Periodic Reporting for period 1 - SAFETYFANS (SAFETYFANS: SAFEty and sustainabiliTY by design: a Framework for Advanced Nano-materials Synthesis.)
Período documentado: 2023-11-01 hasta 2025-10-31
SAFETYFANS emerges in response to critical challenges in nanotechnology, where there is a growing need for safer and more sustainable material designs in line with the EU’s Safe and Sustainable by Design (SSbD) framework. SSbD is a novel, multi-dimensional approach that addresses safety, functionality, and sustainability considerations for advanced materials (AdMa) early in the innovation process. SAFETYFANS, specifically targeting Titanium Dioxide (TiO2), aims to provide a science-based solution for predicting and ranking materials based on their SSbD dimensions. The project is inspired by FAIR (Findable, Accessible, Interoperable, and Reusable) data principles, ensuring that outputs contribute to the broader research community.
In partnership with the CNR-ISSMC, SAFETYFANS will harness a dataset of AdMa. SAFETYFANS will employ New Approach Methodologies (NAMs), specifically Machine Learning (ML) integrated with explainable AI (xAI), to predict the safety, functionality, and sustainability of TiO2. This approach not only leverages data-driven insights but also builds a foundational tool for addressing the ongoing safety controversies associated with TiO2, making SAFETYFANS a pivotal initiative within the SSbD framework.
SAFETYFANS is structured around four key objectives:
1. Objective 1 (Data Foundation): Establish a comprehensive, FAIR-aligned data asset that compiles and organizes essential SSbD indicators for TiO2 from existing knowledge bases and forthcoming EU projects. This foundational dataset will ensure consistency, quality, and accessibility, setting the stage for subsequent project work packages.
2. Objective 2 (ML and xAI Models): Develop a portfolio of ML models tailored to predict the SSbD dimensions (safety, functionality, and sustainability) using the compiled data asset. Through explainable AI (xAI), SAFETYFANS will identify primary determinants that impact each SSbD dimension, allowing for an in-depth understanding of the most influential factors. The resulting insights will offer material designers actionable guidance on minimizing potential human and environmental risks.
3. Objective 3 (Experimental Validation): SAFETYFANS will integrate the experimental and modeling domains by synthesizing and characterizing novel TiO2 formulations and assessing them through toxicological and functional tests. This iterative approach will validate the model predictions against empirical results, ensuring a robust, validated framework that aligns with FAIR data principles and improves communication between experimentalists and modelers.
4. Objective 4 (Composite Indicator for Ranking): The project will culminate in the development of a Composite Indicator (CI) that ranks TiO2 formulations based on their SSbD potential, offering a streamlined approach for material scientists to prioritize designs that maximize functionality while reducing adverse effects. This CI will be a valuable asset for stakeholders, providing a materials sustainability portfolio that serves both environmental and human health needs.
SAFETYFANS is positioned to address the EU’s urgent need for practical, scalable SSbD solutions that resonate across academia, regulatory bodies, and industry. The project moves beyond theoretical frameworks, presenting a real-world instance of SSbD through its data-driven methodologies and predictive modeling. This will be significant, offering the EU a pragmatic, digital asset for enhancing nano-safety and advancing sustainable manufacturing within the nanotechnology sector. Through SSbD integration, SAFETYFANS promotes an economically viable and circular approach to nanotechnology. The ability to rank AdMa by SSbD dimensions during the early design phase offers an economically efficient route to sustainability, aligning nanotechnology with EU goals for circularity. SAFETYFANS mitigates potential disruptions, such as those caused by the ban on TiO2 as a food additive, which has complicated decision-making in the industry. The SSbD framework provides clarity, allowing for proactive compliance beyond regulatory demands, thereby supporting a steady innovation pace without sacrificing safety or functionality.
By the project’s end, SAFETYFANS will leave a lasting impact in the nanotechnology community, addressing the gaps in nanosafety, fostering interdisciplinary collaboration, and promoting a culture of responsible innovation in materials design.
In partnership with the CNR-ISSMC, SAFETYFANS will harness a dataset of AdMa. SAFETYFANS will employ New Approach Methodologies (NAMs), specifically Machine Learning (ML) integrated with explainable AI (xAI), to predict the safety, functionality, and sustainability of TiO2. This approach not only leverages data-driven insights but also builds a foundational tool for addressing the ongoing safety controversies associated with TiO2, making SAFETYFANS a pivotal initiative within the SSbD framework.
SAFETYFANS is structured around four key objectives:
1. Objective 1 (Data Foundation): Establish a comprehensive, FAIR-aligned data asset that compiles and organizes essential SSbD indicators for TiO2 from existing knowledge bases and forthcoming EU projects. This foundational dataset will ensure consistency, quality, and accessibility, setting the stage for subsequent project work packages.
2. Objective 2 (ML and xAI Models): Develop a portfolio of ML models tailored to predict the SSbD dimensions (safety, functionality, and sustainability) using the compiled data asset. Through explainable AI (xAI), SAFETYFANS will identify primary determinants that impact each SSbD dimension, allowing for an in-depth understanding of the most influential factors. The resulting insights will offer material designers actionable guidance on minimizing potential human and environmental risks.
3. Objective 3 (Experimental Validation): SAFETYFANS will integrate the experimental and modeling domains by synthesizing and characterizing novel TiO2 formulations and assessing them through toxicological and functional tests. This iterative approach will validate the model predictions against empirical results, ensuring a robust, validated framework that aligns with FAIR data principles and improves communication between experimentalists and modelers.
4. Objective 4 (Composite Indicator for Ranking): The project will culminate in the development of a Composite Indicator (CI) that ranks TiO2 formulations based on their SSbD potential, offering a streamlined approach for material scientists to prioritize designs that maximize functionality while reducing adverse effects. This CI will be a valuable asset for stakeholders, providing a materials sustainability portfolio that serves both environmental and human health needs.
SAFETYFANS is positioned to address the EU’s urgent need for practical, scalable SSbD solutions that resonate across academia, regulatory bodies, and industry. The project moves beyond theoretical frameworks, presenting a real-world instance of SSbD through its data-driven methodologies and predictive modeling. This will be significant, offering the EU a pragmatic, digital asset for enhancing nano-safety and advancing sustainable manufacturing within the nanotechnology sector. Through SSbD integration, SAFETYFANS promotes an economically viable and circular approach to nanotechnology. The ability to rank AdMa by SSbD dimensions during the early design phase offers an economically efficient route to sustainability, aligning nanotechnology with EU goals for circularity. SAFETYFANS mitigates potential disruptions, such as those caused by the ban on TiO2 as a food additive, which has complicated decision-making in the industry. The SSbD framework provides clarity, allowing for proactive compliance beyond regulatory demands, thereby supporting a steady innovation pace without sacrificing safety or functionality.
By the project’s end, SAFETYFANS will leave a lasting impact in the nanotechnology community, addressing the gaps in nanosafety, fostering interdisciplinary collaboration, and promoting a culture of responsible innovation in materials design.
Summary of the main technical and scientific activities performed and their achievements:
1. Machine Learning for Predictive Modeling in Nanoparticle Synthesis (Furxhi et al., 2024)
Activities: Leveraged machine learning to predict the properties of nanoparticles using synthesis data collected from literature. Data extraction covered synthesis parameters, antibacterial efficiency, and toxicological profiles. Achievements: Regression models accurately predicted core size and antibacterial efficiency, identifying key factors (e.g. synthesis duration, scale, capping agents) through Shapley values. This model advances Safe-by-Design (SbD) approaches by optimizing synthesis processes based on machine learning insights. This paper is a result of an exercise with ISSMC colleagues for data collection and harmonization. This methodology can be applied in any case of different AdMa.
2. Risk Assessment for Silver Nanoparticles in Textiles (Koivisto et al., 2024)
Activities: Conducted risk assessments for dermal and inadvertent (peri-)oral exposure to AgNPs from coated textiles using a mass balance model, evaluating release under conditions like artificial sweat and mechanical stress. Achievements: Identified that AgNP-containing gloves pose a high risk if followed by finger mouthing, emphasizing the need for realistic release test settings and good hygiene practices. The study establishes a foundation for safer design. This methodology can be applied in any case of different AdMa.
3. Development of Safe and Sustainable by Design (SSbD) Roadmap for Silver-Based Textiles (Furxhi et al., 2024)
Activities: Developed a roadmap with quantitative metrics (PCFs, KDFs, KPIs) for implementing SSbD in silver-based antimicrobial textile coatings, aligned with the ASINA project. Achievements: Established guidelines and a decision support tool for SSbD, addressing safety, functionality, and environmental impact. This roadmap promotes a holistic lifecycle approach to SSbD, guiding future nanotechnology projects within a broader EU sustainability context.
4. SSbD Roadmap for TiO2-based Depolluting Surfaces (Furxhi et al., 2024)
Activities: Provided an SSbD framework for TiO2 photocatalytic applications, focusing on depolluting surfaces. Methods included spray-finishing techniques for TiO2 formulations and integrating Key Performance Indicators into the roadmap. Achievements: Demonstrated a quantitative SSbD methodology for environmental applications, creating a benchmark for nanomaterials in various sectors. This contributes to standardized SSbD practices, furthering EU’s Green Deal goals for safer and sustainable innovation.
5. Machine Learning for TiO2 Genotoxicity Prediction (Furxhi et al., 2024)
Activities: Utilized machine learning to predict TiO2-induced DNA damage from comet assay datasets. Models incorporated physicochemical properties and experimental parameters (e.g. concentration, cold lysis conditions). Achievements: Extra Trees and XGB regressors achieved high predictive accuracy, with concentration emerging as a critical factor. This study underscores ML’s role in nanotoxicology, highlighting key predictors of genotoxicity and contributing a FAIR dataset for further research.
6. Bayesian Networks for AgNPs Hazard Prediction in Soil (Furxhi et al., 2024)
Activities: Developed a Bayesian Network model to predict the hazard potential of AdMa in soil environments, incorporating factors like particle characteristics, exposure medium, and ecological data. Achievements: Achieved 82% predictive accuracy for chronic NOEC, identifying significant predictors of ecotoxicological risk. This model aids rapid hazard assessment and reduces dependence on costly lab experiments, aligning with the EU Green Deal’s sustainable practices.
7. Impact of Granulation Parameters on TiO2 Properties (Vespignani, 2024)
Activities: Investigated high-shear wet granulation (HSWG) for TiO2, examining the influence of granulation parameters (e.g. liquid-to-solid ratio, granulation time) on granulate size, porosity, and mechanical strength. Achievements: Identified the liquid-to-solid ratio and granulation time as key factors influencing granulate properties, enabling process optimization. The findings provide actionable insights for tailoring TiO2 granulation processes for specific applications, promoting material quality and stability.
8. High-Shear Granulation for Clay-Based Powders (Vespignani, 2024)
Activities: Assessed HSWG parameters (e.g. liquid addition rate, binder use) on hydrophilic and hydrophobic clays, focusing on granule size distribution and mechanical attributes. Achievements: Revealed the importance of the liquid-to-solid ratio for hydrophilic clays, with granulation time and impeller speed influencing hydrophobic clays. This knowledge supports tailored process design for different clay types, enhancing material performance and applicability.
Summary of Achievements
Collectively, these studies exemplify the integration of machine learning, Bayesian networks, and experimental optimization to advance the SSbD framework for AdMa. Key predictors and process parameters were identified across applications, from nanotoxicology to material synthesis, aiding safer and sustainable product designs. The methods developed, particularly the decision support tools and data-driven approaches, contribute valuable frameworks for reproducible, transparent, and predictive models in nanosafety and advanced material science.
1. Machine Learning for Predictive Modeling in Nanoparticle Synthesis (Furxhi et al., 2024)
Activities: Leveraged machine learning to predict the properties of nanoparticles using synthesis data collected from literature. Data extraction covered synthesis parameters, antibacterial efficiency, and toxicological profiles. Achievements: Regression models accurately predicted core size and antibacterial efficiency, identifying key factors (e.g. synthesis duration, scale, capping agents) through Shapley values. This model advances Safe-by-Design (SbD) approaches by optimizing synthesis processes based on machine learning insights. This paper is a result of an exercise with ISSMC colleagues for data collection and harmonization. This methodology can be applied in any case of different AdMa.
2. Risk Assessment for Silver Nanoparticles in Textiles (Koivisto et al., 2024)
Activities: Conducted risk assessments for dermal and inadvertent (peri-)oral exposure to AgNPs from coated textiles using a mass balance model, evaluating release under conditions like artificial sweat and mechanical stress. Achievements: Identified that AgNP-containing gloves pose a high risk if followed by finger mouthing, emphasizing the need for realistic release test settings and good hygiene practices. The study establishes a foundation for safer design. This methodology can be applied in any case of different AdMa.
3. Development of Safe and Sustainable by Design (SSbD) Roadmap for Silver-Based Textiles (Furxhi et al., 2024)
Activities: Developed a roadmap with quantitative metrics (PCFs, KDFs, KPIs) for implementing SSbD in silver-based antimicrobial textile coatings, aligned with the ASINA project. Achievements: Established guidelines and a decision support tool for SSbD, addressing safety, functionality, and environmental impact. This roadmap promotes a holistic lifecycle approach to SSbD, guiding future nanotechnology projects within a broader EU sustainability context.
4. SSbD Roadmap for TiO2-based Depolluting Surfaces (Furxhi et al., 2024)
Activities: Provided an SSbD framework for TiO2 photocatalytic applications, focusing on depolluting surfaces. Methods included spray-finishing techniques for TiO2 formulations and integrating Key Performance Indicators into the roadmap. Achievements: Demonstrated a quantitative SSbD methodology for environmental applications, creating a benchmark for nanomaterials in various sectors. This contributes to standardized SSbD practices, furthering EU’s Green Deal goals for safer and sustainable innovation.
5. Machine Learning for TiO2 Genotoxicity Prediction (Furxhi et al., 2024)
Activities: Utilized machine learning to predict TiO2-induced DNA damage from comet assay datasets. Models incorporated physicochemical properties and experimental parameters (e.g. concentration, cold lysis conditions). Achievements: Extra Trees and XGB regressors achieved high predictive accuracy, with concentration emerging as a critical factor. This study underscores ML’s role in nanotoxicology, highlighting key predictors of genotoxicity and contributing a FAIR dataset for further research.
6. Bayesian Networks for AgNPs Hazard Prediction in Soil (Furxhi et al., 2024)
Activities: Developed a Bayesian Network model to predict the hazard potential of AdMa in soil environments, incorporating factors like particle characteristics, exposure medium, and ecological data. Achievements: Achieved 82% predictive accuracy for chronic NOEC, identifying significant predictors of ecotoxicological risk. This model aids rapid hazard assessment and reduces dependence on costly lab experiments, aligning with the EU Green Deal’s sustainable practices.
7. Impact of Granulation Parameters on TiO2 Properties (Vespignani, 2024)
Activities: Investigated high-shear wet granulation (HSWG) for TiO2, examining the influence of granulation parameters (e.g. liquid-to-solid ratio, granulation time) on granulate size, porosity, and mechanical strength. Achievements: Identified the liquid-to-solid ratio and granulation time as key factors influencing granulate properties, enabling process optimization. The findings provide actionable insights for tailoring TiO2 granulation processes for specific applications, promoting material quality and stability.
8. High-Shear Granulation for Clay-Based Powders (Vespignani, 2024)
Activities: Assessed HSWG parameters (e.g. liquid addition rate, binder use) on hydrophilic and hydrophobic clays, focusing on granule size distribution and mechanical attributes. Achievements: Revealed the importance of the liquid-to-solid ratio for hydrophilic clays, with granulation time and impeller speed influencing hydrophobic clays. This knowledge supports tailored process design for different clay types, enhancing material performance and applicability.
Summary of Achievements
Collectively, these studies exemplify the integration of machine learning, Bayesian networks, and experimental optimization to advance the SSbD framework for AdMa. Key predictors and process parameters were identified across applications, from nanotoxicology to material synthesis, aiding safer and sustainable product designs. The methods developed, particularly the decision support tools and data-driven approaches, contribute valuable frameworks for reproducible, transparent, and predictive models in nanosafety and advanced material science.
The SAFETYFANS project, alongside the broader efforts documented in the referenced studies above (predictive modeling and SSbD integration for nanomaterials, development of SSbD roadmaps, process optimization for nanomaterial stability, genotoxicity and environmental hazard prediction models), has achieved key results across multiple domains in nanotechnology, contributing to the SSbD framework. These outcomes highlight the technical feasibility and scientific rigor of predictive modelling across the life cycle of AdMa, promoting a safer, more sustainable approach to nanomaterial design
Key needs:
Continued validation of predictive models through larger, diversified datasets and real-world testing is crucial. This involves not only expanding the models for broader applications but also ensuring the findings hold across various conditions. Demonstration projects should be funded to show the practical benefits of SSbD principles in commercial settings.
Collaboration with regulatory bodies, will help create guidelines that support industry compliance. Standardization will also ensure data consistency, which is fundamental for the reproducibility and scalability of predictive models.
To encourage market adoption, the project outcomes need pathways to reach stakeholders in nanotechnology, including investors, manufacturers, and end-users. Grants and financial incentives can help small- and medium-sized enterprises (SMEs) adopt SSbD practices.
Harmonized international frameworks would foster global knowledge sharing. Additionally, integrating the SSbD framework across sectors (e.g. textiles, healthcare, electronics) would enhance its adaptability and relevance.
Key needs:
Continued validation of predictive models through larger, diversified datasets and real-world testing is crucial. This involves not only expanding the models for broader applications but also ensuring the findings hold across various conditions. Demonstration projects should be funded to show the practical benefits of SSbD principles in commercial settings.
Collaboration with regulatory bodies, will help create guidelines that support industry compliance. Standardization will also ensure data consistency, which is fundamental for the reproducibility and scalability of predictive models.
To encourage market adoption, the project outcomes need pathways to reach stakeholders in nanotechnology, including investors, manufacturers, and end-users. Grants and financial incentives can help small- and medium-sized enterprises (SMEs) adopt SSbD practices.
Harmonized international frameworks would foster global knowledge sharing. Additionally, integrating the SSbD framework across sectors (e.g. textiles, healthcare, electronics) would enhance its adaptability and relevance.