Environ. Sci. Technol. 2008, 42, 4088–4092
Preparation of Electrotechnical Products for Reduction of Hazardous Substances Compliance Testing DAMIAN B. GORE,* ERIKA S. HEIDEN, AND RUSSELL J. FIELD Department of Environment & Geography, Macquarie University, NSW 2109, Australia
Received July 16, 2007. Revised manuscript received March 12, 2008. Accepted March 17, 2008.
Potentially toxic substances in electrical and electronic equipment are increasingly regulated, yet there is little guidance regarding appropriate sample preparation and analysis for compliance testing. Printed circuit boards are likely to contain regulated substances. Unfortunately, they are very difficult to break into homogeneous components of a mass allowing ready analysis by X-ray fluorescence (XRF) spectrometry, instrumental neutron activation analysis or dissolution-based techniques such as inductively coupled plasma, yet they must be analyzed despite this heterogeneity. Analysis of unprepared samples of circuit board using microspot, hand-held, benchtop and polarizing XRF spectrometers results in unacceptable analytical precision. Shredding samples to 350 mg/kg; Cd >75 mg/kg) (8), samples should be analyzed further. “Noncompliance” with the RoHS directive is defined as analytes being present at greater than the prescribed critical concentration plus a margin to allow for analytical uncertainty (Hg, Pb >1200 mg/kg; Cd >125 mg/kg). For quantitative reporting of Pb, Hg, Cd, and Cr the sample can be analyzed using XRF or inductively coupled plasma (ICP), with determination of Cr6+ using spectrophotometry, and PBB and PBDE using gas chromatography-mass spectrometry (GCMS) (4). There is uncertainty as to the best preparation and analytical methods for the determination of the RoHS substances (4, 9, 10). The aim of our research was to critically assess the suitability of screening complex electrotechnical products with current methods and how they can be improved. This study will focus on PCBs as exemplar of complex, inhomogeneous electrotechnical materials. XRF screening methods currently used for RoHS compliance testing are reported to have ∼30% relative uncertainty (depending on the instrument used), and they only measure the surface of the product which potentially results in false negative results passing otherwise noncompliant products or parts. For screening quality data, the relative uncertainty of the data can be g20% and the coefficient of determination (r2) >0.7, or >0.85, for potentially definitive level quality data (11–13). We suggest sample preparation commensurate with the quality of data required (screening or quantitative). Sample preparation should be as easy as possible and suited to the accuracy required for compliance testing. For instance, for screening analyses sample preparation should be kept to a minimum; however, more preparation may be required for quantitative analyses. We conducted a comparative study using four energydispersive XRF instruments (microspot, hand-held, benchtop, and laboratory polarizing XRF) on depopulated PCB pieces, in order to determine whether or not adequate quality data could be produced using these methods. Samples were then prepared by various forms of shredding and milling, and the results were compared with samples analyzed using ICP and instrumental neutron activation analysis (INAA) in order to assess data quality resulting from the different preparation methods.
Experimental Section Comparative XRF Analyses of Unprepared Samples. A RoHS-compliant mobile phone PCB (sample “C1”) and two non RoHS-compliant phone PCBs (“N1” and “N2”) were used for the experiments. The PCBs were obtained by dismantling the phone by hand to its constituent parts. Large (>2-3 mm) and, where relevant, RoHS-exempt components (14) were 10.1021/es071748l CCC: $40.75
2008 American Chemical Society
Published on Web 04/25/2008
removed, and the board was depopulated using screwdrivers, a hot air gun, and a soldering iron. The PCBs were gridded at 1 × 2 cm intervals, coded, and cut into 1 × 2 cm pieces using a sheet metal cutter. In order to determine elemental diversity at the subcentimeter scale, fifteen point analyses were made of a 1 cm2 part of RoHS-compliant PCB “C1”, using a microspot energydispersive XRF (Spectro Midex M micro with 0.2-2.0 mm beam). Then, to assess analytical reproducibility, three analyses of the same sample were made using a polarizing XRF (PANalytical Epsilon 5 with 8 mm beam). This single piece of PCB was analyzed, removed from the machine, replaced, and reanalyzed. A comparative study of six adjacent parts of the RoHS-compliant PCB “C1” was then undertaken using three energy dispersive X-ray fluorescence instruments: a hand-held (Niton XLt797), benchtop XRF (PANalytical MiniPal4 RoHS/WEEE), and the polarizing Epsilon 5 XRF used previously. Further details of the spectrometers and analytical conditions are in Table S1 (Supporting Information). Sample Preparation by Shredding and Milling. Non RoHS-compliant and RoHS-compliant PCB samples were used for the shredding and milling experiments. PCBs “C1” and “N1” were used to create a suite of smaller samples, whereas PCB “N2” was used to assess the preparation of a single large sample for the milling experiments. The non RoHS-compliant PCBs weighed 0.39 g per 1 cm2 of depopulated board ((18% relative standard deviation; n ) 50 measurements), and it is this mass per unit area which controls the size of the sample to be prepared for analysis. For the larger samples (∼>2 g), a Fritsch Pulverisette 19 fitted with a 1 mm aperture sieve was operated at 2800 rpm. The rotor, anvil, and sieve are either made from or are coated with tungsten carbide. Initial testing showed that the best results were obtained when the sample was cut into ∼2-5 cm2 pieces prior to shredding. The feeder, chute, shredder, and sieve were removed, blown clean using compressed air, and wiped clean with ethanol between samples. For the comminution of smaller samples (∼
