Hisashi Homma offers a comparison of the fusion method and pressed powder method for cement powder sample analysis
The fusion method and the pressed powder method are established sample preparation techniques in X-ray fluorescence analysis of powder samples. In the fusion method, a sample is fused with some alkali borate flux at high temperature to make a glass bead, enabling a precise analysis as it provides a substantially homogeneous sample specimen for analysis. It is, however, a costly and time-consuming technique, requiring an expensive flux, fusion machine and trained operator.
In the pressed powder method, a briquette is formed by pressing a ground powder sample with a ring or cup using a pressing machine. The pressed powder method is a very simple technique that requires no skills or experience. Therefore, this method is suited for rapid analysis or analysis of large quantities of sample.
In industrial and research fields, the X-ray fluorescence technique is commonly used to control manufacturing processes, to monitor product quality and to determine geochemical characterisation.
This report provides a comparison between analysis results using the fusion method and the pressed powder method applied to cements, focusing on sample preparation methods in XRF analysis and regards the differences observed between the two methods as being mainly caused by differences in sample preparation. The results shown here should be regarded as qualitative observations.
In the analysis, certified reference materials of cement issued by the National Institute of Standards and Technology (NIST) were used. In preparing the sample for the fusion method, 1.0g of sample was ignited at 950°C in advance and 4.0g of lithium tetraborate flux was exactly weighed and then the two were mixed together. The mixture was fused by a fusion machine at 1,200°C. Nine fusion beads were used as calibration standards. Matrix effects were corrected by the theoretical alpha method. The values normalised without ignition loss were used for calibration standard values because all samples were ignited.
The pressed pellets were prepared by grinding with the tungsten carbide vessel. The grinding was performed under wet conditions with n-hexane as a grinding aid. After grinding, well-dried samples were pressed into an aluminum ring at 150kN. Seven pressed pellets were prepared for calibration standards and used to calibrate a Rigaku ZSX Primus IV WDXRF spectrometer. The theoretical alpha correction was applied to correct matrix effects.
Seven Portland cement standards used as calibration standards were analysed for testing. The accuracy of the calibration is summarised in Table 1. The results of the fusion method were recalculated to the values including ignition loss. In cement analysis, the fusion method shows good results for elements lighter than sulphur. No difference is found between the two sample preparation methods for elements heavier than phosphorus.
The results show little difference, except for certain light elements. When the differences among materials observed in this analysis were mainly caused by mineralogical effects, it can be attributed to the manufacturing process whereby limestone, clay, silica and iron oxide are mixed together and heated at high temperature to make clinker and then the clinker is mixed with gypsum to make cement product.
The main component phases comprising the clinker are four minerals (alite, belite, aluminate phase and ferrite phase). The properties of cement depend on the proportion of these four phases. Most types of cement are composed of 70-80% alite+belite.
The results demonstrate that the fusion method and the pressed powder method for cement show no significant difference except for some light elements. Although it is true that the fusion method eliminates grain size and mineralogical effects in powder sample analysis, observations in this report indicate that analytical results for the fusion method are not generally better than the pressed powder method, apart from elements lighter than sulphur.
Hisashi Homma is with Rigaku
