This article presents the results of a comparative study of finding the best detector to identify SS316 in non-aluminum metals and alloys. The results demonstrate that both PIN detectors and SDD detectors are equally good in the identification of several common nonaluminum metals and alloys. However, for this application, the lower cost and more perfect operation of a PIN detector make it a better choice over an SDD detector.
Comparison of Detectors
A comparison of typical SDD and PIN detectors is shown in Table 1. The performance of a typical SDD detector is better than a PIN detector. Besides having a better ultimate energy resolution, SDD detectors can count more X-rays in a given time. The resolution is crucial in determining X-ray events from various elements, while counting is essential to obtain better statistics in less time.
Table 1. A brief comparison between typical SDD and PIN detectors.
||Detector Inner Temp.
||Upper Input Count Rate (ICR) Limit
||10 - 50mm2
||120 - 160eV
||-20 to -40°C
||5 - 15mm2
||150 - 220eV
||-20 to -40°C
A typical PIN detector is less expensive and is often utilized in price sensitive systems such as XRF. PIN detectors are suitable for several XRF applications, including metal and alloy determination, which do not require the performance advantages that the SDD detector provides.
Figure 1 shows a standard XRF set-up used in this experiment, which involved the use of Moxtek’s 50kV, 4 Watt Ultra-Lite X-ray source with a tungsten anode. The source was set at 50kV and 15-20µA emission current. The distance between the source and the sample was 25mm and a 70µm copper filter was placed in front of the source. For each detector, the distance between the sample and detector was 25mm.
Figure 1. On the left is a sketch of the XRF setup, outlining the most critical parts. On the right is an image of the set up where all the components including collimators can be seen.
Moxtek’s MXDPP-50 processed the detector signal. The X-ray source and the detectors validated had aluminum sleeved brass collimator over them. The aluminum sleeve was required to prevent the stray XRF signal from the brass in order to ensure that the XRF signal is only from the sample. Most of the X-rays from the source below ~15keV were eliminated by the copper filter, providing a better signal-to-noise ratio in this region. However, the copper does allow one tungsten Lá line through at ~8.3kV.
The excitation from Nickel and the lower Z elements is improved by the tungsten Lá line, which also produces a non-XRF peak. This may confuse the inexperienced operator or an XRF algorithm. The XRF spectra from a clean plastic sample are illustrated in Figure 2, showing the Compton scattered tungsten Lá line and Compton scattered bremsstrahlung from the source.
Figure 2. Spectra collected from a plastic sample, which shows the Compton scattered background from a clean XRF sample.
The experiment involved the comparison of SDD, XPIN6 and XPIN13 detectors for their XRF performance to determine non-aluminum alloys and metals. The key technical specifications of each detector used in this experiment are summarized in Table 2.
Table 2. A functional comparison between an SDD, XPIN6 and XPIN13 detector
||Fe55 FWHM resolution
||SS316 Spectra counts in 30 sec
||DPP Peaking time
Each detector was run at a 30% dead time by modulating the tube emission current to compare the performance of a quazi-normalized XRF. The SDD has better technical performance when compared to the PIN detectors.
This XRF setup is not suitable to detect elements below calcium on the periodic table. Hence, the elements of chromium and above will only be determined in the SS316. An XRF spectrum from the SS316 source was recorded by each detector for 30 seconds. The full spectrum from the SDD, XPIN6, and XPIN13 is depicted in Figure 3, revealing all of the major elements. The spectral data focused on the region of 5-9keV are shown in Figure 4. As can be seen, the separation of Ká lines of all the elements is well defined for clear identification.
Figure 3. XRF Spectra collected from a 316 stainless sample, over 30 seconds, with all the major peaks labeled. Y axis of counts is in a log scale.
Figure 4. The same XRF Spectra collected in Figure 3, with a linear y-scale and energy range from 5 to 9keV.
The spectra acquired from the SS316 sample were run through an XRF fundamental parameters (FP) routine to convert them into elemental concentrations. Each detector, when properly configured in the FP program, provided almost the same results. The resulting concentrations from each detector are outlined in Table 3. The last column provides the total number of X-ray events identified in each of the spectra.
Table 3. The resulting concentrations from the three compared detectors using the 30 second and 10 second scans at ~30% dead time from a SS316 sample
||Total counts in spectra
Each detector was used to take 30 second and 10 second XRF scans, which were then compared. The results showed that all detectors were able to identify in 10 seconds. The higher count rates of the SDD are not actually required for sub-percent level element identification for metals and alloys. Each detector provided an elemental concentration within 1% or less for each of the elements. This level of accuracy is sufficient for determining SS316.
The results clearly show that both PIN detectors and SDD detectors are equally effective in determining several common nonaluminum metals and alloys. However, for this application, the lower cost and more perfect operation of a PIN detector make it preferable over an SDD detector.
This information has been sourced, reviewed and adapted from materials provided by Moxtek, Inc.
For more information on this source, please visit Moxtek, Inc.