This article was downloaded by: [University of Connecticut] On: 07 October 2014, At: 07:35 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Occupational and Environmental Hygiene Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uoeh20
Performance Evaluation of Currently Used Portable X ray Fluorescence Instruments for Measuring the Lead Content of Paint in Field Samples a
a
a
Yan Muller , Philippe Favreau & Marcel Kohler a
Départment de l'environnement, des transports et de l'agriculture, Direction générale de l'environnement, Service de Toxicologie de l'Environnement Bâti, Geneva, Switzerland Published online: 09 Jun 2014.
To cite this article: Yan Muller, Philippe Favreau & Marcel Kohler (2014) Performance Evaluation of Currently Used Portable X ray Fluorescence Instruments for Measuring the Lead Content of Paint in Field Samples, Journal of Occupational and Environmental Hygiene, 11:8, 528-537, DOI: 10.1080/15459624.2014.880788 To link to this article: http://dx.doi.org/10.1080/15459624.2014.880788
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Occupational and Environmental Hygiene, 11: 528–537 ISSN: 1545-9624 print / 1545-9632 online c 2014 JOEH, LLC Copyright DOI: 10.1080/15459624.2014.880788
Performance Evaluation of Currently Used Portable X ray Fluorescence Instruments for Measuring the Lead Content of Paint in Field Samples Yan Muller, Philippe Favreau, and Marcel Kohler
Downloaded by [University of Connecticut] at 07:35 07 October 2014
Department de l’environnement, des transports et de l’agriculture, Direction gen de l’environnement, ´ ´ erale ´ Service de Toxicologie de l’Environnement Bati, ˆ Geneva, Switzerland
Field-portable X-ray fluorescence (FP-XRF) instruments are important for non-destructive, rapid and convenient measurements of lead in paint, in view of potential remediation. Using real-life paint samples, we compared measurements from three FP-XRF instruments currently used in Switzerland with laboratory measurements using inductively coupled plasma mass spectrometry after complete sample dissolution. Two FP-XRF devices that functioned by lead L shell excitation frequently underestimated the lead concentration of samples. Lack of accuracy correlated with lead depth and/or the presence of additional metal elements (Zn, Ba or Ti). A radioactive source emitter XRF that enabled the additional K shell excitation showed higher accuracy and precision, regardless of the depth of the lead layer in the sample or the presence of other elements. Inspection of samples by light and electron microscopy revealed the diversity of real-life samples, with multi-layered paints showing various depths of lead and other metals. We conclude that the most accurate measurements of lead in paint are currently obtained with instruments that provide at least sufficient energy for lead K shell excitation. Keywords field samples, hand-held X-ray fluorescence, inductively coupled plasma mass spectrometry, lead, lead-based paint XRF, lead diagnostic, scanning electron microscopy
Address correspondence to: Yan Muller, Service de Toxicologie de l’Environnement Bˆati, 23, avenue de Sainte-Clotilde, Case postale 78, CH-1211 Gen`eve 8, Switzerland; e-mail: [email protected]
INTRODUCTION
O
ver the last 20 years, development of field-portable X-ray fluorescence (FP-XRF) instruments has meant that these have now become standard analytical tools for on-site investigations of environmental samples.(1,2) Their wide acceptance mainly stems from their reasonable sensitivity, convenience of use, reliability, and cost-effectiveness for detection and measurement of cations and metal elements. Applications for FP-XRF have been found in many areas, including analysis 528
of toxic elements in food and consumer products,(3) metals in soil,(4) air or dust filter samples,(5) and lead in paint.(6) The detection of lead in paint mainly relies on the use of FP-XRF instruments, which are based on X-ray emitting tubes or radioactive sources. Both types of instruments provide sufficient energy to the substrate to generate X-ray fluorescence; the characteristic fluorescence spectra enables identification and quantification of the elements present in the paint. Instruments containing X-ray emitting tubes are able to excite the L shell of lead at tube energies ranging from 35 to 50 keV. These instruments also permit multi-elemental analysis, which makes them very popular in environmental engineering consulting firms. Alternatively, radioactive sources such as 109Cd or 57Co emit gamma rays that efficiently excite the innermost lead K shell (88.04 keV). Fluorescence spectra derived from L or K shell excitation generate distinct energy patterns and intensities that require different detectors and signal treatment algorithms.(7–11) Due to the low fluorescent energy recovered from L shell excitation, X-ray emitting tubes are limited in their ability to detect heavy metals in complex matrices. In paints, where numerous layers may have been accumulated over time, lead L shell-derived fluorescence is prone to absorption by other elements, especially if the lead is located under multiple layers of paint.(12) In contrast, FP-XRF instruments with a radioactive source permit K shell excitation that is less prone to interference and absorption effects.(13) As a result, there may be discrepancies in paint lead measurements from different FPXRF instruments, especially in real-life samples where paint layers and matrices are extremely diversified. This situation has led to several government guidelines on the use of FP-XRF for lead diagnosis and concentration measurement in paint. For example, the United States Environmental Protection Agency (EPA) requires documented methodology and adequate quality controls, but has no detailed regulatory requirements.(14) It refers to the guidelines of the U.S. Department of Housing and Urban Development (HUD) for the evaluation and control of lead-based paint hazards
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
in housing,(15) which state that only XRF instruments with a performance characteristic sheet (PCS) should be used for detection and measurement of lead in paint.(16) Recently, the French government stipulated that an instrument capable of at least K shell excitation be used for detection of lead in paint.(17) However, in many other countries, including Switzerland, there is currently no legislation for evaluation of lead in paint. The vast majority of Swiss engineering and expert offices currently use L shell excitable XRF instruments, which offer multi-elemental analyses for a large panel of applications without using a radioactive source. The lack of clear and harmonized international regulations or recommendations for lead-based paint evaluation may lead to disparate and inappropriate use of these instruments. Despite the widespread use of FP-XRF instruments for lead measurements, we identified few reports on the capabilities and precision of instruments on test or real-life samples (e.g., lead-based paints) in the scientific literature.(6,18,19) Progress in XRF technology has led to the development of a number of new instruments that offer a large range of analytical environmental applications, including lead-based paint analysis. While these instruments have no major application problems for metal, alloy, or soil testing, some instruments may be inadequate for assessing lead in paint, especially in field samples where a number of parameters are difficult to replicate in standardized laboratory samples. With a view to better understanding of the range of applications and the limitations of some FP-XRF instruments currently used in Switzerland, we measured lead content in real-life paint samples with three different instruments currently used by environmental office experts or laboratory analysts. By comparing the results with inductively coupled plasma mass spectrometry (ICP-MS) data and by visualizing the microstructures of paint layers in all samples we report several key differences in the instruments and give important recommendations for better evaluation of lead in building materials. MATERIALS AND METHODS Sampling Samples consisted of painted building blocks (n = 23) made of concrete, plaster, or wood, with a mean length of 10 cm, width of 10 cm, and depth of 5 cm. These painted blocks were recovered during demolition work in Geneva (Switzerland), before complete destruction. The houses were randomly selected and specifically searched for lead-based-paint in kitchen and bathrooms. These were selected based on a first assessment of minimum lead surface concentration of 0.1 mg/cm2 using the Niton XLp instrument (Thermo Fisher Scientific Inc., Waltham, Mass.). Building blocks were then brought to the laboratory for extensive analytical work with FP-XRF instruments and sample preparation for ICP-MS analysis. For each sample, we chose three circular areas of 1.13 cm2 (spots A, B, and C) approximately 5 cm apart. The center of each spot was
used for hand-held XRF positioning and the complete circular surface was collected for ICP-MS measurements. For each sample, an additional spot was also collected where complete paint was removed to obtain a substrate reference. XRF Analysis Based on the instruments currently used by most expert offices in Switzerland, we selected three FP-XRF instruments to detect and measure lead in paint. Many of these instruments exist on the market, using either X-ray tube or radioactive emitting sources. Two instruments, a Niton Xlt 700 Series, model Xlt792W (Thermo Fisher Scientific Inc., Waltham, Mass.) and Innov-X, model α-4000 (Olympus Corp., Tokyo, Japan) used X-ray emitting-tubes. One device, a Niton Xlp 300 Series (Thermo Fisher Scientific Inc., Waltham, Mass.), used a radioactive emitting source. According to the manufacturer, the Innov-X hand-held XRF spectrometer (Olympus Corp., Tokyo, Japan) is suited for field analysis of alloys, lead-based paint, and environmental soils (Innov-X User Guide, version 2.1). Analyses were performed using Industrial Paint mode with a reading time of 20 sec. Maximum lead concentration that can be measured is 5’000 μg/cm2 and the action level was set to 1 mg/cm2. The manufacturer of the standard Niton XLt 700 Series states that it is not suited for evaluation of lead in paint (Niton XLt 700 User Guide, version 3.7); nevertheless it is used in Switzerland for this purpose by numerous expert offices. The main reason for this stems from the instrument’s extensive and versatile capability for environmental analysis, which enables experts to perform all their field investigations using a single instrument. Analyses were conducted using the Thin Sample Menu, Filter Testing Standard modes, with a reading time of 20 sec. Measurements with the Niton Xlp 300 Series were performed using the Pb Paint, and K and L modes, with a nominal 60-sec reading time for a full-strength radioactive source; the actual analysis time was automatically adjusted by the instrument based on the decay of the radiation source. The action level was set at 1 mg/cm2. At the time of measurements, the 109Cd source was 24 months old and the residual radioactivity was sufficient to perform analyses (minimum value at 75 MBq). The Niton XLp analysis also provided a depth index (DI) value that indicated the depth of the lead in the paint layers. DI values of 1–1.5, 1.5–4, and >4 indicated if the lead was present in the surface, intermediate, or deep layers, respectively. XRF analyses were performed on the same spots on each building block. Six measurements were performed for each spot. Substrate correction was not needed for Innov’X and Niton XLp instruments as stated by their PCS.(16) For comparative purposes, no substrate correction was also applied with the XLt instrument. Before use, instruments were checked with quality controls (0 and 1 mg/cm2 with the Niton calibration set [Thermo Fisher Scientific Inc., Waltham, Mass.]).
Journal of Occupational and Environmental Hygiene
August 2014
529
5000
Innov'x 4500
Niton XLt 4000
Niton XLp
Pb (µg/cm2)
3500
ICP-MS
3000 2500 2000
Downloaded by [University of Connecticut] at 07:35 07 October 2014
1500 1000 500 0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Sample number FIGURE 1. Mean values and standard deviations of lead concentration of paint samples measured with three different XRF instruments and ICP-MS analysis from 23 field-collected building blocks. The standard deviation corresponds to the variation observed from the total 18 measurements performed for each building block (6 replicate measurements on 3 spots).
Sample Preparation and ICP-MS Analysis Destructive samples were prepared by collecting all the paint layers present on each building block by drilling the surface of the spots A, B, and C previously assessed by the XRF instruments. Paint removal was performed including approximately 1 mm of substrate, assuring complete paint collection. Sample collection was also performed after complete paint removal, on each building block spot, to measure any lead contamination from the substrate (substrate reference spot). Materials were then crushed in 50 mL of milliQ water and filtered on 0.8 μm filters (Millipore, Billerica, Mass.) under vacuum. The samples were then dissolved by adding 8 mL of nitric acid (65%) and 2 mL of hydrogen peroxide followed by heating to a maximum temperature of 190◦ C using a microwave (MLS 1200 mega, MLS GmBH, Leutkirch im Allg¨au, Germany) at 200–600 W for 30 min. The resulting supernatant was stored at room temperature until analysis. ICP-MS analysis was carried out using an X series II ICPMS instrument with XS configuration (Thermo Fisher Scientific Inc., Waltham, Mass.), previously calibrated for Pb (20–3500 ppb, ±10% error), Ti (10–300 ppb, ±20% error), Zn (100–5000 ppb, ±10% error), and Ba (20–3000 ppb, ±10% error). The surface content of each metal was calculated in the paint areas collected from spots A, B, and C, as well as for the substrate reference spot.
530
Optical and Scanning Electron Microscopy The building blocks were fractured and the raw side cut was observed using a SZX12 binocular microscope equipped with a colorview camera (Olympus Corp., Tokyo, Japan) to visualize the paint layers with natural colors and textures. Additionally, each sample was cut transversally with a diamond saw and the surface was polished. Samples were mounted on a conductive support (an aluminum stub). A coating of carbon (ca15 nm) was then deposited on the samples by carbon thread evaporation prior to imaging with a Jeol JSM 7001F Scanning Electron Microscope (Department of Geology and Paleontology, Section of Earth and Environmental Sciences, University of Geneva, Switzerland). Semi-quantitative analyses and mapping of various elements (Pb, Ba, Ti, Zn, Na, Si, and Ca) was performed for all samples with a JED2300 EDS detector (JEOL Ltd, Tokyo, Japan).
RESULTS AND DISCUSSION XRF Measurements on Field Samples The mean lead content of each sample obtained with the three FP-XRF instruments and ICP-MS is shown in Figure 1. Values of individual measurements are available in the supplementary information (S1). Measurements of the lead
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
concentration in the 23 samples ranged from approximately 0.1 to >5000 μg/cm2. Values in the sampling set were widely distributed, highlighting the heterogeneity of paint characteristics that were probably related with various construction and renovation events, which was confirmed by visual inspection. Instrumental Precision and Sampling Variation XRF measurements were assessed for instrumental precision by calculating the relative standard deviation of six replicate analyses on each building block spot. Most measurements indicated a mean variation of less than 30% (Table I). The variation of Innov’X instrument was generally in the range of 3–32%, while XLt and Xlp instruments often demonstrated slightly lower variation (0–15% range). These data agree with previous reports on the precision of hand-held XRF instruments.(18,20) Overall, the results did not support the need for multiple XRF readings on the same spot, which is in agreement with the suggestions by the lead-based paint inspection guidelines of HUD.(15) Unexpectedly, the XLp instrument gave six spot analyses (out of 69 total) that showed extreme variation (39–66%). For example, this was observed for sample spots 3C, 7B, 7C, and 9C (Table I). In these cases, it appeared that when acquiring the six replicates on a sample spot, either the K or the L shellderived measurements were selected for the result, leading to notable differences, as observed for sample 3, spot C (Table II). Replicate measurements were made consecutively on the same spot without any change of position or movement between analyses. As the final lead concentration is calculated by a proprietary algorithm that takes K and/or L shell-derived data into account, as well as unknown other parameters, it was difficult to pinpoint the reason for the inconsistency. Such cases should prompt caution, especially when L and K shell-derived values are below and above the decision level, respectively. In a second step, we assessed the inter-spot variation that reflected heterogeneity in lead content on the building block surface. The true lead content of paint layers can vary widely at the macroscopic level due to manual paint application on more or less rough surfaces ( Figure 2A, B). In our case, the inter-spot variation range was similar to the intra-spot variation and was usually in the range of 1–35%, indicating no major heterogeneity between the spots on a same building block. A notable exception was block 12 where a 91% variation was observed between the three spot analyses using the XLp instrument (Table I). Again, this reflected exclusive selection of L or K derived results for spots A/C and B respectively. It is possible that in this particular case, sampling spot heterogeneity might have been observed whatever the instrumentation used; however, we did not confirm this. The three spot measurements were carried out on a relatively small area (usually less than 100 cm2), with a spacing of approximately 5 cm between the spots. In contrast, a relatively important spatial variation was reported during an EPA/HUD field study,(18) with the ratio up to 3.7 between the lowest and highest values measured at a distance of 20 cm.
Comparison Between XRF and ICP-MS Measurements We compared mean XRF readings from the three instruments with measurements from ICP-MS for each building block (Figure 1). ICP-MS analysis was performed on the dissolved paint layers from the spots (A, B, and C) where XRF measurements were carried out. Additionally, ICP-MS data for lead content in the reference substrate were not significant in comparison to the lead paint content, with an average ratio of 0.6% (maximum 2.8%). The tube-emitting XRF devices displayed significant variations in readings from nearly all building blocks. Except for block 2, all data showed a clear trend of underestimation compared with the standard ICP-MS analysis. We assessed the proximity of XRF measurements to the absolute lead content by calculating the ratio between XRF readings and ICP-MS results (Figure 3). The accuracy of each instrument was evaluated by calculating the ratio of the mean reading from a series of six replicate measurements on three spots (n = 18) to the mean of the ICP-MS measurements (3 spots, n = 3). Surprisingly, none of the mean XRF measurements closely agreed with the ICP-MS data; however, we could observe some general trends. For the X-ray tube-emitting instruments, the general trend was consistent underestimation of the lead content in the painted surfaces, irrespective of the true lead concentration in the sample. We observed this with both instruments and it could be directly linked to excitation of only the lead L shell. A trend toward lower values with X-ray tube-emitters, mainly owing to substrate effects, has been reported in other studies.(7,18) This effect was clearly apparent in samples that contained high concentrations of Ti, Ba, or Zn, as shown in Figure 3. Furthermore, the two X-ray tube-emitter instruments demonstrated significantly different ratios of 0.54 and 0.18 (Table III), which indicates that intrinsic capabilities of these instruments in lead measurement is not solely linked to the X-ray emittingdevice, but is also related to numerous other parameters, such as source geometry, signal detection, and data treatment. The radioactive source instrument that produced L and K shell activation provided a mean ratio of 1.13, which represents the highest accuracy observed in this study (Table III), but with the highest variability (standard deviation 0.36). Most data points were not distributed on the median line, showed no clear correlation with the Ti, Ba, or Zn content of the paint or with the lead depth in the paint layers, and showed a slight trend of overestimation. For deeply penetrating Xrays, such as those delivered by an instrument that excites lead K shell, an effect known as substrate effect or “backscatter” phenomenon could potentially affect the results due to the dense nature of the concrete samples under study.(21) This could result in overestimated readings from K shell data. This overestimation effect has been observed previously with other substrate types such as plaster, brick, metal, and wood using a 109Cd-source XRF instrument.(6) For example, the highest values from samples 14 and 15 that largely overestimated the lead paint content were derived from K shell readings whereas
Journal of Occupational and Environmental Hygiene
August 2014
531
TABLE I. Variation of Intra- and Inter-Spot Measurements from Ten Representative Samples Innov’X Sample 1
2
3
Downloaded by [University of Connecticut] at 07:35 07 October 2014
4
5
6
7
8
9
12
Niton XLt
Niton XLp
ICP-MS
Spot
Intra-spot variation
Inter-spot variation
Intra-spot variation
Inter-spot variation
Intra-spot variation
Inter-spot variation
Intra-spot variation
Inter-spot variation
A B C A B C A B C A B C A B C A B C A B C A B C A B C A B C
7.3 6.8 9.0 6.5 6.1 5.3 9.5 4.0 10.6 — — — 22.0 30.6 22.9 10.1 9.9 10.3 50.5 29.7 32.1 13.9 7.4 13.3 17.6 22.7 7.5 15.3 13.4 7.3
2.4
2.6 3.4 8.4 6.9 6.2 6.2 7.5 1.5 1.2 4.9 1.6 0.9 8.3 3.5 4.2 2.0 2.1 7.6 12.0 12.7 12.4 2.3 2.9 3.1 3.5 7.0 1.8 1.8 9.0 3.4
1.6
4.7 2.5 2.4 4.5 3.2 5.0 4.1 4.2 65.7 0.4 0.5 0.4 4.6 5.8 4.0 4.2 2.9 0.0 6.7 44.4 39.0 3.4 2.3 3.3 9.8 6.4 45.2 2.8 3.2 1.9
3.5
— — — — — — — — — — — — — — — — — — — — — — — — — —
3.6
1.4
23.9
—
14.1
12.9
9.2
17.8
2.3
9.8
4.1
8.3
12.4
24.0
15.6
59.6
10.3
6.7
9.4
1.9
28.6
0.3
7.0
5.4
14.1
2.9
32.2
91.4
— — —
13.7
30.6
5.7
5.8
10.4
4.9
11.2
9.4
23.7
Notes: For intra-spot variation, values are relative standard deviation (in %) from six measurements on each spot. For inter-spot variation, values are relative standard deviation (in %) from the 3 spot average readings. Note that for sample 4, no variation could be calculated for Innov’X due to out of range readings (>5000 mg/cm2). Complete details of XRF readings can be found in the supplementary material S1.
TABLE II. Replicate Measurements (n = 6) of Sample Spot 3C with the XLp Instrument Reading 1 2 3 4 5 6 Average Standard deviation
Pb C (mg/cm2)
Pb L (mg/cm2)
Pb K (mg/cm2)
Depth Index
0.3 ± 0.02 0.3 ± 0.02 1.2 ± 0.1 1.2 ± 0.1 0.3 ± 0.02 1.2 ± 0.1 0.75 0.49
0.3 ± 0.02 0.3 ± 0.02 0.4 ± 0.1 0.4 ± 0.1 0.3 ± 0.02 0.4 ± 0.1 0.35 0.05
1.1 ± 0.1 1.1 ± 0.1 1.2 ± 0.1 1.2 ± 0.1 1.1 ± 0.1 1.2 ± 0.1 1.15 0.05
1.81 1.82 1.89 1.85 1.79 1.92 1.84 0.05
Notes: The Pb C value shows the final value given by the instrument, whereas Pb L and Pb K values are readings from the lead L and K shells, respectively, with their instrumental errors.
532
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
FIGURE 2. Light microscopy (A, B) and SEM (C–F) images of a cross-section of sample 1. The region highlighted in A corresponds to image B. Visual inspection of the light microscope image (B) allowed identification of at least four major paint layers on the substrate (S). Panels C to F show elemental detection of Ti, Ba, Zn, and Pb, respectively, with white regions corresponding to the highest density in each image. Note that paint layers 1 and 2 are not discernible in the SEM images.
the lowest ratio (∼0.4) for sample 19 came from an L shell measurement (Figure 3C). Lead Depth The XLp instrument determined a lead DI for each sample, which, represented the quantity of non-leaded paint covering the lead layer (ThermoScientific, User Guide version 5.2).
A DI of 1 corresponded to lead at the surface, whereas an index of 10 indicated a deeply buried or shielded lead layer. This approach can be problematic, as a single value cannot represent the many configurations that occur in real-life samples, especially when lead is present in several layers. Samples 17 and 19 demonstrated two lead layers thus making any clear extrapolation difficult. In addition, we found no clear
Journal of Occupational and Environmental Hygiene
August 2014
533
Downloaded by [University of Connecticut] at 07:35 07 October 2014
FIGURE 3. Accuracies of the three hand-held XRF instruments Innov’X (A), Niton XLt (B) and Niton XLp (C) compared with ICP-MS data on the 23 field samples. Data points are annotated with the sample number and colored according to the Ti, Ba and Zn content of the paint. The green zone indicates the reference value ±20% deviation. In panel C, the use of L or K shell reading in the mean final value is given. “L/K” or “K/L” indicates the higher use of the L or K shell reading respectively in the final measurement. This occurred when the mean value from the 18 measurements per sample resulted from a combination of the L and K atomic layers (See Supplementary information S1 for more details).
534
Journal of Occupational and Environmental Hygiene
August 2014
and colors. The identification of paint layers by SEM relied essentially on the detection of selected cations (Ba, Ti, Zn, Ca, Pb). For this reason, the number of paint layers identified with this method was generally lower due to the presence of similar elements in consecutive layers. In the following sections, layer numbering refers to SEM-observed layers. SEM images and elemental information are reported in Table IV, together with measured DIs and Ba/Ti/Zn total surface concentration. In the 23 field samples, the total paint layer thickness averaged 500 μm and ranged from 250 to 1250 μm, with some layers composed of plaster, characterized by high Ca concentration, instead of paint (Table IV, i.e., samples 7 and 23). The minimum number of layers observed was three and the maximum was eight. Interestingly, we noted a complete lack of interdependence between layer number, thickness, and total Ba, Ti, or Zn content. Sample 1 provided a typical example of lead-containing paint buried below two layers containing Ba/Ti and Zn, respectively. We frequently encountered this configuration, with lead located in a second or third layer under Ba/Ti or Zn-containing layers. XRF measurements, especially from L shell-based instruments, were prone to matrix absorption effects when measuring this configuration (Figure 3A, B). In addition, as the lead-containing paint was often the first paint used, the lead layer was directly in contact with the raw substrate that exhibits a chaotic morphology at a sub-millimeter range (Figure 2). One could therefore expect XRF accuracy and precision readings to be directly modulated by the substrate heterogeneity, as was the case for the 23 field samples examined in this study.
TABLE III. Mean Ratios of XRF Measurements Versus ICP-MS Values
Downloaded by [University of Connecticut] at 07:35 07 October 2014
Mean Standard deviation
Innov’x ratio
Niton XLt ratio
Niton XLp ratio
0.54 (n = 21) 0.27 (n = 21)
0.18 (n = 23) 0.20 (n = 23)
1.13 (n = 23) 0.36 (n = 23)
correlation between the position of the lead layer or its depth in the paint and the DI. For example, sample 1 had a DI of 1.6 with a lead layer at a depth of 250 μm below a highly shielding Ba/Ti layer. In contrast, sample 11 had a DI of 7.5 with a lead layer at a depth of 200 μm moderately shielded by a Ba/Ti layer (Table IV). This illustrates how some parameters that are not accounted for in this study very probably influence DI measurements. Optical and Electron Microscopy of Samples To investigate properties of the samples that could influence the accuracy of XRF measurements, all building blocks were cut to provide a clear fracture allowing observation and analysis by binocular and scanning electron microscopy (SEM). Light microscopy not only permitted visualization of all paint layers on the sample surface, but also provided a useful indication of the morphology of the painted surfaces. A typical example is given by sample 1 (Figure 2), where at least four paint layers could be identified, based on different textures
Innov'x Inconclusive
Niton XLt Niton XLp
False negative
False positive
Correct
0
10
20
30
40
50
60
70
80
90
(%) FIGURE 4. Rates of correct, false positive, and false negative results obtained with the hand-held XRF instruments versus ICP-MS results, based on an action limit of 1 mg/cm2. Inconclusive results with the Innov’X instrument from the PCS statement (0.6–1.4 mg/cm2) are included for reference.
Journal of Occupational and Environmental Hygiene
August 2014
535
TABLE IV.
Main Characteristics of the Paint Layers in Samples
Downloaded by [University of Connecticut] at 07:35 07 October 2014
Sample Substrate
Paint width (μm)
Depth index
[Ba,Ti, Zn] (μg/cm2)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
concrete wood plaster concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete plaster
400 750 700 800 500 700 1250 300 600 600 300 250 350 350 300 250 460 700 750
1.6 1.2 2 4.9 10 7.5 9.2 3.9 8.8 2.8 7.5 2.9 1.6 1.5 3.3 3 9.7 8.4 4
20498 1863 1952 29407 5138 23664 3478 16598 6621 16955 5438 17512 12152 13999 15818 7506 10260 15616 12416
20 21 22 23
concrete concrete concrete concrete
370 370 320 480
4.2 8.6 4.2 2
7353 14772 14078 7205
Layer 1 (width, Layer 2 (width, Layer 3 μm) μm) (width, μm)
Layers 4/5/6/7 (width, μm)
Ba, Ti (150) Zn (100) Pb (150) Ba, Ti (250) Undefined (400) Pb (100) Ba, Ti (100) Zn (10) Pb, Ca (600) Ba, Ti (600) Pb (100) Zn (100) Pb, Ti, Ca (500) Zn, Pb, Ca (100) Ba, Ti (250) Zn, Na (150) Pb, Ca (300) Ca (1000) Ba, Ti (100) Zn, Pb (50) Ba, Ti (100) Ba, Ti (100) Zn (100) Pb (100) Ba, Ti, Ca (300) Zn, Pb (300) Ba, Ti (200) Zn (100) Pb (300) Ba, Ti (200) Zn, Pb (100) Ba, Ti (100) Zn, Pb (150) Ba, Ti (200) Zn, Na (50) Pb (100) Ba, Ti (150) Zn (100) Zn, Pb (100) Ba, Ti (100) Zn, Pb (200) Ba, Ti (50) Zn (50) Zn, Pb (100) Pb (50) Ba, Ti (250) Pb (10) Zn, Na (150) Zn, Na, Pb (50) Ba, Ti (300) Undefined (250) Zn, Pb (150) Ba, Ti (200) Pb, Ca (100) Zn (50) Ba, Ti (50)/Zn, Na (50)/Ca (50)/Zn, Ba, Ti, Pb, Na (100) Ca (20) Ba, Ti, Ca (150) Pb (100) Zn, Na (100) Ca (150) Ba, Ti, Pb (200) Zn, Na (20) Ba, Ti (70) Undefined (100) Ba, Ti, Pb (50) Pb (100) Ca (200) Ba, Ti (100) Undefined Pb, Zn, Cr (20) (80)/Pb, Zn (80)
Notes: Depth index is the value measured by the XLp instrument as detailed in the Materials and methods. Ba, Ti, and Zn concentrations are derived from ICP-MS data. Paint layers and composition were interpreted from SEM observations.
Rates of Correct Results Relative to the Threshold for Action In the field, the ease of use of XRF instruments allows rapid assessment of lead content in paint; however, threshold values that dictate further action must be defined. In 1992, the U.S. Congress defined lead-based paint as a dried paint film containing lead greater than or equal to 1000 μg/cm2 (Title X of the 1992 Housing and Community Development Act). This definition was mainly dictated by the technical limitations of field-portable XRF instrumentation available at the time.(7) Congress also defined lead paint as 0.5% by weight for those surfaces that could not be measured by XRF, but instead would be measured as paint chip samples analyzed in a laboratory. These levels were widely implemented in instrumentation software and thus became accepted by a large community of office experts internationally. In light of the recent decision by the U.S. Centers for Disease Control and Prevention (CDC) to establish a lead in blood reference value of 5 μg/dL, and in light of advances in XRF measurement technologies, exposure limits to lead-based paint hazards should be reconsidered. 536
We calculated the rates of correct, false positive, and false negative results for the three XRF instruments in the present study, based on the current action level of 1000 μg/cm2 (Figure 4) and using ICP-MS data as the reference. Inconclusive rates for the tube-emitting instrument Innov’X based on the PCS of the Innov’X LBP4000 in the range 0.6–1.1 mg/cm2 are also shown. The two tube-emitting devices, Innov’X and XLt, gave a low rate of correct results (57 and 52% respectively). These instruments were characterized by a relatively high rate of false negative results. In practice, such readings would create situations where no abatement would be decided, leading to risk exposure to children living housing and occupational exposure during remediation processes, especially in case of samples 8, 10, 18, 21, or 22 where the Innov’X instrument underestimated the lead content by a factor of 3 to 5. For example, sample 10 was estimated to have a lead content of less than 500 μg/cm2 instead of more than 2500 μg/cm2 in reality. This effect was even more pronounced with the XLt instrument, supporting the manufacturer’s assertion of its inadequacy for lead-based paint analysis.
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
A similarly rate of high false negative results was observed by Schmehl et al.(18) with instrumentation allowing L-X ray emission only. It can thus be concluded that, despite technical progress, currently used L shell-based XRF instruments appear to be poorly suited for assessment of lead in paint, at least on concrete building blocks that constitute the samples in this study. In contrast, the L and K based instrument (XLp) provided conservative measurements with a correct rate of 83% and no false negative measurements. Regarding the four false positive samples (17%), the lead content of these samples was between 500 and 1000 μg/cm2, corresponding to a level that could possibly be considered at risk for human exposure during renovation work. Indeed, preliminary investigation indicated that the use of a grinding disc on such a lead-based paint, in the worst case, can produce 1000–2000 μg/m3 of lead in inhalable dust, which is above the Swiss short-term exposure limit value (800 μg/m3). CONCLUSION
T
he main objective of this study was to evaluate the performance of some currently used FP-XRF instruments to measure the lead content of paint in real-life samples, mainly on concrete substrates. We found a clear difference between instruments exciting only the L shell or both the L and K shells of lead. Field-portable instruments emitting sufficient energy for lead L and K shell fluorescence appeared to be the most precise and accurate. Matrix energy absorption was still a major drawback in using the unique L shell emission energy, as demonstrated by our data and due to paint layer configuration, total Ba/Ti/Zn content, and the buried location of lead layers. Several points must be considered when choosing either a tube or a radioactive source X-ray emitting device for leadbased paint analysis. High acquisition cost, maintenance, source renewal or disposal, and radiation safety procedures are often cited as disadvantages of source-XRFs. However, accuracy should be the key criterion in selecting the appropriate analytical tool, implying the minimum need to excite and detect K-shell X-rays. Devices that excite K-shell only may also provide further improvements that would necessitate additional investigations. Technological developments in Xray fluorescence devices may also bring more reliable data in the future and the findings in this study can serve as a reference for evaluating the performance of the next-generation of field portable XRF instruments. ACKNOWLEDGMENTS
W
e would like to especially thank Jean-Luc Bailly, MarieJos´e Clerc, Nicole Perruchoud, Wanda Stryjenska, and Gilbert Pfister for their help in performing experimental work. We also wish to acknowledge Jasmin Meltretter for useful comments on the project and Agathe Martignier for electron microscopy data as well as Ron Hogg for careful and critical reading of the manuscript.
REFERENCES 1. Kalnicky, D.J., and R. Singhvi: Field portable XRF analysis of environmental samples. J. Hazard. Mater. 83:93–122 (2001). 2. Hou, X.D., Y.H. He, and B.T. Jones: Recent advances in portable X-ray fluorescence spectrometry. Appl. Spectros. Rev. 39(1):1–25 (2004). 3. Palmer, P.T., R. Jacobs, P.E. Baker, K. Ferguson, and S. Webber: Use of field-portable XRF analyzers for rapid screening of toxic elements in FDA-regulated products. J. Agric. Food Chem. 57:2605–2613 (2009). 4. Clark, S., W. Menrath, M. Chen, S. Roda, and P. Succop: Use of a field portable X-ray fluorescence analyzer to determine the concentration of lead and other metals in soil samples. Ann. Agric. Environ. Med. 6:27–32 (1999). 5. Morley, J.C., C.S. Clark, J.A. Deddens, K. Ashley, and S. Roda: Evaluation of a portable X-ray fluorescence instrument for the determination of lead in workplace air samples. Appl. Occup. Environ. Hyg. 14(5):306–316 (1999). 6. Ashley, K., M. Hunter, L.H. Tait, J. Dozier, J.L. Seaman, and P.F. Berry: Field investigation of on-site techniques for the measurement of lead in paint films. Field Anal. Chem. Technol. 2(1):39–50 (1998). 7. DeKosky, R.: Developing chemical instrumentation for environmental use in the late twentieth century: Detecting lead in paint using portable X-ray fluorescence spectrometry. Ambix 56(2):138–162 (2009). 8. Pantazis, T., J. Pantazis, A. Huber, and R. Redus: The historical development of the thermoelectrically cooled X-ray detector and its impact on the portable and hand-held XRF industries (February 2009). X-Ray Spectrom. 39:90–97 (2010). 9. Campbell, J.L., and L.A. Crosse: Portable instrumentation for the determination of lead in painted surfaces. Bull. Environ. Contam. Toxicol. 16:469–473 (1976). 10. Kaplan, E.H., M.D. Lilley, R.F. Schaefer, et al.: Portable X-rayfluorescence instruments for analysis of lead in paints. Pub. Health Rep. 90(3):223–230 (1975). 11. Spurgeon, J.C.: “Response Characteristics of a Portable X-ray Fluorescence Lead Detector. Detection of Lead in Paint.” Springfield, VA: National Bureau of Standards, 1973. 12. Rasberry, S.D.: Investigation of portable X-ray fluorescence analyzers for determining lead on painted surfaces. Appl. Spectrosc. 27:102–108 (1973). 13. Laurer, G.R., T.J. Kneip, R.E. Albert, and F.S. Kent: X-ray fluorescence: Detection of lead in wall paint. Science 172:466–468 (1971). 14. U.S. Environmental Protection Agency: “Lead; Requirements for LeadBased Paint Activities in Target Housing and Child-Occupied Facilities; Final Rule.” EPA Toxic Substances Control Act, Title IV, Sections 402 and 404. Washington, D.C., 1996. 15. U.S. Department of Housing and Urban Development (HUD): “Guidelines for the Evaluation and Control of Lead-based Paint Hazards in Housing.” Office of Healthy Homes and Lead Hazard Control Second Edition. Washington, D.C.: HUD, 2012. 16. “Performance Characteristic Sheet.” Available at http://www.hud.gov/ offices/lead/lbp/hudguidelines/allpcs.pdf (accessed August 8, 2013). 17. sant´e, Minist`ere franc¸ais du travail de l’emploi et de la sant´e: “Arrˆet´e du 19 aoˆut 2011 relatif au constat de risque d’exposition au plomb.” 2011. 18. Schmehl, R.L., D.C. Cox, F.G. Dewalt, et al.: Lead-based paint testing technologies: Summary of an EPA/HUD field study. Am. Ind. Hyg. Assoc. J. 60:444–451 (1999). 19. White, K.T., F.G. Dewalt, D.C. Cox, R.L. Schmehl, W. Friedman, and E.A. Pinzer: Pilot evaluation for lead-based paint proficiency testing of field portable XRF instruments. J. ASTM Inter. 8(3):1–5 (2011). 20. Driscoll, J.N., C. Wood, T. Powell, and J.S. Askew: An extractive x-ray fluorescence method for field screening lead paint sites. Appl. Occup. Environ. Hyg. 9:206–211 (1994). 21. Afshari, S., V. Nagarkar, and M.R. Squillante: Quantitative measurement of lead in paint by XRF analysis without manual substrate correction. Appl. Radiat. Isot. 48:1425–1431 (1997).
Journal of Occupational and Environmental Hygiene
August 2014
537
Journal of Occupational and Environmental Hygiene Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uoeh20
Performance Evaluation of Currently Used Portable X ray Fluorescence Instruments for Measuring the Lead Content of Paint in Field Samples a
a
a
Yan Muller , Philippe Favreau & Marcel Kohler a
Départment de l'environnement, des transports et de l'agriculture, Direction générale de l'environnement, Service de Toxicologie de l'Environnement Bâti, Geneva, Switzerland Published online: 09 Jun 2014.
To cite this article: Yan Muller, Philippe Favreau & Marcel Kohler (2014) Performance Evaluation of Currently Used Portable X ray Fluorescence Instruments for Measuring the Lead Content of Paint in Field Samples, Journal of Occupational and Environmental Hygiene, 11:8, 528-537, DOI: 10.1080/15459624.2014.880788 To link to this article: http://dx.doi.org/10.1080/15459624.2014.880788
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Occupational and Environmental Hygiene, 11: 528–537 ISSN: 1545-9624 print / 1545-9632 online c 2014 JOEH, LLC Copyright DOI: 10.1080/15459624.2014.880788
Performance Evaluation of Currently Used Portable X ray Fluorescence Instruments for Measuring the Lead Content of Paint in Field Samples Yan Muller, Philippe Favreau, and Marcel Kohler
Downloaded by [University of Connecticut] at 07:35 07 October 2014
Department de l’environnement, des transports et de l’agriculture, Direction gen de l’environnement, ´ ´ erale ´ Service de Toxicologie de l’Environnement Bati, ˆ Geneva, Switzerland
Field-portable X-ray fluorescence (FP-XRF) instruments are important for non-destructive, rapid and convenient measurements of lead in paint, in view of potential remediation. Using real-life paint samples, we compared measurements from three FP-XRF instruments currently used in Switzerland with laboratory measurements using inductively coupled plasma mass spectrometry after complete sample dissolution. Two FP-XRF devices that functioned by lead L shell excitation frequently underestimated the lead concentration of samples. Lack of accuracy correlated with lead depth and/or the presence of additional metal elements (Zn, Ba or Ti). A radioactive source emitter XRF that enabled the additional K shell excitation showed higher accuracy and precision, regardless of the depth of the lead layer in the sample or the presence of other elements. Inspection of samples by light and electron microscopy revealed the diversity of real-life samples, with multi-layered paints showing various depths of lead and other metals. We conclude that the most accurate measurements of lead in paint are currently obtained with instruments that provide at least sufficient energy for lead K shell excitation. Keywords field samples, hand-held X-ray fluorescence, inductively coupled plasma mass spectrometry, lead, lead-based paint XRF, lead diagnostic, scanning electron microscopy
Address correspondence to: Yan Muller, Service de Toxicologie de l’Environnement Bˆati, 23, avenue de Sainte-Clotilde, Case postale 78, CH-1211 Gen`eve 8, Switzerland; e-mail: [email protected]
INTRODUCTION
O
ver the last 20 years, development of field-portable X-ray fluorescence (FP-XRF) instruments has meant that these have now become standard analytical tools for on-site investigations of environmental samples.(1,2) Their wide acceptance mainly stems from their reasonable sensitivity, convenience of use, reliability, and cost-effectiveness for detection and measurement of cations and metal elements. Applications for FP-XRF have been found in many areas, including analysis 528
of toxic elements in food and consumer products,(3) metals in soil,(4) air or dust filter samples,(5) and lead in paint.(6) The detection of lead in paint mainly relies on the use of FP-XRF instruments, which are based on X-ray emitting tubes or radioactive sources. Both types of instruments provide sufficient energy to the substrate to generate X-ray fluorescence; the characteristic fluorescence spectra enables identification and quantification of the elements present in the paint. Instruments containing X-ray emitting tubes are able to excite the L shell of lead at tube energies ranging from 35 to 50 keV. These instruments also permit multi-elemental analysis, which makes them very popular in environmental engineering consulting firms. Alternatively, radioactive sources such as 109Cd or 57Co emit gamma rays that efficiently excite the innermost lead K shell (88.04 keV). Fluorescence spectra derived from L or K shell excitation generate distinct energy patterns and intensities that require different detectors and signal treatment algorithms.(7–11) Due to the low fluorescent energy recovered from L shell excitation, X-ray emitting tubes are limited in their ability to detect heavy metals in complex matrices. In paints, where numerous layers may have been accumulated over time, lead L shell-derived fluorescence is prone to absorption by other elements, especially if the lead is located under multiple layers of paint.(12) In contrast, FP-XRF instruments with a radioactive source permit K shell excitation that is less prone to interference and absorption effects.(13) As a result, there may be discrepancies in paint lead measurements from different FPXRF instruments, especially in real-life samples where paint layers and matrices are extremely diversified. This situation has led to several government guidelines on the use of FP-XRF for lead diagnosis and concentration measurement in paint. For example, the United States Environmental Protection Agency (EPA) requires documented methodology and adequate quality controls, but has no detailed regulatory requirements.(14) It refers to the guidelines of the U.S. Department of Housing and Urban Development (HUD) for the evaluation and control of lead-based paint hazards
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
in housing,(15) which state that only XRF instruments with a performance characteristic sheet (PCS) should be used for detection and measurement of lead in paint.(16) Recently, the French government stipulated that an instrument capable of at least K shell excitation be used for detection of lead in paint.(17) However, in many other countries, including Switzerland, there is currently no legislation for evaluation of lead in paint. The vast majority of Swiss engineering and expert offices currently use L shell excitable XRF instruments, which offer multi-elemental analyses for a large panel of applications without using a radioactive source. The lack of clear and harmonized international regulations or recommendations for lead-based paint evaluation may lead to disparate and inappropriate use of these instruments. Despite the widespread use of FP-XRF instruments for lead measurements, we identified few reports on the capabilities and precision of instruments on test or real-life samples (e.g., lead-based paints) in the scientific literature.(6,18,19) Progress in XRF technology has led to the development of a number of new instruments that offer a large range of analytical environmental applications, including lead-based paint analysis. While these instruments have no major application problems for metal, alloy, or soil testing, some instruments may be inadequate for assessing lead in paint, especially in field samples where a number of parameters are difficult to replicate in standardized laboratory samples. With a view to better understanding of the range of applications and the limitations of some FP-XRF instruments currently used in Switzerland, we measured lead content in real-life paint samples with three different instruments currently used by environmental office experts or laboratory analysts. By comparing the results with inductively coupled plasma mass spectrometry (ICP-MS) data and by visualizing the microstructures of paint layers in all samples we report several key differences in the instruments and give important recommendations for better evaluation of lead in building materials. MATERIALS AND METHODS Sampling Samples consisted of painted building blocks (n = 23) made of concrete, plaster, or wood, with a mean length of 10 cm, width of 10 cm, and depth of 5 cm. These painted blocks were recovered during demolition work in Geneva (Switzerland), before complete destruction. The houses were randomly selected and specifically searched for lead-based-paint in kitchen and bathrooms. These were selected based on a first assessment of minimum lead surface concentration of 0.1 mg/cm2 using the Niton XLp instrument (Thermo Fisher Scientific Inc., Waltham, Mass.). Building blocks were then brought to the laboratory for extensive analytical work with FP-XRF instruments and sample preparation for ICP-MS analysis. For each sample, we chose three circular areas of 1.13 cm2 (spots A, B, and C) approximately 5 cm apart. The center of each spot was
used for hand-held XRF positioning and the complete circular surface was collected for ICP-MS measurements. For each sample, an additional spot was also collected where complete paint was removed to obtain a substrate reference. XRF Analysis Based on the instruments currently used by most expert offices in Switzerland, we selected three FP-XRF instruments to detect and measure lead in paint. Many of these instruments exist on the market, using either X-ray tube or radioactive emitting sources. Two instruments, a Niton Xlt 700 Series, model Xlt792W (Thermo Fisher Scientific Inc., Waltham, Mass.) and Innov-X, model α-4000 (Olympus Corp., Tokyo, Japan) used X-ray emitting-tubes. One device, a Niton Xlp 300 Series (Thermo Fisher Scientific Inc., Waltham, Mass.), used a radioactive emitting source. According to the manufacturer, the Innov-X hand-held XRF spectrometer (Olympus Corp., Tokyo, Japan) is suited for field analysis of alloys, lead-based paint, and environmental soils (Innov-X User Guide, version 2.1). Analyses were performed using Industrial Paint mode with a reading time of 20 sec. Maximum lead concentration that can be measured is 5’000 μg/cm2 and the action level was set to 1 mg/cm2. The manufacturer of the standard Niton XLt 700 Series states that it is not suited for evaluation of lead in paint (Niton XLt 700 User Guide, version 3.7); nevertheless it is used in Switzerland for this purpose by numerous expert offices. The main reason for this stems from the instrument’s extensive and versatile capability for environmental analysis, which enables experts to perform all their field investigations using a single instrument. Analyses were conducted using the Thin Sample Menu, Filter Testing Standard modes, with a reading time of 20 sec. Measurements with the Niton Xlp 300 Series were performed using the Pb Paint, and K and L modes, with a nominal 60-sec reading time for a full-strength radioactive source; the actual analysis time was automatically adjusted by the instrument based on the decay of the radiation source. The action level was set at 1 mg/cm2. At the time of measurements, the 109Cd source was 24 months old and the residual radioactivity was sufficient to perform analyses (minimum value at 75 MBq). The Niton XLp analysis also provided a depth index (DI) value that indicated the depth of the lead in the paint layers. DI values of 1–1.5, 1.5–4, and >4 indicated if the lead was present in the surface, intermediate, or deep layers, respectively. XRF analyses were performed on the same spots on each building block. Six measurements were performed for each spot. Substrate correction was not needed for Innov’X and Niton XLp instruments as stated by their PCS.(16) For comparative purposes, no substrate correction was also applied with the XLt instrument. Before use, instruments were checked with quality controls (0 and 1 mg/cm2 with the Niton calibration set [Thermo Fisher Scientific Inc., Waltham, Mass.]).
Journal of Occupational and Environmental Hygiene
August 2014
529
5000
Innov'x 4500
Niton XLt 4000
Niton XLp
Pb (µg/cm2)
3500
ICP-MS
3000 2500 2000
Downloaded by [University of Connecticut] at 07:35 07 October 2014
1500 1000 500 0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Sample number FIGURE 1. Mean values and standard deviations of lead concentration of paint samples measured with three different XRF instruments and ICP-MS analysis from 23 field-collected building blocks. The standard deviation corresponds to the variation observed from the total 18 measurements performed for each building block (6 replicate measurements on 3 spots).
Sample Preparation and ICP-MS Analysis Destructive samples were prepared by collecting all the paint layers present on each building block by drilling the surface of the spots A, B, and C previously assessed by the XRF instruments. Paint removal was performed including approximately 1 mm of substrate, assuring complete paint collection. Sample collection was also performed after complete paint removal, on each building block spot, to measure any lead contamination from the substrate (substrate reference spot). Materials were then crushed in 50 mL of milliQ water and filtered on 0.8 μm filters (Millipore, Billerica, Mass.) under vacuum. The samples were then dissolved by adding 8 mL of nitric acid (65%) and 2 mL of hydrogen peroxide followed by heating to a maximum temperature of 190◦ C using a microwave (MLS 1200 mega, MLS GmBH, Leutkirch im Allg¨au, Germany) at 200–600 W for 30 min. The resulting supernatant was stored at room temperature until analysis. ICP-MS analysis was carried out using an X series II ICPMS instrument with XS configuration (Thermo Fisher Scientific Inc., Waltham, Mass.), previously calibrated for Pb (20–3500 ppb, ±10% error), Ti (10–300 ppb, ±20% error), Zn (100–5000 ppb, ±10% error), and Ba (20–3000 ppb, ±10% error). The surface content of each metal was calculated in the paint areas collected from spots A, B, and C, as well as for the substrate reference spot.
530
Optical and Scanning Electron Microscopy The building blocks were fractured and the raw side cut was observed using a SZX12 binocular microscope equipped with a colorview camera (Olympus Corp., Tokyo, Japan) to visualize the paint layers with natural colors and textures. Additionally, each sample was cut transversally with a diamond saw and the surface was polished. Samples were mounted on a conductive support (an aluminum stub). A coating of carbon (ca15 nm) was then deposited on the samples by carbon thread evaporation prior to imaging with a Jeol JSM 7001F Scanning Electron Microscope (Department of Geology and Paleontology, Section of Earth and Environmental Sciences, University of Geneva, Switzerland). Semi-quantitative analyses and mapping of various elements (Pb, Ba, Ti, Zn, Na, Si, and Ca) was performed for all samples with a JED2300 EDS detector (JEOL Ltd, Tokyo, Japan).
RESULTS AND DISCUSSION XRF Measurements on Field Samples The mean lead content of each sample obtained with the three FP-XRF instruments and ICP-MS is shown in Figure 1. Values of individual measurements are available in the supplementary information (S1). Measurements of the lead
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
concentration in the 23 samples ranged from approximately 0.1 to >5000 μg/cm2. Values in the sampling set were widely distributed, highlighting the heterogeneity of paint characteristics that were probably related with various construction and renovation events, which was confirmed by visual inspection. Instrumental Precision and Sampling Variation XRF measurements were assessed for instrumental precision by calculating the relative standard deviation of six replicate analyses on each building block spot. Most measurements indicated a mean variation of less than 30% (Table I). The variation of Innov’X instrument was generally in the range of 3–32%, while XLt and Xlp instruments often demonstrated slightly lower variation (0–15% range). These data agree with previous reports on the precision of hand-held XRF instruments.(18,20) Overall, the results did not support the need for multiple XRF readings on the same spot, which is in agreement with the suggestions by the lead-based paint inspection guidelines of HUD.(15) Unexpectedly, the XLp instrument gave six spot analyses (out of 69 total) that showed extreme variation (39–66%). For example, this was observed for sample spots 3C, 7B, 7C, and 9C (Table I). In these cases, it appeared that when acquiring the six replicates on a sample spot, either the K or the L shellderived measurements were selected for the result, leading to notable differences, as observed for sample 3, spot C (Table II). Replicate measurements were made consecutively on the same spot without any change of position or movement between analyses. As the final lead concentration is calculated by a proprietary algorithm that takes K and/or L shell-derived data into account, as well as unknown other parameters, it was difficult to pinpoint the reason for the inconsistency. Such cases should prompt caution, especially when L and K shell-derived values are below and above the decision level, respectively. In a second step, we assessed the inter-spot variation that reflected heterogeneity in lead content on the building block surface. The true lead content of paint layers can vary widely at the macroscopic level due to manual paint application on more or less rough surfaces ( Figure 2A, B). In our case, the inter-spot variation range was similar to the intra-spot variation and was usually in the range of 1–35%, indicating no major heterogeneity between the spots on a same building block. A notable exception was block 12 where a 91% variation was observed between the three spot analyses using the XLp instrument (Table I). Again, this reflected exclusive selection of L or K derived results for spots A/C and B respectively. It is possible that in this particular case, sampling spot heterogeneity might have been observed whatever the instrumentation used; however, we did not confirm this. The three spot measurements were carried out on a relatively small area (usually less than 100 cm2), with a spacing of approximately 5 cm between the spots. In contrast, a relatively important spatial variation was reported during an EPA/HUD field study,(18) with the ratio up to 3.7 between the lowest and highest values measured at a distance of 20 cm.
Comparison Between XRF and ICP-MS Measurements We compared mean XRF readings from the three instruments with measurements from ICP-MS for each building block (Figure 1). ICP-MS analysis was performed on the dissolved paint layers from the spots (A, B, and C) where XRF measurements were carried out. Additionally, ICP-MS data for lead content in the reference substrate were not significant in comparison to the lead paint content, with an average ratio of 0.6% (maximum 2.8%). The tube-emitting XRF devices displayed significant variations in readings from nearly all building blocks. Except for block 2, all data showed a clear trend of underestimation compared with the standard ICP-MS analysis. We assessed the proximity of XRF measurements to the absolute lead content by calculating the ratio between XRF readings and ICP-MS results (Figure 3). The accuracy of each instrument was evaluated by calculating the ratio of the mean reading from a series of six replicate measurements on three spots (n = 18) to the mean of the ICP-MS measurements (3 spots, n = 3). Surprisingly, none of the mean XRF measurements closely agreed with the ICP-MS data; however, we could observe some general trends. For the X-ray tube-emitting instruments, the general trend was consistent underestimation of the lead content in the painted surfaces, irrespective of the true lead concentration in the sample. We observed this with both instruments and it could be directly linked to excitation of only the lead L shell. A trend toward lower values with X-ray tube-emitters, mainly owing to substrate effects, has been reported in other studies.(7,18) This effect was clearly apparent in samples that contained high concentrations of Ti, Ba, or Zn, as shown in Figure 3. Furthermore, the two X-ray tube-emitter instruments demonstrated significantly different ratios of 0.54 and 0.18 (Table III), which indicates that intrinsic capabilities of these instruments in lead measurement is not solely linked to the X-ray emittingdevice, but is also related to numerous other parameters, such as source geometry, signal detection, and data treatment. The radioactive source instrument that produced L and K shell activation provided a mean ratio of 1.13, which represents the highest accuracy observed in this study (Table III), but with the highest variability (standard deviation 0.36). Most data points were not distributed on the median line, showed no clear correlation with the Ti, Ba, or Zn content of the paint or with the lead depth in the paint layers, and showed a slight trend of overestimation. For deeply penetrating Xrays, such as those delivered by an instrument that excites lead K shell, an effect known as substrate effect or “backscatter” phenomenon could potentially affect the results due to the dense nature of the concrete samples under study.(21) This could result in overestimated readings from K shell data. This overestimation effect has been observed previously with other substrate types such as plaster, brick, metal, and wood using a 109Cd-source XRF instrument.(6) For example, the highest values from samples 14 and 15 that largely overestimated the lead paint content were derived from K shell readings whereas
Journal of Occupational and Environmental Hygiene
August 2014
531
TABLE I. Variation of Intra- and Inter-Spot Measurements from Ten Representative Samples Innov’X Sample 1
2
3
Downloaded by [University of Connecticut] at 07:35 07 October 2014
4
5
6
7
8
9
12
Niton XLt
Niton XLp
ICP-MS
Spot
Intra-spot variation
Inter-spot variation
Intra-spot variation
Inter-spot variation
Intra-spot variation
Inter-spot variation
Intra-spot variation
Inter-spot variation
A B C A B C A B C A B C A B C A B C A B C A B C A B C A B C
7.3 6.8 9.0 6.5 6.1 5.3 9.5 4.0 10.6 — — — 22.0 30.6 22.9 10.1 9.9 10.3 50.5 29.7 32.1 13.9 7.4 13.3 17.6 22.7 7.5 15.3 13.4 7.3
2.4
2.6 3.4 8.4 6.9 6.2 6.2 7.5 1.5 1.2 4.9 1.6 0.9 8.3 3.5 4.2 2.0 2.1 7.6 12.0 12.7 12.4 2.3 2.9 3.1 3.5 7.0 1.8 1.8 9.0 3.4
1.6
4.7 2.5 2.4 4.5 3.2 5.0 4.1 4.2 65.7 0.4 0.5 0.4 4.6 5.8 4.0 4.2 2.9 0.0 6.7 44.4 39.0 3.4 2.3 3.3 9.8 6.4 45.2 2.8 3.2 1.9
3.5
— — — — — — — — — — — — — — — — — — — — — — — — — —
3.6
1.4
23.9
—
14.1
12.9
9.2
17.8
2.3
9.8
4.1
8.3
12.4
24.0
15.6
59.6
10.3
6.7
9.4
1.9
28.6
0.3
7.0
5.4
14.1
2.9
32.2
91.4
— — —
13.7
30.6
5.7
5.8
10.4
4.9
11.2
9.4
23.7
Notes: For intra-spot variation, values are relative standard deviation (in %) from six measurements on each spot. For inter-spot variation, values are relative standard deviation (in %) from the 3 spot average readings. Note that for sample 4, no variation could be calculated for Innov’X due to out of range readings (>5000 mg/cm2). Complete details of XRF readings can be found in the supplementary material S1.
TABLE II. Replicate Measurements (n = 6) of Sample Spot 3C with the XLp Instrument Reading 1 2 3 4 5 6 Average Standard deviation
Pb C (mg/cm2)
Pb L (mg/cm2)
Pb K (mg/cm2)
Depth Index
0.3 ± 0.02 0.3 ± 0.02 1.2 ± 0.1 1.2 ± 0.1 0.3 ± 0.02 1.2 ± 0.1 0.75 0.49
0.3 ± 0.02 0.3 ± 0.02 0.4 ± 0.1 0.4 ± 0.1 0.3 ± 0.02 0.4 ± 0.1 0.35 0.05
1.1 ± 0.1 1.1 ± 0.1 1.2 ± 0.1 1.2 ± 0.1 1.1 ± 0.1 1.2 ± 0.1 1.15 0.05
1.81 1.82 1.89 1.85 1.79 1.92 1.84 0.05
Notes: The Pb C value shows the final value given by the instrument, whereas Pb L and Pb K values are readings from the lead L and K shells, respectively, with their instrumental errors.
532
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
FIGURE 2. Light microscopy (A, B) and SEM (C–F) images of a cross-section of sample 1. The region highlighted in A corresponds to image B. Visual inspection of the light microscope image (B) allowed identification of at least four major paint layers on the substrate (S). Panels C to F show elemental detection of Ti, Ba, Zn, and Pb, respectively, with white regions corresponding to the highest density in each image. Note that paint layers 1 and 2 are not discernible in the SEM images.
the lowest ratio (∼0.4) for sample 19 came from an L shell measurement (Figure 3C). Lead Depth The XLp instrument determined a lead DI for each sample, which, represented the quantity of non-leaded paint covering the lead layer (ThermoScientific, User Guide version 5.2).
A DI of 1 corresponded to lead at the surface, whereas an index of 10 indicated a deeply buried or shielded lead layer. This approach can be problematic, as a single value cannot represent the many configurations that occur in real-life samples, especially when lead is present in several layers. Samples 17 and 19 demonstrated two lead layers thus making any clear extrapolation difficult. In addition, we found no clear
Journal of Occupational and Environmental Hygiene
August 2014
533
Downloaded by [University of Connecticut] at 07:35 07 October 2014
FIGURE 3. Accuracies of the three hand-held XRF instruments Innov’X (A), Niton XLt (B) and Niton XLp (C) compared with ICP-MS data on the 23 field samples. Data points are annotated with the sample number and colored according to the Ti, Ba and Zn content of the paint. The green zone indicates the reference value ±20% deviation. In panel C, the use of L or K shell reading in the mean final value is given. “L/K” or “K/L” indicates the higher use of the L or K shell reading respectively in the final measurement. This occurred when the mean value from the 18 measurements per sample resulted from a combination of the L and K atomic layers (See Supplementary information S1 for more details).
534
Journal of Occupational and Environmental Hygiene
August 2014
and colors. The identification of paint layers by SEM relied essentially on the detection of selected cations (Ba, Ti, Zn, Ca, Pb). For this reason, the number of paint layers identified with this method was generally lower due to the presence of similar elements in consecutive layers. In the following sections, layer numbering refers to SEM-observed layers. SEM images and elemental information are reported in Table IV, together with measured DIs and Ba/Ti/Zn total surface concentration. In the 23 field samples, the total paint layer thickness averaged 500 μm and ranged from 250 to 1250 μm, with some layers composed of plaster, characterized by high Ca concentration, instead of paint (Table IV, i.e., samples 7 and 23). The minimum number of layers observed was three and the maximum was eight. Interestingly, we noted a complete lack of interdependence between layer number, thickness, and total Ba, Ti, or Zn content. Sample 1 provided a typical example of lead-containing paint buried below two layers containing Ba/Ti and Zn, respectively. We frequently encountered this configuration, with lead located in a second or third layer under Ba/Ti or Zn-containing layers. XRF measurements, especially from L shell-based instruments, were prone to matrix absorption effects when measuring this configuration (Figure 3A, B). In addition, as the lead-containing paint was often the first paint used, the lead layer was directly in contact with the raw substrate that exhibits a chaotic morphology at a sub-millimeter range (Figure 2). One could therefore expect XRF accuracy and precision readings to be directly modulated by the substrate heterogeneity, as was the case for the 23 field samples examined in this study.
TABLE III. Mean Ratios of XRF Measurements Versus ICP-MS Values
Downloaded by [University of Connecticut] at 07:35 07 October 2014
Mean Standard deviation
Innov’x ratio
Niton XLt ratio
Niton XLp ratio
0.54 (n = 21) 0.27 (n = 21)
0.18 (n = 23) 0.20 (n = 23)
1.13 (n = 23) 0.36 (n = 23)
correlation between the position of the lead layer or its depth in the paint and the DI. For example, sample 1 had a DI of 1.6 with a lead layer at a depth of 250 μm below a highly shielding Ba/Ti layer. In contrast, sample 11 had a DI of 7.5 with a lead layer at a depth of 200 μm moderately shielded by a Ba/Ti layer (Table IV). This illustrates how some parameters that are not accounted for in this study very probably influence DI measurements. Optical and Electron Microscopy of Samples To investigate properties of the samples that could influence the accuracy of XRF measurements, all building blocks were cut to provide a clear fracture allowing observation and analysis by binocular and scanning electron microscopy (SEM). Light microscopy not only permitted visualization of all paint layers on the sample surface, but also provided a useful indication of the morphology of the painted surfaces. A typical example is given by sample 1 (Figure 2), where at least four paint layers could be identified, based on different textures
Innov'x Inconclusive
Niton XLt Niton XLp
False negative
False positive
Correct
0
10
20
30
40
50
60
70
80
90
(%) FIGURE 4. Rates of correct, false positive, and false negative results obtained with the hand-held XRF instruments versus ICP-MS results, based on an action limit of 1 mg/cm2. Inconclusive results with the Innov’X instrument from the PCS statement (0.6–1.4 mg/cm2) are included for reference.
Journal of Occupational and Environmental Hygiene
August 2014
535
TABLE IV.
Main Characteristics of the Paint Layers in Samples
Downloaded by [University of Connecticut] at 07:35 07 October 2014
Sample Substrate
Paint width (μm)
Depth index
[Ba,Ti, Zn] (μg/cm2)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
concrete wood plaster concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete plaster
400 750 700 800 500 700 1250 300 600 600 300 250 350 350 300 250 460 700 750
1.6 1.2 2 4.9 10 7.5 9.2 3.9 8.8 2.8 7.5 2.9 1.6 1.5 3.3 3 9.7 8.4 4
20498 1863 1952 29407 5138 23664 3478 16598 6621 16955 5438 17512 12152 13999 15818 7506 10260 15616 12416
20 21 22 23
concrete concrete concrete concrete
370 370 320 480
4.2 8.6 4.2 2
7353 14772 14078 7205
Layer 1 (width, Layer 2 (width, Layer 3 μm) μm) (width, μm)
Layers 4/5/6/7 (width, μm)
Ba, Ti (150) Zn (100) Pb (150) Ba, Ti (250) Undefined (400) Pb (100) Ba, Ti (100) Zn (10) Pb, Ca (600) Ba, Ti (600) Pb (100) Zn (100) Pb, Ti, Ca (500) Zn, Pb, Ca (100) Ba, Ti (250) Zn, Na (150) Pb, Ca (300) Ca (1000) Ba, Ti (100) Zn, Pb (50) Ba, Ti (100) Ba, Ti (100) Zn (100) Pb (100) Ba, Ti, Ca (300) Zn, Pb (300) Ba, Ti (200) Zn (100) Pb (300) Ba, Ti (200) Zn, Pb (100) Ba, Ti (100) Zn, Pb (150) Ba, Ti (200) Zn, Na (50) Pb (100) Ba, Ti (150) Zn (100) Zn, Pb (100) Ba, Ti (100) Zn, Pb (200) Ba, Ti (50) Zn (50) Zn, Pb (100) Pb (50) Ba, Ti (250) Pb (10) Zn, Na (150) Zn, Na, Pb (50) Ba, Ti (300) Undefined (250) Zn, Pb (150) Ba, Ti (200) Pb, Ca (100) Zn (50) Ba, Ti (50)/Zn, Na (50)/Ca (50)/Zn, Ba, Ti, Pb, Na (100) Ca (20) Ba, Ti, Ca (150) Pb (100) Zn, Na (100) Ca (150) Ba, Ti, Pb (200) Zn, Na (20) Ba, Ti (70) Undefined (100) Ba, Ti, Pb (50) Pb (100) Ca (200) Ba, Ti (100) Undefined Pb, Zn, Cr (20) (80)/Pb, Zn (80)
Notes: Depth index is the value measured by the XLp instrument as detailed in the Materials and methods. Ba, Ti, and Zn concentrations are derived from ICP-MS data. Paint layers and composition were interpreted from SEM observations.
Rates of Correct Results Relative to the Threshold for Action In the field, the ease of use of XRF instruments allows rapid assessment of lead content in paint; however, threshold values that dictate further action must be defined. In 1992, the U.S. Congress defined lead-based paint as a dried paint film containing lead greater than or equal to 1000 μg/cm2 (Title X of the 1992 Housing and Community Development Act). This definition was mainly dictated by the technical limitations of field-portable XRF instrumentation available at the time.(7) Congress also defined lead paint as 0.5% by weight for those surfaces that could not be measured by XRF, but instead would be measured as paint chip samples analyzed in a laboratory. These levels were widely implemented in instrumentation software and thus became accepted by a large community of office experts internationally. In light of the recent decision by the U.S. Centers for Disease Control and Prevention (CDC) to establish a lead in blood reference value of 5 μg/dL, and in light of advances in XRF measurement technologies, exposure limits to lead-based paint hazards should be reconsidered. 536
We calculated the rates of correct, false positive, and false negative results for the three XRF instruments in the present study, based on the current action level of 1000 μg/cm2 (Figure 4) and using ICP-MS data as the reference. Inconclusive rates for the tube-emitting instrument Innov’X based on the PCS of the Innov’X LBP4000 in the range 0.6–1.1 mg/cm2 are also shown. The two tube-emitting devices, Innov’X and XLt, gave a low rate of correct results (57 and 52% respectively). These instruments were characterized by a relatively high rate of false negative results. In practice, such readings would create situations where no abatement would be decided, leading to risk exposure to children living housing and occupational exposure during remediation processes, especially in case of samples 8, 10, 18, 21, or 22 where the Innov’X instrument underestimated the lead content by a factor of 3 to 5. For example, sample 10 was estimated to have a lead content of less than 500 μg/cm2 instead of more than 2500 μg/cm2 in reality. This effect was even more pronounced with the XLt instrument, supporting the manufacturer’s assertion of its inadequacy for lead-based paint analysis.
Journal of Occupational and Environmental Hygiene
August 2014
Downloaded by [University of Connecticut] at 07:35 07 October 2014
A similarly rate of high false negative results was observed by Schmehl et al.(18) with instrumentation allowing L-X ray emission only. It can thus be concluded that, despite technical progress, currently used L shell-based XRF instruments appear to be poorly suited for assessment of lead in paint, at least on concrete building blocks that constitute the samples in this study. In contrast, the L and K based instrument (XLp) provided conservative measurements with a correct rate of 83% and no false negative measurements. Regarding the four false positive samples (17%), the lead content of these samples was between 500 and 1000 μg/cm2, corresponding to a level that could possibly be considered at risk for human exposure during renovation work. Indeed, preliminary investigation indicated that the use of a grinding disc on such a lead-based paint, in the worst case, can produce 1000–2000 μg/m3 of lead in inhalable dust, which is above the Swiss short-term exposure limit value (800 μg/m3). CONCLUSION
T
he main objective of this study was to evaluate the performance of some currently used FP-XRF instruments to measure the lead content of paint in real-life samples, mainly on concrete substrates. We found a clear difference between instruments exciting only the L shell or both the L and K shells of lead. Field-portable instruments emitting sufficient energy for lead L and K shell fluorescence appeared to be the most precise and accurate. Matrix energy absorption was still a major drawback in using the unique L shell emission energy, as demonstrated by our data and due to paint layer configuration, total Ba/Ti/Zn content, and the buried location of lead layers. Several points must be considered when choosing either a tube or a radioactive source X-ray emitting device for leadbased paint analysis. High acquisition cost, maintenance, source renewal or disposal, and radiation safety procedures are often cited as disadvantages of source-XRFs. However, accuracy should be the key criterion in selecting the appropriate analytical tool, implying the minimum need to excite and detect K-shell X-rays. Devices that excite K-shell only may also provide further improvements that would necessitate additional investigations. Technological developments in Xray fluorescence devices may also bring more reliable data in the future and the findings in this study can serve as a reference for evaluating the performance of the next-generation of field portable XRF instruments. ACKNOWLEDGMENTS
W
e would like to especially thank Jean-Luc Bailly, MarieJos´e Clerc, Nicole Perruchoud, Wanda Stryjenska, and Gilbert Pfister for their help in performing experimental work. We also wish to acknowledge Jasmin Meltretter for useful comments on the project and Agathe Martignier for electron microscopy data as well as Ron Hogg for careful and critical reading of the manuscript.
REFERENCES 1. Kalnicky, D.J., and R. Singhvi: Field portable XRF analysis of environmental samples. J. Hazard. Mater. 83:93–122 (2001). 2. Hou, X.D., Y.H. He, and B.T. Jones: Recent advances in portable X-ray fluorescence spectrometry. Appl. Spectros. Rev. 39(1):1–25 (2004). 3. Palmer, P.T., R. Jacobs, P.E. Baker, K. Ferguson, and S. Webber: Use of field-portable XRF analyzers for rapid screening of toxic elements in FDA-regulated products. J. Agric. Food Chem. 57:2605–2613 (2009). 4. Clark, S., W. Menrath, M. Chen, S. Roda, and P. Succop: Use of a field portable X-ray fluorescence analyzer to determine the concentration of lead and other metals in soil samples. Ann. Agric. Environ. Med. 6:27–32 (1999). 5. Morley, J.C., C.S. Clark, J.A. Deddens, K. Ashley, and S. Roda: Evaluation of a portable X-ray fluorescence instrument for the determination of lead in workplace air samples. Appl. Occup. Environ. Hyg. 14(5):306–316 (1999). 6. Ashley, K., M. Hunter, L.H. Tait, J. Dozier, J.L. Seaman, and P.F. Berry: Field investigation of on-site techniques for the measurement of lead in paint films. Field Anal. Chem. Technol. 2(1):39–50 (1998). 7. DeKosky, R.: Developing chemical instrumentation for environmental use in the late twentieth century: Detecting lead in paint using portable X-ray fluorescence spectrometry. Ambix 56(2):138–162 (2009). 8. Pantazis, T., J. Pantazis, A. Huber, and R. Redus: The historical development of the thermoelectrically cooled X-ray detector and its impact on the portable and hand-held XRF industries (February 2009). X-Ray Spectrom. 39:90–97 (2010). 9. Campbell, J.L., and L.A. Crosse: Portable instrumentation for the determination of lead in painted surfaces. Bull. Environ. Contam. Toxicol. 16:469–473 (1976). 10. Kaplan, E.H., M.D. Lilley, R.F. Schaefer, et al.: Portable X-rayfluorescence instruments for analysis of lead in paints. Pub. Health Rep. 90(3):223–230 (1975). 11. Spurgeon, J.C.: “Response Characteristics of a Portable X-ray Fluorescence Lead Detector. Detection of Lead in Paint.” Springfield, VA: National Bureau of Standards, 1973. 12. Rasberry, S.D.: Investigation of portable X-ray fluorescence analyzers for determining lead on painted surfaces. Appl. Spectrosc. 27:102–108 (1973). 13. Laurer, G.R., T.J. Kneip, R.E. Albert, and F.S. Kent: X-ray fluorescence: Detection of lead in wall paint. Science 172:466–468 (1971). 14. U.S. Environmental Protection Agency: “Lead; Requirements for LeadBased Paint Activities in Target Housing and Child-Occupied Facilities; Final Rule.” EPA Toxic Substances Control Act, Title IV, Sections 402 and 404. Washington, D.C., 1996. 15. U.S. Department of Housing and Urban Development (HUD): “Guidelines for the Evaluation and Control of Lead-based Paint Hazards in Housing.” Office of Healthy Homes and Lead Hazard Control Second Edition. Washington, D.C.: HUD, 2012. 16. “Performance Characteristic Sheet.” Available at http://www.hud.gov/ offices/lead/lbp/hudguidelines/allpcs.pdf (accessed August 8, 2013). 17. sant´e, Minist`ere franc¸ais du travail de l’emploi et de la sant´e: “Arrˆet´e du 19 aoˆut 2011 relatif au constat de risque d’exposition au plomb.” 2011. 18. Schmehl, R.L., D.C. Cox, F.G. Dewalt, et al.: Lead-based paint testing technologies: Summary of an EPA/HUD field study. Am. Ind. Hyg. Assoc. J. 60:444–451 (1999). 19. White, K.T., F.G. Dewalt, D.C. Cox, R.L. Schmehl, W. Friedman, and E.A. Pinzer: Pilot evaluation for lead-based paint proficiency testing of field portable XRF instruments. J. ASTM Inter. 8(3):1–5 (2011). 20. Driscoll, J.N., C. Wood, T. Powell, and J.S. Askew: An extractive x-ray fluorescence method for field screening lead paint sites. Appl. Occup. Environ. Hyg. 9:206–211 (1994). 21. Afshari, S., V. Nagarkar, and M.R. Squillante: Quantitative measurement of lead in paint by XRF analysis without manual substrate correction. Appl. Radiat. Isot. 48:1425–1431 (1997).
Journal of Occupational and Environmental Hygiene
August 2014
537
