One of the key challenges to good XRF sample preparation is contamination. Contamination comes from two main sources: the sample preparation device and sample to sample cross contamination.
Contamination from the Sample Preparation Device
Sample pulverizers are usually the sample preparation device that has the potential to contribute the most contamination to your sample. Pulverizers usually use a ring and puck style grinding bowl to grind samples from coarse chips to a fine powder. This grinding action can add various elements contained in the grinding vessel to your sample which is why you should be careful about choosing your grinding medium. Steel, tungsten carbide and ceramics such as alumina and zirconia are the most common choices. Steel can add Fe, Ni and Cr, tungsten carbide will add W and alumina and zirconia will add Al and Zr respectively. Think about what elements you are interested in analyzing and make your choice accordingly. Tungsten carbide is often a default choice because of its hardness and because W is usually not a key element for analysis for many applications. The downside is that tungsten carbide is often the most expensive option.
Sample to Sample Cross Contamination
Sample to sample cross contamination in the sample preparation process is by far the biggest contributor to sample contamination. This is particularly true if you are working in an environment where you are analyzing a wide range of sample types.
There are three main ways to combat sample to sample cross contamination
- Bowl Cleaning- Cleaning the sample grinding vessel between each sample. This can be accomplished with vacuum and compressed air. This can be performed manually or automatically in a fully automated system. This works well when preparing samples that are relatively chemically similar.
- Pre-contamination- a split of the sample to be prepared can be ground in the vessel and discarded before grinding the split of the sample to be analyzed. This approach usually removes most of the residue from the previous sample, leaving only residue of the sample to be analyzed. This approach works really well for most samples and can easily be accomplished in automated systems. The diagram below shows the results of an experiment performed using limestone and a steel sinter using a Centaurus automated mill and press. Here iron is the main contaminant. The baseline shows the expected Fe2O3 composition of the limestone. The contamination experiment was conducted by grinding a limestone sample (Fe2O3 = 2.4%) alternated with a steel sinter (Fe2O3 = 84%). The limestone split used to pre-contaminate the grinding vessel has elevated Fe2O3 levels and is discarded. The splits used for the analysis are all within the 3 standard deviation limit for the Fe2O3 composition of the limestone established by the baseline samples. The data show that the pre-contamination cleaning of the vessel was effective.
- Cleaning Media- Another approach to reducing contamination is to clean the grinding vessel in between samples with a hard cleaning media such as quartz sand. Grinding with the sand scours out the mill and then the sand can be discarded prior to grinding the sample to be analyzed. The diagram shows the results of an experiment similar to the previous example with the exception that sand cleaning is used between grinding of the limestone samples without additional pre-contamination of the grinding vessel. The data shows similar results to the previous example with the Fe2O3 content of the limestone samples lie within 3 standard deviations of the baseline value.
All three approaches to reducing cross contamination are valid depending on the needs of your particular laboratory. These methods are not mutually exclusive and can be combined for particularly challenging samples. The choice of your approach should be based on the types of samples you analyze, your analytical goals and the speed required for the overall analysis.