1、Designation: D3975 93 (Reapproved 2014)Standard Practice forDevelopment and Use (Preparation) of Samples forCollaborative Testing of Methods for Analysis of Sediments1This standard is issued under the fixed designation D3975; the number immediately following the designation indicates the year oforig
2、inal adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon () indicates an editorial change since the last revision or reapproval.1. Scope1.1 This practice establishes uniform general procedures forthe dev
3、elopment, (preparation) and use of samples in thecollaborative testing of methods for chemical analysis ofsediments and similar materials.1.2 The principles of this practice are applicable to aqueoussamples with suitable technical modifications.1.3 This standard does not purport to address all of th
4、esafety concerns, if any, associated with its use. 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:2D596 Guide for Reporting Re
5、sults of Analysis of WaterD1129 Terminology Relating to WaterD2777 Practice for Determination of Precision and Bias ofApplicable Test Methods of Committee D19 on WaterD3670 Guide for Determination of Precision and Bias ofMethods of Committee D22D3976 Practice for Preparation of Sediment Samples forC
6、hemical AnalysisE691 Practice for Conducting an Interlaboratory Study toDetermine the Precision of a Test Method3. Terminology3.1 DefinitionsFor definition of terms used in thispractice, refer to Terminology D1129.4. Summary of Practice4.1 Test samples of adequately defined composition andhomogeneit
7、y are required for evaluating the precision and biasof test methods. These samples should be typical in all respectsto the sample for which the test method is applicable. Sampleswith three levels of concentration of the measurand arerecommended to evaluate the linearity of the test method.Acceptable
8、 test materials, in order of preference are: threesamples of different compositions; mixtures of two samples toobtain the desired concentration levels; samples prepared bydilution of a single sample.4.2 In the absence of samples of known composition, theuse of the spiking technique, in which standar
9、d additions ofknown constituents are made by established techniques, will beacceptable for evaluating the linearity and the bias of testmethods. In such a case, the bias statement will consist of theaccuracy of recovery of the spike.4.3 Comparison of a candidate test method with a standardtest metho
10、d of known precision and bias will constitute anacceptable technique for evaluation of precision and bias. Insuch comparative measurements, any convenient test samplesmay be used, provided they are shown to be stable during thetime required to make the intercomparison, and that themeasurement sequen
11、ces are chosen to minimize or eliminateerrors due to sample instability.5. Significance and Use5.1 The objective of this practice is to provide guidelines forthe preparation of samples for use in collaborative tests, toevaluate methods during their development, and for the evalu-ation of the precisi
12、on and bias of proposed test methods.5.2 Statements of the precision and bias are a mandatorypart of ASTM test methods. Such an evaluation is necessary toprovide guidance to the user as to the reliability of measure-ments that can be expected by its use. The statements aredeveloped on the basis of u
13、ser experience (ordinarily collab-orative tests) with the test method.5.3 The availability of test samples is a key requirement forcollaborative evaluation of test methods.6. Rationale6.1 The use of materials of demonstrated adequate homo-geneity and known composition enable evaluation of theprecisi
14、on and bias of a test method. Materials of known1This practice is under the jurisdiction of ASTM Committee D19 on Water andis the direct responsibility of Subcommittee D19.07 on Sediments, Geomorphology,and Open-Channel Flow.Current edition approved Jan. 1, 2014. Published February 2014. Originallya
15、pproved in 1980. Last previous edition approved in 2008 as D3975 93 (2008).DOI: 10.1520/D3975-93R14.2For 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 Do
16、cument Summary page onthe ASTM website.Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States1composition are required to identify biases. Consensus valuesobtained as the result of collaborative testing are not adequateto permit quantitative
17、 evaluation of biases, although they mayshow that such exist.6.2 Materials of adequate homogeneity but uncertain com-position may be used to determine the precision of a given testmethod. Both single-laboratory precision or between operator-laboratory precision can be evaluated.6.3 Samples at only o
18、ne compositional level can provideinformation for performance statements for only the concen-tration level of the test sample. Accordingly, other evidence isrequired to evaluate such factors as linearity and sensitivity.6.4 A series of samples, spanning the useful concentrationrange of the test meth
19、od, are required to evaluate the relation ofthe precision and bias to the concentration level measured.7. Requirements for Test Samples7.1 The collaborative testing of proposed analytical proce-dures ideally requires samples identical to, or closely resem-bling the materials for which the test metho
20、d is designed toanalyze. Because this is not always possible, subrogate mate-rials must often be used. Acceptable kinds of samples may beclassified as:7.1.1 AuthenticSamples identical in all respects to typicaltest specimens.7.1.2 SyntheticSamples synthesized to be equivalent totypical test samples.
21、 Spiked samples are a special class ofsynthetic samples.7.1.3 SimulativeSamples that do not resemble typicalsamples but that possess some parameter of concern of the testmethod.7.2 As one proceeds down the list, the ability of the materialto test the analytical method becomes more controversial.7.3
22、Test samples must have stable compositions during thetest period.7.4 Test samples must be sufficiently homogeneous toevaluate the test method. The degree of homogeneity is relatedto the size of sample analyzed. Hence the minimum size ofsample meeting a specified homogeneity must be stated, andthis m
23、ust be equal to or smaller than the size of samplespecified in the test method.7.5 The sample must be available in sufficient quantity bothfor the requirements of the collaborative test and to permitfurther examination to resolve any operational questions.Preferably, an additional amount should be a
24、vailable forpossible examination by other techniques. A desirable objec-tive would be the establishment of sample banks for future usein: (1) testing refinement of the method and (2) testing othermethodologies.7.6 The uncertainties in homogeneity or composition, orboth, of test samples should not ex
25、ceed one third of themeasurement uncertainties they are expected to evaluate.When this specification is met, sample uncertainties make anegligible contribution to the variance of the test results. Whilereplicate samples may be used in practical analysis to averageout sample variations, this is usual
26、ly not feasible in the case ofcollaborative test samples, hence homogeneity should beattained, as far as possible.8. Preparation of Test Samples8.1 Authentic test samples may be procured from naturalsources or from suppliers of such materials. The specificationsfor such materials must be established
27、 on the basis of therequirements for matrix composition, composition levels, andsample size-homogeneity considerations of the particular testmethod. Ordinarily, such materials will need to bedisaggregated, sieved, and blended to obtain requisite homo-geneity. Pulverizing samples is not recommended b
28、ecause theresulting changes in particle size distribution essentially pre-clude the use of the original size distribution or surface area fornormalizing analytical results. Much of the sample to sampleconcentration variability, commonly observed in naturalsediments, results from differences in the p
29、article size distri-bution. Samples may need preliminary drying (preferablyfreeze-drying) and the drying procedure necessary to obtainconstant weight should be established. Practice D3976 pro-vides guidance in this respect.8.2 Synthetic test samples are prepared to be equivalent totypical analytical
30、 samples as far as practicable. Materialsrequired to prepare the test samples include those to simulatethe matrix and the parameter(s) of the test method. The extentto which these simulate a typical sample will determine thereliability of its use as a test material. Homogeneity of mixingof synthetic
31、 samples must always be verified.8.3 Spiked samples consist of those in which the substanceto be measured is added to an authentic or synthetic testsample, which may or may not already count measurable levelsof the constituent added. The spike may be a pure substance ora mixture or solution containi
32、ng a known amount of thesubstance that is added quantitatively to the dried sediment.The spiking procedure must be developed in accordance withthe requirements of the specific situation. Matrix consider-ations are always of concern in that the added material may notbehave as it would in the case of
33、natural occurrence. Thoroughmixing of the spiked sample is very important, although lesscritical if the entire sample is used in the subsequent measure-ment.8.4 Samples that only possess some parameter of interest,hence classified as “simulative,” should be used only when it isnot feasible to obtain
34、 “authentic” or “synthetic” samples. Theuse of several simulative samples of differing matrices isrecommended to minimize dependence of the test results onmatrix effects.9. Test Levels9.1 The test specimens used in a collaborative test shouldspan the concentration range for which the test method ise
35、xpected to be valid. Three levelsa high, intermediate, andlow levelare recommended. This requirement can be met bythe use of several test samples, by a dilution techniqueinvolving a single sample, by mixing of two samples, or by thespiking technique.9.2 Two samples of differing concentration but wit
36、h similarmatrices may be quantitatively mixed to provide several testD3975 93 (2014)2levels. Samples must be thoroughly blended to obtain therequired homogeneity. In the absence of this, test samples maybe individually prepared and analyzed in their entirety. Theexpression used to calculate the comp
37、osition of a blend of twosamples, A and B, is as follows:A1B5 AWA1BWB/WA1WBwhere:A= weight percent (or ppm) of constituent a in sampleA,B= weight percent (or ppm) or constituent a in sampleB,A+B= weight percent (or ppm) of constituent a in mixture,WA= weight of sample A in mixture, andWB= weight of
38、sample B in mixture.Note that Aand Bmust be substantially different to providea range of concentrations in the mixture. This technique is bestutilized when A and B, respectively represent typical high- andlow-level compositions. One mixture, A+B, will provide thedesired intermediate level.9.3 A samp
39、le containing the desired high-level concentra-tion of constituent a may be diluted with a second samplecontaining an insignificant concentration of a to obtain sampleswith several test levels. The matrix composition of the diluentshould be similar to that of the test sample. The composition ofthe m
40、ixture is calculated from the following:A1D5 AWA/WA1WDwhere:WD= weight of diluent sample mixed with WA.The same considerations for homogeneity apply as discussedin 9.2.9.4 The spiking technique may be used to prepare samplesof any concentration higher than that of the base matrix. Thespike, in suita
41、ble form, is added to the base matrix and theresulting concentration is calculated as follows:Sa5 AWA1100Sa/WA1Wswhere:Sa= absolute weight of a added in the spiking operation,Sa= weight percent of constituent a in spiked sample, andWS= total weight of spiking material added.It may be necessary to ca
42、lculate Safrom the expressionSa5 SaWs/100The same considerations for homogeneity apply as in thecase of 9.2.Recommended levels, Sn, are as follows:S1= WAS2=4 WAS3=9 WAS4=49 WAS5=99 WAOrdinarily, three levels are used, namely: S1, S2, and S3;or,S1,S4, and S5.9.5 The evaluation of the bias of a test m
43、ethod is ordinarilybased on comparison of the “true” value with the measuredvalue. The recovery of a spike may also be used to evaluatebias.10. Verification of Composition of Test Samples10.1 The adequacy of the test samples can never be assumedand must always be verified before they may be used in
44、acollaborative study. This should be done by a single laboratoryof high competence, using a precise technique, so that mea-surement uncertainty is limited to within-laboratory variability.However, the collaborative exercise must be designed to detectany contribution due to variability of the sample.
45、10.2 The homogeneity of a test sample may be verified bymeasurement of an adequate number of subsamples of speci-fied size, using a test method of known requisite precision.10.3 The test method used for establishing the compositionof a test sample must be one of known and adequate accuracy.This norm
46、ally means a standard method of test or its equiva-lent. In the absence of a standard method of test, at least twoindependent test methods of known adequate precision andbelieved to be bias-free should be used and the agreement ofthe results will constitute evidence of accuracy of the reportedvalue.
47、 It is recommended to use standard reference materials ascontrols in each of the above described analytical situations, asfar as they are available.11. Comparison Technique11.1 The problems of an adequate test sample may becircumvented by intercomparing a candidate test method witha standard method
48、of test, using any conveniently available testmaterial. A standard method of test must be available, which isoften not the case, and furthermore the laboratories making thecomparison must be expert in its utilization. Furthermore, theintercomparison must be made according to a statistically validpla
49、n.12. Collaborative Test12.1 Planning the Collaborative Test:12.1.1 Because of the technical diversity of test methodsand practices within the scope of responsibility of CommitteeD19, it is not possible to establish a single rigid protocol forcollaborative testing. Accordingly, the responsibility for plan-ning and conducting an adequate collaborative test is delegatedto the corresponding task group. All aspects including initialplanning, conducting the test program, and analyzing andinterpreting the test results shall be consistent with followingguidelines:
