1、Designation: E 2120 00Standard Practice forthe Performance Evaluation of the Portable X-RayFluorescence Spectrometer for the Measurement of Lead inPaint Films1This standard is issued under the fixed designation E 2120; the number immediately following the designation indicates the year oforiginal ad
2、option or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon (e) indicates an editorial change since the last revision or reapproval.1. Scope1.1 This practice covers portable x-ray fluorescence (XRF)instruments int
3、ended for the measurement of lead in paint. It isintended that manufacturers apply this practice to one unit of aparticular model of an instrument when that model is initiallyavailable. Replicate tests on additional units of the same modelof an instrument are to be performed at the discretion of the
4、manufacturer. This practice also is intended for use by thirdparties performing independent evaluation of portable x-rayfluorescence instruments.1.2 All performance evaluation data are to be in Interna-tional System of Units (SI) units.1.3 Tests of performance are based on replicate measure-ments of
5、 certified reference paint films on a variety of substratematerials. Tests are performed to determine: bias, precision,linearity, limit of detection, interferences, substrate affects, andstability.1.4 This standard does not purport to address all of thesafety concerns, if any, associated with its us
6、e. It is theresponsibility of the user of this standard to establish appro-priate safety and health practices and determine the applica-bility of regulatory limitations prior to use.2. Referenced Documents2.1 ASTM Standards:D 3332 Test Methods for Mechanical Shock Fragility ofProducts, Using Shock M
7、achines2E 344 Terminology Relating to Thermometry and Hydrom-etry3E 456 Terminology Relating to Quality and Statistics4E 1605 Terminology Relating to Abatement of Hazardsfrom Lead-Based Paint in Buildings and Related Struc-tures52.2 ANSI Standards:American National Standard N5381979, “Classification
8、 ofIndustrial Ionizing Radiation Gauging Devices”6American National Standard N5421977, “Sealed Radioac-tive Sources, Classification”62.3 Underwriters Laboratory Standards:Underwriters Laboratory Standard 544, “Medical and Den-tal Equipment”7Underwriters Laboratory Standard 31011, “Chemical Ana-lyzer
9、s”73. Terminology3.1 Definitions:3.1.1 accuracy, nthe theoretical maximum error of ameasurement, expressed as the proportion of the amount beingmeasured without regard for the direction of the error, that isachieved with a given probability (typically 0.95) by themethod.83.1.2 bias, nthe discrepancy
10、 between the mean of thedistribution of measurements from a method and the trueconcentration being measured.3.1.3 limit of detection, nthe smallest (true) signal thatwill be detected with a probability 1 b (b is the probabilityof an error of the second kind, failing to decide that a substanceis pres
11、ent when it is), where the a posteriori decision mecha-nism has a built-in protection level, a (a is the probability ofan error of the first kind, deciding that the substance is presentwhen it is not), against falsely concluding that a blankobservation represents a “real” signal. The b and a termsty
12、pically are 5 % or 1 %, depending on the requirements of thetesting program.93.1.4 precision, nthe closeness of agreement betweenrepetitive test results obtained under prescribed conditions (see1This practice is under the jurisdiction of ASTM Committee E06 on Perfor-mance of Buildings and is the dir
13、ect responsibility of Subcommittee E06.23 onAbatement of Hazards from Lead in Buildings and Related Structures.Current edition approved Nov. 10, 2000. Published March 2001. Originallypublished as PS 116 99. Last previous edition PS 116 99.2Annual Book of ASTM Standards, Vol 15.09.3Annual Book of AST
14、M Standards, Vol 14.03.4Annual Book of ASTM Standards, Vol 14.02.5Annual Book of ASTM Standards, Vol 04.11.6Available from American National Standards Institute, 11 W. 42nd St., 13thFloor, new York, NY 10036.7Available from Underwriters Laboratories, Inc., Research Triangle Park, NC.8Kennedy, E.R.,
15、T.J. Fischbach, R. Song, P.M. Miller, and S.A. Shulman,Guidelines for Air Sampling and Analytical Method Development and EvaluationDHHS (NIOSH) Publication No. 95117. National Institute for Occupational Safetyand Health, Cincinnati, OH 45226, May 1995.9Currie, L.A., “Limits for Qualitative Detection
16、 and Quantitative Determina-tion,” Anal. Chem., 40 (3), pp. 586593, 1968.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.Terminology E 456). The precision of a single instrument is therandom component of its accuracy and is usually i
17、ndicated bythe value of the standard deviation.3.2 The definitions given in Terminologies E 344 andE 1605 shall apply to this practice.3.3 Definitions of Terms Specific to This Standard:3.3.1 battery charger, na means for recharging a portableinstruments self-contained battery pack, usually converti
18、ng110 V AC to low level DC power.3.3.2 cycle or reading time, na period of x-ray datacollection (counting) performed automatically by some instru-ments. Such time may be established by a standardizationprocedure that would adjust for variation in the strength of theradioactive source. Also, it might
19、 be adjustable on someinstruments to achieve different levels of measurement preci-sion. Depending on the instrument model, one “cycle time”may be equivalent to one “measurement time” or several“cycles” may be automatically or manually averaged to equalone “measurement time.” The time begins with th
20、e opening ofthe XRF instrument shutter to expose the paint film surface tothe source radiation and is concluded when the source shutteris closed.3.3.3 display unit, nan electronic device that presents theresults of the measurement to the user. Other parameters suchas total measurement time also may
21、be presented.3.3.4 measurement time, nthe duration of a single mea-surement observed in real time. A measurement may compriseseveral individual readings or cycles.3.3.5 measurement value, nthe readout of a lead concen-tration in mg/cm2obtained at the end of one cycle time (orseveral cycle times if m
22、ultiple readings are averaged) or at theend of one measurement time.3.3.6 probe, na hand-held device containing the radioac-tive source, x-ray detector, and associated mechanical andelectrical components that is placed against the test sample toperform the measurement. The probe may constitute a par
23、t orall of the XRF instrument.3.3.7 radioactive source, na radioactive material (forexample,57Co,109Cd, and241Am) that emits X rays or gammarays that serve to cause ionization of the lead atoms in thesample, and subsequently a cascade of higher energy electronsinto the vacated lower energy shells.As
24、 these electrons fall intothe lower energy orbitals, they emit energy in the form of Xrays that are characteristic of lead.3.3.8 standard paint films, nfree-standing, certified refer-ence paint films, that is, certified reference materials (CRMs),that are acquired from the National Institute of Stan
25、dards andTechnology (NIST) or a commercial vendor. The lead levels inthe standard paint films (CRMs) shall be based on “x” level forlead where“ x” is equal to the appropriate local, state, or federalaction level for lead in coatings (in mg/cm2of lead coating).The paint films shall be as follows:(Fil
26、m 1) 2x3.3.8.1 DiscussionAn example CRM in use is the NISTStandard Reference Material (SRM) 2579, which consists offive films at 2x). Next,perform ten 1-min measurements with each of these five CRMpaint films (2x) on five different, additional substrates. Representativesubstrates and their order of
27、testing are as follows:(1) 0.76-mm (0.030-in.) T6, 6061 aluminum sheet, or45-mm (1.75-in.) hollow core door, or 3.2-mm (0.125-in.)thick veneer paneling (cut to fit in the test box).(2) 13-mm (0.5-in.) thick gypsum board, plasterboard, or15- to 20-mm (0.625- to 0.75-in.) thick birch, fir, or oak boar
28、dscut to fit in the test box.(3) 6-mm (0.25-in.) thick “836” steel plate cut to fit in thetest box.(4) Uncolored cinderblock slab or poured concrete slab50-mm (2-in.) thick cut to fit in the test box.(5) 40-mm (1.5-in.) thick red brick (place three-four bricksside by side in the test box).7.1.2.2 Ca
29、lculate and report the mean and standard devia-tion for each of these thirty measurement conditions. Reportthe differences between each mean value and expected CRMvalue as the bias of the measurement for each substrate andconcentration. Include in the report a description of eachsubstrate material a
30、nd the method of substrate correction.NOTE 6Correction for substrate bias is to be performed according tothe manufacturers standard procedure.NOTE 7Some instruments may display a default value, for example, 1mg/cm2, if the measurement value is within some predetermined range ofan action level or oth
31、er level of concern. The default value cannot be usedin this evaluation; the actual measurement value must be used. If theactual measurement value cannot be determined (displayed), perform thisevaluation with a CRM that is removed from the action level or other levelof concern such that it does not
32、result in the display of a default value.7.1.2.3 PrecisionUsing the data from the measurementsperformed under 7.1.2.1 and 7.1.2.2, calculate the slope,standard deviation of the slope, intercept, and standard devia-tion of the intercept of the measured versus expected valuesusing a linear regression
33、analysis.10Also calculate and reportthe correlation coefficient. Prepare a plot of standard deviationsof the average of each of the ten measurement values versus theconcentrations of the standard films. Include these results in thereport of this evaluation.NOTE 8These results are not to be used as a
34、n indication of theinstruments accuracy. They are solely to characterize the repeatability ofthe readings.NOTE 9Bias and precision, as determined by this ASTM practice,provide only estimates of the bias and precision that may be achieved inthe field. Bias and precision in the field will depend upon
35、the instrumentcalibration, the form and composition of the substrate, and the structure ofthe paint film being analyzed.7.1.2.4 Linearity with Real-World SubstratesThe slope,standard deviation of the slope, and the correlation coefficientfrom 7.1.2.3 are taken as a measure of the linearity of respon
36、seof the instrument.7.1.3 Estimate of Limit of Detection (LOD):7.1.3.1 The procedure presented here for estimating LODhas been developed especially for radiochemical applications.It has the advantage that it allows precision to vary withconcentration, as predicted by counting statistics. This proce-
37、dure is a simplification of the general method described byL.A. Currie.9The determination of the limit of detection isbased on Hypothesis Testing, which is subject to two kinds oferror: deciding that lead is present at or above a specified level,when it is not (error of the first kind), and converse
38、ly, decidingthat lead is not present at or above a specified level, when infact, it is (error of the second kind). In this approach, readingsthat fall above a particular threshold value, xt, are classified as“containing lead,” while readings that fall below the thresholdare classified as “not contai
39、ning lead.” The analyst selects thethreshold value, xt, such that if the true concentration of lead ona surface is zero (0 mg/cm2), 95 % of the readings on thatzero-lead surface will fall below the threshold value, xt, andtherefore will be classified correctly as “not containing lead,”while only 5 %
40、 of the readings on that zero-lead surface will bemisclassified as “containing lead.” The limit of detection, xLOD,is the true concentration of a lead-bearing surface such that95 % of the readings on that surface will fall above thethreshold xtand therefore will be correctly classified as“containing
41、 lead,” while only 5 % of the readings on thatlead-bearing surface will be misclassified as “not containinglead.” Note that xLODis always higher than xt, and is almostalways at least twice xt.7.1.3.2 As noted, this procedure has the advantage that itallows precision to vary with concentration, as pr
42、edicted bycounting statistics, but it does not require intimate knowledgeof the instrument calibration. Simplifying assumptions are thatthe variance of the population, sx2, is linear with concentration(sx2= A + Bx), as is predicted by counting statistics, and thatthe distribution of readings is pure
43、ly Gaussian, rather than themore exact Poisson (see Appendix X1 for additional informa-tion).7.1.3.3 The procedure is as follows:10Hartley, T.F., Computerized Quality Control: Programs for the AnalyticalLaboratory. John Wiley (2) the 0.7x to 1.3x CRM covered bythe 0.2x to 0.5x CRM covered by the 0.2
44、x CRM; all paintedsurfaces toward the XRF unit. Calculate the mean and thestandard deviation for the ten measurements for both setupsand report the difference between the mean values and theexpected sum as bias to the instruments measurements due topaint thickness.NOTE 16Real paint films are not usu
45、ally as thick as the procedureprescribes. The purpose of the test is to show the potential impact, if any,of this particular condition, and also the ability of the instrument tocompensate for this particular condition.7.1.4.4 Effects of Paint LayeringInstruments have thepotential for yielding a low
46、response to lead within theinnermost layers of multi-layered coatings of paint. Thefollowing is performed to test for this potential source of error(bias). Using the system and test conditions described in 7.1.2,perform ten 1-min measurements with the 1-in. white pineboard and CRMs arranged in the f
47、ollowing configurations: the0.7x to 1.3x CRM covered by one, two, and three 0.2x CRMs,and the 1.5x to 2x CRM covered by one, two, and 0.2xCRMs; all painted surfaces toward the XRF units. CalculateE2120006the mean and the standard deviation for the ten measurementsfor all six of the setups and report
48、 the differences between themean values and the expected values as bias to the instrumentsmeasurements due to paint layering.NOTE 17Increasing thickness of paint and paint layering will result infewer lead X rays that are generated within the lowest layer(s) of paintfrom reaching the detector becaus
49、e of absorption of these X rays by upperlayers of paint, and decrease of radiation intensity with distance from thesource (Inverse Square Law). Any instrument showing no effect fromincreasing thickness or layering should be considered suspect.7.1.5 Interferences:7.1.5.1 No standard test for interferences can be prescribedbecause of the lack of standard reference paint films withknown levels of potential interferences. Experience has indi-cated, however, that the following potentially interfering ele-ments may be found in paint, or substrates, or both:
