Most contributions present forward-directed research — from hypothesis through controlled experiment to replication. This contribution turns that direction around and focuses on industrial failure analysis. Here the starting point is not a research question but a failed component whose history is unknown. The problem is an inverse one: to reconstruct a defensible, actionable root cause from a single, non-repeatable event, with incomplete information and under time pressure.

Drawing on anonymized cases reduced to the materials level, this contribution shows how familiar analytical methods (FT-IR, DSC, SEM/EDX, and others) are used for targeted diagnosis rather than exhaustive characterization: converging lines of evidence instead of completeness, and only as much data as a decision requires. Each case takes a different failure in a plastic or composite component and works back to its cause.

This focus on how materials actually behave in service, rather than as idealized laboratory specimens, is what makes the topic valuable for an analytically minded audience. Behind every macroscopic failure lies a chemical story — chain scission or crosslinking, oxidation, the loss of functional additives, the intrusion of contaminants, or the weakening of interfaces — that links the visible damage to changes at the molecular level. Reading these molecular signatures from a degraded part offers insight into universal failure mechanisms, into the limits of individual methods under real-world loading, and into the logic of an analytical approach that reaches defensible conclusions despite uncertainty.