An evaluation of five systems showed broad performance differences.

Energy dispersive x-ray fluorescence (XRF) is ideal for materials screening for RoHS compliance. The fluorescent light is called the characteristic x-ray of the element. In a given element, energies between two specific orbital shells are always the same. Photons emitted when an electron moves between these two levels will always have the same energy. Therefore, it is possible to determine the identity of that element by determining the energy of the x-ray light (photons). There are two types of XRF: handheld and desktop from different vendors. To understand individual XRF capabilities and accurate levels, we conducted this study in two phases.

Phase 1: Evaluate five XRF systems with standard samples; compare Cpk; Gage R&R; stability test and detection level versus acquisition time.

Phase 2: Correlation study of production sample results from XRF against XRF or ICP method from outside two laboratories. The section of experiments and analysis describes experiment design, data collection, and analysis that included the results from five XRF equipments and two test laboratories.

The purpose of this study was not to identify superior equipment, but to show each XRF analyzer’s technical performance compared with external test laboratories.

Experiments and Analysis

We selected two handheld and three desktop XRF spectrometers because of their main specifications, performance and cost. They are Vendor 1-D, Vendor 2-D, Vendor 3-D, Vendor 4-H and Vendor 5-H. We used this XRF equipment for Cpk, stability, Gage R&R, detection level versus acquisition time studies with 12 standard samples, and correlation studies with 11 production samples.

We found:

• Energy dispersive XRF is one of the more suitable analysis methods for parts screening because of its nondestructive, fast and efficient analysis.
  
  
• Each XRF has different accuracy levels for measuring banned substances. We suggest obtaining deviation (different) results for main elements of the XRF in order to learn which samples need to be sent to an outside laboratory for confirmation of RoHS compliance.
  
  
• In general, desktop XRF performed better than handheld XRF. However, the handheld XRF from Vendor 4 performed better than some desktop XRF. Table 19 summarizes the performances.
  

• Choose a reliable test laboratory for identifying samples that are inconclusive (in between the RoHS-compliant and noncompliant ranges).
  
  
• The experiment results listed are primary studies with XRF for RoHS enforcement. The results obtained from this evaluation showed the equipments’ capability to quantify banned substances by RoHS initiative and demonstrated that XRF technique is a feasible testing method for manufacturing to perform quick verifications. However, a safety margin must be chosen to prevent equipment uncertainty from affecting final decisions.
  
  
• Despite the benefits of this technique, important limitations should be considered, such as the inability to quantify PBB, PBDE and chromium VI; the inability to detect banned substances in the inner part of a thick component/part; the inability to differentiate the banned substance contents of a nonhomogeneous component/part (the XRF result is a sum of different homogeneous materials), etc. A trial run is highly recommended to study effectiveness before final approval. Also, a well-trained operator is required to judge the results.

References

1. Rob Rowland, “Demonstrating RoHS Compliance User Perspective,” Web presentation, Aug. 31, 2006.
2. IEC-62321, Ed. 1, Environmental Standardization for Electrical and Electronics Products and Systems, 2005.
3. Flextronics XRF evaluation report, September 2005.
4. Dongkai Shangguan, “Lead-Free Solder Interconnect Reliability,” ASM International, 2005.

Acknowledgments
The team of Flextronics GDL Technology Center, Mexico: Maria Luisa Hernández Eusebio, Juan Coronado, Omar Garcia, Martin Murguia, George Liu, Amador Virgen Oneida, Monreal Zárate Yolanda, Santos Ruiz Luz Arcelia, Aceves, Conrique Carlos Humberto and Jayapaul Basani; the five XRF vendors’ support engineers, and the two external test labs.

Ed.: This article was first published at Apex in February 2007 and is published with permission of the authors.

Hector Rene Marin Hernandez, RefugioVicente Escobedo Alva, Zhen (Jane) Feng, Ph.D., Joao Ofenboeck, and Murad Kurwa are with Flextronics International (flextronics.com); jane.feng@flextronics.com.