Gaseous mixtures of hydrogen and oxygen can ignite, spontaneously producing large amounts of water and heat. Here, we will describe an analysis of the information anatomy of this transient phenomenon -- a set of information-theoretic measures that include multivariate mutual information and transfer entropy. We diagnose this anatomy from symbolic time series that we extract from thousands of ignition events simulated with molecular dynamics and ReaxFF. A prominent feature is the "enigmatic information", which exhibits a minimum value at the time of the earliest observed ignition event. These data show that the onset of ignition can be predicted from the temporal correlations in the fluctuating composition variables of an igniting mixture as much as 0.83 nanoseconds before the mean ignition-delay time. They also point to the possibility of mining robust predictors from the harsh and dynamic conditions of ignition at the atomistic scale.

Information anatomy of igniting hydrogen-oxygen mixtures

256th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.