Improved production control through rapid characterisation of non-metallic inclusions in steel

Non-metallic inclusions can have strong negative effects on the properties of steel. They can act as internal defects in the grain structure and they can cause failures during production process as well as in the semi-finished or finished steel products. Especially in products with extremely small sizes like wire rods, low thickness steel sheets etc. the steel can be damaged during reforming process like deep drawing. Inclusions can as well reduce fatigue properties and therefore decrease the long term usability of the products. Other heterogeneous element distributions like segregations affect the material properties negatively, too. The entrapment of inclusions within the steel matrix never can be excluded but can be minimised by optimisation of the production process in the steel plant. Such clean steels are high-technology products. For process and product control an analytical tool is necessary which enables to characterise the steel cleanness within short time. State of the art methods to characterise the steel cleanness are electron probe microanalysis (EPMA) or optical microscopy - both methods carried out on polished sample surface. Additionally, chemical isolation techniques can be applied to dissolve the steel matrix without attack on the inclusions of interest. Both methods are time consuming so that no on-line analysis is possible for process control. Goal of this project is to develop a less time consuming method to detect and characterise non-metallic inclusions of >1 μm size in steel. Two different optical emission spectrometers using laser-induced plasmas for local element analysis have been applied. The one is able to make spatially resolved measurements by using a moving optics for focussing the laser beam. In the other case the sample is moved in a defined way while a pulsed laser-beam strikes the sample surface in a raster. After each pulse, the emission signals of a couple of elements is registered simultaneously in a commercial OE-spectrometer. Both spectrometers are able to focus a number of laser pulses on the same position of the sample. The process of material ablation and crater formation in the sample surface and of emission signal generation on the plain steel matrix and on inclusions were studied fundamentally. Statistical algorithms and multivariate analysis have been proposed and tested to distinguish between discharges on the matrix and on inclusions. Coincidences of significantly enlarged signals in inclusion specific element channels were evaluated and referred to the chemical composition of different inclusion types. Simulations of signal generation based on simplified models of laser-induced ablation and excitation are presented to be compared with the measured signal distribution. For the investigations of the physical phenomena resulting in different signal distributions, synthetic standards have been used as less complex sample systems. Samples of high grade steels, of electric quality and dynamo sheet, and of low- and micro-alloyed steels from different steps of the production process have been analysed by various current reference methods. Correlation with the results of laser-OES have been leading to first estimations of quantification of inclusion contents.