140 Prakt. Met. Sonderband 46 (2014) 
channel, e.g. Ca channel, that are coincidental or not coincidental with peaks observed on For sevel 
the channel(s) of other element(s), e.g. Al. These numbers are proportional to the number average 
of inclusions containing Ca and Al (e.g. Ca aluminates), respectively to the number of paramete 
inclusions containing Ca but not Al (e.g. Ca oxides). (Equivale 
The simplest Spark-DAT methods consist of evaluating inclusions at a qualitative level or the diame 
relative to reference samples by using these simple count numbers. The samples to intensity 
compare are ideally of the same grade or at least close in their composition. Although analysis f 
qualitative, these methods provide extremely useful information, in particular when 
performed in the production simultaneously with the standard spectrochemical OES 
analysis. An excellent illustration is the morphological control of inclusions at steel casting 
[6]. Ca treatment is performed in order to avoid nozzle clogging by Al.Oz inclusions at 
continuous casting. Ca addition transforms Al.Os inclusions into more globular, less 
clogging Al,O3-CaO inclusions. Excess of Ca additive, further transforms inclusions into 
CaS inclusions that change steel fluidity and cause casting problems. Spark-DAT methods 
can monitor the number of peaks due to Al,Os, Alz03-Ca0 and CaS inclusions. Comparing 
these numbers with ideal values allow very quick adjustment of Ca additive, i.e. add more 
to prevent clogging or less to avoid casting problems. A variety of applications of such 
simple methods exists, for example incoming metals testing and rapid inclusion screening 
in case of problems. 
The qualitative Spark-DAT methods also allow performing simple size distribution Fig.5 
evaluations. Because large inclusions are normally more detrimental to quality than Note 1 
smaller ones, it is very useful to have a method that discriminates inclusions by size. In the 
example of Fig. 4, the algorithms were used for counting the peaks (Ca tot and S tot) and 
coincidental peaks (CaS) at different intensity thresholds. This allowed counting the 5.2 QUA: 
inclusion belonging to three intensity classes, corresponding to three size classes “small’, ’ 
“medium” and “large”. This type of qualitative classification of inclusions is sufficient in the In man 
context of a production line where samples of the same grade have to be controlled. 0pm 
small medium large possible 
on calibre 
\ allow per 
1560 the quan 
se oxide inc 
SC 
GET Ca 
TERRA 
Fig.4: Qualitative size classification of CaS in a low alloy sample. 
5. ADVANCED SPARK-DAT INCLUSION ANALYSIS METHODS nd 
5.1 QUANTITATIVE INCLUSION SIZE ANALYSIS mold 
Recently we introduced innovative Spark-DAT methods able to determine the inclusions This me 
quantitatively [5]. determin