The first step towards using AES for scrap sorting was taken with manual spectroscopes but with their low D and the relatively high demands made by users prevented general application. The more convenient table spectroscopes, which had higher D and were able to compare spectra, allowed sorting of high-grade scrap such as eg. bronzes, admittedly with wide tolerances.

  Photographic spectra recording and evaluation, which became very important in the analytical field, had to be ruled out because of the long access time for results and the high cost involved in sample preparation for sorting purposes.

Although the introduction of photoelectric measurements of line intensities by means of laboratory spectrometers allowed rapid spectra evaluation, sample preparation, as before still proved to be an obstacle for sorting applications.

  The development of mobile test probes with lightguides opened the way for widespread use of AES for testing and sorting large quantities of materials.

  Developments in the laboratory AE spectrometers with SDAR for metal analyses had led to RSD≈1%, SR 1-5% and LOD of 10µg/g. In a well organized and suitably equipped spectrometer laboratory for melt control, a time of 3 min was achieved from sample preparation to analysis output.

  Because of the influence of sample structure on spectrometer results, the analysis of semi-finished and finished material was only carried out after remelting in remelting furnaces specially developed for this purpose. The structural isoforming process developed about 1970 using HEPS led to increased use of laboratory spectrometers with SDAR for outgoing and incoming inspection. Today, more than half of these spectrometers are used in this field, more than are for melt control.

  Incoming and outgoing inspection was not carried out systematically for each piece, but only when it was suspected that material had been mixed-up. This inspection and the organization required were expensive: samples had to be taken from the material, they had to be numbered, transported for preparation, analysed and re-allocated at the production plant. Mix-up often occurred with this procedure.

  I can still hear what a Plant Engineer at a steelworks said to me regarding the importance of finished material inspection: “ To us at the plant, what use are the best spectrometers in the Laboratory with an accuracy of 1% rel. if 10% of the qualities are mixed up by the time they are loaded?”

  Inspection can only be sensible if the test equipment goes to the job in the plant, not if the job goes to the test equipment in the Laboratory.

  Classical methods such as magnets, spark grinding and the spectroscope had proved satisfactory for practical applications, but no longer met most requirements because of their “lack of definition”.

  It should not go without mention that analysts at large plants often have little appreciation of the work done by Quality Control staff because they do not understand the work. When an analyst has to make an analysis, the more elements with more digits after the point the better. Quality identification by a Quality Control man can often be done by a simple statement of “yes” or “no”, of which some examples are:

  Al-, Si-killed steel

  Pb-containing free-cutting steel

  Mo-, Ti-, Nb-stabilised CrNi-steel.