1、Designation: D 6620 06Standard Practice forAsbestos Detection Limit Based on Counts1This standard is issued under the fixed designation D 6620; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revision. A number in
2、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 presents the procedure for determining thedetection limit (DL)2for measurements of fibers or structures3using microscopy methods.1
3、.2 This practice applies to samples of air that are analyzedeither by phase contrast microscopy (PCM) or transmissionelectron microscopy (TEM), and samples of dust that areanalyzed by TEM.1.3 The microscopy methods entail counting asbestos struc-tures and reporting the results as structures per cubi
4、c centime-ter of air (str/cc) or fibers per cubic centimeter of air (f/cc) forair samples and structures per square centimeter of surface area(str/cm2) for dust samples.2. Referenced Documents2.1 ASTM Standards:4D 1356 Terminology Relating to Sampling and Analysis ofAtmospheresD 5755 Test Method for
5、 Microvacuum Sampling and Indi-rect Analysis of Dust by Transmission Electron Micros-copy for Asbestos Structure Number Surface LoadingD 6281 Test Method for Airborne Asbestos Concentrationin Ambient and Indoor Atmospheres as Determined byTransmission Electron Microscopy Direct Transfer (TEM)D 6480
6、Test Method for Wipe Sampling of Surfaces, Indi-rect Preparation, and Analysis for Asbestos StructureNumber Surface Loading by Transmission Electron Mi-croscopyE 456 Terminology Relating to Quality and Statistics3. Terminology3.1 Definitions of Terms Specific to This Standard:3.1.1 average, nthe sum
7、 of a set of measurements(counts) divided by the number of measurements in the set.3.1.1.1 DiscussionThe average is distinguished from themean. The average is calculated from data and serves as anestimate of the mean. The mean (also referred to as thepopulation mean, expected value,orfirst moment) i
8、s a param-eter of the underlying statistical distribution of counts.3.1.2 background, na statistical distribution of structuresintroduced by (i) analyst counting errors and (ii) contaminationon an unused filter or contamination as a consequence of thesample collection and sample preparation steps.3.
9、1.2.1 DiscussionThis definition of background is spe-cific to this practice. The only counting errors considered inthis definition of background are errors that result in anover-count (that is, false positives). Analyst counting errors areerrors such as, determining the length of structures or fiber
10、sand whether, based on length, they should be counted; countingartifacts as fibers; determining the number of structures pro-truding from a matrix; and interpreting a cluster as one, two, ormore structures that should be counted only as zero or onestructure. For purposes of developing the DL, assume
11、 thatbackground contamination sources have been reduced to theirlowest achievable levels.3.1.3 blank, na filter that has not been used to collectasbestos from the target environment.3.1.3.1 DiscussionBlanks are used in this practice todetermine the degree of asbestos contamination that is reflectedi
12、n asbestos measurements. Contamination may be on the virginfilter or introduced in handling the filter in the field or whenpreparing it for inspection with a microscope. The datarequired to determine the degree of contamination consists,therefore, of measurements of field blanks that have experi-enc
13、ed the full preparation process.3.1.4 decision value, na numerical value used as a bound-ary in a statistical test to decide between the null hypothesisand the alternative hypothesis.3.1.4.1 DiscussionIn the present context, the decisionvalue is a structure count that defines the boundary between“be
14、low detection” (the null hypothesis) and “detection” (thealternative hypothesis). If a structure count were larger than thedecision value, then one would conclude that detection has1This practice is under the jurisdiction of ASTM Committee D22 on Air Qualityand is the direct responsibility of Subcom
15、mittee D22.07 on Sampling and Analysisof Asbestos.Current edition approved Oct. 1, 2006. Published October 2006. Originallyapproved in 2000. Last previous edition approved in 2000 as D 6620 00.2The DL also is referred to in the scientific literature as Limit of Detection(LOD), Method Detection Limit
16、 (MDL), and other similar descriptive names.3For purposes of general exposition, the term “structures” will be used in placeof “fibers or structures.” In the examples in Section 8, the specific term, “fiber” or“structure,” is used where appropriate. These terms are defined separately in Section3.4Fo
17、r referenced ASTM standards, visit the ASTM website, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For Annual Book of ASTMStandards volume information, refer to the standards Document Summary page onthe ASTM website.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C70
18、0, West Conshohocken, PA 19428-2959, United States.been achieved (that is, the sample is from a distribution otherthan the background distribution). If the count were less than orequal to the decision value, the result would be reported as“below detection,” which means that the sample cannot bediffe
19、rentiated from a sample that would have been collectedfrom the background distribution.3.1.5 detection limitthe mean of a structure count popu-lation that is sufficiently large so a measurement from thispopulation would have a high probability (for example, 0.95 orlarger) of exceeding the decision v
20、alue that determines detec-tion.3.1.5.1 DiscussionThe DL is the value of a parameter, thetrue mean of a structure count population in the statisticalhypothesis testing problem, that underlies the DL concept.Specifically, it is the true mean of the alternative hypothesisthat ensures a sufficiently hi
21、gh power for the statistical test thatdetermines detection.3.1.6 count, nthe number of fibers or structures identifiedin a sample.3.1.7 fiber, nany of various discrete entities with essen-tially parallel sides counted by a particular method thatspecifies length, width, and aspect ratio.3.1.7.1 Discu
22、ssionThe definitions of “fiber” and “struc-ture” are similar because the measurement method employedspecifies the shape, length, width, and aspect ratio.3.1.8 mean, nthe mean value of the number of structuresin the population of air or dust sampled.3.1.8.1 DiscussionThe mean in this definition is in
23、tendedto be the population mean, expected value, or first moment ofa statistical distribution. It is a theoretical parameter of thedistribution that may be estimated by forming an average ofmeasurements (refer to Terminology E 456 for definition ofpopulation).3.1.9 power, nthe probability that a cou
24、nt exceeds thedecision value for a sample that was obtained from a popula-tion other than the background population.3.1.9.1 DiscussionPower is the probability of selecting,based on a statistical test, the alternative hypothesis when it istrue. In the present context, this means the probability ofmak
25、ing the correct decision to report a structure concentrationfor a sample that was collected from a population other than thebackground population. The power of the statistical test equals1 minus the type II error rate.3.1.10 replicate, na second measurement is a replicate ofthe initial measurement i
26、f the second measurement is obtainedfrom an identical sample and under identical conditions as theinitial measurement.3.1.10.1 Discussion“Identical,” as applied to sample, canmean“ same subsample preparation,” “separate preparation ofa distinct subsample,” or a distinct sample obtained from thesame
27、population as the initial sample. For this practice,“identical” means distinct sample obtained from the samepopulation as the initial sample.3.1.11 sample, nthe segment of the filter that is inspected,and thereby, embodies the air or dust that was collected and thesubset of structures that were capt
28、ured on the portion of thefilter subjected to microscopic inspection (also, see Terminol-ogy D 1356).3.1.12 sensitivity, nthe structure concentration corre-sponding to a count of one structure in the sample.3.1.13 structure, nany of various discrete entities countedby a particular method that specif
29、ies shape, length, width, andaspect ratio.3.1.14 type I error, nchoosing, based on a statistical test,the alternative hypothesis over the null hypothesis when thenull hypothesis is, in fact, true; a false positive outcome of astatistical test.3.1.14.1 DiscussionA type I error would occur if thecount
30、 for a sample exceeded the decision value, but the samplewas, in fact, obtained from the background population. Theanalyst erroneously would be led by the statistical test to reporta structure concentration (that is, choose the alternative hy-pothesis of the statistical test), where the result shoul
31、d bereported as “below the detection limit” (that is, the nullhypothesis of the statistical test is true).3.1.15 type II error, nchoosing, based on a statistical test,the null hypothesis over the alternative hypothesis when thealternative hypothesis is, in fact, true; a false negative outcomeof a st
32、atistical test.3.1.15.1 DiscussionA type II error would occur if thecount for a sample does not exceed the decision value, but thesample was, in fact, obtained from a population other than thebackground population. The analyst would erroneously be ledby the statistical test to report a “below the de
33、tection limit”result (that is, choose the null hypothesis of the statistical test),where the result should be reported as a structure concentration(that is, the alternative hypothesis of the statistical test is true).3.1.16 type I error rate, nthe probability of a type I error(also referred to as th
34、e significance level, a-level,orp-value ofthe statistical test).3.1.17 type II error rate, nthe probability of a type II error(also referred to as the b-level of the statistical test).3.1.18 llambda, the Greek letter used to represent thepopulation mean of a Poisson distribution.3.1.19 l0the populat
35、ion mean of the Poisson distributionof background counts.3.1.19.1 Discussionl0is the population mean of thePoisson distribution under the null hypothesis in the statisticalhypothesis testing problem that defines the DL.3.1.20 l1the population mean of the Poisson distributionunder the alternative hyp
36、othesis in the statistical hypothesistesting problem that defines the DL (DL = l1).3.1.21 x0decision value for determining detection. If thecount in a measurement is not greater than x0, the measurementis reported as “below detection.”3.1.22 Xa Poisson distributed random variable used todenote the n
37、umber of structures (fibers) counted in a sample.3.1.23 Athe area of the filter inspected to obtain astructure count.3.1.24 P(Xx/l,A)the Poisson probability of a structurecount exceeding x structures (fibers) when the population meanis equal to l and an area, A, of the filter is inspected.4. Signifi
38、cance and Use4.1 The DL concept addresses potential measurement inter-pretation errors. It is used to control the likelihood of reportinga positive finding of asbestos when the measured asbestos levelD6620062cannot clearly be differentiated from the background contami-nation level. Specifically, a m
39、easurement is reported as being“below the DL” if the measured level is not statisticallydifferent than the background level.4.2 The DL, along with other measurement characteristicssuch as bias and precision, is used when selecting a measure-ment method for a particular application. The DL should bee
40、stablished either at the method development stage or prior toa specific application of the method. The method developersubsequently would advertise the method as having a certainDL. An analyst planning to collect and analyze samples would,if alternative measurement methods were available, want tosel
41、ect a measurement method with a DL that was appropriatefor the intended application.5The most important use of theDL, therefore, takes place at the planning stage of a study,before samples are collected and analyzed.5. Descriptive Terms and Procedures5.1 Introduction:5.1.1 The DL is one of a number
42、of characteristics used todescribe the expected performance of a measurement method.6The DL concept addresses certain potential measurementinterpretation errors. Specifically, a measurement is reported asbeing “below the DL” if the measured level cannot bedistinguished from zero or from the randomly
43、 varying back-ground contamination level. Stated differently, the DL providesprotection against a false positive finding. When a measuredvalue is less than an appropriately specified decision value, theanalyst is instructed to disregard the measured value and reportthe result only as “below the DL.”
44、5.1.2 The DL concept for asbestos measurements, which arebased on microscopy, is simpler than the DL concept formeasurement methods that depend, for example, on spectros-copy. For asbestos, the measurement is derived from a directcount of discrete structures using a microscope. For spectros-copy met
45、hods, the measurement is indirect requiring a calibra-tion curve, and is subject to interferences and unspecifiedbackground signals that could be responsible for measurementvalues that are false positives.5.1.3 The sources of false positives for asbestos counts are(i) analyst errors (for example, de
46、termining the length ofstructures or fibers and whether, based on length, they shouldbe counted; counting artifacts as fibers; determining the num-ber of structures protruding from a matrix; interpreting acluster as one, two, or more structures that should be countedonly as zero or one), and (ii) co
47、ntamination (for example,virgin filter contamination or contamination introduced duringsample collection or sample preparation). Collectively, thesesources are referred to subsequently as “background.” Forpurposes of developing the DL, assume that each backgroundsource has been reduced to its lowest
48、 achievable level.5.2 DLGeneral Discussion:5.2.1 DLs often have been misspecified and misinterpretedbecause the DL concept has not been defined with sufficientclarity for translation into operational terms; however, the DLconcept and operational implementation have been presentedcorrectly in the sci
49、entific literature by a number of authors.7These authors describe the DL as a theoretical value, specifi-cally the true mean concentration of a substance in a sampledmedium. This true mean, the DL, must be large enough toensure a high probability (for example, 0.95 or larger) ofconcluding based on one or more measurements from a sampleof the medium that the true concentration in the medium is, infact, greater than zero or greater than an appropriately definedbackground level. The DL, therefore, is a parameter in thestatistical decision that determi
