The EU Chemical Strategy for Sustainability (CSS), Zero Pollution Action Plan and Circular Economy Action Plan are all pertaining to the European Green Deal and/or the ‘Fit for 55’ Package to deliver the EU’s 2030 Climate Target on the way to climate neutrality, by calling for a twin green & digital transition to a sustainable, climate-neutral and circular, as well as globally competitive and resilient European economy. With these ambitious targets in mind, the achievement of the market goals for the development of Safe-and-Sustainable-by-Design (SSbD) (advanced) materials and chemicals (AdMas&Chems) is a central requirement for reaching the multiple goals of strengthening critical raw materials resilience, decreasing energy consumption and greenhouse gas emissions, and reaching the zero-pollution goal and a toxic-free environment. The resulting SSbD AdMas&Chems need to provide the high functionality required for their advanced (and sometimes smart) applications, whilst simultaneously exhibiting improved safety and sustainability performances that take into account the complete value chain and life cycle, as outlined in the SSbD framework by the EU Joint Research Centre and adopted in the Commission Recommendation of 8 Dec. 2022, referred to in short as “EU SSbD Framework” in this proposal.
The integration of the SSbD Framework into the development cycle of AdMas&Chems (lower part) will be achieved by solving the multi-objective optimisation problem to improve and balance the four requirement categories (i.e. functionality, cost-efficiency, safety and sustainability, middle part) at each stage of the development. This will be achieved by integrating existing data and filling data gaps using innovative modelling and simulation approaches from the complete life cycle and value chain (upper part). All five steps of the SSbD (see more details below) will be integrated in each stage, even if starting with a limited set of evaluation criteria or rough estimates for each step at early development. PINK will then be able to use higher-tiered methods in later material development stages, providing higher confidence in the predictions by producing increasing information on a constantly reduced set of better performing candidates.
The PINK Project combines computational models and a decision support system (DSS) that exploit the combined power of first-principles simulation and pre-existing data, which – in itself – is further improved by advanced artificial intelligence (AI) technology; this provides PINK with the capability to significantly decrease the time needed for SSbD AdMas&Chems development. However, this requires the integration of tools from different and (hitherto) partly independently developed areas. PINK provides this integration in the form of an open innovation platform, the PINK In Silico Hub (PINKISH) based on an advanced Interoperability Framework, giving access to all information and knowledge, and executing SSbD workflows customisable to (a) the application area of the AdMas&Chems, (b) their safety and sustainability concerns of the existing materials, and (c) the status of the relevant development project (from early design ideas to registration and market entry). This is achieved by following a tiered approach, with respect to the throughput and confidence levels of the methods used. I
PINK will push current boundaries of the integration of computational approaches in the development of SSbD AdMas&Chems in industry and especially SMEs, based on the amalgamation of different technologies into intelligent, tiered functionality, safety and sustainability evaluation and assessment approaches that address market-relevant AdMas&Chems at all stages of their lifecycle and along the complete value chain. Together, these activities will form the foundations for the three components of the three PINK Business Case Innovations (see Section 2.2.9): 1) SSbD modelling services; 2) Computational SSbD Consultancy Services; and 3) Matchmaker and marketplace platform.
The EU SSbD Framework is based on a hierarchical approach with two high-level components, which are executed after each other in an iterative manner: (i) the (re-)design phase, and (ii) the safety and sustainability assessment phase. This is also labelled as a “stage-gate approach”, where AdMas&Chems are first designed and then they are assessed if they are SSbD.
PINK will follow the guidelines and recommendations of the Framework and further increase their practicability by implementing the evaluations criteria into a flexible computational SSbD workflow and an AI-driven decision support system (see details in Subsection III below) exploiting computational new-approach methodologies (NAMs) as well as data retrieved from in vitro NAMs, integrated approaches for testing and assessment (IATAs) and automatic data curation from databases and literature. To acknowledge the need of industry and specifically SMEs, PINK will deviate from the workflow concepts outlined in the EU SSbD Framework in three aspects:
Considering these three modifications to the concept, PINK will develop a Tiered Approach, where each development cycle will start with mining and creating data on the functionality, safety, and sustainability, as well as cost/economic feasibility, which will be provided simultaneously as input for the decision support workflow.
The figure above provides a schematic representation of the PINK Tiered Approach in comparison to the stage-gate approach of the EU SSbD Framework.
The middle part describes how different properties advance as the design project move from lower to higher PINK Tiers: the criteria to be included and the methods evaluate them will depend on the amount of AdMas&Chems candidates identified in a specific development stage and the knowledge about them. The PINK Tiered Approach will categorise all computational methods into three tiers that parallel the material development TRLs, with respect to their data needs, throughput, and confidence levels.
At earlier development stages (ideation/design phase of material development), where less data is available, high-throughput computational approaches of low tiers will be applied to avoid limiting the exploration and missing promising design routes due to time constraints (human and computer) and costs. This will allow prioritisation of selected routes and candidates, which can then be further investigated using more expensive but also more accurate methods of higher tier. Suggestions for experimental data generation (in vitro NAMs and later in vivo and pilot-plant level) as input for the next cycle will also be proposed.
To integrate the different data, models and software, as well as the services building on top of these (chemical/material dashboard, decision support system and Generative AI) into one Open Innovation Platform, a major part of the PINK Project is devoted to achieving technical and semantic interoperability. This will be based, on the one hand, on microservices providing the data resource, model and software and, on the other hand, on data documentation and annotation on different levels based on clearly defined data models, combination and harmonisation of chemical, material, biological and medical ontologies as well as ontologies to annotate (meta)data models. The resulting semantically annotated application programming interfaces (APIs) will be able to provide the input for (i) the PINK Registry, which indexes all available services and data, both publicly available as well as confidential when deploying PINKISH as an in-house software, respectively, (ii) workflows combining different services to evaluate complex criteria (e.g. using physics-based simulations to provide input for LCA analysis or implementing IATAs), (iii) integrated knowledge graphs construction in form of the PINK Knowledge Base (PINK KB) as well as the Chemical Impact Assessment knowledge graph, and (iv) the components of the central PINKISH platform, the decision support workflow, the Generative AI and the Dashboard. A schematic representation of the modelling and interoperability software stack is presented in the figure to the right.
