conditions. Regarding the additives and contaminants, a large amount of different metals and halogen atoms in the ppm range were detected in end-of-life HDPE, of which a small amount was still found in the pyrolysis oil.

We present a group contribution method (SoluteGC) and a machine learning model (SoluteML) to predict the Abraham solute parameters, as well as a machine learning model (DirectML) to predict solvation free energy and enthalpy at 298 K. The proposed group contribution method uses atom-centered functional groups with corrections for ring and polycyclic strain whilst the machine learning models adopt a directed message passing neural network. The solute parameters predicted from SoluteGC and SoluteML are used to calculate solvation energy and enthalpy via linear free energy relationships. Extensive data sets containing 8366 solute parameters, 20253 solvation free energies, and 6322 solvation enthalpies are compiled in this work to train the models. The three models are each evaluated on the same test sets using both random and substructure-based solute splits for solvation energy and enthalpy predictions. The results show that the DirectML model is superior to the SoluteML and SoluteGC models for both predictions and can provide accuracy comparable to that of advanced quantum chemistry methods. Yet, even though the DirectML model performs better in general, all three models are useful for various purposes. Uncertain predicted values can be identified by comparing the 3 models, and when the 3 models are combined together, they can provide even more accurate predictions than any one of them individually. Finally, we present our compiled solute parameter, solvation energy, and solvation enthalpy databases (SoluteDB, dGsolvDBx, dHsolvDB) and provide public access to our final prediction models through a simple web-based tool, software package, and source code.

Automatically generated kinetic networks are ideally validated against a large set of accurate, reproducible, and easy-to-model experimental data. However, although this might seem simple, it proves to be quite challenging. QUANTIS, a publicly available Python package, is specifically developed to evaluate both the precision and accuracy of experimental data and to ensure a uniform, quick processing, and storage strategy that enables automated comparison of developed kinetic models. The precision is investigated with two clustering techniques, PCA and t-SNE, whereas the accuracy is probed with checks for the conservation laws. First, the developed tool processes, evaluates, and stores experimental yield data automatically. All data belonging to a given experiment, both unprocessed and processed, are stored in the form of an HDF5 container. The demonstration of QUANTIS on three different pyrolysis cases showed that it can help in identifying and overcoming instabilities in experimental datasets, reduce mass and molar balance closure discrepancies, and, by evaluating the visualized correlation matrices, increase understanding in the underlying reaction pathways. Inclusion of all experimental data in the HDF5 file makes it possible to automate simulating the experiment with CHEMKIN. Because of the employed InChI string identifiers for molecules, it is possible to automate the comparison experiment/simulation. QUANTIS and the concepts demonstrated therein is a potentially useful tool for data quality assessment, kinetic model validation, and refinement.

Aliphatic amines are an important class of nitrogen-containing compounds present in renewable fuels such as bio-oils. Conversion of this fuel-bound nitrogen can lead to the formation of HCN and NH3, as well as NOx emissions. In this work, the combustion chemistry of small aliphatic amines is investigated via a combination of quantum chemical calculations, chemical kinetic modeling and experimental validation. The influence of the degree of substitution on the nitrogen atom and the alkyl chain length on the reactivity and product distribution is studied via three model compounds, i.e. ethylamine (EA, CH3CH2NH2), dimethylamine (DMA, (CH3)2NH) and diethylamine (DEA, (CH3CH2)2NH). A detailed kinetic model containing 258 species and 2274 reactions is developed to describe their combustion over a wide range of conditions. The proposed model captures the trends in ignition delay times and species concentrations over a temperature range from 500 to 2000 K and pressures from 4 to 170 kPa. The ignition delay data are predicted with an average deviation of 10%. The difference between experimental and simulated species concentrations from laminar premixed flames is for the major species on average 50%, while the agreement for the JSR is even better, with an average deviation of 10%. The dominant decomposition pathway under all the studied conditions is a set of hydrogen abstractions from the Cα and N positions followed by β-scission of the fuel radicals. Among the unimolecular decomposition pathways, the homolytic C C and C N scissions and four-centered elimination play a minor role. HCN is the main intermediate and at temperatures above 1100 K the amines are completely converted to N2 and NO.

Computer-aided synthesis has received much attention in recent years. It is a challenging topic in itself, due to the high dimensionality of chemical and reaction space. It becomes even more challenging when the aim is to suggest syntheses that can be performed in continuous flow. Though continuous flow offers many potential benefits, not all reactions are suited to be operated continuously. In this work, three machine learning models have been developed to provide an assessment of whether a given reaction may benefit from continuous operation, what the likelihood of success in continuous flow is for a certain set of reaction components (i.e. reactants, reagents, solvents, catalysts, and products) and, if the likelihood of success is low, which alternative reaction components can be considered. The first model uses an abstract version of a reaction template, obtained via gaussian mixture modelling, to quantify its relative increase in publishing frequency in continuous flow, without relying on potentially ambiguously defined reaction templates. The second model is an artificial neural network that categorizes feasible and infeasible reaction components with a 75 % success rate. A set of reaction components is considered to be feasible if there is an explicit reference to it being used in continuous synthesis in the database; all other reaction components are considered infeasible. While several cases that are ‘infeasible’ by this definition, are classified as feasible by the neural network, further analysis shows that for many of these cases, it is at least plausible that they are in fact feasible – they simply have not been tested to (dis)prove this. The final model suggests alternative continuous flow components with a top-1 accuracy of 95%. Combined, they offer a black-box evaluation of whether a reaction and a set of reaction components can be considered promising for continuous syntheses.

International audience; The thermal decomposition of furfural is investigated experimentally and through theoretical calculations at the CBS-QB3 level of theory. Furfural is a major product observed during biomass pyrolysis, but despite its importance there are many speculations about the thermal decomposition channels of this compound. To address these open questions new experiments are performed in a jet-stirred reactor at atmospheric pressure and temperatures ranging from 900 to 1100 K with a furfural inlet mole fraction of 0.005 and He as diluent. The residence time is set to 2 s. The main products observed by GC analysis are CO, CO2, α-pyrone, furan, 3-furaldehyde, propyne, propadiene, acetylene, methane and benzene. Small amounts of other aromatics, e.g. toluene, styrene, benzaldehyde and phenol are observed as well. Theoretical calculations at the CBS-QB3 level are used to extend the furfural potential energy surface and to identify possible reaction pathways to the observed products. The unimolecular non-radical decomposition channel through α-pyrone as proposed in literature is confirmed as the main channel, but carbene pathways are found to make small contributions as well. Furthermore, pericyclic reactions are suggested to contribute to the molecular elimination of CO in open-chain molecules and Diels Alder reactions are found to be important for the formation of CO2 and aromatics. Finally, even radical chemistry initiated by homolytic scission of the approximately 380 kJ/mol strong C-H bond in the furfural carbonyl group has a non-negligible influence.

The pyrolysis of undiluted cyclohexane has been studied in a continuous flow tubular reactor at temperatures from 913 to 1073 K and inlet feed flow rates in the range 288–304 g·h–1 at 0.17 MPa reactor pressure with average reactor residence time of 0.5 s calculated based on the pressure in the reactor, the temperature profile along the reactor, and the molar flow rate along the reactor estimated by the logarithmic average of the inlet and outlet molar flows. The reactions lead to conversions between 2% and 95%. Forty-nine products were identified and quantified using two-dimensional gas chromatography equipped with thermal conductivity and flame ionization detectors. The products with molecular weights between those of hydrogen and naphthalene constitute more than 99 mass % of the total products. A kinetic mechanism composed exclusively of elementary step reactions with high pressure limit rate coefficients has been generated with the automatic network generation tool “Genesys”. The kinetic parameters for...

International audience; The pyrolysis and low-to intermediate temperature oxidation chemistry of dimethoxymethane (DMM), the simplest oxymethylene ether, is studied theoretically and experimentally in a JSR setup. The potential energy surfaces for peroxy species relevant during the low-temperature oxidation of dimethoxymethane are studied at the CBS-QB3 level of theory and the results are used to calculate thermodynamic properties of the main species as well as rate expressions for important reactions. An elementary step model for DMM pyrolysis and oxidation is built with the automatic kinetic model generation software Genesys. To describe the chemistry of small species not directly related to DMM, the AramcoMech 1.3 mechanism developed by Metcalfe et al. is used. If the more recently extended version of this mechanism, i.e. the propene oxidation mechanism published by Burke et al., was used as alternative base mechanism, large discrepancies for the mole fractions of CO2, methyl formate and methanol during the pyrolysis of DMM were observed. The validation of the new DMM model is carried out with new experimental data that is acquired in an isothermal quartz jet-stirred reactor at low and intermediate temperatures. Different equivalence ratios, = 0.25, = 1.0, = 2.0 and = ∞, are studied in a temperature range from 500 K to 1100 K, at a pressure of 1.07 bar and with an inlet DMM mole fraction of 0.01. The experimental trends are well predicted by the model without any tuning of the model parameters although some improvements are possible to increase quantitative agreement. The largest discrepancies are observed at fuel lean conditions for the hydrocarbon mole fractions, and at low-temperatures as can be noticed by the over prediction of formaldehyde and methyl formate. The kinetic model is also validated against plug flow reactor, jet-stirred reactor and lean and rich premixed flames data from the literature. Rate of production analyses are performed to identify important pathways for low-and intermediate-temperature oxidation and pyrolysis.

A kinetic modeling study is presented for batch nitroxide mediated polymerization (NMP) of methyl methacrylate(MMA; nitroxide: N-tert-butyl-N-[1-diethylphosphono-(2,2-dimethylpropyl)] (SG1)). Arrhenius parameters for SG1 dis-proportionation (A 5 1.4 107Lmol21s21;Ea5 23 kJ mol21) are reported, based on homopolymerization data account-ing for unavoidable temperature variations with increasing time, that is, nonisothermicity. For low targeted chain lengths(TCLs  300), this nonisothermicity is also relevant for NM P of MMA with a small amount of styrene. Parameter tuningto copolymerization data confirms a penultimate monomer unit effect for activation (sa25 ka12/ka2256.7; 363 K;1: MMA; 2: styrene). To obtain, for a broad TCL range (up to 800), a dispersity well below 1.3 an initial styrene massfraction of ca. 10% is required. An interpretation of the comonomer incorporation is performed by calculating the frac-tions of activation-growth-deactivation cycles with a given amount of monomer units and the copolymer compositiondistribution.

The pyrolysis of cyclopentane is not well established although it is an abundant compound in typical naphtha feedstocks and can be considered a model compound for cyclic fuels. The studies in literature so far have focused primarily on the initial decomposition of cyclopentane in shock tubes. This article therefore explores the pyrolysis of cyclopentane in a continuous flow tubular reactor with pure cyclopentane feed at reactor conditions 0.17 MPa, 973–1073 K, and a residence time of 0.5s. Conversions of 5% to 75% were realized while the product concentrations were quantified using two dimensional gas chromatography. A mechanism composed of elementary high pressure limit reactions has been generated using the automatic network generation tool “Genesys”. Kinetics of the reactions originate from high level ab-initio calculations and new group additive values derived from ab-initio kinetic data in literature. Overall the Genesys model outperforms the models available in literature and there is a good agreement between model calculated mass fraction profiles and experimental data for 22 products ranging from hydrogen to naphthalene without any adjustments of the kinetic parameters. Reaction path analysis reveals that cyclopentane consumption is initiated by the unimolecular isomerization to 1-pentene, but overall dominated by hydrogen abstraction reactions by allyl radicals and hydrogen atoms to give cyclopentyl radicals, whose ring opening and further scissions lead to smaller molecules. Dominant routes for the major products are discussed.

International audience; Resonantly stabilized radicals play an important role in the formation of aromatics. In this work, the pyrolysis (φ = ∞) and oxidation (φ = 1 and 2) of 1,5-hexadiene, diluted in He, has been studied experimentally in a jet-stirred reactor at atmospheric pressure. The temperature was varied between 500 and 1100 K and the residence time was fixed at 2 s. Gas chromatography was used to determine the reactor effluent composition. The dedicated analysis section allowed the identification and quantification of many hydrocarbon and oxygenated product species up to naphthalene. The pyrolysis of 1,5-hexadiene results in the formation of small alkenes and cyclic hydrocarbons, with a particularly high selectivity towards 1,3-cyclopentadiene and benzene. In the presence of molecular oxygen, various oxygenated intermediates, including acrolein, prop-2-en-1-ol and but-3-enyl-oxirane, were detected in the outlet gases, besides the pyrolysis products. A detailed kinetic model was developed, mainly with an automatic network generation tool, to simulate and interpret the performed experiments. The kinetic model includes molecular weight growth chemistry to predict mole fractions of the main aromatic species. Model calculated and experimental mole fraction profiles are in relatively good agreement. At low-temperature pyrolysis conditions, 1,5-hexadiene is in quasi-equilibrium with allyl radicals. Hydrogen abstraction from 1,5-hexadiene by allyl radicals has the strongest effect on conversion. The resulting hexa-2,5-dien-1-yl radical can react by intramolecular radical addition and eventually form 1,3-cyclopentadiene and benzene. Recombination of cyclopentadienyl with alkyl radicals followed by hydrogen abstraction and ring enlargement is an important formation path to aromatics. At oxidizing conditions, the pyrolysis reaction pathways are in competition with reactions involving hydroxyl and hydroperoxy radicals, as well as molecular oxygen. Above 900 K, 1,5-hexadiene is mainly consumed by C-C scission. The conversion and product distribution in 1,5-hexadiene oxidation are found to be sensitive to the branching ration of the reactions of allyl with hydroperoxy radicals. Formation of hydroxyl and allyloxy radicals increases the reactivity while the propene and molecular oxygen channel decreases the number of radicals in the system.

Data scarcity, bias, and experimental noise are all frequently encountered problems in the application of deep learning to chemical and material science disciplines. Transfer learning has proven effective in compensating for the lack in data. The use of quantum calculations in machine learning enables the generation of a diverse dataset and ensures that learning is less affected by noise inherent to experimental databases. In this work, we propose a transfer learning approach for the prediction of solvation free energies that combines fundamentals from quantum calculations with the higher accuracy of experimental measurements. The employed model architecture is based on the directed-message passing neural network for the molecular embedding of solvent and solute molecules. A significant advantage of models pre-trained on quantum calculations is demonstrated for small experimental datasets and for out-of-sample predictions. The improved out-of-sample performance is shown for new solvents, for new solute elements, and for the extension to higher molar mass solutes. The overall performance of the pre-trained models is limited by the noise in the experimental test data, known as the aleatoric uncertainty. On a random test split, a mean absolute error of 0.21 kcal/mol is achieved. This is a significant improvement compared to the mean absolute error of the quantum calculations (0.40 kcal/mol). The error can be further reduced to 0.09 kcal/mol if the model performance is assessed on a more accurate subset of the experimental data., Comment: Submitted to Chemical Engineering Journal, 27 Nov 2020

The presence of large amounts of oxygen in the molecular structure of triglyceride and fatty acid based feedstocks makes direct use in conventional steam crackers impossible without substantial modifications to the cold section. Full or partial catalytic deoxygenation has potential to resolve this, giving a mixture which consists primarily of normal and branched alkanes. Two of these deoxygenated mixtures have been investigated theoretically and experimentally in a dedicated bench setup (P = 0.17 MPa, T = 1050–1150 K, FHC = 4.17 10−2 g s−1, steam dilution of 0.3 and 0.5 gH2O/gHC). Furthermore, the degree of branching of the hydrocarbon mixtures impacts the product distribution, in particular the alkene selectivity. The newly generated, validated detailed kinetic model shows that small alkenes are formed by hydrogen abstraction and successive C C β-scission reactions. In the studied temperature range mono-aromatics are formed by three competing pathways: a series of recombination reactions of allylic radicals followed by hydrogen abstraction and intramolecular radical additions, additions of allylic and vinyl radicals on dienes followed by intramolecular radical addition, and finally recombination reactions of carbon-centered radicals with 1,3-cyclopentadienyl followed by hydrogen abstraction and ring enlargement.

International audience; The pyrolysis and oxidation chemistry of diethylamine (DEA), a nitrogen-containing bio-oil model compound, is investigated theoretically and experimentally at low to intermediate temperatures. The pyrolysis of DEA is studied using two experimental units, i.e. a jet-stirred reactor and a tubular reactor. Oxidation experiments are performed at three different equivalence ratios, i.e. φ = 1.0, 2.0 and 0.5 in the jet-stirred reactor unit. The temperature ranges from 500 K to 1100 K, at a pressure of 1.07 bar, and with a space time of 2 s. An elementary step kinetic model for DEA pyrolysis and oxidation is built using the automatic kinetic model generator Genesys with a base mechanism extracted from Glarborg et al. (2019) which describes the oxidation of the small nitrogen-containing species. Important thermodynamic and kinetic parameters for the DEA decomposition chemistry are obtained from quantum chemical calculations. The experimental trends are well predicted by the model, even without any fitting of the model thermodynamic or kinetic parameters. Rate of production analyses reveal the important pathways for the pyrolysis and low-and intermediate-temperature oxidation to hydrogen cyanide, acetonitrile, NOx and others.

Modelling case study on the role of pressure dependence in single event kinetic modelling for steam cracking of both ethane and propane. Results are validated with in-house generated experimental data. Renewable cracking feedstocks from plastic waste and the need for novel reactor designs related to electrification of steam crackers drives the development of accurate and fundamental kinetic models for this process, despite its large scale implementation for more than half a century. Pressure dependent kinetics have mostly been omitted in fundamental steam cracking models, while they are crucial in combustion models. Therefore, we have assessed the importance of pressure dependent kinetics for steam cracking via in-depth modelling and experimental studies. In particular we have studied the influence of considering fall-off on the product yields for ethane and propane steam cracking. A high-pressure limit fundamental kinetic model is generated, based on quantum chemical data and group additive values, and supplemented with literature values for pressure dependent kinetic parameters for beta-scission reactions and homolytic bond scissions of C-2 and C-3 species. Model simulations with high-pressure limit rate coefficients and pressure dependent kinetics are compared to new experimental measurements. Steam cracking experiments for pure ethane and propane feeds are performed on a tubular bench-scale reactor at 0.17 MPa and temperatures ranging from 1058 to 1178 K. All important product species are identified using a comprehensive GC x GC-FID/q-MS. For homolytic bond scissions, the inclusion of pressure dependent kinetics has a significant effect on the conversion profile for ethane steam cracking. On the other hand, pressure dependence of C-2 beta-scissions significantly influences conversion and product species profiles for both ethane and propane steam cracking. The C-3 beta-scissions pressure dependence has a negligible effect in ethane steam cracking, while for propane steam cracking the effect is non-negligible on the product species profiles.