Fresenius Zeitschrift fiir Fresenius Z Anal Chem (1988) 331:725-729 9 Springer-Verlag1988 Critical comparison of decomposition procedures for atomic absorption spectrometric analysis of prehistorical ceramics J. A. Stratis 1, G. A. Zachariadis 1, E. A. Dimitrakoudi l, and V. Simeonov 2 1 Laboratory of Analytical Chemistry, Aristotelian University, GR-540.06 Thessaloniki, Greece 2 Chair of Analytical Chemistry, University of Sofia, A. Ivanov 1, BG-1126 Sofia, Bulgaria Kritiseher Vergleich yon AufschluBmethoden fiir die AAS-Analyse von pr~ihistorischer Keramik Summary. This paper describes an effort to compare acid and fusion decomposition techniques for ceramic samples taken from a Neolithic settlement. An interpretation of the results is given following two statistical ways, the analysis of variance (ANOVA) and rank correlation. Also an ideo- graphic approach is performed. Finally, conclusions are pre- sented about comparability of the above representative di- gestions. Introduction Chemical analysis of ceramic findings has always attracted the attention of many workers. The methods applied are quite numerous and they use the advantages of such tech- niques as NAA, XFR, AAS, AES etc. The most difficult and time-consuming step in the analytical procedure is the sample preparation. Generally, ceramics are consisting of oxides which have high melting points and low solubility, even in strong acids [4, 6, 12, 13]. The analytical scheme of ceramic analysis is given as block diagram in Fig. 1. The variety of fusion techniques is often confusing since there is a lack of critical comparisons of the possibilities of different techniques. Fusion techniques are often classified [1] in five general types. Alkaline fusions are using carbonate salts of sodium, potassium etc. or borate salts of lithium. In acidic fusions the currently used fluxes are LiF, KF, H3BO3- Na2B407 etc. In oxidizing fusions the unique flux is Na202 or its mixture with NaOH. Reducing and thioalkaline fusions are rarely reported for ceramic analysis. On the other hand, acids have been used for ceramic decomposition. The use of acids has the advantage to keep the final solution practically free of flux interferences. In this field hydroflu- oric acid has solved many analytical problems and is used successfully until now in mixture with other acids, such as HNO3, HC104, H2804, H3BO3 etc. All the known classical and electrochemical methods were used in silicate analysis [4, 17, 20, 22]. But the recent development of spectrochemical and radiochemical tech- niques has simplified the analytical scheme and reduced the demand of time. Among these techniques AAS has faced the greatest application in multielement analysis of a wide variety of samples [8, 23, 24]. Neutron activation [7] and X-ray fluorescence [2, 21] are also very popular techniques but they need more expensive equipment. In this paper AAS was chosen as suitable to interface all the compared decomposition techniques according to the black line of the diagram in Fig. 1. It is obvious that a lot of features have to be taken into account in order to choose the most appropriate sample preparation procedure. The aim of the present study is to compare six different sample preparation methods for ceramic analysis by various means - ANOVA technique, rank correlation and ideo- graphic approach. It makes possible to choose one of the Classicat AcJ_d ottec___kk Electrochemical ~ Alkaline fusion . Spectroscopic \\/ Other fusion Rodioehemical Measurement ] Fig. 1 General scheme for ceramic analysis Offprint requests to: V. Simeonov