PART 3000

METALS

Part CoordinatorsRandy A. Gottler and Marvin D. Piwoni

3010 Introduction Brian J. Condike3020 Quality Assurance/Quality Control David W. Tucker, Cindy A. Ziernicki3030 Preliminary Treatment of Samples .Jonathan Talbott3113 MetaIs by Electrothermal Atomic Absorption Spectrometry Raymond J. Lovett3125 MetaIs by Inductively Coupled PlasmaIMass Spectrometry William R. Kammin3130 MetaIs by Anodic Stripping Voltammetry Malgorzata Ciszkowska3500-( all) Brian J. Condike3500-Cr Chromium Rock J. Vitale

Chromium (3500-Cr) contains updated information on the stability of hexavalent chromium.Quality Assurance/Quality Control (3020) was revised to make it consistent with current regulatoryrequirements.

The effects of metais in water and wastewater range frombeneficial through troublesome to dangerously toxic. Some met-ais are essential to plant and animal growth while others mayadversely affect water consumers, wastewater treatment systems,and receiving waters. The benefits versus toxicity of some metaisdepend on their concentrations in waters.

Preliminary treatment is often required to present the metaIs tothe analytical methodology in an appropriate formo Altemativemethods for pretreatment of samples are presented in Section3030.

Metais may be determined satisfactorily by a variety of meth-ods, with the choice often depending on the precision and sen-sitivity required. Part 3000 describes colorimetric methods aswell as instrumental methods, i.e., atomic absorption spectrom-etry, including ftame, electrothermal (fumace), hydride, and coldvapor techniques; ftame photometry; inductively coupled plasmaemission spectrometry; inductively coupled plasma mass spec-trometry, and anodic stripping voltammetry. Flame atomic ab-sorption methods generally are applicable at moderate (0.1- to10-mglL) concentrations in clean and complex-matrix samples.Electrothermal methods generally can increase sensitivity if ma-trix problems do not interfere. Inductively coupled plasma emis-

sion techniques are applicable over a broad linear range and areespecially sensitive for refractory elements. Inductively coupledplasma mass spectrometry offers significantly increased sensi-tivity for some elements (as low as 0.01 /LglL) in a variety ofenvironmental matrices. Flame photometry gives good results athigher concentrations for several Group 1 and 11 elements. An-odic stripping offers high sensitivity for several elements inrelatively clean matrices. Colorimetric methods are applicable tospecific metal determinations where interferences are known notto compromise method accuracy; these methods may providespeciation information for some metais. Table 3010:1 lists themethods available in Part 3000 for each metal.

a. Dissolved metaIs: Those metais in an unacidified samplethat pass through a 0.45-/Lm membrane filter.

b. Suspended metaIs: Those metaIs in an unacidified samplethat are retained by a 0.45-/Lm membrane filter.

C. Total metaIs: The concentration of metais determined in anunfiltered sample after vigorous digestion, or the sum of theconcentrations of metais in the dissolved and suspended frac-tions. Note that total metais are defined operationally by thedigestion procedure.

d. Acid-extractable metaIs: The concentration of metaIs insolution after treatment of an unfiltered sample with hot dilutemineral acid. To determine either dissolved or suspended metais,filter sample immediately after collection. Do not preserve withacid until after filtration.

TABLE3010:1. ApPLlCABLEMETHODSFORELEMENTALANALYSlS

Flame Flame InductivelyAtomic Atomic Electrotherma1 Hydride/Co1d Coupled ICPlMass Anodic

Absorption Absorption Flame Atomic Vapor Atomic Plasma Spectrometry Stripping AlternativeElement (Direct) (Extracted) Photometry Absorption Absorption (lCP) (ICPIMS) Vo1tammetry Methodst

A]uminum 3111D 3111E 3113B 3120A 3125 3500-AI.BAntimony 311lB 3113B 3120A 3125Arsenic 3]]3B 3]]4B 3120A 3125 3500-As.BBarium 31]JD 3111E 3]]3B 3120A 3125Beryllium 311JD 3111E 3113B 3120A 3125Bismuth 311lB 3113B 3125*Boron 3]20A 3125* 4500-B.B,CCadmium 311lB 311lC 3]]3B 3]20A 3125 3]30BCalcium 31] lB,D 3111E 3120A 3125* 3500-Ca.BCesium 3]] IB 3125*Chromium 311lB 3111C 3113B 3120A 3125 3500-Cr.B,CCobalt 3111B 31 JlC 3113B 3120A 3125Copper 311lB 3111C 3113B 3120A 3125 3500-Cu.B,CGallium 3113B 3125*Germanium 3]]3B 3]25*Gold 311lB 31 ]3B 3125*lndium 3]]3B 3125*lridium 311lB 3125*

3-1

3-2 METAL8 (3000)

TABLE3010:1. CONTo

Flame Flame InductivelyAtomic Atomic Electrothermal Hydride/Cold Coupled ICPlMass Anodic

Absorption Absorption Flame Atomic Vapor Atomic Plasma Spectrometry Stripping AltemativeElement (Direct) (Extracted) Photometry Absorption Absorption (ICP) (lCPIMS) Voltammetry Methodst

Iron 3111B 311lC 3113B 3120A 3125* 3500-Fe.BLead 3111B 3111C 3113B 3120A 3125 3130B 3500-Pb.BLithium 3111B 3500-Li.B 3120A 3125*Magnesium 3111B 3120A 3125* 3500-Mg.B,CManganese 31l1B 3111C 3113B 3120A 3125 3500-Mn.BMercury 3112B 3125*Molybdenum 311lD 3111E 3113B 3120A 3125Nickel 3111B 3111C 3113B 3120A 3125Osmium 311lD 3111E 3125*Palladium 3111B 3125*Platinum 3111B 3125*Potassium 3111B 3500-K.B 3120A 3125* 3500-K.CRhenium 3111D 3111E 3125*Rhodium 3111B 3125*Ruthenium 31lIB 3125*Selenium 3113B 3114B,C 3120A 3125 3500-Se.C,D,ESilicon 311lD 31l1E 3120A 3125*Silver 3111B 3111C 31l3B 3120A 3125Sodium 3111B 3500-Na.B 3120A 3125*Strontium 31l1B 3500-Sr.B 3120A 3125Tellurium 3113B 3125*Thallium 3111B 3113B 3120A 3125Thorium 311lD 3111E 3125*Tin 3111B 3113B 3125*Titanium 311lD 3111E 3125*Uranium 3125Vanadium 311lD 311IE 3113B 3120A 3125 3500-V.BZinc 3111B 311IC 3113B 3120A 3125 3130B 3500-Zn.B

* Metal is not specificallymentionedin the method,but 3125 may be used successfullyin most cases.t Additionalaltemativemethods for aluminum,beryllium,cadmium,mercury,selenium,silver, and zinc may be found in the 19thEditionof Standard Methods.

3010 B. Sampling and Sample Preservation

Before collecting a sample, decide what fraction is to beanalyzed (dissolved, suspended, total, or acid-extractable). Thisdecision will determine in part whether the sample is acidifiedwith or without filtration and the type of digestion required.

Serious errors may be introduced during sampling and storagebecause of contamination from sampling device, failure to re-move residues of previous samples from sample container, andloss of metaIs by adsorption on and/or precipitation in samplecontainer caused by failure to acidify the sample properly.

The best sample containers are made of quartz or TFE. Be-cause these containers are expensive, the preferred sample con-tainer is made of polypropylene or linear polyethylene with apolyethylene capoBorosilicate glass containers also may be used,but avoid soft glass containers for samples containing metaIs inthe microgram-per-liter range. Store samples for determinationof silver in light-absorbing containers. Use only containers andfilters that have been acid rinsed.

Preserve samples immediately after sampling by acidifyingwith concentrated nitric acid (HN03) to pH <2. Filter samplesfor dissolved metaIs before preserving (see Section 3030). Usu-ally 1.5 mL cone HNOiL sample (or 3 mL 1 + 1 HNOiLsample) is sufficient for shorHerm preservation. For sampleswith high buffer capacity, increase amount of acid (5 mL may berequired for some alkaline or highly buffered samples). Usecommercially available high-purity acid* or prepare high-purityacid by sub-boiling distillation of acid.

After acidifying sample, preferably store it in a refrigerator atapproximately 4°C to prevent change in volume due to evapo-ration. Under these conditions, samples with metal concentra-tions of several milligrams per liter are stable for up to 6 months(except mercury, for which the limit is 5 weeks). For microgram-per-liter metal levels, analyze samples as soon as possible aftersample collection.

AIternatively, preserve samples for mercury analysis by add-ing 2 rnLlL 20% (w/v) K2Cr207 soIution (prepared in I + IHN03). Store in a refrigerator not contaminated with mercury.(CAUTION: Mercury concentrations may increase in samplesstored in pIastic bottles in mercury-contaminated Iaboratories.)

STRUEMPLER,A.W. 1973. Adsorption characteristics of silver, lead, cal-cium, zinc and nickel on borosilicate glass, polyethylene andpolypropylene container surfaces. Ana!. Chem. 45:2251.

FELDMAN,C. 1974. Preservation of dilute mercury solutions. Ana!.Chem. 46:99.

K!NG,W.O., J.M. RODRlGUEZ& C.M. WAI. 1974. Losses of trace con-centrations of cadmium from aqueous solution during storage inglass containers. Ana!. Chem. 46:771.

BATLEY,O.E. & O. OARDNER.1977. Sampling and storage of naturalwaters for trace metal analysis. Water Res. 11:745.

SUBRAMANIAN,K.S., C.L. CHAKRABARTI,J.E. SUETIAS& I.S. MAINES.1978. Preservation af some trace metais in samples of naturalwaters. Ana!. Chem. 50:444.

BERMAN,S. & P. YEATS.1985. Sampling of seawater for trace metais.Crit. Rev. Ana!. Chem. 16:1.

WENDLANDT,E. 1986. Sample containers and analytical accessoriesmade of modem plastics for trace analysis. Gewaess. Wass. Ab-wass. 86:79.

Avoid introducing contarninating metais from containers, dis-tilled water, or membrane filters. Some plastic caps or cap Iinersmay introduce metal contamination; for exampIe, zinc has beenfound in bIack bakelite-type screw caps as well as in many rubberand pIastic products, and cadmium has been found in pIastic pipettips. Lead is a ubiquitous contaminant in urban air and dust.

ThoroughIy clean sample containers with a metaI-free non-ionic detergent soIution, rinse with tap water, soak in acid, andthen rinse with metaI-free water. For quartz, TFE, or gIassmateriaIs, use I + I HN03, I + 1 HCI, or aqua regia (3 partscone HCI + I part cone HN03) for soaking. For plastic material,use I + I HN03 or 1 + 1 HCI. ReliabIe soaking conditions are24 h at 70°C. Chromic acid or chromium-free substitutes* maybe used to remove organic deposits from containers, but rinse

containers thoroughly with water to remove traces of chromium.Do not use chromic acid for plastic containers or if chromium isto be determined. Always use metaI-free water in analysis andreagent preparation (see 3111B.3c). In these methods, the word"water" means metaI-free water.

For analysis of microgram-per-liter concentrations of metaIs,airborne contaminants in the form of volatiIe compounds, dust,soot, and aerosols present in Iaboratory air may become signif-icant. To avoid contamination use "clean Iaboratory" facilitiessuch as commercially available Iaminar-flow clean-air benchesor custom-designed work stations and analyze blanks that reflectthe complete procedure.

MITCHELL,J.W. 1973. Ultrapurity in trace analysis. Anal. Chem. 45:492A.OARDNER,M., O. HUNT& O. TOPPING.1986. Analytical quality control

(AQC) for monitoring trace metais in the coastal and marine envi-ronment. Water Sei. Teehno!. 18:35.

General information and recommendations for quality assur-ance (QA) and quality control (QC) are provided in Sections1020 Quality Assurance, 1030 Data Quality, and 1040 MethodDevelopment and Evaluation. This section discusses QAJQCrequirements that are common to the analytical methods pre-sented in Part 3000.

The requirements described in this section are recommendedminimum QAJQC activities; refer to individual methods andregulatory program requirements for method-specific QAJQCrequirements. NOTE: If an individual method in Part 3000 hasrequirements more stringent than those shown here, follow re-

Joint Task Group: Cindy A. Ziemicki (chair), Katherine B. Adams, Myriam E.Cardenas, David Eugene Kimbrough; 20th Edition-David W. Tucker (chair),Nilda B. Cox, David W. Haddaway, Daniel A. McLean, Jonalea V. Ostlund, JohnT. Pivinski.

quirements of that method. If analyses are to be conducted forregulatory purposes and QAJQC requirements are set in regula-tions or in a reference method, follow those requirements.

Always consider the overall purpose of analyses. QAJQCmeasures and substantiation for operational-control determina-tions may differ significantly from those for determinations oftrace metaIs at water quality criteria IeveIs. LeveIs of tracemetaIs in environmental samples may be orders of magnitudeIower than in potential sources of contamination.

Use replicates of measurable concentration to establish preci-sion and known-additions recovery to determine bias. Useblanks, calibrations, controI charts, known additions, standards,and other ancillary measurement tools as appropriate. Provideadequate documentation and record keeping to satisfy clientrequirements and performance criteria established by the labo-ratory.

a. lnitial demonstration of capability: Verify analyst capabil-ity before analyzing any samples and repeat periodically todemonstrate proficiency with the analytical method. Verify thatthe method being used provides sufficient sensitivity for thepurpose of the measurement. Test analyst capability by analyz-ing at least four reagent water portions containing known addi-tions of the analyte of interest. Confirm proficiency by generat-ing analytical results that demonstrate precision and bias withinacceptable limits representative of the analytical method.

b. Method detection levei (MDL): Before samples are ana-lyzed, determine the MDL for each analyte by the procedures ofSection 1030, or other applicable procedure. I Determine MDL atleast annually for each method and major matrix category. Ver-ify MDL for a new analyst or whenever instrument hardware ormethod operating conditions are modified. Analyze samples forMDL determinations over a 3- to 5-d period to generate arealistic value. Preferably use pooled data from several analystsrather than data from a single analyst.

c. Dynamic range (DR): Before using a new method, deter-mine the dynamic range, i.e., the concentration range over whicha method has an increasing response (linear or second-order), foreach analyte by analyzing several standard solutions that bracketthe range of interest. Each standard measurement should bewithin 10% of the true value for acceptance into the DR deter-mination. Take measurements at both the low and high end of thecalibration range to determine method suitability. Analyticalinstrumentation with curve-fitting features may al10w utilizationof nonlinear instrument response.

a. lnitial calibration: Calibrate initially with a minimum of ablank and three calibration standards of the analyte(s) of interest.Select calibration standards that bracket the expected concentra-tion of the sample and that are within the method's dynamicrange. The number of calibration points depends on the width of thedynarnic range and the shape of the calibration curve. One calibra-tion standard should be at or below the reporting limit for themethod. As a general rule, differences between calibration standardconcentrations should not be greater than one order of magnitude(i.e., I, 10, 100, 1000). Apply linear or polynomial curve-fittingstatistics, as appropriate, for analysis of the concentration-instru-ment response relationship. The appropriate linear or nonlinearcorrelation coefficient for standard concentration to instrument re-sponse should be 2: 0.995. Use initial calibration for quantitation ofanalyte concentration in samples. Use calibration verification, CJ[ bbelow, only for checks on the initial calibration and not for samplequantitation. Repeat initial calibration daily and whenever calibra-tion verification acceptance criteria are not satisfied.

b. Calibration verification: Calibration verification is the pe-riodic confirmation that instrument response has not changedsignificantly from the initial calibration. Verify calibration byanalyzing a midpoint calibration standard (check standard) andcalibration blank at the beginning and end of a sample run,periodically during a run (normal1y after each set of ten sam-pIes). A check standard determination outside 90 to 110% of theexpected concentration indicates a potential problem. If a checkstandard determination is outside 80 to 120% of the expectedconcentration, immediately cease sample analyses and initiate cor-

rective action. For instrumental methods (3111, 3113, 3120, and3125), cease analysis and initiate corrective action if check stan-dards exceed 90 to 110%. Repeat initial calibration and sampledeterminations since the last acceptable calibration verification. Usecalculated control lirnits (Section 1020B) to provide better indica-tions of system performance and to provide tighter lirnits.

c. Quality contrai sample: Analyze an externally generatedquality control sample of known concentration at least quarterlyand whenever new calibration stock solutions are prepared.Obtain this sample from a source external to the laboratory orprepare it from a source different from those used to prepareworking standards. Use to validate the laboratory's workingstandards both qualitatively and quantitatively.

a. Method blank (MB): A method blank (also known as reagentblank) is a portion of reagent water treated exactly as a sample,including exposure to all equipment, glassware, procedures, andreagents. The MB is used to assess whether analytes or interferenceare present within the analytical process or system. No analyte ofinterest should be present in the MB at a warning leveI based on theend user's requirements. Undertake immediate corrective action forMB measurements above the MDL. Include a rninimum of one MBwith each set of 20 or fewer samples.

b. Laboratory-fortified blank (LFB): The laboratory-fortifiedblank (also known as blank spike) is a method blank that hasbeen fortified with a known concentration of analyte. It is usedto evaluate ongoing laboratory performance and analyte recov-ery in a c1ean matrix. Prepare fortified concentrations approxi-mating the rnidpoint of the calibration curve or lower with stocksolutions prepared from a source different from those used todevelop working standards. Calculate percent recovery, plotcontrol charts, and determine controllimits (Section 1020B) forthese measurements. Ensure that the LFB meets performancecriteria for the method when such criteria are specified. Establish

corrective actions to be taken in the event that LFB does notsatisfy acceptance criteria. Include a rninimum of one LFB witheach set of 20 or fewer samples.

c. Duplicates: Use duplicate samples of measurable concen-tration to measure precision of the analytical processo Randomlyselect routine samples to be analyzed twice. Process duplicatesample independently through entire sample preparation andanalytical processo Include a minimum of one duplicate for eachmatrix type with each set of 20 or fewer samples.

d. Laboratory-fortified matrix (LFM)/Laboratory-fortified ma-trix duplicate: Use LFM (also known as matrix spike) and LFMduplicate to evaluate the bias and precision, respectively, of themethod as infiuenced by a specific matrix. Prepare by adding aknown concentration of analytes to a randomly selected routinesample. Prepare addition concentrations to approximately doublethe concentration present in the original sample. If necessary,dilute sample to bring the measurement within the establishedcalibration curve. Limit addition volume to 5% or less of samplevolume. Calculate percent recovery and relative percent differ-ence, plot control charts, and determine control limits (Section1020B). Ensure the performance criteria for the method aresatisfied. Process fortified samples independent!y through entiresample preparation and analytical processo Include a minimum ofone LFMlLFM duplicate with each set of 20 or fewer samples.

e. Method ofknown additions: To analyze a new or unfamiliarmatrix, use the method of known additions (Section 1020B) todemonstrate freedom from interference before reporting concen-tration data for the analyte. Verify absence of interferences byanalyzing such samples undiluted and in a 1:10 dilution; resultsshould be within 10% of each other. Lirnit known-additionvolume to 10% or less of the sample volume.

4. Reference

I. D.S. ENVIRONMENTALPROTECfIONAGENCY.1995. Definition and pro-cedure for the determination of the method detection limit, revision1.11. 40 CFR Part 136, Appendix B. Federal Register 5:23703.

Samples contaJmng particulates or organic material generallyrequire pretreatment before spectroscopic analysis. "Total metals"includes ali metals, inorganically and organically bound, both dis-solved and particulate. Colorless, transparent samples (primarilydrinking water) having a turbidity of < I NTU, no odor, and singlephase may be analyzed directly by atornic absorption spectroscopy(fiame or electrothermal vaporization) or inductively coupled

* Approved by Standard Methods Cornmillee, 1997.Joint Task Group: 20th Edition-Jonathan Talboll (chair), Paul R. Fritschel, EllyM. Gabrielian, David Eugene Kirnbrough, H.M. Kingston, Nimi Kocherlakota,Dennis Neuin, Mark E. Tatro, Mark M. Vltis, Melissa A. Weekley, Aaron D.Weiss, Ruth E. Wolf.

plasma spectroscopy (atomic ernission or mass spectrometry) fortotal metals without digestion. For further verification or if changesin existing matrices are encountered, compare digested and undi-gested samples to ensure comparable results. On collection, acidifysuch samples to pH <2 with conc nitric acid (1.5 mL HNOiL isusually adequate for drinking water) and analyze directly. Digest aliother samples before deterrnining total metals. To analyze for dis-solved metals, filter sample, acidify filtrate, and store until analysescan be performed. To determine suspended metals, filter sample,digest filter and the material on it, and analyze. To determineacid-extractable metais, extract metals as indicated in Sections3030E through K and analyze extract.

This section describes general pretreatment for samples inwhich metaIs are to be determined according to Sections 3110through 3500-Zn with several exceptions. The special digestiontechniques for mercury are given in Sections 3112Bo4b and c,and those for arsenic and selenium in Sections 3114 and 3500-Se.

Take care not to introduce metals into samples during prelim-inary treatrnent. During pretreatment avoid contact with rubber,metal-based paints, cigarette smoke, paper tissues, and ali metalproducts including those made of stainless steel, galvanizedmetal, and brass. Conventional fume hoods can contribute sig-nificantly to sample contamination, particularly during acid di-gestion in open containers. Keep vessels covered with watchglasses and tum spouts away from incoming air to reduce air-bome contamination. Plastic pipet tips often are contaminated

with copper, iron, zinc, and cadmium; before use soak in 2N HCIor HN03 for several days and rinse with deionized water. Avoidusing colored plastics, which can contain metaIs. Use certifiedmetal-free plastic containers and pipet tips when possible. Avoidusing glass if analyzing for aluminum or silica.

Use metal-free water (see 3111B.3c) for alI operations. Checkreagent-grade acids used for preservation, extraction, and diges-tion for purity. If excessive metal concentrations are found,purify the acids by distillation or use ultra-pure acids. Induc-tively coupled plasma mass spectrometry (ICP-MS) may requireuse of ultra-pure acids and reagents to avoid measurable con-tamination. Process blanks through alI digestion and filtrationsteps and evaluate blank results relative to corresponding sampleresults. Either apply corrections to sample results or take othercorrective actions as necessary or appropriate.

If dissolved or suspended metais (see Section 301OA) are to bedetermined, filter sample at time of collection using a precondi-tioned plastic filtering device with either vacuum or pressure,containing a filter support of plastic or fiuorocarbon, through aprewashed ungridded 004- to 0045-I-Lm-pore-diammembrane fil-ter (polycarbonate or cellulose esters). Before use filter a blankconsisting of metal-free (deionized) water to insure freedomfrom contarnination. Precondition filter and filter device by rins-ing with 50 mL deionized water. If the filter blank containssignificant metaIs concentrations, soak membrane filters in ap-proximately 0.5N HCI or lN HN03 (recommended for electro-therrnal and ICP-MS analyses) and rinse with deionized waterbefore use.

Before filtering, centrifuge highly turbid samples in acid-washed fiuorocarbon or high-density plastic tubes to reduceloading on filters. Stirred, pressure filter units foul less readilythan vacuum filters; filter at apressure of 70 to 130 kPa. Afterfiltration acidify filtrate to pH 2 with cone HN03 and store untilanalyses can be perforrned. If a precipitate forrns on acidifica-tion, digest acidified filtrate before analysis as directed (seeSection 3030E). Retain filter and digest it for direct determina-tion of suspended metaIs.

If it is not possible to field-filter the sample without contam-inating it, obtain sample in an "unpreserved" bottle as above and

promptly cool to 4°C. Do not acid-preserve the sample. Then,without delay, filter sample under cleaner conditions in thelaboratory .

Test pH of a portion of aqueous sample upon receipt in thelaboratory to ensure that the sample has been properly filteredand acid-preserved. I

NOTE:Different filters display different sorption and filtrationcharacteristics2; for trace analysis, test filter and filtration sytemto verify complete recovery of metaIs.

If suspended metaIs (see Section 301OA) are to be determined,filter sample as above for dissolved metaIs, but do not centrifugebefore filtration. Retain filter and digest it for direct deterrnina-tion of suspended metaIs. Record sample volume filtered andinclude a filter in determination of the blank.

CAUTION: Do not use perchloric acid to digest membranefilters. (See 3030H for more information on handling HCI04).

1. D.S. ENVIRONMENTALPROTECTIONAGENCY. 1994. Sample PreparationProcedure for Spectrochemical Determination of Total RecoverableElements, Method 200.2. Environmental Monitoring Systems Lab.,Cincinnati, Ühio.

2. HOROWlTZ,A.J., K.R. LUM, J.R. GARBARlNO,G.E.M. HALL, C. LE-MIEUX& C.R. DEMAS. 1996. Problems with using filtration to definedissolved trace element concentrations in natural water samples.Environ. Sei. Teehnol. 30: 954.

Extractable metaIs (see Section 301OA) are lightly adsorbedon particulate material. Because some sample digestion may beunavoidable use rigidly controlled conditions to obtain meaning-fuI and reproducible results. Maintain constant sample volume,

acid volume, and contact time. Express results as extractablemetaIs and specify extraction conditions.

At collection, acidify entire sample with 5 mL cone HN03/Lsample. To prepare sample, mix well, transfer 100 mL to a

beaker or f1ask, and add 5 rnL 1 + 1 high-purity HCi. Heat 15min on a steam bath. Filter through a membrane filter (precon-ditioned as in Section 3030B) and carefully transfer filtrate to atared volumetric f1ask.Adjust volume to 100 rnL with metal-free

water, mix, and analyze. If volume is greater than 100 rnL,detem1ine volume to nearest 0.1 rnL by weight, analyze, andcorrect final concentration measurement by multiplying by thedilution factor (final volume -;- 100).

To reduce interference by organic matter and to convert metaisassociated with particulates to a form (usually the free metal)that can be determined by atomic absorption spectrometry orinductively-coupled plasma spectroscopy, use one of the diges-tion techniques presented below. Use the least rigorous digestionmethod required to provide acceptable and consistent recoverycompatible with the analytical method and the metal beinganalyzed.I-3

Nitric acid will digest most samples adequately (Section3030E). Nitrate is an acceptable matrix for both f1ame andelectrothermal atomic absorption and the preferred matrix forICP-MS.4 Some samples may require addition of perchloric,hydrochloric, hydrofluoric, or sulfuric acid for complete di-gestion. These acids may interfere in the analysis of somemetais and ali provide a poorer matrix for both electrothermaland ICP-MS analysis. Confirm metal recovery for each diges-tion and analytical procedure used. Use Table 3030:1 as aguide in determining which acids (in addition to HN03) to usefor complete digestion. As a general mie, HN03 alone isadequate for c1ean samples or easily oxidized materiais;HN03-H2S04 or HN03-HCI digestion is adequate for readilyoxidizable organic matter; HNOrHCI04 or HN03-HCI04-HFdigestion is necessary for difficult-to-oxidize organic matteror minerais containing silicates. Although dry ashing is notgenerally recommended because of the loss of many volatileelements, it may be helpful if large amounts of organic matterare present.

Dilute samples with Ag concentrations greater than 1 mg/L tocontain less than 1 mg Ag/L for f1ame atomic absorption meth-

TABLE 3030:1. ACIOS USED IN CONJUNCTION WITH HN03 FOR SAMPLE

PREPARATlON

Recommendedfor

Not Recommendedfor

May Be Helpfulfor

Th,PbAg, Pb, Ba

OrganicmateriaIs

SiliceousmateriaIs

ods and 25 J.Lg/Lor less for electrothermal analysis?.5,6 Toaddress problems with silver halide solubility in HN03, digestusing method 3030F.3b.

Report digestion technique used.Acid digestion techniques (Sections 3030E through I) gener-

ally yield comparable precision and bias for most sample typesthat are totally digested by the technique. Because acids used indigestion wiU add metais to the samples and blanks, minimizethe volume of acids used.

Because the acid digestion techniques (3030E and F) normallyare not total digestions, the microwave digestion procedure(3030K) may be used as an altemative. The microwave methodis a c1osed-vessel procedure and thus is expected to provideimproved precision when compared with hot-plate techniques.Microwave digestion is recommended for samples being ana-Iyzed by ICP-MS. The microwave digestion method is recom-mended for the analysis of Ag, AI, As, Ba, Be, Ca, Cd, Co, Cr,Cu, Fe, K, Mg, Mn, Mo, Na, Ni, Pb, Sb, Se, TI, V, and Zn.Microwave digestion may be acceptable for additional analytesprovided its performance for those elements is validated.

Suggested sample volumes are indicated below for f1ameatomic absorption spectrometry. Lesser volumes, to a minimumof 5 rnL, are appropriate for graphite fumace, ICP, and ICP-MS.Do not subsample volumes less than 5 rnL, especially whenparticulates are present. Instead dilute samples with elevatedanalyte concentrations after digestion. If the recommended vol-ume exceeds digestion vessel capacity, add sample as evapora-tion proceeds. For samples containing particulates, wide-borepipets may be useful for volume measurement and transfer.

When samples are concentrated during digestion (e.g., >100rnL sample used) determine metal recovery for each matrixdigested, to verify method validity. Using larger samples willrequire additional acid, which also would increase the concen-tration of impurities.

Estimated MetalConcentration

mg/L

<0.10.1-1010-100+

Sample VoJume*mL

1000100

10

BMetal concentration, mgJL = A X C

A = concentration of metal in digested solution, mg/L,B = final volume of digested solution, mL, andC = sample size, roL.

Prepare solid samples or liquid sludges with high solids con-tents on a weight basis. Mix sample and transfer a suitableamount (typically 1 g of a sludge with 15% total solids) directlyinto a preweighed digestion vessel. Reweigh and calculateweight of sample. Proceed with one of the digestion techniquespresented below. However, as these digestion methods are pre-dominantly for dissolved and extractable metais in aqueoussamples, other approaches may be more appropriate for solidsamples. For complete mineralization of solid samples, consultmethods available elsewhere. J .4.6.7 Report results on wet- ordry-weight basis as follows:

AXBMetal concentration, mg/kg (wet-weight basis) = ---

g sample

A X B 100Metal concentration, mg/kg (dry-weight basis) = --- X -

g sample D

where:A = concentration of metal in digested solution, mgIL,B = final volume of digested solution, mL, andD = total solids, % (see Section 25400).

Always prepare acid blanks for each type of digestion per-formed. Although it is always best to eliminate all relevant

sources of contamination, a reagent blank prepared with thesame acids and subjected to the same digestion procedure as thesample can correct for impurities present in acids and reagentwater. However, blank correction is not recommended for anyother sources of contamination such as impurities adsorbed onglassware.

1. BOUMANS,P.W.J.M., ed. 1987. Inductively Coupled Plasma EmissionSpectroscopy, Part lI: Applications and FundamentaIs. John Wiley &Sons, New York, N.Y.

2. U.S. ENVlRONMENTALPROTECTIONAGENCY. 1992. Test Methods forEvaluating Solid Waste, Physical/Chemical Methods, SW-846, 3rded. Update I, Methods 3005A, 301OA, 3020A & 3050A. Off. SolidWaste & Emergency Response, Washington, D.e.

3. HOENlG,M. & A.M. DEKERSABJEC.1996. Sample preparation steps foranalysis by atomic spectroscopy methods: Present status. Spectro-chim. Acta. B51: 1297.

4. JARV1S,K.E., A.L. ORAY& R.S. HOUK, eds. 1992. Sample preparationfor ICP-MS. Chapter 7 in Handbook of Inductively Coupled PlasmaMass Spectrometry. Blackie, Olasgow & London, U.K.

5. U.S. ENVlRONMENTALPROTECTJONAGENCY. 1994. Sample PreparationProcedure for Spectrochemical Determination of Total RecoverableElements, Method 200.2. Environmental Monitoring Systems Lab.,Cincinnati, Ohio.

6. K.rNGSTON,H.M. & S. HASWELL, eds. 1997. Microwave EnhancedChemistry: FundamentaIs, Sample Preparation and Applications.American Chemical Soc., Washington, D.C.

7. BOCK, R. 1979. A Handbook of Decomposition Methods in Analyt-ical Chemistry. Blackie, Olasgow, U.K.

Because of the wide variation in concentration levels detectedby various instrumental techniques and the need to deal ade-quately with sources of contamination at trace levels, thismethod presents one approach for high-Ievel analytes (>0.1mg/L) and another for trace levels (:::;0.1 mg/L).

1. Digestion for Flame Atomic Absorption and High-LevelConcentrations

a. Apparatus:1) Rot plate.2) Conical (erlenmeyer) fiasks, 125-rnL, or Griffin beakers,

150-rnL, acid-washed and rinsed with water.3) Volumetric fiasks, 100-rnL.4) Watch glasses, ribbed and unribbed.b. Reagent:Nitric acid, HN03, conc, analytical ar trace-metais grade.c. Procedure: Transfer a measured volume (100 rnL recom-

mended) of well-mixed, acid-preserved sample appropriate forthe expected metais concentrations to a ftask or beaker (see3030D for sample volume). ln a hood, add 5 rnL conc HN03. Ifa beaker is used, cover with a ribbed watch glass to minimizecontamination. Boiling chips, glass beads, or Hengar granulesmay be added to aid boiling and minimize spatter when high

concentration levels (> 10 mg/L) are being deterrnined. Bring toa slow boi! and evaporate on a hot plate to the 10west volumepossible (about 10 to 20 rnL) before precipitation occurs. Con-tinue heating and adding cone HN03 as necessary until digestionis complete as shown by a light-colored, clear solution. Do notlet sample dry during digestion.

Wash down ftask or beaker walls and watch glass cover (ifused) with metal-free water and then filter if necessary (seeSection 3030B). Transfer filtrate to a 100-rnL volumetric ftaskwith two 5-rnL portions of water, adding these rinsings to thevolumetric ftask. Cool, dilute to mark, and mix thoroughly. Takeportions of this solution for required metal determinations.

2. Digestion for Trace-Levei (::sO.1 mg/L) Concentrations forICP and ICP-MS1

a. Apparatus:1) Block heater, dry, with temperature control.2) Polypropylene tubes*, graduated, round-bottom tubes with

caps, 17 X 100 mm, acid-washed and rinsed with metal-freewater. Preferably use tubes that simultaneously match the anal-

ysis instrument autosampler and the block digester. A fit with thecentrifuge is secondary but also desirable.

3) Pipettors, assorted sizes or adjustable.4) Pipet tips.5) Centrifuge.b. Reagent:Nitric acid, HN03, cone, double distilled. tc. Procedure: Soak new polypropylene tubes and caps over-

night or for several days in 2N HN03. Triple rinse with metal-free water, and preferably dry in poly rackets or baskets in alow-temperature oven ovemight. Store c1eaned tubes in plasticbags before use. Pipet tips also may need to be c1eaned; evaluatebefore use.

Pipet 10 mL well-mixed, acid-preserved sample into a pre-c1eaned, labeled tube with a macropipet. With a minimum vol-ume change «0.5 mL), add appropriate amount of analyte formatrix fortified samples. With a pipet, add 0.5 mL cone HN03

(or 1.0 mL 1 + 1 HN03) to ali samples, blanks, standards, andquality control samples.

Place tubes in block heater in a hood and adjust temperature to105°C. Orape caps over each tube to allow escape of acid vaporswhile preventing contamination. NOTE:00 not screw on caps atthis time. Oigest samples for a minimum of 2 h. 00 not letsamples boi!. Add more cone nitric acid as necessary untildigestion is complete by observation of a c1ear solution.

Remove tubes from heat and coo!. Oilute back to original 10mL volume with metal-free water. Adjust over-volume samplesto next convenient gradation for calculations and note volume.(Apply concentration correction from Section 30300.) If tubescontain particulates, centrifuge and decant c1ear portion intoanother precleaned tube. Tighten screw caps and store at 4°Cuntil ready for analysis.

3. Reference

I. JARVIS,K.E., A.L. GRAY& R.S. HOUK,eds. 1992.Samplepreparationfor ICP-MS.Chapter7 in Handbookof InductivelyCoupledPlasmaMass Spectrometry.Blackie& Son, Ltd., Glasgow& London,U.K.

See 3030E.1a. The following also may be needed:Steam bath.

a. Nitric acid, HN03, cone, analytical grade or better (seeSection 3030E).

b. Hydrochloric acid, HCl, 1 + 1.c. Nitric acid, HN03, 1 + 1.

a. Total HNOiHCl: Transfer a measured volume of well-mixed, acid-preserved sample appropriate for the expected met-ais concentrations to a flask or beaker (see 30300 for samplevolume). In a hood add 3 mL cone HN03 and cover with aribbed watch glass. Place flask or beaker on a hot plate andcautiously evaporate to less than 5 mL, making certain thatsample does not boi1 and that no area of the bottom of thecontainer is allowed to go dry. Coo!. Rinse down walls ofbeakerand watch glass with a minimum of metal-free wat«r and add 5

mL cone HN03. Cover container with a nonribbed watchglass and return to hot plate. Increase temperature of hot plateso that a gentle reflux action occurs. Continue heating, addingadditional acid as necessary, until digestion is complete (gen-erally indicated when the digestate is light in color or does notchange in appearance with continued refluxing). Coo!. Add 10mL 1 + 1 HCI and 15 mL water per 100 mL anticipated finalvolume. Heat for an additional 15 min to dissolve any pre-cipitate or residue. Cool, wash down beaker walls and watchglass with water, filter to remove insoluble material that couldc10g the nebulizer (see Section 3030B), and transfer filtrate toa 100-mL volumetric flask with rinsings. Alternatively cen-trifuge or let settle overnight. Adjust to volume and mixthoroughly.

b. Recoverable HN03/HCl: For this less rigorous digestionprocedure, transfer a measured volume of well-mixed, acid-preserved sample to a flask or beaker. Add 2 mL 1 + 1 HN03and 10 mL 1 + 1 HCI and cover with a ribbed watch glass. Heaton a steam bath or hot plate until volume has beeil reduced tonear 25 mL, making certain sample does not boi!. Cool and filterto remove insoluble material or altematively centrifuge or letsettle overnight. Quantitatively transfer sample to volumetricflask, adjust volume to 100 mL, and mix.

For trace-levei digestion, use precautionary measures similarto those detailed in Section 3030E.

a. Nitric acid, HN03, conc. (See 3030E for acid grades.)b. Sulfuric acid, HZS04, conc.

Transfer a measured volume of well-mixed, acid-preservedsample appropriate for the expected metais concentrations toa f1ask or beaker (see 3030D for sample volume). Add 5 mL

conc HN03 and cover with a ribbed watch glass. Bring toslow boil on hot plate and evaporate to 15 to 20 mL. Add 5mL conc HN03 and 10 mL conc HZS04, cooling f1ask orbeaker between additions. Evaporate on a hot plate until densewhite fumes of S03 just appear. If solution does not clear, add10 mL conc HN03 and repeat evaporation to fumes of S03.Heat to remove all HN03 before continuing treatment. AllHN03 will be removed when the solution is clear and nobrownish fumes are evident. Do not let sample dry duringdigestion.

Cool and dilute to about 50 rnL with water. Heat to almostboiling to dissolve slowly soluble salts. Filter if necessary, thencomplete procedure as directed in Section 3030E.lc beginningwith, "Transfer filtrate ... "

See 3030E.la. The following also are needed:a. Safety shield.b. Safety goggles.c. Watch glasses.

a. Nitric acid, HN03, conc.b. Perchloric acid, HCI04.

c. Ammonium acetate solution: Dissolve 500 g NH4CzH30Zin 600 rnL water.

CAUTION: Heated mixtures of HCl04 and organic matter mayexplode violently. Avoid this hazard by taking the followingprecautions: (a) do not add HC104 to a hot solution containingorganic matter; (b) always pretreat samples containing organicmatter with HN03 before adding HC104; (c) avoid repeatedfuming with HC104 in ordinary hoods (For routine operations,use a water pump attached to a glass fÚme eradicator. Stainlesssteel fume hoods with adequate water washdown facilities areavailable commercially and are acceptable for use with HClO 4);

and (d) never let samples being digested with HC104 evaporateto dryness.

Transfer a measured volume of well-mixed, acid-preservedsample appropriate for the expected metais concentrations to af1askar beaker (see 3030D for sample volume). In a hood add 5rnL conc HN03 and cover with a ribbed watch glass. Evaporatesample to 15 to 20 rnL on a hot plate. Add 10 rnL each of concHN03 and HCI04, cooling f1ask or beaker between additions.Evaporate gently on a hot plate until dense white fumes ofHCI04 just appear. If solution is not clear, keep solution justboiling until it clears. If necessary, add 10 rnL conc HN03 tocomplete digestion. Cool, dilute to about 50 mL with water, andboil to expel any chlorine or oxides of nitrogen. Filter, thencomplete procedure as directed in 3030E.lc beginning with,"Transfer filtrate ... "

If lead is to be determined in the presence of high amounts ofsulfate (e.g., determination of Pb in power plant f1yash samples),dissolve PbS04 precipitate as follows: Add 50 rnL ammoniumacetate solution to f1askor beaker in which digestion was carriedout and heat to incipient boiling. Rotate container occasionally towet all interior surfaces and dissolve any deposited residue.Reconnect filter and slowly draw solution through it. Transferfiltrate to a 100-rnL volumetric f1ask, cool, dilute to mark, mixthoroughly, and set aside for determination of lead.

a. Hot plate.b. TFE beakers, 250-mL, acid-washed and rinsed with water.c. Volumetric flasks, 100-rnL, polypropylene ar other suitable

plastic.

a. Nitric acid, HN03, conc and I + 1.b. Perchloric acid, HCI04.

c. Hydrofluoric acid, HF, 48 to 51%.

CAUTION:See precautions for using HCL04 in 3030H; handleHF with extreme care and provide adequate ventilation, espe-cially for the heated solution. Avoid ali contact with exposedskin. Provide medical attention for HF burns.

Transfer a measured volume of well-mixed, acid-preservedsample appropriate for the expected metais concentrations into a

250-rnL TFE beaker (see 3030D for sample volume). Evaporateon a hot plate to 15 to 20 rnL. Add 12 mL conc HN03 andevaporate to near dryness. Repeat HN03 addition and evapora-tion. Let solution cool, add 20 rnL HCI04 and 1 rnL HF, and boiluntil solution is clear and white fumes of HCI04 have appeared.Cool, add about 50 rnL water, filter, and proceed as directed in3030E.l c beginning with, "Transfer fiItrate ... "

a. Microwave unit with programmable power (minimum 545W) to within ± 10 W of required power, having a corrosion-resistant, welJ-ventilated cavity and having ali electroilics pro-tected against corrosion for safe operation. Use a unit having arotating turntable with a minimum speed of 3 rpm to insurehomogeneous distribution of microwave radiation. Use onlylaboratory-grade microwave equipment and closed digestioncontainers with pressure relief that are specifically designed forhot acid.'

b. Vessels: Construction requires an inner liner of perflu-oroalkoxy (PFA) Teflon™,* other TFE, or composite fluorinatedpolymers,t capable of withstanding pressures of at least 760 ±70 kPa (110 ± 10 psi), and capable of controlled pressure reliefat the manufacturer's maximum pressure rating.

Acid wash ali digestion vessels and rinse with water ('l[ 2a).For new vessels or when changing between high- and low-concentration samples, clean by leaching with hot:j:hydrochloricacid (1:1) for a minimum of 2 h and then with hot nitric acid(1:1) for a minimum of 2 h; rinse with water and dry in a cleanenvironment. Use this procedure whenever the previous use ofdigestion vessels is unknown or cross-contamination from ves-seIs is suspected.

c. Temperature feedback control system, using shielded ther-mocouple, fiber-optic probe, or infrared detector.

d. Bottles, polyethylene, 125-mL, with caps.e. Thermometer, accurate to ± O.I°C.f Balance, large-capacity (1500 g), accurate to 0.1 g.g. Filtration or centrifuge equipment (optional).h. Plastic container with cover, l-L, preferably made of PFA

Teflon™§.

* Or equivalent.t Such as TFMTMor equivalent.

.:j: At temperatures greater than sooe, but not boiling.§ Or equivalem.

b. Nitric acid, HN03, cone, sub-boiling distilled. Non-sub-boiling acids can be used if they are shown not to contributeblanks.

NOTE:For microwave units equipped with temperature feed-back electronic controls, calibration of the microwave unit is notrequired provided performance specifications can be duplicated.

For cavity-type microwave equipment, evaluate absolutepower (watts) by measuring the temperature rise in 1 kg waterexposed to microwave radiation for a fixed time. With thismeasurement, the relationship between available power (W) andthe partial power setting (%) of the unit can be estimated, andany absolute power in watts may be transferred from one unit toanother. The calibration format required depends on type ofelectronic system used by manufacturer to provide partial mi-crowave power. Few units have an accurate and precise linearrelationship between percent power settings and absorbed power.Where linear circuits have been used, determine calibrationcurve by a three-point calibration method; otherwise, use themultiple-point calibration method.

a. Three-point calibration method: Measure power at 100%and 50% power using the procedure described in 'l[ 3c andcalculate power setting corresponding to required power in wattsas specified in the procedure from the two-point line. Measureabsorbed power at the calculated partial power setting. If themeasured absorbed power does not correspond to the calculatedpower within ± 10 W, use the multiple-point calibration method,'l[ 3b. Use this point periodically to verify integrity of calibration.

b. Multiple-point calibration method: For each microwaveunit, measure the following power settings: 100, 99, 98, 97, 95,90, 80, 70, 60, 50, and 40% using the procedure described in 'l[3c. These data are clustered about the customary working powerranges. Nonlinearity commonly is encountered at the upper endof the calibration curve. If the unit's electronics are known tohave nonlinear deviations in any region of proportional powercontrol, make a set of measurements that bracket the power to beused. The final calibration point should be at the partial powersetting that will be used in the test. Check this setting periodi-

cally to eva1uate the integrity of the ca1ibration. li a significantchange (± 10 W) is detected, re-evaluate entire calibration.

c. Equilibrate a 1arge volume of water to room temperature (23± 2°C). Weigh 1 kg water (l000 g ± 1 g) or measure (l000 rnL± 1 rnL) into a plastic, not glass, container, and measure thetemperature to ± 0.1°e. Condition microwave unit by heating aglass beaker with 500 to 1000 rnL tap water at full power for 5min with the exhaust fan on. Loosely cover p1astic container toreduce heat loss and place in normal sample path (at outer edgeof rotating turntab1e); circu1ate continuous1y through the micro-wave field for 120 s at desired power setting with exhaust fan onas it will be during normal operation. Remove plastic containerand stir water vigorously. Use a magnetic stirring bar insertedimmediately after microwave irradiation; record maximum tem-perature within the first 30 s to ± O.loC. Use a new sample foreach additional measurement. If the water is reused, return bothwater and beaker to 23 ± 2°C. Make three measurements at eachpower setting. When any part of the high-voltage circuit, powersource, or control components in the unit have been serviced orreplaced, recheck calibration power. If power output haschanged by more than ± 10 W, re-evaluate entire calibration.

Compute absorbed power by the following relationship:

(K) (Cp) (m) (!J.Dp=------

t

where:

P = apparent power absorbed by sample, W,K = conversion factor for thennochemical calories sec - I to

watts (4.184),Cp = heat capacity, thennal capacity, or specific heat (cal g-I

De-I) of water,m = mass of water sample, g,

!J.T = final temperature minus initial temperature, De, andt = time, s.

For the experimental conditions of 120 s and 1 kg water (Cpat 25°C = 0.9997), the calibration equation simplifies to:

Stable line voltage within the manufacturer' s specification isnecessary for accurate and reproducible calibration and opera-tion. During measurement and operation it must not vary bymore than ± 2 V. A constant power supply may be neceSSarY ifline voltage is unstable.

CAUTION: This method is designedfor microwave digestion ofwaters only. It is not intended for the digestion of solids, forwhich high concentrations of organic compounds may result inhigh pressures and possibly unsafe conditions.

CAUTION: As a safety measure, never mix di.fferent manufac-turers' vessels in the same procedure. Vessels constructed dif-ferently wili retain heat at di.fferent rates; control of heatingconditions assumes that ali vessels have the same heat-transfercharacteristics. Inspect casements for cracks and chemical cor-rosion. Failure to maintain the vessels' integrity may result incatastrophic failure.

Both prescription controls and performance controls are pro-vided for this procedure. Performance controls are the mostgeneral and most accurate. When equipment capability permits,use the performance criterion.

a. Performance criterion: The following procedure is basedon heating acidified samples in two stages where the first stageis to reach 160 ± 4° in 10 min and the second stage is to permita slow rise to 165 to l70°C during the second 10 mino Thisperformance criterion is based on temperature feedback controlsystem capability that is implemented in various ways by dif-ferent manufacturers. Because the temperature of the acid con-trols the reaction, this is the essential condition that will repro-duce results in this preparation method. Verification of temper-ature conditions inside the vessel at these specific times issufficient to verify the critical procedura1 requirements.

b. Prescription criterion: For ali PFA vesse1s without 1iners,a verified prograrn that meets the performance-based tempera-ture-time proti1e is 545 W for 10 min followed by 344 W for 10min using tive sing1e-wall PFA Teflon™II digestion vesse1s.zAny verified program for a given microwave unit depends onunit power and operationa1 power settings, heating times, num-ber, type, and placement of digestion vessels within the unit, andsarnple and acid volumes. The change in power, time, andtemperature profile is not directly proportional to the change inthe number of sample vessels. Any deviations from the verifiedprograrn conditions will require verification of the time-temper-ature profile to conform to the given two-stage profile. This maybe done by laboratory personnel if suitable test equipment isavailab1e, or by the manufacturer of the microwave equipment.

C. General conditions: Weigh entire digestion vesse1 assem-bly to 0.1 g before use and record (A). Accurately transfer 45 rnLof well-shaken samp1e into the digestion vessel. Pipet 5 rnL concHN03 into each vessel. Attach all safety equipment required forappropriate and safe vessel operation following manufacturer' sspecifications. Tighten cap to manufacturer' s specifications.Weigh each capped vesse1 to the nearest 0.1 g (B).

Place appropriate number of vesse1s even1y distributed in thecarousel. Treat sample b1anks, known additions, and duplicatesin the same manner as samples. For prescription contro1 on1y,when fewer samples than the appropriate number are digested,fill remaining vessels with 45 rnL water and 5 rnL conc HN03 toobtain full complement of vessels for the particular programbeing used.

Place carousel in unit and seat it carefully on turntable. Pro-gram microwave unit to heat samples to 160 ± 4°C in 10 minand then, for the second stage, to permit a slow rise to 165 tol70°C for 10 mino Start microwave generator, making sure thatturntable is turning and that exhaust fan is on.

At completion of the microwave program, let vessels cool forat least 5 min in the unit before removal. Cool samples furtheroutside the unit by removing the carousel and letting them coolon a bench or in a water bath. When cooled to room temperature,weigh each vessel (to 0.1 g) and record weight (C).

If the net weight of sample plus acid decreased by more than10%, discard sample.

Complete sample preparation by carefully uncapping andventing each vessel in a fume hood. Follow individual manufac-

turer's specifications for relieving pressure in individual vesseltypes. Transfer to acid-cleaned noncontaminating plastic bottles.If the digested sample contains particulates, filter, centrifuge, orsettle ovemight and decanto

5. Calculations

a. Dilution correction: Multiply results by 50/45 or 1.11 toaccount for the dilution caused by the addition of 5 mL acid to45 mL sample.

b. Discarding of sample: To determine if the net weight ofsample plus acid decreased by more than 10% during the diges-tion process, use the following calculation

[(B - A) - (C - A))(B _ A) x 100> 10% (l % for multilayervessels)

NOTE: When nitric acid digestion is used, recoveries of silverand antimony in some matrices may be unacceptably low. Verifyrecoveries using appropriate known additions.

Preferably include a quality-control sample in each loadedcarousel. Prepare samples in batches including preparationblanks, sample duplicates, and pre-digestion known additions.Determine size of batch and frequency of quality-control sam-ples by method of analysis and laboratory practice. The power ofthe microwave unit and batch size may prevent including one ormore of the quality-control samples in each carousel. Do notgroUP quality-control samples together but distribute themthroughout the various carousels to give the best monitoring ofdigestion.

I. KrNGSTON,H.M. & S. HAswELL,eds. 1997. Microwave EnhancedChemistry: FundamentaIs, Sample Preparation and Applications.AmericanChemicalSoe., Washington,D.e.

2. D.S. ENVfRONMENTALPROTECTIONAGENCY.1990.Microwaveassistedacid digestion of aqueous samples and extracts. SW-846 Method3015,Test Methodsfor EvaluatingSolidWaste.D.S. EnvironmentalProtectionAgency,Washington,D.C.

Because requirements for determining metais by atomic ab-sorption spectrometry vary with metal and/or concentration to bedetermined, the method is presented as follows:

Section 3111, Metais by Flame Atomic Absorption Spectrom-etry, encompasses:

• Determination of antimony, bismuth, cadmium, calcium,cesium, chromium, cobalt, copper, gold, iridium, iron, lead,lithium, magnesium, manganese, nickel, palladium, platinum,potassium, rhodium, ruthenium, silver, sodium, strontium, thal-lium, tin, and zinc by direct aspiration into an air-acetylene flame(3111B),

• Determination of low concentrations of cadmium, chro-mium, cobalt, copper, iron, lead, manganese, nickel, silver, andzinc by chelation with ammonium pyrrolidine dithiocarbamate(APDC), extraction into methyl isobutyl ketone (MIBK), andaspiration into an air-acetylene flame (3111C),

• Determination of aluminum, barium, beryllium, calcium,molybdenum, osmium, rhenium, silicon, thorium, titanium, and

vanadium by direct aspiration into a nitrous oxide-acetylenefiame (311lD), and

• Determination of low concentrations of aluminum and be-ryllium by chelation with 8-hydroxyquinoline, extraction intoMIBK, and aspiration into a nitrous oxide-acetylene flame(3111E).

Section 3112 covers determination of mercury by the coldvapor technjque.

Section 3113 concerns determination of micro quantities ofaluminum, antimony, arsenic, barium, beryllium, cadmium,chromium, cobalt, copper, iron, lead, manganese, molybdenum,nickel, selenium, silver, and tin by electrothermal atomic absorp-tion spectrometry.

Section 3114 covers determination of arsenic and selenium byconversion to their hydrides and aspiration into an argon-hydro-gen or nitrogen-hydrogen flame.

In flame atomic absorption spectrometry, a sample is aspiratedinto a flame and atomized. A light beam is directed through theflame, into a monochromator, and onto a detector that measures

the amount of light absorbed by the atomized element in theflame. For some metaIs, atomic absorption exhibits superiorsensitivity over flame emission. Because each metal has its owncharacteristic absorption wavelength, a source lamp composed ofthat element is used; this makes the method relatively free fromspectral or radiation interferences. The amount of energy at thecharacteristic wavelength absorbed in the flame is proportionalto the concentration of the element in the sample over a limited

concentration range. Most atomic absorption instruments alsoare equipped for operation in an emission mode, which mayprovide better linearity for some elements.

a. Chemical interference: Many metaIs can be deterrnined bydirect aspiration of sample into an air-acetylene flame. The mosttroublesome type of interference is termed "chemical" and re-sults from the lack of absorption by atoms bound in molecularcombination in the flame. This can occur when the flame is nothot enough to dissociate the molecules or when the dissociatedatom is oxidized immediately to a compound that will notdissociate further at the flame temperature. Such interferencesmay be reduced or eliminated by adding specific elements orcompounds to the sample solution. For example, the interferenceof phosphate in the magnesium determination can be overcomeby adding lanthanum. Similarly, introduction of calcium elimi-nates silica interference in the determination of manganese.However, silicon and metaIs such as aluminum, barium, beryl-Iium, and vanadium require the higher-temperature, nitrous ox-ide-acetylene flame to dissociate their molecules. The nitrousoxide-acetylene flame also can be useful in minimizing certaintypes of chemical interferences encountered in the air-acetyleneflame. For exampIe, the interference caused by high concentra-tions of phosphate in the determination of calcium in the air-acetylene flame is reduced in the nitrous oxide-acetylene flame.

MIBK extractions with APDC (see 3111C) are particularlyuseful where a salt matrix interferes, for example, in seawater.This procedure also concentrates the sample so that the detectionlimits are lowered.

Brines and seawater can be analyzed by direct aspiration butsample dilution is recommended. Aspiration of solutions con-taining high concentrations of dissolved solids often results insolids buildup on the bumer head. This requires frequent shut-down of the flame and cleaning of the burner head. Preferablyuse background correction when analyzing waters that contain inexcess of 1% solids, especially when the primary resonance lineof the element of interest is below 240 nm. Make more frequentrecovery checks when analyzing brines and seawaters to insureaccurate results in these concentrated and complex matrices.

Barium and other metaIs ionize in the flame, thereby reducingthe ground state (potentialIy absorbing) population. The additionof an excess of a cation (sodium, potassium, or lithium) havinga similar or lower ionization potential will overcome this prob-lem. The wavelength of maximum absorption for arsenic is193.7 nm and for selenium 196.0 nm-wavelengths at which theair-acetylene flame absorbs intensely. The sensitivity for arsenicand selenium can be improved by conversion to their gaseoushydrides and analyzing them in either a nitrogen-hydrogen or anargon-hydrogen flame with a quartz tube (see Section 3114).

b. Background correction: Molecular absorption and lightscattering caused by solid particles in the flame can cause erro-neously high absorption vaIues resulting in positive errors. Whensuch phenomena occur, use background correction to obtainaccurate values. Use any one of three types of background

correction: continuum-source, Zeeman, or Smith-Hieftje correc-tion.

1) Continuum-source background correction-A continuum-source background corrector utilizes either a hydrogen-filledhollow cathode lamp with a metal cathode or a deuterium arclamp. When both the line source hollow-cathode lamp and thecontinuum source are placed in the same optical path and aretime-shared, the broadband background from the elemental sig-nal is subtracted electronically, and the resultant signal will bebackground-compensated.

Both the hydrogen-filled hollow-cathode lamp and deuteriumarc lamp have lower intensities than either the line source hol-low-cathode lamp or electrodeless discharge lamps. To obtain avalid correction, match the intensities of the continuum sourcewith the line source hollow-cathode or electrodeless dischargelamp. The matching may result in lowering the intensity of theline source or increasing the slit width; these measures have thedisadvantage of raising the detection limit and possibly causingnonlinearity of the calibration curve. Background correctionusing a continuum source corrector is susceptible to interferencefrom other absorbing lines in the spectral bandwidth. Miscorrec-tion occurs from significant atomic absorption of the continuumsource radiation by elements other than that being determined.When a line source holIow-cathode lamp is used without back-ground correction, the presence of an absorbing line from an-other element in the spectral bandwidth will not cause an inter-ference unless it overlaps the line of interest.

Continuum-source background correction will not remove di-rect absorption spectral overlap, where an element other than thatbeing deterrnined is capable of absorbing the line radiation of theelement under study.

2) Zeeman background correction- This correction is basedon the principIe that a magnetic field splits the spectral line intotwo linearly polarized Iight beams paralIel and perpendicular tothe magnetic field. One is called the pi (7T) component and theother the sigma (O') component. These two light beams haveexactly the same wavelength and differ only in the plane ofpolarization. The 7T line will be absorbed by both the atoms ofthe element of interest and by the background caused by broad-band absorption and light scattering of the sample matrix. The O'

line will be absorbed only by the background.Zeeman background correction provides accurate background

correction at much higher absorption levels than is possible withcontinuum source background correction systems. It also virtu-alIy eliminates the possibility of error from structured back-ground. Because no additional light sources are required, thealignment and intensity Iimitations encountered using continuumsources are eliminated.

Disadvantages of the Zeeman method include reduced sensi-tivity for some elements, reduced linear range, and a "rollover"effect whereby the absorbance of some elements begins to de-crease at high concentrations, resulting in a two-sided calibrationcurve.

3) Smith- Hieftje background correction- This correction isbased on the principIe that absorbance measured for a specificelement is reduced as the current to the hollow cathode lamp isincreased while absorption of nonspecific absorbing substancesremains identical at all current levels. When this method isapplied, the absorbance at a high-current mode is subtractedfrom the absorbance at a low-current mode. Under these condi-

FLAME ATOMIC ABSORPTION SPECTROMETRY (3111)/lntroduction 3-15

TABLE31 11:I. ATOMICABSORPTlONCONCENTRATlONRANGESWITH detection levei is defined here as the concentration that producesDIRECTASPIRATIONATOMlCABSORPTION absorption equivalent to twice the magnitude of the background

lnstrument Optimum ftuctuation. Sensitivity and detection levels vary with the instru-

Wave- Detection Concentration ment, lhe element determined, the complexity of the matrix, and

length F1ame LeveI Sensitivity Range the technique selected. The optimum concentration range usuallyElement nm Gases* mglL mg/L mg/L starts from the concentration of several times the detection levei

and extends to the concentration at which the calibration curveAg 328.1 A-Ac 0.01 0.06 0.1-4

starts to ftatten. To achieve best results, use concentrations ofAI 309.3 N-Ac 0.1 I 5-100 samples and standards within the optimum concentration rangeAu 242.8 A-Ac 0.01 0.25 0.5-20Ba 553.6 N-Ac 0.03 0.4 1-20 of the spectrometer. See Table 3111:1 for indication of concen-

Be 234.9 N-Ac 0.005 0.03 0.05-2 tration ranges measurable with conventional atomization. In

Bi 223.1 A-Ac 0.06 0.4 l-50 many instances the concentration range shown in Table 3111:1Ca 422.7 A-Ac 0.003 0.08 0.2-20Cd 228.8 A-Ac 0.002 0.025 0.05-2Co 240.7 A-Ac 0.03 0.2 0.5-10 TABLE311l:II. lNTERLABORATORYPREClSlONANDBlASDATAFORCr 357.9 A-Ac 0.02 0.1 0.2-10Cs 852.1 A-Ac 0.02 0.3 0.5-15

ATOMICABSORPTIONMETHoDs-DlRECTASPIRATION

Cu 324.7 A-Ac 0.01 0.1 0.2-10ANDEXTRACTEDMETALS

Fe 248.3 A-Ac 0.02 0.12 0.3-10 RelativeIr 264.0 A-Ac 0.6 8 Relative SD Error No.ofK 766.5 A-Ac 0.005 0.04 0.1-2 Metal Conc.* SD* % % ParticipantsLi 670.8 A-Ac 0.002 0.04 0.1-2Mg 285.2 A-Ac 0.0005 0.007 0.02-2 Direct determination:Mn 279.5 A-Ac 0.01 0.05 0.1-10 Aluminum' 4.50 0.19 4.2 8.4 5Mo 313.3 N-Ac 0.1 0.5 1-20 Barium2 1.00 0.089 8.9 2.7 11Na 589.0 A-Ac 0.002 0.015 0.03-1 Beryllium' 0.46 0.0213 4.6 23.0 1INi 232.0 A-Ac 0.02 0.15 0.3-10 Cadmium3 0.05 0.Q108 21.6 8.2 26Os 290.9 N-Ac 0.08 I Cadmium' 1.60 0.1 I 6.9 5.1 16Pbt 283.3 A-Ac 0.05 0.5 1-20 Calcium' 5.00 0.21 4.2 0.4 8Pt 265.9 A-Ac 0.1 2 5-75 Chromium' 3.00 0.301 10.0 3.7 9Rh 343.5 A-Ac 0.5 0.3 Cobalt' 4.00 0.243 6.1 0.5 14Ru 349.9 A-Ac 0.07 0.5 Coppe~ 1.00 0.112 11.2 3.4 53Sb 217.6 A-Ac 0.07 0.5 1-40 Copper' 4.00 0.331 8.3 2.8 15Si 251.6 N-Ac 0.3 2 5-150 lron' 4.40 0.260 5.8 2.3 16Sn 224.6 A-Ac 0.8 4 10---200 lron3 0.30 0.0495 16.5 0.6 43Sr 460.7 A-Ac 0.03 0.15 0.3-5 Leadl 6.00 0.28 4.7 0.2 14Ti 365.3 N-Ac 0.3 2 5-100 Magnesium3 0.20 0.021 10.5 6.3 42V 318.4 N-Ac 0.2 1.5 2-100 Magnesium' 1.10 0.116 10.5 10.0 8Zn 213.9 A-Ac 0.005 0.02 0.05-2 Manganese' 4.05 0.317 7.8 1.3 16

Manganese3 0.05 0.0068 13.5 6.0 14* A-Ac = air-acetylene;N-Ac = nitrousoxide-acetylene.. Nickel' 3.93 0.383 9.8 2.0 14t The more sensitive217.0 nm wavelengthis recommendedfor instruments Silver3 0.05 0.0088 17.5 10.6 7with backgroundcorrectioncapabilities.

Silverl 2.00 0.07 3.5 1.0 10Copyright© ASTM.Reprintedwith permission.Sodium' 2.70 0.122 4.5 4.1 12Strontiuml 1.00 0.05 5.0 0.2 12Zinc3 0.50 0.041 8.2 0.4 48

tions, any absorbance due to nonspecific background is sub- Extracted determination:Aluminum2 300 32 10.7 0.7 15

tracted out and corrected for. Beryllium2 5 1.7 34.0 20.0 9Smith-Hieftje background correction provides a number of Cadrnium3 50 21.9 43.8 13.3 12

advantages over continuum-source correction. Accurate correc- Cobaltl 300 28.5 9.5 1.0 6tion at higher absorbance levels is possible and error from Copper' 100 71.7 71.7 12.0 8structured background is virtually eliminated. In some cases, Ironl 250 19.0 7.6 3.6 4spectral interferences also can be eliminated. The usefulness of Manganesel 21.5 2.4 11.2 7.4 8Smith-Hieftje background correction with e1ectrodeless dis- Molybdenum I 9.5 1.1 11.6 1.3 5charge lamps has not yet been established. Nickell 56.8 15.2 26.8 13.6 14

Lead3 50 11.8 23.5 19.0 8

4. Sensitivity, Detection Levels, and Optimum Silverl 5.2 1.4 26.9 3.0 7

Concentration Ranges * For direct deterrninations,mg/L; for extracteddetenninations,1J.g/L.Superscriptsrefer to referencenumbers.

The sensitivity of ftame atomic absorption spectrometry isSource:AMERICANSOCIETYFORTESTINGANDMATERIALS.1986.AnnualBookof

ASTMStandards.Volume 11.01,Water and EnvironmentalTechnology.Amer-defined as the metal concentration that produces an absorption of ican SocoTesting & MateriaIs,Philadelphia,Pa. Copyright© ASTM.Reprinted1% (an absorbance of approximately 0.0044). The instrument with perrnission.

TABLE3111:III. SINGLE-OPERATORPRECISIONANORECOMMENDEDCONTROLRANGESFORATOMICABSORPTIONMETHODS-DIRECTASPIRATIONANDEXTRACTEDMETALS

Relative SD No.of QCMetal Conc.* SD* % Participants Std.* Acceptable Range*

Direct determination:Aluminuml 4.50 0.23 5.1 15 5.00 4.3-5.7Berylliuml 0.46 0.012 2.6 10 0.50 0.46-0.54Calciuml 5.00 0.05 1.0 8 5.00 4.8-5.2Chromiuml 7.00 0.69 9.9 9 5.00 3.3-6.7Cobalt' 4.00 0.21 5.3 14 4.00 3.4-4.6Copperl 4.00 0.115 2.9 15 4.00 3.7-4.3Iranl 5.00 0.19 3.8 16 5.00 4.4-5.6Magnesiuml 1.00 0.009 0.9 8 1.00 0.97-\.03Nickel4 5.00 0.04 0.8 5.00 4.9-5.1Silverl 2.00 0.25 12.5 10 2.00 1.2-2.8Sodium4 8.2 0.1 1.2 5.00 4.8-5.2Strontiuml \.00 0.04 4.0 12 1.00 0.87-1.13Potassium4 1.6 0.2 12.5 1.6 1.0-2.2Molybdenum4 7.5 0.07 0.9 10.0 9.7-10.3Tin4 20.0 0.5 2.5 20.0 18.5-21.5Titanium4 50.0 0.4 0.8 50.0 48.8-51.2Vanadium 50.0 0.2 0.4 50.0 49.4-50.6

Extracted determination:Aluminuml 300 12 4.0 15 300 264- 336Cobaltl 300 20 6.7 6 300 220- 380Copperl 100 21 21 8 100 22- 178Ironl 250 12 4.8 4 250 180- 320Manganesel 21.5 202 10.2 8 25 17- 23Molybdenuml 9.5 1.0 10.5 5 10 5.5-14.5Nickell 56.8 9.2 16.2 14 50 22-78Silverl 5.2 1.2 23.1 7 5.0 0.5-9.5

* For direct determinations,mg/L; for extracteddeterminations,JLg/L.Superscriptsrefer to refereneenumbers.Souree:AMERICANSOCIETYFORTESTINGANDMATERIALS.1986.AnnualBookof ASTMStandards.Volume11.01,Waterand EnvironrnentalTeehnology.AmerieanSoe.

Testing & Materiais,Philadelphia,Pa. Copyright© ASTM. Reprintedwith permission.

may be extended downward either by scale expansion ar byintegrating the absorption signal over a long time. The range maybe extended upward by diJution, using a less sensitive wave-length, rotating the bumer head, or utilizing a microprocessar tolinearize the calibration curve at high concentrations.

Prepare standard solutions of known metal concentrations inwater with a matrix similar to the sample. Use standards thatbracket expected sample concentration and are within the meth-od's working range. Very di lute standards should be prepareddaily from stock solutions in concentrations greater than 500mglL. Stock standard solutions can be obtained from severalcommercial sources. They also can be prepared from NationalInstitute of Standards and Technology (NIST) reference materi-aIs or by procedures outlined in the following sections.

For samples containing high and variable concentrations ofmatrix materiaIs, make the major ions in the sample and thedilute standard similar. If the sample matrix is complex andcomponents cannot be matched accurately with standards, usethe method of standard additions, 3113BAd2), to correct formatrix effects. If digestion is used, carry standards through thesame digestion procedure used for samples.

a. Atomic absorption spectrometer, conslstlllg of a lightsource emitting the line spectrum of an element (hollow-cathodelamp or electrodeless discharge lamp), a device for vaporizingthe sample (usually a flame), a means of isolating an absorptionline (monochromatar ar filter and adjustable slit), and a photo-electric detector with its associated electronic amplifying andmeasuring equipment.

b. Bumer: The most common type of bumer is a premix,which introduces the spray into a condensing chamber for re-moval of large droplets. The bumer may be fitted with a con-ventional head containing a single slot; a three-slot Boling head,which may be preferred for direct aspiration with an air-acety-Iene flame; or a special head for use with nitrous oxide andacetylene.

c. Readout: Most instruments are equipped with either a dig-ital or null meter readout mechanism. Most modem instrumentsare equipped with microprocessors or stand-alone control com-puters capable of integrating absorption signals over time andlinearizing the calibration curve at high concentrations.

d. Lamps: Use either a hollow-cathode lamp ar an electrode-less discharge lamp (EDL). Use one lamp for each element beingmeasured. Multi-element hollow-cathode lamps generally pro-

vide lower sensitivity than single-element lamps. EDLs take alonger time to warm up and stabilize.

e. Pressure-reducing valves: Maintain supplies of fuel andoxidant at pressures somewhat higher than the controlled oper-ating pressure of the instrument by using suitable reducingvalves. Use a separate reducing valve for each gas.f Vent: Place a vent about 15 to 30 cm above the bumer to

remove fumes and vapors from the fiame. This precaution pro-tects laboratory personnel from toxic vapors, protects the instru-ment from cOITosive vapors, and prevents fiame stability frombeing affected by room drafts. A damper or variable-speedblower is desirable for modulating air fiow and preventing fiamedisturbance. Select blower size to provide the air fiow recom-mended by the instrument manufacturer. ln laboratory locationswith heavy particulate air pollution, use clean laboratory facili-ties (Section 301OC).

Some data typical of the precision and bias obtainablewith the methods discussed are presented in Tables 3111:IIand III.

Analyze a blank between sample or standard readings to verifybaseline stability. Rezero when necessary.

To one sample out of every ten (or one sample from eachgroup of samples if less than ten are being analyzed) add aknown amount of the metal of interest and reanalyze toconfirm recovery. The amount of metal added should beapproximately equal to the amount found. If little metal ispresent add an amount close to the middle of the linear rangeof the testo Recovery of added metal should be between 85and 115%.

Analyze an additional standard solution after every tensamples or with each batch of samples, whichever is less, toconfirm that the test is in control. Recommended concentra-tions of standards to be run, limits of acceptability, and

reported single-operator precision data are listed in Table311l:III.

See Section 3020 for additional recommended quality controlprocedures.

8. References

1. AMERICANSOCIETYFORTESTINGANDMATERIALS.1986. Annual Bookof ASTM Standards, Volume 11.01, Water and Environmental Tech-nology. American Soe. Testing & Materiais, Philadelphia, Pa.

2. D.S. DEPARTMENTHEALTH,EDUCATIONANDWELFARE.1970. WaterMetais No. 6, Study No. 37. D.S. Public Health ServoPubl. No. 2029,Cincinnati, Ohio.

3. D.S. DEPARTMENTHEALTH,EDUCATIONANDWELFARE.1968. WaterMetais No. 4, Study No. 30. D.S. Public Health Servo Publ. No.999-DTH-8, Cincinnati, Ohio.

4. D.S. ENVIRONMENTALPROTECTIONAGENCY.1983. Methods for Chem-ical Analysis of Water and Wastes. Cincinnati, Ohio.

KAHN,H.L. 1968. Principies and Practice of Atomic Absorption. Advan.Chem. Ser. No. 73, Div. Water, Air & Waste Chemistry, AmericanChemical Soe., Washington, D.C.

RAMIRIZ-MuNOZ,J. 1968. Atomic Absorption Spectroscopy and Analysisby Atomic Absorption Flame Photometry. American EIsevier Pub-lishing Co., New York, N.Y.

SLAVIN,W. 1968. Atomic Absorption Spectroscopy. John Wiley & Sons,New York, N.Y.

PAUS,P.E. 1971. The application of atomic absorption spectroscopyto the analysis of natural waters. Atomic Absorption Newsletter10:69.

EDIGER,R.D. 1973. A review of water analysis by atomic absorption.Atomic Absorption Newsletter 12:151.

PAUS,P.E. 1973. Determination of some heavy metais in seawater byatomic absorption spectroscopy. Fresenius Zeitschr. Anal. Chem.264:118.

BURRELL,D.C. 1975. Atomic Spectrometric Analysis of Heavy-MetalPollutants in Water. Ann Arbor Science Publishers, Inc., AnnArbor, Mich.

This method is applicable to the determination of antimony,bismuth, cadmium, calcium, cesium, chromium, cobalt, cop-per, gold, iridium, iron, lead, lithium, magnesium, manga-nese, nickel, palladium, platinum, potassium, rhodium, ruthe-nium, silver, sodium, strontium, thallium, tin, and zinco

Atomic absorption spectrometer and associated equipment:See Section 3111A.6. Use bumer head recommended by themanufacturer.

a. Air, cleaned and dried through a suitable filter to removeoil, water, and other foreign substances. The source may be acompressor or commercially bottled gas.

b. Acetylene, standard commercial grade. Acetone, which al-ways is present in acetylene cylinders, can be prevented fromentering and damaging the burner head by replacing a cylinderwhen its pressure has fallen to 689 kPa (100 psi) acetylene.

CAUTION: Acetylene gas represents an explosive hazard in thelaboratory. Follow instrument manufacturer's directions inplumbing and using this gas. Do not allow gas contact withcopper, brass with >65% copper, silver, or liquid mercury; donot use copper or brass tubing, regulators, or fittings.

C. Metal-free water: Use metal-free water for preparing alireagents and calibration standards and as dilution water. Prepare

metal-free water by deionizing tap water anel/or by using one ofthe following processes, depending on the metal concentration inthe sample: single distillation, redistillation, or sub-boiling. Al-ways check deionized or distilled water to determine whether theelement of interest is present in trace amounts. (NOTE:If thesource water contains Hg or other volatile metais, single- orredistilled water may not be suitable for trace analysis becausethese metais distill over with the distilled water. In such cases,use sub-boiling to prepare metal-free water).

d. Calcium solution: Dissolve 630 mg calcium carbonate,CaC03, in 50 mL of 1 + 5 HCI. If necessary, boil gently toobtain complete solution. Cool and dilute to 1000 mL withwater.

e. Hydrochloric acid, HCI, 1%, 10%, 20% (all v/v), 1 + 5, 1+ 1, and cone.f Lanthanum solution: Dissolve 58.65 g lanthanum oxide,

LaZ03, in 250 mL cone HCI. Add acid slowly until the materialis dissolved and dilute to 1000 mL with water.

g. Hydrogen peroxide, 30%.h. Nitric acid, HN03, 2% (v/v), 1 + 1, and cone.i. Aqua regia: Add 3 volumes cone HCI to 1 volume cone

HN03·

j. Standard metal solutions: Prepare a series of standard metalsolutions in the optimum concentration range by appropriatedilution of the following stock metal solutions with water con-taining 1.5 mL cone HNOiL. Stock standard solutions areavailable from a number of commercial suppliers. Altematively,prepare as described below. Thoroughly dry reagents before use.In general, use reagents of the highest purity. For hydrates, usefresh reagents.

1) Antimony: Dissolve 0.2669 g K(SbO)C4H406 in water, add10 mL 1 + 1 HCI and dilute to 1000 mL with water; 1.00 mL =

100 J.Lg Sb.2) Bismuth: Dissolve 0.100 g bismuth metal in a minimum

volume of 1 + 1 HN03. Dilute to 1000 mL with 2% (v/v) HN03;

1.00 mL = 100 J.Lg Bi.3) Cadmium: Dissolve 0.100 g cadmium metal in 4 mL cone

HN03. Add 8.0 mL cone HN03 and dilute to 1000 mL withwater; 1.00 mL = 100 J.Lg Cd.

4) Calcium: Suspend 0.2497 g CaC03 (dried at 180° for 1 hbefore weighing) in water and dissolve cautiously with a mini-mum amount of 1 + 1 HN03. Add 10.0 mL cone HN03 anddilute to 1000 mL with water; 1.00 mL = 100 J.Lg Ca.

5) Cesium: Dissolve 0.1267 g cesium chloride, CsCl, in 1000mL water; 1.00 mL = 100 J.Lg Cs.

6) Chromium: Dissolve 0.1923 g Cr03 in water. When solu-tion is complete, acidify with 10 mL cone HN03 and dilute to1000 mL with water; 1.00 mL = 100 J.Lg Cr.

7) Cobalt: Dissolve 0.1000 g cobalt metal in a minimumamount of 1 + 1 HN03. Add 10.0 mL 1 + 1 HCl and dilute to1000 mL with water; 1.00 mL = 100 J.Lg Co.

8) Copper: Dissolve 0.100 g copper metal in 2 mL coneHN03, add 10.0 mL cone HN03 and dilute to 1000 mL withwater; 1.00 mL = 100 J.Lg Cu.

9) Gold: Dissolve 0.100 g gold metal in a minimum volumeof aqua regia. Evaporate to dryness, dissolve residue in 5 mLcone HCl, cool, and dilute to 1000 mL with water; 1.00 mL =

100 J.Lg Au.

10) Iridium: Dissolve 0.1147 g ammonium chloroiridate,(NH4)zIrCI6, in a minimum volume of 1% (v/v) HCl and diluteto 100 mL with 1% (v/v) HCI; 1.00 mL = 500 J.Lg Ir.

11) Iron: Dissolve 0.100 g iron wire in a mixture of 10 mL 1+ 1 HCl and 3 mL cone HN03. Add 5 mL cone HN03 and diluteto 1000 mL with water; 1.00 mL = 100 J.Lg Fe.

12) Lead: Dissolve 0.1598 g lead nitrate, Pb(N03h, in aminimum amount of 1 + 1 HN03, add 10 mL cone HN03, anddilute to 1000 mL with water; 1.00 mL = 100 J.Lg Pb.

13) Lithium: Dissolve 0.5323 g lithium carbonate, LizC03, ina minimum volume of 1 + 1 HN03. Add 10.0 mL cone HN03and dilute to 1000 mL with water; 1.00 mL = 100 J.Lg Li.

14) Magnesium: Dissolve 0.1658 g MgO in a minimumamount of 1 + 1 HN03. Add 10.0 mL cone HN03 and dilute to1000 mL with water; 1.00 mL = 100 J.Lg Mg.

15) Manganese: Dissolve 0.1000 g manganese metal in 10mL cone HCl mixed with 1 mL cone HN03. Dilute to 1000 mLwith water; 1.00 mL = 100 J.Lg Mn.

16) Nickel: Dissolve 0.1000 g nickel metal in 10 mL hot coneHN03, cool, and dilute to 1000 mL with water; 1.00 mL = 100J.Lg Ni.

17) Palladium: Dissolve 0.100 g palladium wire in a mini-mum volume of aqua regia and evaporate just to dryness. Add 5mL cone HCl and 25 mL water and warm until dissolution iscomplete. Dilute to 1000 mL with water; 1.00 mL = 100 J.Lg Pd.

18) Platinum: Dissolve 0.100 g platinum metal in a minimumvolume of aqua regia and evaporate just to dryness. Add 5 mLcone HCl and 0.1 g NaCl and again evaporate just to dryness.Dissolve residue in 20 mL of 1 + 1 HCl and dilute to 1000 mLwith water; 1.00 mL = 100 J.Lg Pt.

19) Potassium: Dissolve 0.1907 g potassium chloride, KCI,(dried at 110°C) in water and make up to 1000 mL; 1.00 mL =100 J.Lg K.

20) Rhodium: Dissolve 0.386 g ammonium hexachlororho-date, (NH4)3RhCI6' 1.5HzO, in a minimum volume of 10%(v/v) HCI and dilute to 1000 mL with 10% (v/v) HCl; 1.00 mL= 100 J.Lg Rh.

21) Ruthenium: Dissolve 0.205 g ruthenium chloride, RuCI3,in a minimum volume of 20% (v/v) HCI and dilute to 1000 mLwith 20% (v/v) HCI; 1.00 mL = 100 J.Lg Ru.

22) Silver: Dissolve 0.1575 g silver nitrate, AgN03, in 100mL water, add 10 mL cone HN03, and make up to 1000 mL;1.00 mL = 100 J.Lg Ag.

23) Sodium: Dissolve 0.2542 g sodium chloride, NaCl, driedat 140°C, in water, add 10 mL cone HN03 and make up to 1000mL; 1.00 mL = 100 J.Lg Na.

24) Strontium: Suspend 0.1685 g SrC03 in water and dissolvecautiously with a minimum amount of 1 + 1 HN03. Add 10.0mL cone HN03 and dilute to 1000 mL with water: 1 mL = 100J.Lg Sr.

25) Thallium: Dissolve 0.1303 g thallium nitrate, TIN03, inwater. Add 10 mL cone HN03 and dilute to 1000 mL with water;1.00 mL = 100 J.Lg TI.

26) Tin: Dissolve 1.000 g tin metal in 100 mL cone HCI anddilute to 1000 mL with water; 1.00 mL = 1.00 mg Sn.

27) Zinc: Dissolve 0.100 g zinc metal in 20 mL 1 + 1 HCland dilute to 1000 mL with water; 1.00 mL = 100 J.Lg Zn.

a. Sample preparation: Required sample preparation dependson the metal form being measured.

If dissolved metais are to be determined, see Section 3030Bfor sample preparation. lf total or acid-extractable metais are tobe determined, see Sections 3030C through K. For ali samples,make certain that the concentrations of acid and matrix modifiersare the same in both samples and standards.

When determining Ca or Mg, dilute and mix 100 mL sampleor standard with 10 mL lanthanum solution (<j[ 31) before aspi-rating. When determining Fe or Mn, mix 100 mL with 25 mL ofCa solution (<j[ 3d) before aspirating. When determining Cr, mixI mL 30% H202 with each 100 mL before aspirating. Alterna-tively use proportionally smaller volumes.

b. Instrument operation: Because of differences betweenmakes and models of atomic absorption spectrometers, it is notpossible to formulate instructions applicable to every instrumentoSee manufacturer's operating manual. ln general, proceed ac-cording to the following: lnstall a hollow-cathode lamp for thedesired metal in the instrument and roughly set the wavelengthdial according to Table 3111:1. Set slit width according tomanufacturer's suggested setting for the element being mea-sured. Tum on instrument, apply to the hollow-cathode lamp thecurrent suggested by the manufacturer, and let instrument warmup until energy source stabilizes, generally about 10 to 20 minoReadjust current as necessary after warmup. Optimize wave-length by adjusting wavelength dial until optimum energy gain isobtained. Align lamp in accordance with manufacturer's instruc-tions.

lnstall suitable burner head and adjust burner head position.Turn on air and adjust flow rate to that specified by manufacturerto give maximum sensitivity for the metal being measured. Turnon acetylene, adjust flow rate to value specified, and ignite flame.Let flame stabilize for a few minutes. Aspirate a blank consistingof deionized water containing the same concentration of acid instandards and samples. Zero the instrumento Aspirate a standardsolution and adjust aspiration rate of the nebulizer to obtainmaximum sensitivity. Adjust burner both vertically and horizon-tally to obtain maximum response. Aspirate blank again andrezero the instrumento Aspirate a standard near the middle of the

linear range. Record absorbance of this standard when freshlyprepared and with a new hollow-cathode lamp. Refer to thesedata on subsequent determinations of the same element to checkconsistency of instrument setup and aging of hollow-cathodelamp and standard.

The instrument now is ready to operate. When analyses arefinished, extinguish flame by turning offfirst acetylene and then air.

c. Standardization: Select at least three concentrations of eachstandard metal solution (prepared as in <j[ 3j above) to bracket theexpected metal concentration of a sample. Aspirate blank andzero the instrumento Then aspirate each standard in turn intoflame and record absorbance.

Prepare a calibration curve by plotting on linear graph paperabsorbance of standards versus their concentrations. For instru-ments equipped with direct concentration readout, this step isunnecessary. With some instruments it may be necessary toconvert percent absorption to absorbance by using a table gen-erally provided by the manufacturer. Plot calibration curves forCa and Mg based on original concentration of standards beforedilution with lanthanum solution. Plot calibration curves for Feand Mn based on original concentration of standards beforedilution with Ca solution. Plot calibration curve for Cr based onoriginal concentration of standard before addition of H202.

d. Analysis of samples: Rinse nebulizer by aspirating watercontaining 1.5 mL conc HNOiL. Aspirate blank and zero in-strument. Aspirate sample and determine its absorbance.

Calculate concentration of each metal ion, in micrograms perliter for trace elements, and in milligrams per liter for morecommon metais, by referring to the appropriate calibration curveprepared according to <j[ 4c. Alternatively, read concentrationdirectly from the instrument readout if the instrument is soequipped. lf the sample has been diluted, multiply by the appro-priate dilution factor.

WILLlS, J.B. 1962. Determination of lead and other heavy metaIs in urineby atomic absorption spectrophotometry. Ana!. Chem. 34:614.

Also see Section 3111A.8 and 9.

This method is suitable for the determination of low concen-trations of cadmium, chromium, cobalt, copper, iron, lead, man-ganese, nickel, silver, and zincoThe method consists of chelationwith ammonium pyrrolidine dithiocarbamate (APDC) and ex-traction into methyl isobutyl ketone (MIBK), followed by aspi-ration into an air-acetylene flame.

a. Atomic absorption spectrometer and associated equip-ment: See Section 3111A.6.

b. Bumer head, conventional. Consult manufacturer's oper-ating manual for suggested burner head.

a. Air: See 3111B.3a.b. Acetylene: See 31 [IB.3b.c. Metal-free water: See 3111B.3c.d. Methyl isobutyl ketone (MIBK), reagent grade. For trace anal-

ysis, purify MlBK by redistillation or by sub-boiling distillation.e. Ammonium pyrrolidine dithiocarbamate (APDC) solution:

Dissolve 4 g APDC in 100 mL water. If necessary, purify APDC

with an equal volume of MIBK. Shak:e30 s in a separatory funnel,let separate, and withdraw lower portion. Discard MIBK layer.f Nitric acid, HN03, cone, ultrapure.g. Standard metal solutions: See 3111B.3j.h. Potassium permanganate solution, KMn04, 5% (w/v)

aqueous.i. Sodium sulfate, Na2S04, anhydrous.j. Water-saturated MIBK: Mix one part purified MIBK with

one part water in a separatory funnel. Shak:e30 s and let separate.Discard aqueous layer. Save MIBK layer.

k. Hydroxylamine hydrochloride solution, 10% (w/v). Thissolution can be purchased commercially.

a. Instrument operation: See Section 3111BAb. After finaladjusting of bumer position, aspirate water-saturated MIBK intoflame and gradually reduce fuel flow until flame is similar to thatbefore aspiration of solvent.

b. Standardization: Select at least three concentrations ofstandard metal solutions (prepared as in 3111B.3j) to bracketexpected sample metal concentration and to be, after extraction,in the optimum concentration range of the instrumento Adjust100 mL of each standard and 100 mL of a metal-free water blankto pH 3 by adding lN HN03 or lN NaOH. For individualelement extraction, use the following pH ranges to obtain opti-mum extraction efficiency:

pH Range for OptimumExtraction

2-5 (complexunstable)1-6

2-103-9

0.1-82-5

2-4 (complexunstable)2-4

0.1-62-6

NOTE:For Ag and Pb extraction the optimum pH value is 2.3::!:: 0.2. The Mn complex deteriorates rapidly at roomtemperature, resulting in decreased instrument response.Chilling the extract to O°C may preserve the complex for afew hours. If this is not possible and Mn cannot be analyzedimmediately after extraction, use another analytical procedure.

Transfer each standard solution and blank to individual200-mL volumetric flasks, add 1 mL APDC solution, and shaketo mix. Add 10 mL MIBK and shak:e vigorously for 30 s. (Themaximum volume ratio of sample to MIBK is 40.) Let contentsof each flask separate into aqueous and organic layers, thencarefully add water (adjusted to the same pH at which theextraction was carrled out) down the side of each flask to bringthe organic layer into the neck and accessible to the aspiratingtube.

Aspirate organic extracts directly into the flame (zeroing in-strument on a water-saturated MIBK blank) and record absor-bance.

Prepare a calibration curve by plotting on linear graph paperabsorbances of extracted standards against their concentrationsbefore extraction.

c. Analysis of samples: Prepare samples in the same manneras the standards. Rinse atomizer by aspirating water-saturatedMIBK. Aspirate organic extracts treated as above directly intothe flame and record absorbances.

With the above extraction procedure only hexavalent chro-mium is measured. To determine total chromium, oxidizetrivalent chromium to hexavalent chromium by bringing sam-pIe to a boil and adding sufficient KMn04 solution dropwiseto give a persistent pink color while the solution is boiled for10 mino Destroy excess KMn04 by adding 1 to 2 dropshydroxylamine hydrochloride solution to the boiling solution,allowing 2 min for the reaction to proceed. If pink colorpersists, add 1 to 2 more drops hydroxylamine hydrochloridesolution and wait 2 mino Heat an additional 5 mino Cool,extract with MIBK, and aspirate.

During extraction, if an emulsion forms at the water-MIBKinterface, add anhydrous Na2S04 to obtain a homogeneous or-ganic phase. In that case, also add Na2S04 to alI standards andblanks.

To avoid problems associated with instability of extractedmetal complexes, determine metaIs immediately after extraction.

Calculate the concentration of each metal ion in microgramsper liter by referring to the appropriate calibration curve.

ALLAN, I.E. 1961. The use of organic solvents in atomic absorptionspectrophotometry.Speetroehim. Aeta 17:467.

SACHDEV,S.L. & P.W. WEST. 1970. Concentrationof trace metais bysolvent extraction and their determinationby atomic absorptionspectrophotometry.Environ. Sei. Teehnol. 4:749.

This method is applicable to the determination of aluminum,barium, beryllium, calcium, molybdenum, osmium, rhenium,silicon, thorium, titanium, and vanadium.

a. Atomic absorption spectrometer and associated equip-ment: See Section 3111A.6.

b. Nitrous oxide burner head: Use special bumer head as

suggested in manufacturer's manual. At roughly 20-min inter-vaIs of operation it may be necessary to dislodge the carbon crustthat forms along the slit surface with a carbon rod or appropriatealtemative.

c. T-junction valve or other switching valve for rapidly chang-ing from nitrous oxide to air, so that flame can be tumed on oroff with air as oxidant to prevent flashbacks.

a. Air: See 3111B.3a.b. Acetylene: See 3111B.3b.c. Metal-free water: See 3111B.3c.d. Hydrochloric acid, HCl, IN, 1+ 1, and cone.e. Nitric acid, HN03, cone.f Sulfuric acid, HZS04, 1% (v/v).g. Hydrofiuoric acid, HF, lN.h. Nitrous oxide, commercially available cylinders. Fit nitrous

oxide cylinder with a special nonfreezable regulator or wrap aheating coil around an ordinary regulator to prevent flashback atthe bumer caused by reduction in nitrous oxide flow through afrozen regulator. (Most modem atomic absorption instrumentshave automatic gas control systems that will shut down a nitrousoxide-acetylene flame safely in the event of a reduction innitrous oxide flow rate.)

CAUTION:Use nitrous oxide with strict adherence to manufac-turer's directions. Improper sequencing of gas fiows at startupand shutdown of instrument can produce explosions from fiash-back.

i. Potassium chloride solution: Dissolve 250 g KCl in waterand dilute to 1000 roL.

j. Aluminum nitrate solution: Dissolve 139 g Al(N03)3 . 9HzOin 150 mL water. Acidify slightly with cone HN03 to precludepossible hydrolysis and precipitation. Warm to dissolve completely.Coal and dilute to 200 roL.

k. Standard metal solutions: Prepare a series of standardmetal solutions in the optimum concentration ranges by appro-priate dilution of stock metal solutions with water containing 1.5roL cone HNOiL. Stock standard solutions are available from anumber of commercial suppliers. Altematively, prepare as de-scribed below.

1) Aluminum: Dissolve 0.100 g aluminum metal in an acidmixture of 4 roL 1 + 1 HCl and 1 roL cone HN03 in a beaker.Warm gently to effect solution. Transfer to a l-L flask, add 10roL 1 + 1 HCl, and dilute to 1000 roL with water; 1.00 roL =100 J.Lg AI.

2) Barium: Dissolve 0.1516 g BaClz (dried at 2500 for 2 h), inabout 10 roL water with 1 roL 1+ 1 HCI. Add 10.0 roL 1 + 1HCI and dilute to 1000 roL with water; 1.00 roL = 100 J.Lg Ba.

3) Beryllium: Do not dry. Dissolve 1.966 g BeS04 . 4HzO inwater, add 10.0 roL cone HN03, and dilute to 1000 roL withwater; 1.00 roL = 100 J.Lg Be.

4) Calcium: See 311IB.3j4).5) Molybdenum: Dissolve 0.2043 g (NH4h Mo04 in water

and dilute to 1000 roL; 1.00 roL = 100 J.Lg Mo.6) Osmium: Obtain standard O.lM osmium tetroxide solu-

tion* and store in glass bottle; 1.00 roL = 19.02 mg Os. Make

dilutions daily as needed using 1% (v/v) HZS04. CAUTlON:OS04is extremely toxic and highly volatile.

7) Rhenium: Dissolve 0.1554 g potassium perrhenate,KRe04, in 200 roL water. Dilute to 1000 roL with 1% (v/v)HZS04; 1.00 roL = 100 J.Lg Re.

8) Silica: Do not dry. Dissolve OA730 g NazSi03 . 9HzO inwater. Add 10.0 roL cone HN03 and dilute to 1000 roL withwater. 1.00 roL = 100 J.Lg SiOz' Store in polyethylene.

9) Thorium: Dissolve 0.238 g thorium nitrate, Th(N03)4 . 4HzOin 1000 roL water; 1.00 roL = 100 J.Lg Th.

10) Titanium: Dissolve 0.3960 g pure (99.8 or 99.9%) tita-nium chloride, TiCI4,t in a mixture of equal volumes of 1N HCIand 1N HF. Make up to 1000 roL with this acid mixture; 1.00 roL= 100 J.Lg Ti.

11) Vanadium: Dissolve 0.2297 g ammonium metavanadate,NH4V03, in a minimum amount of cone HN03. Heat to dissolve.Add 10 roL cone HN03, and dilute to 1000 mL with water; 1.00roL = 100 J.Lg V.

a. Sample preparation: See Section 3111BAa.When deterrnining AI, Ba, or Ti, mix 2 roL KCl solution into

100 roL sample and standards before aspiration. When determin-ing Mo and V, mix 2 roL AI(N03)3 . 9Hz°into 100 roL sampleand standards before aspiration.

b. Instrument operation: See Section 3111BAb. After adjustingwavelength, install a nitrous oxide burner head. Turn on acetylene(without igniting flame) and adjust flow rate to value specified bymanufacturer for a nitrous oxide-acetylene flame. Turn off acety-Iene. With both air and nitrous oxide supplies tumed on, set T-junction valve to nitrous oxide and adjust flow rate according tomanufacturer's specifications. Turn switching valve to the air posi-tion and verify that flow rate is the same. Turn acetylene on andignite to a bright yellow flame. With a rapid motion, tum switchingvalve to nitrous oxide. The flame should have a red cone above theburner. If it does not, adjust fuel flow to obtain red cone. Afternitrous oxide flame has been ignited, let bumer come to thermalequilibrium before beginning analysis.

Aspirate a blank consisting of deionized water containing 1.5mL cone HN03/L and check aspiration rate. Adjust if necessaryto a rate between 3 and 5 roL/ min. Zero the instrumento Aspiratea standard of the desired metal with a concentration near themidpoint of the optimum concentration range and adjust bumer(both horizontally and vertically) in the light path to obtainmaximum response. Aspirate blank again and re-zero the instru-ment. The instrument now is ready to run standards and samples.

To extinguish flame, tum switching valve from nitrous oxide toair and tum off acetylene. This procedure eliminates the danger offlashback that may occur on direct ignition or shutdown of nitrousoxide and acetylene. (See also discussion in 3111BAb.)

C. Standardization: Select at least three concentrations ofstandard metal solutions (prepared as in 'J[ 3k) to bracket theexpected metal concentration of a sample. Aspirate each in tuminto the flame and record absorbances.

Most modem instruments are equipped with microprocessorsand digital readout which perrnit calibration in direct concentra-

t Alpha Ventron, P.O. Box 299, 152 Andover St., Danvers, MA 01923, orequivalent.

tion terms. If instrument is not so equipped, prepare a calibrationcurve by plotting on linear graph paper absorbance of standardsversus concentration. Plot calibration curves for AI, Ba, and Tibased on original concentration of standard before adding KCIsolution. Plot calibration curves for Mo and V based on originalconcentration of standard before adding AI(N03)3 solution.

d. Analysis of samples: Rinse atomizer by aspirating watercontaining 1.5 rnL conc HNOiL and zero instrumento Aspiratea sample and determine its absorbance.

Calculate concentration of each metal ion in micrograms perliter by referring to the appropriate calibration curve preparedaccording to '11 4c.

Alternatively, read the concentration directly from the instru-ment readout if the instrument is so equipped. If sample has beendiluted, multiply by the appropriate dilution factor.

WLLLlS, I.B. 1965. Nitrous oxide-acetylene f1ame in atomic absorptionspectroscopy. Nature 207:715.

AIso see Section 3111A.8 and 9.

a. Application: This method is suitable for the determinationof aluminum at concentrations less than 900 /Lg/Land berylliumat concentrations less than 30 /Lg/L. The method consists ofchelation with 8-hydroxyquinoline, extraction with methylisobutyl ketone (MIBK), and aspiration into a nitrous oxide-acetylene f1ame.

b. lnterferences: Concentrations of Fe greater than 10 mgILinterfere by suppressing AI absorption. Iron interference can bemasked by addition of hydroxylarnine hydrochloride/I,lO-phenan-throline. Mn concentrations up to 80 mgIL do not interfere ifturbidity in the extract is allowed to settle. Mg forrns an insolublechelate with 8-hydroxyquinoline at pH 8.0 and tends to remove AIcomplex as a coprecipitate. However, the Mg complex forrnsslowly over 4 to 6 min; its interference can be avoided if the solutionis extracted immediately after adding buffer.

Atomic absorption spectrometer and associated equipment:See Section 311IA.6.

a. Air: See 3IIIB.3a.b. Acetylene: See 3IIIB.3b.c. Ammonium hydroxide, NH40H, conc.d. Buffer: Dissolve 300 g ammonium acetate, NH4C2H30z, in

water, add 105 rnL conc NH40H, and dilute to I L.e. Metal-free water: See 3111B.3c.f Hydrochloric acid, HCI, conc.g. 8-Hydroxyquinoline solufion: Dissolve 20 g 8-hy-

droxyquinoline in about 200 mL water, add 60 rnL glacial aceticacid, and dilute to 1 L with water.

h. Methyl isobutyl ketone: See 3111C.3d.i. Nitric acid, HN03, conc.

j. Nitrous oxide: See 311ID.3h.k. Standard metal solutions: Prepare a series of standard metal

solutions containing 5 to 1000 /LgIL by appropriate dilution of thestock metal solutions prepared according to 311ID.3k.

l. lron masking solution: Dissolve 1.3 g hydroxylamine hy-drochloride and 6.58 g 1,1O-phenanthroline monohydrate inabout 500 rnL water and dilute to 1 L with water.

a. lnstrument operation: See Sections 3111BAb, CAa, andDAb. After final adjusting of burner position, aspirate MIBKinto f1ameand gradually reduce fuel f10wuntil f1ameis similar tothat before aspiration of solvent. Adjust wavelength setting ac-cording to Table 3111:1.

b. Standardization: Select at least three concentrations of stand-ard metal solutions (prepared as in '113k) to bracket the expectedmetal concentration of a sample and transfer 100 rnL of each (and100 rnL water blank) to four different 2oo-rnL volumetric f1asks.Add 2 rnL 8-hydroxyquinoline solution, 2 rnL masking solution (ifrequired), and 10 rnL buffer to one f1ask,immediately add 10 rnLMIBK, and shake vigorously. The duration of shaking affects theforrns of aluminum complexed. A fast, lO-s shaking time favorsmonomeric AI, whereas 5 to 10 min of shaking also will complexpolymeric species. Adjustment of the 8-hydroxyquinoline to sampleratio can improve recoveries of extremely high or low concentra-tions of aluminum. Treat each blank, standard, and sample insimilar fashion. Continue as in Section 3111C.4b.

c. Analysis of samples: Rinse atomizer by aspirating water-saturated MIBK. Aspirate extracts of samples treated as above,and record absorbances.

Calculate concentration of each metal in micrograms per liter byreferring to the appropriate calibration curve prepared according to'114b.

For general introductory material on atomic absorption spec-trometric methods, see Section 3111A.

When possible, dedicate glassware for use in Hg analysis.Avoid using glassware previously exposed to high levels of Hg,such as those used in COD, TKN, or Cl- analysis.

a. Atomic absorption spectrometer and associated equipment:See Section 3111A.6. Instruments and accessories specificallydesigned for measurement of mercury by the cold vapor tech-nique are available commercially and may be substituted.

b. Absorption ceU,a glass or plastic tube approximately 2.5 cm indiarneter. Ao 11.4-cm-Iong tube has been found satisfactory but a15-cm-Iong tube is preferred. Grind tube ends perpendicular to thelongitudinal axis and cement quartz windows in place. Attach gasinlet and outlet ports (6.4 mm diarn) 1.3 cm from each end.

c. CeU support: Strap cell to the flat nitrous-oxide bumer heador other suitable support and align in light beam to give maxi-mum transmittance.

d. Air pumps: Use any peristaltic pump with electronic speedcontrol capable of delivering 2 L air/min. Any other regulatedcompressed air system ar air cylinder also is satisfactory.

e. Flowmeter, capable of measuring an air flow of 2 L/min.I Aeration tubing, a straight glass frit having a coarse porosity

for use in reaction flask.g. Reaction flask, 250-rnL erlenmeyer flask or a BOD bottle,

fitted with a rubber stopper to hold aeration tube.h. Drying tube, 150-mm X 18-mm-diam, containing 20 g Mg

(CI04)z. A 60-W light bulb with a suitable shade may be sub-stituted to prevent condensation of moisture inside the absorp-tion cell. Position bulb to maintain cell temperature at 10°Cabove ambiento

i. Connecting tubing, glass tubing to pass mercury vapor fromreaction flask to absorption cell and to interconnect ali other com-ponents. Clear vinyl plastic* tubing may be substituted for glass.

• Tygon or equivalent.t Use specially prepared reagents low in mercury.

b. Stock mercury solution: Dissolve 0.1354 g mercuric chlo-ride, HgClz, in about 70 rnL water, add I rnL conc HN03, anddilute to 100 rnL with water; 1.00 rnL = 1.00 mg Hg.

c. Standard mercury solutions: Prepare a series of standardmercury solutions containing O to 5 J.LglL by appropriate dilutionof stock mercury solution with water containing 10 rnL concHN031L. Prepare standards daily.

d. Nitric acid, HN03, conc.e. Potassium permanganate solution: Dissolve 50 g KMn04 in

water and dilute to 1 L.I Potassium persulfate solution: Dissolve 50 g KZSZOg in

water and dilute to I L.g. Sodium chloride-hydroxylamine sulfate solution: Dissolve

120 g NaCI and 120 g (NHzOH)2 . H2S04 in water and dilute toI L. A 10% hydroxylamine hydrochloride solution may besubstituted for the hydroxylamine sulfate.

h. Stannous ion (Snz+) solution: Use either stannous chloride,<j[ I), or stannous sulfate, <j[ 2), to prepare this solution containingabout 7.0 g Sn2+/100 rnL.

1) Dissolve 10 g SnClz in water containing 20 rnL conc HCIand dilute to 100 rnL.

2) Dissolve 11 g SnS04 in water containing 7 rnL conc HZS04

and dilute to 100 rnL.Both solutions decompose with aging. If a suspension forros,

stir reagent continuously during use. Reagent volume is suffi-cient to process about 20 samples; adjust volumes prepared toaccommodate number of samples processed.

i. Sulfuric acid, H2S04, conc.

a. Instrument operation: See Section 311lB.4b. Set wavelength to253.7 um. Install absorptioncell and align in light path to give maxi-mum transmission.Connect associated equipment to absorption cellwith glassor vinylplastictubing as indicatedin Figure3112:1.Tum onair and adjust flow rate to 2 Umin. Allow air to flow continuously.Altematively, follow manufacturer's directions for operation. NOTE:Auorescent lighting may increasebaseline noise.

b. Standardization: Transfer 100 rnL of each of the 1.0, 2.0,and 5.0 J.LglL Hg standard solutions and a blank of 100 rnL waterto 250-rnL erlenmeyer reaction flasks. Add 5 rnL conc HZS04

and 2.5 roL conc HN03 to each flask. Add 15 rnL KMn04

solution to each flask and let stand at least 15 min. Add 8 rnL

Absorption cellwith quartz windows

+

Hollow cathodemercury lampVent to hood

with tra p for Hg

\Drying tube filled

with Mg(CIO.h; Rotameteri - 2 L air/min

Compressed air2 L air/min

Figure 3112:1. Schematic arrangement of equipment for measurementof mercury by cold-vapor atomic absorption technique.

K2S20g solution to each ftask and heat for 2 h in a water bath at95°C. Cool to room temperature.

Treating each flask individually, add enough NaCl-hydroxyl-amine solution to reduce excess KMn04, then add 5 rnL SnCI2

or SnS04 solution and immediately attach flask to aerationapparatus. As Hg is volatilized and carried into the absorptioncell, absorbance will increase to a maximum within a few sec-onds. As soon as recorder returns approximately to the base line,remove stopper holding the frit from reaction flask, and replacewith a ftask containing water. Flush system for a few secondsand run the next standard in the same manner. Construct astandard curve by plotting peak height versus micrograms Hg.

c. Analysis af samples: Transfer 100 rnL sample or portiondiluted to 100 rnL containing not more than 5.0 JLg Hg/L to areaction flask. Treat as in <j[ 4b. Seawaters, brines, and efftuentshigh in chlorides require as much as an additional 25 rnL KMn04

solution. During oxidation step, chlorides are converted to freechlorine, which absorbs at 253 nm. Remove alI free chlorinebefore the Hg is reduced and swept into the cell by using anexcess (25 rnL) of hydroxylamine reagent.

Remove free chlorine by sparging sample gently with air ornitrogen after adding hydroxylamine reducing solution. Use aseparate tube and frit to avoid carryover of residual stannouschloride, which could cause reduction and loss of mercury.

TABLE 3112:1. INTERLABORATORYPRECISIONANDBIAS OF COlO-VAPORATOMICABSORPTIONSPECTROMETRICMETHODFORMERCURyI

Relative RelativeConc. SD SD EITor No.of

Form fLglL fLglL % % Participants

Inorganic 0.34 0.077 22.6 21.0 23Inorganic 4.2 0.56 13.3 14.4 21Organic 4.2 0.36 8.6 8.4 21

Determine peak height of sample from recorder chart and readmercury value from standard curve prepared according to 1: 4b.

Data on interlaboratory precision and bias for this method aregiven in Table 3112:1.

1. Kopp, J.F., M.C. LONGBOTTOM& L.B. LOBRlNG. 1972. "Cold vapor"method for determining mercury. J. Amer. Water Works Assoe.64:20.

HATCH,W.R. & W.L. OTT. 1968. Determination of submicrogram quan-tities of mercury by atomic absorption spectrophotometry. Ana!.Chem. 40:2085.

UTHE, J.F., F.AJ. ARMSTRONG& M.P. STAINTON.1970. Mercury deter-mination in fish samples by wet digestion and f1ameless atomicabsorption spectrophotometry. J. Fish. Res. Board Cano 27:805.

FElOMAN, C. 1974. Preservation of dilute mercury solutions. Ana!.Chem. 46:99.

BOTHNER,M.H. & D.E. ROBERTSON.)975. Mercury contamination of seawater samples stored in polyethylene containers. Ana!. Chem. 47:592.

HAWLEY,J.E. & J.D. 1NGLE,JR. 1975. 1mprovements in cold vapor atomicabsorption determination of mercury. Ana!. Chem. 47:719.

Lo, J.M. & C.M. WAL. 1975. Mercury loss from water during storage:Mechanisms and prevention. Ana!. Chem. 47:1869.

EL-AwADY, A.A., R.B. MILLER& MJ. CARTER.1976. Automated methodfor the determination of total and inorganic mercury in water andwastewater samples. Anal. Chem. 48: 110.

ODA, C.E. & J.D. INGLE,JR. 1981. Speciation of mercury by cold vaporatomic absorption spectrometry with selective reduction. Ana!.Chem. 53:2305.

SUDDENDORF,R.F. 1981. Interference by selenium or tellurium in thedetermination of mercury by cold vapor generation atomic absorp-tion spectrometry. Ana!. Chem. 53:2234.

HEIDEN, R.W. & D.A. AlKENS. 1983. Humic acid as a preservative fortrace mercury (lI) solutions stored in polyolefin containers. Ana!.Chem. 55:2327.

CHOU, H.N. & C.A. NALEWAY.1984. Determination of mercury by coldvapor atomic absorption spectrometry. Ana!. Chem. 56: 1737.

Electrothermal atomic absarption permits determination ofmost metallic elements with sensitivities and detection limitsfrom 20 to 1000 times better than those of conventional flametechniques without extraction or sample concentration. Thisincrease in sensitivity results from an increase in atom densitywithin the furnace as compared to flame atomic absorption.Many elements can be determined at concentrations as low as1.0 J-Lg/L.An additional advantage of electrothermal atomicabsorption is that only a very small volurne of sample isrequired.

The electrothermal technique is used only at concentrationlevels below the optimum range of direct flame atomic absorp-tion because it is subject to more interferences than the f1ameprocedure and requires increased analysis time. The method ofstandard additions may be required to insure validity of data.Because of the high sensitivity of this technique, it is extremelysusceptible to contarnination; extra care in sample handling andanalysis may be required.

Electrothermal atomic absorption spectroscopy is based on thesame principie as direct f1ame atomization but an electricallyheated atomizer or graphite fumace replaces the standard bumerhead. A discrete sample volume is dispensed into the graphitesample tube (or cup). Typically, determinations are made byheating the sample in three or more stages. First, a low currentheats the tube to dry the sample. The second, ar charring, stagedestroys organic matter and volatilizes other matrix componentsat an intermediate temperature. Finally, a high current heats thetube to incandescence and, in an inert atmosphere, atomizes theelement being determined. Additional stages frequently areadded to aid in drying and charring, and to clean and cool thetube between samples. The resultant ground-state atornic vaporabsorbs monochromatic radiation fram the source. A photoelec-tric detector measures the intensity of transrnitted radiation. Theinverse of the transmittance is related logarithmically to theabsorbance, which is directly proportional to the number densityof vaporized ground-state atoms (the Beer-Lambert law) over alirnited concentration range.

Electrothermal atomization determinations may be subject tosignificant interferences fram molecular absorption as well aschernical and matrix effects. Molecular absorption may occurwhen components of the sample matrix volatilize during atom-

* Approved by Standard Methods Comminee, 1999.Joint Task Group: 20th Edition-Raymond J. Lovett (chair), David J. Kaptain.

ization, resulting in broadband absorption. Several backgroundcorrection techniques are available commercially to compensatefor this interference. A continuum source such as a deuterium arccan correct for background up to absorbance levels of about 0.8.Continuum lamp intensity diminishes at long wavelengths anduse of continuum background correction is limited to analyticalwavelengths below 350 nm. Zeeman effect background correc-tars can handle background absarbance up to 1.5 to 2.0. TheSmith-Hieftje correction technique can accommodate back-ground absorbance levels as large as 2.5 to 3.0 (see Section3111A.3). Both Zeeman and Smith-Hieftje background correc-tions are susceptible to rollover (development of a negativeabsorbance-concentration relationship) at high absorbances. Therollover absorbance for each element should be available in themanufacturer' s literature. Curvature due to rollover should be-come apparent during ca1ibration; dilution produces a morelinear calibration plot. Use background correction when analyz-ing samples containing high concentrations of acid or dissolvedsolids and in determining elements for which an absorption linebelow 350 nm is used.

Matrix modification can be useful in rninimizing interferenceand increasing analytical sensitivity. Determine need for a mod-ifier by evaluating recovery of a sample with a known addition.Recovery near 100% indicates that sample matrix does not affectanalysis. Chemical modifiers generally modify relative volatili-ties of matrix and metal. Some modifiers enhance matrix re-moval, isolating the metal, while other modifiers inhibit metalvolatilization, allowing use of higher ashing/charring tempera-tures and increasing efficiency of matrix removal. Chernicalmodifiers are added at high concentration (percent levei) and canlead to sample contamination from impurities in the modifiersolution. Heavy use of chernical modifiers may reduce the usefullife (normally 50 to 100 firings) of the graphite tube. Somespecific chemical modifiers and appraximate concentrations areIisted in Table 3113:1.

Addition of a chemical modifier directly to the sample beforeanalysis is restricted to inexpensive additives (e.g. phosphoricacid). Use of palladium salts for matrix modification normallyrequires methods of co-addition, in which sample and modifierare added consecutively to the fumace either manually or, pref-erab1y, with an automatic sampler. Palladium salts (nitrate ispreferred, chloride is acceptable) are listed in Table 3113:I as amodifier for many metais. The palladium solution (50 to 2000mglL) generally includes citric ar ascarbic acid, which aidsreduction of palladium in the furnace. Citric acid levels of 1 to2% are typical. Use of hydrogen (5%) in the coo1ant gas (avail-able commercially as a mixture) also reduces palladium, elimi-nating need for organic reducing acids. CAUTlON:Do not miJehydrogen and other gases in the laboratory; hydrogen gas isvery flammable-handle with caution. Use low levels of palla-dium (50 to 250 mglL) for normal samp1es and higher leve1s forcomplex samples. Addition of excess palladium modifier maywiden atomization peaks; in such cases peak area measurements

TABLE3113:1. POTENTIALMATRlXMODIFIERSFORELECfROTHERMALATOMICABSORYfIONSPECfROMETRY*

Analyses for WhichModifier

May Be Useful

1500 mg PcIIL + 1000 mgMg(N03h/L1

500-2000 mg PcIIL + reducingagent2t

5000 mg Mg(N03)2/L I

100-500 mg PcIIL2

50 mg Ni/L22% pot + 1000 mg Mg (N03)2/L1

Ag, As, Au, Bi, Cu, Ge, Mn,Hg, In, Sb, Se, Sn, Te, TI

Ag, As, Bi, Cd, Co, Cr, Cu, Fe,Hg, Mn, Ni, Pb, Sb

Be, Co, Cr, Fe, Mn, VAs, Ga, Ge, SnAs, Se, SbCd, Pb

*Assumes 10 ,."L modifier/IO ,."L sample.tCitrie aeid (1-2%) preferred; aseorbie aeid or H2 aeeeptable.

may provide higher quality results. The recommended mo de ofmodifier use is through co-addition to the fumace of about 10 pLof the palladium (or other) modifier solution. Palladium may notbe the best modifier in all cases and cannot be recommendedunconditionally. Test samples requiring a modifier first withpalladium; test other modifiers only if palladium is unsuccessfulor to minimize modifier cost. See Section 3ll3B.3 for prepara-tion of modifier solution.

Temperature ramping, i.e., gradual heating, can be used todecrease background interferences and permits analysis of sam-pIes with complex matrices. Ramping permits a controlled, con-tinuous increase of furnace temperature in any of the varioussteps of the temperature sequence. Ramp drying is used forsamples containing mixtures of solvents or for samples with ahigh salt content (to avoid spattering). If spattering is suspected,develop drying ramp by visual inspection of the drying stage,using a mirro r. Samples that contain a complex mjxture of matrixcomponents sometimes require ramp charring to effect con-trolled, complete thermal decomposition. Ramp atomization mayminimize background absorption by permitting volatilization ofthe element being determined before the matrix. This is espe-cially applicable in the determination of such volatile elementsas cadmium and lead. Use of time-resolved absorbance profiles(available on most modem instruments) greatly aids methoddevelopment. Changes in atomization, notably the element peakappearance time and magnitude of background and metal absor-bances, can be monitored directly.

Improve analysis by using a graphite platform, inserted intothe graphite tube, as the atomization site. The platform is notheated as directly by the current f10wing through the graphitetube; thus the metal atomizes later and under more uniformconditions.

Use standard additions to compensate for matrix interferences.When mabng standard additions, determine whether the addedmetal and that in the sample behave similarly under the specifiedconditions. [See Section 3113BAd2)]. In the extreme, test everysample for recovery (85 to 115% recovery desired) to determineif standard addition is needed. Test every new sample type forrecovery. Recovery of only 40 to 85% generally indicates thatstandard addition is required. Often, as long as the samples arefrom sources of consistent properties, a representative recoverycan be used to characterize the analysis and determine the

TABLE3113:11. DETECfIONLEVELSANDCONCENTRATIONRANGESFORELECfROTHERMALATOMIZATIONATOMICABSORYfIONSPECfROMETRY

Wavelengthnm

EstimatedDetection

LeveijLg/L

33120.20.12IIII0.2II20.25

OptimumConcentration

RangejLg/L

20-20020-3005-100

10-2001-30

0.5-105-1005-1005-1005-1005-1001-303-605-1005-1001-25

20-300

309.3217.6193.7553.6234.9228.8357.9240.7324.7248.3283.3279.5313.3232.0196.0328.1224.6

*The more sensitive 217.0-nm wavelength is reeommended for instruments withbaekground eorreetion eapabilities.

necessity of standard addition. Test samples of unknown originor of complex composition (digestates, for example) individuallyfor metal recovery. Ideally, chemical modifiers and graphiteplatforms render the sample tit to be analyzed using a standardanalytical calibration curve. Always verify this assumption;however, a properly developed method with judicious use ofchemical modifiers should eliminate the necessity for standardaddition in all but the most extreme samples.

Chemical interaction of the graphite tube with various ele-ments to form refractory carbides occurs at high charring andatomization temperatures. Elements that form carbides are bar-ium, molybdenum, nickel, titanium, vanadium, and silicon. Car-bide formation is characterized by broad, tailing atomizationpeaks and reduced sensitivity. Using pyrolytically coated tubesfor these metaIs minimizes the problem.

4. Sensitivity, Detection Levels, and OptimumConcentration Range

Estimated detection levels and optimum concentration rangesare listed in Table 3113 :II. These values may vary with thechemical form of the element being determined, sample compo-sition, or instrumental conditions.

For a given sample, increased sensitivity may be achieved byusing a larger sample volume or by reducing f10w rate of thepurge gas or by using gas interrupt during atomization. Note,however, that these techniques also will increase the effects ofany interferences present. Sensitivity can be decreased by dilut-ing the sample, reducing sample volume, increasing purge-gasf1ow, or using a less sensitive wavelength. Use of argon, ratherthan nitrogen, as the purge gas generally improves sensitivityand reproducibility. Hydrogen mixed with the inert gas maysuppress chemical interference and increase sensitivity by acting

as a reducing agent, thereby aiding in producing more ground-state atoms. Pyrolytically coated graphite tubes can increasesensitivity for the more refractory elements and are recom-mended. The optical pyrometer/maximum power accessoryavailable on some instruments also offers increased sensitivitywith lower atomization temperatures for many elements.

Using the Stabilized Temperature Platforrn Fumace (STPF)technique, which is a combination of individual techniques, alsooffers significant interference reduction with improved sensitiv-ity. Sensitivity changes with sample tube age. Discard graphitetubes when significant variations in sensitivity or poor reproduc-ibility are observed. The use of high acid concentrations, brinesamples, and matrix modifiers often drastically reduces tube life.Preferably use the graphite platforrn in such situations.

1. PERKIN-ELMERCORPo1991. Summary of Standard Conditions forGraphite Fumace. Perk.in-Elmer Corp., Norwalk, Conn.

2. ROTHERY,E., ed. 1988. Analytical Methods for Graphite Tube Atom-izers. Varian Techtron Pty, Ltd., Mulgrave, Victoria, Australia.

FERNANDEZ,FJ. & D.e. MANNlNG.1971. Atomic absorption analyses ofmetal pollutants in water using a heated graphite atomizer. AtomicAbsorption Newsletter 10:65.

SEGAR,D.A. & 1.G. GONZALEZ.1972. Evaluation of atomic absorptionwith a heated graphite atomizer for the direct determination of tracetransition metaIs in sea water. Ana!. Chim. Acta 58:7.

BARNARD,W.M. & MJ. FISHMAN.1973. Evaluation of the use of heatedgraphite atomizer for the routine determination of trace metais inwater. Atomic Absorption Newsletter 12:118.

KAHN,H.L. 1973. The detection of metallic elements in wastes andwaters with the graphite fumace. /nt. J. Environ. Ana!. Chem.3:121.

WALSH,P.R., 1.L. FASCHING& R.A. DucE. 1976. Matrix effects and theircontrol during the f1ameless atomic absorption determination ofarsenic. Ana!. Chem. 48: 1014.

HENN,E.L. 1977. Use of Molybdenum in Eliminating Matrix Interfer-ences in Flameless Atomic Absorption. Spec. Tech. Publ. 618,American Soc. Testing & Materials, Philadelphia, Pa.

MARTlN,T.D. & 1.F. Kopp. 1978. Methods for MetaIs in Drinking Water.U.S. Environmental Protection Agency, Environmental Monitoringand Support Lab., Cincinnati, ühio.

HYDES,DJ. 1980. Reduction of matrix effects with a soluble organicacid in the carbon fumace atomic absorption spectrometric deter-mination of cobalt, copper, and manganese in seawater. Anal.Chem. 52:289.

SOTERA,1.1. & H.L. KAHN.1982. Background correction in AAS. Amer.Lab. 14:100.

SMITH,S.B. & G.M. HIEFrJE. 1983. A new background-correctionmethod for atomic absorption spectrometry. App!. Spectrosc. 37:419.

GROSSER,Z. 1985. Techniques in Graphite Fumace Atomic AbsorptionSpectrophotometry. Perk.in-Elmer Corp., Ridgefield, Conn.

SLAVIN,W. & G.R. CARNICK.1985. A survey of applications of thestabilized temperature platform fumace and Zeeman correction.Atomic Spectrosc. 6:157.

BRUEGGEMEYER,T. & F. FRICKE.1986. Comparison of furnace & atom-ization behavior of aluminum from standard & thorium-treatedL'vov platforms. Anal. Chem. 58: 1143.

This method is suitable for determination of micro quantitiesof aluminum, antimony, arsenic, barium, beryllium, cadmium,chromium, cobalt, copper, iron, lead, manganese, molybdenum,nickel, selenium, silver, and tin. It is also applicable to analysisof bismuth, gallium, gerrnanium, gold, indium, mercury, tellu-rium, thallium, and vanadium, but precision and accuracy dataare not yet available.

a. Atomic absorption spectrometer: See Section 3111A.6a.The instrument must have background correction capability.

b. Source lamps: See Section 3111A.6d.C. Graphite furnace: Use an electrically heated device with

electronic control circuitry designed to carry a graphite tube orcup through a heating program that provides sufficient therrnalenergy to atomize the elements of interest. Fumace heat control-lers with only three heating steps are adequate only for freshwaters with low dissolved solids content. For salt waters, brines,and other complex matrices, use a fumace controller with up toseven individually programmed heating steps. Fit the fumaceinto the sample compartment of the spectrometer in place of the

conventional bumer assembly. Use argon as a purge gas tominimize oxidation of the fumace tube and to prevent the for-mation of metallic oxides. Use graphite tubes with platforrns tominimize interferences and to improve sensitivity.

d. Readout: See Section 3111A.6c.e. Sample dispensers: Use microliter pipets (5 to 100 J1,L) or

an automatic sampling device designed for the specific instru-ment.

f Vent: See Section 3111A.6fg. Cooling water supply: Cool with tap water fiowing at I to

4 Llmin or use a recirculating cooling device.h. Membrane filter apparatus: Use an all-glass filtering device

and 0.45-lLm or smaller-pore-diameter membrane filters. Fortrace analysis of aluminum, use polypropylene or TFE devices.

a. Metal-free water: See Section 311lB.3c.b. Hydrochloric acid, HCI, 1 + I and conc.C. Nitric acid, HN03, 1 + 1 and conc.d. Matrix modifier stock solutions:1) Magnesium nitrate, 10 000 mg MgIL: Dissolve 10.5 g

Mg(N03h . 6H20 in water. Dilute to 100 rnL.

2) Nickel nitrate, 10 000 mg NilL: Dissolve 4.96 gNi(N03)2 . 6H20 in water. Dilute to 100 mL.

3) Phosphoric acid, 10% (v/v): Add 10 mL conc H3P04 towater. Dilute to 100 rnL.

4) Palladium nitrate, 4000 mg PdlL: Dissolve 8.89 gPd(N03h .H20 in water. Dilute to I L.

5) Citric acid, 4%: Dissolve 40 g citric acid in water. Dilute toIL.

NOTE:Ali of the modifier solutions recommended in Table3113:1 can be prepared with volumetric combination of theabove solutions and water. For preparation of other matrix mod-ifiers, see references or follow manufacturers' instructions.

e. Stock metal solutions: Refer to Sections 3111B and 3114.f Chelating resin: 100 to 200 mesh* purified by heating at

60°C in 1ON NaOH for 24 h. Cool resin and rinse 10 times eachwith alternating portions of lN HCI, metal-free water, lN NaOH,and metal-free water.

g. Metal-free seawater (or brine): Fill a lA-cm-ID X 20-cm-long borosilicate glass column to within 2 cm of the top withpurified chelating resin. Elute resin with successive 50-mL por-tions of IN HCI, metal-free water, IN NaOH, and metal-freewater at the rate of 5 rnL/min just before use. Pass salt water orbrine through the column at a rate of 5 mL/min to extract tracemetaIs present. Discard the first 10 bed volumes (300 rnL) ofeluate.

a. Sample pretreatment: Before analysis, pretreat ali samplesas indicated below. Rinse ali glassware with 1 + 1 HN03 andwater. Carry out digestion procedures in a clean, dust-free labo-ratory area to avoid sample contamination. For digestion of tracealuminum, use polypropylene or TFE utensils to avoid leachablealuminum from glassware.

1) Dissolved metals-See Section 3030B. For samples requir-ing arsenic and/or selenium analysis add 3 mL 30% hydrogenperoxide/l 00 mL sample and an appropriate volume of nickelnitrate solution (see Table 3113:1) before analysis. For ali othermetaIs no further pretreatment is required except for adding anoptional matrix modifier.

2) Total recoverable metais (AI, Sb, Ba, Be, Cd, Cr, Co, Cu,Fe, Pb, Mn, Mo, Ni, Ag, and Sn)-NOTE: Sb and Sn are notrecovered unless HCI is used in the digestion. See Section3030D. Quantitatively transfer digested sample to a 100-rnLvolumetric flask, add an appropriate amount of matrix modifier(see Table 3113:1), and dilute to volume with water.

3) Total recoverable metaIs (As, Se)-Transfer 100 rnL ofshaken sample, 1 rnL conc HN03, and 2 rnL 30% H202 to aclean, acid-washed 250-mL beaker. Heat on a hot plate withoutallowing solution to boiI until volume has been reduced to about50 mL. Remove from hot plate and let cool to room temperature.Add an appropriate concentration of nickel (see Table 3113:1),and dilute to volume in a 100-rnL volumetric flask with water.Substitution of palladium is uneconomical. Nickel may be de-leted if palladium is co-added during analysis. Simultaneouslyprepare a digested blank by substituting water for sample andproceed with digestion as described above.

* Chelex 100, or equivalem, available from Bio-Rad Laboratories, Richmond,CA.

b. Instrument operation: Mount and align furnace deviceaccording to manufacturer' s instructions. Tum on instrument anddata collection system. Select appropriate light source and adjustto recommended electrical setting. Select proper wavelength andset ali conditions according to manufacturer' s instructions, in-cluding background correction. Background correction is impor-tant when elements are determined at short wavelengths or whensample has a high leveI of dissolved solids. Background correc-tion normally is not necessary at wavelengths longer than 350nm. If background correction above 350 nm is needed deuteriumarc background correction is not useful and other types must beused.

Select proper inert- or sheath-gas flow. In some cases, it isdesirable to interrupt the inert-gas flow during atomization. Suchinterruption results in increased sensitivity by increasing resi-dence time of the atomic vapor in the optical path. Gas inter-ruption also increases background absorption and intensifiesinterference effects, but modern background correction methodsusually eliminate these problems. Consider advantages and dis-advantages of this option for each matrix when optimizinganalytical conditions.

To optimize graphite fumace conditions, carefully adjust fumacetemperature settings to maximize sensitivity and precision and tominimize interferences. Follow manufacturer's instructions.

Use drying temperatures slightly above the solvent boilingpoint and provide enough time and temperature for completeevaporation without boiling or spattering.

Select atomization temperature by determining the lowesttemperature providing maximum sensitivity without signifi-cantly eroding precision. Optimize by a series of successivedeterminations at various atomization temperatures using a stan-dard solution giving an absorbance of 0.2 to 0.5.

The charring temperature must be high enough to maximizevolatilization of interfering matrix components yet too low tovolatilize the element of interest. With the drying and atomiza-tion temperatures set to their optimum values, analyze a standardsolution at a series of charring temperatures in increasing incre-ments of 50 to 100°C. When the optimum charring temperatureis exceeded, there will be a significant drop in sensitivity. Plotcharring temperature versus sample absorbance: the optimumcharring temperature is the highest temperature without reducedsensitivity. Verify optimization with major changes in samplematrix.

C. Instrument calibration: Prepare standard solutions for in-strument calibration by diluting metal stock solutions. Preparestandard solutions fresh daily.

Prepare a blank and at least three calibration standards in theappropriate concentration range (see Table 3113:11) for correlat-ing element concentration and instrument response. Match thematrix of the standard solutions to those of the samples asclosely as possible. In most cases, this simply requires matchingthe acid background of the samples. For seawaters or brines,however, use the metal-free matrix (9! 3g) as the standard solu-tion diluent. ln addition, add the same concentration of matrixmodifier (if required for sample analysis) to the standard solu-tions.

lnject a suitable portion of each standard solution, in order ofincreasing concentration. Analyze each standard solution in trip-licate to verify method precision.

ELECTROTHERMALATOMIC ABSORPTION SPECTROMETRY (3113)/Electrothermal Atomic Absorption Spectrometric Method 3-29

Construct an ana1ytical curve by plotting the average peak reproducible results are obtained. A variation of s; 10% isabsorbances ar peak areas of the standard solution versus con- considered acceptable reproducibility. Average replicate values.centration on linear graph paper. Altematively, use electronic I) Direct determination-Inject a measured portion of pre-instrument calibration if the instrument has this capability. treated sample into the graphite fumace. Use the same volume as

d. Sample analysis: Analyze all samples except those demon- was used to prepare the calibration curve. Usually add modifierstrated to be free of matrix interferences (based on recoveries of immediately after the sample, preferably using an automatic85% to 115% for known additions) using the method of standard sampler or a micropipet. Some methods require modifier to beadditions. Analyze all samples at least in duplicate or until injected before the sample. Use the same volume and concen-

TABLE3113:III. INTERLABORATORYSINGLE-ANALYSTPRECISIONDATAFORELECfROTHERMALATOMIZATlONMETHODS'

Single-Analyst Precision%RSD

LabConcentration Pure Drinking Surface Effluent Effluent Effluent

Element J.LgIL Water Water Water 1 2 3

AI 28 66 108 70 66125 27 35 24 34

11 000 11 2258300 27 19

460 9 302180 28 4

10.5 20 13 13 13 56 18230 10 18 13 21 94 14

As 9.78 40 25 15 74 23 11227 10 6 8 11 15 6

Ba 56.5 36 21 29 59 23 27418 14 12 20 24 24 18

Be 0.45 18 27 15 30 2 1110.9 14 4 9 7 12 12

Cd 0.43 72 49 1 121 35 2712 11 17 22 14 11 15

Cr 9.87 24 33 10 23 15 10236 16 7 11 13 16 7

Co 29.7 10 17 10 19 24 12420 8 11 13 14 9 5

Cu 10.1 49 47 17 17 30234 8 15 6 21 11300 6 11

1670 11 6Fe 26.1 144 52 153 124

455 48 37 45 311030 17 305590 6 32

370 14 192610 9 18

Pb 10.4 6 19 17 21 19 33243 17 7 17 18 12 16

Mn 0.44 187 180 27514.8 32 19 1891.0 15 48

484.0 4 12111.0 12 21666.0 6 20

Ni 26.2 20 26 25 24 18 9461.0 15 11 9 8 11 4

Se 10.0 12 27 16 35 41 13235.0 6 6 15 6 13 14

Ag 8.48 10 15 27 1656.5 14 7 16 230.45 27 166 48

13.6 15 4 10

3-30 METALS (3000)

tration of modifier for ali standards and samples. Dry, char, and tion) or peak area of the most concentrated standard solution,atomize according to the preset programo Repeat until reproduc- dilute sample and reanalyze. If very large dilutions are re-ible results are obtained. quired, another technique (e.g., tlame AA or ICP) may be

Compare the average absorbance value or peak area to the more suitable for rhis sample. Large dilution factors magnifycalibration curve to determine concentration of the element of small errors on final calculation. Keep acid background andinterest. Altematively, read results directly if the instrument is concentration of matrix modifier (if present in the solutions)equipped with this capability. If absorbance (or concentration) or constant. Dilute the sample in a blank solution of acid andpeak area of the sample is greater than absorbance (concentra- matrix modifiers.

TABLE3113:IV. INTERLABORATORYOVERALLPRECISIONDATAFORELECI"ROTHERMALATOMIZATIONMETHODs'

Overall Precision%RSD

LabConcentration Pure Drinking Surface Effluent Effluent Effluent

E1ement JLg/L Water Water Water 1 2 3

AI 28 99 114 124 131125 45 47 49 40

11 000 19 4358300 31 32

460 20 472180 30 15

10.5 37 19 22 50 103 39230 26 16 16 17 180 21

As 9.78 43 26 37 72 50 39227 18 12 13 20 15 14

Ba 56.5 68 38 43 116 43 65418 35 35 28 38 48 16

Be 0.45 28 31 15 67 50 3510.9 33 15 26 20 9 19

Cd 0.43 73 60 5 88 43 6512 19 2S 41 26 20 27

Cr 9.87 30 53 24 60 41 23236 18 14 24 20 14 20

Co 29.7 13 26 17 18 21 17420 21 21 17 18 13 13

Cu 10.1 58 82 31 32 74234 12 33 19 21 26300 13 14

1670 12 13Fe 26.1 115 93 306 204

455 53 46 53 441030 32 255590 10 43

370 28 222610 13 22

Pb 10.4 27 42 31 23 28 47243 18 19 17 19 19 25

Mn 0.44 299 272 24814.8 52 41 2991.0 16 45

484.0 5 17111.0 15 17666.0 8 24

Ni 26.2 35 30 49 35 37 43461.0 23 22 15 12 21 17

Se 10.0 17 48 32 30 44 51235.0 16 18 18 17 22 34

Ag 8.48 23 16 35 3456.5 15 24 32 280.45 57 90 368

13.6 19 19 59

Proceed to lJ[ 5a below.2) Method of standard additions-Refer to lJ[ 4c above. The

method of standard additions is valid only when it falls in thelinear portion of the calibration curve. Once instrument sensitiv-ity has been optimized for the element of interest and the linearrange for the element has been established, proceed with sampleanalyses.

Inject a measured volume of sample into furnace device. Dry,char or ash, and atomize samples according to preset programoRepeat until reproducible results are obtained. Record instrumentresponse in absorbance or concentration as appropriate. Add aknown concentration of the element of interest to a separate portionof sample so as not to change significantly the sample volume.Repeat the determination.

TABLE 3113:V. INTERLABORATORYRELATIVE ERROR DATA FOR ELECTROTHERMALATOMIZATIONMETHoDs'

Relative Error%

LabConcentration Pure Drinking Surface Effluent Effluent Effluent

Element fLg/L Water Water Water 1 2 3

AI 28.0 86 150 54 126125.0 4 41 39 30

11000.0 2 1458300.0 12 7

460.0 2 112180.0 11 9

Sb 10.5 30 32 28 24 28 36230.0 35 14 19 13 73 39

As 9.78 36 1 22 106 13 16227.0 3 7 10 19 6 13

Ba 56.5 132 54 44 116 59 40418.0 4 O O 13 6 60

Be 0.45 40 16 11 16 10 1510.9 13 2 9 7 8 8

Cd 0.43 58 45 37 66 16 1912.0 4 6 5 22 18 3

Cr 9.87 10 9 4 2 5 15236.0 11 O 9 13 5 8

Co 29.7 7 7 1 6 3 13420.0 12 8 8 11 5 18

Cu 10.1 16 48 2 5 15234.0 8 7 O 4 19300.0 4 21

1670.0 6 2Fe 26.1 85 60 379 158

455.0 43 22 31 181030.0 8 85590.0 2 12

370.0 4 112610.0 35 2

Pb 10.4 16 10 17 1 34 14243.0 5 15 8 18 15 29

Mn 0.44 332 304 55614.8 10 1 3691.0 31 10

484.0 42 4111.0 I 29666.0 6 23

Ni 26.2 9 16 10 7 33 54461.0 15 19 18 31 16 18

Se 10.0 12 9 6 36 17 37235.0 7 7 O 13 10 17

Ag 8.48 12 1 51 2056.5 16 8 51 220.45 34 162 534

13.6 3 12 5

Add a known concentration (preferably twice that used in thefirst addition) to a separate sample portion. Mix well and repeatthe determination.

Using linear graph paper, plot average absorbance or instru-ment response for the sample and the additions on the verticalaxis against the concentrations of the added element on thehorizontal axis, using zero as the concentration for the sample.Draw a straight line connecting the three points and extrapolateto zero absorbance. The intercept at the horizontal axis is thenegative of the element concentration in the sample. The con-centration axis to the left of the origin should be a mirror imageof the axis to the right.

where:

C = metal concentration as read directly from the instrument orfrom the calibration curve, fLg/L, and

F = dilution factor.

where:

C = metal concentration as read from the method of additionsplot, fLg/L, and

F = dilution factor.

Data typical of the precision and bias obtainable are presentedin Tables 3113:III, IV, and V.

See Section 3020 for specific quality control procedures tobe followed during analysis. Although previous indicationswere that very low optimum concentration ranges were at-tainable for most metaIs (see Table 3113:II), data in Table3113:111 using variations of these protocols show that thismay not be so. Exercise extreme care when applying this

method to the lower concentration ranges. Verify analystprecision at the beginning of each analytical mn by makingtriplicate analyses. Verify autosampler precision by checkingvolumes (by weight) delivered by the autosampler at routinelyused injection volume settings.

1. COPELAND,T.R. & J.P. MANEY.1986. EPA Method Study 31: TraceMetais by Atomic Absorption (Furnace Techniques). EPA-600/S4-85-070, U.S. Environmental Protection Agency, EnvironmentalMonitoring and Support Lab., Cincinnati, ühio.

RENSHAW,G.D. 1973. The determination of barium by flameless atomicabsorption spectrophotometry using a modified graphite tube atom-izer. Atomic Absorption Newsletter 12:158.

YANAGlSAWA,M., T. TAKEUCHI& M. SUZUKJ.1973. Flameless atomicabsorption spectrometry of antimony. Ana!. Chim. Acta 64:381.

RATIONETIl,A. 1974. Determination of soluble cadmium, lead, silverand indium in rainwater and stream water with the use of flamelessatomic absorption. Ana!. Chem. 46:739.

HENN,E.L. 1975. Determination of selenium in water and industrialeffluents by flameless atomic absorption. Ana!. Chem. 47:428.

MARTIN,T.D. & J.F. Kopp. 1975. Determining selenium in water, waste-water, sediment and sludge by flameless atomic absorption spec-trometry. Atomic Absorption Newsletter 14:109.

MARUTA,T., K. MINEGISHI& G. SUDüH.1976. The flameless atomicabsorption spectrometric determination of aluminum with a carbonatomization system. Ana!. Chim. Acta 81:313.

CRANSTON,R.E. & J.W. MURRAY.1978. The determination of chromiumspecies in natural waters. Ana!. Chim. Acta 99:275.

HOFFMEISTER,W. 1978. Determination of iron in ultrapure water byatomic absorption spectroscopy. Z. Ana!. Chem. 50:289.

LAGAS,P. 1978. Determination of beryllium, barium, vanadium andsome other elements in water by atomic absorption spectrometrywith electrothermal atomization. Ana!. Chim. Acta 98:261.

CARRüNDü,M.J.T., J.N. LESTER& R. PERRY.1979. Electrothermalatomic absorption determination of total aluminum in waters andwaste waters. Ana!. Chim. Acta lII:29I.

NAKAHARA,T. & c.L. CHAKRABARTI.1979. Direct determination of tracesof molybdenum in synthetic sea water by atomic absorption spec-trometry with electrothermal atomization and selective volatiliza-tion of the salt matrix. Ana!. Chim. Acta 104:99.

TIMINAGA,M. & Y. UMEZAKI.1979. Determination of submicrogramamounts of tin by atomic absorption spectrometry with electrother-mal atomization. Ana!. Chim. Acta 110:55.

3114 ARSENIC AND SELENIUM BY HYDRIDE GENERATION/ATOMICABSORPTION SPECTROMETRY*

3114 A. Introduction

For general introductory material on atomic absorption spec-trometric methods, see Section 3111A.

Two methods are presented in this section: A manual method anda continuous-flow method especially recommended for selenium.Continuous-flow automated systems are preferable to manual hy-dride generators because the effect of sudden hydrogen generationon light-path transparency is removed and any blank response fromcontarnination of the HCI reagent by the elements being deterrninedis incorporated into the background base line.

a. PrincipIe: This method is applicable to the determinationof arsenic and selenium by conversion to their hydrides bysodium borohydride reagent and transport into an atomic absorp-tion atornizer.

Arsenous acid and selenous acid, the As(I1I) and Se(IV) oxi-dation states of arsenic and selenium, respectively, are instanta-neously converted by sodium borohydride reagent in acid solu-tion to their volatile hydrides. The hydrides are purged contin-uously by argon or nitrogen into a quartz cell heated electricallyor by the fiame of an atomic absorption spectrometer and con-verted to the gas-phase atoms. The sodium borohydride reducingagent, by rapid generation of the elemental hydrides in an ap-propriate reaction cell, minimizes dilution of the hydrides by thecarrier gas and provides rapid, sensitive determinations of ar-senic and selenium.

CAUTION: Arsenic and selenium and their hydrides are toxic.Handle with care.

At room temperature and solution pH values of I or less,arsenic acid, the As(V) oxidation state of arsenic, is reducedrelatively slowly by sodium borohydride to As(I1I), which isthen instantaneously converted to arsine. The arsine atomicabsorption peaks commonly are decreased by one-fourth toone-third for As(V) when compared to As(III). Determinationof total arsenic requires that ali inorganic arsenic compoundsbe in the As(llI) state. Organic and inorganic forms of arsenicare first oxidized to As(V) by acid digestion. The As(V) thenis quantitatively reduced to As(lIl) with sodium or potassiumiodide before reaction with sodium borohydride.

Selenic acid, the Se(VI) oxidation state of selenium, is notmeasurably reduced by sodium borohydride. To determinetotal selenium by atomic absorption and sodium borohydride,first reduce Se(VI) formed during the acid digestion procedureto Se(IV), being careful to prevent reoxidation by chlorine.Efficiency of reduction depends on temperature, reductiontime, and HCI concentration. For 4N HCI, heat 1 h at 100°C.For 6N HCI, boiling for 10 min is sufficient.1-3 Alternatively,autoclave samples in sealed containers at 121°C for 1 h.NOTE: Autoclaving in sealed containers may result in incom-plete reduction, apparently due to the buildup of chlorine gas.To obtain equal instrument responses for reduced Se(VI) andSe (IV) solutions of equal concentrations, manipulate HCIconcentration and heating time. For further details, see Sec-tion 3500-Se.

b. Equipment selection: Certain atomic absorption atomiz-ers and hydride reaction cells are available commercially for usewith the sodium borohydride reagent. A functional manual sys-tem that can be constructed in the laboratory is presented inFigure 3114:1. Irrespective of the hydride reaction cell-atornizersystem selected, it must meet the following quality-control COl\.-

siderations: (a) it must provide a precise and reproducible stan-

Gas dispersiontube

Outlet tubeto quartz ceU

Rubberstopper

200-mL Berzelius beaker or300-mL beaker

dard curve between O and 20 /-tgAs or SelL and an instrumentaldetection limit between 0.1 and 0.5 /-tg As or SelL; (b) whencarried through the entire procedure, oxidation state couples [As(I1I) - As (V) or Se (IV) - Se (VI)] must cause equal instrumentresponse; and (c) sample digestion must yield 80% or greaterrecovery of added cacodylic acid (dimethyl arsinic acid) and90% or greater recovery of added As(I1I), As(V), Se(VI), orSe(IV).

Quartz atomization cells provide for the most sensitive arsenicand selenium hydride determinations. The quartz cell can beheated electrically or by an air-acetylene fiame in an atornicabsorption unit.

c. Digestion techniques: Waters and wastewaters may containvarying amounts of organic arsenic compounds and inorganiccompounds of As(I1I), As(V), Se(IV), and Se(VI). To measuretotal arsenic and selenium in these samples requires sampledigestion to solubilize particulate forms and oxidize reducedforms of arsenic and selenium and to convert any organic com-pounds to inorganic ones. Organic selenium compounds rarely

BIBLIOTECA CEFET·RS

have been demonstrated in wateI. It is left to the expeliencedanalyst's judgment whether sample digestion is required.

Various alternative digestion procedures are provided in 1[s4cand d below. Consider sulfuric-nitric-perchloric acid digestion (1[4c) or sulfuric-nitric acid digestion (3030F) as providing a mea-sure of total recoverable arsenic rather than total arsenic becausethey do not completely convert certain organic arsenic com-pounds to As(V). The sulfuric-nitric-perchloric acid digestioneffectively destroys organics and most particulates in untreatedwastewaters or solid samples, but does not convert all organicarsenicals to As(V). The potassium persulfate digestion (1[4d) iseffective for converting organic arsenic and selenium com-pounds to As(V) and Se(VI) in potable and surface waters and inmost wastewaters.4

The HCl-autoclave reduction of Se(VI) described above and in1[4f is an effective digestion procedure for total inorganic sele-nium; however, it has not been found effective for convertingbenzene substituted selenium compounds to inorganic selenium.In all cases, verify the effectiveness of digestion methods bycarrying samples with known additions of organic As or Se(IV)through the entire procedure.

d. lnterferences: Interferences are minimized because the Asand Se hydrides are removed from the solution containing mostpotential interfering substances. Slight response variations occurwhen acid matrices are varied. Control these variations by treat-ing standards and samples in the same manneI. Low concentra-tions of noble metaIs (approximately 100 /LglL of Ag, Au, Pt, Pd,etc.), concentrations of copper, lead, and nickel at or greater than1 mg/L, and concentrations between 0.1 and 1 mglL of hydride-forming elements (Bi, Sb, Sn, and Te) may suppress the responseof As and Se hydrides. Interference by transition metaIs dependsstrongly on HCl concentration. Interferences are less pronouncedat 4 to 6N HCl than at lower concentrations.5 The presence of Asor Se in each other' s matrices can cause similar suppression.Reduced nitrogen oxides resulting from HN03 digestion andnitlite also can suppress instrumental response for both elements.Large concentrations of iodide interfere with the Se determina-tion by reducing Se to its elementaI formo Do not use anyglassware for determining Se that has been used for iodidereduction of As(V).

To prevent chlorine gas produced in the reduction of Se(VI) toSe(IV) from reoxidizing the Se(IV), generate the hydride withina few hours of the reduction steps or purge the chlorine from thesamples by sparging.6

Interferences depend on system design and defy quantitativedescription because of their synergistic effects. Certain watersand wastewaters can contain interferences in sufficient concen-tration to suppress absorption responses of As and Se. Forrepresentative samples in a given laboratory and for initial anal-yses of unknown wastewaters, add appropriate inorganic formsof As or Se to digested sample portions and measure recovery. Ifaverage recoveries are less than 90%, consider using alternativeanalytical procedures.

e. Detection levei and optimum concentration range: For botharsenic and selenium, analyzed by aspiration into a nitrogen-hydrogen flame after reduction, the method detection leveI is 2/LglL or lower and the optimum concentration range 2 to 20/LglL.

a. Atomic absorption spectrometer equipped with air-acety-Iene flame and quartz cell with mounting bracket or an electri-cally heated quartz cell, As and Se electrodeless discharge lampswith power supply, background correction at measurementwavelengths, and appropriate strip-chart recordeI. A good-qual-ity 1O-mV recorder with high sensitivity and a fast response timeis needed.

b. Atomizer: Use one of the following:1) Cylindrical quartz cell, 10 to 20 cm long, bracket-mount-

able above air-acetylene burneI.2) Cylindrical quartz cell, 10 to 20 cm long, electrically heated

by external nichrome wire to 800 to 900°C?3) Cylindrical quartz cell with internal fuel rich hydrogen-

oxygen (air) flame.sThe sensitivity of quartz cells deteriorates over several months

of use. Sensitivity sometimes may be restored by treatment with40% HF. CAUTION:HF is extremely corrosive. Avoid all contactwith exposed skin. Handle with care.

C. Reaction cell for producing As or Se hydrides: See Figure3114:1 for an example of a manual, laboratory-made system. Acommercially available system is acceptable if it utilizes liquidsodium borohydride reagents; accepts samples digested in ac-cordance with 1[s 4c, d, and e; accepts 4 to 6N HCl; and isefficiently and precisely stirred by the purging gas and/or amagnetic stirreI.

d. Eye dropper or syringe capable of delivering 0.5 to 3.0 roLsodium borohydride reagent. Exact and reproducible addition isrequired so that production of hydrogen gas does not varysignificantly between determinations.

e. Vent: See Section 3111A.6f

a. Sodium borohydride reagent: Dissolve 8 g NaBH4 in 200roL O.IN NaOH. Prepare fresh daily.

b. Sodium iodide prereductant solution: Dissolve 50 g NaI in500 mL wateI. Prepare fresh daily. Alternatively use an equiv-alent KI solution.

C. Sulfuric acid, H2S04, I8N.d. Sulfuric acid, 2.5N: Cautiously add 35 roL conc H2S04 to

about 400 roL water, let cool, and adjust volume to 500 roL.e. Potassium persulfate, 5% solution: Dissolve 25 g K2S20S in

water and dilute to 500 roL. Store in glass and refrigerate.Prepare weekly.f Nitric acid, HN03, conc.g. Perchloric acid, HCI04, conc.h. Hydrochloric acid, HCl, conc.i. Argon (or nitrogen), commercial grade.j. Arsenic(lll) solutions:1) Stock As(lll) solution: Dissolve 1.320 g arsenic trioxide,

AS203, in water containing 4 g NaOH. Dilute to 1 L; 1.00 mL =1.00 mg As(III).

2) lntermediate As(llI) solution: Dilute 10 mL stock As solu-tion to 1000 mL with water containing 5 mL conc HCl; 1.00 mL= 10.0 /LgAs(III).

3) Standard As(lll) solution: Dilute 10 mL intermediateAs(III) solution to 1000 mL with water containing the sameconcentration of acid used for sample preservation (2 to 5 mL

cone HN03); 1.00 rnL = 0.100 J.Lg As(I1I). Prepare dilutedsolutions daily.

k. Arsenic(V) solutions:1) Stock As(V) solution: Dissolve 1.534 g arsenie pentoxide,

As20S' in distilled water eontaining 4 g NaOH. Dilute to 1 L;1.00 mL = 1.00 mg As(V).

2) Intermediate As(V) solution: Prepare as for As(I1I) above;1.00 rnL = 10.0 J.Lg As(V).

3) Standard As(V) solution: Prepare as for As(I1I) above; 1.00rnL = 0.100 J.Lg As(V).

l. Organic arsenic solutions:1) Stock organic arsenic solution: Dissolve 1.842 g dimethyl-

arsinie aeid (eaeodylie aeid), (CH3)2AsOOH, in water eontaining4 g NaOH. Dilute to 1 L; 1.00 rnL = 1.00 mg As. [NOTE:Cheekpurity of eaeodylie aeid reagent against an intermediate arseniestandard (50 to 100 mg As/L) using flame atomie absorption.]

2) Interrnediate organic arsenic solution: Prepare as forAs(I1I) above; 1.00 rnL = 10.0 J.Lg As.

3) Standard organic arsenic solution: Prepare as for As(I1I)above; 1.00 rnL = 0.100 J.Lg As.

m. Selenium(IV) solutions:1) Stock Se(IV) solution: Dissolve 2.190 g sodium selenite,

Na2Se03, in water eontaining 10 rnL HCI and dilute to 1 L; 1.00rnL = 1.00 mg Se(IV).

2) Intermediate Se(N) solution: Dilute 10 rnL stoek Se(lV) to1000 rnL with water eontaining 10 rnL cone HCI; 1.00 rnL =10.0 J.Lg Se(lV).

3) Standard Se(IV) solution: Dilute 10 mL intermediateSe(IV) solution to 1000 mL with water eontaining the sameeoneentration of aeid used for sample preservation (2 to 5 rnLcone HN03). Prepare solution daily when eheeking the equiva-leney of instrument response for Se(IV) and Se(VI); 1.00 mL =0.100 J.Lg Se(IV).

n. Selenium(VI) solutions:1) Stock SerVI) solution: Dissolve 2.393 g sodium selenate,

Na2Se04, in water eontaining 10 rnL cone HN03. Dilute to 1 L;1.00 rnL = 1.00 mg Se(VI).

2) Intermediate SerVI) solution: Prepare as for Se(lV) above;1.00 rnL = 10.0 J.Lg Se (VI).

3) Standard SerVI) solution: Prepare as for Se(IV) above; 1.00rnL = 0.100 J.Lg Se(VI).

a. Apparatus setup: Either see Figure 3114:1 or follow man-ufaeturer's instruetions. Conneet inlet of reaetion eell with aux-iliary purging gas eontrolled by flow meter. If a drying eellbetween the reaetion eelI and atomizer is neeessary, use onlyanhydrous CaCI2 but not CaS04 beeause it may retain SeH2.Before using the hydride generation/analysis system, optimizeoperating parameters. Align quartz atomizers for maximum ab-sorbanee. Aspirate a blank until memory effeets are removed.Establish purging gas flow, eoneentration and rate of addition ofsodium borohydride reagent, solution volume, and stirring ratefor optimum instrument response for the ehemieal speeies to beanalyzed. Optimize quartz eelI temperature. If sodium borohy-dride reagent is added toa quiekly, rapid evolution of hydrogenwill unbalanee the system. If the volume of solution beingpurged is toa large, the absorption signal wilI be deereased.

Reeommended wavelengths are 193.7 and 196.0 nm for As andSe, respeetively.

b. Instrument calibration standards: Transfer 0.00, 1.00, 2.00,5.00, 10.00, 15.00, and 20.00 rnL standard solutions of As(I1I) orSe(lV) to 100-rnL volumetrie flasks and bring to volume withwater eontaining the same aeid eoneentration used for samplepreservation (eommonly 2 to 5 rnL cone HN03/L). This yieldsblank and standard solutions of O, 1,2,5, 10, 15, and 20 J.Lg Asor Se/L. Prepare fresh daily. In all cases, standards must beearried through the same digestion protoeol as the samples tomonitor digestion effeetiveness.

c. Preparation of samples and standards for total recoverablearsenic and selenium: Use digestion proeedure deseribed in3030F for samples and standards. Altematively, add 50 rnLsample, As(I1I), or Se(IV) standard to 200-rnL Berzelius beakeror 100-rnL miero-kjeldahl flask. Add 7 rnL 18N H2S04 and 5 rnLcone HN03. Add a smalI boiling ehip or glass beads if neeessary.Evaporate to S03 fumes. Maintain oxidizing eonditions at alItimes by adding smalI amounts of HN03 to prevent solutionfrom darkening. Maintain an exeess of HN03 until alI organicmatter is destroyed. Complete digestion usually is indicated by alight-colored solution. Cool slightly, add 25 rnL water and 1 rnLcone HCI04 and again evaporate to S03 fumes to expel oxides ofnitrogen. CAUTION:See Section 3030H for cautions on use ofHClO4• Monitor effeetiveness of either digestion proeedure usedby adding 5 rnL of standard organie arsenic solution or 5 rnL ofa standard selenium solution to a 50-rnL sample and measuringreeovery, earrying standards and the sample with known additionthrough entire proeedure. To report total reeoverable arsenie astotal arsenie, average reeoveries of eaeodylie aeid must exeeed80%. After final evaporation of S03 fumes, dilute to 50 rnL forarsenie measurements or to 30 rnL for selenium measurements.For analysis of both elements in a single sample, inerease samplevolume to 100 rnL and double the volumes of aeids used in thedigestion. Adjust final digestate volume to 100 rnL. Use 50 rnLfor As and 30 rnL for Se determinations, making appropriatevolume eorreetions in ealculating results.

d. Preparation of samples and standards for total arsenic andselenium: Add 50 rnL undigested sample or standard to a200-rnL Berzelius beaker or 100-rnL miero-kjeldahl flask. Add 1mL 2.5N H2S04 and 5 rnL 5% K2S20S. Boil gently on apre-heated hot plate for approximately 30 to 40 min or until afinal volume of 10 rnL is reaehed. Do not let sample go todryness. Alternatively heat in an autoclave at 121°C for 1 h ineapped eontainers. After manual digestion, dilute to 50 rnL forsubsequent arsenie measurements and to 30 mL for seleniummeasurements. Monitor effeetiveness of digestion by measuringreeovery of As or Se as above. If poor reeovery of arsenie addedas eaeodylie aeid is obtained, reanalyze using double the amountof K2S20S' For analysis of both elements in a single sample,inerease sample volume to 100 rnL and double the volumes ofaeids used in the digestion. Adjust final digestate volume to 100rnL. Use 50 rnL for As and 30 rnL for Se determinations, makingappropriate volume eorreetions in ealculating results.

e. Deterrnination of arsenic with sodium borohydride: To 50mL digested standard or sample in a 200-rnL Berzelius beaker(see Figure 3114:1) add 5 rnL cone HCI and mix. Add 5 rnL Na!prereduetant solution, mix, and wait at least 30 mino (NOTE:TheNal reagent has not been found neeessary for eertain hydridereaetion eelI designs if a 20 to 30% loss in instrument sensitivity

is not important and variables of solution acid conditions, tem-peratures, and volumes for production of As(V) and arsine canbe controlled strictly. Such control requires an automated deliv-ery system; see Section 3114C.)

Attach one Berzelius beaker at a time to lhe rubber sloppercontaining the gas dispersion tube for the purging gas, thesodium borohydride reagent inlet, and the outlet to the quartzcell. Tum on strip-chart recorder and wait until the base line isestablished by the purging gas and all air is expelled from thereaction cell. Add 0.5 mL sodium borohydride reagent. After theinstrument absorbance has reached a maximum and returned tothe base line, remove beaker, rinse dispersion tube with water,and proceed to the next sample or standard. Periodically comparestandard As(I1I) and As(V) curves for response consistency.Check for presence of chemical interferences that suppress in-strument response for arsine by treating a digested sample with10 J-tglL As(I1I) or As(V) as appropriate. Average recoveriesshould be not less than 90%.f Determination of selenium with sodium borohydride: To 30

rnL digested standard or sample in a 200-rnL Berzelius beaker or100-rnL micro-kjeldahl flask, add 15 mL conc HCl and mix.Heat for a predetermined period at 90 to 100°C. Alternativelyautoclave at 121°C in capped containers for 60 min, or heat fora predetermined time in open test tubes using a 90 to 100°C hotwater bath or an aluminum block digester. Check effectivenessof the selected heating time by demonstrating equal instrumentresponses for calibration curves prepared either from standardSe(IV) or from Se(VI) solutions. Effective heat exposure forconverting Se(VI) to Se(IV), with no loss of Se(IV), rangesbetween 5 and 60 min when open beakers or test tubes are used.Establish a heating time for effective conversion and apply thistime to ali samples and standards. Do not digest standard Se(lV)and Se(VI) solutions used for this check of equivalency. Afterprereduction of Se(VI) to Se(IV), attach Berzelius beakers, oneat a time, to the purge apparatus. For each, turn on the strip-chartrecorder and wait until the base line is established. Add 0.50 rnLsodium borohydride reagent. After the instrument absorbancehas reached a maximum and returned to the base line, removebeaker, rinse dispersion tube with water, and proceed to the nextsample or standard. Check for presence of chemical interferencesthat suppress selenium hydride instrument response by treating adigested sample with 10 JLg Se (lV)IL. Average recoveriesshould be not less than 90%.

Construct a standard curve by plotting peak heights or areas ofstandards versus concentration of standards. Measure peakheights or areas of samples and read concentrations from curve.lf sample was diluted (or concentrated) before sample digestion,apply an appropriate facto r. On instruments so equipped, readconcentrations directly after standard calibration.

Single-laboratory, single-operator data were collected forAs(I1I) and organic arsenic by both manual and automatedmethods, and for the manual determination of selenium. Recov-ery values (%) from seven replicates are given below:

As(III)

91.8109.499.892.5

Manual with digestionManual without digestionAutomated with digestionAutomated without digestion

87.319.498.410.4

1. VlJAN, P.N. & O. LEUNG. 1980. Reduetion of ehemieal interfereneeand speeiation studies in the hydride generation-atomie absorptionmethod for selenium. Anal. Chim. Aeta 120: 141.

2. VOTH-BEACH, L.M. & O.E. SHRADER. 1985. Reduetion of interfer-enees in the determination of arsenie and selenium by hydride gen-eration. Speetroseopy 1:60.

3. JULSHAMN,K., O. RINGDAL,K.-E. SUNNING& O.R. BRAEKKAN.1982.Optimization of the determination of selenium in marine samples byatomie absorption speetrometry: Comparison of a flameless graphitefurnaee atomie absorption system with a hydride generation ato mieabsorption system. Speetroehim. Aeta. 37B:473.

4. NYGAARD,0.0. & J.H. LOWRY. 1982. Sample digestion proeeduresfor simultaneous determination of arsenic, antimony, and seleniumby inductively coupled argon plasma emission spectrometry withhydride generation. Anal. Chem. 54:803.

5. WELZ, B. & M. MELCHER. 1984. Meehanisms of transition metalinterferences in hydride generation atomic-absorption spectrometry.Part I. Influence of cobalt, copper, iron and nickel on seleniumdetermination. Analyst 109:569.

6. KR!VAN, V., K. PETRICK,B. WELZ & M. MELCHER. 1985. Radiotracererror-diagnostie investigation of selenium determination by hydride-generation atomic absorption spectrometry involving treatment withhydrogen peroxide and hydrochloric aeid. Anal. Chem. 57: 1703.

7. CHU, R.C., O.P. BARRON & P.A.W. BAUMGARNER.1972. Arsenicdetermination at submicrogram levels by arsine evolution and f1ame-less atomic absorption spectrophotometric technique. Anal. Chem.44:1476.

8. SIEMER,0.0., P. KOTEEL& V. JARJWALA.1976. Optimization of arsinegeneration in atomie absorption arsenie determinations. Anal. Chem.48:836.

FERNANDEZ,FJ. & O.C. MANNING.1971. The determination of arsenie atsubmicrogram levels by atomic absorption spectrophotometry.Atomie Absorption Newsletter 10:86.

JARREL-AsHCORPORATION.1971. High Sensitivity Arsenie Oeterminationby Atomic Absorption. larrel-Ash Atomie Absorption ApplieationsLaboratory Buli. No. As-3.

MANNING, O.C. 1971. A high sensitivity arsenic-selenium samplingsystem for atomie absorption speetroscopy. Atomie AbsorptionNewsletter 10: 123.

BRAMAN,R.S. & C.c. FOREBACK.1973. Methylated forms of arsenic inthe environrnent. Seienee 182: 1247.

CALDWELL,J.S., RJ. LISHKA& E.F. McFARREN. 1973. Evaluation of alow-eost arsenic and selenium determination of microgram-per-literlevels. J. Amer. Water Works Assoe., 65:731.

AGGETT,J. & A.C. ASPELL. 1976. The determination of arsenic (III) andtotal arsenic by atomie-absorption spectroscopy. Analyst 101 :341.

FJüRINO, J.A., J.W. JONES& S.O. CAPAR. 1976. Sequential determinationof arsenic, selenium, antimony, and tellurium in foods via rapidhydride evolution and atomic absorption spectrometry. Anal. Chem.48:120.

CUITER,O.A. 1978. Species determination of selenium in natural waters.Ana!. Chem. Acta 98:59.

OODDEN,R.O. & D.R. THOMERSON.1980. Oeneration of covalent hy-drides in atomic absorption spectroscopy. Analyst 105:1137.

BROWN,R.M., lR., R.C. FRY,l.L. MOYERS,S.J. NORTHWAY,M.B. DENToN& O.S. WILSON.1981. Interference by volatile nitrogen oxides andtransition-metal catalysis in the preconcentration of arsenic andselenium as hydrides. Ana!. Chem. 53: 1560.

SINEMUS,H.W., M. MELCHER& B. WELZ..1981. Intluence of valencestate on the determination of antimony, bismuth, selenium, andtellurium in lalce water using the hydride AA technique. AtomicSpectrosc. 2:81.

DEDINA,l. 1982. Interference of volatile hydride forming e1ements inselenium determination by atomie absorption spectrometry withhydride generation. Ana!. Chem. 54:2097.

The continuous hydride generator offers the advantages ofsimplicity in operation, excellent reproducibility, low detectionlimits, and high sample volume throughput for selenium analysisfollowing preparations as described in 3S00-Se.B or 3114BAcand d.

a. Principie: See Section 3114B.b. lnterferences: Free chlorine in hydrochloric acid is a com-

mon but difficult-to-diagnose interference. (The amount of chlo-rine varies with manufacturer and with each lot from the samemanufacturer.) Chlorine oxidizes the hydride and can contami-nate the hydride generator to prevent recoveries under any con-ditions. When interference is encountered, or preferably beforeusing each new bottle of HCI, eliminate chlorine from a 2.3-Lbottle of cone HCI by bubbling with helium (commercial grade,100 rnL/min) for 3 h.

Excess oxidant (peroxide, persulfate, or permanganate) fromthe total selenium digestion can oxidize the hydride. Followprocedures in 3S00-Se.B.2, 3, or 4 to ensure removal of alloxidizing agents before hydride generation.

Nitrite is a common trace constituent in natural and wastewaters, and at levels as low as 10 fLglL nitrite can reduce therecovery of hydrogen selenide from Se(IV) by over SO%. More-over, during the reduction of Se (VI) to Se(IV) by digestion withHCI (3S00-Se.B.S), some nitrate is converted to nitrite, whichsubsequently interferes. When this interference is suspected, addsulfanilamide after sample acidification (or HCI digestion). Thediazotization reaction between nitrite and sulfanilamide com-pletely removes the interferent effect (i.e., the standard additionslope is normal).

a. Continuous hydride generator: The basic unit is composedof two parts: a precision peristaltic pump, which is used to meterand mix reagents and sample solutions, and the gas-liquid sep-arator. At the gas-liquid separator a constant flow of argon stripsout the hydrogen and metal hydride gases formed in the reactionand carries them to the heated quartz absorption cell (3114B.lband 2b), which is supported by a metal bracket mounted on topof the regular air acetylene burner head. The spent liquid flows

out of the separator via a constant leveI side drain to a wastebucket. Schematics and operating parameters are shown in Fig-ure 3114:2.

Check flow rates frequently to ensure a steady flow; anuneven flow in any tubing will cause an erratic signal. Re-move tubings from pump rollers when not in use. Typical flowrates are: sample, 7 mUmin; acid, I mUmin; borohydridereagent, I mUmin. Argon flow usually is pre-fixed, typicallyat 90 mUmin.

b. Atomic absorption spectrometric equipment: See Section3111A.6.

a. Hydrochloric acid, HCI, S + I: Handle cone HCI under afume hood. li neeessary, remove free Cl2 by stripping cone HCIwith helium as deseribed above.

b. Borohydride reagent: Dissolve 0.6 g NaBH4 and O.S gNaOH in 100 mL water. CAUTION:Sodium borohydride is toxic,flammable, and corrosive.

c. Selenium reference standard solution, 1000 mglL: Useeommercially available standard; verify that selenium is Se(IV).

d. lntermediate standard solution, I mglL: Dilute 1 mL ref-erence standard solution to 1 L in a volumetric flask withdistilled water.

e. Working standard solutions, 5, 10, 20, 30, and 40 J.LglL:Dilute 0.5, 1.0, 2.0, 3.0, and 4.0 mL intermediate standardsolution to 100 mL in a volumetric flask.f Sulfanilamide solution: Prepare a 2.5% (w/v) solution daily;

add several drops cone HCI per 50 mL solution to facilitatedissolution.

a. Samp!e preparation: See Section 3500-Se or 3114BAc andd for preparation steps for various Se fractions or total Se.

b. Preconditioning hydride generator: For newly installedtubing, tum on pump for at least 10 to 15 min before instrumentcalibration. Sample the highest standard for a few minutes to letvolatile hydride react with the reactive sites in the transfer linesand on the quartz absorption cell surfaces.

c. lnstrument calibration: Depending on total void volume insampJe tubing, sampling time of 15 to 20 s generally is sufficientto obtain a steady signal. Between samples, submerge uptaketube in rinse water. Calibrate instrument daily after a 45-minlamp warmup time. Use either the hollow cathode or the elec-trodeless discharge lamp.

d. Antifoaming agents: Certain samples, particularly waste-water samples containing a high concentration of proteinaceoussubstances, can cause excessive foaming that could carry theliquid directly into the heated quartz absorption cell and causesplattering of salty deposits onto the windows of the spectrom-eter. Add a drop of antifoarning agent* to eliminate this problem.

e. Nitrite remova!: After samples have been acidified, or afteracid digestion, add 0.1 mL sulfanilamide solution per 10 mLsample and let react for 2 min.f Ana!ysis: Follow manufacturer's instructions for operation

of analytical equipment.

5. Calculation

Construct a calibration curve based on absorbance vs. standardconcentration. Apply dilution factors on diluted samples.

6. Precision and Bias

Working standards were analyzed together with batches ofwater samples on a routine production basis. The standards were

compounded using chemically pure sodium selenite and sodiumselenate. The values of Se(IV) + Se(VI) were determined byconverting Se(VI) to Se(IV) by digestion with HCI. Results aretabulated below.

No. Mean Se (IV) ReI. Dev. Se (IV) + Se(VI) ReI. DeI.Analyses J.LgIL % J.LgIL %

21 4.3 12 10.3 726 8.5 12 19.7 622 17.2 7 39.2 820 52.8 5 106.0 6

REAMER,D.C. & C. VElLOON. 1981. Preparation of biological materiaisfor determination of selenium by hydride generation-AAS. Anal.Chem. 53:1192.

SINEMUS, H.W., M. MELCHER & B. WELZ. 1981. lnfluence of valencestate on the determination of antirnony, bisrnuth, selenium, andtellurium in lake water using the hydride AA technique. AtomicSpectrosc. 2:81.

RODEN, D.R. & D.E. TALLMAN. 1982. Deterrnination of inorganic sele-niurn species in groundwaters containing organic interferences byion chromatography and hydride generationlatomic absorptionspectrometry. Ana!. Chem. 54:307.

CUITER, G. 1983. Elimination of nitrite interference in the determinationof selenium by hydride generation. Anal. Chim. Acta 149:391.

NARASAKI,H. & M. IKEDA. 1984. Automated deterrnination of arsenicand selenium by atomic absorption spectrometry with hydride gen-eration. Ana!. Chem. 56:2059.

WELZ, B. & M. MELCHER. 1985. Decomposition of marine biologicaltissues for determination of arsenic, selenium, and mercury usinghydride-generation and cold-vapor atomic absorption spectrome-tries. Ana!. Chem. 57:427.

EBDoN, L. & S.T. SPARKES.1987. Determination of arsenic and seleniumin environmental samples by hydride generation-direct current plas-ma-atomic emission spectrometry. Microchem. J. 36: 198.

EBDoN, L. & I.R. WrLKINSON. 1987. The determination of arsenic andselenium in coal by continuous flow hydride-generation atomicabsorption spectroscopy and atomic fluorescence spectrometry.Ana!. Chim. Acta. 194: 177.

VOTH-BEACH,L.M. & D.E. SHRADER.1985. Reduction of interferences inthe deterrnination of arsenic and selenium by hydride generation.Spectroscopy 1:60.

Emission spectroscopy using inductively coupled plasma(ICP) was developed in the mid-1960'sl.2 as a rapid, sensitive,

and convenient method for the determination of metaIs in waterand wastewater samples.3-6 Dissolved metaIs are determined infiltered and acidified samples. Total metaIs are determined afterappropriate digestion. Care must be taken to ensure that potentialinterferences are dealt with, especially when dissolved solidsexceed 1500 mglL.

I. GREENFlELD,S., 1.L. IONES & C.T. BERRY.1964. High-pressure plas-ma-spectroscopic emission sources. Analyst 89: 713.

2. WENDT,R.H. & V.A. FASSEL.1965. Induction-coupled plasma spec-trometric excitation source. Ana!. Chem. 37:920.

3. D.S. ENVIRONMENTALPROTECTIONAGENCY.1994. Method 200.7. In-ductively coupled plasma-atomic emission spectrometric method fortrace element analysis of water and wastes. Methods for the Deter-mination of Metais in Environmental Samples-Supplement I. EPA600/R-94-111, May 1994.

4. AMERICANSOCIETYFORTESTINGANDMATERIALS.1987. Annua1 Bookof ASTM Standards, Vol. 11.01. American Soe. Testing & Materials,Philadelphia, Pa.

5. FISHMAN,M.I. & W.L. BRADFORD,eds. 1982. A Supplement to Meth-ods for the Determination of Inorganic Substances in Water andFluvial Sediments. Rep. No. 82-272, D.S. Geological Survey, Wash-ington, D.e.

6. GARBARINO,I.R. & H.E. TAYLoR.1985. Trace Analysis. Recent De-velopments and Applications of Inductively Coupled Plasma Emis-sion Spectroscopy to Trace Elemental Analysis of Water. Volume 4.Academic Press, New York, N.Y.

a. Principie: An ICP source consists of a flowing stream ofargon gas ionized by an applied radio frequency field typi-cally oscillating at 27.1 MHz. This field is inductively cou-pled to the ionized gas by a water-cooled coil surrounding aquartz "torch" that supports and confines the plasma. A sam-pIe aerosol is generated in an appropriate nebulizer and spraychamber and is carried into the plasma through an injectortube located within the torch. The sample aerosol is injecteddirectly into the ICP, subjecting the constituent atoms totemperatures of about 6000 to SOOO°K. I Because this resultsin almost complete dissociation of molecules, significant re-duction in chemical interferences is achieved. The high tem-perature of the plasma excites atomic emission efficiently.lonization of a high percentage of atoms produces ionicemission spectra. The ICP provides an optically "thin" sourcethat is not subject to self-absorption except at very highconcentrations. Thus linear dynamic ranges of four to sixorders of magnitude are observed for many elements.z

The efficient excitation provided by the ICP results in lowdetection limits for many elements. This, coupled with the ex-tended dynamic range, perrnits effective multielement determi-nation of metals.3 The light emitted from the ICP is focused ontothe entrance slit of either a monochromator or a polychromatorthat effects dispersion. A precisely aligned exit slit is used toisolate a portion of the emission spectrum for intensity measure-ment using a photomultiplier tube. The monochromator uses asingle exit slit/photomultiplier and may use a computer-con-trolled scanning mechanism to examine emission wavelengthssequentially. The polychromator uses multiple fixed exit slits andcorresponding photomultiplier tubes; it simultaneously monitorsall configured wavelengths using a computer-controlled readoutsystem. The sequential approach provides greater wavelengthselection while the simultaneous approach can provide greatersample throughput.

b. Applicable metais and analyticallimits: Table 3120:1 listselements for which this method applies, recommended analyticalwavelengths, and typical estimated instrument detection levelsusing conventional pneumatic nebulization. Actual working de-tection levels are sample-dependent. Typical upper limits forlinear calibration also are included in Table 3120:1.

c. Interferences: lnterferences may be categorized as follows:1) Spectral interferences-Light emission from spectral

sources other than the element of interest may contribute toapparent net signal intensity. Sources of spectral interferenceinclude direct spectral line overlaps, broadened wings of intensespectral lines, ion-atom recombination continuum emission, mo-lecular band emission, and stray (scattered) light from the emis-sion of elements at high concentrations.4 Avoid line overlaps byselecting altemate analytical wavelengths. Avoid or minimizeother spectral interference by judicious choice of backgroundcorrection positions. A wavelength scan of the element lineregion is useful for detecting potential spectral interferences andfor selecting positions for background correction. Make correc-tions for residual spectral interference using empirically deter-mined correction factors in conjunction with the computer soft-ware supplied by the spectrometer manufacturer or with thecalculation detailed below. The empirical correction methodcannot be used with scanning spectrometer systems if the ana-Iytical and interfering lines cannot be precisely and reproduciblylocated. ln addition, if using a polychromator, verify absence ofspectral interference from an element that could occur in asample but for which there is no channel in the detector array. Dothis by analyzing single-element solutions of 100 mg/L concen-tration and noting for each element channel the apparent con-centration from the interfering substance that is greater than theelement's instrument detection limit.

2) Nonspectral interferencesa) Physical interferences are effects associated with sample

nebulization and transport processes. Changes in the physicalproperties of samples, such as viscosity and surface tension, cancause significant error. This usually occurs when samples con-taining more than 10% (by volume) acid or more than 1500 mgdissolved solids/L are analyzed using calibration standards con-taining ~ 5% acid. Whenever a new or unusual sample matrix isencountered, use the test described in CJ( 4g. If physical interfer-ence is present, compensate for it by sample dilution, by usingmatrix-matched calibration standards, or by applying the methodof standard addition (see CJ( 5d below).

High dissolved solids content also can contribute to instru-mental drift by causing salt buildup at the tip of the nebulizer gasorifice. Using prehumidified argon for sample nebulization less-ens this problem. Better control of the argon flow rate to the

3-40 METALS (3000)

TABLE 3120:1. SUGGESTEDWAVELENGTHS,ESTIMATEDDETECTIONLEVELS, ALTERNATEWAVELENGTHS,CALlBRATIONCONCENTRATlONS,ANDUPPER LIMITS

EstimatedSuggested Detection Alternate Calibration Upper Limit

Wavelength LeveI Wavelength* Concentration ConcentrationtElement 11m /LgIL 11m mglL mglL

Aluminum 308.22 40 237.32 10.0 100Antimony 206.83 30 217.58 10.0 100Arsenic 193.70 50 189.04t 10.0 100Barium 455.40 2 493.41 1.0 50Beryllium 313.04 0.3 234.86 1.0 10Boron 249.77 5 249.68 1.0 50Cadmium 226.50 4 214.44 2.0 50Calcium 317.93 10 315.89 10.0 100Chromium 267.72 7 206.15 5.0 50Cobalt 228.62 7 230.79 2.0 50Copper 324.75 6 219.96 1.0 50lron 259.94 7 238.20 10.0 100Lead 220.35 40 217.00 10.0 100Lithium 670.78 4§ 5.0 100Magnesium 279.08 30 279.55 10.0 100Manganese 257.61 2 294.92 2.0 50Molybdenum 202.03 8 203.84 10.0 100Nickel 231.60 15 221.65 2.0 50Potassium 766.49 100§ 769.90 10.0 100Selenium 196.03 75 203.99 5.0 100Silica (Si02) 212.41 20 251.61 21.4 100Silver 328.07 7 338.29 2.0 50Sodium 589.00 30§ 589.59 10.0 100Strontium 407.77 0.5 421.55 1.0 50Thallium 190.86t 40 377.57 10.0 100Vanadium 292.40 8 1.0 50Zinc 213.86 2 206.20 5.0 100

* Olher wavelenglhs may be substituted if lhey provide lhe needed sensitivily and are correcled for spectral inlerference.t Defines the 10p end of the effective calibration range. Do not extrapolate to concentrations beyond highest standard.t Available with vacuum or inert gas purged oplical path.§ Sensitive 10 operating conditions.

nebulizer using a mass flow eontroller improves instrumentperformance.

b) Chemical interferences are eaused by molecular compoundformation, ionization effects, and thermochemical effeets asso-ciated with sample vaporization and atomization in the plasma.Normally these effects are not pronouneed and can be minimizedby careful selection of operating conditions (incident power,plasma observation position, ete.). Chemical interferences arehigWy dependent on sample matrix and element of interest. Aswith physicaJ interferences, compensate for them by using ma-trix matched standards or by standard addition (lJISd). To deter-mine the presence of chemicaJ interference, follow instructionsin lJI4g.

a. ICP source: The ICP source eonsists of a radio frequency(RF) generator capable of generating at least 1.1 kW of power,torch, tesla coil, load coil, impedanee matching network, nebu-lizer, spray chamber, and drain. High-quality flow regulators arerequired for both the nebulizer argon and the plasma support gasflow. A peristaltic pump is recommended to regulate sample flowto the nebulizer. The type of nebulizer and spray chamber used

may depend on the samples to be analyzed as well as on theequipment manufacturer. In general, pneumatic nebulizers of theconcentric or cross-flow design are used. Viscous samples andsamples containing particulates or high dissolved solids content(>Sooo mg/L) may require nebulizers of the Babington type.5

b. Spectrometer: The speetrometer may be of the simulta-neous (polychromator) or sequential (monochromator) type withair-path, inert gas purged, or vacuum optics. A spectral bandpassof O.OS nm or less is required. The instrument should permitexamination of the spectral baekground surrounding the emis-sion lines used for metais determination. It is necessary to beable to measure and correct for spectral background at one ormore positions on either side of the analytical lines.

Use reagents that are of ultra-high-purity grade or equivalent.Redistilled acids are aeceptable. Except as noted, dry all salts atlOsoC for 1 h and store in a desiceator before weighing. Usedeionized water prepared by passing water through at least twostages of deionization with mixed bed cation!anion exchangeresins.6 Use deionized water for preparing all calibration stan-dards, reagents, and for dilution.

a. Hydrochloric acid, HCI, conc and 1+ 1.b. Nitric acid, HN03, conc.c. Nitric acid, HN03, 1+ 1: Add 500 rnL conc HN03 to 400

rnL water and dilute to 1 L.d. Standard stock solutions: See 3111B, 3111D, and 3114B.

CAUTION: Many metal salts are extremely toxic and may be fatalif swallowed. Wash hands thoroughly afier handling.

1) Aluminum: See 3111D.3kl).2) Antimony: See 311lB.3jl).3) Arsenic: See 3114B.3kl).4) Barium: See 311lD.3k2).5) Beryllium: See 311lD.3k3).6) Boron: Do not dry but keep bottle tightly stoppered and

store in a desiccator. Dissolve 0.5716 g anhydrous H3B03 inwater and dilute to 1000 rnL; 1 rnL = 100 fLg B.

7) Cadmium: See 3111B.3j3).8) Calcium: See 3111B.3j4).9) Chromium: See 3111B.3j6).10) Cobalt: See 3111B.3j7).11) Copper: See 311lB.3j8).12) lron: See 311IB.3jll).13) Lead: See 311lB.3jI2).14) Lithium: See 3111B.3jI3).15) Magnesium: See 311lB.3jI4).16) Manganese: See 3111B.3jI5).17) Molybdenum: See 311lD.3k4).18) Nickel: See 3111B.3jI6).19) Potassium: See 3111B.3jI9).20) Selenium: See 3114B.3nl).21) Silica: See 311lD.3k7).22) Si/ver: See 3111B.3j22).23) Sodium: See 3111B.3j23).24) Strontium: See 3111B.3j24).25) Thallium: See 3111B.3j25).26) Vanadium: See 311lD.3klO).27) Zinc: See 3111B.3j27).e. Calibration standards: Prepare mixed calibration standards

containing the concentrations shown in Table 3120:1 by com_obining appropriate volumes of the stock solutions in 100-rnLvolumetric flasks. Add 2 rnL 1+1 HN03 and 10 mL 1+1 HCIand dilute to 100 mL with water. Before preparing mixed stan-dards, analyze each stock solution separately to determine pos-sible spectral interference or the presence of impurities. Whenpreparing mixed standards take care that the elements are com-patible and stable. Store mixed standard solutions in an FEPfluorocarbon or unused polyethylene bottle. Verify calibrationstandards initially using the quality control standard; monitorweek1y for stability. The following are recommended combina-tions using the suggested analytical lines in Table 3120:1. Alter-native combinations are acceptable.

1) Mixed standard solution I: Manganese, beryllium, cad-mium, lead, selenium, and zinco

2) Mixed standard solution II: Barium, copper, iron, vana-dium, and cobalt.

3) Mixed standard solution III: Molybdenum, silica, arsenic,strontium, and lithium.

4) Mixed standard solution IV: Calcium, sodium, potassium,aluminum, chromium, and nickel.

5) Mixed standard solution V: Antimony, boron, magne-sium, silver, and thallium. If addition of silver results in an

initial precipitation, add 15 mL water and warm flask untilsolution clears. Cool and di lute to 100 mL with water. For thisacid combination limit the silver concentration to 2 mg/L.Silver under these conditions is stable in a tap water matrixfor 30 d. Higher concentrations of silver require additionalHCI.f Calibration blank: Dilute 2 rnL 1+ 1 HN03 and 10 rnL 1+ 1

HCI to 100 rnL with water. Prepare a sufficient quantity to beused to flush the system between standards and samples.

g. Method blank: Carry a reagent blank through entire samplepreparation procedure. Prepare method blank to contain the sameacid types and concentrations as the sample solutions.

h. lnstrument check standard: Prepare instrument check stan-dards by combining compatible elements at a concentration of 2mglL.

i. lnstrument quality control sample: Obtain a certified aque-ous reference standard from an outside source and prepare ac-cording to instructions provided by the supplier. Use the sameacid matrix as the calibration standards.

j. Method quality control sample: Carry the instrument qual-ity control sample (<j[ 3i) through the entire sample preparationprocedure.

k. Argon: Use technical or welder's grade. lf gas appears to bea source of problems, use prepurified grade.

a. Sample preparation: See Section 3030F.b. Operating conditions: Because of differences among

makes and models of satisfactory instruments, no detailed oper-ating instructions can be provided. Follow manufacturer' s ir1-structions. Establish instrumental detection limit, precision, op-timum background correction positions, linear dynamic range,and interferences for each analyticalline. Verify that the instru-ment configuration and operating conditions satisfy the analyti-cal requirements and that they can be reproduced on a day-to-daybasis. An atom-to-ion emission intensity ratio [Cu(I) 324.75nrnlMn(lI) 257.61 nm] can be used to reproduce optimum con-ditions for multielement analysis precisely. The CulMn intensityratio may be incorporated into the calibration procedure, includ-ing specifications for sensitivity and for precision? Keep daily orweekly records of the Cu and Mn intensities and/or the intensi-ties of critical element lines. AIso record settings for opticalalignment of the polychromator, sample uptake rate, powerreadings (incident, reflected), photomultiplier tube attenuation,mass flow controller settings, and system maintenance.

c. lnstrument calibration: Set up instrument as directed (<j[ b).Warm up for 30 mino For polychromators, perform an opticalalignment using the profile lamp or solution. Check alignment ofplasma torch and spectrometer entrance slit, particularly if main-tenance of the sample introduction system was performed. MakeCulMn or similar intensity ratio adjustment.

Calibrate instrument according to manufacturer's recom-mended procedure using calibration standards and blank. Aspi-rate each standard or blank for a minimum of 15 s after reachingthe plasma before beginning signal integration. Rinse with cal-ibration blank or similar solution for at least 60 s between eachstandard to eliminate any carryover from the previous standard.Use average intensity of multiple integrations of standards orsamples to reduce random error.

Before analyzing samples, analyze instrument check standard.Concentration values obtained should not deviate from the actualvalues by more than ±5% (or the established control Iimits,whichever is lower).

d. Analysis of samples: Begin each sample run with ananalysis of the calibration blank, then analyze the methodblank. This permits a check of the sample preparation re-agents and procedures for contamination. Analyze samples,alternating them with analyses of calibration blank. Rinse forat least 60 s with dilute acid between samples and blanks.After introducing each sample or blank let system equilibratebefore starting signal integration. Examine each analysis ofthe calibration blank to verify that no carry-over memoryeffect has occurred. If carry-over is observed, repeat rinsinguntil proper blank values are obtained. Make appropriatedilutions and acidifications of the sample to determine con-centrations beyond the linear calibration range.

e. Instrumental quality control: Analyze instrument checkstandard once per 10 samples to determine if significantinstrument drift has occurred. If agreement is not within ±5% of the expected values (or within the established controllimits, whichever is lower), terminate analysis of samples,correct problem, and recalibrate instrumento If the intensityratio reference is used, resetting this ratio may restore cali-bration without the need for reanalyzing calibration standards.Analyze instrument check standard to confirm proper recali-bration. Reanalyze one or more samples analyzed just beforetermination of the analytical run. Results should agree towithin ± 5%, otherwise ali samples analyzed after the lastacceptable instrument check standard analysis must be rean-alyzed.

Analyze instrument quality control sample within every run.Use this analysis to verify accuracy and stability of the calibra-tion standards. If any result is not within ± 5% of the certifiedvalue, prepare a new calibration standard and recalibrate theinstrumento If this does not correct the problem, prepare a newstock solution and a new calibration standard and repeat calibra-tion.f Method quality control: Analyze the method quality control

sample within every run. Results should agree to within ± 5% ofthe certified values. Greater discrepancies may reflect losses orcontamination during sample preparation.

g. Test for matrix interference: When analyzing a new orunusual sample matrix verify that neither a positive nornegative nonlinear interference effect is operative. If theelement is pres- ent at a concentration above 1 mg/L, useserial dilution with calibration blank. Results from the anal-yses of a dilution should be within ± 5% of the original resultoAlternately, or if the concentration is either below 1 mg/L ornot detected, use a post-digestion addition equal to 1 mg/L.Recovery of the addition should be either between 95% and105% or within established control Iimits of ± 2 standarddeviations around the mean. If a matrix effect causes testresults to fali outside the criticallimits, complete the analysisafter either diluting the sample to eliminate the matrix effectwhile maintaining a detectable concentration of at least twicethe detection Iimit or applying the method of standard addi-tions.

a. Blank correction: Subtract result of an adjacent calibrationblank fram each sample result to make a baseline drift correc-tion. (Concentrations printed out should include negative andpositive values to compensate for positive and negative baselinedrift. Make certain that the calibration blank used for blankcorrection has not been contaminated by carry-over.) Use theresult of the method blank analysis to correct for reagent con-tamination. Alternatively, intersperse method blanks with appro-priate samples. Reagent blank and baseline drift correction areaccomplished in one subtraction.

b. Dilution correction: If the sample was diluted or concen-trated in preparation, multiply results by a dilution factor (DF)calculated as follows:

Finalweightor volumeDF = --------

Initialweightor volume

c. Correction for spectral inteiference: Correct for spectralinterference by using computer software supplied by the instru-ment manufacturer or by using the manual method based oninterference correction factors. Determine interference correc-tion factors by analyzing single-e1ement stock solutions of ap-propriate concentrations under conditions matching as c10sely aspossible those used for sample analysis. Unless analysis condi-tions can be reproduced accurately from day to day, or for longerperiods, redetermine interference correction factors found toaffect the results significantly each time samples are analyzed.7• 8

Calculate interference correction factors (K;) from apparent con-centrations observed in the analysis of the high-purity stocksolutions:

Apparentconcentrationof elementiK=-------------'J Actualconcentrationof interferingelementj

wbere the apparent concentration of element i is the differ-ence between the observed concentration in the stock solutionand the observed concentration in the blank. Correct sampleconcentrations observed for element i (already corrected forbaseline drift), for spectral interferences fram elements j, k,and l; for example:

Observedconcentration - (K;)

of i

Observedconcentrationof interfering

elementj

Observedconcentrationof interfering

elementk

Observedconcentrationof interfering

elementI

Interference correction factors may be negative if backgroundcorrection is used for element i. A negative Kij can resultwhere an interfering line is encountered at the backgroundcorrection wavelength rather than at the peak wavelength.Determine concentrations of interfering elements j, k. and lwithin their respective linear ranges. Mutual interferences (i

ConcentrationRange Total Digestion*

Element fLg/L fLg/L

Aluminum 69--4792 X 0.9273C + 3.6S 0.0559X + 18.6SR = 0.0507X + 3.5

Antimony 77-1406 X = 0.7940C - 17.0S = 0.1556X - 0.6SR = 0.1081X + 3.9

Arsenic 69-1887 X = 1.0437C - 12.2S = 0.1239X + 204SR = 0.0874X + 604

Barium 9-377 X = 0.7683C + 0047S = 0.1819X + 2.78SR = 0.1285X + 2.55

Beryllium 3-1906 X = 0.9629C + 0.05S = 0.0136X + 0.95SR = 0.0203X - 0.07

Boron 19-5189 X = 0.8807C + 9.0S = 0.1150X + 14.1SR = 0.0742X + 23.2

Cadmium 9-1943 X = 0.9874C - 0.18S = 0.0557X + 2.02SR= 0.0300X + 0.94

Calcium 17--47 170 X = 0.9182C - 2.6S = 0.1228X + 10.1SR = 0.0189X + 3.7

Chromium 13-1406 X = 0.9544C + 3.1S = 0.0499X + 404SR = 0.0009X + 7.9

Cobalt 17-2340 X = 0.9209C - 4.5S = 0.0436X + 3.8SR= 0.0428X + 0.5

Copper 8-1887 X = 0.9297C -0.30S = 0.0442X + 2.85SR = 0.0128X + 2.53

Iron 13-9359 X = 0.8829C + 7.0S = 0.0683X + 11.5SR = -0.0046X + 10.0

Lead 42--4717 X = 0.9699C - 2.2S = 0.0558X + 7.0SR = 0.0353X + 3.6

Magnesium 34-13 868 X = 0.9881C - 1.1S = 0.0607X + 11.6SR = 0.0298X + 0.6

Manganese 4-1887 X = 0.9417C + 0.13S = 0.0324X + 0.88SR = 0.0153X + 0.91

Molybdenum 17-1830 X = 0.9682C + 0.1S = 0.0618X + 1.6SR= 0.0371X + 2.2

Nickel 17--47 170 X = 0.9508C + 004S = 0.0604X + 404SR = 0.0425X + 3.6

Potassium 347-14 151 X = 0.8669C - 3604S = 0.0934X + 77.8SR = -0.0099X + 144.2

Selenium 69-1415 X = 0.9363C - 2.5S = 0.0855X + 17.8SR = 0.0284X + 9.3

Recoverable Digestion*fLg/L

X = 0.9380C + 22.1S = 0.0873X + 31.7SR = 0.0481X + 18.8X = 0.8908C + 0.9S = 0.0982X + 8.3SR = 0.0682X + 2.5X = 1.0175C + 3.9S = 0.1288X + 6.1SR = 0.0643X + 10.3X = 0.8380C + 1.68S = 0.2540X + 0.30SR = 0.0826X + 3.54X = 1.01 77 C - 0.55S = 0.0359X + 0.90SR = 0.0445X - 0.10X = 0.9676C + 18.7S = 0.1320X + 16.0SR = 0.0743X + 21.1X = 1.0137C - 0.65S = 0.0585X + 1.15SR = 0.0332X + 0.90X = 0.9658C + 0.8S = 0.0917X + 6.9SR = 0.0327X + 10.1X = 1.0049C - 1.2S = 0.0698X + 2.8SR = 0.0571X + 1.0X = O.9278C - 1.5S = 0.0498X + 2.6SR = 0.0407X + 004X = 0.9647C - 3.64S = 0.0497X + 2.28SR = 0.0406X + 0.96X = 0.9830C + 5.7S = 0.1024X + 13.0SR = 0.0790X + 11.5X = 1.0056C + 4.1S = 0.0799X + 4.6SR = 0.0448X + 3.5X = 0.9879C + 2.2S = 0.0564X + 13.2SR = 0.0268X + 8.1X = 0.9725C + 0.07S = 0.0557X + 0.76SR = 0.0400X + 0.82X = 0.9707C - 2.3S = 0.0811X + 3.8SR = 0.0529X + 2.1X = 0.9869C + 1.5S = 0.0526X + 5.5SR = 0.0393X + 2.2X = 0.9355C - 183.1S = 0.0481X + 177.2SR = 0.0329X + 60.9X = 0.9737C - 1.0S = 0.1523X + 7.8SR = 0.0443X + 6.6

ConcentrationRangeJ.LgIL

189-9434

Total Digestion*J.LglL

X = 0.5742C - 35.6S = 0.4160X + 37.8SR = 0.1987X + 8.4X = 0.4466C + 5.07S = 0.5055X - 3.05SR = 0.2086X - 1.74X = 0.9581C + 39.6S = 0.2097X + 33.0SR = 0.0280X + 105.8X = 0.9020C - 7.3S = O.I004X + 18.3SR = 0.0364X + 11.5X = 0.9615C - 2.0S = 0.0618X + 1.7SR = 0.0220X + 0.7X = 0.93-6C - 0.30S = 0.0914X + 3.75SR =-0.0130X + 10.07

Recoverable Digestion*J.LgIL

X = 0.9737C - 60.8S = 0.3288X + 46.0SR = 0.2133X + 22.6X = 0.3987C + 8.25S = 0.5478X - 3.93SR = 0.1836X - 0.27X = 1.0526C + 26.7S = 0.1473X+27.4SR = 0.0884X + 50.5X = 0.9238C + 5.5S = 0.2156X + 5.7SR =-0.0106X + 48.0X = 0.9551C + 0.4S = 0.0927X + 1.5SR = 0.0472X + 0.5X = 0.9500C + 1.22S = 0.0597X + 6.50SR = 0.0153X + 7.78

*X = mean recovery, !!g/L,C = true value, !!gIL,S = multi-laboratory standard deviation, !!gIL,

SR = single-analyst standard deviation, !!g/L.

interferes with j and j interferes with i) require iterative ormatrix methods for calculation.

d. Correction for nonspectral inteiference: If nonspectral in-terference correction is necessary, use the method of standardadditions. It is applicable when the chemical and physicalform of the element in the standard addition is the same as inthe sample, or the lep converts the metal in both sampleand addition to the same form; the interference effect isindependent of metal concentration over the concentrationrange of standard additions; and the analytical calibrationcurve is linear over the concentration range of standard ad-ditions.

Use an addition not less than 50% nor more than 100% ofthe element concentration in the sample so that measurementprecision will not be degraded and interferences that dependon element/interferent ratios will not cause erroneousresults. Apply the method to all elements in the sampleset using background correction at carefully chosen off-linepositions. Multielement standard addition can be used ifit has been determined that added elements are not interfer-ents.

e. Reporting data: Report analytical data in concentrationunits of milligrams per liter using up to three significant figures.Report result below the detennined detection limit as not de-tected less than the stated detection lirrút corrected for sampledilution.

As a guide to the generally expected precision and bias, see thelinear regression equations in Table 3120:11.9 Additional inter-laboratory information is available.1O

I. FAlRES,L.M., B.A. PALMER,R. ENGLEMAN,IR. & T.M. NIEMCZYK.1984. Temperature determinations in the inductively coupledplasma using a Fourier transform spectrometer. Spectrochim. Acta39B:819.BARNES,R.M. 1978. Recent advances in emission spectroscopy:inductively coupled plasma discharges for spectrochemical analysis.CRC Cril. Rev. Ana!. Chem. 7:203.

3. PARSO5, M.L., S. MAJOR& A.R. FOR5TER.1983. Trace elementdetermination by atomic spectroscopic methods - State of the art.Appl. Spectrosc. 37:411.

4. LARSON,G.F., V.A. FASSEL,R. K. WINGE& R.N. KNISELEY.1976.Ultratrace analysis by optical emission spectroscopy: The stray lightproblem. App!. Spectrosc. 30:384.

5. GARBARINO,I.R. & H.E. TAYLOR.1979. A Babington-type nebulizerfor use in lhe analysis of natural water samples by inductivelycoupled plasma spectrometry. App!. Speetrosc. 34:584.

6. AMERJCANSOCIETYFORTESTINGANDMATERJALS.1988. Standardspecification for reagent water, D1l93-77 (reapproved 1983). An-nual Book of ASTM Standards. American Soe. Testing & Materiais,Philadelphia, Pa.

7. BOTTO,R.I. 1984. Quality assurance in operating a multielementICP ernission spectrometer. Spectrochim. Acta 39B:95.

8. Borro, R.I. 1982. Long-term stability of spectral interference cali-brations for inductively coupled plasma atornic emission spectrom-etry. Ana!. Chem. 54:1654.

9. MAXFJELD,R. & B. MINDAK.1985. EPA Method Study 27, Method200. 7 (Trace MetaIs by ICP). EPA-600/S4-85/05. National Tech-rucal lnformation Serv., Springfield, Va.

10. GARBARINO,I.R., B.E. IONEs, G. P. STEIN,W.T. BELSER& H.E.TAYLOR.1985. Statistical evaluation of an inductively coupledplasma atomic emission spectrometric method for routine waterquality testing. Appl. Spectrosc. 39:53.

This method is used for the determination of trace metaIsand metalloids in surface, ground, and drinking waters byinductively coupled plasma/mass spectrometry (ICP/MS). Itmay also be suitable for wastewater, soils, sediments, sludge,and biological samples after suitable digestion followed bydilution and/or cleanup.l,2 Additional sources of informationon quality assurance and other aspects of ICP/MS analysis ofmetaIs are available?-5

The method is intended to be performance-based, allowingextension of the elemental analyte list, implementation of"clean" preparation techniques as they become available, andother appropriate modifications of the base method as tech-nology evolves. Preferably validate modifications to the basemethod by use of the quality control standards specified in themethod.

Instrument detection levels for many analytes are between 1and 100 nglL. The method is best suited for the determination ofmetaIs in ambient or pristine fresh-water matrices. More com-plex matrices may require some type of cleanup to reduce matrixeffects to a manageable leveI. Various cleanup techniques areavailable to reduce matrix interferences and/or concentrate ana-lytes of interest,6-1O

This method is ideally used by analysts experienced in the useof ICP/MS, the interpretation of spectral and matrix interference,and procedures for their correction. Preferably demonstrate an-alyst proficiency through analysis of a performance evaluationsample before the generation of data.

* Approvedby StandardMethodsCommittee,1997.Joint Task Group: 20th Edition-William R. Kammin (chair), John R. Barnett,Isabel C. Chamberlain,Robert Henry, James O. Ross, Ruth E. Wolf, Cindy A.Ziemicki.

L MONTASER,A. & D.W. GOLlGHTLY,eds. 1992. Inductively CoupledPlasmas in Analytical Atomic Spectrometry, 2nd ed. VCH Publisb-ers, Inc., New York, N.Y.

2. DATE,AR. & AL. GRAY.1989. Applications of Inductively Coup1edPlasma Mass Spectrometry, BIackie & Son, Ltd., Glasgow, u.K.

3. V.S. ENVlRONMENTALPROTECTIONAGENCY.1994. Determination oftraceelements in waters and wastes by inductively coupled plasma-massspectrometry, Metbod 200.8. V.S. Environmental Protection Agency,Environmental Monitoring Systems Lab., Cincinnati, Oruo.

4. LONGBOTTOM,J.E., T.D. MARTIN,K.W. EDGELL,S.E. LONG,M.R.PLANTZ& B.E. WARDEN.1994. Determination of trace e1ements inwater by inductively coupled plasma-mass spectrometry: coIlabo-rative study. J. AOAC Internat. 77:1004.

5. V.S. ENVIRONMENTALPROTECTIONAGENCY.1995. Method 1638: De-termination of trace elements in ambient waters by inductivelycoupled plasma-mass spectrometry. V.S. Environmental ProtectionAgency, Off. Water, Washington, D.C.

6. McLAREN,J.W., AP. MYKYTlUK,S.N. WILLlE& S. S. BERMAN.1985.Determination of trace metaIs in seawater by inductively coupledplasma mass spectrometry with preconcentration on silica-immobi-lized 8-hydroxyquinoline. Ana!. Chem. 57:2907.

7. BURBA,P. & P.G. WILLMER.1987. Multielement preconcentrationfor atomic spectroscopy by sorption of dithiocarbamate metal com-plexes (e.g., HMDC) on cellulose collectors. Fresenius Z. Ana!.Chem, 329:539.

8. WANG,X. & R.M. BARNES.1989. Chelating resins for on-line flowinjection preconcentration with inductively coupled plasma atornicernission spectroscopy. J. Ana!. Atam. Spectram. 4:509.

9. SIRlRAKS,A., H.M. KINGSTON& J.M. RIVIELLO.1990. Chelation ionchromatography as a method for trace elemental analysis in com-plex environmental and biological samples. Anal. Chem. 62: 1185.

10. PUGETSOUNDWATERQUALlTYAUTHORITY.1996. RecommendedGuidelines for Measuring Metais in Puget Sound Marine Water,Sediment and Tissue Samples. Appendix D: Alternate Methods forthe Analysis of Marine Water Samples. Puget Sound Water QualityAuthority, Olympia, Wash.

a, Principie: Sample material is introduced into an argon-based, high-temperature radio-frequency plasma, usually bypneumatic nebulization. Energy transfer from the plasma to thesample stream causes desolvation, atomization, and ionization of

target elements. Ions generated by these energy-transfer pro-cesses are extracted from the plasma through a differentialvacuum interface, and separated on the basis of their mass-to-charge ratio by a mass spectrometer. The mass spectrometerusually is of the quadrupole or magnetic sector type. The ionspassing through the mass spectrometer are counted, usually by

TABLE3125:1. RECOMMENDEDANALYTEMASSES, INSTRUMENTALDETECTIONLEVELS(IDL), ANDINTERNALSTANDARDS

IDL*Mg/L

0.0250.030.020.040.030.0020.0020.0040.0250.0030.0040.0170.0200.0250.0930.0640.0030.0020.0060.0030.070.070.030.030.0050.0320.0010.003t0.008tO.OOU

ReeommendedInterna1Standard

* IDLs were detennined on a Perkin Elmer Elan 6000 ICPIMS using sevenreplicate analyses of a 1% nitric acid solution, at Manchester EnvironrnentalLaboratory, July 1996.t From EPA Melhod 200.8 for the Analysis ofDrinking Waters-Application Note,Order No. ENV A-300A, The Perkin Elmer Corporation, 1996.j: From Perkin Elrner Technical Summary TSMS-12.

an electron multiplier detector, and the resulting informationprocessed by a computer-based data-handling system.

b. Applicable elements and analytical limits: This method issuitable for aluminum, antimony, arsenie, barium, berylIium,cadmium, chromium, cobalt, copper, lead, manganese, molyb-denum, nickel, selenium, silver, strontium, thallium, uranium,vanadium, and zinco The method is also acceptable for otherelemental analytes as long as the same quality assurance prac-tices are folIowed. The basic element suíte and recommendedanalytical masses are given in Table 3125:1.

Typical instrument detection levels (IDL)I,2 for method ana-lytes are presented in Table 3125:1. Determine the IDL andmethod detection leveI (or limit) (MDL) for alI analytes beforemethod implementation. Section 1030 contains additional infor-mation and approaches for the evaluation of detection capabili-ties.

The MDL is defined in Section 1010C and elsewhere.2 Deter-mination of the MDL for each element is critical for complexmatrices such as seawater, brines, and industrial effluents, TheMDL wilI typicalIy be higher than the IDL, because of back-

ground analyte in metaIs preparation and analysis laboratoriesand matrix-based interferences. Determine both IDL and MDLupon initial implementation of this method, and then yearly orwbenever the instrument configuration changes or major main-tenance occurs, whichever comes first.

Determine linear dynamic ranges (LDR) for alI method ana-lytes, LDR is defined as the maximum concentration of analyteabove the highest calibration point where analyte response iswithin ::!: 10% of the theoretical response. When determininglinear dynamic ranges, avoid using unduly high concentrationsof analyte that might damage the detector. Determine LDR onmultielement mixtures, to account for possible interelement ef-fects. Determine LDR on initial implementation of this method,and then yearly,

C. Interferences: ICPIMS is subject to severa1 types of inter-ferences.

1) Isotopes of different elements that form ions of the samenominal mass-to-charge ratio are not resolved by the quadrupolemas spectrometer, and cause isobaric elemental interferences.TypicalIy, ICPIMS instrument operating software wilI have alI

TABLE 3125:II. ELEMENTALABUNDANCEEQUATIONSANDCOMMONMOLECULARION CORRECTIONEQUATIONS

Li 6Be 9Al 27Se 45V 51Cr 52Cr 53Mn 55Co 59i 60i 62

Cu 63Cu 65Zn 66Zn 68As 75Se 77Se 82Sr 88Mo 98Rh 103Ag 107Ag 109Cd 111Cd 114Sb 121Sb 123Ba 135Ho 165TI 203TI 205Pb 208Tb 232U 238

= C 6= C9= C27= C45= C 51 - (3.l27)((C 53) - (0.113 X C 52)]= C 52= C 53= C 55= C 59= C60= C62= C63= C65= C66= C68= C 75 - (3.127)((C 77) - (0.815 X C 82)]=C77= C 82 - (1.008696 X C 83)= C 88= C 98 - (0.110588 X C 101)= C 103= C 107= C 109= C 111 - (l.073)((C 108) - (0.712 X C 106)]= C 114 - (0.02686 X C 118)= C 121= C 123 - (0.l27189 X C 125)= C 135= C 165= C 203= C 205= C 208 + (1 X C 206) + (l X C 207)= C 232= C 238

* C = calibration blank corrected counts at indicated mass.t From EPA Method 200.8 for lhe Analysis of Drinking Waters - ApplicationNote, Order No. ENVA-3OOA, The Perkin Elmer Corporation, 1996.

TABLE 3125:III.COMMONMOLECULARlON lNrERFERENCESlN lCPIMS I

Element MeasurementMoleeular lon Mass Affeeted by lnterferenee

Baekground moleeular ions:NH+ 15OH+ 17OH + 18C1- 24 Mg2CN+ 26 MgCO+ 28 SiN + 28 Si2

NzH+ 29 SiNO+ 30

NOH+ 31 P0+ 32 Sz

OzH+ 3336ArH+ 37 Cl38ArH+ 39 K4oArH+ 41CO + 44 Ca

CO ~H 45 SezArC+, ArO+ 52 Cr

ArN+ 54 CrArNH+ 55 MnArO+ 56 FeArH+ 57 Fe

4°Ar36Ar+ 76 Se4oAr38Ar 78 Se4°Arz + 80 Se

Matrix moleeular ions:Bromide:

81BrH+ 82 Se79BrO+ 95 Mo81BrO+ 97 Mo

8IBrOH+ 98 MoAr81Br+ 121 Sb

Chloride:35C10+ 51 V

35CIOH+ 52 Cr37C10+ 53 Cr

37ClOH+ 54 CrAr35CI+ 75 AsA?7Cl+ 77 Se

Sulfate:32S0+ 48 Ti

32SOH+ 4934S0+ 50 V, Cr

34SOH+ 51 VSOz +, S2 + 64 Zn

Ar3ZS+ 72 GeAr34S+ 74 Ge

Phosphate:PO+ 47 Ti

POH+ 48 TiP02+ 63 CuArP+ 71 Ga

Group I & IImetaIs:ArNa+ 63 CuArK+ 79 BrArCa + 80 Se

Matrix oxides*TiO 62-66 Ni, Cu, ZnZrO 106-112 Ag,CdMoO 108-116 CdNbO 109 Ag

* Oxide interferences norrnally will be very small and will affect lhe melhod elementsonly when oxide-producing elements are present at relatively high concentrations, orwhen lhe instrurnent is improperly lUnedor maintaíned. Preferably monítor Ti and Zrisotopes for soi!, sediment, or solid waste samples, because lhese samples potentiallycontaín high levels of lhese ínterfering elements.

known isobaric interferences entered, and will perform necessarycalculations automatically. Table 3125:II shows many of thecommonly used corrections. Monitor the fol1owing additionalmasses: 83Kr, 99Ru, 118Sn,and 1Z5Te.It is necessary to monitorthese masses to correct for isobaric interference caused by 82Kron 82Se, by 98Ru on 98Mo, by I14Snon 114Cd,and by 123Teon123Sb.Monitor ArCl at mass 77, to estimate chloride interfer-ences. Verify that aIl elemental and molecular correction equa-tions used in this method are correct and appropriate for the massspectrometer used and sample matrix.

2) Abundance sensitivity is an analytical condition in wbichthe tails of an abundant mass peak contribute to or obscureadjacent masses. Adjust spectrometer resolution to minimizethese interferences.

3) Polyatomic (molecular) ion interferences are caused byions consisting of more than one atom and having the samenominal mass-to-charge ratio as the isotope of interest. Most ofthe common molecular ion interferences have been identifiedand are listed in Table 3l25:IIl. Because of the severity ofchloride ion interference on important analytes, particularly ar-senic and selenium, hydrochloric acid is not recommended foruse in preparation of any samples to be analyzed by ICPIMS.The mathematical corrections for chloride interferences onlycorrect chloride to a concentration of 0.4%. Because chloride ionis present in most environmental samples, it is critical to usechloride correction equations for affected masses. A high-reso-lution ICPIMS may be used to resolve interferences caused bypolyatomic ions. Polyatomic interferences are strongly influ-enced by instrument design and plasma operating conditions, andcan be reduced in some cases by careful adjustment of nebulizergas fiow and other instrument operating parameters.

4) PhysicaI interferences include differences in viscosity, sur-face tension, and dissolved solids between samples and calibra-tion standards. To minimize these effects, dissolved solid levelsin analytical samples should not exceed 0.5%. Dilute water andwastewater samples containing dissolved solids at or above 0.5%before anaJysis. Use internal standards for correction of physicaJinterferences. Any internaJ standards used should demonstrate com-parable anaJyticaJbehavior to the elements being deterrnined.

5) Memory interferences occur when analytes from a previ-ous sample or standard are measured in the current sample. Usea sufficiently long rinse or fiush between samples to minimizethis type of interference. lf memory interferences persist, theymay be indications of problems in the sample introduction sys-tem. Severe memory interferences may require disassembly andcleaning of the entire sample introduction system, including theplasma torch, and the sampler and skimmer cones.

6) lonization interferences result when moderate (0.1 to 1%)amounts of a matrix ion change the analyte signal. This effect,which usually reduces the analyte signal, also is known as"suppression." Correct for suppression by use of internal stan-dardization techniques.

a. Inductively coupled plasmalmass spectrometer: lnstrumen-tation, available from several manufacturers, includes a massspectrometer detector, inductively coupled plasma source, massfiow controllers for regulation of ICP gas flows, peristaltic pumpfor sample introduction, and a computerized data acquisition and

instrument control system. An x-y autosampler also may be usedwith appropriate control software.

b. Laboratory ware: Use precleaned plastic laboratory warefor standard and sample preparation. Teflon, * either tetrafluoro-ethylene hexafluoropropylene-copolymer (FEP), polytetrafluo-roethylene (PTFE), or perfluoroalkoxy PTFE (PFA) is preferredfor standard preparation and sample digestion, while high-densitypolyethylene (HDPE) and other dense, meta!-free plastics may beacceptable for interna! standards, known-addition solutions, etc.Check each new lot of autosampler tubes for suitability, and pre-clean autosampler tubes and pipettor tips (see Section 301OC.2).

c. A ir displacement pipets, 10 to 100 JLL,100 to 1000 JLL,andI to 10 mL size.

d. Analytical balance, accurate to 0.1 mg.e. Sample preparation apparatus, such as hot plates, micro-

wave digestors, and heated sand baths. Any sample preparationdevice has the potential to introduce trace levels of target ana-Iytes to the sample.f Clean hood (optiona!), Class 100 (certified to contain less

than 100 particles/m3), for sample preparation and manipulation.Preferably perform ali sample manipulations, digestions, dilu-tions, etc. in a certified Class 100 environment. Alternatively,handle samples in glove boxes, plastic fume hoods, or otherenvironments where random contamination by trace metais canbe minimized.

a. Acids: Use ultra-high-purity grade (or equivalent) acids toprepare standards and to process sample. Redistilled acids areacceptable if each batch is demonstrated to be free from con-tamination by target analytes. Use extreme care in the handlingof acids in the laboratory to avoid contamination of the acidswith trace levels of metais.

I) Nitric acid, HN03, conc (specific gravity 1041).2) Nitric acid, I + I: Add 500 mL conc HN03 to 500 mL

reagent water.3) Nitric acid, 2%: Add 20 mL conc HN03 to 100 mL reagent

water; dilute to 1000 mL.4) Nitric acid, 1%: Add 10 mL conc HN03 to 100 mL reagent

water; dilute to 1000 mL.b. Reagent water: Use water of the highest possible purity for

blank, standard, and sample preparation (see Section 1080).Alternati vely, use the procedure described below to producewater of acceptable quality. Other water preparation regimesmay be used, provided that the water produced is metal-free.Reagent water containing trace amounts of analyte elements willcause erroneous results.

Produce reagent water using a softener/reverse osmosis unitwith subsequent UV sterilization. After the general deionizationsystem use a dual-column strong acid/strong base ion exchangesystem to polish laboratory reagent water before production ofmetal-free water. Use a multi-stage reagent water system, withtwo strong acid/strong base ion exchange columns and an acti-vated carbon filter for organics removal for final polishing oflaboratory reagent water. Use only high-purity water for prepa-ration of samples and standards.

c. Stock, standard, and other required solutions: See3120B.3d for preparation of standard stock solutions from ele-mental materiais (pure metais, salts). Preferably, purchase high-purity commercially prepared stock solutions and dilute to re-quired concentrations. Single- or multi-element stock solutions(1000 mglL) of the following elements are required: aluminum,antimony, arsenic, barium, beryllium, cerium, cadmium, chro-mium, coba!t, copper, germanium, indium, lead, magnesium,manganese, molybdenum, nickel, rhodium, scandium, selenium,silver, strontium, terbium, thallium, thorium, uranium, vana-dium, and zinco Prepare internal standard stock separately fromtarget element stock solution. The potential for incompatibilitybetween target elements and/or internal standards exists, andcould cause precipitation or other solution instability.

1) Internal standard stock solution: Lithium, scandium, ger-manium, indium, and thorium are suggcsted as internal stan-dards. The following masses are monitored: 6Li, 45SC, 72Ge,1151n,and 232Th. Add to ali samples, standards, and qualitycontrol (QC) samples a levei of internal standard that will give asuitable counts/second (cps) signal (for most internal standards,200 000 to 500 000 cps; for lithium, 20 000 to 70 000 cps).Minimize error introduced by dilution during this addition byusing an appropriately high concentration of internal standardmix solution. Maintain volume ratio for ali internal standardadditions.

Prepare internal standard mix as follows: Prepare a nominal50-mglL solution of 6Li by dissolving 0.15 g 6Li2C03 (isotopi-ca!ly pure, i.e., 95% or greater purityt) in a minimal amount ofl: I HN03. Pipet 5.0 mL 1000-mglL scandium, germanium,indium, and thorium standards into the lithium solution, diluteresulting solution to 500.0 mL, and mix thoroughly. The result-ant concentration of Sc, Ge, In, and Th will be 10 mglL. Olderinstruments may require higher levels of internal standard toachieve acceptable levels of precision.

Other internal standards, such as rhodium, yttrium, terbium,holmium, and bismuth may also be used in this method. Ensurethat interna! standard mix used is stable and that there are noundesired interactions between elements.

Screen ali samples for internal standard elements before anal-ysis. The analysis of a few representative samples for internalstandards should be sufficient. Analyze samples "as received" or"as digested" (before addition of internal standard), then addinterna! standard mix and reanalyze. Monitor counts at the in-terna! standard masses. If the "as received" or "as digested"samples show appreciable detector counts (10% or higher ofsamples with added internal standard), dilute sample or use analternate interna! standard. If the internal standard response ofthe sample with the addition is not within 70 to 125% of theresponse for a calibration blank with the internal standard added,either dilute the sample before analysis, or use an alternateinternal standard. During actual analysis, monitor internal stan-dard masses and note ali internal standard recoveries over 125%of internal standard response in calibration blank. Interpret re-sults for these samples with caution.

The internal standard mix may be added to blanks, standards,and samples by pumping the solution so it is mixed with thesample stream in the sample introduction processo

RinseReagent blankReagent blank5-J.Lg/Lstandard10-J.Lg/Lstandard20- J.Lg/Lstandard50-J.Lg/Lstandard100-J.Lg/LstandardRinsernitial calibration verification, 50

J.Lg/LInitial calibration blank0.30-J.Lg/Lstandard1.0-J.Lg/LstandardExternal reference materialContinuing calibration verificationContinuing blank calibrationProject sample method blankProject sample laboratory-fortified

blankProject sample 1-4Project sample 5Project sample 5 with known

additionProject sample 5 duplicate with

known additionContinuing calibration verificationContinuing calibration blank

Check mass calibration andresolution

Optimize instrument for maximumrhodium counts while keepingoxides, double charged ions,and background withininstrument specifications

Check for contaminationCalibration standard blank

Low-Ievel calibration verificationLow-Ievel calibration verificationNr5T 1643c ar equivaJent

Mass resolutionMass calibrationBa2+/Ba+CeO/CeBackground counts at mass 220Correlation coefficientCalibration blanksCalibration verification standardsLaboratory fortified blank (control

sample)Precision

Known-addition recovery0.3 and 1.0 J.Lg/L standards

Manufacturer's specificationManufacturer' s specificationManufacturer' s specificationManufacturer' s specificationManufacturer' s specification2:0.995< Reporting limit± 10% of true vaJue

±30% of true value±20% relative percent difference

for lab duplicates75-125%Dependent on data quality

objectivesDependent on data quality

objectives70-125% of response in

calibration blank with knownaddition

2) lnstrument optimization/tuning solution, containing thefollowing elements: barium, beryllium, cadmium, cerium, co-balt, copper, germanium, indium, magnesium, rhodium, scan-dium, terbium, thallium, and lead. Prepare this solution in 2%HN03. This mix includes ali common elements used in optimi-zation and tuning of the various ICPIMS operational parameters.It may be possible to use fewer elements in this solution, de-pending on the instrument manufacturer's recommendations.

3) Calibration standards, O, 5, 10, 20, 50, and 100 fLgIL.+Other calibration regimes are acceptable, provided the full suíteof quality assurance samples and standards is mn to validatethese method changes. Fewer standards may be used, and atwo-point blank/mid-range calibration technique commonly usedin ICP optical methods should also produce acceptable results.Calibrate ali analytes using the selected concentrations. Prepareali calibration standards and blanks in a matrix of 2% nitric acid.Add internal standard mix to ali calibration standards to provideappropriate count rates for interference correction. NOTE: Alistandards and blanks used in this method have the internalstandard mix added at the same ratio.

4) Method blank, consisting of reagent water (lJ[ 3b) takenthrough entire sample preparation processo For dissolved sam-pIes, take reagent water through same filtration and preservationprocesses used for samples. For samples requiring digestion,process reagent water with the same digestion techniques assamples. Add internal standard mix to method blank.

Reference material['J[3c9)]

Greater of: onceper samplebatch, ar 5%

Greater of: onceper samplebatch, or 5%

Dependent on dataquality objectives

Preparatory /methodblank ['J[ 3c4)]

± Absolute vaJueaf instrumentdetection limit;± absolute valueof laboratoryreporting limit orMDLisacceptable

± 30% of truevalue

Laboratory fartifiedblank ['J[ 3c7)]

Greater of: onceper samplebatch, or 5%

Greater of: onceper samplebatch, or 5%

10%

± 20% relativepercentdifference

± 10% of knownconcentration

Duplicate known-addition samples

Continuing calibrationverification standards['J[ 3c5)]

Continuing caJibrationverification blank['J[ 3c6)]

± Absolute valueof instrumentdetection limit;± absolute vaJueof laboratoryreporting limit orMDLisacceptable

Continuing Calibration Verification Standard (N = 44) Initial Calibration Verification Standard (N = 12)

Standard Relative Standard Standard Relative StandardMean Recovery Mean Deviation Deviation Mean Recovery Mean Deviation Deviation

Element Mass % J-LgIL J-LgIL % % J-LgIL J-LgIL %

Be 9 98.71 49.35 3.43 6.94 100.06 50.03 1.90 3.80AI 27 99.62 49.81 2.99 6.01 98.42 49.21 1.69 3.44V 51 100.97 50.48 1.36 2.68 99.91 49.96 1.23 2.47Cr 52 101.39 50.70 1.86 3.66 99.94 49.97 1.47 2.95Cr 53 100.68 50.34 1.91 3.79 99.13 49.56 1.44 2.90Mn 55 101.20 50.60 1.98 3.91 99.48 49.74 1.40 2.82Co 59 101.67 50.83 2.44 4.79 99.44 49.72 1.61 3.24Ni 60 99.97 49.99 2.14 4.28 97.98 48.99 1.70 3.47Ni 62 99.79 49.89 2.09 4.18 97.57 48.79 1.32 2.71Cu 63 100.51 50.25 2.19 4.36 97.87 48.93 1.63 3.33Cu 65 100.39 50.19 2.26 4.51 98.34 49.17 1.58 3.20Zn 66 101.07 50.53 1.93 3.82 98.75 49.38 0.87 1.76Zn 68 100.42 50.21 1.89 3.77 97.75 48.87 0.50 1.02As 75 100.76 50.38 1.15 2.28 98.83 49.41 0.89 1.80Se 77 101.71 50.85 1.43 2.81 99.54 49.77 1.01 2.03Se 82 101.97 50.98 1.50 2.95 99.76 49.88 0.94 1.89Ag 107 101.50 50.75 1.68 3.30 99.27 49.63 1.17 2.36Ag 109 101.65 50.83 1.68 3.31 99.66 49.83 1.54 3.08Cd 111 100.92 50.46 1.94 3.84 98.61 49.30 1.36 2.77Cd 114 100.90 50.45 2.07 4.10 99.20 49.60 1.41 2.84Sb 121 100.14 50.07 2.39 4.77 99.38 49.69 1.38 2.78Sb 123 99.98 49.99 2.48 4.97 99.09 49.54 1.34 2.71TI 203 101.36 50.68 1.64 3.23 100.05 50.02 1.01 2.01TI 205 102.40 51.20 1.93 3.78 101.23 50.62 1.45 2.87Pb 208 101.21 50.61 1.65 3.25 99.33 49.67 0.84 1.69U 238 101.54 50.77 1.93 3.80 99.80 49.90 1.36 2.72

* Single-laboratory, single-operator, single-instrumem data, determined using a SO-JLgILstandard prepared from sources independem of calibration standard source. Dataacquired January-November )996 during actual sample determinations. Performance of continuing calibration verification standards at different levels may vary.Perkin-Elmer Elan 6000 ICPIMS used for determination.

5) Calibration verification standard: Prepare a mid-rangestandard, from a source different from the source of the calibra-tion standards, in 2% HN03, with equivalent addition of internalstandard.

6) Calibration verification blank: Use 2% HN03.7) Laboratory fortified blank (optional): Prepare solution with

2% nitric acid and method analytes added at about 50 J-LglL.Thisstandard, some times called a laboratory control sample (LCS), isused to validate digestion techniques and known-addition levels.

8) Reference materials: Externally prepared reference mate-rial, preferably from National Institute of Standards and Tech-nology (NIST) 1643 series or equivalent.

9) Known-addition solution for samples: Add stock standardto sample in such a way that volume change is less than 5%. Inthe absence of information on analyte levels in the sample,prepare known additions at around 50 J-LglL.If analyte concen-tration levels are known, add at 50 to 200% of the sample levels.For samples undergoing digestion, make additions before diges-tion. For the determination of dissolved metais, make additionsafter filtration, preferably immediately before analysis.

10) Low-level standards: Use both a 0.3- and a 1.0-J-LglLstandard when expected analyte concentration is below 5 J-LglL.Prepare both these standards in 2% nitric acid.

Prepare volumetrically a mixed standard containing themethod analytes at desired concentration(s) (0.30 J-LglL,1.0J-LgIL,or both). Prepare week1y in 100-rnL quantities.

d. Argon: Use a prepurified grade of argon unless it can bedemonstrated that other grades can be used successfully. The useof prepurified argon is usually necessary because of the presenceof krypton as an impurity in technical argon. 82Krinterferes withthe determination of 82Se. Monitor 83Kr at ali times.

a. Sample preparation: See Sections 301Oand 3020 for generalguidance regarding sampling and quality control. See Section3030E for recommended sample digestion technique for ali analytesexcept silver and antimony. If silver and antimony are target ana-Iytes, use method given in 3030F, paying special attention to inter-ferences caused by chloride ion, and using ali applicable elementalcorrections. Alternative digestion techniques and additional guid-ance on sample preparation are available.3,4

Ideally use a "clean" environment for any sample handling,manipulation, or preparation. Preferably perform ali sample ma-nipulations in a Class 100 clean hood or room to minimizepotential contamination artifacts in digested or filtered samples.

b. Instrument operating conditions: Follow manufacturer'sstandard operating procedures for initialization, mass calibration,gas flow optimization, and other instrument operating condi-tions. Maintain complete and detailed information on the oper-ational status of the instrument whenever it is used.

c. Analytical run sequence: A suggested analytical run se-quence, including instrument tuning/optimization, checking ofreagent blanks, instrument calibration and calibration verifica-tion, analysis of samples, and analysis of quality control samplesand blanks, is given in Table 3125:IV.

d. Instrument tuning and optimization: Follow manufactur-er' s instructions for optimizing instrument performance. Themost important optimization criteria include nebulizer gas f1ows,detector and lens voltages, radio-frequency forward power, andmass calibration. Periodically check mass calibration and instru-ment resolution. Ideally, optimize the instrument to minimizeoxide formation and doubly-charged species formation. Measurethe CeO/Ce ratio to monitor oxide formation, and measuredoubly-charged species by determination of the Ba2+ lBa + ratio.Both these ratios should meet the manufacturer's criteria befareinstrument calibration. Monitor background counts at mass 220after optimization and compare with manufacturer' s criteria. Asummary of perfarmance criteria related to optimization andtuning, calibration, and analytical performance for this method isgiven in Table 3125:V.

e. lnstrument calibration: After optimization and tuning, cal-ibrate instrument using an appropriate range of calibration stan-dards. Use appropriate regression techniques to determine cali-bration lines or curves for each analyte. For acceptable calibra-tions, correlation coefficients for regression curves are ideally0.995 or greater.

Immediately after calibration, run initia! calibration verifica-tion standard, lj[ 3c5); acceptance criteria are :l: 10% of knownanalyte concentration. Next run initial calibration verificationblank, lj[ 3c6); acceptance criteria are ideally :l: the absolute valueof the instrument detection limit for each analyte, but in practice,:l: the absolute value of the laboratory reporting limit or thelaboratory method detection limit for each analyte is acceptable.Verify low-Ievel calibration by running 0.3- and/or 1.0-/-LglLstandards, if analyte concentrations are less than 5 /-LglL.f Sample analysis: Ensure that alI vessels and reagents are

free from contamination. During analytical run (see Table 3125:IV), include quality control analyses according to schedule ofTable 3125:VI, ar follow project-specific QNQC protocols.

Interna! standard recoveries must be between 70% and 125% ofinternal standard response in the laboratory-fortified blank; other-wise, dilute sample, add internal standard mix, and reana!yze.

Make known-addition analyses for each separate matrix in adigestion or filtration batch.

Configure instrument software to report internal standard cor-rected results. For water samples, preferably report results inmicrograms per liter. Report appropriate number of significantfigures.

Total Recoverable Metalst

Relative StandardMean Recovery Deviation

Element Mass % %

Be 9 89.09 5.77V 51 87.00 8.82Cr 52 87.33 8.42Cr 53 86.93 7.90Mn 55 91.81 10.12Co 59 87.67 8.92Ni 60 85.07 8.42Ni 62 84.67 8.21Cu 63 84.13 8.46Cu 65 84.37 8.05Zn 66 86.14 23.01Zn 68 81.95 20.31As 75 90.43 4.46Se 77 83.09 4.76Se 82 83.42 4.73Ag 107Ag 109Cd 111 91.37 5.47Cd 114 91.47 6.04Sb 121 94.40 5.24Sb 123 94.56 5.36TI 203 97.24 5.42TI 205 98.14 6.21Pb 208 96.09 7.08

Dissolved Metalst

Relative StandardDeviation

%Mean Recovery

%

88.38 6.4388.52 5.95

89.31 5.7089.00 5.8288.55 8.3388.26 7.8095.59 13.8191.94 13.2797.30 8.84

105.36 10.80105.36 10.7591.98 5.0692.25 4.9696.91 6.0397.03 5.42

* Single-Iaboratory, single-operator, single-instrument data. Samples were Washington State surface waters fram various locations. Data acquired January-November 1996during acrual sample determinations. Performance of known additions at different levels may vary. Perkin-Elrner Elan 6000 ICP/MS used for determination.t Known-addition levei 20 JLgIL. Additions made before preparation according to Section 3030E (rnodified by cleanhood digestion in TFE beakers). N = 20.:j: Known-addition leveI for Cd and Pb I JLgIL; for other analytes 10 JLglL. Additions made after filtration through 1:I HNO, precleaned 0.45-JLrn filters. N = 28.

-~

3-52 METAL8 (3000)

TABLE3125:IX. METHODPERFORMANCEWITHLow-LEVELCHECKSTANDARDS*

1.0-p,g/L Standard 0.3-p,g/L Standard

Mean Standard Relative Standard Mean Standard Relative StandardRecovery Mean Deviation Deviation Recovery Mean Deviation Deviation

Element Mass % p,glL p,g/L % % p,g/L p,g/L %

Be 9 97 0.97 0.06 6.24 95 0.284 0.03 12.11AI 27 121 1.21 0.32 26.49 196 0.588 0.44 74.30V 51 104 1.04 0.06 5.83 111 0.332 0.10 28.96Cr 52 119 1.19 0.34 28.62 163 0.490 0.37 75.90Cr 53 102 1.02 0.36 35.54 113 0.338 0.32 93.70Mn 55 103 1.03 0.07 6.55 110 0.329 0.08 25.64Co 59 103 1.03 0.07 6.42 102 0.307 0.04 12.53Ni 60 101 1.01 0.05 5.24 107 0.321 0.05 14.14Ni 62 102 1.02 0.06 5.42 109 0.326 0.05 15.94Cu 63 107 1.07 0.09 8.78 118 0.355 0.06 18.29Cu 65 107 1.07 0.10 9.05 117 0.352 0.06 17.69Zn 66 117 1.17 0.51 43.52 182 0.547 0.68 124.13Zn 68 116 1.16 0.50 42.90 179 0.537 0.66 122.12As 75 97 0.97 0.05 5.23 101 0.302 0.06 18.29Se 77 89 0.89 0.08 8.72 88 0.265 0.08 29.07Se 82 92 0.92 0.14 15.50 106 0.317 0.14 43.91Ag 107 101 1.01 0.05 4.53 94 0.282 0.04 15.74Ag 109 103 1.03 0.07 6.57 92 0.277 0.04 13.68Cd 111 98 0.98 0.04 3.80 96 0.288 0.03 8.74Cd 114 100 1.00 0.03 3.39 98 0.293 0.03 8.70Sb 121 94 0.94 0.05 5.28 93 0.280 0.06 21.89Sb 123 94 0.94 0.05 5.36 93 0.278 0.06 22.39TI 203 101 1.01 0.04 3.57 98 0.294 0.03 11.89TI 205 104 1.04 0.05 5.15 100 0.300 0.03 10.43Pb 208 104 1.04 0.04 3.65 104 0.312 0.03 11.13U 238 106 1.06 0.05 4.64 102 0.307 0.03 9.92

* Single-laboratory,single-operator,single-instrumentdata. N = 24 for botb standards.

a. Correction for dilutions and solids: Correct all results fordilutions, and raise reporting limit for all analytes reported fromthe diluted sample by a corresponding amount. Sirnilarly, ifresults for solid samples are to be deterrnined, use Method2540B to determine total solids. Report results for solid samplesas micrograms per kilogram, dry weight. Correct all results forsolids content of solid samples. Use the following equation tocorrect solid or sediment sample results for dilution duringdigestion and moisture content:

Runcorr x VR =-----corr W x % TS/l 00

Rcorr = corrected result, p,gIkg,Runcorr = uncorrected elemental result, p,g/L,

V = volume of digestate (after digestion), L,W = mass of the wet sample, kg, and

% TS = percent total solids determined in the solid sample.

b. Compensation for interferences: Use instrument softwareto correct for interferences listed previously for this method. SeeTable 3125:III for a listing of the most common molecular ioninterferences.

c. Data reporting: Establish appropriate reporting lirnits formethod analytes based on instrument detection limits and the

laboratory blank. For regulatory programs, ensure that reportinglirnits for method analytes are a factor of three below relevantregulatory criteria.

li method blank contamination is typically random, sporadic,or otherwise not in statistical control, do not correct results forthe method blank. Consider the correction of results for labora-tory method blanks only if it can be demonstrated that theconcentration of analytes in the method blank is within statisticalcontrol over a period of months. Report all method blank dataexplicitly in a manner identical to sample reporting procedures.

d. Documentation: Maintain documentation for the following(where applicable): instrument tuning, mass calibration, calibra-tion verification, analyses of blanks (method, field, calibration,and equipment blanks), IDL and MDL studies, analyses ofsamples and duplicates with known additions, laboratory andfield duplicate information, serial dilutions, internal standardrecoveries, and any relevant quality control charts.

Also maintain, and keep available for review, all raw datagenerated in support of the method.5

Table 3125:1 presents instrument detection limit (lDL) datagenerated by this method; this represents optimal state-of-the-art instrument detection capabilities, not recommended

method detection or reporting limits. Tables 3125: VII throughIX contain single-laboratory, single-operator, single-instru-ment performance data generated by this method for calibra-tion verification standards, low-level standards, and known-addition recoveries for fresh-water matrices. Performancedata for this method for some analytes are not currentlyavailable. However, performance data for similar ICPIMSmethods are available in the literature.1•4

1. V.S. ENVIRONMENTALPROTECTlONAGENCY.1994. Determination oftrace elements in waters and wastes by inductively coupled plas-ma-mass spectrometry, Method 200.8. V.S. Environmental Pro-tection Agency, Environmenta1 Monitoring Systems Lab., Cincin-nati,Ohio.

2. V.S. ENVIRONMENTALPROTECTIONAGENCY.1984. Definition and pro-cedure for the determination of the method detection limit, revision1.11. 40 CFR 136, Appendix B.

3. V.S. ENVIRONMENTALPROTECTlONAGENCY.1991. Methods for the'determination of metaIs in environmental samples. V.S. Environ-mental Protection Agency, Off. Research & Development, Wash-ington D.e.

4. V.S. ENVIRONMENTALPROTECTIONAGENCY.1995. Method 1638: De-termination of trace elements in ambient waters by inductively cou-pled plasma mass spectrometry. V.S. Environmental ProtectionAgency, Off. Water, Washington, D.C.

5. V.S. ENVIRONMENTALPROTECTIONAGENCY.1995. Guidance on theDocumentation and Evaluation of Trace MetaIs Data Collected forClean Water Act Compliance Monitoring. V.S. Environrnental Pro-tection Agency, Off. Water, Washington, D.C.

GRAY,A.L. 1974. A plasma source for mass analysis. Prac. Soe. Anal.Chem. 11:182.

HAYHURST,A.N. & N.R. TELFORD.1977. Mass spectrometric sampling ofions from atmospheric pressure flames. I. Characteristics and cali-bration of the sampling system. Combust. Flame. 67.

HOUK,R.S., V.A. FASSEL,G.D. FLESCH,H.1. SVEC,AL GRAY& C.E.TAYLOR.1980. Inductively coupled argon plasma as an ion source

for mass spectrometric determination of trace elements. Ana/.Chem. 52:2283.

DOUGLAs,D.1. & 1.B. FRENCH.1981. Elemental analysis with a micro-wave-induced plasma/quadrupole mass spectrometer system. Anal.Chem. 53:37.

HOUK,R.S., V.A. FASSEL& H.1. SVEC.1981. lnductively coupled plas-ma-mass spectrometry: Sample introduction, ionization, ion extrac-tion and analytical results. Dyn. Mass Speetrom. 6:234.

OUVARES,1.A. & R.S. HOUK.1985. Ion sampling for inductively coupledplasma mass spectrometry. Anal. Chem. 57:2674.

HOUK,R.S. 1986. Mass spectrometry of inductively coupled plasmas.Anal. Chem. 58:97.

THOMPSON,1.1. & R.S. HOUK.1986. lnductively coupled plasma massspectrometric detection for multielement flow injection analysis andelemental speciation by reversed-phase liquid chromatography.Anal. Chem. 58:2541.

VAUGHAN,M.A. & G. HORLICK.1986. Oxide, hydroxide, and doublycharged analyte species in inductively coupled plasma/mass spec-trometry. Appl. Speetrose. 40:434.

GARBARINO,1.R. & H.E. TAYLOR.1987. Stable isotope dilution analysisof hydrologic samples by inductively coupled plasma mass spec-trometry. Anal. Chem. 59:1568.

BEAUCHEMIN,D., J.W. McLAREN,A.P. MYKYTIUK& S.S. BERMAN.1987.Determination of trace metals in a river water reference material byinductively coupled plasma mass spectrometry. Anal. Chem. 59:778.

THOMPSON,1.1. & R.S. HOUK.1987. A study of internal standardizationin inductively coupled plasma-mass spectrometry. Appl. Spectrose.41:801.

1ARVIS,K.E., A.L. GRAY& R.S. HOUK. 1992. lnductively CoupledPlasma Mass Spectrometry. Blackie Academic & Professional,Chapman & Hall, New York, N.Y.

TAYLOR,D.B., H.M. KINGSTON,D.1. NOGAY,D. KOLLER& R. HmoN.1996. On-line solid-phase chelation for the determination of eightmetals in environmental waters by ICP-MS. 1MS 11:187.

KINGSTON,H.M.S. & S. HASWELL,eds. 1997. Microwave EnhancedChemistry: Fundamentals, Sample Preparation, and Applications.ACS Professional Reference Book Ser., American Chemical Soc.,Washington, D.C.

V.S. ENVIRONMENTALPROTECTIONAGENCY.1998. lnductively coupledplasma-mass spectrometry, Method 6020. In Solid Waste Methods.SW846, Vpdate 4, V.S. Environmental Protection Agency, Envi-ronrnenta1 Monitoring Systems Lab., Cincinnati, Ohio.

Anodic stripping voltammetry (ASV) is one of the most sen-sitive metal analysis techniques; it is as much as 10 to 100 timesmore sensitive than electrothermal atomic absorption spectros-copy for some metaIs. This corresponds to detection limits in thenanogram-per-liter range. The technique requires no sample

* Approved by Standard Methods Committee, 1997.Joint Task Group: 20th Edition-Ma1gorzata Ciszkowska (chair), Margaret M.Goldberg, Janet G. Osteryoung, Robert S. Rodgers, Marek Wojciechowski.

extraction or preconcentration, it is nondestructive, and it allowssimultaneous determination of four to six trace metaIs, utilizinginexpensive instrumentation. The disadvantages of ASV are thatit is restricted to amalgam-forming metais, analysis time islonger than for spectroscopic methods, and interferences andhigh sensitivity can present severe limitations. The analysisshould be performed only by analysts skilled in ASV method-ology because of the interferences and potential for trace back-ground contamination.

a. Principie: Anodic stripping voltammetry is a two-step elec-troanalytical technique. In the preconcentration step, metal ionsin the sample solution are reduced at negative potential andconcentrated into a mercury electrode. The concentration of themetal in the mercury is 100 to lOOOtimes greater than that of themetal ion in the sample solution. The preconcentration step isfollowed by a stripping step applying a positive potential scan.The amalgamated metal is oxidized rapidly and the accompany-ing current is proportional to metal concentration.

b. Detection limits and working range: The limit of detectionfor metal determination using ASV depends on the metal deter-mined, deposition time, stirring rate, solution pH, sample matrix,working electrode (hanging mercury drop electrode, HMDE, orthin mercury film electrode, TMFE), and mode of the strippingpotential scan (square wave or differential pulse). Cadmium,lead, and zinc are concentrated efficiently during pre-electrolysisbecause of their high solubility in mercury and thus have lowdetection limits « I J,LgIL). Long deposition times and highstirring rates increase the concentration of metal preconcentratedin the mercury phase and reduce detection limits. The effects ofsolution pH and matrix are more complicated. In general, add ahigh concentration of inert electrolyte to samples to maintain ahigh, constant ionic strength. Acidify sample to a low pH or adda pH buffer. If the pH buffer or other component of the samplematrix complexes the metal (3130B.lc), detection limits oftenare increased.

The choice of working electrode is determined largely by theworking range of concentration required. The HMDE is bestsuited for analysis from approximately I J,LgIL to 10 mgIL, whilethe TMFE is superior for detection below 1 J,LgIL.

c. lnteiferences: Major interferences include intermetalliccompound formation, overlapping stripping peaks, adsorption oforganics, and complexation. Intermetallic compounds can formin the mercury phase when high concentrations of certain metaIsare present simultaneously. Zinc forms intermetallic compoundswith cobalt and nickel, and both zinc and cadmium form inter-metallic compounds with copper, silver, and gold. As a result,the stripping peak for the constituent metais may be severelydepressed or shifted and additional peaks due to intermetalliccompound stripping may be observed. Minimize or avoid inter-metallic compound formation by use of a hanging mercury dropelectrode instead of a thin film mercury electrode when metalconcentrations are above 1 J,LgIL, application of a preconcentra-tion potential sufficiently negative to reduce the desired but notthe interfering metal, and use of a relatively short preconcentra-tion period followed by a relatively large pulse modulation (50mV) during the stripping stage. In general, suspect formation ofintermetallic compounds if metaIs are present in concentrationsabove I mgIL. li metaIs are present at concentrations above 10mgIL, do not use anodic stripping voltammetry. Concentrationsabove 10 mgIL usually can be quantitated by methods such asthose given in Sections 3111 and 3120.

Separate over1apping stripping peaks by various methods,including appropriate choice of buffer and electrolyte. 1-3 If onlyone of the metal peaks is of interest, eliminate interfering peaks

by selective complexation with a suitable ligand, such as EDTA.ludicious choice of preconcentration potential can result in thedeposition of the selected metal but not the interfering metal inthe mercury electrode. Selection of buffer/ligand also may helpto distinguish metaIs during the preconcentration step. Alterna-tively, use "medi um exchange," in which preconcentration isperformed with the electrodes in the sample and stripping isperformed in a different electrolyte solution. In this procedure,metaIs are deposited from the sample into the amalgam as usual,but they may be stripped into a medium that provides differentstripping peak potentials for the over1apping metaIs.

Minimize interferences from adsorption of organic com-pounds and complexation by removal of the organic matter.Digest samples with high-purity acids as described in Section3030. Make standard additions to determine if complexation oradsorption remains a problem. Analyze a metal-free solutionwith a matrix similar to that of the sample both before and afteraddition of known quantities of standard. Repeat procedure forsample. If the slope of the stripping current versus added metalis significantly different in the sample reiative to the metal-freesolution, digest sample further. The choice of stripping wave-form also is important. While both square-wave and differential-pulse stripping attempt to minimize the contribution of adsorp-tion currents to the total measured stripping current, square-wavestripping does this more effectively. Thus, use square-wavestripping instead of differential-pulse stripping when adsorptionoccurs.

a. Electrochemical analyzer: The basic electrochemical ana-lyzer for ASV applications contains a three-electrode poten-tiostat, which very precisely controls potential applied to theworking electrode relative to the reference electrode, and asensitive current measuring device. It is capable of deliveringpotential pulses of various amplitudes and frequencies, and pro-vides several scan rates and current ranges. More advanced ASVinstruments offer automated timing, gas purge and stirring, anddata processing routines including curve smoothing, baselinecorrection, and background subtraction.

Two variations of stripping waveforms are commonly used:differential pulse (DPASV) and square wave (SWASV) wave-forms. The differential-pulse waveform consists of a series ofpulses superimposed on a linear voltage ramp, while the square-wave waveform consists of a series of pulses superimposed on astaircase potential waveform. Square-wave stripping is signifi-cantly faster than differential-pulse stripping and is typically tentimes more sensitive. Most commercially available ASV instru-ments perform both differential-pulse and square-wave stripping.

b. Electrodes and cell: Provide working, reference, and aux-iliary electrodes. Working electrodes are either hanging mercurydrop or thin mercury film electrodes. Hanging mercury dropelectrodes must be capable of dispensing mercury in very pre-cisely controlled drop sizes. Three types of electrodes meet thisrequirement: static mercury drop electrodes, controlled growthmercury drop electrodes, or Kemula-type electrodes. In any case,

use a drop knocker to remove an old drop before dispensing afresh mercury drop.

When the lowest detection limits are required, a thin mercuryfilm electrode is preferred. This electrode consists of a rotatingglassy carbon disk plated with mercury in situ during precon-centration of the analyte. A high-precision, constant-speed rota-tor controls the rotation rate of the electrode and provides repro-ducible mass transporto

Reference electrodes may be either saturated calomel or silverlsilver chloride electrodes. Use a platinum wire for the auxiliaryelectrode.

Use cells constructed of glass, or preferably fused silica orTFE, because they are more resistant to solution adsorption orleaching. Cover cell with a lid that provides reproducible place-ment of the electrodes and gas purging tubes. Provide an addi-tional hole in the cell lid for addition of standards. Most com-mercially available mercury drop electrodes include electrolyticcells, reference and auxiliary electrodes, and gas purging tubes.

Use a constant-speed stirring mechanism to provide reproduc-ible mass transport in samples and standards.

Locate the cell in an area where temperature is relativelyconstant. Altematively, use a constant-temperature water bathand cell jacket.

c. Oxygen-removal apparatus: Oxygen interferes in electro-chemical analyses; remove it from solution before preconcentra-tion by purging with nitrogen or argon. Provide two gas inlettubes through the cell lid: one extends into the solution and thesecond purges the spaee above the solution. A gas outlet hole inthe lid provides for removal of oxygen and excess purging gas.

d. Recording device: If the electrochemical analyzer is notequipped with a digital data acquisition system, use an XYplotter to record stripping voltammograms.

e. Timer: If preconcentration and equilibration periods are notcontrolled by the instrument, use an accurate timing device.f Polishing wheel: To obtain the high polish required for a

glassy carbon disk electrode, use a motorized polishing wheel.

CAUTION:Follow proper practices for disposal of any solu-tions containing mercury.

a. Metal-free water: Use deionized water to prepare buffers,electrolytes, standards, etc. Use water with at least 18megohm-em resistivity (see Seetion 1080).

b. Nitric acid, HN03, conc, high-purity.*C. Nitric acid, HN03, 6N, 1.6N (10%), and 0.01N.d. Purging gas (nitrogen or argon), high-purity. Remove

traces of oxygen in nitrogen or argon gas before purging thesolution. Pass gas through sequential scrubbing columns con-taining vanadous chloride in the first, deionized water in thesecond, and buffer (or electrolyte) solution in the third column.

e. Metal standards: Prepare stock solutions containing I mgmetal/mL in polyethylene bottles. Purchase these solutions com-mercially or prepare as in Section 3111. Daily prepare dilutionsof stock standards in a matrix similar to that of the samples tocover the concentration range desired.

f Electrolytelbuffer: Use one of the following:1) Acetate buffer, pH 4.5: Dissolve 16.4 g anhydrous sodium

acetate, NaC2H302, in 800 mL water. Adjust to pH 4.5 withhigh-purity glacial acetic acid. * Dilute to 1 L with water.

2) Citrate buffer, pH 3: Dissolve 42.5 g citric acid (monohy-drate) in 700 mL water. Adjust to pH 3 with high-purityNH40H. * Dilute to I L with water.

3) Phosphate buffer, pH 6.8: Dissolve 24 g NaH2P04 in 500mL water. Adjust to pH 6.8 with IN NaOH. Dilute to I L withwater.

g. Mercury: Use commercially available triply distilled me-tallic mercury for hanging mercury drop electrodes. CAUTION:Mercury vapors are highly toxic. Use only in well-ventilatedarea.

h. Mercuric nitrate solution, Hg(N03)2: For thin mercury filmelectrodes, dissolve 0.325 g Hg(N03h in 100 mL O.OIN HN03.

i. Reference eleetrode filling solution: Available from elec-trode manufacturer.

j. Amalgamated zinc: Dissolve 2 g Hg(N03h in 25 mL concHN03; dilute to 250 mL with water. In a separate beaker, cleanapproximately 50 g mossy zinc by gently oxidizing the surfacewith 10% HN03 and rinse with water. Add Hg(N03h solution tocleaned zinc and stir with a glass rodoIf barely visible bubbles donot appear' add a small amount of 6N HN03. Zine should rapidlyacquire a shiny, metallic appearance. Decant solution and storefor amalgamating future batches of zinco Rinse amalgamatedzine copiously with water and transfer to a gas scrubbing col-umn.

k. Hydrochloric acid, HCI, conc.l. Vanadous chloride: Add 2 g ammonium metavanadate,

NH4V03, to 25 mL cone HCI and heat to boiling. Solutionshould tum blue-green. Dilute to 250 mL with water. Poursolution into gas scrubbing column packed with amalgamatedzinc and bubble purging gas through it until the solution tums aclear violet color. When the violet color is replaced by a blue,green, or brown color, regenerate vanadous chloride by addingHCI.

m. Siliconizing solution: Preferably use commercially avail-able solutions in sealed ampules for siliconizing capillaries usedfor hanging mercury drops. CAUTION:Most commercial siliconiz-ing reagents contain CCI4, a toxic and cancer-suspect agent.Handle with gloves and avoid breathing vapors.

n. Alumina suspensions, I, 0.3, and 0.05 /Lm. Use eommer-cially available alumina suspensions in water, or make a suspen-sion by adding a small amount of water to the alumina.

O. Hydrofluoric acid, HF, 5%: Dilute 5 mL conc HF to 100mL with water.

p. Methanol.q. Sodium hydroxide, NaOH, lN.

a. Sample preparation and storage: Colleet samples in pre-cleaned, acid-soaked polyethylene or TFE bottles. Add 2 mLconc HNOiL sample and mix well. Cap tightly and store inrefrigerator or freezer until ready for analysis.

b. Cell preparation: Soak clean cell in 6N HN03 ovemightanà rinse well with water before use.

c. Electrode preparation:1) HMDE-Follow manufacturer's guidance for capillary clean-

ing. If not available, use the following procedure. Remove allmercury from the capillary. Aspirate the following through thecapillary in the order listed: 6N HN03, water, 5% HF, water,methanol, and air. Dry capillary at 100°C for 1 h. Siliconize cooledcapillary using a siliconizing solution. Between uses, fill capillarywith clean mercury and irnrnerse tip in clean mercury. If the cap-illary fails to suspend a drop of mercury, repeat cleaning.

2) TMFE-Polish glassy carbon disks used for thin mercuryfilm electrodes to a high metallic sheen with alumina suspen-sions, progressively decreasing particle size from 1 fLm to 0.05fLm. Use a motorized polishing wheel for best results. Com-pletely rinse off all traces of alurnina with water. Check diskfrequently for etching or pitting; repolish as necessary to main-tain reproducible mercury film deposition.

d. Instrumental conditions: Use the following conditions:

Initial potentialFinal potentialEquilibration potentialHMDE drop sizeTMFE rotation rateDPASV:

Pulse amplitudePulse periodPulse widthSample widthScan rate

SWASV:SW amplitudeStep potentialFrequency

-1.00 V (Pb, Cd); -1.20 V (Zn)0.00 V-1.00 V (Pb, Cd); -1.20 V (Zn)medium2000 rpm

25 mV0.5 s50 ms17 ms5 mV/s

25 mV4mV100 Hz

e. Deoxygenation: Pipet 2 mL sample and 3 mL electrolyte/buffer into cell. If using a TMFE, add 10 fLL Hg(N03)2 solution.Place electrodes in cell and secure cell lidoDeoxygenate solutionwith purified purging gas for 10 min while stirring. When solutionpurge is completed, purge the space above the solution with purifiedpurging gas. Continue head-space purge throughout analysis.

f Preconcentration: If using a HMDE, dispense a new mer-cury drop. Start preconcentration, stirring and timing simulta-neously. Precisely control and keep constant preconcentrationtimes and stirring rates for solutions and standards. Generally use120 s and a rotation rate of 2000 rpm for the TMFE.

Afier the metal is sufficient1y concentrated in the amalgam,stop stirring or TMFE rotation for an equilibration period ofprecisely 30 s.

g. Anodic stripping: After equilibration period, begin anodicstripping without stirring and make potential at the workingelectrode progressively more positive as a function of time.Monitor stripping current and plot as a function of appliedpotential in stripping voltammograms. Use peak current to quan-tify metal concentration and peak potential to identify the metal.

h. Add 5 to 50 fLL standard solution and repeat analysis,beginning with deoxygenation of sample. Adjust volume ofadded standard solution to obtain 30 to 70% increase of thestripping peak. If 50 fLL addition is not sufficient, use standardsolution with a higher concentration of metal. Shorten deoxy-genation step to 1 min after initial gas purge.

Calculate the concentration of metal in the original sampleusing the following equation:

Cs x Vs ioCo = --- x ---Vo (is io)

where:Co = concentration of metal in sample, mg/ L,Cs = concentration of metal in standard solution, mgIL,io = stripping peak height in original sample,is = stripping peak height in sample with standard addition,

Vo = volume of sample, rnL, andVs = volume of standard solution added, rnL.

Follow quality control guidelines outlined in Section 3020with respect to use of additions, duplicates, and blanks for best

TABLE 3130:1. PRECISIONOF CD, PB, ANDZN ANALYSISBY ASV

Metal Concentration RSD/-Lg/L %

Sample Electrode ASV Mode Cd Pb Zn Cd Pb Zn

Tap water #14 HMDE SW 0.068 0.57 4.2 4.8Tap water #24 HMDE SW 2.50 5.1Seawater #15 TFME DP 0.0121 0.0086 10.7 8.1Seawater #25 TFME DP 0.032 0.032 6.3 6.3Soil extract #16 HMDE SW 189 11.8 2.5 5.6Soil extract #16 HMDE DP 186 11.9 2.5 4.0Deionized water4 HMDE SW 0.13 0.79 5.5 2.2Wastewater # 17 HMDE DP 74 26 4.3 4.5Wastewater #27 HMDE DP 47 86 5.2 6.3Wastewater #37 HMDE DP 46 65 4.6 6.2Wastewater #48 TFME SW 5.2 60 12 5.2 6.1 7.4

results. Blanks are critical because of the high sensitivity of themethod.

Table 3130:1 gives precision data for analyses of samples withvarious matrices.

1. WANG,J. 1985. Stripping Analysis: PrincipIes, Instrumentation, andApplications. VCH PubIishers, Inc., DeerfieId Beach, FIa.

2. BRAlNINA,K.H. & E. NEYMAN.1993. EIectroanalytical StrippingMethods. John WiIey & Sons, New York, N.Y.

3. KrSSINGER,PT. & W.R. HEINEMAN.1996. Laboratory Techniques inElectroanaIytical Chemistry, 2nd ed. MareeI Dekker, Inc., New York,N.Y.

4. MARTIN-OOLDBERG,M. 1989. Unpublished data. ResearchTriangleInstitute, Metais Analysis Facility. ResearchTriangle Park, N.C.

5. BRULAND,K.W., K.H. COALE& L. MART.1985. Analysis of seawaterfor dissolved Cd, Cu, and Pb: An intercomparison of voltammetricand atomic absorption methods. Mar. Chem.17:285.

6. OSTAPCZUK,P., P. VALENTA& H.W. NURNBERG.1986. Square wavevoltammetry-a rapid and reliable determination method of Zn, Cd,Pb, Ni, and Co in biological and environmental samples. J. Electro-ana!. Chem. 214:51.

Aluminum (AI) is the second element in Group IIIA of theperiodic table; it has an atomic number of 13, an atomicweight of 26.98, and a valence of 3. The average abundancein the earth's crust is 8.1 %; in soils it is 0.9 to 6.5%; instreams it is 400 fLglL; in V.S. drinking waters it is 54 fLglL,and in groundwater it is <0.1 fLg/L. Aluminum occurs in theearth's crust in combination with silicon and oxygen to form-feldspars, micas, and clay mineraIs. The most important min-eraIs are bauxite and corundum, which is used as an abrasive.Aluminum and its alloys are used for heat exchangers, aircraftparts, building materiaIs, containers, etc. Aluminum potas-sium sulfate (alum) is used in water-treatment processes toflocculate suspended particles, but it may leave a residue ofaluminum in the finished water.

* Approved by Standard Methods Committee, 200\.Joint Task Group: 20th Edition-Brian 1. Condike (chair), Deanna K. Anderson,Anthony Bright, Richard A. Cahi 11 , Alois F. Clary, C. Ellen Gomer, Peter M.Grohse, Daniel C. Hillman, Alben C. Holler, Amy Hughes, J. Charles Jennen,Roger A. Minear, Marlene O. Moore, Gregg L. Oelker, S. Kusum Perera, JamesG. Poff, Jeffrey G. Skousen, Michael D. Wichman, John L. Wuepper.

7. CLARK,B.R., D.W. DEPAOLI,D.R. McTAGGART& B.D. PATTON.1988.An on-line voltammetric analyzer for trace metaIs in wastewater.Ana!. Chim. Acta 215:13.

8. BRETT,C.M.A., A.M. OLIVElRA-BRETT& L. TUGULEA.1996. Anodicstripping voItammetry of trace metaIs by batch injection anaIysis.Ana!. Chim. Acta 322:151.

9. Bibliography

E.O. & O. PRINCETONApPLIEDRESEARCHCORPo1980. DifferentiaI pulseanoclic stripping voltammetry of water and wastewater. ApplicationNote W-1. Princeton, NJ.

PETERSON,W.M. & R.V. WONG.1981. Fundamentais of stripping voltam-metry. Amer. Lab. 13(11):116.

E.O. & O. PRINCETONAPPLlEDRESEARCHCORPo1982. Basics of voltam-metry and polarography. Application Note P-2. Princeton, NJ.

NURNBERG,H.W. 1984. The voltammetric approach in trace metal chem-istry of natural waters and atmospheric precipitation. Ana!. Chim.Acta 164:1.

SHUMAN,M.S. & MARTIN-OOLDBERG,M. 1984. Electrochemical meth-ods: Anodic stripping. In R. Minear & L. Keith, eds. Water Anal-ysis: Inorganic Species. 11:345. Academic Press, Orlando, FIa.

FLORENcE,T.M. 1986. Electrochemical approaches to trace eIementspeciation in waters: A review. Analyst 111:489.

FLORENcE,T.M. 1992. Trace element speciation by anodic strippingvoltammetry. Analyst 117:551.

TERClER,M.-L. & J. BUFFLE.1993. In situ voltammetric measurements innatural waters: Future prospects and challenges. Electroanalysis5:187.

Alurninum's occurrence in natural waters is controlled by pHand by very finely suspended mineral particles. The cation AI3+predominates at pH less than 4. Above neutral pH, the predom-inant dissolved form is AI(OH)4 -. Aluminum is nonessential forplants and animais. Concentrations exceeding 1.5 mglL consti-tute a toxicity hazard in the marine environment, and levelsbelow 200 fLglL present a rninimal risk. The Vnited NationsFood and Agriculture Organization's recommended maximumlevei for irrigation waters is 5 mglL. The possibility of a linkbetween elevated alurninum levels in brain tissues and Alzheimer'sdisease has been raised. The U.S. EP A secondary drinking waterregulations list an optimal secondary maximum contaminant levei(SMCL) of 0.05 mgIL and maximum SMCL of 0.2 mgIL.

The atomic absorption spectrometric methods (3111D andE, and 3113B) and the inductively coupled plasma methods(3120 and 3125) are free from such common interferences asfluoride and phosphate, and are preferred. The Eriochromecyanine R colorimetric method (B) provides a means forestimating aluminum with simpler instrumentation.

a. PrincipIe: With Eriochrome cyanine R dye, dilute alumi-num solutions buffered to a pH of 6.0 praduce a red to pinkcomplex that exhibits maximum absorption at 535 nm. Theintensity of the developed color is intluenced by the aluminumconcentration, reaction time, temperature, pH, alkalinity, andconcentration of other ions in thesample. To compensate forcolo r and turbidity, the aluminum in one portion of sample iscomplexed with EDTA to pravide a blank. The interference ofiran and manganese, two elements commonly found in waterwhen aluminum is present, is eliminated by adding ascorbic acid.The optimum aluminum range lies between 20 and 300 J.Lg/Lbutcan be extended upward by sample dilution.

b. Interference: Negative errors are caused by both tluorideand polyphosphates. When the tluoride concentration is con-stant, the percentage error decreases with increasing amountsof aluminum. Because the tluoride concentration often isknown or can be determined readily, fairly accurate resultscan be obtained by adding the known amount of tluoride to aset of standards. A simpler correction can be determined framthe family of curves in Figure 3500-AI:1. A procedure isgiven for the removal of complex phosphate interference.Orthophosphate in concentrations under 10 mg/L does notinterfere. The interference caused by even small amounts ofalkalinity is removed by acidifying the sample just beyond theneutralization point of methyl orange. Sulfate does not inter-fere up to a concentration of 2000 mg/L.

c. Minimum detectabIe concentration: The minimum alumi-num concentration detectable by this method in the absence oftluorides and complex phosphates is approximately 6 J.Lg/L.

d. SampIe handling: Collect samples in clean, acid-rinsedbottles, preferably plastic, and examine them as soon as possibleafter collection. If only soluble aluminum is to be determined,filter a portion of sample through a 0.45-J.Lm membrane filter;discard first 50 roL of filtrate and use succeeding filtrate for thedetermination. Do not use filter paper, absorbent couon, or glasswool for filtering any solution that is to be·tested for aluminum,because they will remove most of the soluble aluminum.

a. Colorimetric equipment: One of the following is required:1) Spectrophotometer, for use at 535 nm, with a light path of

I cm or longer.2) Filta photometer, providing a light path of I cm or longer

and equipped with a green filter with maximum transmittancebetween 525 and 535 nm.

3) Nessler tubes, 50-mL, tall form, matched.b. Glassware: Treat all glassware with warm I + I HCI and

rinse with aluminum-free distilled water to avoid errors due tomateriaIs absorbed on the glass. Rinse sufficiently to remove allacid.

Use reagents low in aluminum, and aluminum-free distilledwater.

a. Stock aluminum solution: Use either the metal (I) or thesalt (2) for preparing stock solution; 1.00 roL = 500 J.Lg AI:

I) Dissolve 500.0 mg aluminum metal in 10 roL conc HCI byheating gent1y. Dilute to 1000 roL with water, or

2) Dissolve 8.791 g aluminum potassium sulfate (alsocalled potassium aIum), AIK(S04h' 12H20, in water anddilute to 1000 mL. Correct this weight by dividing by thedecimal fraction of assayed AIK(S04h . 12H20 in the reagentused.

b. Standard aluminum solution: Dilute 10.00 roL stock alu-minum solution to 1000 roL with water; 1.00 roL = 5.00 J.Lg AI.Prepare daily.

c. Sulfuric acid, H2S04, 0.02N and 6N.d. Ascorbic acid solution: Dissolve 0.1 g ascorbic acid in

water and make up to 100 roL in a volumetric tlask. Prepare freshdaily.

e. Buffer reagent: Dissolve 136 g sodium acetate,aCZH30Z • 3HzO, in water, add 40 roL IN acetic acid, and

dilute to I L.i Stock dye solution: Use any of the following products:I) Solochrome cyanine R-200* or Eriochrome cyanine:t Dis-

solve 100 mg in water and dilute to 100 roL in a volumetric tlask.This solution should have a pH of about 2.9.

2) Eriochrome cyanine R::j: Dissolve 300 mg dye in about 50roL water. Adjust pH fram about 9 to about 2.9 with I + I aceticacid (approximately 3 mL will be required). Dilute with water to100 roL.

3) Eriochrome cyanine R:§ Dissolve 150 mg in about 50 roLwater. Adjust pH fram about 9 to about 2.9 with I + I aceticacid (appraximately 2 roL will be required). Dilute with water to100 roL.

Stock solutions have excellent stability and can be kept for atleast a year.

g. Working dye solution: Dilute 10.0 roL of selected stock dyesolution to 100 roL in a volumetric f1ask with water. Workingsolutions are stable for at least 6 months.

h. Methyl orange indicator solution, or bromcresol green in-dicator solution specified in the total alkalinity determination(Section 2320B.3d).

i. EDTA (sodium sal! of ethylenediamine-tetraacetic acid di-hydrate), O.OIM: Dissolve 3.7 g in water, and dilute to I L.

j. Sodium hydroxide, NaOH, IN and O.IN.

4. Procedure

a. Preparation of calibration curve:I) Prepare a series of aluminum standards from O to 7 J.Lg (O

to 280 J.LgfL based on a 25-roL sample) by accurately measuringthe calculated volumes of standard aluminum solution into50-roL volumetric f1asks or nessler tubes. Add water to a totalvolume of approximately 25 roL.

2) Add I mL 0.02N H2S04 to each standard and mix. Add IroL ascorbic acid solution and mix. Add 10 roL buffer solution

* Arnold Hoffman & Co., Providence, RI.t K & K Laboratories, K & K Lab. Div., Life Sciences Group, Plainview, NY.:j: Pfaltz & Bauer, Inc., Starnford, CT.§ EM Science, Gibbstown, NJ.

0.35 J;.q.~O

-J '~O

~~ -10-

c: 0.30O:;:Jl..'?

+"c:Q)u 0.25c:O(J

«"OQ)

0.20...:J(J)coQ)

~0.15

Figure 3500-AI:1. Correction curves for estimation of aluminum in the presence of fluoride. Above the mg F- fL present, locate the point correspondingto the apparent mg AIfL measured. From this point interpolate between the curves shown, if the point does not fali directly on one of thecurves, to read the !rue mg AIfL on the ordinate, which corresponds to 0.00 mg F- fL. For example, an apparent 0.20 mg AIfL in a samplecontaining 1.00 mg F- fL would actually be 0.30 mg AlfL if no ftuoride were present to interfere.

and mix. With a volumetric pipet, add 5.00 rnL working dyereagent and mix. Immediately make up to 50 rnL with distilledwater. Mix and let stand for 5 to 10 minoThe color begins to fadeafter 15 mino

3) Read transmittance or absorbance on a spectrophotometer,using a wavelength of 535 nm or a green filter providing max-imum transmittance between 525 and 535 nm. Adjust instrumentto zero absorbance with the standard containing no aluminum.

Plot concentration of AI (micrograms AI in 50 rnL finalvolume) against absorbance.

b. Sample treatment in absence of fiuoride and complex phos-phates: Place 25.0 rnL sample, or a portion diluted to 25 rnL, ina porcelain dish or fiask, add a few drops of methyl orangeindicator, and titrate with 0.02N H2S04 to a faint pink color.Record reading and discard sample. To two similar samples atroom temperature add the same amount of 0.02N H2S04 used inthe titration and 1 rnL in excesso

To one sample add 1 rnL EDTA solution. This will serve as ablank by complexing any aluminum present and compensatingfor color and turbidity. To both samples add 1 rnL ascorbic acid,

10 mL buffer reagent, and 5.00 mL working dye reagent asprescribed in <j[ a2) above.

Set instrument to zero absorbance or 100% transmittanceu ing the EDTA blank. After 5 to 10 min contact time, readtransmittance or absorbance and determine aluminum concen-tration from the calibration curve previously prepared.

c. Visual comparison: If photometric equipment is not avail-able, prepare and treat standards and a sample, as describedabove, in 50-mL nessler tubes. Make up to mark with water andcompare sample color with the standards after 5 to 10 mincontact time. A sample treated with EDTA is not needed whennessler tubes are used. If the sample contains turbidity or color,the use of nessler tubes may result in considerable error.

d. Removal ofphosphate inteiference: Add 1.7 mL 6NHzS04

to 100 mL sample in a 200-mL erlenmeyer f1ask. Heat on a hotplate for at least 90 min, keeping solution temperature just belowthe boiling point. At the end of the heating period solutionvolume should be about 25 mL. Add water if necessary to keepit at or above that volume.

After cooling, neutralize to a pH of 4.3 to 4.5 with NaOH,using IN NaOH at the start and O.IN for the final fine adjust-ment. Monitor with a pH meter. Make up to 100 mL with water,mix, and use a 25-mL portion for the aluminum test.

Run a blank in the same manner, using 100 mL distilled waterand 1.7 mL 6N HZS04. Subtract blank reading from samplereading or use it to set instrument to zero absorbance beforereading the sample.

e. Correction for samples containing fiuoride: Measure sam-pie f1uoride concentration by the SPADNS or electrode method.Either:

I) Add the same amount of f1uoride as in the sample to eachaluminum standard, or

2) Determine f1uoride correction from the set of curves inFigure 3500-AI: 1.

5. Calculation

""g AI (io 50 rnL final volume)mg AlJL = ----------

rnLsample

Antimony (Sb) is the fourth element in Group VA in theperiodic table; it has an atomic number of 51, an atomic weightof 121.75, and valences of 3 and 5. The average abundance of Sbin the earth's crust is 0.2 ppm; in soils it is 1 ppm; in streams itis I IJ-glL, and in groundwaters it is <0.1 mglL. Antimony issometimes found native, but more commonly in stibnite (SbZS3).

lt is used in alloys af lead and in batteries, bullets, solder,pyrotechnics, and semiconductors.

A synthetic sample containing 520 IJ-gAlIL and no interfer-ence in distilled water was analyzed by the Eriochrome cyanineR method in 27 laboratories. Relative standard deviation was34.4% and relative error 1.7%.

A second synthetic sample containing 50 IJ-gAlIL, 500 IJ-gBaIL, and 5 IJ-g BelL in distilled water was analyzed in 35laboratories. Relative standard deviation was 38.5% and relativeerror 22.0%.

A trurd synthetic sample containing 500 IJ-gAlIL, 50 IJ-gCdIL,11O IJ-gCrlL, 1000 IJ-gCulL, 300 IJ-gFelL, 70 IJ-gPblL, 50 IJ-gMnIL, ISO IJ-gAglL, and 650 IJ-gZnlL in distilled water wasanalyzed in 26 laboratories. Relative standard deviation was28.8% and relative error 6.2%.

A fourth synthetic sample containing 540 IJ-gAlIL and 2.5 mgpolyphosphatelL in distilled water was analyzed in 16 laborato-ries that hydrolyzed the sample in the prescribed manner. Rela-tive standard deviation was 44.3% and relative error 1.3%. ln 12laboratories that applied no corrective measures, the relativestandard deviation was 49.2% and the relative error 8.9%.

A fifth synthetic sample containing 480 IJ-gAIIL and 750 IJ-gFIL in distilled water was analyzed in 16 laboratories that reliedon the curve to correct for the f1uoride content. Relative standarddeviation was 25.5% and reiative error 2.3%. The 17 laboratoriesthat added f1uoride to the aluminum standards showed a reiativestandard deviation of 22.5% and a relative error of 7.1%.

SHULL,K.E. & G.R. GUTHAN.1967. Rapid modified Eriochrome cyanineR method for determination of aluminum in water. J. Amer. WaterWorks Assoe. 59:1456.

The common aqueous species are Sb02, HSbOz, and com-plexes with carbonate and sulfate. Soluble salts of antimony aretoxic. The U.S. EPA primary drinking water standard MCL is 6IJ-glL.

The electrothermal atomic absorption spectrometric method(3113B) or the inductively coupled plasma/mass spectrometricmethod (3125) are the methods of choice because of their sen-sitivity. Altematively use the f1ame atomic absorption spectro-metric method (3111B) or the inductively coupled plasmamethod (3120) when high sensitivity is not required.

Arsenic (As) is the third element in Group VA of the periodictable; it has an atomic number of 33, an atol1Úcweight of 74.92,and valences of 3 and 5. The average abundance of As in theearth's crust is 1.8 ppm; in soils it is 5.5 to 13 ppm; in streamsit is less than 2 fLglL, and in groundwater it is generally less than100 fLglL. It occurs naturally in sulfide I1Úneralssuch as pyrite.Arsenic is used in alloys with lead, in storage batteries, and inammunition. Arsenic compounds are widely used in pesticidesand in wood preservatives.

Arsenic is nonessential for plants but is an essential traceelement in severa I animal species. The predol1Únant form be-tween pH 3 and pH 7 is HzAs04 -, between pH 7 and pH 11 itis HAsO/-, and under reducing conditions it is HAsOz(aq) (orH3As03). Aqueous arsenic in the form of arsenite, arsenate, andorganic arsenicals may result from I1Úneraldissolution, industrialdischarges, or the application of pesticides. The chel1Úcalformof arsenic depends on its source (inorganic arsenic from miner-aIs, industrial discharges, and pesticides; organic arsenic fromindustrial discharges, pesticides, and biological action on inor-ganic arsenic).

Severe poisoning can arise from the ingestion of as little as100 mg arsenic trioxide; chronic effects may result from theaccumulation of arsenic compounds in the body at low intakelevels. Carcinogenic properties also have been imputed to ar-senic compounds. The toxicity of arsenic depends on its chem-ical formo Arsenite is many times more toxic than arsenate. For

* Approved by Standard Metbods Committee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

the protection of aquatic life, the average concentration of As3+in water should not exceed 72 fLglL and the maximum should notexceed 140 fLglL. The United Nations Food and AgricultureOrganization's recommended maximum levei for irrigation wa-ters is 100 fLglL. The U.S. EPA primary drinking water standardMCL is 0.05 mglL.

2. Selection of Method

Methods are available to identify and deterIlÚne total arsenic,arsenite, and arsenate. Unpolluted fresh water normally does notcontain organic arsenic compounds, but may contain inorganicarsenic compounds in the form of arsenate and arsenite. The elec-trothermal atol1Úcabsorption spectrometric method (3113B) is themethod of choice in the absence of overwhelrrúng interferences.The hydride generation-atol1Úcabsorption method (3114B) is pre-ferred when interferences are present that cannot be overcome bystandard electrothermal techniques (e.g., matrix modifiers, back-ground correction). The silver diethylditlúocarbamate method (B),in which arsine is generated by reaction with sodium borohydride inacidic solution, is applicable to the deterIlÚnationof total inorganicarsenic when interferences are absent and when the sample containsno methylarsenic compounds. This method also provides the ad-vantage of being able to identify and quantify arsenate and arseniteseparately by generating arsine at different pHs. The inductivelycoupled plasma (ICP) el1Ússion spectroscopy method (3120) isuseful at higher concentrations (greater than 50 fLgIL) while theICP-mass spectrometric method (3125) is applicable at lower con-centrations if cWoride does not interfere. When measuring arsenicspecies, document that speciation does not change over time. Nouniversal preservative for speciation measurements has been iden-tified.

a. Principie: Arsenite, containing trivalent arsenic, is reducedselectively by aqueous sodium borohydride solution to arsine,AsH3, in an aqueous medium of pH 6. Arsenate, methylarsonicacid, and dimethylarsenic acid are not reduced under theseconditions. The generated arsine is swept by a stream of oxygen-free nitrogen from the reduction vessel through a scrubber con-taining glass wool or cotton impregnated with lead acetate so-lution into an absorber tube containing silver diethyldithiocar-bamate and morpholine dissolved in chloroform. The intensity ofthe red color that develops is measured at 520 nm. To determinetotal inorganic arsenic in the absence of methylarsenic com-pounds, a sample portion is reduced at a pH of about l. Alter-natively, arsenate is measured in a sample from which arsenitehas been removed by reduction to arsine gas at pH 6 as above.

The sample is then acidified with hydrochloric acid and anotherportion of sodium borohydride solution is added. The arsineformed from arsenate is collected in fresh absorber solution.

b. Interferences: Although certain metals-chrol1Úum, co-balt, copper, mercury, molybdenum, nickel, platinum, silver, andselenium-influence the generation of arsine, their concentra-tions in water are seidom high enough to interfere, except in theinstance of acid rock drainage. HzS interferes, but the interfer-ence is removed with lead acetate. Antimony is reduced tostibine, which forms a colored complex with an absorptionmaximum at 510 nm and interferes with the arsenic determina-tion. Methylarsenic compounds are reduced at pH I to methyl-arsines, which form colored complexes with the absorber solu-tion. If methylarsenic compounds are present, measurements oftotal arsenic and arsenate are unreliable. The results for arseniteare not influenced by methylarsenic compounds.

Joint forconnection toCaCl, drying tube

/'

Absorbertube

1.8 em ID-- 2.0 em OD

PbAe-moistenedglass wool

Inert gasdelivery tube

"" Silver diethyldithiocarbamatereagent

NaBH, andsample port

a. Arsine generator, scrubber, and absorption tube: See Fig-ure 3500-As:1. Use a 200-rnL three-necked f1askwith a sidearm(19/22 or similar size female ground-glass joint) through whichthe inert gas delivery tube reaching almost to the bottom of thef1ask is inserted; a 24/40 female ground-glass joint to carry thescrubber; and a second side arm closed with a rubber septum, orpreferably by a screw cap with a hole in its top for insertion ofa TFE-faced silicone septum. Place a small magnetic stirring barin the f1ask. Fit absorber tube (20 rnL capacity) to the scrubberand fill with silver diethyldithiocarbamate solution. Do not userubber or cork stoppers because they may absorb arsine. Cleanglass equipment with coneentrated nitrie acid.

b. Fume hood: Use apparatus in a well-ventilated hood withf1ask secured on top of a magnetic stirrer.

c. Photometric equipment:I) Spectrophotometer, for use at 520 nm.2) Filter photometer, with green filter having a maXlmum

transmittance in the 500- to 540-nm range.3) Cells, for spectrophotometer or filter photometer, I-cm,

clean, dry, and each equipped with a tightly fitting cover (TFEstopper) to prevent chloroform evaporation.

b. Acetate buffer, pH 5.5: Mix 428 rnL 0.2M sodium acetate,NaCzH30Z' and 72 rnL 0.2M acetic acid, CH3COOH.

c. Sodium acetate, 0.2M: Dissolve 16.46 g anhydrous sodiumacetate or 27.36 g sodium acetate trihydrate, NaCzH30Z • 3HzO,in water. Dilute to 1000 rnL with water.

d. Acetic acid, CH3COOH, 0.2M: Dissolve 11.5 rnL glacialacetic acid in water. Dilute to 1000 rnL.

e. Sodium borohydride solution, I %: Dissolve 0.4 g sodiumhydroxide, NaOH (4 pellets), in 400 rnL water. Add 4.0 gsodium borohydride, NaBH4 (check for absence of arsenic).Shake to dissolve and to mix. Prepare fresh every few days.f Hydrochloric acid, HCI, 2M: Dilute 165 rnL cone HCI to

1000 rnL with water.g. Lead acetate solution: Dissolve 10.0 g Pb(CH3COO)z . 3HzO

in 100 rnL water.h. Silver diethyldithiocarbamate solution: Dissolve 1.0 rnL

morpholine (CAUTION:Corrosive-avoid contact with skin) in 70rnL chloroform, CHCI3. Add 0.30 g silver diethyldithiocarbam-ate, AgSCSN(CzHs)z; shake in a stoppered f1ask until most isdissolved. Dilute to 100 rnL with chloroform. Filter and store ina tightly c10sed brown bottle in a refrigerator.

i. Standard arsenite solution: Dissolve 0.1734 g NaAsOz inwater and dilute to 1000 rnL with water. CAUTION:Toxic-avoidcontact with skin and do not ingest. Dilute 10.0 rnL to 100 rnLwith water; dilute 10.0 mL of this intermediate solution to 100mL with water; 1.00 rnL = 1.00 J-Lg As.

j. Standardarsenatesolution:Dissolve0.416 g NazHAs04 • 7HzOin water and dilute to 1000 rnL. Dilute 10.0 rnL to 100 rnL withwater; dilute 10 rnL of this intermediate solution to 100 rnL; 1.00mL = 1.00 J-Lg As.

a. Arsenite:I) Preparation of scrubber and absorber-Dip glass wool into

lead acetate solution; remove excess by squeezing glass wool.Press glass wool between pieces of filter paper, then f1uff it.Alternatively, if cotton is used treat it similarly but dry in a desic-cator and f1uff thoroughly when dry. Place a plug of loose glasswool or cotton in scrubber tube. Add 4.00 rnL silver diethyldithio-carbamate solution to absorber tube (5.00 rnL may be used toprovide enough volume to rinse spectrophotometer cell).

2) Loading of arsine generator-Pipet not more than 70 rnLsample containing not more than 20.0 J-Lg As (arsenite) into thegenerator f1ask. Add 10 rnL acetate buffer. If necessary, adjusttotal volume of Iiquid to 80 rnL. Flush f1askwith nitrogen at therate of 60 rnL/min.

3) Arsine generation and measurement-While nitrogen ispassing through the system, use a 30-rnL syringe to injectthrough the septum 15 rnL 1% sodium borohydride solutionwithin 2 minoStir vigorously with magnetic stirrer. Pass nitrogenthrough system for an additional 15 min to f1ush arsine intoabsorber solution. Pour absorber solution into a clean and dryspectrophotometric cell and measure absorbance at 520 nmagainst chloroform. Determine concentration from a caJibrationcurve obtained with arsenite standards. If arsenate also is to bedeterrnined for this sample by using the same sample portion,save the liquid in the generator f1ask.

4) Preparation of standard curves- Treat standard arsenitesolution containing 0.0, 1.0, 2.0, 5.0, 10.0, and 20.0 J-Lg As

described in CJ[s1) through 3) above. Plot absorbance versusmicrograms arsenic in the standard.

b. Arsenate: After removal of arsenite as arsine, treat sampleto convert arsenate to arsine:

If the lead acetate-impregnated glass wool has become inef-fective in removing hydrogen sulfide (if it has become gray toblack) replace glass wool [see CJ[aI)]. Pass nitrogen throughsystem at the rate of 60 rnL/min. Cautiously add 10 rnL 2.0NHCI. Generate arsine as directed in CJ[4a3) and prepare standardcurves with standard solutions of arsenate according to proce-dure of CJ[4a4).

c. Total inorganic arsenic: Prepare scrubber and absorber asdirected in CJ[4al) and load arsine generator as directed in CJ[4a2)using 10 rnL 2.0N HCI instead of acetate buffer. Generate arsineand measure as directed in CJ[4a3). Prepare standard curvesaccording to CJ[4a4). Curves obtained with standard arsenitesolution are almost identical to those obtained with arsenatestandard solutions. Therefore, use either arsenite or arsenatestandards.

Calculate arsenite, arsenate, and total inorganic arsenic fromreadings and caIibration curves obtained in 4a, b, and c, respec-tively, as follows:

Barium (Ba) is the fifth element in Group IIA in theperiodic table; it has an atomic number of 56, an atomicweight of 137.33, and a valence of 2. The average abundanceof Ba in the earth' s crust is 390 ppm; in soils it is 63 to 810ppm; in streams it is 10 mg/L; in V.S. drinking waters it is 49iLg/L; and in groundwaters it is 0.05 to I mg/L. It is foundchiefly in barite (BaS04) or in witherite (BaC03). Barium'smain use is in mud slurries used in drilling oil and explorationwells, but it is also used in pigments, rat poisons, pyrotech-nics, and in medicine.

Beryl1ium (Be) is the first element in Group IIA of the periodictable; it has an atomic number of 4, an atomic weight of 9.01,

p..g As (from calibration curve)mg As/L = -----------

rnL sample in generator flask

Interlaboratory comparisons are not available. The relativestandard deviation of results obtained with arsenite/arsenate mix-tures containing approximately 10 iLg arsenic was less than 10%.

PEOPLES,S.A., J. LAKSO & T. LAls. 1971. The simultaneous determina-tion of methylarsonic acid and inorganic arsenic in urine. Proe.West. Pharmacol. Soe. 14:178.

AGGETI, J. & A.C. ASPELL. 1976. Determination of arsenic (Ill) and totalarsenic by the silver diethyldithiocarbamate method. Analyst 101:912.

HOWARD, A.O. & M.H. ARBAB-ZAVAR. 1980. Sequential spectrophoto-metric determination of inorganic arsenic (III) and arsenic (V)species. Analyst 105 :338.

PANDE,S.P. 1980. Morpholine as a substitute for pyridine in determina-tion of arsenic in water. J. Inst. Chem. (India) 52:256.

IRGOLIC,KJ. 1986. Arsenic in the environment. ln A. V. Xavier, ed.Frontiers in Bioinorganic Chernistry. VCH Publishers, Weinheim,Oermany.

IRGOLIC,KJ. 1987. AnalyticaJ procedures for the determination of or-ganic compounds of metais and metalloids in environmental sam-pies. Sei. Total Environ. 64:61.

The solubility of barium in natural waters is controlled by thesolubility of BaS04, and somewhat by its adsorption on hydrox-ides. High concentrations of barium occur in some brines. Con-centrations exceeding I mg/L constitute a toxicity hazard in themarine environment. The V.S. EPA primary drinking waterstandard MCL is I mg/L.

Perform analyses by the atomic absorption spectrometricmethods (3111D or E), the electrothermal atomic absorptionmethod (3113B), or the inductively coupled plasma methods(3120 or 3125).

and a valence of 2. The average abundance of Be in the earth'scrust is 2 ppm; in soils it is 0.8 to 1.3 ppm; in streams it is 0.2iLg/L, in V.S. drinking waters and in groundwaters it is typically<0.1 iLg/L. Beryllium occurs in nature in deposits of beryls ingranitic rocks. Beryllium is used in high-strength alloys of cop-

per and nickel, windows in X-ray tubes, and as a moderator innuclear reactors.

Beryllium solubility is controlled in natural waters by thesolubility of beryllium hydroxides. The solubility at pH 6.0 isapproximately 0.1 J.Lg/L. It is nonessential for plants and animais.Acute toxicity occurs at 130 J.Lg/L, and chronic toxicity at 5 J.Lg/Lin freshwater species. The United Nations Food and AgricultureOrganization recommended maximum levei for irrigation watersis 100 J.Lg/L. The U.S. EPA primary drinking water standardMCL for beryllium is 4 J.Lg/L.

Bismuth (Bi) is the fifth element in Group VA in the periodictable; it has an atomic number of 83, an atomic weight of 208.98,and valences of 3 and 5. The average abundance of Bi in theearth's crust in 0.08 ppm; in streams it is <0.02 mg/L, and ingroundwaters it is <0.1 mg/L. Bismuth occurs in associationwith lead and silver ores, and occasionally as the native element.2IOBi,212Bi, and 214Biare naturally occuring radioisotopes pro-duced in the decay of uranium and thorium. The metal is used inalloys of lead, tin, and cadmium, and in some pharmaceuticals.

The atornic absorption spectrometric methods (3111D and E,and 3113B) and the inductively coupled plasma (ICP) methods(3120 and 3125) are the methods of choice. If atomic absorptionor ICP instrumentation is not available, the aluminon colorimet-ric method detailed in the 19th Edition of Standard Methods maybe used. This method has poorer precision and bias than themethods of choice.

In natural water, Bi3+ ion will occur, and complex ions withnitrate and chloride also might be expected. The iodide andtelluride compounds are toxic by ingestion or inhalation.

Perform analyses by the atornic absorption spectrometricmethod (3111B) or by the electrothermal atomic absorptionmethod (3113B). The inductively coupled plasma mass spectro-metric method (3125) also may be applied successfully in mostcases (with lower detection levels), even though bismuth is notspecifically listed as an analyte in the method.

Cadmium (Cd) is the second element in Group IIB of theperiodic table; it has an atomic number of 48, an atomic weightof 112.41, and a valence of 2. The average abundance of Cd inthe earth's crust is 0.16 ppm; in soils it is 0.1 to 0.5 ppm; instreams it is 1 J.Lg/L, and in groundwaters it is from 1 to 10 J.Lg/L.Cadrnium occurs in sulfide minerais that also contain zinc, lead,or copper. The metal is used in electroplating, batteries, paintpigments, and in alloys with various other metais. Cadmium isusually associated with zinc at a ratio of about 1 part cadrniumto 500 parts zinc in most rocks and soils.

The solubility of cadrnium is controlled in natural waters bycarbonate equilibria. Guidelines for maximum cadrnium concen-trations in natural water are linked to the hardness or alkalinityof the water (i.e., the softer the water, the lower the permittedlevei of cadmium). It is nonessential for plants and animais.

Cadrnium is extremely toxic and accumulates in the kidneysandliver, with prolonged intake at low levels sometimes leadingto dysfunction of the kidneys. The United Nations Food andAgriculture Organization recommended maximum leveI for cad-rnium in irrigation waters is 10 J.Lg/L. The U.S. EPA primarydrinking water standard MCL is 10 J.Lg/L.

The electrothermal atomic absorption spectrometric method(3l13B) is preferred. The f1ame atomic absorption methods(3l11B and C) and inductively coupled plasma methods (3120and 3125) provide acceptable precision and bias, with higherdetection limits. Anodic stripping voltammetry (3130B) canachieve superior detection lirnits, but is susceptible to interfer-ences from copper, silver, gold, and organic compounds. Whenatornic absorption spectrometric or inductively coupled plasmaapparatus is unavailable and the desired precision is not as great,the dithizone method detailed in the 19th Edition of StandardMethods is suitable.

Calcium (Ca) is the third element in Group lIA of the periodictable; it has an atomic number of 20, an atomic weight of 40.08, anda valence of 2. The average abundance of Ca in the earth' s crust is4.9%; in soils it is 0.07 to 1.7%; in streams it is about 15 mg!L; andin groundwaters it is from 1 to > 500 mg!L. The most commonforms of calcium are calcium carbonate (calcite) and calcium-magnesium carbonate (dolomite). Calcium compounds are widelyused in pharmaceuticals, photography, lime, de-icing salts, pig-ments, fertilizers, and plasters. Calcium carbonate solubility is con-trolled by pH and dissolved COz. The COz, HC03 -, and C03 z-equilibrium is the major buffering mechanism in fresh waters.Hardness is based on the concentration of calcium and magnesiumsalts, and often is used as a measure of potable water quality.

NOTE:Calcium is necessary in plant and animal nutrition andis an essential component of bones, shells, and plant structures.The presence of calcium in water supplies results from passageover deposits of limestone, dolomite, gypsum, and gypsiferousshale. Small concentrations of calcium carbonate combat corro-

* Approved by Standard Methods Committee, 1997.Joint Task Group: 20th Edition-See 3500-AJ.

sion of metal pipes by laying down a protective coating. Becauseprecipitation of calcite in pipes and in heat-exchangers can causedamage, the amount of calcium in domestic and industrial watersis often controlled by water softening (e.g., ion exchange, re-verse osmosis). Calcium carbonate saturation and water hardnessare discussed in Sections 2330 and 2340, respectively.

Calcium contributes to the total hardness of water. Chemicalsoftening treatment, reverse osmosis, electrodialysis, or ion ex-change is used to reduce calcium and the associated hardness.

The atomic absorption methods (3111B, D, and E) and induc-tively coupled plasma method (3120) are accurate means ofdetermining calcium. The EDTA titration method gives goodresults for control and routine applications, but for samplescontaining high P levels (>50 mglL) only the atomic absorptionor atomic emission methods are recommended because of inter-ferences often encountered with EDTA indicators.

The customary precautions are sufficient if care is taken toredissolve any calcium carbonate that may precipitate on standing.

a. Principie: When EDTA (ethylenediaminetetraacetic acidor its salts) is added to water containing both calcium andmagnesium, it combines first with the calcium. Calcium can bedetermined directly, with EDTA, when the pH is made suffi-ciently high that the magnesium is largely precipitated as thehydroxide and an indicator is used that combines with calciumonly. Several indicators give a color change when all of thecalcium has been complexed by the EDTA at a pH of 12 to 13.

b. Interference: Under conditions of this test, the followingconcentrations of ions cause no interference with the calciumhardness deterrnination: Cuz+, 2 mglL; FeH, 20 mglL; FeH, 20mglL; MnH, 10 mglL; Znz+, 5 mglL; Pbz+, 5 mglL; AIH, 5mglL; and Sn4+, 5 mglL. Orthophosphate precipitates calcium atthe pH of the testo Strontium and barium give a positive inter-ference and alkalinity in excess of 300 mglL may cause anindistinct end point in hard waters.

b. Indicators: Many indicators are available for the calciumtitration. Some are described in the literature (see Bibliography);others are commercial preparations and also may be used. Murexide(ammonium purpurate) was the first indicator available for detectingthe calcium end point; directions for its use are presented in thisprocedure. Individuals who have difficulty recognizing the murex-ide end point may find the indicator Eriochrome Blue Black R(color index number 202) or Solochrome Dark Blue an improve-ment because of the color change from red to pure blue. EriochromeBlue Black R is sodium-I-(2-hydroxy-I-naphthylazo)-2-naphthol-4-sulfonic acid. Other indicators specifica11ydesigned for use asend-point detectors in EDTA titration of calcium may be used.

I) Murexide (ammoniumpurpurate) indica to r: This indicatorchanges from pink to purple at the end point. Prepare by dis-solving 150 mg dye in 100 g absolute ethylene glycol. Watersolutions of the dye are not stable for longer than 1 d. A groundmixture of dye powder and sodium chloride (NaCl) provides astable form of the indicator. Prepare by mixing 200 mg murexidewith 100 g solid NaCl and grinding the mixture to 40 to 50 mesh.Titrate immediately after adding indicator because it is unstableunder alkaline conditions. Facilitate end-point recognition bypreparing a color comparison blank containing 2.0 mL NaOHsolution, 0.2 g solid indicator mixture (or 1 to 2 drops if a

olurion is used), and sufficient standard EDT A titrant (O.OS to0.10 mL) to produce an unchanging colar.

2) Eriochrome Blue Black R indicator: Prepare a stable formof the indicator by grinding together in a mortar 200 mg pow-dered dye and 100 g solid NaCI to 40 to SO mesh. Store in atightly stoppered bottle. Use 0.2 g of ground mixture for thetitration in the same manner as murexide indicator. Duringtitration the colar changes from red through purple to bluishpurple to apure blue with no trace of reddish or purple tinto ThepH of some (not all) waters must be raised to 14 (rather than 12to 13) by the use of 8N NaOH to get a good calor change.

c. Standard EDTA titrant, O.OlM: Prepare standard EDTA titrantand standardize against standard calcium solution as described inSection 2340C to obtain EDTA/CaC03 equivalence. StandardEDTA titrant, O.OlOOM,is equivalent to 1.000 mg CaC03/1.00 mL;use titrated equivalent for B in the calculations in 4.

a. Pretreatment ofwater and wastewater samples: Follow theprocedure described in Section 3030E ar I if samples requirepreliminary digestion.

b. Sample preparation: Because of the high pH used in thisprocedure, titrate immediately after adding alkali and indicator.Use SO.O rnL sample, ar a smaller portion diluted to SO mL sothat the calcium content is about S to 10 mg. Analyze hard waterswith alkalinity higher than 300 mg CaC031L by taking a smallerportion and diluting to 50 rnL. Altematively, adjust sample pHinto the acid range (pH <6), boil for I min to dispel CO2, andcoa I before beginning titration.

c. Titration: Add 2.0 rnL NaOH solution ar a volume sufficientto produce a pH of 12 to 13. Stir. Add 0.1 to 0.2 g indicator mixtureselected (ar 1 to 2 drops if a solution is used). Add EDTA titrantslowly, with continuous stirring to the proper end point. When usingmurexide, check end point by adding 1 to 2 drops of titrant in excessto make certain that no further colar change occurs.

Cesium (Cs) is the sixth element in Group IA of the periodictable; it has an atomic number of S5, an atomic weight of 132.90,and a valence of 1. The average abundance of Cs in the earth'scrust is 2.6 ppm; in soils it is 1 to S ppm; in streams it is 0.02mgIL; and in groundwaters it is generally <0.1 mglL. Cesium isfound in lepidolite and in the water of certain mineral springs.137Cs, with a 33-year half-life, is widely dispersed on the earth's

A x B x 400.8mg Ca/L = -----

rnL sample

A x B x 1000Calcium hardness as mg CaCOiL = -----

rnLsample

where:

A = rnL titrant for sample andB = mg CaC03 equivalent to 1.00 mL EDTA titrant at the

calcium indicator end poinl.

A synthetic sample containing 108 mg CaIL, 82 mg MglL, 3.1mg KIL, 19.9 mg NaIL, 241 mg Cl-IL, 1.1 mg N03 - -NIL, 0.25mg N02 - -NIL, 259 mg S042-1L, and 42.S mg total alkalinitylL(contributed by NaHC03) in distilled water was analyzed in 44laboratories by the EDTA titrimetric method, with a relativestandard deviation of 9.2% and a relative errar of 1.9%.6. Bibliography

DLEHL,H. & J.L. ELLINGBOE.1956. Indicator for titration of calcium inthe presence of magnesium using disodium dihydrogen ethylenedi-amine tetraacetate. Ana!. Chem. 28:882.

HILDEBRAND,G.P. & C.N. REILLEY.1957. New indicator for complexo-metric titration of calcium in the presence of magnesium. Ana!.Chem. 29:258.

PAITON,J. & W. REEDER.1956. New indicator for titration of calciumwith (ethylenedinitrilo) tetraacetate. Ana!. Chem. 28:1026.

SCHWARZENBACH,G. 1957. Complexometric Titrations. Interscience Pub-lishers, New York, N.Y.

FURMAN,N.H. 1962. Standard Methods of Chemical Analysis, 6th ed. D.Van Nostrand Co., Inc., Princeton, NJ.

KATZ,H. & R. NAVONE.1964. Method for simultaneous determination ofcalcium and magnesium. J. Amer. Water Works Assoe. 56:121.

surface as a result of the radioactive fallout from the atmospherictesting of nuclear weapons. Cesium compounds are used inphotoelectric cells, as a catalyst, and in brewing. Some cesiumcompounds are fire hazards.

Perform analyses by the f1ame atomic absorption spectromet-ric method (3lllB). The inductively coupled plasma/mass spec-trometric method (3125) also may be applied successfully inmost cases (with lower detection levels), even though cesium isnot specifically listed as an analyte in the method.

Chromium (Cr) is the first element in Group VIB in the periodictable; it has an atomic number of 24, an atomic weight of 51.99, andvalences of O and 2 through 6. The average abundance of Cr in theearth's crust is 122 ppm; in soils Cr ranges from 11 to 22 ppm; instreams it averages about 1 J.Lg/L, and in groundwaters it is generally100 J.Lg/L. Chromium is found chiefly in chrome-iron ore(FeO·CrZ03). Chromium is used in alloys, in electroplating, and inpigments. Chromate compounds frequently are added to coolingwater for corrosion controI.

In natural waters trivalent chromium exists as Cr3+,Cr(OH)z+, Cr(OH)z +, and Cr(OH)4-; in the hexavalent formchromium exists as CrO/- and as CrzO/-. Cr3+ would beexpected to form strong complexes with amines, and would beadsorbed by clay mineraIs.

Chromium is considered nonessential for plants, but an essen-tial trace element for animaIs. Hexavalent compounds have beenshown to be carcinogenic by inhalation and are corrosive totissue. The chromium guidelines for natural water are linked tothe hardness or alkalinity of the water (i.e., the softer the water,the lower the permitted leveI for chromium). The United NationsFood and Agriculture Organization recommended maximumleveI for irrigation waters is 100 J.LglL. The U.S. EPA primarydrinking water standard MCL is 100 J.LglL for total chromium.

* Approved by Standard Methods Committee, 2001.Joint Task Group: Rock J. Vitale (chair), C. Ellen Gonter, Timothy S. Oostdyk; 20thEdition-See 3500-AI.

The colorimetric method (B) is useful for the determination ofhexavalent chromium in a natural or treated water in the rangefrom 100 to 1000 J.LglL. This range can be extended by appro-priate sample dilution or concentration and/or use of longer cellpaths. The ion chromatographic method with photometric detec-tion (C) is suitable for determining dissolved hexavalent chro-mium in drinking water, groundwater, and industrial wastewatereffluents from 0.5 to 5000 J.LglL. The electrothermal atomicabsorption spectrometric method (3113B) is suitable for deter-mining low levels of total chromium « 50 J.LglL) in water andwastewater, and the f1ameatomic absorption spectrometric meth-ods (3111B and C) and the inductively coupled plasma methods(3120 and 3125) are appropriate for measuring total chromiumconcentrations up to milligram-per-liter levels.

If only the dissolved chromium content is desired, filter sam-pIe immediately through a 0.45-J.Lm membrane filter at time ofcollection, and acidify filtrate with conc nitric acid (HN03) to pH<2. If only dissolved hexavalent chromium is desired, adjust pHof filtrate to 9 with lN sodium hydroxide solution and refrigerateto 4 ± 2°C. If the total chromium content is desired, acidifyunfiltered sample at time of collection with conc HN03 to pH<2. If total hexavalent chromium is desired, adjust pH of unfil-tered sample to 9 with lN sodium hydroxide and refrigerate to 4± 2°C.

a. Principie: This procedure measures only hexavalent chro-miunf, Cr(VI). For total chromium determination, acid-digest thesample (see Section 3030) and follow with a suitable instrumen-tal analysis technique. The hexavalent chromium is determinedcolorimetrically by reaction with diphenylcarbazide in acid so-lution. A red-violet colored complex of unknown composition isproduced. The reaction is very sensitive, the molar absorptivitybased on chromium being about 40000 L g-J cm-1 at 530 or540 nm. NOTE: Validation data for Method 3500-Cr.B weredeveloped at 540 nm. Validation data for Method 3500-Cr.Cwere developed at 530 nm.

b. lnterferences: The reaction with diphenylcarbazide is nearlyspecific for chromium. Hexavalent molybdenum and mercurysalts will react to form color with the reagent but the intensitiesare much lower than that for chromium at the specified pH.Concentrations of Mo or Hg as high as 200 mglL can betolerated. Vanadium interferes strongly but concentrations up to

10 times that of chromium will not result in significant analyticalerror. Iron in concentrations greater than 1 mglL may produce ayellow color but the femc ion (Fe3+) color is not strong and nodifficulty is encountered normally if the absorbance is measuredphotometrically at the appropriate wavelength.

2. Apparatus

a. Colorimetric equipment: One of the following is required:1) Spectrophotometer, for use at 530 or 540 nm, with a light

path of 1 cm or longer.2) Filter photometer, providing a light path of 1 cm or longer

and equipped with a greenish yellow filter having maximumtransmittance at 530 or 540 nm.

b. Laboratory ware: Soak alI reusable items (glass, plastic,etc.), including sample containers, overnight in laboratory-gradedetergent, rinse, and soak for 4 h in a mixture of nitric acid (lpart), hydrochloric acid (2 parts), and reagent water (9 parts).Rinse with tap water and reagent water. NOTE:Never use chro-mic acid cleaning solution.

Use reagent water (see Section 1080) for reagent preparationand analytical procedure.

a. Stock chromium solution: Dissolve 141.4 mg K2Cr207 inwater and dilute to 100 rnL; 1.00 rnL = 500 J-tg Cr. CAUTION:Hexavalent chromium is toxic and a suspected carcinogen. Han-dle with care.

b. Standard chromium solution: Dilute 1.00 mL stock chro-mium solution to 100 rnL; 1.00 rnL = 5.00 fLg Cr. Preparecalibration standards at time of analysis.

c. Nitric acid, HN03, conc.d. Sulfuric acid, H2S04, conc, 18N, and 6N.e. Sulfuric acid, H2S04, 0.2N: Dilute 17 rnL 6N H2S04 to 500

mL with water.f Phosphoric acid, H3P04, conc.g. Diphenylcarbaúde solution: Dissolve 250 mg I,5-diphe-

nylcarbazide (I,5-diphenylcarbohydrazide) in 50 rnL acetone.Store in a brown bottle. Prepare weekly. Discard if the solutionbecomes discolored.

h. Sodium hydroxide, IN: Dissolve 40 g NaOH in 1 L water.Store in plastic bottle.

a. Preparation of calibration curve: To compensate for pos-sible slight losses of chromium during analytical operations, treatstandards and samples with the same procedure. Accordingly,pipet measured volumes of standard chromium solution (5 J-tg/rnL) ranging from 2.00 to 20.0 rnL, to give standards for 10 to100 J-tg Cr, into 250-rnL beakers or conical f1asks.Depending onpretreatrnent used in q[ b below, proceed with subsequent treat-ment of standards as if they were samples.

Develop color as for samples, transfer a suitable portion ofeach colored solution to a I-cm absorption ceU, and measureabsorbance at 540 nm, using reagent water as reference. Correctabsorbance readings of standards by subtracting absorbance of areagent blank carried through the method.

Construct a calibration curve by plotting corrected absorbancevalues against micrograms chromium in 102 rnL final volume.

b. Treatment of sample: If sample has been filtered and/or onlyhexavalent chromium is desired, start analysis within 24 h ofcoUection and proceed to q[ 4c. NOTE:Recent evidence suggeststhat preserved samples can be held for 30 d without substantialchanges to Cr (VI) concentrations.I•2

c. Color development and measurement: Add 0.25 rnL (5drops) H3P04. Use 0.2N H2S04 and a pH meter to adjustsolution to pH 2.0 ::!: 0.5. Transfer solution to a 100-rnL volu-metric f1ask, dilute to 100 rnL, and mix. Add 2.0 rnL diphenyl-carbazide solution, mix, and let stand 5 to 10 min for fuU colordevelopment. Transfer an appropriate portion to a I-cm absorp-tion ceU and measure its absorbance at 530 or 540 nm, usingreagent water as reference. Correct absorbance reading of sampleby subtracting absorbance of a blank carried through the method(see also note below). From the corrected absorbance, detenninemicrograms chromium present by reference to the calibration curve.

NOTE:If the solution is turbid after dilution to 100 rnL in q[ cabove, take an absorbance reading before adding carbazide re-agent and correct absorbance reading of final colored solution bysubtracting the absorbance measured previously.

J.Lg Cr (in 102 mL final volume)mg CrlL = A

where:A = mL original sample.

6. Precision and Bias

Collaborative test data from 16 laboratories were obtained onreagent water, tap water, 10% NaCI solution, treated water fromsynthetic organic industrial waste, EPA extraction leachate, processwater, lake water, and effluent from a steel pickle liquor treatmentplant.3 The test data yielded the following relationships: '

S, = 0.037x + 0.006So = 0.022x + 0.004

Drinking or wastewater:SI 0.067x + 0.004So = 0.037x + 0.002

SI 0.032x + 0.007So 0.017x + 0.004

where:S, = overall precision,So = single-operator precision, andx = chromium concentration, mglL.

1. D.S. ENVIRONMENTALPROTECl10N AGENCY. 1996. Determination ofhexavalent chromium by ion chromatography. Method 1636. EPA821-R-96-003, D.S. Environmental Protection Agency, Washington,D.C.

2. EATON, A., A. HAGHANI& L. RAMIREZ.2001. The Erin BrockovichfactoL In Proe. Water Quality Technology Conf. (Nashville, Tenn.,November 11-15, 2001). American Water Works Assoe., Denver, Colo.

3. AMERICAl'-1SOCIETYFOR TESTING AND MATERlALS. 1996. Chromium,total. DI687-92, Annual Book of ASTM Standards, Vol. 11.01.American Soe. Testing & Materiais, Philadelphia, Pa.

ROWLAND,G.P., IR. 1939. Photoelectric colorimetry-Optical study ofperrnanganate ion and of chromium-diphenylcarbazide system.Ana!. Chem. 11 :442.

DRONE, P.F. 1955. Stability of colorimetric reagent for chromium, 5-di-phenylcarbazide, in various solvents. Ana!. Chem. 27: 1354.

ALLEN, T.L. 1958. Microdetermination of chromium with 1,5-diphenyl-carbohydrazide. Ana!. Chem. 30:447.

ON1SHl,H. 1979. Colorimetric Deterrnination ofTraces ofMetals, 4th ed.Interscience Publishers, New York, N.Y.

a. PrincipIe: This method is applicable to determination of dis-solved hexavalent chromium in drinking water, groundwater, andindustrial wastewater effluents. An aqueous sample is filtered andits pH adjusted to 9 to 9.5 with a concentrated buffer. This pHadjustrnent reduces the solubility of trivalent chromium and pre-serves the hexavalent chromium oxidation state. The sample isintroduced into the instrument's eluent stream of ammonium sulfateand ammonium hydroxide. Trivalent chromium in solution is sep-arated from the hexavalent chromium by the column. After sepa-ration, hexavalent chromiurn reacts with an azide dye to produce achromogen that is measured at 530 or 540 um. NOTE:Validationdata for 3500-Cr.C were developed at 530 um. Hexavalent chro-mium is identified on the basis of retention time.

Although this method was developed using specific commer-cial equipment, use of another manufacturer's equipment shouldbe acceptable if appropriate adjustments are made.

b. Interferences: Interferences may come from severalsources. Use analytical grade salts for the buffer because traceamounts of chromium may be included.

Several soluble species of trivalent chromium in the samplemay be oxidized to the hexavalent forro in an alkaline medium inthe presence of such oxidants as hydrogen peroxide, ozone, andmanganese dioxide. The hexavalent forro can be reduced to thetrivalent in the presence of reducing species in an acid medium.

High ionic concentration may cause column overload. Sam-pies high in chloride and/or sulfate might show this phenome-non, which is characterized by a change in peak geometry.

Interfering organic compounds are removed by the guardcolumn.

c. Minimum detectable concentrations: The method detectionlevels obtained in a single laboratory with a 250-J..loLloop were asfollows:

ReagentwaterDrinkingwaterGroundwaterPrimary wastewatereftluentElectroplatingwaste

0.4 p.gIL0.3 p.gIL0.3 p.g/L0.3 p.g/L0.3 p.g/L

d. Sample preservation and holding time: Filter samplethrough a 0.45-J..lom filter. Use a portion of sample to rinse syringefilter unit and filter, then collect the required volume of filtrate.Adjust pH to 9 to 9.5 by adding buffer solution dropwise whilechecking pH with a pH meter.

Ship and store sample at 4°C. Bring to room temperaturebefore analysis. Analyze samples within 24 h of collection.

a. Ion chromatograph equipped with a pump capable of pre-cisely delivering a f10wof 1 to 5 rnL/min. The metallic parts ofthe pump must not contact sample, eluent, ar reagent. Sampleloops should be available ar the instrument should be capable ofdelivering from 50 to 250-J..loLinjections of sample. The visibleabsorption cell should not contain metallic parts that contact theeluent-sample f1ow. The cell must be usable at 530 nm. Use

plastic pressurized containers to deliver eluent and post-columnreagent. Use high-purity helium (99.995%) to pressurize theeluent and post-column reagent vessel.

b. Guard column, to be placed before the separator column,containing an adsorbent capable of adsorbing organic com-pounds and particulates that would damage or interfere with theanalysis or equipment. *

C. Separator column, packed with a high-capacity anion-ex-change resin capable of resolving chromate from other sampleconstituents. t

d. Recorder, integrator, or computer for receiving signalsfrom the detector as a function of time.

e. Laboratory ware: Soak ali reusable items (glass, plastic,etc.) including sample containers, ovemight in laboratory-gradedetergent, rinse, and soak for 4 h in a mixture of nitric acid (1part), hydrochloric acid (2 parts), and reagent water (9 parts).Rinse with tap water and reagent water. NOTE:Never use chro-mic acid cleaning solution.f Syringe, equipped with male luer-type fitting and a capacity

of at least 3 rnL.

a. Reagent water: Deionized or distilled water free frominterferences at the minimum detection levei of each constituent,filtered through a 0.2-J..lom membrane filter and having a conduc-tance of less than 0.1 J..loS/cm. Use for preparing ali reagents (seeSection 1080).

b. Cr(VI) stock solution, 100 mg Cr6+ IL: Prepare from pri-mary standard grade potassium dichromate. Dissolve 0.1414 gK2Cr207 in water and dilute to 500 rnL in a volumetric f1ask.pHadjustment is not required. Store in plastic. CAUTION:Hexavalentchromium is toxic and a suspected carcinogen; handle with care.

c. Eluent: Dissolve 33 g ammonium sulfate (NH4hS04' in 500rnL water and add 6.5 rnL conc ammonium hydroxide, NH40H.Dilute to 1 L with water.

d. Post-column reagent: Dissolve 0.5 g 1,5-diphenylcarbazidein 100 rnL HPLC-grade methanol. Add with stirring to 500 rnLwater containing 28 rnL conc H2S04. Dilute to 1 L with water.Reagent is stable for 4 or 5 d; prepare only as needed.

e. Buffer solution: Dissolve 33 g ammonium sulfate,(NH4hS04' in 75 rnL water and add 6.5 rnL conc ammoniumhydroxide, NH40H. Dilute to 100 rnL with water.

a. Instrument setup: Establish ion chromatograph operatingconditions as indicated in Table 3500-Cr:I. Set f10w rate of theeluent pump at 1.5 rnL/min and adjust pressure of reagentdelivery module so that the system flow rate, measured after thedetector, is 2.0 rnL/min. Measure system f10w rate using agraduated cylinder and stopwatch. Allow approximately 30 minafter adjustment before measuring flow.

* IonPac NGl, Dionex, 4700 Lakeside Drive, Sunnyvale, CA 94086, or equivalent.t IonPac AS7, Dionex, or equivalent.

Ouard columnSeparator columnEluent

Dionex IonPac NO IDionex IonPac AS7250 mM (NH4)2S04100 mM NH40H1.5 mL/min2 mM diphenylcarbohydrazide10% v/v CH30HIN H2S04

0.5 mL/minVisible @ 530 nm3.8 min

Eluent f10w ratePost-column reagent

Post-column reagent f10w rateDetectorRetention time

Use an injection loop size based on required sensitivity. A50-p,L loop is sufficient, although a 250-p,L loop was used todetermine the method detection leveI.

b. Calibration: Before sample analysis, construct a calibrationcurve using a minimum of a blank and three standards thatbracket the expected sample concentration range. Prepare cali-bration standards from the stock standard (3b) by appropriatedilution with reagent water in volumetric flasks. Adjust to pH 9to 9.5 with buffer solution (3e) before final dilution. Injectionvolumes of standards should be about 10 times the injection loopvolume to insure complete loop flushing.

c. Sample analysis: Bring chilled, pH-adjusted sample toambient temperature. Fill a clean syringe with sample, attach a0.45-p,m syringe filter, and inject 10 times the sample loopvolume into the instrumento Dilute any sample that has a con-centration greater than the highest calibration standard.

Determine area or height of the Cr(VI) peak in the calibrationstandard chromatograms. Establish a calibration curve by per-forming a regression analysis of peak height or peak area againststandard concentration in mg/L. If correlation coefficient is lessthan 0.995, do not use these data. Check analytical processo

For samples, measure area (or height) of Cr(VI) peak insample chromatogram, as determined by retention time. Calcu-late Cr(VI) concentration by interpolating from the calibrationline. Correct data for any dilutions made.

Currently available instrumentation automates the emire mea-surement process (peak measurement, calibration, and sample mea-surements and calculations). Ensure that enough quality controlsamples are analyzed to monitor the instrumental processes.

a. Initial demonstration of peiformance: Before sample anal-ysis, set up instrument and analyze enough known samples todetermine estimates for the method detection leveI and linearcalibration range. Use the initial demonstration of performanceto characterize instrument performance, i.e., method detectionlevels (MDLs) and linear calibration range.

b. Initial and continuing calibration peiformance: Initially,after every 10 samples, and after the final sample, analyze anindependent check sample and a calibration blank. The concentra-tion of the calibration check sample should be near the mid-cali-

METALS (3000)

TABLE3500-Cr:II. SINGLE-LABORATORYPRECISIONANDElAS

Concentration* Mean RecoverySample Type IJ-g/L % RPDt

Reagent water 100 100 0.81000 100 0.0

Drinking water 100 105 6.71000 98 1.5

Oroundwater 100 98 0.01000 96 0.8

Primary wastewater 100 100 0.7efftuent 1000 104 2.7

Electroplating 100 99 0.4effluent 1000 101 0.4

* Sample fonified at this concentration leveI.t RPD - relative percenl difference between fortified duplicates.

bration range; prepare from a source independent of the calibrationstandards. Use acceptance criteria for check standard recovery andcalibration standard concentration based on project goals for preci-sion and accuracy. Typical values for recovery of the check stan-dard range from 90 to 110%. The acceptance criteria for the cali-bration blank are typically set at :!: the nominal MDL.

C. Reagent blank analysis: Analyze one laboratory reagentblank with each batch of samples. Significant Cr(VI) detected inthe reagent blank is a sign of contamination. Identify source andeliminate contamination. ~

d. LaboratoryJortified matrix (also known as matrix spike) anal-ysis: To a portion of a sample, add a known quantity of Cr(VI).After analysis, calculate percent recovery of the known addition. Ifthe recovery falls outside the controllimits (typically 75 to 125%),the matrix may be interfering with the analysis. If matrix interfer-ences are thought to be present, use the method of standard addi-tions, if appropriate, to minirnize the effect of interferences. NOTE:There are situations where low recoveries of analyte additions canbe overcome, or where the non-detected analyte results can befurther validated by the application of three-point method of stan-dard addition. The decision to apply method of standard additionsshould be a function of achieving desired data quality objectives.Analyze fortified matrix samples as frequently as dictated by projectgoals and anticipated similarity of matrices in the sample set.

e. Laboratory contral sample: Analyze a laboratory controlsample (LCS) from an external source with every sample batch.Process LCS and samples identically, including filtering and pHadjustment. Base acceptance criteria for LCS recovery on projectgoals for precision and bias. Typical values for acceptable re-covery range from 90 to 110%.

The instrument operating conditions and data from a single-laboratory test of the method are shown in Tables 3500-Cr:I and3500-Cr:II, respectively.

Multilaboratory test data are shown in Table 3500-Cr:III.:j:Fifteen laboratories analyzed samples ranging from 1.2 to 960p,g Cr/L.

t The multilaboratory precision and bias dala cited in this method were lhe resullor a collaborative study carried out jointly between U.S. EPA EnvironmentalMonitoring Systems Laboratory (Cincinnati) and Committee D-19 or ASTM.

COPPER (3500-Cu)/lntroduction

TABLE3500-Cr:I1I. MULTILABORATORYDETERMINATIONOFBIASFORHEXAV ALENTCHROMIUM*

Amount AmountAdded Found Bias

Water J.Lg/L J.Lg/L S, So %

Reagent 6.00 6.68 1.03 0.53 +11.38.00 8.64 1.10 +8.0

16.0 17.4 2.25 0.77 +8.820.0 21.4 2.31 +7.0

100 101 1.91 3.76 +1.0140 143 5.52 +2.1800 819 24.3 12.7 +2.4960 966 18.5 +7.3

Waste 6.0 5.63 1.17 0.55 -6.28.0 7.31 1.91 -8.6

16.0 15.1 2.70 1.85 -5.620.0 19.8 1.01 -1.0

100 98.9 4.36 3.31 -1.1140 138 8.39 -1.4800 796 60.6 27.1 -0.5960 944 72.1 -1.7

* Each Youdenpair was used to calculateone laboratorydata point (Sol.

For reagent water matrix:

So = 0.033x + 0.106

where:

So = single-operator precision,S, = overall precision, andx = added amount.

So = 0.041x + 0.039S, = 0.059x + 1.05

Mean recovery = 0.989x - 0.41

The eleven water samples consisted of a reagent water blankand five Youden pairs. The nine wastewater samples consisted ofa wastewater blank and four Youden pairs.

V.S. ENVIRONMENTALPROTECTIONAGENCY.1991. Methods for the Determi-nation of Metals in Environrnental Samples. Method 218.6, EPA-600/4-91-010, Environrnental Monitoring Systems Lab., Cincinnati, Ohio.

DIONEX.1990. Technical Note No. 26. Dionex, Sunnyvale, Calif.EDGEL,K.W., I.A. LONGBOITOM& R.A. IOYCE.1994. Determination of

dissolved hexavalent chrornium in drinking water, ground water,and industrial wastewater effluents by ion chromatography: Collab-orative study. J. Assoe. Offie. Ana!. Chem. 77:994.

Cobalt (Co) is the second element in Group VIII in theperiodic table; it has an atomic number of 27, an atomic weightof 58.93, and valences of 1,2, and 3. The average abundance ofCo in the earth's crust is 29 ppm; in soils it is 1.0 to 14 ppm; instreams it is 0.2 fLglL; and in groundwaters it is 1 to 10 fLglL.Cobalt occurs only sparingly in ores, usually as the sulfide or thearsenide. It is widely used in alloys of various steels, in electro-plating, in fertilizers, and in porcelain and glass.

The solubility of cobalt is controlled by coprecipitation oradsorption by oxides or manganese and iron, by carbonate pre-

Copper (Cu) is the first element in Group IB in the periodictable; it has an atomic number of 29, an atomic weight of 63.54,

* Approvedby StandardMethodsComminee, 1999.Joint Task Group: 20th Edition-See 3500-AI.

cipitation, and by the formation of complex ions. Cobalt dust isflammable and is toxic by inhalation. Cobalt is considered es-sential for algae and some bacteria, nonessential for higherplants, and an essential trace element for animaIs. The UnitedNations Food and Agriculture Organization recommended max-imum leveI for irrigation waters is 100 fLglL.

Perform analyses by the flame atomic absorption spectromet-ric methods (3111B and C), by the electrothermal atomic ab-sorption method (3113B), or by the inductively coupled plasmamethods (3120 and 3125).

and valences of 1 and 2. The average abundance of Cu in theearth's crust is 68 ppm; in soils it is 9 to 33 ppm; in streams itis 4 to 12 fLglL; and in groundwater it is <0.1 mglL. Copperoccurs in its native state, but is also found in many mineraIs, themost important of which are those containing sulfide compounds(e.g., chalcopyrite), but also those with oxides and carbonates.Copper is widely used in e1ectrical wiring, roofing, various

alloys, pigments, cooking utensils, piping, and in the chemicalindustry. Copper salts are used in water supply systems tocontrol biological growths in reservoirs and distribution pipesand to catalyze the oxidation of manganese. Copper forms anumber of complexes in natural waters with inorganic and or-ganic ligands. Among the common aqueous species are Cu2+,Cu(OH)z, and CuHC03 +. Corrosion of copper-containing alloysin pipe fittings may introduce measurable amounts of copper intothe water in a pipe system.

Copper is considered an essential trace element for plants andanimais. Some compounds are toxic by ingestion or inhalation.The United Nations Food and Agriculture Organization recom-mended maximum levei for irrigation waters is 200 J.LglL. Underthe lead-copper mie, the U.S. EPA drinking water 90th percen-tile action levei is 1.3 mglL.

The atomic absorption spectrometric methods (311lB and C),the inductively coupled plasma methods (3120 and 3125), andthe neocuproine method (B) are recommended because of theirfreedom from interferences. The electrothermal atomic absorp-tion method (3113B) also may be used with success with anappropriate matrix modifier. The bathocuproine method (C) maybe used for potable waters.

3. Sampling and Storage

Copper ion tends to be adsorbed on the surface of sample con-tainers. Therefore, analyze samples as soon as possible after collec-tion. If storage is necessary, use 0.5 rnL I + 1 HClIlOOrnL sample,or acidify to pH <2 with HN03, to prevent this adsorption.

a. Principie: Cuprous ion (Cu+) in neutral or slightly acidicsolution reacts with 2,9-dimethyl-l, IO-phenanthroline (neocu-proine) to form a complex in which 2 moles of neocuproine arebound by I mole of Cu + ion. The complex can be extracted bya number of organic solvents, induding a chloroform-methanol(CHClrCH30H) rnixture, to give a yellow solution with a molarabsorptivity of about 8000 at 457 nm. The reaction is virtuallyspecific for copper; the color follows Beer's law up to a con-centration of 0.2 mg Cu/25 rnL solvent; full color developmentis obtained when the pH of the aqueous solution is between 3 and9; the color is stable in CHCI3-CH30H for several days.

The sample is treated with hydroxylamine-hydrochloride toreduce cupric ions to cuprous ions. Sodium citrate is used tocomplex metallic ions that rnight precipitate when the pH israised. The pH is adjusted to 4 to 6 with NH40H, a solution ofneocuproine in methanol is added, and the resultant complex isextracted into CHCI3. After dilution of the CHCI3 to an exactvolume with CH30H, the absorbance of the solution is measuredat 457 nm.

b. lnteiference: Large amounts of chrornium and tin mayinterfere. Avoid interference from chromium by adding sulfu-rous acid to reduce chromate and complex chromic ion. ln thepresence of much tin or excessive amounts of other oxidizingions, use up to 20 rnL additional hydroxylamine-hydrochloridesolution.

Cyanide, sulfide, and organic matter interfere but can beremoved by a digestion procedure (see Section 3030).

c. Minimum detectable concentration: The rninimum detect-able concentration, corresponding to 0.01 absorbance or 98%transrnittance, is 3 J.Lg Cu when a I-cm cell is used and 0.6 J.LgCu when a 5-cm cell is used.

a. Colorimetric equipment: One of the following is required:1) Spectrophotometer, for use at 457 nm, providing a light

path of I cm or longer.

2) Filter photometer, providing a light path of I cm or longerand equipped with a narrow-band violet filter having maximumtransrnittance in the range 450 to 460 nm.

b. Separatory funnels, 125-mL, Squibb form, with glass orTFE stopcock and stopper.

a. Redistilled water, copper-free: Because most ordinary dis-tilled water contains detectable amounts of copper, use redistilledwater, prepared by distilling singly distilled water in a resistant-glass still, or distilled water passed through an ion-exchange unit,to prepare ali reagents and dilutions.

b. Stock copper solution: To 200.0 mg polished electrolyticcopper wire or foil in a 250-rnL conical flask, add 10 rnL waterand 5 rnL conc HN03. After the reaction has slowed, warrngently to complete dissolution of the copper and boi] to expel oxidesof nitrogen, using precautions to avoid loss of copper. Cool, addabout 50 rnL water, transfer quantitatively to a l-L volumetric flask,and dilute to the mark with water; 1 rnL = 200 J.Lg Cu.

c. Standard copper solution: Dilute 50.00 rnL stock coppersolution to 500 rnL with water; 1.00 rnL = 20.0 J.Lg Cu.

d. Sulfuric acid, H2S04, conc.e. Hydroxylamine-hydrochloride solution: Dissolve 50 g

NH20H . HCI in 450 rnL water.f Sodium citrate solution: Dissolve 150 g Na3C6H507 • 2H20 in

400 rnL water. Add 5 rnL NH20H . HCI solution and 10 rnLneocuproine reagent. Extract with 50 rnL CHCI3 to remove copperimpurities and discard CHCI3 layer.

g. Ammonium hydroxide, NH40H, 5N: Dilute 330 rnL concNH40H (28-29%) to 1000 rnL with water. Store in a polyethyl-ene bottle.

h. Congo red paper, or other pH test paper showing a colorchange in the pH range of 4 to 6.

i. Neocuproine reagent.· Dissolve 100 mg 2,9-dimethyl-I,IO-phenanthroline hemihydrate* in 100 mL methanol. This

solution is stable under ordinary storage conditions for amonth or more.

j. Chloroform, CHCI3: Avoid or redistill material that comesin containers with metal-lined caps.

k. Methanol, CH30H, reagent grade.l. Nitric acid, HN03, conc.m. Hydrochloric acid, HCI, conc.

a. Preparation of calibration curve: Pipet 50 rnL water into a125-rnL separatory funnel for use as a reagent blank. Preparestandards by pipetting 1.00 to 10.00 rnL (20.0 to 200 JLg Cu)standard copper solution into a series of 125-rnL separatoryfunnels, and dilute to 50 rnL with water. Add 1 rnL conc H2S04

and use the extraction procedure given in lJ[ 4b below.Construct a calibration curve by plotting absorbance versus

micrograms of copper.To prepare a calibration curve for smaller amounts of copper,

dilute 10.0 rnL standard copper solution to 100 rnL. Carry 1.00-to 1O.00-rnL volumes of this diluted standard through the pre-viously described procedure, but use 5-cm cells to measureabsorbance.

b. Treatment of sample: Transfer 100 rnL sample to a 250-rnLbeaker, add 1 rnL conc H2S04 and 5 rnL conc HN03. Add a fewboiling chips and cautiously evaporate to dense white S03 fumeson a hot plate. If solution remains colored, cool, add another 5rnL conc HN03, and again evaporate to dense white fumes.Repeat, if necessary, until solution becomes colorless.

Cool, add about 80 rnL water, and bring to a boil. Cool andfilter into a 100-rnL volumetric flask. Make up to 100 rnL withwater using mostly beaker and filter washings.

Pipet 50.0 rnL or other suitable portion containing 4 to 200 JLgCu, from the solution obtained from prelirninary treatment, intoa 125-rnL separatory funnel. Dilute, if necessary, to 50 rnL withwater. Add 5 rnL NH20H . HCl solution and 10 rnL sodiumcitrate solution, and rnix thoroughly. Adjust pH to approximately

4 by adding 1-rnL increments of NH40H until Congo red paperis just definitely red (or other suitable pH test paper indicates avalue between 4 and 6).

Add 10 rnL neocuproine reagent and 10 rnL CHC13. Stopperand shake vigorously for 30 s or more to extract the copper-neocuproine complex into the CHC13. Let mixture separate intotwo layers and withdraw lower CHC13 layer into a 25-rnLvolumetric flask, taking care not to transfer any of the aqueouslayer. Repeat extraction of the water layer with an additional 10rnL CHC13 and combine extracts. Dilute combined extracts to 25rnL with CH30H, stopper, and mix thoroughly.

Transfer an appropriate portion of extract to a suitable absorp-tion cell (1 cm for 40 to 200 JLg Cu; 5 cm for lesser amounts) andmeasure absorbance at 457 nm or with a 450- to 460-nm filter.Use a sample blank prepared by carrying 50 rnL water throughthe complete digestion and analytical procedure.

Determine micrograms copper in final solution by reference tothe appropriate calibration curve.

p.g Cu (in 25 mL final volume)mgCufL =. .

mL portlOn taken for extractlOn

SMITH, G.F. & W.H. MCCURDY. 1952. 2,9-Dimethyl-I,IO-phenanthro-line: New specific in spectrophotometric determination of copper.Ana/. Chem. 24:371.

LUKE, C.L. & M.E. CAMPBELL. 1953. Determination of impurities ingermanium and silicon. Ana/. Chem. 25:1586.

GAHLER, A.R. 1954. Colorimetric determination of copper with neocu-proine. Ana/. Chem. 26:577.

FULTON,J.W. & J. HASTINGS.1956. Photometric determinations of cop-per in alurninum and lead-tin solder with neocuproine. Ana/. Chem.28:174.

FRANK,A.J., A.B. GOULSTON& A.A. DEACUTIS.1957. Spectrophotomet-ric determination of copper in titanium. Ana/. Chem. 29:750.

a. Principle: Cuprous ion forms a water-soluble orange-col-ored chelate with bathocuproine disulfonate (2,9-dimethyl-4,7-diphenyl-l,lO-phenanthrolinedisulfonic acid, disodium salt).While the color forms over the pH range 3.5 to 11.0, therecommended pH range is between 4 and 5. The sample isbuffered at a pH of about 4.3 and reduced with hydroxylarninehydrochloride. The absorbance is measured at 484 nm. Themethod can be applied to copper concentrations up to at least 5mglL with a sensitivity of 20 JLglL.

b. Inteiference: The following substances can be toleratedwith an error of less than ± 2%: Cyanide, thiocyanate, persul-fate, and EDTA also can interfere.

c. Minimum detectable concentration: 20 JLglL with a 5-cmcell.

ConcentrationmglL

CationsAlurninumBerylliumCadmiumCalciumChrornium (I1I)Cobalt (11)Iron (11)Iron (I1I)LithiumMagnesiumManganese (11)Nickel (11)Sodium

10010

1001000

105

100100500100500500

1000

ConcentrationmglL

StrontiumThorium (IV)Zinc

AnionsChlorateChlorideFluorideNitrateNitriteOrthophosphatePerchlorateSulfate

CompoundsResidual chlorineLinear alkylate sulfonate (LAS)

10001000500200200

100010001000

a. Colorimetric equipment: One of the fol1owing, with a lightpath of 1 to 5 cm (unless nessler tubes are used):

1) Spectrophotometer, for use at 484 nm.2) Filter photometer, equipped with a blue-green filter exhib-

iting maximum light transmission near 484 nm.3) Nessler tubes, matched, 100-rnL, tal1 formob. Acid-washed glassware: Rinse all glassware with conc HCI

and then with copper-free water.

a. Copper-free water: See Method B, <J[ 3a.b. Stock copper solution: Prepare as directed in Method B, <J[

3b, but use 20.00 mg copper wire or foil; 1.00 rnL = 20.00 fLgCU.

C. Standard copper solution: Dilute 250 rnL stock coppersolution to 1000 rnL with water; 1.00 rnL = 5.00 fLg CU. Preparedaily.

d. Hydrochloric acid, HCI, 1 + 1.e. Hydroxylamine hydrochloride solution: See Method B, <J[ 3e.f Sodium citrate solution: Dissolve 300 g Na3C6Hs07 • 2H20 in

water and make up to 1000 mL.

Gallium (Ga) is the third element in Group IIIA in the periodictable; it has an atomic number of 31, an atomic weight of 67.72,and valences of 1, 2, and 3. The average abundance of Ga in theearth's crust is 19 ppm; in soils it is 1.9 to 29 ppm; in streams itis 0.09 fLg/L; and in groundwaters it is <0.1 mg/L. Galliumoccurs in many zinc ores, and nearly always in bauxite. Galliumcompounds are used in semiconducting devices.

g. Disodium bathocuproine disulfonate solution: Dissolve1.000 g Cl2H4N2(CH3hCC6H4h (S03Nah in water and make upto 1000 rnL.

Pipet 50.0 rnL sample, ar a suitable portion diluted to 50.0 rnL,into a 250-rnL erlenmeyer flask. In separate 250-mL erlenmeyerflasks, prepare a 50.0-rnL water blank and a series of 50.0-rnLcopper standards containing 5.0, 10.0, 15.0, 20.0, and 25.0 fLgCU. To sample, blank, and standards add, rnixing after eachaddition, 1.00 rnL 1 + 1 HCI, 5.00 rnL NH20H . HCI solution,5.00 rnL sodium citrate solution, and 5.00 rnL disodium batho-cuproine disulfonate solution. Transfer to cel1s and read sampleabsorbance against the blank at 484 nm. Plot absorbance againstmicrograms Cu in standards for the calibration curve. Estimateconcentration from the calibration curve.

p,g Cu (in 66 mL final volume)mg Cu/L = -----------

mLsample

A synthetic sample containing 1000 fLg Cu/L, 500 fLg Al/L, 50fLg Cd/L, 110 fLg Cr/L, 300 fLg Fe/L, 70 fLg Pb/L, 50 fLg Mn/L,150 fLg Ag/L, and 650 fLg Zn/L was analyzed in 33 laboratoriesby the bathocuproine method, with a reI ative standard deviationof 4.1 % and a relative error of 0.3%.

SMITH,G.F. & D.H. WILKINS.1953. New colorimetric reagent specificfor copper. Ana!. Chem. 25:510.

BORCHARDT,L.G. & J.P. BUTLER.1957. Determination of trace amountsof copper. Anal. Chem. 29:414.

ZAK,B. 1958. Simple procedure for the single sample determination ofserum copper and iron. Clinica Chim. Acta 3:328.

BLAIR,D. & H. DlEHL. 1961. Bathophenanthrolinedisulfonic acid andbathocuproinedisulfonic acid, water soluble reagents for iron andcopper. Talanta 7: 163.

The element exists as Ga3+ in natural water, and its solubilityis controlled by formation of the hydroxide. It is considerednonessential for plants and animaIs.

Perform analyses by the electrothermal atomic absorptionmethod (3113B). The inductively coupled plasma/mass spectro-metric method (3125) also may be applied successful1y in mostcases (with lower detection levels), even though gallium is notspecifical1y listed as an analyte in the method.

Germanium (Ge) is the third element in Group IVA in theperiodic table; it has an atomic number of 32, an atomic weightof 72.59, and valences of 2 and 4. The average abundance of Gein the earth's crust is 1.5 ppm; in streams it is 0.03 to 0.1 fJ-glL;and in groundwaters it is <0.1 mglL. Germanium is found ingermanite, in certain zinc ores, and in elevated levels in certainhot spring waters. Germanium alloys are used in transistars, goldalloys, phosphors, and semiconducting devices.

Gold (Au) is the third element in Group IB in the periodictable; it has an atomic number of79, an atomic weight of 196.97,and valences of 1 and 3. The average abundance of Au in theearth's crust is 0.004 ppm; in streams it is 2 fJ-glL; and ingroundwater it is <0.1 mglL. Gold occurs in the native form, andis associated with quartz or pyrite. The main uses of gold are injewelry, dentistry, electronics, and the aerospace industry.

Gold solubility is restricted to acidic waters in the presence ofoxidizing agents and chloride, ar in alkaline solutions in the

Indium (In) is the fourth element in Group IIIA in the periodictable; it has an atomic number of 49, an atomic weight of 114.82,and valences of 1, 2, and 3. The average abundance of indium inthe earth's crust is 0.19 ppm; in streams it is <0.01 fJ-glL; and ingroundwaters it is <0.1 mglL. Indium often occurs in combina-tion with zinc ores, and sometimes with pyrites and siderite.Indium is used in alloys for bearings, brazing, solder, and inelectrical devices.

Iridium (Ir) is the eighth element in Group VIII of the periodictable; it has an atomic number of 77, an atomic weight of 192.2,and valences of 1, 3, and 4. The average abundance of Ir in theearth's crust is probably <0.001 ppm, and in groundwaters it is<0.1 mglL. Iridium occurs uncombined with platinum and othermetaIs. It is used in alloys with platinum in catalysts, thermo-couples, electrodes, and wires.

Germanium is present in natural waters in the tetravalent state,and its distribution in natural waters probably is controlled byadsorption on clay mineral surfaces. It is nonessential for plantsand animaIs.

Perform analyses by the electrothermal atomic absorptionmethod (3113B). The inductively coupled plasma/mass spectro-metric method (3125) also may be applied successfully in mostcases (with lower detection levels), even though germanium isnot specifically listed as an analyte in the method.

presence of hydrogen sulfide. Its solubility may be influenced bynatural organic acids. Compounds of gold containing thiosulfateand cyanide have some human toxicity.

Perform analyses by the atomic absorption spectrometricmethod (3111B) ar by the electrothermal atomic absorptionmethod (3113B). The inductively coupled plasma mass spectro-metric method (3125) also may be applied successfully in mostcases (with lower detection levels), even though gold is notspecifically listed as an analyte in the method.

Indium exists as In3+ and as a number of complex ions. Itssolubility is controlled by the formation of the insoluble hydrox-ide. The metal and its compounds are toxic by inhalation.

Perform analyses by the electrothermal atomic absorptionmethod (3113B). The inductively coupled plasma mass spectro-metric method (3125) also may be applied successfully in mostcases (with lower detection levels), even though indium is notspecifically listed as an analyte in the method.

The aqueous chemistry is controlled by complex compounds,although the solubility in natural waters is relatively unknown.

Perform analyses by the flame atomic absorption spectromet-ric method (3111B). The inductively coupled plasma/mass spec-trometric method (3125) also may be applied successfully inmost cases (with lower detection levels), even though iridium isnot specifically listed as an analyte in the method.

Iron (Fe) is the first element in Group VIII of the periodictable; it has an atomic number of 26, an atomic weight of 55.85,and common valences of 2 and 3 (and occasionally valences of1,4, and 6). The average abundance of Fe in the earth's crust is6.22%; in soils Fe ranges from 0.5 to 4.3%; in streams itaverages about 0.7 mglL; and in groundwater it is 0.1 to 10mglL. Iron occurs in the mineraIs hematite, magnetite, taconite,and pyrite. It is widely used in steel and in other alloys.

The solubility of ferrous ion (Fe2+) is controlled by thecarbonate concentration. Because groundwater is often anoxic,any soluble iron in groundwater is usually in the ferrous state. Onexposure to air or addition of oxidants, ferrous iron is oxidizedto the femc state (Fe3+) and may hydrolyze to form red, insol-uble hydrated femc oxide. In the absence of complex-formingions, femc iron is not significantly soluble unless the pH is verylow.

Elevated iron levels in water can cause stains in plumbing,laundry, and cooking utensils, and can impart objectionabletastes and colors to foods. The United Nations Food and Agri-culture Organization recommended leveI for irrigation waters is5 mglL. The U.S. EPA secondary drinking water standard MCLis 0.3 mglL.

Sensitivity and detection levels for the atomic absorptionspectrometric methods (31l1B and C), the inductively coupledplasma method (3120), and the phenanthroline colorimetric pro-cedure described here (B) are similar and generally adequate foranalysis of natural or treated waters. Lower detection levels canbe achieved with electrothermal atomic absorption spectrometry(31l3B) when an appropriate matrix modifier is used. The com-plexing reagents used in the colorimetric procedures are specificfor ferrous iron but the atomic absorption procedures are notoHowever, because of the instability of ferrous iron, which ischanged easily to the femc form in solutions in contact with airdetermination of ferrous iron requires special precautions andmay need to be done in the field at the time of sample collection.

The procedure for deterrnining ferrous iron using 1,10-phenanthroline (3500-Fe.BAc) has a somewhat limited applica-bility; avoid long storage time or exposure of samples to light. Arigorous quantitative distinction between ferrous and femc ironcan be obtained with a special procedure using bathophenanth-roline. Spectrophotometric methods using bathophenanthro-linel-6 and other organic complexing reagents such as ferrozine7

or TPTZ8 are capable of determining iron concentrations as lowas 1 fLglL. A chemiluminescence procedure9 is stated to have a

* Approvedby StandardMethodsCommittee,1997.Joint Task Group: 20th Edition-See 3500-AI.

detection limit of 5 nglL. Additional procedures are describedelsewhere.lO-13

Plan in advance the methods of collecting, storing, and pre-treating samples. Clean sample container with acid and rinsewith reagent water. Equipment for membrane filtration of sam-ples in the field may be required to determine iron in solution(dissolved iron). Dissolved iron, considered to be that passingthrough a OA5-fLm membrane filter, may include colloidal iron.The value of the deterrnination depends greatly on the care takento obtain a representative sample. Iron in well or tap watersamples may vary in concentration and forrn with duration anddegree of flushing before and during sampling. When taking asample portion for determining iron in suspension, shake thesample bottle often and vigorously to obtain a uniform suspen-sion of precipitated iron. Use particular care when colloidal ironadheres to the sample bottle. This problem can be acute withplastic bottles.

For a precise determination of total iron, use a separate con-tainer for sample collection. Treat with acid at the time ofcollection to place the iron in solution and prevent adsorption ordeposition on the walls of the sample container. Take account ofthe added acid in measuring portions for analysis. The additionof acid to the sample may eliminate the need for adding acidbefore digestion (3500-Fe.BAa).

1. LEE,G.F. & W. STUMM.1960. Determination of ferrous iron in thepresence of ferric iron using bathophenanthroline. J. Amer. WaterWorks Assoe. 52:1567.

2. GHOSH,M.M., J.T. O'CONNOR& R.S. ENGELBRECHT.1967. Batho-phenanthroline method for the determination of ferrous iron.J. Amer. Water Works Assoe. 59:897.

3. BLAIR,D. & H. DIEHL.1961. Bathophenanthroline-disulfonic acidand bathocuproine-disulfonic acid, water soluble reagents for ironand copper. Talanta 7: 163.

4. SHAPIRO,J. 1966. On the measurement of ferrous iron in naturalwaters. Limnol. Oeeanogr. 11:293.

5. McMAHON,J.W. 1967. The influence of light and acid on themeasurement of ferrous iron in lake water. Limno!. Oceanogr.12:437.

6. McMAHON,J.W. 1969. An acid-free bathophenanthroline methodfor measuring dissolved ferrous iron in lake water. Water Res.3:743.

7. GIBBS,C. 1976. Characterization and application of ferrozine ironreagent as a ferrous iron indicator. Ana!. Chem. 48:1197.

8. DOUGAN,W.K. & A.L. WILSON.1973. Absorbtiometric determina-tion of iron with TPTZ. Water Treat. Exam. 22:110.

9. SEITZ, W.R. & D.M. HERCULES.1972. Determination of traceamounts of iron (11)using chemiluminescence analysis. Ana!. Chem.44:2143.

10. Moss, M.L. & M.G. MELLON.1942. Co1orimetric detennination ofiron with 2,2'-bipyridine and with 2,2 ',2"-tripyridine. Ind. Eng.Chem., Ana!. Ed. 14:862.

11. WELCHER,FJ. 1947. Organic Ana1ytica1Reagents. D. Van NostrandCo., Princeton, NJ., Vol. 3, pp. 100-104.

12. MORRlS,R.L. 1952. Detennination of iron in water in the presenceof heavy metaIs. Ana!. Chem. 24:1376.

13. DOIG,M.T., III & D.F. MARTIN.1971. Effect of humic acids on imnana1yses in natural water. Water Res. 5:689.

1. General Discussion c. Separatory funnels: 125-rnL, Squibb form, with ground-glass or TFE stopcocks and stoppers.

a. PrincipIe: Iron is brought into solution, reduced to the ferrousstate by boiling with acid and hydroxylamine, and treated with1,IO-phenanthroline at pH 3.2 to 3.3. Three molecules of phenan-throline chelate each atom of ferrous iron to form an orange-redcomplex. The colored solution obeys Beer's law; its intensity isindependent of pH from 3 to 9. A pH between 2.9 and 3.5 insuresrapid color development in the presence of an excess of phenanth-roline. Color standards are stable for at least 6 months.

b. Interference: Among the interfering substances are strongoxidizing agents, cyanide, nitrite, and phosphates (polyphos-phates more so than orthophosphate), chromium, zinc in con-centrations exceeding 10 times that of iron, cobalt and copper inexcess of 5 mg/L, and nickel in excess of 2 mg/L. Bismuth,cadmium, mercury, molybdate, and silver precipitate phenanth-roline. The initial boiling with acid converts polyphosphates toorthophosphate and removes cyanide and nitrite that otherwisewould interfere. Adding excess hydroxylarnine eliminates errorscaused by excessive concentrations of strong oxidizing reagents.In the presence of interfering metal ions, use a larger excess ofphenanthroline to replace that complexed by the interferingmetais. Where excessive concentrations of interfering metal ionsare present, the extraction method may be used.

If noticeable amounts of color or organic matter are present, itmay be necessary to evaporate the sample, gently ash the resi-due, and redissolve in acid. The ashing may be carried out insilica, porcelain, or platinum crucibles that have been boiled forseveral hours in 6N HCI. The presence of excessive amounts oforganic matter may necessitate digestion before use of the ex-traction procedure.

C. Minimum detectable concentration: Dissolved or total con-centrations of iron as low as 10 fLg/Lcan be deterrnined with aspectrophotometer using cells with a 5 cm or longer light path.Carry a blank through the entire procedure to allow for correc-tion.

a. Colorimetric equipment: One of the following is required:1) Spectrophotometer, for use at 510 nm, providing a light

path of I cm or longer.2) Fi/ter photometer, providing a light path of 1 cm or longer

and equipped with a green filter having maximum transmittancenear 510 nm.

3) Nessler tubes, matched, 100-rnL, tall formob. Acid-washed glassware: Wash ali glassware with conc

hydrochloric acid (HCl) and rinse with reagent water before useto remove deposits of iron oxide.

Use reagents low in iron. Use reagent water (see lOS0 and3111B.3c) in preparing standards and reagent solutions and inprocedure. Store reagents in glass-stoppered bottles. The HCIand ammonium acetate solutions are stable indefinitely if tightlystoppered. The hydroxylarnine, phenanthroline, and stock ironsolutions are stable for several months. The standard iron solu-tions are not stable; prepare daily as needed by diluting the stocksolution. Visual standards in nessler tubes are stable for severalmonths if sealed and protected from light.

a. Hydrochloric acid, HCI, conc, containing less than 0.5 ppmiron.

b. Hydroxylamine solution: Dissolve 10 g NHzOH'HCI in 100rnL water.

C. Ammonium acetate buffer solution: Dissolve 250 gNH4CzH30Z in 150 rnL water. Add 700 mL conc (glacial) aceticacid. Because even a good grade of NH4CzH30Z contains asignificant amount of iron, prepare new reference standards witheach buffer preparation.

d. Sodium acetate solution: Dissolve 200 g NaCzH30Z'3HzOin SOOrnL water.

e. Phenanthroline solution: Dissolve 100 mg I,lO-phenanth-roline monohydrate, ClzH8Nz'HzO, in 100 rnL water by stirringand heating to SO°C. Do not boil. Discard the solution if itdarkens. Heating is unnecessary if 2 drops conc Hei are addedto the water. (NOTE:One milliliter of this reagent is sufficient forno more than 100 fLg Fe.)f Potassium permanganate, 0.02M: Dissolve 0.316 g KMn04

in reagent water and dilute to 100 rnL.g. Stock iron solution: Use metal (1) ar salt (2) for preparing

the stock solution.1) Use electrolytic iron wire, or "iron wire for standardizing,"

to prepare the solution. If necessary, clean wire with fine sand-paper to remove any oxide coating and to produce a brightsurface. Weigh 200.0 mg wire and place in a 1000-rnL volumet-ric flask. Dissolve in 20 rnL 6N sulfuric acid (HZS04) and diluteto mark with water; 1.00 rnL = 200 fLg Fe.

2) If ferrous ammonium sulfate is preferred, slowly add 20mL conc HZS04 to 50 rnL water and dissolve 1.404 gFe(NH4h(S04)z·6HzO. Slowly add potassium permanganate (~1j) until a faint pink color persists. Add the last few millilitersof the solution dropwise. Approximately 50 rnL of the potassiumpermanganate will be required. Dilute to 1000 rnL with waterand mix; 1.00 rnL = 200 fLg Fe.

h. Standard iron solutions: Prepare daily for use.I) Pipet 50.00 rnL stock solution into a 1000-rnL volumetric

flask and dilute to mark with water; 1.00 rnL = 10.0 J.Lg Fe.2) Pipet 5.00 rnL stock solution into a 1000-rnL volumetric

flask and dilute to mark with water; 1.00 rnL = 1.00 J.Lg Fe.i. Diisopropyl or isopropyl ether. CAUTION: Ethers may form

explosive peroxides; test before using.

a. Total iron: Mix sample thoroughly and measure 50.0 rnLinto a 125-rnL erlenmeyer flask. lf this sample volume containsmore than 200 J.Lg iron use a smaller accurately measured portionand dilute to 50.0 rnL. Add 2 rnL conc HCI and I rnLNH20H·HCI solution. Add a few glass beads and heat to boiling.To insure dissolution of all the iron, continue boiling untilvolume is reduced to 15 to 20 mL. (lf the sample is ashed, takeup residue in 2 rnL conc HCI and 5 rnL water.) Cool to roomtemperature and transfer to a 50- or 100-rnL volumetric flask ornessler tube. Add 10 rnL NH4C2H302 buffer solution and 4 rnLphenanthroline solution, and dilute to mark with water. Mixthoroughly and allow a minimum of 10 min for maximum colordevelopment.

b. Dissolved iron: Immediately after collection filter samplethrough a 0.45-J.Lm membrane filter into a vacuum flask contain-ing I rnL conc HCl/100 rnL sample. Analyze filtrate for totaldissolved iron (lJ[ 4a) and/or dissolved ferrous iron (lJ[ 4c). (Thisprocedure also can be used in the laboratory if it is understoodthat normal sample exposure to air during shipment may result inprecipitation of iron.)

Calculate suspended iron by subtracting dissolved from totaliron.

c. Ferrous iron: Determine ferrous iron at sampling site be-cause of the possibility of change in the ferrous-ferric ratio withtime in acid solutions. To determine ferrous iron only, acidify aseparate sample with 2 rnL conc HCl/100 rnL sample at time ofcollection. Fill bottle directly from sampling source and stopper.Immediately withdraw a 50-rnL portion of acidified sample andadd 20 rnL phenanthroline solution and 10 rnL NH4C2H302

solution with vigorous stirring. Dilute to 100 rnL and measurecolor intensity within 5 to 10 mino Do not expose to sunlight.(Color development is rapid in the presence of excess phenan-throline. The phenanthroline volume given is suitable for lessthan 50 J.Lg total iron; if larger amounts are present, use acorrespondingly larger volume of phenanthroline or a moreconcentrated reagent.)

Calculate ferric iron by subtracting ferrous from total iron.d. Color measurement: Prepare a series of standards by accu-

rately pipetting calculated volumes of standard iron solutions[use solution described in lJ[ 3h2) to measure 1- to 10-J.Lg por-tions] into 125-rnL erlenmeyer flasks and diluting to 50 rnL byadding measured volumes of water. Add 2 rnL conc HCI and IrnL NH20H·HCI solution. Carry out the steps in lJ[ 4a beginningwith transfer to a 100-rnL volumetric flask or nessler tube.

For visual comparison, prepare a set of at least 10 standards,ranging from I to 100 J.Lg Fe in the final 100-rnL volume.Compare colors in 100-rnL tall-form nessler tubes.

For photometric measurement, use Table 3500-Fe:I as a roughguide for selecting proper light path at 510 nm. Read standards

TABLE 3500-FE:I. SELECTIONOF LIGHT PATH LENGTH FOR VARIOUS 1RONCONCENTRATIONS

50-mL 100-rnL LightFinal Final Path

Volume Volume em

50-200 100--400 125-100 50-200 210--40 20-80 55-20 10--40 10

against water set at zero absorbance and plot a calibration curve,including a blank (see lJ[ 3c and General Introduction).

lf samples are colored or turbid, carry a second set of samplesthrough all steps of the procedure without adding phenanthro-line. Instead of water, use the prepared blanks to set photometerto zero absorbance and read each sample developed withphenanthroline against the corresponding blank without phenan-throline. Translate observed photometer readings into iron valuesby means of the calibration curve. This procedure does notcompensate for interfering ions.

e. Samples containing organic interferences: Digest samplescontaining substantial amounts of organic substances accordingto the directions given in Sections 3030G or H.

I) lf a digested sample has been prepared according to thedirections given in Section 3030G or H, pipet 10.0 rnL or othersuitable portion containing 20 to 500 J.Lg Fe into a 125-rnLseparatory funnel. If the volume taken is less than 10 rnL, addwater to make up to 10 rnL. To the separatory funnel add 15 rnLconc HCI for a lO-mL aqueous volume; or, if the portion takenwas greater than 10.0 rnL, add 1.5 rnL conc HCl/rnL sample.Mix, cool, and proceed with 4e3) below.

2) To prepare a sample solely for determining iron, measurea suitable volume containing 20 to 500 J.Lg Fe and carry itthrough ·the digestion procedure described in either Section3030G or H. However, use only 5 rnL H2S04 or HCI04 and omitH202. When digestion is complete, cool, dilute with 10 rnLwater, heat almost to boiling to dissolve slowly soluble salts,and, if the sample is still c1oudy, filter through a glass-fiber,sintered-glass, or porcelain filter, washing with 2 to 3 rnL water.Quantitatively transfer filtrate or c1ear solution to a 25-rnL volu-metric flask and make up to 25 rnL with water. Empty flask intoa 125-rnL separatory funnel, rinse flask with 5 rnL conc HCI andadd to the funnel. Add 25 mL conc HCI measured with the sameflask. Mix and cool to room temperature.

3) Extract the iron from the HCI solution in the separatoryfunnel by shabng for 30 s with 25 mL isopropyl ether (CAU-TION). Draw off lower acid layer into a second separatory funnel.Extract acid solution again with 25 rnL isopropyl ether, drainacid layer into a suitable c1ean vessel, and add ether layer to theether in the first funnel. Pour acid layer back into second sepa-ratory funnel and re-extract with 25 rnL isopropyl ether. With-draw and discard acid layer and add ether layer to first funnel.Persistence of a yellow color in the HCI solution after threeextractions does not signify incomplete separation of iron be-cause copper, which is not extracted, gives a similar yellowcolar.

Shake combined ether extracts with 25 mL water to retum ironto aqueous phase and transfer lower aqueous layer to a 100-mLvolumetric f1ask. Repeat extraction with a second 25-mL portionof water, adding this to the first aqueous extract. Discard etherlayer.

4) Add 1 mL NH20H'HCI solution, 10 mL phenanthrolinesolution, and 10 mL NaC2H302 solution. Dilute to 100 mL withwater, mix thoroughly, and let stand for a minimum of 10 minoMeasure absorbance at 510 nm using a 5-cm absorption cell foramounts of iron less than 100 J.Lg or 1-cm cell for quantities from100 to 500 J.Lg. As reference, use either water or a sample blankprepared by carrying the specified quantities of acids through theentire analytica1 procedure. If water is used as reference, correctsamp1e absorbance by subtracting absorbance of a samp1e blank.

Determine micrograms of iron in the sample from the absor-bance (corrected, if necessary) by reference to the calibrationcurve prepared by using a suitab1e range of iron standardscontaining the same amounts of phenanthroline, hydroxylamine,and sodium acetate as the sample.

p,g Fe (in 100 mL final volume)mg FelL = -----------

mLsample

When the sample has been treated according to 4e1):

p,g Fe (in 100 mL final volume) 100mg FelL = -----------X ----

mL sample mL portion

Report details of sample collection, storage, and pretreatmentif they are pertinent to interpretation of results.

Precision and bias depend on the method of sample collectionand storage, the method of color measurement, the iron concen-tration, and the presence of interfering color, turbidity, and

Lead (Pb) is the fifth element in Group IV A in the periodictable; it has an atomic number of 82, an atomic weight of 207.19,and valences of 2 and 4. The average abundance of Pb in the

*Approved by Standard Methods Cornrninee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

foreign ions. In general, optimum reliability of visual compari-son in ness1er tubes is not better than 5% and often on1y 10%,whereas, under optimum conditions, photometric measurementmay be reliable to 3% or 3 J.Lg, whichever is greater. Thesensitivity limit for visual observation in nessler tubes is approx-imately 1 J.Lg Fe. Sample variability and instability may affectprecision and bias of this determination more than will the errorsof ana1ysis. Serious divergences have been found in reports ofdifferent 1aboratories because of variations in methods of col-lecting and treating samples.

A synthetic sample containing 300 J.Lg FelL, 500 J.Lg AIIL, 50J.Lg CctIL, 110 J.Lg CrlL, 470 J.Lg CulL, 70 J.Lg PblL, 120 J.Lg MnIL,150 J.Lg AglL, and 650 J.Lg ZnIL in distilled water was ana1yzedin 44 1aboratories by the phenanthroline method, with a re1ativestandard deviation of 25.5% and a reI ative error of 13.3%.

CHRONHElM,G. & W. WINK.1942. Determination of divalent iron (byo-nitrosophenol). Ind. Eng. Chem., Anal. Ed. 14:447.

MEHLIG,R.P. & R.H. HULETI.1942. Spectrophotometric determinationof iron with o-phenanthroline and with nitro-o-phenanthroline. Ind.Eng. Chem., AnaI. Ed. 14:869.

CALDWELL,D.H. & R.B. ADAMs.1946. Colorimetric determination ofiron in water with o-phenanthroline. J. Amer. Water Works Assoe.38:727.

WELCHER,FJ. 1947. Organic Analytical Reagents. D. Van Nostrand CO.,Princeton, NJ., VoI. 3, pp. 85-93.

KOLTHOFF,I.M., T.S. LEE & D.L. LEUSSING.1948. Equilibrium andkinetic studies on the formation and dissociation of ferroin andferrin. Anal. Chem. 20:985.

RYAN,I.A. & G.H. BOTHAM.1949. Iron in aluminum alloys: Colorimet-ric determination using 1,lO-phenanthroline. Anal. Chem. 21: 1521.

RErTZ,L.K., A.S. O'BRIEN& T.L. DAvrs. 1950. Evaluation of three ironmethods using a factorial experiment. Ana!. Chem. 22:1470.

SANDELL,E.B. 1959. Chapter 22 in Colorimetric Determination ofTraces of Metais, 3rd ed. Interscience Publishers, New York, N.Y.

SKOUGSTAD,M.W., M.I. FrSHMAN,L.C. FRIEDMAN,D.E. ERDMANN& S.S.DUNCAN.1979. Methods for determination of inorganic substancesin water and fluvial sediment. Chapter AI in Book 5, Techniques ofWater Resources Investigations of the United States GeologicalSurvey. U.S. Geological Surv., Washington, D.C.

earth's crust is 13 ppm; in soi1s it ranges from 2.6 to 25 ppm; instreams it is 3 J.LglL, and in groundwaters it is generally <0.1mglL. Lead is obtained chiefly from ga1ena (PbS). It is used inbatteries, ammunition, solder, piping, pigments, insecticides, andalloys. Lead a1so was used in gasoline for many years as ananti-knock agent in the form of tetraethy1 lead.

The common aqueous species are Pb2+ and hydroxide andcarbonate complexes. Lead in a water supply may come fromindustrial, mine, and smelter discharges or from the dissolu-

lion of plumbing and plumbing fixtures. Tap waters that are inher-endy noncorrosive or not suitably treated may contain lead resultingfrom an attack on lead service pipes, lead interior plumbing, brassfixtures and fittings, or solder pipe joints.

Lead is nonessential for plants and animaIs. It is toxic byingestion and is a cumulative poison. The Food and Dmg Ad-ministration regulates lead content in food and in house paints.Under the lead-copper mIe, the U.S. EPA drinking water 90thpercentile action leve! is 15 J.tg/L.

2. Selection of MethodThe atomic absorption spectrometric method (3UIB) has a

relatively high detection limit in the f1ame mode and requires

a. Principle: An acidified sample contammg microgramquantities of lead is mixed with ammoniacal citrate-cyanidereducing solution and extracted with dithizone in chloroform(CHCI3) to form a cherry-red lead dithizonate. The color of themixed color solution is measured photometricalIy.1.2 Samplevolume taken for analysis may be 2 L when digestion is used.

b. lnterference: In a weakly ammoniacal cyanide solution (pH8.5 to 9.5) dithizone forms colored complexes with bismuth,stannous tin, and monovalent thalIium. ln strongly ammoniacalcitrate-cyanide solution (pH 10 to 11.5) the dithizonates of theseions are unstable and are extracted only partially? This methoduses a high pH, mixed color, single dithizone extraction. lnter-ference from stannous tin and monovalent thallium is reducedfurther when these ions are oxidized during preliminary diges-tion. A modification of the method allows deteetion and elimi-nation of bismuth interference. Excessive quantities of bismuth,thalIium, and tin may be removed.4

Dithizone in CHCl3 absorbs at 510 nm; control its interferenceby using nearly equal coneentrations of excess dithizone insamples, standards, and blank.

The method is without interferenee for the determination of0.0 to 30.0 J.tgPb in the presence of 20 J.tgTI+, 100 J.tgSn2+,200 J.tg ln3+, and 1000 J.tg each of Ba2+, Cd2+, C02+, Cu2+,Mg2+, Mn2+, Hg2+, Sr2+, 2n2+, AI3+, Sb3+, As3+, Cr3+, Fe3+,V3+, PO/-, and SO/-. Gram quantities of alkali metaIs do notinterfere. A modifieation is provided to avoid interference fromexeessive quantities of bismuth or tino

C. Preliminary sample treatment: At time of colIection acidifywith cone HN03 to pH <2 but avoid exeess HN03. Add 5 rnLO.IN iodine solution to avoid losses of volatile organo-Ieadeompounds during handling and digesting of samples. Prepare ablank of lead-free water and carry through the procedure.

d. Digestion of samples: Unless digestion is shown to beunnecessary, digest all samples for dissolved or total lead asdescribed in 3030H or K.

an extraction procedure (3111 C) for the low eoncentrationscommon in potable water. The eleetrothermal atomie absorp-tion (AA) method (3113B) is more sensitive for low eoncen-trations and does not require extraction. The inductively cou-pled plasma/mass spectrometric method (3125) is even moresensitive than the electrothermal AA method. The inductivelycoupled plasma method (3120) has a sensitivity similar to thatof the f1ame atomie absorption method. Anodie strippingvoltammetry (3130B) can achieve superior deteetion levels,but is suseeptible to interferences from copper, silver, gold,and organie eompounds. The dithizone method (B) is sensi-tive and speeifie as a eolorimetrie proeedure.

e. Minimum detectable concentration: 1.0 J.tg PbllO rnL di-thizone solution.

a. Spectrophotometer for use at 510 nm, providing a lightpath of 1 em or longer.

b. pH meter.C. Separatory funnels: 250-rnL Squibb type. Clean alI glass-

ware, including sample botdes, with 1 + 1 HN03. Rinse thor-oughly with reagent water.

d. Automatic dispensing burets: Use for all reagents to min-imize indeterminate contamination errors.

Prepare alI reagents in lead-free water.a. Stock lead solution: Dissolve 0.1599 g lead nitrate,

Pb(N03h (minimum purity 99.5%), in approximately 200 rnLwater. Add 10 rnL cone HN03 and dilute to 1000 rnL with water.Altematively, dissolve 0.1000 g pure Pb metal in 20 rnL I + 1HN03 and dilute to 1000 mL with water; 1.00 rnL = 100 J.tgPb.

b. Working lead solution: Dilute 2.0 rnL stoek solution to 100rnL with water; I rnL = 2.00 J.tgPb.

C. Nitric acid, HN03, 1 + 4: Dilute 200 rnL cone HN03 to 1L with water.

d. Ammonium hydroxide, NH40H, 1 + 9: Dilute 10 rnL coneNH40H to 100 rnL with water.

e. Citrate-cyanide reducing solution: Dissolve 400 g dibasicammonium citrate, (NH4)2HC6HS07,20 g anhydrous sodium sul-fite, Na2S03, 10 g hydroxylarnine hydrocWoride, NH20H . HCI,and 40 g potassium cyanide, KCN (CAlITION:Poison) in water anddilute to 1 L. Mix this solution with 2 L cone NH40H. Do not pipetby mouth. Prepare solution in a fume hood.f Stock dithizone solution: The dithizone coneentration in the

stoek dithizone solutions is based on having a 100% pure dithi-zone reagent. Some commercial grades of dithizone are eontam-

inated with the oxidation product diphenylthiocarbodiazone orwith metais. Purify dithizone as directed below. For dithizonesolutions not stronger than 0.001 % (w/v), calculate the exactconcentration by dividing the absorbance of the solution in a1.00-cm cell at 606 nm by 40.6 X 103, the molar absorptivity.

In a fume hood, dissolve 100 mg dithizone in 50 rnL CHCl3 ina 150-mL beaker and filter through a 7-cm-diam paper. * Receivefiltrate in a 500-rnL separatory funnel or in a 125-rnL erlenmeyerflask under slight vacuum; use a filtering device designed tohandle the CHCl3 vapor. Wash beaker with two 5-rnL portionsCHCI3, and filter. Wash the paper with three 5-rnL portionsCHCI3, adding final portion dropwise to edge of paper. If filtrateis in flask, transfer with CHCl3 to a 500-rnL separatory funnel.

Add 100 rnL 1 + 99 NH40H to separatory funnel and shakemoderately for 1 min; excessive agitation produces slowlybreaking emulsions. Let layers separate, swirling funnel gentlyto submerge CHCl3 droplets held on surface of aqueous layer.Transfer CHCl3 layer to 250-rnL separatory funnel, retaining theorange-red aqueous layer in the 500-rnL funneI. Repeat extrac-tion, receiving CHCI3 layer in another 250-rnL separatory funneland transferring aqueous layer, using 1 + 99 NH40H, to the500-rnL funnel holding the first extract. Repeat extraction, trans-ferring the aqueous layer to 500-rnL funneI. Discard CHCI3

layer.To combined extracts in the 500-rnL separatory funnel add 1

+ I HCI in 2-rnL portions, mixing after each addition, untildithizone precipitates and solution is no longer orange-red. Ex-tract precipitated dithizone with three 25-rnL portions CHCI3.

Dilute combined extracts to 1000 rnL with CHCI3; 1.00 mL =100 J.Lg dithizone.

g. Dithizone working solution: Dilute 100 rnL stock dithizonesolution to 250 rnL with CHCI3; 1 rnL = 40 J.Lg dithizone.

h. Special dithizone solution: Dissolve 250 mg dithizone in250 rnL CHCI3. This solution may be prepared without purifi-cation because ali extracts using it are discarded.

i. Sodium sulfite solution: Dissolve 5 g anhydrous Na2S03 in100 rnL water.

j. Iodine solution: Dissolve 40 g KI in 25 rnL water, add12.7 g resublimed iodine, and dilute to 1000 rnL.

a. With sample digestion: CAUTION: Perform the foUowingprocedure (excluding use of spectrophotometer) in a fume hood.To a digested sample containing not more than 1 rnL cone acidadd 20 rnL 1 + 4 HN03 and filter through lead-free filter papertand filter funnel directly into a 250-rnL separatory funneI. Rinsedigestion beaker with 50 rnL water and add to filter. Add 50 rnLammoniacal citrate-cyanide solution, mix, and cool to roomtemperature. Add 10 rnL dithizone working solution, shakestoppered funnel vigorously for 30 s, and let layers separate.Insert lead-free cotton in stem of separatory funnel and draw offlower layer. Discard 1 to 2 rnL CHCl3 layer, then fill absorption

* Whatman No. 42 or equivalent.t Whatman No. 541 or equivalent.

celI. Measure absorbance of extract at 510 nm, using dithizoneworking solution, CJ[ 3g, to zero spectrophotometer.

b. Without sample digestion: To 100 rnL acidified sample (pH2) in a 250-rnL separatory funnel add 20 rnL 1 + 4 HN03 and50 mL citrate-cyanide reducing solution; mix. Add 10 rnL dithi-zone working solution and proceed as in CJ[ 4a.

c. Calibration curve: Plot concentration of at least five stan-dards and a blank against absorbance. Determine concentrationof lead in extract from curve. Ali concentrations are J.Lg PbllOrnL final extract.

d. Removal of excess interferences: The dithizonates of bis-muth, tin, and thallium differ from lead dithizonate in maximumabsorbance. Detect their presence by measuring sample absor-bance at 510 nm and at 465 nm. Calculate corrected absorbanceof sample at each wavelength by subtracting absorbance of blankat same wavelength. Calculate ratio of corrected absorbance at510 nm to corrected absorbance at 465 nm. The ratio of correctedabsorbances for lead dithizonate is 2.08 and for bismuth dithi-zonate is 1.07. If the ratio for the sample indicates interference,i.e., is markedly less than 2.08, proceed as follows with a new100-rnL sample: If the sample has not been digested, add 5 rnLNa2S03 solution to reduce iodine preservative. Adjust sample topH 2.5 using a pH meter and 1 + 4 HN03 or 1 + 9 NH40H asrequired. Transfer sample to 250-rnL separatory funnel, extractwith a minimum of three 10-rnL portions special dithizonesolution, or until the CHCl3 layer is distinctly green. Extract with20-rnL portions CHCl3 to remove dithizone (absence of green).Add 20 mL I + 4 HN03, 50 rnL citrate-cyanide reducingsolution, and 10 rnL dithizone working solution. Extract as in CJ[

4a and measure absorbance.

MgPb (in 10roL,fram calibrationcurve)mg PbIL = -------------

roLsample

Single-operator precision in recovering 0.0104 mg PblL fromMississippi River water was 6.8% relative standard deviationand -1.4% relative error. At the level of 0.026 mg PblL,recovery was made with 4.8% relative standard deviation and15% relative error.

1. SNYDER,L.J. 1947.Improveddithizonemethod for determinationoflead-mixed color method at high pH. Ana!. Chem. 19:684.

2. SANDELL,E.B. 1959.Co1orimetricDeterminationofTraces ofMeta1s,3rd ed. Interscience,New York, N.Y.

3. WICHMANN,H.I. 1939.Isolationand determinarionof trace metals-the dithizonesystem. Ind. Eng. Chem., Ana!.Ed. 11:66.

4. AMERlCANSOClETYFORTESTlNGANOMATERlALS.1977.AnnualBookof ASTM Standards. Part 26, Method D3112-77, American Soe.Testing & Materials,Philadelphia,Pa.

BIBLIOTECA CEFET· RS

Lithium (Li) is the second element in Group IA of theperiodic table; it has an atomic number of 3, an atomic weightof 6.94, and a valence of 1. The average abundance of Li inthe earth's crust is 18 ppm; in soils it is 14 to 32 ppm; instreams it is 3 fLg/L, and in groundwaters it is <0.1 mg/L. Themore important mineraIs containing lithium are lepidolite,spodumene, petalite, and amblygonite. Lithium compoundsare used in pharmaceuticals, soaps, batteries, welding flux,ceramics, reducing agents (e.g., lithium aluminum hydride),and cosmetics.

Many lithium salts are only slightly soluble, and the metal'sconcentration in water is controlled by incorporation in clay

* Approved by Standard Methods Committee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

mineraIs of soils. Lithium is considered nonessential for plantsand animaIs, but it is essential for some microorganisms. Somelithium salts are toxic by ingestion. The United Nations Food andAgriculture Organization recommended maximum leveI for lith-ium in irrigation waters is 2.5 mglL.

The atomic absorption spectrometric method (3111B) and theinductively coupled plasma method (3120) are preferred. Theflame emission photometric method (B) also is available forlaboratories not equipped to use preferred methods. The induc-tively coupled plasma/mass spectrometric method (3125) may beapplied successfully in most cases (with lower detection levels),even though lithium is not specifically listed as an analyte in themethod.

a. PrincipIe: Lithium can be determined in trace amounts byflame photometric methods at a wavelength of 670.8 nm.

b. Inteiference: A molecular band of strontium hydroxidewith an absorption maximum at 671.0 nm interferes in the flamephotometric determination of lithium. Ionization of lithium canbe significant in both the air-acetylene and nitrous oxide-acety-Iene flames and can be suppressed by adding potassium. SeeSection 3500-Na.B.lb for additional information on minimizinginterferences in fiame photometry.

c. Minimum detectable concentration: The minimum lithiumconcentration detectable is approximately 0.1 fLglL for reagentwater analyzed on an atomic absorption spectrophotometer in theemission mode with an air-acetylene fiame, or 0.03 fLglL with anitrous oxide-acetylene flame.

d. Sampling and storage: Preferably collect sample in a poly-ethylene bottle, although borosilicate glass containers aIs o maybe used. At time of collection adjust sample to pH <2 with nitricacid (HN03).

Flame photometer: A flame photometer or an atomic absorp-tion spectrometer operating in the emission mo de using a leanair-acetylene flame is recommended.

Use reagent water (see 311IB.3c) in reagent preparation andanalysis.

a. Potassium ionization suppressant: Dissolve 95.35 g KCIdried at 110°C and dilute to 1000 mL with water; 1.00 mL = 50mg K.

b. Stock lithium solution: Dissolve 152.7 mg anhydrous lith-ium chloride, LiCI, in water and dilute to 250 mL; 1.00 mL =100 fLg Li. Dry salt overnight in an oven at 105°C. Cool in adesiccator and weigh immediately after removal from desiccator.Alternatively, purchase prepared stock from a reputable supplier.

c. Standard lithium solution: Dilute 10.00 mL stock LiCIsolution to 500 mL with water; 1.00 mL = 2.0 fLg Li.

a. Pretreatment of polluted water and wastewater samples:Choose digestion method appropriate to matrix (see Section3030).

b. Suppressing ionization: If necessary, filter sample throughmedium-porosity paper, add 1.0 mL potassium ionization sup-pressant to 50 mL volumetric flask, and dilute with sample forflame photometric determination. Sample solution will be in a0.1% K matrix.

c. Treatment of standard solutions: Prepare dilutions of the Listandard solution to bracket sample concentration or to establishat least three points on a calibration curve of emission intensity

against Li concentration. Prepare standards by adding appropri-ate volumes of standard lithium solution to 25 rnL water + 1.0rnL potassium ionization suppressant reagent in a 50-rnL volu-metric flask. Dilute to 50.0 rnL and mix. Both samples andstandards will be in a 0.1% K matrix to suppress ionization oflithium.

d. Flame photometric measurement: Determine lithium con-centration by direct intensity measurements at a wavelength of670.8 nm. The bracketing method (Section 3500-Na.BAd) canbe used with some photometric instruments, while the construc-tion of a calibration curve is necessary with others. Run sample,water, and lithium standard as nearly simultaneously as possible.For best results, average several readings on each solution.

Follow the manufacturer' s instructions for instrument operation.

mL sample + mL watermL sample

Magnesium (Mg) is the second element in Group lIA of theperiodic table; it has an atomic number of 12, an atomicweight of 24.30, and a valence of 2. The average abundanceof Mg in the earth's crust is 2.1 %; in soils it is 0.03 to 0.84%;in streams it is 4 mg/L, and in groundwaters it is >5 mg/L.Magnesium occurs commonly in the mineraIs magnesite anddolomite. Magnesium is used in alloys, pyrotechnics, flashphotography, drying agents, refractories, fertilizers, pharma-ceuticals, and foods.

The common aqueous species is Mg2+. The carbonateequilibrium reactions for magnesium are more complicatedthan for calcium, and conditions for direct precipitation ofdolomite in natural waters are not common. Important con-tributors to the hardness of a water, magnesium salts breakdown when heated, forming scale in boilers. Chemical soft-

* Approved by Standard Methods Committee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

Process a QC standard through entire analytical protocol as away of determining systematic bias. The control limits for pre-cision of duplicate determinations at concentrations (in water) of4.0 fLglL and 10.0 fLglL were 4.09 ± 0.056 fLglL and 9.96 ±0.094 fLglL, respectively. The single-operator RSD was 1.38%for a lithium solution containing 10 fLglL.

7. Bibliography

FISHMAN,MJ. 1962. Flame photometric determination of lithium inwater. J. Amer. Water Works Assoe. 54:228.

PICKETT,E.E. & S.R. KOIRTYOHANN.1968. The nitrous oxide-acetyleneflame in emission analysis-I. General characteristics. Speetroehem.Aeta. 23B:235.

KOIRTYOHANN,S.R. & E.E. PICKETT.1968. The nitrous oxide-acetyleneflame in emission analysis-II. Lithium and the alkaline earths.Speetrochem. Acta, 23B:673.

URE, A.M. & R.L. MITCHELL.1975. Lithium, sodium, potassium,rubidium, and cesium. In J.A. Dean & T.C. Rains, eds. FlameEmission and Atomic Absorption Spectrometry. Dekker, NewYork, N.Y.

THOMPSON,K.C. & RJ. REYNOLDS.1978. Atomic Absorption Fluores-cence, and Flame Spectroscopy-A Practical Approach, 2nd ed.John Wiley & Sons, New York, N.Y.

WILLARD,H.H., L.L. MERRIT,J.A. DEAN& F.A. SETTLE.,JR. 1981.Instrumental Methods of Analysis, 6th ed. Wadsworth PublishingCo., Belmont, Calif.

ening, reverse osmosis, or ion exchange reduces magnesiumand associated hardness to acceptable levels.

Magnesium is an essential element in chlorophyll and in redblood cells. Some salts of magnesium are toxic by ingestion orinhalation. Concentrations greater than 125 mglL also can havea cathartic and diuretic effect.

The methods presented are applicable to waters and waste-waters. Direct determinations can be made with the atomicabsorption spectrometric method (3l1lB) and inductivelycoupled plasma method (3120). The inductively coupledplasma mass spectrometric method (3125) also may be ap-plied successfully in most cases (with lower detection levels),even though magnesium is not specifically listed as an analytein the method. These methods can be applied to most con-centrations encountered, although sample dilution may berequired. Choice of method is largely a matter of personalpreference and analyst experience. A calculation method (B)also is available.

Magnesium may be estimated as the difference between hard- mgMg/L = [totalhardness(as mg CaC03/L)ness and calcium as CaC03 if interfering metaIs are present innoninterfering concentrations in the calcium titration (Section - calciumhardness(as mgCaC03/L)] X 0.2433500-Ca.B) and suitable inhibitors are used in the hardnesstitration (Section 2340C).

Manganese (Mn) is the first element in Group VIIB in theperiodic table; it has an atomic number of 25, an atomic weightof 54.94, and common valences of 2,4, and 7 (and more rarely,valences of 1, 3, 5, and 6). The average abundance of Mn in theearth's crust is 1060 ppm; in soils it is 61 to 1010 ppm; instreams it is 7 p,g/L, and in groundwaters it is <0.1 mg/L.Manganese is associated with iron mineraIs, and occurs in nod-ules in ocean, fresh waters, and soils. The common ores arepyrolusite (MnOz) and psilomelane. Manganese is used in steelalloys, batteries, and food additives.

The common aqueous species are the reduced Mnz+ and the

oxidized Mn4+. The aqueous chemistry of manganese is similarto that of iron. Since groundwater is often anoxic, any solublemanganese in groundwater is usually in the reduced state(Mnz+). Upon exposure to air or other oxidants, groundwatercontaining manganese usually will precipitate black MnOz. EI-evated manganese levels therefore can cause stains in plumbing/laundry, and cooking utensils. It is considered an essential traceelement for plants and animaIs. The United Nations Food and

* Approved by Standard Methods Commitlee, 1999.Joint Task Group: 20th Edition-See 3500-Al.

a. Principie: Persulfate oxidation of soluble manganous com-pounds to form permanganate is carried out in the presence ofsilver nitrate. The resulting color is stable for at least 24 h ifexcess persulfate is present and organic matter is absent.

b. Interference: As much as 0.1 g chloride (Cl-) in a 50-mLsample can be prevented from interfering by adding 1 g mercuricsulfate (HgS04) to form slightly dissociated complexes. Bro-

Agriculture Organization recommended maximum leveI formanganese in irrigation waters is 0.2 mg/L. The U.S. EPAsecondary drinking water standard MCL is 50 p,g/L.

The atomic absorption spectrometric methods (3111B and C),the electrothermal atomic absorption method (3113B), and theinductively coupled plasma methods (3120 and 3125) permitdirect determination with acceptable sensitivity and are themethods of choice. Of the various colorimetric methods, thepersulfate method (B) is preferred because the use of mercuricion can control interference from a limited chloride ion concen-tration.

Manganese may exist in a soluble form in a neutral waterwhen first collected, but it oxidizes to a higher oxidation stateand precipitates or becomes adsorbed on the container walls.Determine manganese very soon after sample collection. Whendelay is unavoidable, total manganese can be determined if thesample is acidified at the time of collection with HN03 to pH<2. See Section 301OB.

mide and iodide still will interfere and only trace amounts maybe present. The persulfate procedure can be used for potablewater with trace to small amounts of organic matter if the periodof heating is increased after more persulfate has been added.

For wastewaters containing organic matter, use preliminarydigestion with nitric and sulfuric acids (HN03 and HZS04) (seeSection 3030G). If large amounts of CI- also are present, boilingwith HN03 helps remove it. Interfering traces of Cl- are elim-inated by HgS04 in the special reagent.

Colored solutions from other inorganic ions are compensatedfor in the final colorimetric step.

Samples that have been exposed to air may give low resultsdue to precipitation of manganese dioxide (MnOz). Add I drop30% hydrogen peroxide (HzOz) to the sample, after adding thespecial reagent, to redissolve precipitated manganese.

c. Minimum detectable concentration: The molar absorptivityof permanganate ion is about 2300 L g- I cm -I. This corre-sponds to a minimum detectable concentration (98% transmit-tance) of 210 /LgMnJL when a I-cm cell is used or 42 /LgMnJLwhen a 5-cm cell is used.

Colorimetric equipment: One of the following is required:a. Spectrophotometer, for use at 525 nm, providing a light

path of 1 cm or longer.b. Filter photometer, providing a light path of 1 cm or longer

and equipped with a green filter having maximum transmittancenear 525 nm.

c. Nessler tubes, matched, 100-mL, tall formo

a. Special reagent: Dissolve 75 g HgS04 in 400 mL concHN03 and 200 mL distilled water. Add 200 mL 85% phosphoricacid (H3P04), and 35 mg silver nitrate (AgN03). Dilute thecooled solution to I L.

b. Ammonium persulfate, (NH4hSzOg, solid.c. Standard manganese solution: Prepare a O.IN potassium

permanganate (KMn04) solution by dissolving 3.2 g KMn04 indistilled water and rnaking up to 1 L. Age for several weeks insunlight or heat for several hours near the boiling point, thenfilter through a fine fritted-glass filter crucible and standardizeagainst sodium oxalate as follows:

Weigh severall00- to 200-mg samples ofNazCz04 to 0.1 mgand transfer to 400-mL beakers. To each beaker, add 100 mLdistilled water and stir to dissolve. Add 10 mL I + 1 HZS04 andheat rapidly to 90 to 95°C. Titrate rapidly with the KMn04solution to be standardized, while stirring, to a slight pinkend-point color that persists for at least I mino Do not lettemperature fall below 85°C. lf necessary, warm beaker contentsduring titration; 100 mg Na2CZ04 will consume about 15 mLpermanganate solution. Run a blank on distilled water andHZS04•

g Na2C104Normalityof KMn04 = (A _ B) x 0.06701

where:A = mL titrant for sample andB = mL titrant for blank.

Average results of several titrations. Calculate volume of thissolution necessary to prepare I L of solution so that 1.00 mL =50.0 /Lg Mn, as follows:

4.55mL KMn04 = ------

normalityKMn04

To this volume add 2 to 3 mL conc H2S04 and NaHS03

solution dropwise, with stirring, until the permanganate colordisappears. Boil to remove excess SOz, cool, and dilute to 1000mL with distilled water. Dilute this solution further to measuresmall amounts of manganese.

d. Standard manganese solution (alternate): Dissolve 1.000 gmanganese metal (99.8% min.) in 10 mL redistilled HN03.

Dilute to 1000 mL with 1% (v/v) HCI; 1 mL = 1.000 mg Mn.Dilute 10 mL to 200 mL with distilled water; 1 mL = 0.05 mgMn. Prepare dilute solution daily.

e. Hydrogen peroxide, HZ02, 30%.f Nitric acid, HN03, conc.g. Sulfuric acid, HZS04, conc.h. Sodium nitrite solution: Dissolve 5.0 g NaN0z in 95 mL

distilled water.i. Sodium oxalate, NaZCZ04, primary standard.j. Sodium bisulfite: Dissolve 10 g NaHS03 in 100 mL distilled

water.

a. Treatment of sample: If a digested sample has been preparedaccording to directions for reducing organic matter and/or excessivecWoridesin Section 30300, pipet a portion containing0.05 to 2.0 mgMn into a 250-mLconical flask.Add distilledwater, if necessary,to 90mL and proceed as in lJ[ b.

b. To a suitable sample portion add 5 mL special reagentand 1 drop HzOz. Concentrate to 90 mL by boiling or dilute to90 mL. Add 1 g (NH4hSzOg, bring to a boil, and boil for 1mino Do not heat on a water bath. Remove from heat source,let stand 1 min, then cool under the tapo (Boiling toa longresults in decomposition of excess persulfate and subsequentloss of permanganate color; cooling toa slowly has the sameeffect.) Dilute to 100 mL with distilled water free fromreducing substances and mix. Prepare standards containing O,5.00, ... 1500 /LgMn by treating various amounts of standardMn solution in the same way.

C. Nessler tube comparison: Use standards prepared as in lJ[ 4band containing 5 to 100 /LgMn/l00 mL final volume. Comparesamples and standards visually.

d. Photometric determination: Use a series of standards fromO to 1500 /Lg Mn/100 mL final volume. Make photometricmeasurements against a distilled water blank. The followingtable shows Iight path length appropriate for various amounts ofmanganese in 100 mL final volume:

Mn Rangep,g

5-20020-40050-1000100-1500

Light Pathem

Prepare a calibration curve of manganese concentration vs.absorbance from the standards and determine Mn in the samplesfrom the curve. lf turbidity or interfering color is present, makecorrections as in lJ[ 4e.

e. Correction for turbidity or inteifering color: Avoid filtra-tion because of possible retention of some permanganate on thefilter paper. If visual comparison is used, the effect of turbidity

only can be estimated and no correction can be made for inter-fering colored ions. When photometric measurements are made,use the following "bleaching" method, which also corrects forinterfering color: As soon as the photometer reading has beenmade, add 0.05 mL H202 solution directly to the sample in theoptical cell. Mix and, as soon as the perrnanganate color hasfaded completely and no bubbles remain, read again. Deductabsorbance of bleached solution from initial absorbance to ob-tain absorbance due to Mn.

f.Lg Mn/lOOmL 100mg Mn/L = ----- x ----

mLsample mL portion

b. When a portion of the digested sample (100 mL finalvolume) is taken for analysis:

f.Lg Mn (in 100mL final volume)mgMn/L = -----------

mL sample

Mercury (Hg) is the third element in Group I1B in the periodictable; it has an atomic number of 80, an atomic weight of 200.59,and valences of 1 and 2. The average abundance of Hg in theearth's crust is 0.09 ppm; in soils it is 30 to 160 ppb; in streamsit is 0.07 JLglL, and in groundwaters it is 0.5 to 1 JLgIL. Mercuryoccurs free in nature, but the chief source is cinnabar (HgS).Mercury is used in amalgams, mirror coatings, vapor lamps,paints, measuring devices (therrnometers, barometers, manome-ters), pharmaceuticals, pesticides, and fungicides. It is often usedin paper mills as a mold retardant for paper.

The common aqueous species are Hg2+, Hg(OHho, HgO,andstable complexes with organic ligands. Inorganic mercury can bemethylated in sediments when sulfides are present to form di-methyl mercury, (CH3hHg, which is very toxic and can concen-trate in the aquatic food chain. Mercury poisoning occurred inJapan in the 1950s as the result of consumption of shellfish that

A synthetic sample containing 120 JLg MnIL, 500 JLg AIIL, 50JLg Cd/L, 110 JLg CrlL, 470 JLg CulL, 300 JLg FelL, 70 JLg PblL,150 JLg AglL, and 650 JLg ZnIL in distilled water was analyzedin 33 laboratories by the persulfate method, with a relativestandard deviation of 26.3% and a relative error of 0%.

A second synthetic sample, similar in all respects except for 50JLg MnIL and 1000 JLg CulL, was analyzed in 17 laboratories bythe persulfate method, with a relative standard deviation of50.3% and a relative error of 7.2%.

RICHARDS,M.D. 1930. Colorimetric detennination of manganese inbiologica1material.Ana/yst 55:554.

NYDAHL,F. 1949.Determinationof manganeseby thepersu1fatemethod.Ana!. Chem. Acta. 3:144.

M!LLS,S.M. 1950.Elusive manganese.Water Sewage Works 97:92.SANDELL,E.B. 1959. ColorimetricDeterminationof Traces of Metais,

3rd ed. IntersciencePublishers,New York, N.Y., Chapter26.DELFINO,IJ. & G.F. LEE.1969.Co1orimetricdeterminationof manga-

nese in lake waters. Environ. Sei. Techno!. 3:761.

bad accumulated mercury. In times past, mercury was used in thehaberdashery industry to block hats (the cause of the "madbatter" syndrome).

Mercury is considered nonessential for plants and animals.Tbe U.S. EPA primary drinking water standard MCL is 2 JLglL.

The cold-vapor atomic absorption metbod (3112B) is themethod of choice for all samples. The inductively coupledplasma mass spectrometric method (3125) also may be appliedsuccessfully in some cases, even though mercury is not specif-ically listed as an analyte in the method. The dithizone methoddetailed in the 19th edition of Standard Methods can be used fordeterrnining high levels of mercury (>2 JLglL) in potable waters.

Because mercury can be lost readily from samples, preservethem by treating with HN03 to reduce the pH to <2 (see Section1060). Glass starage containers are preferred to plastic, becausethey can extend the holding time to 30 d, rather than only the14 d allowed in plastic containers.

Molybdenum (Mo) is the second element in Group VIB in theperiodic table; it has an atomic number of 42, and atomic weight of95.95, and valences of 2, 3, 4, 5, and 6. Tbe average abundance ofMo in the earth's crust is 1.2 ppm; in soils it is 2.5 ppm; in streamsit is 1 JLgIL, and in groundwaters it is <0.1 mgIL. Molybdenumoccurs naturally as molybdenite (MoS2)and wulfenite (PbMo04). Itis used in alloys, ink pigments, catalysts, and lubricants.

The common aqueous species are HMo04 -, Mo04 2-, andorganic complexes. It is considered an essential trace element forplants and animais. The United Nations Food and AgricultureOrganization recommended maximum levei for irrigation watersis 0.01 mgIL.

Use one of the ftame atomic absorption spectrometric methods(3111D or E), the electrothermal atomic absorption spectromet-ric method (3113B), or one of the inductively coupled plasmamethods (3120 ar 3125).

Nickel (Ni) is the third element in Group VIII in the periodictable; it has an atornic number of 28, an atornic weight of 58.69,and a common valence of 2 and less commonly 1, 3, or 4. Theaverage abundance of Ni in the earth's crust is 1.2 ppm; in soilsit is 2.5 ppm; in streams it is I J.LglL, and in groundwaters it is<0.1 mglL. Nickel is obtained chiefly from pyrrhotite and gar-nierite. Nickel is used in alIoys, magnets, protective coatings,catalysts, and batteries.

The common aqueous species is Ni2+. In reducing conditionsinsoluble sulfides can form, while in aerobic conditions nickel

Osmium (Os) is the seventh element in Group VIII in theperiodic table; it has an atomic number of 76, an atomic weightof 190.2, and valences of 3,4, and 6, and less commonly 1, 2, 5,7, and 8. The average abundance of Os in the earth's crust isprobably <0.005 ppm, and in groundwaters it is <0.01 mglL.

Palladium (Pd) is the sixth element in Group VIII of the periodictable; it has an atomic number of 46, an atomic weight of 106.42,and valences of 2 and 4. Palladium occurs with platinum in nature.It is used in alloys to make electrical relays, catalysts, in the makingf "white gold," and in protective coatings.

Platinum (Pt) is the ninth element in Group VIII of theperiodic table; it has an atomic number of 78, an atornic weightof 195.1, and valences of 2 and 4. The average abundance of Ptin the earth's crust is probably <0.01 ppm, and in groundwatersit is <0.1 mglL. Platinum is usualIy found in its native state, butalso may be found as sperrylite (PtAs2). Platinum is used as acatalyst and in laboratory ware, jewelry, and surgical wire.

complexes with hydroxide, carbonates, and organic ligands canformo It is suspected to be an essential trace element for someplants and animais. The United Nations Food and AgricultureOrganization recommended maximum leveI for irrigation watersis 200 J.LglL. The U.S. EPA prirnary drinking water standardMeL is 0.1 mglL.

The atomic absorption spectrometric methods (3111B and C),the inductively coupled plasma methods (3120 and 3125), andthe e1ectrothermal atornic absorption spectrometric method(3113B) are the methods of choice for alI samples.

Osmium occurs in iridosime and in platinum-bearing river sands.Osrnium is used as a hardener with iridium and as a catalyst withplatinum.

The aqueous chemistry is controlled by complex com-pounds, although the solubility in natural waters is relativelyunknown.

Analyze by f1ame atornic absorption methods (3111D and E).

PalIadium has no known toxic effects. The United NationsFood and Agriculture Organization recommended maximumlevei for irrigation waters is 5 mglL.

Preferably analyze by f1ame atornic absorption method(3111B). The inductively coupled plasma mass spectrometricmethod (3125) also may be applied successful1y in most cases(with lower detection levels), even though palIadium is notspecificalIy Iisted as an analyte in the method.

The aqueous chemistry in natural waters is relatively un-known, although its solubility is probably controlIed by complexcompounds. In powder form, platinum can be f1ammable, and itssoluble salts are toxic by inhalation.

Preferably analyze by f1ame atomic absorption method(3111B). The inductively coupled plasma mass spectrometricmethod (3125) also may be applied successfulIy in most cases(with lower detection levels), even though platinum is not spe-cificalIy Iisted as an analyte in the method.

Potassium (K) is the fourth element in Group IA of theperiodic table; it has an atomic number of 19, an atomic weightof 39.10, and a valence of 1. The average abundance of K in theearth's crust is 1.84%; in soils it has a range of 0.1 to 2.6%; instreams it is 2.3 mglL, and in groundwaters it has a range of 0.5to 10 mglL. Potassium is commonly associated with alumino-silicate mineraIs such as feldspars. 40K is a naturally occurringradioactive isotope with a half-life of 1.3 x 109 years. Potassiumcompounds are used in glass, fertilizers, baking powder, softdrinks, explosives, electroplating, and pigments. Potassium is anessential element in both plant and human nutrition, and occursin groundwaters as a result of mineral dissolution, from decom-posing plant material, and from agricultural runoft.

The common aqueous species is K+. Unlike sodium, it doesnot remain in solution, but is assimilated by plants and isincorporated into a number of clay-mineral structures.

* Approved by Standard Methods Committee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

Methods for the determination of potassium include f1ameatomicabsorption (311lB), inductively coupled plasma (3120), f1amepho-tometry (B), and selective ion electrode (C). The inductively cou-pled plasma/mass spectrometric method (3125) usually may beapplied successfully (with lower detection levels), even thoughpotassium is not specifically listed as an analyte in the method. Thepreferred methods are rapid, sensitive, and accurate; selection de-pends on instrument availability and analyst choice.

Do not store samples in soft-glass bottles because of thepossibility of contamination from leaching of the glass. Useacid-washed polyethylene or borosilicate glass bottles. Adjustsample to pH < 2 with nitric acid. This will dissolve potassiumsalts and reduce adsorption on vessel walls.

a. Principie: Trace amounts of potassium can be determinedin either a direct-reading or internal-standard type of f1amephotometer at a wavelength of 766.5 nm. Because much of theinformation pertaining to sodium applies equally to the potas-sium determination, carefully study the entire discussion dealingwith the f1ame photometric determination of sodium (Section3500-Na.B) before making a potassium determination.

b. lnterference: Interference in the intemal-standard methodmay occur at sodium-to-potassium ratios of 5:1 or greater. Cal-cium may interfere if the calcium-to-potassium ratio is 10:1 ormore. Magnesium begins to interfere when the magnesium-to-potassium ratio exceeds 100:1.

c. Minimum detectable concentration: Potassium levels ofapproximately 0.1 mglL can be deterrnined.

2. Apparatus

See Section 3500-Na.B.2.

3. Reagents

To minimize potassium pickup, store all solutions in plasticbottles. Shake each container thoroughly to dissolve accumu-lated salts from walls before pouring.

a. Reagent water: See Section 1080. Use this water for pre-paring ali reagents and calibration standards, and as dilutionwater.

b. Stock potassium solution: Dissolve 1.907 g KCI dried at110°C and dilute to 1000 mL with water; 1 mL = 1.00 mg K.

c. lntermediate potassium solution: Dilute 10.0 mL stock potas-siurn solution with water to 100 mL; 1.00 mL = 0.100 mg K.

Use this solution to prepare calibration curve in potassiumrange of 1 to 10 mglL.

d. Standard potassium solution: Dilute 10.0 mL intermediatepotassium solution with water to 100 mL; 1.00 mL = 0.0 10 mgK. Use this solution to prepare calibration curve in potassiumrange of 0.1 to 1.0 mglL.

4. Procedure

Make deterrnination as described in Section 3500-Na.B.4, butmeasure emission intensity at 766.5 nm.

5. Calculation

See Section 3500-Na.B.5.

6. Precision and Bias

A synthetic sample contallllllg 3.1 mg K+/L, 108 mgCa2+1L, 82 mg Mg2+/L, 19.9 mg Na+/L, 241 mg Cl-/L, 0.25

mg N02 - -NIL, 1.1 mg N03 - -NIL, 259 mg S042-/L, and 42.5mg total alkalinity/L (contributed by NaHC03) was analyzedin 33 labaratories by the flame photometric method, with arelative standard deviation of 15.5% and a relative errar of2.3%.

MEHLICH,A.& R.J.MONROE.1952.Reportonpotassiumana1ysesby meansof flamephotometermethods.J. Assoe.Dific. Agr. Chem.35:588.

AIso see 3500-Na.B.7.

a. Principie: Potassium ion is measured potentiometrically byusing a potassium ion-selective electrade and a double-junction,sleeve-type reference electrade. The analysis is performed witheither a pH meter having an expanded millivolt scale capable ofbeing read to the nearest 0.1 mV or a specific ion meter havinga direct concentration scale for potassium.

Before measurement, an ionic strength adjustor reagent isadded to both standards and samples to maintain a constant ionicstrength. The electrode response is measured in standard solutionswith potassium concentrations spanning the range of interest usinga calibration line derived either by the instrurnent meter or manu-ally. The electrode response in sample solutions is measured fol-lowing the same procedure and potassium concentration determinedfrom the calibration line or instrurnent direct readout.

b. Interferences: Although most sensitive to potassium, thepotassium electrode will respond to other cations at highconcentrations; this can result in a positive bias. Table 3500-K:I lists the concentration of common cations causing a 10%errar at various concentrations of potassium chloride with abackground ionic strength of 0.12N sodium chlaride. Of thecations Iisted, ammonium ion is most often present in samplesat concentrations high enough to result in a significant bias. Itcan be converted to gaseous ammonia by adjusting to pH >10.

An electrode exposed to interfering cations tends to drift andrespond sluggishly. To restore normal performance soak elec-trode for 1 h in distilled water and then for several hours in astandard potassium solution.

TABLE3500-K:I. CONCENTRATIONOFCATIONSOOERFERINGATVARIOUSCONCENTRATIONSOFPOTASStUM

ConcentrationCausing 10%Errormg/L

Cation K cone = 1 mg/L K cone = 10 mg/L K cone = 100 mg/L

Cs+ 1.0 10 100NH4+ 2.7 27 270Tl+ 31.4 314 3140Ag+ 2765 27650 276500Tris+ 3105 31050 310500Li+ 356 3560 35600Na+ 1 179 11790 117900H+ 3.6* 2.6* 1.6*

* pH.

c. Detection limits: Samples containing from 0.1 to 1000 mgK+IL may be analyzed. To measure higher concentrations dilutethe sample.

a. Expanded-scale or digital pH meter or ion-selective meter.b. Potassium ion-selective electrode.c. Sleeve-type double-junction reference electrode: Fill outer

sleeve with reference electrode filling solution (see 'li 3b). Fillinner sleeve with inner filling solution provided with the elec-trode.

d. pH electrode.e. Mixer, magnetic, with a TFE-coated stirring bar.

a. Ionic strength adjustor (lSA): Dissolve 29.22 g NaCI inreagent water and dilute to 100 rnL.

b. Reference electrode outer sleeve filling solution: Dilute 2rnL ISA solution to 100 rnL with reagent water.

c. Stock potassium solution: See 3500-K.B.3b.d. Sodium hydroxide, NaOH, 6N.e. Reagent water: See Section 1080.

a. Preparation of standards: Prepare a series of standardscontaining 100.0, 10.0, 1.0, and 0.1 mg K+IL by making serialdilutions of the stock potassium solution as in Section 3500-K.B.3c and 3d.

b. Instrument calibration: Fill reference electrode according tothe manufacturer' s instructions using reference electrode fillingsolution. Transfer 100 rnL 0.1 mg K+/L standard into a 150-rnLbeaker and add 2 rnL ISA. Raise pH to about 11. Stir gently withmagnetic mixer. Immerse electrodes, wait approximately 2 min forpotential stabilization and record meter reading. Thoroughly rinseelectrodes and blot dry. Repeat for each standard solution in orderof increasing concentration. Prepare calibration curve on semiloga-rithmic graph paper by plotting observed potential in millivolts(linear scale) against concentration (log scale). Altematively, calcu-late calibration line by regression analysis.

c. Analysis of samples: Transfer 100 rnL sample into a150-rnL beaker and follow procedure applied to standards in 14babove. From the measured response, calculate K+ concentrationfrom calibration curve.

Reproducibility of potential measured, over the method' srange, can be expected to be ± 0.4 mV, corresponding to about± 2.5% in concentration.

The slope of the calibration line should be -56 mV/lO-fold con-centrationchange. If the slope is outsidethe range of - 56 ±3 mV, theelectrode may require maintenance (replace filling solutions). If theproper electrode response cannot be obtained, replace electrode.

Analyze an independent check standard with a mid-range potas-sium concentration throughout analysis of a series, initially, every

Rhenium (Re) is the third element in Group VIIB in theperiodic table; it has an atomic number of 75, an atomic weightof 186.21, and valences of I through 7, with 7 being the moststable. The average abundance of Re in the earth's crust is 7ppm, and in groundwaters it is <0.1 mglL. Rhenium is found incolumbite, tantalite, and wolframite, as well as in molybdenum

Rhodium (Rh) is the fifth element in Group VIII in theperiodic table; it has an atomic number of 45, an atomic weightof 102.91, and valences of I through 6, the most common beingI and 3. Rhodium is found in its native state in platinum-bearingsands. It is used in platinum alloys for thermocouples, electricalcontacts, and jewelry.

Ruthenium (Ru) is the fourth element in Group VIII in theperiodic table; it has an atomic number of 44, an atomic weightof 101.07, and valences of I through 7, the most common being2, 3, and 4. The average abundance of Ru in the earth's crust isprobably <0.01 ppm, and in groundwaters it is <0.1 mglL. Itoccurs in its native state in platinum-bearing river sands. It is

ten samples, and after final sample. If the value has changed bymore than 5%, recalibrate electrode. Analyze a reagent blank at thesame frequency. Readings must represent a lower concentrationthan the lowest concentration standard (0.1 mgIL).

PIODA, L., V. STANKOVA& W. SIMON. 1969. Highly selective potassiumion responsive liquid membrane electrode. Ana!. Lett. 2 (12): 665.

MIDGLEY,D. & K. TORRANCE.1978. Potentiometric Water Analysis. JohnWiley & Sons, New York, N.Y.

BAlLEY, P.L. 1980. Ana1ysis with Ion-Selective Electrodes. Heyden &Son Ltd., Philadelphia, Pa.

ore concentrates. It is used in tungsten-molybdenum-based al-loys, thermocouples, filaments, and flash bulbs. Rhenium in thepowder form can be f1ammable.

For analysis methods, see f1ame atomic absorption methods(3111D and E). The inductively coupled plasma mass spectro-metric method (3125) also may be applied successfully in mostcases (with lower detection levels), even though rhenium is notspecifically listed as an analyte in the method.

The aqueous chemistry in natural waters is relatively un-known. The metal is f1ammable in the powder form, and its saltsare toxic by inhalation.

For analysis see f1ameatomic absorption method (3111B). Theinductively coupled plasma/mass spectrometric method (3125)also may be applied successfully in most cases (with lowerdetection levels), even though rhodium is not specifically listedas an analyte in the method.

used in jewelry with platinum, in electrical contacts, and as acatalyst.

The aqueous chemistry in natural waters is relatively un-known. Ruthenium has no known toxic effects.

For analysis see f1ameatomic absorption method (3111B). Theinductively coupled plasma mass spectrometric method (3125)also may be applied successfully in most cases (with lowerdetection levels), even though ruthenium is not specifically listedas an analyte in the method.

1. Occurrence and Significance

Selenium (Se) is the third element in Group VIA in theperiodic table; it has an atomic number of 34, an atomic weightof 78.96, and valences of 2, 4, or 6. The average abundance ofSe in the earth's crust is 0.2 ppm; in soils it is 0.27 to 0.74 ppm;in streams it is 0.2 JLgfL, and in groundwaters it is <0.1 mgfL.Selenium is used in electronics, ceramics, and shampoos.

The inorganic fraction of dissolved selenium consists predom-inantly of selenium as the selenate ion (SeO/-), designated hereas Se(VI), and selenium as the selenite ion (Se032-), Se(IV).Other common aqueous species include Se2-, HSe -, and SeGoSelenium is considered a nonessential trace element for mostplants, but is an essential trace nutrient for most animaIs, andselenium deficiency diseases are well known in veterinary med-icine. Above trace levels, ingested selenium is toxic to animaIsand may be toxic to humans. While the selenium concentrationof most natural waters is low, the pore water in seleniferous soilsin semiarid areas may contain up to hundreds or thousands ofmicrograms dissolved selenium per liter. Certain plants thatgrow in such areas accumulate large concentrations of seleniumand may poison livestock that graze on them. Water drainedfrom such soil may cause severe environmental pollution andwildlife toxicity. Selenopolysulfide ions (SSe2-) may occur inthe presence of hydrogen sulfide in waterlogged, anoxic soils.Selenium derived from microbial degradation of seleniferousorganic matter includes selenite, selenate, and the volatile or-ganic compounds dimethylselenide and dimethyldiselenide.Nonvolatile organic selenium compounds may be released towater by microbial processes. Soluble selenium may be leachedfrom coal ash and fly ash at electric power plants that bumseleniferous coal.

The United Nations Food and Agriculture Organization rec-ommended maximum leveI for selenium in irrigation waters is20 JLgfL. The U.S. EPA primary drinking water standard MCL is50 JLgfL.

The selenium methods using hydride generation atomic ab-sorption (31l4B and C), electrothermal atomic absorption(31l3B), and derivatization colorimetry (C) are the most sensi-tive currently available. For determination of selenium at higherconcentrations, the inducti vely coupled plasma methods (3120and 3125) may be used.

By using suitable preparatory steps to convert other chemicalspecies to Se(IV), it is possible to distinguish the chemicalspecies in the sample. In drinking water and most surface andground waters, Se(IV), Se(VI), and particulate selenium fre-~uen~ly .are the only significant species. However, when specia-tIon 1S nnportant, for example, when a new matrix is beinganalyzed, the general analytical scheme shown in Figure 3500-Se: I may be carried out as follows: Determine volatile seleniumby stripping sample with nitrogen or air and collecting selenium

* Approved by Standard Methods Committee 1999Joint Task Group: 20th Edition-See 3500-Ai. .

in alkaline hydrogen peroxide (see Method D). To obtain anestimate of selenium in suspended particles, determine totalselenium, filter sample, and make a second determination of totalselenium. In any case, filter sample. Occasionally, a filteredsample may have the odor of hydrogen sulfide and a yellowcolor; such a sample may contain selenopolysulfides, which maybe estimated by comparing results of total selenium analysesbefore and after acidification, stripping with nitrogen, settling for10 min, and refiltration. Determine selenite, Se(lV), by analyzingfiltered water sample directly by Methods 3114B or C, or by3500-Se.C or E. In principIe, sample digestion with HCI willconvert Se(VI) to Se(IV), and the value determined will equalthe sum of the two species. In practice, samples frequentlycontain an unknown masking agent that produces an unduly lowresulto Test for this effect by analyzing samples with knownadditions of both species. If recovery is good, the HCI digestionfollowed by analyses will yield reliable results. If recovery ispoor and organic selenium is to be determined subsequently,attempt to remove the interference by sample pretreatment withresin (B.l). Interference also can be eliminated by digestion with

Acidify,strip with N2•let stand 20 min,filter

Treat sampleto removeinterference(3500-Se.6.1)

Se(VI) = [Se(VI)+Se(IV)]-Se(IV)

Organic Se = Total Se- [Se(VI)+Se(lV)]

an oxidizing agent (B.2, 3, and 4), but these procedures preventdistinguishing of Se(Vl) and organic selenium and also oxidizemany organic seIenium compounds. To measure nonvoIatiIeorganic seIenium compounds, use Method E.

The choice of digestion method for oxidizing interferencesand organic seIenium depends on sample matrix. The methodsdescribed in B.2, 3, and 4, in order of increasing compIexity anddigestive ability, use ammonium (or potassium) persuIfate, hy-drogen peroxide, and potassium permanganate. Ammonium per-sulfate digestion is adequate for most filtered ground, drinking,and surface water. Hydrogen peroxide digestion may be requiredif organic selenium compounds are present, and potassium per-manganate digestion may be needed with unfiltered samples orthose containing refractory organic seIenium compounds. Con-firm results obtained with one digestion method by using a morerigorous method when characterizing a new matrix.

lnterferences are found in certain reagents, as well as insamples. Recognition of the presence of an interferant is critical,especially when unknown sample matrices are being anaIyzed.RoutineIy add Se(lV) and Se(Vl) to test for interference. Ifpresent, characterize the interference and correct by the methodof standard additions. A slope Iess than one indicates interfer-ence. ln cases of mild interference (recoveries reduced by 25%or Iess), the standard additions method will largely correct de-termined vaIues.

Because the hydride atomic absorption method is extremeIysensitive, sampIes frequentIy need to be diluted to bring themwithin the linear range of the instrument. DiIuting a filtered watersample frequently will eliminate sample-related interferences.

lnclude full reagent bIanks in each run to ensure absence ofcontamination from reagents. Hydride generator atomic absorp-tion is susceptible to common interference problems related tonitrite in the sample or free chIorine in the reagent HCI (seeMethods 3114B and C).

4. Bibliography

V.S. NATIONALACADEMYOF SCIENCES.1976. Selenium: Medical andBiological Effects of Environmental Pollutants. National Academyof Sciences, Washington, D.C.

CUrrER,G.A. 1978. Species determination of selenium in natural waters.Ana!. Chim. Acta 98:59.

ROBBERECHT,H. & R. VANGRIEKEN.1982. Selenium in environmentalwaters: Determination, speciation and concentration levels. Talanta29:823.

KUBOTA,J. & E.E. CARY.1982. Cobalt, molybdenum and selenium. lnA.L. Page et aI., eds. Methods of soil analysis, Part 2, 2nd ed.Agronomy 9:485.

RAPTIS,S. et aI. 1983. A survey of selenium in the environment and acritical review of its determination at trace levels. Fresenius Z.Ana!. Chem. 316:105.

CUITER,G. 1983. Elimination of nitrite interference in the determinationof selenium by hydride generation. Ana!. Chim. Aeta 149:391.

CAMPBELL,A. 1984. Critical evaluation of analytical methods for thedetermination of trace elements in various matrices. Part L Deter-mination of selenium in biological materials and water: Ana!. Chem.56:645.

LEMLY,A.D. 1985. Ecological basis for regulating aquatic emissionsfrom the power industry: The case with selenium. Regu!. Toxie.Pharm. 5:465.

OHLENDORF,H.M., DJ. HOFFMAN,M.K. SAIKI& T.W. ALDRlCH.1986.Embryonic mortality and abnormalities of aquatic birds: Apparentimpacts of seienium from irrigation drainwater. Sei. Total Environ.52:49.

1. Removal of Organic and lron Interference by ResinPretreatment

lnterferences are common in seIenium anaIysis, particularlywhen chemical speciation is attempted. Routine pretreatment ofwater samples as described here is not necessary. The methods inthis section should be tried when poor recovery of standardadditions indicates a problem.

Many waters contain iron and/or dissolved organic matter(humic acid) in quantities sufficient to interfere. Reduction ofSe(Vl) to Se(IV) usually is nonquantitative. When Se (VI)standard additions show poor recovery, treat the sampIe be-fore analysis. To remove dissolved organic compounds, passan acidified sample through a resin. Because dissolved or-ganic seIenium compounds also may be removed by thistreatment, aIso determine total selenium in the untreatedwater sampIe (see 2, 3, or 4 below). To remove iron use astrong base ion exchange resin. * Iron is removed as the

anionic chIora complexo ln this treatment the acidity and ionexchanger do not alter speciation; complete speciation ofselenium is possible.

a. Apparatus:I) Chromatography column for organics removal, glass,

about 0.8 cm ID x30 cm long, with fluorocarbon metering valve.2) Chromatography column for ion exchange, disposable

poIyethyIene. t3) pH meter.b. Reagents:1) Organics-removal resin: Thoroughly rinse 16 to 50 mesh

resin:l: with deionized water and remove resin fines by decanting.Rinse three times with pH 12 soIution. Store resin in pH 12soIution and refrigerate to prevent bacteriaI growth.

2) Anion exchange resin: Add 100 to 200 mesh anion ex-change resin* to a beaker and thoroughly rinse with deionized

t Bio-Rad Econo-Columns or equivalent.:I: Amberlile XAD-S, Supelco, or equivalent.

water. Cover resin with 4N HCI, stir, and let settle. Decant andrepeat acid rinse twice more. Store resin in 4N HCI.

3) Hydrochlorie aGid, conc: Before use, bubble heliumthrough the acid for 3 h at rate of 100 rnL/min. (CAUTION:Use afume hood.)

4) pH 1.6 solution: Adjust pH of deionized water to 1.6 withHCI.

5) pH 12 solution: Adjust pH of deionized water to 12 withKOH.

e. Procedure:1) Organic removal-Place 5 cm washed resin in a 0.8-cm-ID

column. Precondition column, at 1 mLlmin, with 30 mL pH 12solution and 20 rnL pH 1.6 solution. Using HCI and a pH meteradjust sample to pH 1.6 to 1.8. Pass sample through precondi-tioned column at rate of 1 rnL/min. Discard first 10 mL and usenext 11 to 50 rnL collected for Se(IV) determinations by Meth-ods 3114B or C, or 3500-Se.C preceded, if Se(VI) aIso is to bedetermined, by preparatory step B.5. If more than 50 rnL sampleare needed, use another column or use a column with twice asmuch resin.

2) Iron removal-Place 4 cm prepared resin in a small chro-matographic column (add resin to column filled with 4N HCI toavoid air bubbles). Rinse column with 10 rnL 4N HCI at flow rate< 6 mLlmin. Let solution drain to top of resin, but do not let thecolumn mn dry. Adjust sample to 4N HCI and pour into column.Discard first 10 rnL and collect the next 11 to 100 rnL for Se(lV)analysis by Methods 3114B or C, or 3500-Se.C preceded, ifnecessary, by preparatory step B.2 and if Se(VI) also is to bedetermined, by preparatory step B.5 below.

The combination of this procedure with step B.5 below andMethods 3114B or C, or 3500-Se.C is, in most cases, the pre-ferred method for determining total selenium in filtered water. Asmall amount of ammonium or potassium persulfate is added tothe mixture of sample and HCI to remove interference fromreducing agents and to oxidize relatively labile organic seleniumcompounds such as selenoamino acids and methaneseleninicacid.

If the sample contains hydrogen sulfide or a large concentrationof organic matter or is otherwise suspect or to confirm methodaccuracy, reanalyze sample using digestion procedure 3 or 4 below.

Prepare 2% potassium or amnl0nium persulfate solution bydissolving 2.0 g in 100 rnL deionized water (prepare weekly).Add 0.2 rnL persulfate solution to the mixed sample and HCI of<j[ B.5e before heating and proceeding with pretreatment andanalysis.

After completing analysis multiply concentration of seleniumdetermined in the acidified sample by 2.04 to obtain total sele-nium in original sample.

3. Removal of Interference by Alkaline Hydrogen PeroxideDigestion

Occasionally, digestion with persulfate gives incomplete re-covery of total selenium. In this case, digestion with hydrogenperoxide is used to remove ali reducing agents that might inter-fere and to fully oxidize organic selenium to Se(VI). The result-

ing solution can be analyzed for total selenium after pretreatmentaccording to step B.5 below.

This method is suitable for determining total selenium inunfiltered water samples, where particulate selenium is present.When working with a new matrix, confirm results obtained byreanalyzing the sample using digestion procedure 4, below.

a. Apparatus:1) Beakers, 150-rnL.2) Wateh glasses.3) Hot plate.4) Pipettor, 1-mL, and tips.5) Graduated eylinder, 25-rnL.b. Reage/7fs:1) Hydrogen peroxide, H20b 30%. Keep refrigerated.2) Sodium hydroxide, NaOH, 1N.3) Hydroehlorie aGid, HCI, 1.5N: Dilute 125 rnL conc HCI to

1 L with deionized water.e. Procedure: Add 2 rnL 30% H202 and I rnL 1 N NaOH to

25 mL sample in a beaker. Cover beaker to control spattering andsimmer on hot plate until fine bubbles characteristic of H202

decomposition subside and are replaced by ordinary boiling. Add1 rnL 1.5N HCI to redissolve any precipitate that may haveformed, 1et cool, and pour into graduated cylinder. Rinse beakerwith deionized water into graduated cylinder and make volumeup to 25 rnL. Proceed to B.5 and chosen analytical method.

This digestion method utilizes potassium permanganate tooxidize selenium and remove interfering organic compounds.Excess KMn04 and Mn02 are removed by reaction with hydrox-ylamine. HCI digestion is included here, because it is conve-niently performed in the same reaction vial. Selenite may then bedetermined directly by Methods 3114B or C, or 3500-Se.C.

Verify recovery for the given matrix. This method gives goodrecovery even with heavily contaminated water samples thatcontain organic selenium compounds, dissolved organic matter,and visible suspended material.

Permanganate may oxidize chloride ion to free chlorine. Partof the chlorine (whieh interferes with hydride analysis) is re-moved by reaction with hydroxylamine, but the best way toeliminate free chlorine is by prolonged heating in an open vialduring the digestion step. Excess hydroxylamine may redueereeovery by reducing selenium to Se(O).

a. Apparatus:1) Oven with thermostat, for continuous operation at 110 ±

5°e.2) Digestion vials, 40-rnL, with fluorocarbon-lined screw

eaps.3) Metal support raek to hold 40 digestion vials.b. Reagents:1) Hydroehlorie aGid, HCI, cone. (See B.lb3)].2) Hydroehlorie aGid, HCI, lON: Dilute 1000 rnL conc HCI to

1200 rnL with deionized water.3) Sulfurie aGid, H2S04, cone and 25%. NOTE:Many brands of

H2S04 are contarninated with selenium. Run reagent blanks whenstarting a new bottle. Make 25% v/v solution by adding 250 rnLconc H2S04 to 500 rnL deionized water, and diluting to 1 L.

4) Potassium permanganate solution, KMn04, 5% (w/v):Dissolve 50 g KMn04 in 1000 rnL deionized water.

5) Hydroxylamine hydrochloride solution: Dissolve 100 gNH20H . HC! in 1000 rnL deionized water.

c. Procedure: Pipet 5 rnL samp!e into digestion via!, add 5rnL 25% H2S04 and 1 rnL KMn04 solution. Lightly screw capon and place in preheated oven at 110°C for 1 h. CAUTION:Excessive pressure may build up in tightly capped vials. Care-fully remove tray with vials from the oven (CAUTION:potentialfor acid gas vapor release) and cool to room temperature. Openvial, careful1y add a few drops hydroxylamine hydrochloridesolution, mix, and wait unti! sample is decolorized and residualmanganese dioxide is dissolved. Avoid excess hydroxylaminesolution, which can cause a low reading. Add 10 rnL cone HC!to the sample and heat vial 60 min at 95°C without capoLet coolto room temperature. Transfer sample to a 25-rnL volumetricflask or graduated cylinder, rinse vial into flask, dilute to mark,and mix well. Proceed to analyze by Methods 3114B or C, or3500-Se.C. If Method 3114 B or C is used, multiply spectrom-eter readings by the dilution factor as follows:

final volumeConcentration,ILg/L= ------ X readingvolumeof sample

5. Reduction of Se(Vl) to Se(lV) by Hydrochloric AcidDigestion

Se(VI) is reduced to Se(IV) by digestion with HCI. DetermineSe(IV) + Se(VI) by either hydride generation atomic absorptionspectrometer (Method 3114B or C) or as an organic derivative(Method C).

Test any given sample matrix to ensure recovery of addedSe(VI). If recovery is poor, try to remove interference by theprocedure of lJ[ B.I, above, or if at least 75% recovery isachieved, use the method of standard additions. The methoddescribed here is of limited utility for direct analysis of watersamples, but is useful as a step in determining total selenium ina sample where selenium has been oxidized to Se(VI).

NOTE:Available literature does not provide definitive criteriafor collection and preservation of samples for which speciation isdesired. li speciation is needed, fully evaluate collection andpreservation as part of the data quality objectives.

a. Apparatus:I) Dispenser, bottle type, 5-rnL, suitable for dispensing con-

centrated HCI.2) Pipettors, 0.2- and 5-rnL.3) Screw-cap culture tubes, borosilicate glass, 25- X 150-

mm.4) Boiling water bath, suitable for heating culture tubes; a l-L

beaker on a hot plate is suitable.b. Reagents:1) Sodium selenate additions solution: Dilute 1000 mglL

stock selenate solution with deionized water to prepare a so!utionof I to 10 mglL, such that the concentration of the additionssolution will be approximately 50 times greater than anticipatedtotal selenium in the sample to be ana!yzed.

2) Hydrochloric acid, HCI, cone: See Rlb3).C. Procedure: Calibrate acid dispenser using water. Preheat

water bath. Pipet 5 rnL filtered sample into a culture tube. Add5 rnL cone HCI. Loosely cap tube (do not tighten) and place inboiling water bath for 20 mino Let tube cool and tighten capoDetermine total Se(IV) by Methods 3114B or C, or 3500-Se.c.

Add 0.200 rnL additions solution with a microliter pipet tosample and proceed as above. Ana!yze a deionized water blankand a blank with the addition to ensure absence of contaminationand to determine the true value of the addition.

Multiply the concentration of selenium determined in theacidified sample by 2.00 to obtain total concentration of Se(IV)+ Se(VI). Multiply reading obtained for sample with addition by2.04.

JANGHORBANI, M., B. TING, A. NAHAPETLAN & R. YOUNG. 1982.Conver-sion of urinary selenium to selenium IV by wet oxidation.Ana!.Chem. 54: 1188.

ÁDELOJU, S.B. & A.M. BOND. 1984. Critical evaluation of some wetdigestionmethods for the strippingvoltammetricdeterminationofseleniumin biologicalmateriais.Ana!. Chem. 56:2397.

BRJMMER, S.P., W.R. FAWCETI & K.A. KULHAVY. 1987. Quantitativereduction of selenate ion to selenite in aqueous samples. Ana!.Chem. 59:1470.

a. Principle: This method is specific to determining seleniteion in aqueous solution. Selenite ion reacts with 2,3-diamino-naphthalene to produce a brightly colored and strongly fluores-cent piazselenol compound, which is extracted in cyc10hexaneand measured colorimetrically.

The optimum pH for formation of the piazselenol complex isapproximate!y 1.5 but should not be above 2.5 because above pH2, the rate of formation of the colored compound is criticallydependent on pH. When indicators are used to adjust pH, resultsfrequently are erratic; results can be improved when pH ismonitored electrochemically.

b. Interference: No inorganic compounds are known to give apositive interference. Colored organic compounds extractable bycyclohexane may be encountered, but usually they are absent orcan be removed by oxidizing the sample (see B.2, 3, or 4) or bytreating it to remove dissolved organics (see RI). Negativeinterference results from compounds that reduce the concentra-tion of diaminonaphthalene by oxidizing it. Addition of EDTAeliminates negative interference from at least 2.5 mg Fe2+.

C. Minimum detectable quantity: 10 JLg SelL.

a. Colorimetric equipment: A spectrophotometer, for use at480 nm, providing a light path of 1 em or longer.

b. SeparatOlY funnel, 250-rnL, preferably with a fiuorocarbonstopcock.

c. Thermostatieally controlled water bath (50°C) with cover.d. pH meter.e. Centrifuge, with rotor for 50-rnL tubes (optional).f Centrifuge bottles, 60-mL, screw-capped, fiuorocarbon.g. Shaker, suitable for separatory funnel (optiona!).

Use reagent water (see Section 1080) in preparing reagents.a. Selenium standard referenee solution: Dissolve 2.190 g

sodium selenite, Na2Se03, in water containing 10 rnL HCI anddilute to 1 L. 1.00 rnL = 1.00 mg Se(IV).

b. Working standard selenium solutions: Dilute selenium ref-erence standard solution with water or suitable background so-lution to produce a series of working standards spanning theconcentration range of interest.

c. Hydrochloric acid: HCl, conc and O.IN.d. Ammonium hydroxide, NH40H, 50% v/v.e. Cyclohexane, C6H12.

f 2,3-Diaminonaphthalene (DAN) solution: Dissolve 200 mgDAN in 200 rnL O.IN HCI. Shake 5 minoExtract three times with2S-rnL portions of cyclohexane, retaining agueous phase anddiscarding organic portions. Filter into opague container* andstore in cool, dark place for no longer than 8 h. CAUTION:Toxic,handle with extreme care.

g. Hydroxylamine-EDTA solution (HA-EDTA): Dissolve4.5 g Na2EDTA in approximately 450 rnL water. Add 12.5 ghydroxylamine hydrochloride (NH20H . HCl) and adjust vol-ume to 500 mL.

a. FOlmation ofpiazselenol: Add 2 rnL HA-EDTA solution to10 rnL sample in 60-rnL centrifuge bottle (filtered if Se(IV) is tobe detennined; oxidized using Method B.2, 3, or 4, then reducedusing Method B.5 for total Se). Adjust to pH 1.5 ::+:: 0.3 with O.INHCl and SO% NH40H, using a pH meter. Add S rnL DANsolution and heat in a covered water bath at SO°C for 30 mino

b. Extraction of piazselenol: Cool and add 2.0 mL cyclohex-ane. Cap container securely and shake vigorously for S minoLetsolution stand for S min or until cyclohexane layer becomes wellseparated. If separation is slow, centrifuge for 5 min at 2000 rpm.Place bottle in a clamp on a ringstand at a 4So angle to thevertical. Remove agueous phase using a disposable pipet at-tached to a vacuum line. Transfer organic phase to a smallcapped container using a clean disposable pipet, or to the spec-trophotometer cuvette if absorbance is to be read immediately.

c. Determination of absorbanee: Read absorbance at 480 nmusing a zero standard. The piazselenol color is very stable butevaporation of the cyclohexane concentrates the color unless thecontainer is capped. CAUTION:Avoid inhaling cyclohexane va-pors. Beer's Law is obeyed up to 2 mglL.

Construct calibration curve using at least a three-point stan-dard curve to bracket the expected sample concentration. Plotabsorbance vs. concentration. Correct for digestion blank andany reagent blank.

Three standard reference materiais (wheat fiour, water, and acommercial standard) were used to evaluate Se recovery.1 Thewheat fiour sample was digested using HN03 and HCI04 toconvert total selenium to Se(VI), digested with HCl to convertSe(VI) to Se(IV) and finally, the colorimetric method was used.Results were as follows:

Selenium ConcentrationJLg SelL

NBS, SRM 1567, wheat flourtNBS, SRM 1543ib, waterFisher Certified AAS Standard

1097::+:: 1979.7::+:: 0.5

1002:': 8

1113 ::+::88.7:': O

1002:': O

* Analyses in triplicate.t Dry weight basis.

1. HOLTZcLAw, K.M., R.H. NEAL, G. SPOSITO& S.J. TRAINA. 1987. Asensitive colorimetric method for lhe quantitation of selenite in soilsolutions and natural waters. Soil Sei. Soe. Amer. J. 51:75.

HOSTE, l. & l. GlLLlS. 1955. Spectrophotometric determination of tracesof selenium with 3,3' -diaminobenzidine. Ana/. Chim. Aeta. 12: 158.

CHENG, K. 1956. Determination of traces of selenium. Ana/. Chem.28: 1738.

MAGIN, G.B. et aI. 1960. Suggested modified method for colorimetricdetermination of selenium in natural waters. 1. Amer. Water WorksAssoe. 52: 1199.

RossuM, l.R. & P.A. VILLARRUZ.1962. Suggested methods for deter-mining selenium in water. J. Amer. Water Works Assoe. 54:746.

OLSEN, O.E. 1973. Simplified spectrophotometric analysis of plants forselenium. J. Assoe. Offie. Anal. Chem. 56:1073.

Dimethylselenide and dimethyldiselenide are low boiling, ex-tremely malodorous organic compounds sparingly soluble inwater. They are produced by microbial processes in seleniferoussoil and decaying seleniferous organic matter, and occasionalIyare present in natural waters. They are readily air stripped froma water sample and can be colIected with high efficiency in analkaline solution of hydrogen peroxide, which oxidizes themquantitatively to Se(VI). Total selenium is determined by diges-tion with HCI and analysis by the hydride atomic absorption(3114B or C) or colorimetríc (3S00-Se.C) methods.

Either nitrogen or air may be used to strip the sample. Pref-erably use nitrogen if air-sensitive compounds (e.g., selenopo-lysulfides) are suspected.

Because volatile selenium can be lost in the course of samplecolIection and handling, preferably air-stríp the sample in thefield immediately after it is colIected. After boiling to decomposeH202, retum the alkaline peroxide solution to the laboratory foranalysis.

AlI apparatus required for selenate reduction (B.S) and Meth-ods 3114B or C, or 3S00-Se.C, plus:

a. Gas washing bottles, borosilicate glass, 2S0-rnL, withcoarse porous glass gas dispersion frit. Mark 100-rnL leveI onside of bottle.

b. Rotameter, to measure 3 Llmin air ftow.c. Gas flow regulatord. Hot plate.e. Graduated cylinder, lOO-rnL.f Beakers, 2S0-rnL.g. Rubber tubing, to interconnect gas washing bottles and

other gas equipment.h. Rubber gloves.

AlI reagents required for selenate reduction (B.S) and Methods3114B or C, or 3S00-Se.C, plus:

a. Hydrogen peroxide, H202, 30%. Refrigerate.b. Sodium hydroxide solution, NaOH, lN.c. Compressed air or nitrogen.

Set up air-ftow train in this order: Regulated air supply ~rotameter ~ gas washing bottle 1 ~ gas washing bottle 2.

Prepare alkaline peroxide solution immediately before use bypouring 20 rnL 30% H202 into a 100-rnL graduated cylinder,adding SOto 60 rnL deionized water and S mL lN NaOH, andmaking up to 100 rnL. CAUTlON:Alkaline H20Z is unstable. Donot keep in glass bottle; hold at about O °C in oversized plasticbottle. Solution is corrosive; protect eyes and skin. Pour into gaswashing bottle 2. Pour approximately 100 rnL freshly colIectedsample into gas washing bottle 1. Do not attempt to measuresample volume accurately before volatile selenium determina-tion, as unnecessary handling may cause volatile selenium to belost.

Connect and check alI air lines, tum on air and adjust ftow to3 Llmin. Stríp for 30 min or more. After 30 min, tum off air,disconnect gas washing bottle 2, and place it on the hot plate.Adjust heat to produce a gentle simmering of oxygen bubblesfrom decomposition of H202• Continue heating until the char-acteristic effervescence of oxygen subsides and is replaced byordinary boiling. Remove from hot plate and let cool. Poursolution into beaker. (Volume wilI be very near to 100 rnL, andcorrection wilI usualIy be unnecessary.) Analyze for total sele-nium using HCI digestion (B.S) and Methods 3114B or C, or3S00-Se.c. Once boiled, this solution may be safely stored andtransported in plastic bottles. Measure volume of sample in gaswashing bottle 1.

The concentration of volatile selenium compounds in theoriginal water sample can be calculated as:

100C = f . x cone of Se in solution

volume o onginal sample

Approximately 90% of dimethylselenide in samples wilI berecovered with 30 min air stripping. The recovery of dimethyl-diselenide is not known. Loss of gases to the atmosphere duringsampling and handling that precede analysis may cause a sig-nificant negative error.

In principIe, the total amount of dissolved organic seleniumplus polysulfidic selenium may be estimated by comparing "totalSe," determined by oxidation and HCI digestion (B.2 or 3 andB.S, or BA), folIowed by Methods 3114B or C, or 3S00-Se.C,

with Se(IV) + Se(VI) determined by HCI digestion (B.S) andMethod 3114C, or 3S00-Se.C. In practice, this wilI give a mean-ingful estimate only if a known addition of Se(VI) is fulIyrecovered. Even if recovery is good, this estimate may be unre-Iiable, because it is the difference of two (frequently larger)numbers determined by slightly different methods. Comparing

total Se before and after treatment with resin [B.leI)] gives asimilarly unreliable estimate of nonvolatile organic Se.

It is preferable to separate and directIy determine nonvolatileorganic selenium. One method involves adsorption of dissolvedorganic matter onto a C-18 reverse phase HPLC resin, elutionwith an organic solvent, and determination of selenium in thisfraction. While this technique is relatively simple, it is affectedby pH and small organic molecules (e.g., individual seleniumcontaining amino acids) are not retained by the resins. Adjustsample pH to 1.5 to 2.0 before using the column, but because thelatter problem cannot be solved easiJy, the use of organic adsarbentsprovides only an estimate of organic selenium concentration.

Altematively, isolate specific compounds and determine theirselenium content. In some natural waters selenium may beassociated with dissolved polypeptides or small proteins, andeven small amounts of free selenoamino acids may be present.Because selenoamino acids are the most toxic form of theelement, a direct determination is sometimes desirable.

To detennine selenium in dissolved peptides, hydrolyze with acidand isolate the free amino acids via ligand exchange chromatogra-phy. Elute the selenoamino acids from the colurnn and detennineselenium. Selenoamino acids are unstable during acid hydrolysisand even using nonoxidizing methyl sulfonic acid and nitrogen-purged glass ampules, selenoarnino acid recoveries are only 50 to80%. This method is good only for estimating protein-bound seJe-nium. A somewhat more reliable estimate of free selenoamino acidsand selenium associated with small oligopeptides is obtained byperfonning a similar procedure without the hydrolysis step.

While these methods are toa intricate for routine use, semi-quantitative, and sensitive only to certain classes of organiccompounds, at present they are the only ones available with anypractical experience.

Imperfect separation of organic selenium compounds frominorganic forms of selenium may cause interference. In parallelwith the actual determination, always perform the procedureusing a solution compounded to resemble the actual matrix andcontaining a similar amount of selenium, but in the form ofSe(IV) and Se(VI) to determine degree of interference.

a. Rotary evaporator, with temperature-control bath and30-roL pear-shaped flasks.

b. Glass ampule sealing apparatus, ar oxygen-gas torch.c. Heating block, 100°C, or pressure cooker.d. Glass chromatography columns, 15 cm long, 0.7 cm ID.*e. Glass syringe, 50-roL.f Glass ampules, 10- ar 20- roL. Clean by heating in a muffle

fumace at 400°C for 24 h.g. pH meter.

a. Hydrochloric acid, HCI, lN.b. Methyl sulfonic acid, conc.c. Ammonium hydroxide, NH40H, 1.5N: Dilute 100 roL conc

NH40H to 1 L with deionized water.

d. Sodium hydroxide, NaOH pellets and IN solution.e. pH 1.6 solution: Adjust pH of deionized water to 1.6 using

HCI.f pH 9.0 solution: Adjust pH of deionized water to 9.0 using

NaOH.g. Copper sulfate solution, CuS04, 1M: Dissolve 25 g

CuS04 • 5H20 in deionized water and dilute to 100 roL.h. Methanol, purified.i. C-18 cartridges:t Using a glass syringe, pass the following se-

quence of reagents through the cartridge to c1ean the resin (6 mUminar less): 10 mL deionized water; 20 mL lN HCI; 10 mL deionizedwater; 20 mL methanol; 10 mL deionized water; and 20 mL pH 1.6solution. Refrigerate but do not freeze c1eaned cartridges.

j. Ligand-exchange chromatographic resin, 100/200 mesh:Rinse resint with deionized water to remove fines. Then rinseresin three times in the following sequence: lN HCI; 1.5NNH40H; and deionized water. Store resin wet.

k. Copper-treated ligand-exchange chromatographic resin,100/200 mesh: Rinse resint with deionized water to removefines. Then rinse three times with lN HCI, followed by deionizedwater. Using NaOH adjust pH of supernatant above the resin toabout 7. Add CuS04 solution to the resin and stir. After settling,decant supematant and add more CuS04 solution. Decant CuS04

solution and rinse with deionized water until no copper is no-ticeable in the supernatant. Rinse resin three times with 1.5NNH40H and three times with deionized water. Store resin wet.

a. Extractable organic selenium: Adjust sample (5 to 50 mL) topH 1.5 to 2.0 using HCI and place in a c1ean glass syringe with anattached c1eaned C-18 cartridge. Push sample through the cartridgeat a rate of 6 mL/min. After removing the cartridge, draw 2 roL pH1.6 solution into syringe as a rinse, reattach cartridge, and push therinse through the cartridge. Repeat two additional times. The car-tridge can be refrigerated for storage. To elute organic selenium,push 10 mL methanol through the cartridge at rate of 2 mUmin andcoUect eluate in a 30-mL pear-shaped flask. Remove methanol byrotary evaporation, with the water bath temperature less than 40°C.Use deionized water to solubilize and transfer the residue into thevessel used for total selenium digestions. Determine total dissolvedselenium by digestion with persulfate (B.2) or peroxide (B.3),reduction of Se(VI) (B.5), and analysis by Methods 3114B or C, or3500-Se.C.

b. Hydrolysis of protein-bound selenium: Place filtered sam-pIe in a 10- or 20-mL glass ampule (depending on desiredvolume), and add conc methyl sulfonic acid to adjust concentra-tion to 4M. Purge acidified sample with nitrogen for 10 min andseal top with a tarch. Heat sealed vial at 100°C for 24 h in aheating block ar pressure cooker. Transfer cooled hydrolysissolution with deionized water rinses to a 50-mL beaker and placein an ice bath. Using NaOH pellets and lN NaOH, adjust to pH9.0, taking care not to allow solution to heat to boiling.

c. Determination of selenoamino acids: If sample is not hy-drolyzed as in lJ[ 4b, filter, and adjust pH to 9.0 using IN NaOH.

Fill an empty chromatographic column with deionized waterand add ammonium-form resin to a depth of 2 cm. Add copper-

t Sep-Pak. Waters Assaciates. ar equivalent.:j: Chelex 100 (ammanium rorm) Bia Rad, ar equivalent.

treated resin to form a 12-cm length of resin. (The ammonium-form resin removes any copper that bleeds from the copper-treated resin above it). Rinse with deionized water until the pHof the effluent is 9.0; maintain flow through the column bygravity. Pass sample through column, rinse sample beaker with5 mL pH 9 solution, and place the rinse on column (after the lastof the sample reaches the top of the resin). Rinse beaker twicemore. Discard flow through column.

Place clean beaker under column and add 20 mL 1.5N NH40Hto the column. Neutralize NH40H eluate with 2.5 mL cone HCI.Determine total dissolved selenium by digestion with persulfate(B.2) or peroxide (B.3), reduction of Se(VI) (B.5) and analysisby Methods 3114B or C, or 3500-Se.c.

Silver (Ag) is the second element in Group IB of theperiodic table; it has an atomic number of 47, an atomicweight of 107.87, and valences of I and 2. The averageabundance of Ag in the earth's crust is 0.08 ppm; in soils it is<0.01 to 0.5 ppm; in streams it is 0.3 pog/L; in D.S. drinkingwaters it is 0.23 pog/L, and in groundwater it is <0.1 pog/L.Silver occurs in its native state and in combination with manynonmetallic elements such as argentite (Ag2S) and hom silver(AgCI). Lead and copper ores also may yield considerablesilver. Silver is widely used in photography, silverware, jew-elry, mirrors, and batteries. Silver iodide has been used in theseeding of clouds, and silver oxide to a limited extent is usedas a disinfectant for water.

In acidic water Ag+ would predominate, and in high-chloride water a series of complexes would be expected.Silver is nonessential for plants and animais. Silver can causeargyria, a permanent, blue-gray discoloration of the skin andeyes that imparts a ghostly appearance. Concentrations in therange of 0.4 to I mg/L have caused pathological changes inthe kidneys, liver, and spleen of rats. Toxic effects on fish infresh water have been observed at concentrations as low as

Sodium (Na) is the third element in Group IA of the periodictable; it has an atomic number of 11, an atomic weight of 22.99,

* Approved by Standard Methods Commillee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

These procedures are only semiquantitative. Typical relativestandard deviation is 12% for the C-18 isolation of dissolvedorganic selenium, and 15% for protein-bound selenium.

CUITER, G. 1982. Selenium in reducing waters. Seienee 217:829.COOKE, T.D. & K.W. BRULAND.1987. Aquatic chemistry of selenium:

evidence ofbiomethylation. Environ. Sei. Teehnol. 21:1214.

0.17 poglL. For freshwater aquatic life, total recoverable silvershould not exceed 1.2 mglL.

The atomic absorption spectrometric methods (3111 B andC) and the inductively coupled plasma methods (3120 and3125) are preferred. The electrothermal atomization method(3113B) is the most sensitive for determining silver in naturalwaters. The dithizone method detailed in the 19th edition ofStandard Methods can be used when an atomic absorptionspectrometer is unavailable. A method suitable for analysis ofsilver in industrial or other wastewaters at levels above ImglL is available. I

If total silver is to be determined, acidify sample with conenitric acid (HN03) to pH <2 at time of collection. If samplecontains particulate matter and only the "dissolved" metal con-tent is to be determined, filter through a 0.45-pom membranefilter at time of collection. After filtration, acidify filtrate withHN03 to pH <2. Complete analysis as soon after collection aspossible. Some samples may require special storage and diges-tion; see Section 3030D.

1. V.S. ENVIRONMENTALPROTEcnON AGENCY.1994. Approved InorganicTest Procedures. Federal Register 59(20):4504.

and a valence of 1. The average abundance of Na in the earth'scrust is 2.5%; in soils it is 0.02 to 0.62%; in streams it is 6.3mglL, and in groundwaters it is generally >5 mglL. Sodiumoccurs with silicates and with salt deposits. Sodium compoundsare used in many applications, including caustic soda, salt,fertilizers, and water treatment chemicals.

Sodium is very soluble, and its monovalent ion Na+ can reachconcentrations as high as 15 000 mg/L in equilibrium with sodium

bicarbonate. The ratio of sodium to total cations is important inagriculture and human physiology. Soil permeability can be harmedby a high sodium ratio. In large concentrations it may affect personswith cardiac difficulties. A Jimiting concentration of 2 to 3 mgIL isrecommended in feedwaters destined for high-pressure boilers.When necessary, sodium can be removed by the hydrogen-ex-change process or by distillation. The U.S. EPA advisory limit forsodium in drinking water is 20 mg/L.

Method 3111B uses an atomic absorption spectrometer in theftame absorption mode. Method 3120B uses inductively coupledplasma; this method is not as sensitive as the other methods, butusually this is not important. Method 3500-Na.B uses either a ftame

photometer or an atomic absorption spectrometer in the ftameemission mode. The inductively coupled plasma/mass spectromet-ric method (3125) also may be applied successfully in most cases(with lower detection limits), even though sodium is not specificallylisted as an analyte in the method. When ali of these instruments areavailable, the choice will depend on factors including relative qual-ity of the instruments, precision and sensitivity required, number ofsamples and analytes per sample, matrix effects, and reiative ease ofinstrument operation. If an atomic absorption spectrometer is used,operation in the emission mode is preferred.

Store alkaline samples or samples contammg low sodiumconcentrations in polyethylene botdes to eliminate the possibilityof sample contamination due to leaching of the glass container.

a. Principie: Trace amounts of sodium can be determined byftame emission photometry at 589 nm. Sample is nebulized intoa gas ftame under carefully controlled, reproducible excitationconditions. The sodium resonant spectral line at 589 nm isisolated by interference filters or by light-dispersing devices suchas prisms or gratings. Emission light intensity is measured by aphototube, photomultiplier, or photodiode. The light intensity at589 nm is approximately proportional to the sodium concentra-tion. Alignment of the wavelength dispersing device and wave-length readout may not be precise. The appropriate wavelengthsetting, which may be slightly more ar less than 589 nm, can bedetermined from the maximum emission intensity when aspirat-ing a sodium standard solution, and then used for emissionmeasurements. The calibration curve may be linear but has atendency to levei off or even reverse at higher concentrations.Wark in the linear to near-linear range.

b. lnterferences: Minimize interference by incorporation ofone ar more of the following:

1) Operate at the lowest practical concentration range.2) Add releasing agents, such as strontium or lanthanum at

1000 mg/L, to suppress ionization and anion interference.Among common anions capable of causing interference are CI-,S042- and RC03 - in relatively large amounts.

3) Matrix-match standards and samples by adding identicalamounts of interfering substances present in the sample to cali-bration standards.

4) Apply an experimentally determined correction in thoseinstances where the sample contains a single important interfer-ence.

5) Remove interfering ions.6) Remove bumer-clogging particulate matter from the sam-

pie by filtration through a filter paper of medium retentiveness.7) Use the standard addition technique as described in the

ftame photometric method for strontium (3500-Sr.B). Themethod involves preparing a calibration curve using the samplematrix as a diluent, and determining the sample concentration

either mathematically or graphically.8) Use the internal standard technique. Potassium and cal-

cium interfere with sodium determination by the intemal-stan-dard method if the potassium-to-sodium ratio is ~5:1 and thecalcium-to-sodium ratio is ~ IO: 1. When these ratios are ex-ceeded, determine calei um and potassium concentrations andmatrix-match sodium calibration standards by addition of ap-proximately equivalent concentrations of interfering ions. Inter-ference from magnesium is not significant until the magnesium-to-sodium ratio exceeds 100, arare occurrence.

c. Minimum detectabie concentration: Better ftame photom-eters or atomie absorption spectrometers operating in the emis-sion mode can be used to determine sodium levels approximat-ing 5 j.Lg/L.

a. Fiame photometer (either direct-reading ar intemal-stan-dard type) or atomic absorption spectrometer operating in theftame emission mode.

b. Giassware: Rinse alI glassware with 1 + 15 RN03 fol-lowed by several portions of reagent water (~ 3a).

To minimize sodium contamination, store alI solutions inplastic bottles. Use small containers to reduce the amount of dryelement that may be picked up from the botde walls when thesolution is poured. Shake each eontainer vigorously to washaccumulated salts from walls befare pouring solution.

a. Reagent water: See Section 1080. Use reagent water toprepare ali reagents and calibration standards, and as dilutionwater.

b. Stock sodium soiution: Dissolve 2.542 g NaCI dried at140°C to constant weight and dilute to 1000 mL with water; 1.00mL = 1.00 mg Na.

c. lntermediate sodium soiution: Dilute 10.00 mL stock so-dium solution with water to 100.0 mL; 1.00 mL = 0.10 mg Na

(1.00 mL = 100 JLg Na). Use this intermediate solution toprepare calibration curve in sodium range of 1 to 10 mg/L.

d. Standard sodium solution: Dilute 10.00 mL intermediatesodium solution with water to 100 mL; 1.00 mL = 10.0 JLg Na.Use this solution to prepare calibration curve in sodium range of0.1 to 1.0 mg/L.

a. Pretreatment of polluted water and wastewater samples:Follow the procedures described in Section 3030.

b. Instrument operation: Because of differences betweenmakes and models of instruments, it is impossible to formulatedetailed operating instructions. Follow manufacturer' s recom-mendation for selecting proper photocell and wavelength, ad-justing slit width and sensitivity, appropriate fuel and oxidant gaspressures, and the steps for warm-up, correcting for interferencesand flame background, rinsing of burner, igniting flame, andmeasuring emission intensity.

c. Direct-intensity measurement: Prepare a blank and so-dium calibration standards in stepped amounts in any of thefollowing applicable ranges: O to 1.0, O to 10, ar O to 100mg/L. Determine emission intensity at 589 nm. Aspirate cal-ibration standards and samples enough times to secure areliable average reading for each. Construct a calibrationcurve from the sodium standards. Determine sodium concen-tration of sample from the calibration curve. Where a largenumber of samples must be run routinely, the calibrationcurve provides sufficient accuracy. If greater precision andless bias are desired and time is available, use the bracketingapproach described in <J[ 4d below.

d. Bracketing approach: From the calibration curve, selectand prepare sodium standards that immediately bracket the emis-sion intensity of the sample. Determine emission intensities ofthe bracketing standards (one sodium standard slightly less andthe other slightly greater than the sample) and the sample asnearly simultaneously as possible. Repeat the deterrnination onbracketing standards and sample. Calculate the sodium concen-tration by the equation in <J[ 5b and average the findings.

[(B-A)(S-a) ]

mgNaIL= -----+A D(b - a)

where:

B = mg NaIL in upper bracketing standard,A = mg NaIL in lower bracketing standard,b = emission intensity of upper bracketing standard,a = emission intensity of lower bracketing standard,s = emission intensity of sample, and

D = dilution ratio

mL sample + mL watermL sample

A synthetic sample contammg 19.9 mg Na+/L, 108 mgCaz+/L, 82 mg Mgz+/L, 3.1 mg K+/L, 241 mg Cl-/L, 0.25 mgNOz - -N/L, 1.1 mg N03 - -N/L, 259 mg SO/-/L, and 42.5 mgtotal alkalinity/L (as CaC03) was analyzed in 35 laboratories bythe flame photometric method, with a reI ative standard deviationof 17.3% and a relative error of 4.0%.

WEST,P.W., P. FOLSE& D. MONTGOMERY.1950. Application of flamespectrophotometry to water analysis. Ana/. Chem. 22:667.

COLLINS,C.O. & H. POLKlNHORNE.1952. An investigation of anionicinterference in the determination of sma1J quantities of potassiumand sodium with a new flame photometer. Analyst 77:430.

MELOCHE,V.W. 1956. Flame photometry. Ana/. Chem. 28:1844.BURRIEL-MARTI,F. & J. RAMlREZ-MuNOZ.1957. F1ame Photometry: A

Manual ofMethods and Applications. D. Van Nostrand Co., Prince-ton, NJ.

DEAN,J.A. 1960. Flame Photometry. McOraw-HiJI Publishing Co., NewYork, N.Y.

URE,A.M. & R.L. MITCHELL.1975. Lithium, sodium, potassium, rubid-ium, and cesium. ln J.A. Dean & T.e. Rains, eds. F1ame Emissionand Atomic Absorption Spectrometry. Dekker, New York, N.Y.

THOMPSON,K.C. & REYNOLOS,RJ. 1978. Atomic Absorption, Fluores-cence, and Flame Spectroscopy-A Practical Approach, 2nd ed.John Wiley & Sons, New York, N.Y.

WlLLARD,H.H., L.L. MERRIT,JR., J.A. DEAN& F.A. SETTLE,JR. 1981.Instrumental Methods of Analysis, 6th ed. Wadsworth PublishingCo., Belmont, Calif.

AMERICANSOClETYFORTESTINGANDMATERIALS.1988. Method D 1428-82: Standard test methods for sodium and potassium in water andwater-formed deposits by flame photometry. Annual Book ofASTM Standards, Vol. 11.01. American Soc. Testing & Materials,Philadelphia, Pa.

Strontium (Sr) is the fourth element in Group lIA of theperiodic table; it has an atomic number of 38, an atomic weightof 87.62, and a valence of 2. The average abundance of Sr in theearth's crust is 384 ppm; in soi1sSr ranges from 3.6 to 160 ppm;in streams it averages 50 J.LglL, and in groundwaters it rangesfrom 0.01 to 10 mg/L. Strontium is found chiefiy in celestite(SrS04) and in strontianite (SrC03). Strontium compounds areused in pigments, pyrotechnics, ceramics, and fiares. 90Sr is afission product of nuclear reactor fuels, and was widely distrib-uted on the earth' s surface as a result of fallout from nuclearweapons testing.

The common aqueous species is Sr2+. The solubility ofstrontium is controlled by carbonate and sulfate. Some com-pounds are toxic by ingestion and inhalation. Although thereis no U.S. EPA drinking water standard MCL for concentra-

* Approved by Standard Methods Committee, 1997.Joint Task Group: 20th Edition-See 3500-AI.

tion of strontium, strontium-90 measurements are requiredwhen the gross beta activity of a water sample is greater than50 pCi/L. The U.S. EPA primary drinking water standardMCL for 90Sr is 8 pCi/L.

A method for determination of 90Sr is found in Section 7500-Sr.

The atomic absorption spectrometric method (3111B) andinductively coupled plasma methods (3120 and 3125) are pre-ferred. The fiame emission photometric method (B) also isavailable for those laboratories that do not have the equipmentneeded for one of the preferred methods.

Polyethylene bottles are preferable for sample storage, al-though borosilicate glass containers aIso may be used. At time ofcollection adjust sample to pH <2 with nitric acid (HN03).

a. Principle: The fiame photometric method can be used forthe determination of strontium in the concentration range prev-alent in natural waters. The strontium emission is measured at awavelength of 460.7 nm, while the background intensity ismeasured at a wavelength of 466 nm. The difference in readingsobtained at these two wavelengths measures the light intensityemitted by strontium.

b. Interference: Emission intensity is a linear function ofstrontium concentration and concentration of other constituents.The standard addition technique distributes the same ionsthroughout the standards and the sample, thereby equalizing theradiation effect of possible interfering substances. A very low pH« 1) could produce an interference, but sample dilution shouldeliminate this interference.

c. Minimum detectable concentration: Strontium levels ofabout 0.2 mglL can be detected by the fiame photometric methodwithout prior sample concentration.

Spectrophotometer, equipped with photomultiplier tube andfiame accessories; or an atomic absorption spectrophotometercapable of operation in fiame emission mode.

a. Stock strontium solution: Dissolve 2.415 g strontium ni-trate, Sr(N03h, dried to constant weight at 140°C, in 1000 mL1% (v/v) HN03; 1.00 mL = 1.00 mg Sr.

b. Standard strontium solution: Dilute 25.00 mL stock stron-tium solution to 1000 mL with water; 1.00 mL = 25.0 J.Lg Sr.Use this solution for preparing Sr standards in the 0.2- to25-mglL range.

c. Nitric acid, HN03, conc.

a. Pretreatment of polluted water and wastewater samples:Select an appropriate procedure from Section 3030.

b. Preparation of strontium standards: Dilute samples, ifnecessary, to contain less than 400 mg Ca or BaIL and less than40 mg SrlL. Add 25.0 mL sample (or a lesser but consistentvolume to keep all standards in the linear range of the instru-ment) to 25.0 mL of each of a series of four or more strontiumstandards containing from O mglL to a concentration exceedingthat of the sample. For most natural waters O, 2.0, 5.0, and 10.0mg Sr/L standards are sufficient. A broader range curve might bepreferable for brines. Dilute the brine sufficiently to eliminateburner splatter and clogging.

c. Concentration of low-level strontium samples: Concentratesamples containing less than 2 mg SrlL. Polluted water or

wastewater samples can be concentrated during digestion bystarting with a larger volume (see Section 3030D). For othersamples, add 3 to 5 drops cone HN03 to 250 rnL sample andevaporate to about 25 rnL. Cool and make up to 50.0 rnL withdistilled water. Proceed as in CJ[ b. The HN03 concentration in thesample prepared for atomization can approach 0.4 rnL/50 rnLwithout producing interference.

d. Flame photometric measurement: Measure emission inten-sity of prepared samples (standards plus sample) at wavelengthsof 460.7 and 466 nm. Follow manufacturer's instructions forcorrect instrument operation. Use a fuel-rich nitrous oxide-acet-ylene flame, if possible.

a. Using a calculator or computer with linear regression ca-pability, enter the net intensity (reading at 460.7 nm minusreading at 466 nm) versus concentration added to the sample andsolve the equation for zero emissions. The negative of thisnumber multiplied by any dilution factor is the sample concen-tration.

b. Plot net intensity (reading at 460.7 nm minus reading at466 nm) against strontium concentration added to the sample.Because the plot forms a straight calibration line that intersectsthe ordinate, strontium concentration can be calculated from theequation:

A-B DmgSrlL = -- x-C E

where:A = sample emission-intensity reading of sample plus O mglL at

460.7 nm,B = background radiation reading at 466 nm, andC = slope of calibration line.

Use the ratio D/E only when E rnL of sample are concentratedto a final volume D rnL (typically 50 rnL).

c. Graphical method: Strontium concentration also can beevaluated by the graphical method illustrated in Figure 3500-Sr: 1. Plot net intensity against strontium concentration addedto sample. If the line intersects the ordinate at Yemissions, thestrontium concentration is where the abscissa value of thepoint on the calibration line has an ordinate value of 2Yemissions due to the two-fold dilution with standards (ifsample and standards are mixed in equal volumes). Thecalibration line in the example intersects the ordinate at 12.Thus, Y = 12 and 2Y = 24. The strontium concentration ofthe sample is the abscissa value of the point on the calibrationline having an ordinate value of 24. In the example, thestrontium concentration is 9.0 mg/L.

d. Report strontium concentrations below 10 mg/L to the nearest0.1 mg/L and above 10 mg/L to the nearest whole number.

.q 30(/)cQ).•...cco 20'Vi(/)

'EUJ.•...Q)

z T10

>-

5 10 15Strontium Concentration, mglL

Figure 3500-8r:1. Graphical method of computing strontium concentra-tion.

See Part 1000 and Section 3020 for specific quality controlprocedures to be followed during sample preparation and anal-ysis.

Strontium concentrations in the range 12.0 to 16.0 mglL canbe determined with an accuracy within ± 1 to 2 mglL.

CHOW, T.J. & T.G. ThOMPSON.1955. Flame photometric determinationof strontium in sea water. Ana!. Chem. 27:18.

NrcHoLS,M.S. & D.R. McNALL.1957. Strontium content of Wisconsinmunicipal waters. J. Amer. Water Works Assoe. 49: 1493.

HORR,C.A. 1959. A survey of analytical methods for the determinationof strontium in natural water. V.S. Geol. Surv. Water Supply PapoNo. 1496A.

Tellurium (Te) is the fourth element in Group VIA in the periodictable; it has an atomic number of 52, an atomic weight of 127.60,and valences of 2, 4, and 6. The average abundance of Te in theearth's crust is 0.002 ppm; in soils it is 0.001 to 0.01 ppm; and ingroundwaters it is <0.1 mglL. Tellurium is found in its native stateand as the telluride of gold and other metals. It is used in alloys,catalysts, batteries, and as a coloring agent in glass and ceramics.

Thallium (TI) is the fifth element in Group IIIA in the periodictable; it has an atomic number of 81, an atomic weight of 204.38,and valences of 1 and 3. The average abundance in the earth's crustis 0.07 ppm, and in groundwaters it is <0.1 mgIL. The metal occurschiefly in pyrites. Thallium is used in the production of glasses androdenticides, in photoelectric applications, and in electrodes fordissolved oxygen meters.

Thorium (Th) is the first element in the actinium series of theperiodic table; it has an atomic number of 90, an atomic weightof 232.04, and a valence of 4. The average abundance in theearth's crust is 8.1 ppm; in soils it is 13 ppm; in streams it is 0.1fLglL, and in groundwaters it is <0.1 mglL. Thorium is aradioactive element, with 232Thhaving a half-life of IA X 1010

years. It is widely distributed in the earth, with the principalmineral being monazite. Thorium is used in sun lamps, photo-electric cells, incandescent lighting, and gas mantles.

Tin (Sn) is the fourth element in Group IVA in the periodictable; it has an atomic number of 50, an atomic weight of 118.69,and valences of 2 and 4. The average abundance in the earth'scrust is 2.1 ppm; in soils it is 10 ppm; in streams it is 0.1 fLglL,and in groundwaters it is <0.1 mglL. Tin is found mostly in themineral cassiterite (Sn02), in association with granitic rocks. Tinis used in reducing agents, solder, bronze, pewter, and coatingsfor various metais.

The common aqueous species are Sn4+, Sn(OH)4,SnO(OHh, and SnO(OHh -. Tin is adsorbed to suspended

The common aqueous species is TeO/-. The metal and itscompounds are toxic by inhalation.

Perform analyses by the electrothermal atomic absorptionmethod (3113B). The inductively coupled plasma mass spectro-metric method (3125) also may be applied successfully in mostcases (with lower detection limits), even though tellurium is notspecifically listed as an analyte in the method.

The common aqueous species is TI+. It is nonessential forplants and animais. Compounds of thallium are toxic on contactwith moisture, and by inhalation. The D.S. EPA primary drink-ing water standard MCL is 2 fLglL.

For analysis,use one of the atomicabsorptionspectrometricmethods(311lB ar 3113B), or one of the inductivelycoupled plasma methods(3120 or 3125), depending upon sensitivityrequirements.

The aqueous chemistry of thorium is controlled by the Th4+ion, which forms a set of complex species with hydroxides.Thorium's radioactive decay isotopes are dangerous when in-haled or ingested as thorium dust particles.

Either of the f1ame atomic absorption spectrometric methods(3111D or E) may be used for analysis. The inductively coupledplasma mass spectrometric method (3125) also may be appliedsuccessfully in most cases (with lower detection limits), eventhough thorium is not specifica1lylisted as an analyte in the method.

solids, sulfides, and hydroxides. Tin can be methylated insediments. Tributyl tin undergoes biodegradation quickly.Organo-tin compounds are toxic. Tin is considered nonessen-tial for plants and animaIs.

Either the f1ame atomic absorption method (3111B) or theelectrothermal atomic absorption method (3113B) may be usedsuccessfully for analyses, depending upon the sensitivity desired.The inductively coupled plasma/mass spectrometric method(3125) also may be applied successfully in most cases (withlower detection limits), even though tin is not specifically listedas an analyte in the method.

Titanium (Ti) is the first element in Group IVB in the periodictable; it has an atomic number of 22, an atomic weight of 47.88, andvalences of 2, 3, and 4. The average abundance of Ti in the earth'scrust is 0.6%; in soils it is 1700 to 6600 ppm; in streams it is 3 /-Lg/L,and in groundwaters it is <0.1 mg/L. The element is commonlyassociated with iron mineraIs. Titanium is used in alloys for aircrafi,marine, and food-handling equipment. Compounds of the metal areused in pigments and as a reducing agent.

Uranium (U) is the third element in the actinide series of theperiodic table; it has an atomic number of 92, an atomic weightof 238.04, and valences of 3, 4, and 6. The average abundance ofU in the earth's crust is 2.3 ppm, and in soils it is 1.8 ppm.Concentrations of uranium in drinking waters usually are ex-pressed in terms of picocuries per liter, but that is now beingreplaced by Becquerel per liter (Bq/L). The approximate con-version factor, assuming equilibrium between 234Uand 238U,isI /-Lguranium equals 0.67 pCi. The mean concentration ofuranium in drinking water is 1.8 pCi/L. The chief ore is ura-ninite, or pitchblende, uranous uranate [U(U04h). Uranium isknown mainly for its use in the nuclear industry, but has alsobeen used in glass, ceramics, and photography.

Uranium compounds are radioactive and are thereby toxicby inhalation and ingestion. There are three natural radioiso-

Vanadium (V) is the first element in Group VB in the periodictable; it has an atomic number of 23, an atomic weight of 50.94,and valences of2, 3, 4, and 5. The average abundance ofV in theearth's crust is 136 ppm; in soils it ranges from 15 to 110 ppm;in streams it averages about 0.9 /-Lg/L,and in groundwaters it isgenerally <0.1 mg/L. Though relatively rare, vanadium is foundin a variety of mineraIs; most important among these are vana-dinite [Pb5(V04hCI), and patronite (possibly VS4), occurringchiefly in Peru. Vanadium complexes have been noted in coal

* Approved by Standard Methods Cornrnittee, 1997.loint Task Group: 20th Edition-See 3500-AI.

Titanium species are usually insoluble in natural waters, with theTi4+ species being the most common ion when found. Some com-pounds are toxic by ingestion and the pure metal is flarrunable.

Either of the flame atomic absorption spectrometric methods(3111D or E) may be used. The inductively coupled plasma massspectrometric method (3125) also may be applied successfully inmost cases (with lower detection limits), even though titanium isnot specifically Iisted as an analyte in the method.

topes of uranium. Uranium-238 has a half-Iife of 4.5 X109years, represents 99% of uranium's natural abundance,and is not fissionable, but can be used to form plutonium-239,which is fissionable. Uranium-235 has a half-life of 7.1 X 108years, represents 0.75% of uranium's natural abundance, isreadily fissionable, and was the energy source in the originalatomic bombs. Uranium-234 has a half-life of 2.5 X 105 yearsand represents only 0.006% of uranium's natural abundance.

The common forms in natural water are U4+ and U022+. lnnatural waters below pH 5, UO/+ would dominate; in the pHrange of 5 to 10, soluble carbonate complexes predominate.Although there is no U.S. EPA drinking water standard MCLfor uranium, an analysis for uranium is required if the grossalpha activity of a water sample is greater than 15 pCi/L.

Perform analyses by the inductively coupled plasma/massspectrometry method (3125) or by one ofthe methods in 7500-U(for regulatory compliance purposes).

and petroleum deposits. Vanadium is used in steel alloys and asa catalyst in the productian af sulfuric acid and synthetic rubber.

The dominant farm in natural waters is V5+. It is associatedwith organic complexes and is insoluble in reducing enviran-ments. It is considered nonessential for most higher plants andanimaIs, althaugh it may be an essential trace element for somealgae and microorganisms. Laboratary and epidemiological ev-idence suggests that vanadium may play a beneficial role in thepreventian af heart disease. ln water supplies in New Mexico,which has a low incidence af heart disease, vanadium has beenfound in concentrations af 20 to 150 /-Lg/L.ln a state whereincidence of heart disease is high, vanadium was not found inwater supplies. Hawever, vanadium pentoxide dust causes gas-trointestinal and respiratory disturbances. The United Nations

Food and Agriculture Organization recommended maxlmumleveI for irrigation waters is 0.1 mglL.

The atomic absorption spectrometric methods (3111D and E),the electrothermal atomic absorption method (3113B), the in-

ducti vel y coupled plasma methods (3120 and 3125), and gallicacid method (3500- V.B) are suitable for potable water samples.The atomic absorption spectrometric and inductively coupledplasma methods are preferred for polluted samples. The electro-thermal atomic absorption method also may be used successfullywith an appropriate matrix modifier.

a. Principle: The concentration of trace amounts of vanadiumin water is determined by measuring the catalytic effect it exertson the rate of oxidation of gallic acid by persulfate in acidsolution. Under the given conditions of concentrations of reac-tants, temperature, and reaction time, the extent of oxidation ofgallic acid is proportional to the concentration of vanadium.Vanadium is deterrnined by measuring the absorbance of thesample at 415 nm and comparing it with that of standard solu-tions treated identically.

b. Interference: The substances listed in Table 3500- V:I willinterfere in the deterrnination of vanadium if the specified con-centrations are exceeded. This is not a serious problem for Cr6+,Co2+, M06+, N?+, Ag +, and U6+ because the tolerable concen-tration is greater than that commonly encountered in fresh water.However, in some samples the tolerable concentration of Cu2+,Fe2+, and Fe3+ may be exceeded. Because of the high sensitivityof the method, interfering substances in concentrations onlyslightly above tolerance limits can be rendered harmless bydilution.

Traces of Br- and 1- interfere seriously and dilution alonewill not always reduce the concentration below tolerance limits.Mercuric ion may be added to complex these halides and min-imize their interference; however, mercuric ion itself inteIferes ifin excesso Adding 350 /Lg mercuric nitrate, Hg(N03h, per sam-pIe permits determination of vanadium in the presence of up to100 mg Cl-IL, 250 /Lg Br -IL, and 250 /Lg I-IL. Dilute samplescontaining high concentrations of these ions to concentrationsbelow the values given above and add Hg(N03)2'

TABLE 3500-V:I. CONCENTRATION AT WH1CH VARIOUS IONS INTERFERE IN

THE DETERMINATIDN DF V ANADlUM

ConcentrationmglL

1.01.0

0.05

0.30.50.13.02.03.00.1

100.00.001

Cr6+

Co2+

Cu2+Fe2+Fe3+

Mo6+Ni2+

A<>+'"U6+

Br-CI-1-

c. Minimum detectable concentration: 0.025 /Lg V in approx-imately 13 mL final volume or approximately 2 /Lg VIL.

a. Water bath, eapable of being operated at 25 ::t: 0.5°e.b. Colorimetric equipment: One of the following is required:I) Spectrophotometer, for measurements at 415 nm, with a

light path of 1 to 5 em.2) Fi/ter photometer, providing a light path of I to 5 em and

equipped with a violet filter with maximum transmittanee near415 nm.

Use reagent water (see Section IOS0) in preparation of re-agents, for dilutions, and as blanks.

a. Stock vanadium solution: Dissolve 229.6 mg ammoniummetavanadate, NH4 V03, in a volumetrie flask eontaining ap-proximately SOOmL water and 15 mL I + 1 nitric aeid (HN03).Dilute to 1000 rnL; 1.00 mL = 100 /Lg V.

b. Intermediate vanadium solution: Dilute 1.00 mL stoekvanadium solution with water to 100 mL; 1.00 mL = 1.00/LgV.

C. Standard vanadium solution: Dilute 1.00 rnL intermedi-ate vanadium solution with water to 100 mL; 1.00 mL =omo /Lg V.

d. Mercuric nitrate solution: Dissolve 350 mgHg(N03h . H20 in 1000 mL water.

e. Ammonium persulfate-phosphoric acid reagent: Dissolve2.5 g (NH4hS20S in 25 mL water. Bring just to a boil, removefrom heat, and add 25 mL cone H3P04. Let stand approximately24 h before use. Diseard after 4S h.f Gallic acid solution: Dissolve 2 g H6C70S in 100 mL

warm water, heat to a temperature just below boiling, andfilter through filter paper. * Prepare a fresh solution for eachset of samples.

a. Preparation of standards and sample: Prepare both blankand suffieient standards by diluting 0- to S.O-rnL portions (O toO.OS /Lg V) of standard vanadium soJution to 10 rnL with water.Pipet sample (10.00 mL maximum) containing less than O.OS /LgV into a suitable eontainer and adjust volume to 10.0 mL withwater. Filter eolored or turbid samples. Add 1.0 mL Hg(N03h

solution to each blank, standard, and sample. Place containers ina water bath regulated to 25 ::'::0.5°C and allow 30 to 45 min forsamples to come to the bath temperature.

b. Colar development and measurement: Add 1.0 mL ammo-nium persulfate-phosphoric acid reagent (temperature equili-brated), swirl to mix thoroughly, and retum to water bath. Add1.0 mL gallic acid solution (temperature equilibrated), swirl tomix thoroughly, and retum to water bath. Add gallic acid tosuccessive samples at intervals of 30 s or longer to permitaccurate control of reaction time. Exactly 60 min after addinggallic acid, remove sample from water bath and measure its absor-bance at 415 nm, using water as a reference. Subtract absorbance ofblank from absorbance of each standard and sample. Construct acalibration curve by plotting absorbance values of standards versusmicrograms vanadium. Determine amount of vanadium in a sampleby referring to the corresponding absorbance on the calibrationcurve. Prepare a calibration curve with each set of samples.

Zinc (Zn) is the first element in Group IIB in the periodictable; it has an atomic number of 30, an atomic weight of 65.38,and a valence of 2. The average abundance of Zn in the earth'scrust is 76 ppm; in soils it is 25 to 68 ppm; in streams it is 20/Lg/L, and in groundwaters it is <0.1 mglL. The solubility ofzinc is controlled in natural waters by adsorption on mineralsurfaces, carbonate equilibrium, and organic complexes. Zinc isused in a number of alloys such as brass and bronze, and inbatteries, fungicides, and pigments. Zinc is an essential growthelement for plants and animais but at elevated levels it is toxic tosome species of aquatic life. The United Nations Food andAgriculture Organization recommended levei for zinc in irriga-tion waters is 2 mglL. The U.S. EPA secondary drinking water

p.,gV (in 13mLfinal volume)mg VIL = ----------

originalsamplevolume,mL

In a synthetic sample containing 6 /Lg VIL, 40 /Lg AslL, 250/Lg BelL, 240 /Lg BIL, and 20 /Lg SelL in distilled water,vanadium was measured in 22 laboratories with a relative stan-dard deviation of 20% and no relative error.

FISHMAN,M.J. & M.V. SKOUGSTAD.1964. Catalytic determinationofvanadiumin water.Anal. Chem. 36:1643.

standard MCL is 5 mglL. Concentrations above 5 mglL cancause a bitter astringent taste and an opalescence in alkalinewaters. Zinc most commonly enters the domestic water supplyfrom deterioration of galvanized iron and dezincification ofbrass. In such cases lead and cadmium also may be presentbecause they are impurities of the zinc used in galvanizing. Zincin water also may result from industrial waste pollution.

2. Selection of Method

The atomic absorption spectrometric methods (311lB and C)and inductively coupled plasma methods (3120 and 3125) arepreferred. The zincon method (B), suitable for analysis of bothpotable and polluted waters, may be used if instrumentation forthe preferred methods is not available.

* Approvedby StandardMethodsCommittee,1997.JointTaskGroup:20thEdition-See3500-AI. See Section 301OB.2 for sample handling and storage.

a. Principie: Zinc forms a blue complex with 2-carboxy-2'-hydroxy-5'-sulfoformazyl benzene (zincon) in a solution buff-ered to pH 9.0. Other heavy metais likewise form colored com-plexes with zincon. Cyanide is added to complex zinc and heavymetais. Cyclohexanone is added to free zinc selectively from itscyanide complex so that it can be complexed with zincon to forma blue color. Sodium ascorbate reduces manganese interference.The developed color is stable except in the presence of copper(see table below).

b. Interferences: The following ions interfere at concentra-tions exceeding those listed:

lon mg/L lon mg/L

Cd2+ 1 C~+ 10AI3+ 5 Ni2+ 20Mn2+ 5 Cu2+ 30Fe3+ 7 Co2+ 30Fe2+ 9 Cr04

2- 50

c. Minimum detectable concentration: 0.02 mg ZnIL.

a. Colorimetric equipment: One of the following is required:1) Spectrophotometer, for measurements at 620 nm, provid-

ing a light path of 1 cm 01'10nger.2) Filter photometer, providing a light path of 1 cm 01'10nger

and equipped with a red fi1ter having maximum transmittancenear 620 nm. Deviation from Beer's Law occurs when the filterband pass exceeds 20 nm.

b. Graduated cylinders, 50-mL, with ground-glass stoppers,C1ass B or better.

c. Erlenmeyer flasks, 50-mL.d. Filtration apparatus: 0.45-fLm filters and filter ho1ders.

a. Metal-free water: See Section 3111B.3c. Use water forrinsing apparatus and preparing solutions and dilutions.

b. Stock zinc solution: Dissolve 1000 mg (1.000 g) zinc metalin 10 mL 1 + 1 HN03. Dilute and boil to expe1 oxides ofnitrogen. Dilute to 1000 mL; 1.00 mL = 1.00 mg Zn.

c. Standard zinc solution: Dilute 10.00 mL stock zinc solutionto 1000 mL; 1.00 mL = 10.00 fLg Zn.

d. Sodium ascorbate, fine granular powder, USP.e. Potassium cyanide solution: Dissolve 1.00 g KCN in ap-

proximately 50 mL water and dilute to 100 mL. CAUTlON:Potassium cyanide is a deadly poison. Avoid skin contact orinhalation of vapors. Do not pipet by mouth or bring in contactwith acids.f Buffer solution, pH 9.0: Dissolve 8.4 g NaOH pellets in

about 500 mL water. Add 31.0 g H3B03 and swirl 01' stir todissolve. Di1ute to 1000 mL with water and mix thorough1y.

g. Zincon reagent: Dissolve 100 mg zincon (2-carboxy-2'-hydroxy-5'-su1foformazy1 benzene) in 100 mL methanol. Be-cause zincon dissolves slow1y, stir and/or 1et stand overnight.

h. Cyclohexanone, purified.i. Hydrochloric acid, HC1, cone and lN.j. Sodium hydroxide, NaOH, 6N and 1N.

a. Preparation of colorimetric standards: Accurate1y deliverO, 0.5, 1.0, 3.0, 5.0, 10.0, and 14.0 mL standard zinc solution toa series of 50-mL graduated mixing cylinders. Di1ute each to20.0 mL to yie1d solutions containing O, 0.25, 0.5, 1.5, 2.5, 5.0,and 7.0 mg ZnIL, respectively. (Lower-range standards may beprepared to extend the quantitation range. Longer optical pathcells can be used. Verify linearity of response in this lowerconcentration range.) Add the following to each solution insequence, mixing thoroughly after each addition: 0.5 g sodium

ascorbate, 5.0 mL buffer solution, 2.0 mL KCN solution, and 3.0mL zincon solution. Pipet 20.0 mL of the solution into a clean50-mL erlenmeyer f1ask.Reserve remaining solution to zero theinstrument. Add 1.0 mL cyclohexanone to the erlenmeyer f1ask.Swirl for 10 s and note time. Transfer portions of both solutionsto clean sample cells. Use solution without cyclohexanone tozero colorimeter. Read and record absorbance for solution withcyclohexanone after 1 mino The calibration curve does not passthrough zero because of the color enhancement effect of cyclo-hexanone on zincon.

b. Treatment of samples: To determine readily acid-extract-able total zinc, add 1 mL conc HCI to 50 mL sample and mixthoroughly. Filter and adjust to pH 7. To determine dissolvedzinc, filter sample through a 0.45-fLm membrane filter. Adjust topH 7 with 1N NaOH 01' IN HCI if necessary after filtering.

C. Sample analysis: Cool samples to less than 30°C if neces-sary. Analyze 20.0 mL of prepared sample as described in <j[ 4aabove, beginning with "Add the following to each solution ... "If the zinc concentration exceeds 7 mg ZnIL prepare a sampledilution and analyze a 20.0-mL portion.

Read zinc concentration (in milligrams per liter) directly fromthe calibration curve.

A synthetic sample containing 650 fLg ZnIL, 500 fLg AIIL, 50fLg CdlL, 110 fLg CrlL, 470 fLg CulL, 300 fLg FelL, 70 fLg PblL,120 fLg MnIL, and 150 fLg AglL in doubly demineralized waterwas analyzed in a single laboratory. A series of 10 replicatesgave a relative standard deviation of 0.96% and a relative errorof 0.15%. A wastewater sample from an industry in StandardIndustrial Classification (SIC) No. 3333, primary smelting andrefining of zinc, was analyzed by 10 different persons. The meanzinc concentration was 3.36 mg ZnlL and the reIative standarddeviation was 1.7%. The reIative error compared to results froman atomic absorption analysis of the same sample was -1.0%.

PLATIE,I.A. & V.M. MARCY.1959.Photometricdeterminationof zincwith zincon.Ana!. Chem. 31:1226.

RusH,R.M. & I.H. YOE.1954.Colorimetricdeterminationof zinc andcopper with 2-carboxy-2'-hydroxy-5'-su1foformazyl-benzene.Ana!. Chem. 26:1345.

MILLER,D.G. 1979.Co1orimetricdeterminationof zinc with zinconandcyclohexanone.J. Water Pollut. Contrai Fed. 51:2402.

PANDE,S.P. 1980.Studyon the determinationof zinc in drinkingwater.J. IWWA XII (3):275.