HAL-INERIS
Not a member yet
8807 research outputs found
Sort by
Aerosol source apportionment uncertainty linked to the choice of input chemical components
No full textInternational audienceFor a Positive Matrix Factorization (PMF) aerosol source apportionment (SA) studies there is no standard procedure to select the most appropriate chemical components to be included in the input dataset for a given site typology, nor specific recommendations in this direction. However, these choices are crucial for the final SA outputs not only in terms of number of sources identified but also, and consequently, in the source contributions estimates. In fact, PMF tends to reproduce most of PM mass measured independently and introduced as a total variable in the input data, regardless of the percentage of PM mass which has been chemically characterized, so that the lack of some specific source tracers (e.g. levoglucosan) can potentially affect the results of the whole source apportionment study. The present study elaborates further on the same concept, evaluating quantitatively the impact of lacking specific sources’ tracers on the whole source apportionment, both in terms of identified sources and source contributions. This work aims to provide first recommendations on the most suitable and critical components to be included in PMF analyses in order to reduce PMF output uncertainty as much as possible, and better represent the most commons PM sources observed in many sites in Western countries. To this aim, we performed three sensitivity analyses on three different datasets across EU, including extended sets of organic tracers, in order to cover different types of urban conditions (Mediterranean, Continental, and Alpine), source types, and PM fractions. Our findings reveal that the vehicle exhaust source resulted to be less sensitive to the choice of analytes, although source contributions estimates can deviate significantly up to 44 %. On the other hand, for the detection of the non-exhaust one is clearly necessary to analyze specific inorganic elements. The choice of not analysing non-polar organics likely causes the loss of separation of exhaust and non-exhaust factors, thus obtaining a unique road traffic source, which provokes a significant bias of total contribution. Levoglucosan was, in most cases, crucial to identify biomass burning contributions in Milan and in Barcelona, in spite of the presence of PAHs in Barcelona, while for the case of Grenoble, even discarding levoglucosan, the presence of PAHs allowed identifying the BB factor. Modifying the rest of analytes provoke a systematic underestimation of biomass burning source contributions. SIA factors resulted to be generally overestimated with respect to the base case analysis, also in the case that ions were not included in the PMF analysis. Trace elements were crucial to identify shipping emissions (V and Ni) and industrial sources (Pb, Ni, Br, Zn, Mn, Cd and As). When changing the rest of input variables, the uncertainty was narrow for shipping but large for industrial processes. Major and trace elements were also crucial to identify the mineral/soil factor at all cities. Biogenic SOA and Anthropogenic SOA factors were sensitive to the presence of their molecular tracers, since the availability of OC alone is unable to separate a SOA factor. Arabitol and sorbitol were crucial to detecting fungal spores while odd number of higher alkanes (C27 to C31) for plant debris
Exposure to PM2.5 modulate the pro-inflammatory and interferon responses against influenza virus infection in a human 3D bronchial epithelium model
No full textInternational audienceEpidemiological studies showed a positive association between exposure to PM2.5 and the severity of influenza virus infection. However, the mechanisms by which PM2.5 can disrupt antiviral defence are still unclear. From this perspective, the objective of this study was to evaluate the effects of PM2.5 on antiviral signalling in the respiratory epithelium using the bronchial Calu-3 cell line grown at the air-liquid interface. Pre-exposure to PM2.5 before infection with the influenza virus was investigated, as well as a co-exposure. Although a physical interaction between the virus and the particles seems possible, no effect of PM2.5 on viral replication was observed during co-exposure, although a downregulation of IFN-β release was associated to PM2.5 exposure. However, pre-exposure slightly increased the viral nucleoprotein production and the pro-inflammatory response. Conversely, the level of the myxovirus resistance protein A (MxA), an interferon-stimulated gene (ISG) induced by IFN-β, was reduced. Therefore, these results suggest that pre-exposure to PM2.5 could alter the antiviral response of bronchial epithelial cells, increasing their susceptibility to viral infection
Significant impact of urban tree biogenic emissions on air quality estimated by a bottom-up inventory and chemistry transport modeling
No full textInternational audienceBiogenic volatile organic compounds (BVOCs) are emitted by vegetation and react with other compounds to form ozone and secondary organic matter (OM). In regional air quality models, biogenic emissions are often calculated using a plant functional type approach, which depends on the land use category. However, over cities, the land use is urban, so trees and their emissions are not represented. Here, we develop a bottom-up inventory of urban tree biogenic emissions in which the location of trees and their characteristics are derived from the tree database of the Paris city combined with allometric equations. Biogenic emissions are then computed for each tree based on their leaf dry biomass, tree-species-dependent emission factors, and activity factors representing the effects of light and temperature. Emissions are integrated in WRF-CHIMERE air quality simulations performed over June–July 2022. Over Paris city, the urban tree emissions have a significant impact on OM, inducing an average increase in the OM of about 5 %, reaching 14 % locally during the heatwaves. Ozone concentrations increase by 1.0 % on average and by 2.4 % during heatwaves, with a local increase of up to 6 %. The concentration increase remains spatially localized over Paris, extending to the Paris suburbs in the case of ozone during heatwaves. The inclusion of urban tree emissions improves the estimation of OM concentrations compared to in situ measurements, but they are still underestimated as trees are still missing from the inventory. OM concentrations are sensitive to terpene emissions, highlighting the importance of favoring urban tree species with low-terpene emissions
Fire hazards of carbonate-based electrolytes for sodium-ion batteries: What changes from lithium-ion batteries?
No full textInternational audienceCarbonate-based electrolytes are often employed as the preferred electrolyte for both Li-ion and Na-ion cells. Toinvestigate the fire risk during abuse conditions in real-life scenarios covering the full value chain, not only celllevelstudies but also component-level investigation is crucial. Hence, Na-ion advanced electrolyte combustiontests are performed employing the fire propagation apparatus also called Tewarson calorimeter. Heat andcombustion products releases are measured, making use of fire calorimetry laws and analytical techniques suchas FTIR, NDIR, FID, or paramagnetic analyzers, and optical measurement. Thermal and chemical impacts of Naionelectrolytes combustion and fires are assessed under well-ventilated and under-ventilated environments. Dataare then compared against a carbonate-based electrolyte used in Li-ion batteries to create a comparative studybetween these technologies. Overall, the heat released rate majorly depends upon the solvents used and is lessimpacted by inorganic Li or Na salts while the emitted gases depend on both solvent and salt chemistry. Anotherkey observation lies in the different fate of the fluorine element chemically bound to the concerned salts: insimilar burning conditions, F from NaPF6 decomposition is preferably converted in F-containing solid species inthe residues whilst LiPF6 gives off more gaseous species such as HF
Study of the behaviour of walls and doors of 20 ft ISO containers through real-scale explosion tests
No full textInternational audienceISO containers are increasingly used to house processes involving a potential explosion risk such asBattery Energy Storage Systems or hydrogen-based systems. For both examples, explosions may bepowerful because of the potentially large obstruction in the container promoting flame accelerationand/or the high dynamics of the hydrogen flames.An important part of the modelling strategy for predicting residual pressure effects of an explosion inan ISO container is the knowledge of the mechanical behaviour of the structure initially confining theflammable cloud.In order to improve knowledge regarding the mechanical behaviour of the container and get adatabase suitable for validation purposes of the modelling tool, hydrogen explosions tests in 20-footISO containers were performed at INERIS testing facilities. The goal was to generate distinctivepressure signals and to collect comprehensive measurements related to the mechanical behaviour ofthe doors and walls.The paper details the experimental set-up, the metrology dedicated to mechanical measurements andthe main results. First conclusions of the experimental campaign are also supplied.Thus, the comparative analyses of dynamic explosion load tests (20%, 22%) highlight the significantdifferences in the intensity of the stress imposed on the structure of the container. The 20% testpresents less severe conditions with reduced deformations and recorded stresses compared to the 22%test. These results underline the importance of considering the variable nature of dynamic loads whenevaluating container resistance and offer valuable insights for improving the design of protectivebarriers (like explosion vent panels for example) in similar conditions
Une Approche quantitative de l'exposition fœtale au plomb : développement d'un modèle pharmacocinétique basé sur la physiologie de la grossesse (p-PBPK) et application aux données de biosurveillance
No full textLead (Pb) is a highly potent neurotoxin, especially for children. A vast epidemiological literature documents the effects of Pb exposure in young children. Pb can cross the placental barrier to induce early fetal exposure. This prenatal exposure is associated with impaired cognitive function in children, resulting in learning problems, decreased intelligence quotient (IQ) and behavioral problems. For obvious ethical reasons, it is not possible to measure Pb concentrations in the fetus during pregnancy. Consequently, a single measurement obtained from umbilical cord blood at delivery is available to assess prenatal exposure. Thankfully, Pregnancy physiologically based pharmacokinetic (p-PBPK) models can be used to simulate prenatal exposures throughout pregnancy. The aim of this thesis is to provide a tool capable of simulating internal fetal exposures, from incomplete data, such as those obtained during biomonitoring campaigns.A literature search was carried out on the impact of pregnancy on the ADME processes of Pb in both the mother and fetus. This research highlighted an increase in maternal bone turnover (i.e. cycle of formation and resorption) during pregnancy. Indeed, urinary markers of bone formation and resorption linearly increase from the 2nd trimester until the end of pregnancy. Furthermore, data on the kinetic of Pb during pregnancy reveal a distinctive U-shaped trend, an initial phase of decrease during the first 4 months followed by a second phase of increase in maternal blood lead levels. The parameter values modeling bone turnover and placental transfer of Pb were calibrated using literature data. The predictive performance of our p-PBPK model was assessed using paired concentrations of Pb in umbilical cord and maternal blood at delivery, demonstrating satisfactory performance between the model predictions and the observed data. A global sensitivity analysis of the model highlighted à strong influence on fetal exposures from the Pb binding to erythrocytes, Pb accumulation in bones and the partition coefficient in the brain.Pb concentrations measured in the umbilical cord at delivery, from the French Longitudinal Study since Infancy (Elfe), were used in the p-PBPK model to simulate Pb concentrations in the fetal brain. Simulations carried out on the Elfe cohort reveal variations in predicted Pb concentrations in fetal brain over the course of pregnancy. By the end of pregnancy, almost half the fetuses in the Elfe cohort had predicted brain concentrations likely to induce in vitro and ex vivo effects on neurodevelopmental.In conclusion, this thesis highlights the potential of p-PBPK models for assessing prenatal exposure. We have demonstrated the ability of our model to simulate fetal exposure indicators, taking into account the physiological mechanisms essential to Pb toxicokinetics during pregnancy. In the future, this model can be coupled with an effect model such as an Adverse Outcome Pathway (AOP) to help predict the effects of Pb on the neurodevelopment of children.Le plomb (Pb) est une neurotoxine très puissante, notamment pour les enfants. Une vaste littérature épidémiologique documente les effets de l'exposition au Pb chez les jeunes enfants. Le Pb est capable de traverser la barrière placentaire pour induire une exposition précoce du fœtus. Cette exposition prénatale est associée à une altération des fonctions cognitives des enfants qui se traduit par des problèmes d'apprentissage, une diminution du quotient intellectuel (QI) et des problèmes de comportement. Pour des raisons d'éthique évidentes, il n'est pas possible de mesurer la concentration de Pb chez le fœtus pendant la grossesse. Ainsi, une seule mesure obtenue dans le sang du cordon ombilical à l'accouchement est disponible pour évaluer l'exposition prénatale. Les modèles pharmacocinétiques basés sur la physiologie de la grossesse (p-PBPK) peuvent être utilisés pour simuler les expositions prénatales tout au long de la grossesse. L'objectif de cette thèse est de fournir un outil capable de simuler les expositions internes du fœtus à partir de données parcellaires, comme par exemple, celles obtenues lors de campagnes de biosurveillance.Une recherche bibliographique a été menée sur l'impact de la grossesse sur les processus ADME du Pb chez la mère et le fœtus. Cette recherche a mis en évidence une augmentation du renouvellement (i.e. cycle de formation et de résorption) de l'os maternel pendant la grossesse. Les marqueurs de la formation et de la résorption de l'os, dosés dans les urines, augmentent linéairement à partir du 2nd trimestre jusqu'à la fin de grossesse. Par ailleurs, les données sur la cinétique du Pb pendant la grossesse chez la mère révèlent une courbe en forme de U : une première phase de diminution lors des 4 premiers mois et suivie d'une seconde phase d'augmentation de la plombémie maternelle. Les valeurs des paramètres modélisant le renouvellement de l'os et le transfert placentaire du Pb ont été calibré à partir des données de la littérature. La performance prédictive de notre modèle a été évaluée en utilisant les rapports de concentrations appariées de Pb dans le sang du cordon ombilical et maternel à l'accouchement, démontrant ainsi une performance satisfaisante entre les prédictions du modèle et les données observées. Une analyse de sensibilité globale du modèle a mis en évidence une forte influence sur les expositions fœtales de la liaison du Pb aux érythrocytes, l'accumulation du Pb dans les os et du coefficient de partage dans le cerveau.Les concentrations de Pb mesurées dans le cordon ombilical à l'accouchement, provenant de l'étude longitudinale française depuis l'enfance (Elfe), ont été utilisées dans le modèle p-PBPK pour simuler les concentrations de Pb dans le cerveau des fœtus. Les simulations effectuées sur la cohorte Elfe révèlent des variations des concentrations prédites de Pb dans le cerveau des fœtus au cours de la grossesse. À la fin de la grossesse, presque la moitié des fœtus de la cohorte Elfe présentaient des concentrations prédites dans le cerveau susceptibles d'induire in vitro et ex vivo des effets sur le neurodéveloppement.Pour conclure, ce travail de thèse met en lumière le potentiel des modèle p-PBPK pour l'évaluation des expositions prénatales. Nous avons démontré la capacité de notre modèle à simuler des indicateurs d'exposition chez le fœtus en tenant compte des mécanismes physiologiques essentiels à la toxicocinétique du Pb pendant la grossesse. Ce modèle peut à l'avenir être couplé à un modèle d'effet tel qu'un chemin d'effet néfaste (AOP) pour aider à prédire les effets du Pb sur le neurodéveloppement des enfants
Identification des sources des particules sur la zone littorale des Hauts de France et de la Normandie
No full textIn Hauts-de-France and Normandy, an increase was observed between 2018 and 2019 in the number of days on which the regulatory value for airborne dust less than or equal to 10 μm - PM10 (> 50 μg/m3) was exceeded for a majority of stations located in coastal areas. This increase was mainly observed during the spring period. It is in this context that, in 2021, Atmo Hauts-de-France and Atmo Normandie, with the support of the LCSQA (Laboratoire Central de Surveillance de la Qualité de l'Air), have set up a study aimed at improving understanding of these PM10 episodes in coastal areas. This study was carried out as part of the CARA programme (chemical characterisation of particles) and co-financed by the Hauts-de-France Region, the Agence Régionale de Santé Normandie, the Institut des Géosciences de l'Environnement (IGE) and the Laboratoire Central de Surveillance de la Qualité de l'Air (LCSQA). Its main initial objective was to understand the origin of spring and coastal episodes of particulate pollution by analysing the composition of PM10 particles at 4 coastal stations ( English Channel and the North Sea).Four monitoring stations were selected for the study: two in Hauts-de-France, Calais Berthelot (urban) and Arrest (rural), and two in Normandy, Phare d'Ailly (rural) and Le Havre ville Haute (urban). At the 4 sites investigated, daily filters were collected over the 5 months of the study (between 1 February and 29 June 2021). A detailed chemical characterisation was carried out on all these filters, the results of which were published in a previous report1.This study is based on the use of a statistical tool called Positive Matrix Factorization (PMF), which enables the contribution of the main sources and mechanisms of particle formation to be determined. Initially, two approaches were used and compared: ‘single-site’ PMF and ‘multi-site’ PMF. The multi-site MFP was chosen in order to better identify source profiles with strong seasonal characteristics, which are difficult to characterise by an ‘individualised’ MFP due to a lack of data on certain sites.The multi-site MFP identified 11 factors that could be associated with PM10 source profiles. To simplify interpretation, some factors were grouped together. Thus, in fine, 9 sources will be used for the interpretation part (marine salts, primary biogenic, secondary biogenic, primary road traffic, heavy fuel oil combustion, mineral dusts, industries, biomass combustion, and nitrate-rich). A high contribution of marine salts and secondary aerosols enriched in ammonium nitrate (nitrate-rich) to the PM10 mass was observed. These two sources together account for an average of around 50 to 60% of the total mass of PM10 particles, depending on the site, over the study period.The temporal variability of the main sources identified is then discussed for each site. Average monthly concentrations of PM10 in cold periods (February - March) are higher than those observed in warmer periods (April - June) at all the sites studied. 6 sources show marked monthly variability over the study period.The temporal variability of the main sources identified is then discussed for each site. Average monthly PM10 concentrations in cold periods (February - March) are higher than those observed in warmer periods (April - June) at all the sites studied. 6 sources show marked monthly variability over the study period.Over the study period, 4 exceedances of the regulatory daily value for PM10 (> 50 μg/m3) were observed at the study sites, i.e. 1 exceedance at Calais and 3 exceedances at Arrest. No exceedances were observed at the two sites in Normandy. The days on which high daily concentrations were observed were analysed in detail. Thus, 3 sources (biomass combustion, nitrate-rich and primary road traffic) were identified as the main sources during these winter and spring episodes.In addition, an analysis of the geographical origins is presented for 4 of the most abundant factors (sources), confirming the mainly local origins of primary road traffic and biomass combustion. Unsurprisingly, the highest concentrations of sea salts are associated with oceanic air masses, while the factor rich in ammonium nitrate (nitrate-rich) appears to have essentially (supra-)regional origins, with maximum concentrations during anticyclonic periods under north-easterly winds.Finally, a summary of the origins of particulate matter and the priority actions to be considered by sector are proposed in the form of conclusions.En Hauts-de-France et en Normandie, il a été observé entre 2018 et 2019, une augmentation du nombre de jours de dépassement de la valeur réglementaire relative aux poussières en suspension dans l’air de taille inférieure ou égale à 10 μm - PM10 (> 50 μg/m3) pour une majorité des stations situées en zone côtière. Cette augmentation a principalement été constatée en période printanière. C’est dans ce contexte qu’en 2021, Atmo Hauts-de-France et Atmo Normandie, avec l’appui du LCSQA (Laboratoire Central de Surveillance de la Qualité de l'Air), ont mis en place une étude visant à améliorer la compréhension de ces épisodes de PM10 en zone littorale. Cette étude a été réalisée dans le cadre du programme CARA (caractérisation chimique des particules) et cofinancée par la Région Hauts-de-France, L’Agence Régionale de Santé Normandie, l'Institut des Géosciences de l'Environnement (IGE) et le Laboratoire Central de Surveillance de la Qualité de l'Air (LCSQA). Son principal objectif initial était, par l’analyse de la composition des particules PM10 sur 4 stations littorales (Manche et Mer du Nord) de comprendre l’origine des épisodes printaniers et côtiers de pollution particulaire.Quatre stations de mesure ont été sélectionnées pour cette étude : deux en Hauts-de-France, Calais Berthelot (urbaine) et Arrest (rurale), et deux en Normandie, Phare d’Ailly (rurale) et Le Havre ville Haute (urbaine). Sur les 4 sites investigués, des filtres journaliers ont été collectés durant les 5 mois de l’étude (entre le 1er février et le 29 juin 2021). Une caractérisation chimique détaillée a été réalisée sur l’ensemble de ces filtres dont les résultats ont été publiés dans un précédent rapport1.Cette étude est basée sur l’utilisation d’un outil statistique appelé Positive Matrix Factorization (PMF), qui permet la détermination de la contribution des principales sources et mécanismes de formation des particules. Dans un premier temps, deux approches, la PMF « mono-site » et la PMF « multi-sites » ont été réalisées et comparées. La PMF multi-sites a été retenue pour mieux identifier les profils de sources présentant une forte saisonnalité, qui sont difficilement caractérisables par une PMF « individualisée » en raison d’un manque de données sur certains sites.La PMF multi-sites a permis de distinguer 11 facteurs qui ont pu être associés à des profils de sources des particules PM10. Pour simplifier l’interprétation, certains facteurs ont été regroupés. Ainsi, in fine, 9 sources seront utilisées pour la partie d’interprétation (sels marins, biogénique primaire, biogénique secondaire, trafic routier primaire, combustion de fuel lourd, poussières minérales, industries, combustion de biomasse, et nitrate-rich). Une forte contribution des sels marins et des aérosols secondaires enrichis en nitrate d’ammonium (nitrate-rich) à la masse des PM10 a été observée. Ces deux sources représentent ensemble en moyenne environ 50 à 60% de la masse totale des particules PM10 selon les sites sur la période d’étude.La variabilité temporelle des principales sources identifiées est ensuite discutée pour chaque site. Les concentrations moyennes mensuelles de PM10 en période froide (févier – mars) sont plus élevées que celles observées en périodes plus chaudes (avril – juin) sur l’ensemble des sites étudiés. 6 sources présentent une variabilité mensuelle marquée au cours de la période d’étude.Sur la période étudiée, 4 dépassements de la valeur journalière réglementaire en PM10 (> 50 μg/m3) ont été constatés sur les sites d’étude, à savoir 1 dépassement à Calais et 3 dépassements à Arrest. Aucun dépassementn'a été observé pour les deux sites normands. Les journées pour lesquelles des fortes concentrations journalières ont été observées ont fait l’objet d’une analyse approfondie. Ainsi, 3 sources (combustion de biomasse, nitrate-rich et trafic routier primaire) ont été identifiées comme sources majoritaires pendant ces épisodes hivernaux et printaniers.Par ailleurs, une analyse des origines géographiques est présentée pour 4 facteurs (sources) parmi les plus abondants, confirmant les origines principalement locales du trafic routier primaire et de la combustion de biomasse. Sans surprise, les plus fortes concentrations de sels marins sont associées à des masses d’air océaniques, alors que le facteur riche en nitrate d’ammonium (nitrate-rich) apparaît comme ayant des origines essentiellement (supra-)régionales, avec des concentrations maximales en période anticyclonique sous vents de nord-est.Enfin, un résumé des origines des particules et les actions prioritaires à envisager par secteur sont proposés sous forme de conclusions
Prediction of the accumulation behavior of pesticides in PDMS-coated stir bars used as passive samplers in freshwaters
No full textInternational audiencePassive samplers accumulate organic contaminants at rates that depend on in-field exposure conditions such as freshwater flow velocity, water temperature and water quality. Time-weighted average concentrations can be determined by using a correction process such as the performance reference compound (PRC) method. This study presented a new approach to predict the accumulation behavior of pesticides in polydimethylsiloxane-coated stir bars under different exposure conditions and assign a specific PRC to each pesticide for quantitative purposes. We used an experimental design with eight simultaneous accumulation kinetics of 13 pesticides and elimination kinetics of three PRC candidates run in a flow-through system to determine the effects of flow velocity, water temperature and dissolved organic matter on the kinetic constants. We identified the parameters that had a significant effect on the accumulation of each pesticide and assigned a PRC candidate to each pesticide. We then used a discriminant function analysis to find the parameters that had a significant effect on accumulation of the 13 pesticides via their physical-chemical properties and to predict through a stochastic approach the parameters for seven other pesticides. This approach provides a better framework for identifying a PRC than conventional methods to determine unbiased concentrations in future monitoring efforts
Size characterization and quantification of E171 titanium dioxide particles in food sauces using spICP-MS
No full textInternational audienceA sample preparation method associated to a single particle inductively coupled plasma mass spectrometry (spICP-MS) analysis was developed to ensure a quantitative characterization of TiO2 particles in food sauces. The method was optimized with a fat sauce matrix for which the pristine E171 additive was available. The highest extracted amount of TiO2 particles was obtained using an ultrasound-assisted dispersion in tetramethylammonium hydroxide (TMAH). TEM analysis confirmed these results, showing that these conditions effectively solubilised the fat matrix. The trueness of the method evaluated using a Ti-free mayonnaise spiked with the E171 additive or the NM-100 reference material (NM-100 RM) proved satisfactory with quantitative mass recoveries (>90%). A good agreement in terms of particle sizes was also obtained between those of the spiked mayonnaise and the TiO2 suspensions used for spiking. However, as also shown in the optimization study, the presence of fatty compounds tended to improve particle dispersion. The application of this procedure to E171 labelled food sauces purchased from the EU market showed concentrations of TiO2 particles in the (0.6 – 6) g/kg range with median diameters between 170 and 320 nm. The TiO2 nanoparticle fraction (size < 100 nm) was evaluated between 2% and 12%, depending on the sample. No titanium was detected in the food sauces not labelled E171. Conclusively, the developed method can provide an efficient and reliable analytical approach for TiO2 particles determination in fat food samples which complies with the EU current regulations
Estimating the dynamic early life exposure to PFOA and PFOS of the HELIX children: Emerging profiles via prenatal exposure, breastfeeding, and diet
No full textInternational audienceIn utero and children’s exposure to per- and polyfluoroalkyl substances (PFAS) is a major concern in health risk assessment as early life exposures are suspected to induce adverse health effects. Our work aims to estimate children’s exposure (from birth to 12 years old) to PFOA and PFOS, using a Physiologically-Based Pharmacokinetic (PBPK) modelling approach.A model for PFAS was updated to simulate the internal PFAS exposures during the in utero life and childhood, and including individual characteristics and exposure scenarios (e.g., duration of breastfeeding, weight at birth, etc.). Our approach was applied to the HELIX cohort, involving 1,239 mother–child pairs with measured PFOA and PFOS plasma concentrations at two sampling times: maternal and child plasma concentrations (6 to 12 y.o).Our model predicted an increase in plasma concentrations during fetal development and childhood until 2 y.o when the maximum concentrations were reached. Higher plasma concentrations of PFOA than PFOS were predicted until 2 y.o, and then PFOS concentrations gradually became higher than PFOA concentrations. From 2 to 8 y.o, mean concentrations decreased from 3.1 to 1.88 µg/L or ng/mL (PFOA) and from 4.77 to 3.56 µg/L (PFOS). The concentration–time profiles vary with the age and were mostly influenced by in utero exposure (on the first 4 months after birth), breastfeeding (from 5 months to 2 (PFOA) or 5 (PFOS) y.o of the children), and food intake (after 3 (PFOA) or 6 (PFOS) y.o of the children). Similar measured biomarker levels can correspond to large differences in the simulated internal exposures, highlighting the importance to investigate the children’s exposure over the early life to improve exposure classification.Our approach demonstrates the possibility to simulate individual internal exposures using PBPK models when measured biomarkers are scarce, helping risk assessors in gaining insight into internal exposure during critical windows, such as early life