Predictive methods for determining the decomposition properties of hazardous substances: from development to experimental verification — HAZPRED

The problem addressed

Up to now, experimental characterization is used by laboratories to gather data considering physico-chemical hazardous properties of energetic substances. Due to the fact that these tests are complex, costly and require a large amount of substances (more than 1 kg for hazardous properties such as explosivity) and dedicated facilities (both small and large scale), alternative or predictive methods are more and more encouraged. In particular, the development and use of QSPR models are recommended by the REACH regulation. Moreover, QSPR or small-scale tests (requiring lower quantities of materials) present a real interest for screening purposes in R&D development phases (even when chemicals are not yet synthesized) or when they are very expensive or dangerous for the users (e.g. toxic). These both predictive methods can be used to replace huge experimental characterization and prevent that valuable time and costs are spent on potential late-stage failures. In this context, the HAZPRED project addresses the development and valorization (use) of predictive methods, in particular QSPR and small-scale tests, for investigating hazardous properties and behaviours of hazardous substances (like organic peroxides and self-reactive substances) in a regulatory or risk assessment context.

Keywords: hazardous property testing REACH

Research questions

Within the framework of the European REACH regulation1, a large number of (eco)-toxicological but also physico-chemical properties are required for the registration of substances in the perspective of industrial scale production. For reasons of time, cost, availability of laboratories and ethics (in case of experiment on animals), companies are encouraged to present alternative testing strategies, such as computational chemistry and molecular modelling tools. Indeed, the REACH regulation recommended the use of alternative methods to experiment and in particular QSPR methods that can predict properties using only the molecular structure of substances. These methods also benefit from recognized validation principles, established by OECD, to ensure their transparency, reproducibility and accuracy, in terms of fitting, robustness and predictive capabilities and to allow their use in a regulatory context.

Beyond this regulatory context and considering the fast development and availability of computers, such models allow industries to qualify new hazardous substances in a risk assessment process and developers of new chemicals to integrate safety considerations in the early steps of their development, even before the synthesis of identified molecules of interest. It can facilitate the rational design of new or alternative processes and products, for example when testing is expensive or impractical or when ‘green’ alternatives are a must. In this way, R&D can focus on the most promising candidates and this will cut costs and development time, allowing timely elimination of future failures. So, the integration of predictive and experimental knowledge will aid decision-making. While such models have already been developed for some physico-chemical properties, only few of them followed the different OECD validation criteria, in particular for hazardous physico-chemical properties (flammability and explosivity). First attempts were made to predict reactivity hazards of organic peroxides using the QSPR approach but a huge need is still growing to enhance the safety knowledge of these reactive substances. This proposal addresses the development of two complementary predictive methods, QSPR and small-scale tests, for investigating hazardous properties of these chemicals based on solid experimental databases.

Keywords: QSPR predictive method

Scientific disciplines: chemistry/chemical engineering, medicine/public health

Expected outputs

Results of the project will be helpful for different stakeholders such as regulators of chemicals and industry. Predicted data will guide decisions in the R&D of new (hazardous) substances. In this way, spending valuable time and money on potential future failures can be avoided or at least be reduced considerably. In particular, QSPR models will allow the a priori estimation of physico-chemical properties without any experimental testing, even before their synthesis, to identify at early steps of development high potential candidates and to eliminate potentially hazardous compounds. This in silico approach presents a great interest for industries in the context of the screening of properties but also for substitution or elimination of dangerous chemicals in the context of REACH regulation. The outcome of this work will also be beneficial for industries to become REACH-compliant.

Keywords: screening


The project will be divided into 4 technical work packages and another for the project management, that can be summarized as follows:

  • initial database set-up from existing information (WP 1)
  • consolidation of the database by experimental tests (WP 2)
  • development and validation of models (WP 3)
  • valorization and implementation of predicted data (WP4)
  • project management (WP5)

Associated deliverables

HAZPRED: presentation at fourth SAF€RA symposium
Presentation CC BY published on 2016-08-26
A presentation of project progress at the fourth SAF€RA symposium, organized in Athens in April 2016.

Download deliverable

Participating researchers

Patricia Rotureau (Direction des risques accidentels, Ineris, France) — project coordinator

Guillaume Fayet (Direction ds risques accidentels, Ineris, France)

Anka Berger (BAM, Germany)

Petr Lepik (Faculty of Safety Engineering, VSB – Technical University of Ostrava, Czech Republic)

Marola van Lipzig (TNO, The Netherlands)

Wim Mak (TNO, The Netherlands)

Antoine van de Heijden (TNO, The Netherlands)

Funding organizations

TNO (The Netherlands)

BAM (Germany)

INERIS (France)

CZ-TPIS (Czech Republic)

More details

Duration 2014-04 to 2017-06
Contact email
More information

Information last updated on 2016-08-26.

Return to the list of projects.