TECHNICAL NOTE Ludwig Niewoehner, 1 Ph.D.; Jan Andrasko, 2 Ph.D.; Jan Biegstraaten, 3 B.Sc.; Lawrence Gunaratnam, 4 M.Sc.; Sylvia Steffen, 1 Ph.D.; Steffen Uhlig, 5 Ph.D.; and Sabine Antoni, 5 M.Sc. GSR2005—Continuity of the ENFSI Proficiency Test on Identification of GSR by SEM ⁄ EDX ABSTRACT: Within the framework of the ENFSI Expert Working Group ‘‘Firearms’’ a proficiency test on the detection and identification of GSR by energy-dispersive scanning electron microanalysis (SEM ⁄ EDX) is carried out in a 2 years term. The latest test was performed in 2005 ⁄ 2006 and was denoted as GSR2005. Seventy-five laboratories from 28 countries participated in this proficiency test and submitted in total 83 independent data-sets. The participating laboratories were requested to determine the total number of PbSbBa containing particles on a synthetic test sample fol- lowing their own laboratory specific methods of automated GSR particle search and detection by SEM ⁄ EDX. Furthermore size and position of the detected particles had to be reported by the laboratories and were evaluated statistically. The results were compiled by means of z-scores according to the IUPAC and EURACHEM guidelines—assessing individual laboratory achievements (inter-laboratory) as well as intra-laboratory perfor- mance—and were compared to the results of the previous proficiency test run GSR2003 (1). The comparison shows that there is a noticeable improvement in the method's detection capability. KEYWORDS: forensic science, gunshot residues, scanning electron microscopy, GSR, SEM ⁄ EDX, ENFSI, proficiency testing, z-scores The ENFSI proficiency test program on identification of GSR (gunshot residues) is an ongoing proficiency test organized by the ENFSI expert working group ‘‘Firearms’’. The history of these tests, GSR2001 and GSR2003, can be read in previous reports (1,2). The aim of the GSR proficiency tests is not a competition between the participating laboratories. The aim is to detect defici- encies in instrumental settings, hardware and software used, defi- ciencies which result in either missing particles or multiple detection of particles. The participating laboratories are asked to use their standard operating procedures for the detection of GSR particles by scanning electron microscopy equipped with X-ray microanalysis (SEM ⁄ EDX). Materials and Methods The test items for the GSR2005 proficiency test consisted of spe- cially prepared, identical samples in accordance with the ISO 5725 for the performance of proficiency tests (3). The samples were prepared artificially (1) and consisted of a number of particles with the composition characteristic for real GSR—namely of lead (Pb), antimony (Sb), and barium (Ba). As in previous tests, submicron particles have been included in order to evaluate the capability of some laboratories in detecting submicron particles using their standard parameter settings in routine analysis. Each of the samples contained in total 100 particles in four differ- ent size classes: 0.5, 0.8, 1.2, and 2.4 lm in diameter. For further statistical evaluations only the number of particles per size category was slightly changed but remained comparable to that used in the previous test (GSR2003) (1). Three different layouts were applied within the sample production, where the location of the 100 PbSbBa-particles was varied. Sets of identical samples containing synthetic GSR particles were dispatched to the participating labora- tories. The number, the size, and the location of the particles had to be reported to the organizers within a limited time period. Test samples were sent to 81 laboratories, which had accepted the conditions of participation (Table 1). This number is remark- ably higher than the 56 laboratories in the previous test (1). Not only European laboratories, but also 25 non-European laboratories participated in this test. Altogether 83 data sets from 75 laboratories were considered in the statistical evaluation of the test (some labo- ratories submitted more than one result for different SEM ⁄ EDX systems). After the test, the participating laboratories received a complete evaluation of their own results in terms of finding and ⁄ or missing particles. In this way, the participants are given the opportunity to comment on their performance assessment. Additionally, laborato- ries with poor results may question their settings and procedures and, as a consequence, improve them. Laboratory evaluation and data assessment were performed in the same way as reported in the previous studies (1) according to ISO 13528 (4). An assessment of the laboratory's capability to detect GSR particles by SEM ⁄ EDX was carried out using z-scores according to IUPAC and EURACHEM (5,6). Therefore the robust statistical method of ‘‘Q ⁄ Hampel’’ was used which meets the standard DIN 38402 A45 (= ISO ⁄ DIS 20612) (7,8). The values of all characteristics used for the calculation of the z-scores are shown in Table 2. The evaluated data of all participants are sum- marized in Table 3. 1 Forensic Science Institute, BKA, D-65173 Wiesbaden, Germany. 2 National Laboratory of Forensic Science, SKL, SE-58194 Linkoeping, Sweden. 3 Netherlands Forensic Institute, NFI, NL-2497 GB Den Haag, The Netherlands. 4 National Bureau of Investigation, NBI, PL 285, FIN-01301 Vantaa, Finland. 5 QuoData Ges. für Qualitaetsmanagement und Statistik mbH, Kaitzer Str. 135, D-01187 Dresden, Germany. Received 15 Jan. 2007; and in revised form 1 June 2007; accepted 3 June 2007. J Forensic Sci, January 2008, Vol. 53, No. 1 doi: 10.1111/j.1556-4029.2008.00594.x Available online at: www.blackwell-synergy.com 162 Ó 2008 American Academy of Forensic Sciences