The predominant method for positive material and contaminant identification is Fourier Transform Infrared (FTIR) Spectroscopy. However, FTIR’s lack of sensitivity towards inorganic and metallic components can contribute to both ingenuine materials having a positive identification, and the inability to identify metallic contaminants. To fill the void, another method can be incorporated, one that has an increased sensitivity towards metallic components. There are several potential candidates, including atomic absorption, inductively coupled plasma mass spectrometry, and atomic emission spectroscopy; however, X-ray fluorescence (XRF) has attributes that make it ideal for positive material and contaminant identification.

XRF is the Perfect Complement to FTIR for Positive Material and Contaminant Identification

Most analytical techniques that are sensitive towards metals are destructive, time-demanding, consumable intensive, and have a large footprint, however, XRF is none of these. XRF is a nondestructive, inexpensive, rapid – often on the scale of seconds – technique that has the footprint of anywhere from handheld to floor standing.

XRF determines the elemental composition of a sample by measuring the fluorescent X-rays generated from a primary X-ray source, typically an X-ray tube. A typical XRF spectrometer can measure elements from sodium to uranium, with detection limits for metallic elements being in the ppm to sub-ppm level. XRF’s intrinsic metallic sensitivity fills the void that FTIR leaves, making it the complementary tool for positive material and contaminant identification.

To illustrate the utility of XRF, the infrared spectra of two batches of polyethylene and the accompanying XRF spectra are presented. The FTIR spectra are identical; however, as apparent from the XRF spectra, one sample is loaded with ppm levels of metals. In this specific example, it is not hard to imagine that if FTIR was only used, the material may have passed a positive identification.

In addition to positive material identification the XRF-FTIR combination also offers a solution for contaminant identification. FTIR, like XRF, needs little sample for analysis, which benefits contaminant analysis since sample amount is often limited. There are sample holders on the market that let the same exact sample be measured by both XRF and FTIR, and since both techniques are nondestructive, the sample can be stored for archiving. These type of sample holders have an adhesive coating that prevents the sample from moving, which is necessary for FTIR measurements in the attenuated total reflectance configuration. The sample holders close like a sandwich for the XRF measurement. Since the sample holder closes, both liquid and powder samples can be measured and archived.

Reducing the Workload: Library Search-Match Algorithms

In an ideal situation, every contaminant and raw material would have its own XRF and FTIR spectrum, which can be used to compare to unknown contaminants or incoming materials. As could be imagined, visually inspecting the unknown with a library of FTIR and XRF data can be laborious and time consuming. To alleviate the workload, library software incorporating search-match algorithms allow users to both search-match spectra and generate a “hit-list” of material and contaminant identification matches, thereby bypassing the manual search and comparison process.

The crux of library search software is that a library of data, in this case FTIR and XRF data, is used to compare to an unknown material or contaminant. Using a metric that compares the similarity between the unknown and all the entries in the library, a hit-list is generated that is indexed according to the similarity metric. For standalone users of FTIR and XRF, search-match software is a common feature that comes with instrument software.  However, there are search-match software packages that combine data from both FTIR and XRF, and output a “composite index,” an index that integrates both XRF and FTIR data.

Computing the Composite Similarity Index

To generate a composite similarity index, the search-match software algorithm first looks at the XRF data and determines whether the material is inorganic, organic, or a mix. The determination is done by measuring the background in the XRF spectrum. In short, the background in XRF is based on Compton scattering amplitude. Materials that are composed of light elements, namely organics, have a much larger Compton scattering amplitude compared to heavy, metallic-like, materials. Hence XRF can distinguish whether a material is organic, or inorganic in composition.

After the algorithm measures the XRF background and bins the sample into inorganic, organic, or mix, the algorithm will measure both the FTIR and XRF spectrum and generate a respective similarity index for both, as if they were used in an independent fashion. Before we discuss how the composite index is generated from a combination of background classification, and the two similarity indexes (FTIR and XRF), it is useful to discuss the nature of the similarity index and how it is calculated.

How the actual algorithm calculates a similarity index is different between FTIR and XRF. For FTIR, the similarity index is calculated as a sum of the square of the residuals between two different spectra, then normalized to a specific value, often one. For XRF, the similarity index is calculated from the elemental composition differences between samples and, unlike FTIR, the raw spectrum.

The point regarding how the similarity index for XRF is calculated cannot be understated. The XRF similarity index is based on elemental composition, which has near to no dependence on sample quantity. XRF insensitivity towards sample amount is due to the elemental composition being determined via the ratio of peaks, instead of absolute peak area. For contaminant analysis where sample volume can range from grams of powder to microliters of liquid, it is important for the XRF similarity index to be independent of sample quantity.

After the two similarity indexes are determined, the composite index, which is a weighted average of the two indexes, is calculated. The weight is based on the previously mentioned XRF background data, and as a result, whether the material is inorganic, organic or a mix. For materials that are organic, the composite index is weighted towards the FTIR data, and for inorganic material, the composite index is weighted towards XRF data. For a mix, both XRF and FTIR are equally weighted.

The overall effect of the composite index is to provide for both better selectivity and sensitivity towards library matches. In other words, if the data were not weighted, there would be more ambiguity in the library hit-list. The implementation of the composite index avoids this ambiguity.

Looking at the Big Picture: XRF, FTIR and Search-Match Software

In a globalized world where multiple vendors can be supplying the same material to one customer, it is possible that vendors will change the material for the purpose of reducing cost, a phenomenon known as “silent change.” Silent change can have far-reaching consequences, both fiscal and societal. The ever-increasing complexity of manufacturing lines also introduces a new scope of potential contaminants that previous manufacturing sites would never encounter. There is a litany of analytical tools for both positive material identification and contaminant analysis, but realistically, none of them are as simple, rapid, inexpensive, and user-friendly as XRF and FTIR.

With regards to identification, XRF and FTIR are not useful without proper search-match software. With modern search-match algorithms, which were traditionally used by other analytical methodologies, but have now spanned into FTIR and XRF, contaminant and material identification became faster and easier. The next step was to develop a software algorithm that utilized both FTIR’s sensitivity towards organics, and XRF’s sensitivity towards metals, and manifest it into a so-called composite index.

By utilizing software that integrates both methods, XRF and FTIR has now become more than the sum of their parts, and the ultimate tool for positive material and contaminant identification.