Non-Invasive Detection of Corrosion Through Odour Emissions: Sniffing Out Rust

Corrosion poses a significant challenge in various industries, necessitating continual monitoring to avert disastrous consequences. Our objective is to devise an alternative method for corrosion detection by identifying by-products resulting from the redox reactions or any catalytic reactions of rust with its surroundings, leading to the release of odorant chemicals. Subsequently, using an electronic nose (AlphaMOS FOX4000), we explored differences in the chemical headspace between rusted samples and controls. The results indicated discernible variations in the chemical compounds within the rusted sample environments compared to the control samples, including metal, air, and water blanks, over specified corrosion time intervals ranging from one hour to two months. Application of Principal Component Analysis (PCA) unveiled distinctive clusters among corroded samples at different time points, illustrating notable disparities between rusted and non-rusted samples, as well as variations among the corroded samples based on the duration of corrosion.

I. Introduction

Corrosion’s impact on critical systems, like nuclear power plants, was evident in EDF France’s 2021 incident where corrosion on safety injection system pipes led to 12 reactors being shut down. The repercussions span from decreased efficiency to costly repairs and safety hazards, emphasizing the need for non-visual corrosion detection methods [1]. The electrochemical process behind corrosion, particularly iron oxide (rust) formation on iron alloys, is important to detect as its porous nature allows continued corrosion by enabling water and air penetration, furthering the development of corrosion [2]. Various detection techniques exist, categorized as non-invasive and invasive, with non-invasive methods favoured for preserving structural integrity [3]. Amongst these, Fjellanger Detection and Training Academy, used trained canines to detect Corrosion Under Insulation. This innovative method showed the potential in detecting corrosion, albeit with some challenges in chemical identification through gas chromatography. The use of dogs demonstrated a 59% proficiency in corrosion detection, supported by double-blind tests [4]. Due to the challenges associated with accessing all areas of a plant/factory and the possible safety implications for the animals, we investigating if this concept can be replicated using an electronic nose (eNose). In the future, this eNose could be made to be small and portable to detect odours in areas challenging to access or that has safety risks to workers. To achieve this, this paper aims to investigate the presence of rust using an eNose in a laboratory environment as a first step towards this goal.