How to Analyze XRF Spectra

To elaborate on the usefulness of technology in the field of conservation would take more than this blogger could handle. So, I will limit it to one, and that is the usefulness of X-ray fluorescence. One thing that needs to be mentioned is the goal of conservation over the years has been to be able to analyse an object via non-destructive techniques. XRF is one such technique, which is fantastic. However, it does have its limitations, that being it is only useful if you are looking for elemental compositions of elements that are heavier and have a higher atomic number. Now if you remember way back when you took chemistry, you might be able to remember that the atomic number is the number given to an element that tells how many protons are in its nucleus, and therefore, how many electrons in an uncharged atom that are in its orbitals (this is also the number that gives the order of the periodic table, hydrogen having an atomic number of 1, helium having an atomic number of 2, etc.) Because of this difference, the energy from the x-ray beam that passes through the first few microns of material excites the different electrons in the different elements. Each element absorbs and releases different energies, and therefore appears in a different position on the spectrum that is produced using a specialty computer program. To put it a little more simply, each element is excited at a characteristic energy, which is recorded on a graph to make it simple to read.

For the pewter collection at the MCC, I have been working on analyzing the data collected a few weeks ago. So here is a sample spectrum:Now, in order to understand what you are looking at, you need to know what the graph is telling us. The y-axis (vertical axis) is giving the intensity of the energies being released by the individual elements. Meaning, the higher concentration of elements will release more energy giving a taller stronger peak. The x-axis (horizontal axis) gives a reading of the energy (in KeV). So, if we know what elements we are expecting, and what energy their excited electrons give off, we will be able to identify if they are present.  Originally you would see just the red line. But using a specifically designed computer program that is able to interpret the data produced by the XRF and give it to us in the graph form, there is a way to identify the different peaks and label them. As I was analyzing pewter, for this graph I knew from research that the elements of interest were tin (Sn), lead (Pb), copper (Cu), and antimony (Sb), different metals used in to make pewter alloys. I have iron (Fe) labeled as well, as one of my samples came back with a strong Fe peak. The information on the bottom is simply giving the machine specifications used for this sample; the important information being the Lsec: 30.16 (the amount of time the machine ran for, 30 seconds), and Kev:19.6635 (the central energy of the graph, which can be seen as a short black line coming from the x-axis). Other information that is important is the information regarding the handheld XRF used: including the energy of the beam coming from the XRF which is 40 kV; that a yellow filter was used; and that the machine was not used in a vacuum. This information is important in case someone else wants to recreate the experiment, they will be able to use the same settings on the XRF and hopefully get the same results. 

Now that you know what the graph is showing, I can talk you through what I have gotten from this graph. First off, I am only looking for qualitative data, which is just data thatcan be observed but not measured. The other type of data that can be obtained from XRF is quantitative, meaning it can me measured, whether it be length, percentage, weight, etc. This means that I was simply trying to see what elements were in the various samples of pewter to ascertain if there was a specific element that could tell if one of the samples was fake pewter or not. Using the same computer program, one way of testing this, was to see if the graphs had peaks at the same energies (in other words, if the same metals were in all the samples). This was completed by overlaying the graphs, putting one on top of another and see how they lined up.

 The red and the blue lines are two different samples’ spectra. From this, you can see that even though the peeks are different heights, all the peaks show up in the same place. From this I can gather that the same allow composition was used, a combination of Pb, Sn, Cu, Sb, and maybe even some Fe. One thing you might notice, however, is that the red spectrum has a higher/stronger Pb peak, whereas the blue spectrum has a higher/stronger Sn peak. From this it might be possible to get a semi-qualitative analysis, by saying that in the blue sample, there is a higher percentage of tin, and the red has a higher percentage of lead. But I am unsure if that is what is causing the difference. It could also just be from how much contact the xray beams got with the sample. Either way, There is not much more that can be gathered from this. If a quantitative analysis was desired, then there are other programs that can be installed that it is possible to get quantitative results. But for me, qualitative is enough.

As you can see, this sort of analysis can be a very fast and useful analysis technique when dealing with inorganic material like metals. Not only does it have an ease of use (a point and shoot), but it also has easy enough data to interpret, once you get the hang of it. And with it being a nondestructive technique, it is extremely useful for conservation.