23 June 2011 - Preparing to XRF
This week in the Regenstein Lab, we have been preparing to analyze a collection of carved and painted shields from Papua New Guinea using p-XRF. P-XRF stands for portable X-ray fluorescence. By bombarding the pigments on the shield that we want to identify with X-rays, we can see the chemical composition of the paint. The electrons in the cloud around the nucleus of each atom respond to the X-rays by jumping to a lower energy level and giving off some energy of their own. P-XRF measures the energy released and identifies the type of element in the sample.
But before we can actually use p-XRF on the shields, we have to make sure that we have perfected our technique. While p-XRF is a non-destructive technique, we can use the XRF machine to both measure samples in situ or remove a small amount of pigment and measure it in a small cup. We will be using both methods and before we remove any pigment from the shields, we want to be certain that we are taking off the smallest amount possible. First, we need very fine needles to remove a small bit of the paint. We made our own.
By dipping tungsten wire into a beaker filled with sodium hydroxide with a current running through it, you can electrolytically shave off tiny amounts of the wire. I repeatedly dipped the wire in the solution in increasingly smaller lengths, which narrowed the wire to a very sharp point. These needles can be used to pick up tiny samples that other, bigger tools might miss or destroy.
Next I optimized the technique that we would use to measure the small amounts of pigments removed from the shields. I measured increasing amounts of known chemicals like table salt and calcium carbonate to see exactly how much of a sample we would need to remove from the shields to get a good reading from the p-XRF machine. Not surprisingly larger samples give more accurate readings. But it does seem that we can get good results from a sample about the size of two or three grains of salt. That is promising considering we would like to take as little of the pigment off the shields as possible.
After all this preparation, the next step is to actually test the shields. First we tested them in situ with the p-XRF. We took a couple of reading from each of the colors of pigment on the shield and from the bare wood as a comparative sample.
Most of the shields are decorated with red, white and dark brown pigments. In the white pigments, calcium predominates, but we have also found a small amount of strontium. This confirms that the white paint is probably made from marine shells or coral since strontium is present in seawater. In addition, we have been finding some manganese along with iron in the red pigments. It will be interesting to see if there are any differences between the shields with manganese present in the red pigment and those without the element.