Method 1639: Determination of Trace
Elements in Ambient Waters by
Stabilized Temperature Graphite
Furnace Atomic Absorption
                            ) Printed on Recycled Paper

Method 1639
Method 1639 was prepared under the direction of William A. TeUiard of the U.S.  Environmental
Protection Agency's (EPA's) Office of Water (OW), Engineering and Analysis Division (BAD).  The
method was prepared under EPA Contract 68-C3-0337 by the DjmCorp Environmental Programs Division
with assistance from Interface, Inc.

The following researchers in marine chemistry contributed to the philosophy behind this  method.  Their
contribution is gratefully acknowledged.                                                       ;

Shier Herman, National Research Council, Ottawa, Ontario, Canada
Nicholas Bloom, Frontier Geosciences, Inc., Seattle, Washington
Paul Boothe and Gary Steinmetz, Texas A&M University, College Station, Texas
Eric Crecelius, Battelle Marine Sciences Laboratory, Sequim, Washington
Russell Regal, University of California at Santa Cruz, California
Gary Gill, Texas A&M University at Galveston, Texas
Carlton Hunt and Dion Lewis, Battelle Ocean Sciences, Duxbiiiry, Massachusetts
Carl Watras, Wisconsin Department of Natural Resources, Botilder Junction, Wisconsin
Herb Windom and Ralph Smith, Skidaway Institute of Oceanography, Savannah, Georgia

In addition, J.T. Creed, L.B. Lobring, T.D. Martin, and J.W. O'Dell of the EPA Office of Research and
Development's  Environmental Monitoring Systems  Laboratory in  Cincinnati, Ohio, are gratefully
acknowledged for developing the analytical procedures described in this method.
This method has been reviewed and approved for publication by the Engineering and Analysis Division
of the U.S. Environmental Protection Agency. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
 Questions concerning this method or its application should be addressed to

 W.A. Telliard
 USEPA Office of Water
 Analytical Methods Staff
 Mail Code 4303
 401 M Street, SW
 Washington, DC 20460
 Phone: 202/260-7120
 Fax:   202/260-7185
                                                                                 January 1996

                                                                                      Method 1639

 This analytical method was designed to support water quality monitoring programs authorized under the
 Clean Water Act Section 304(a) of the Clean Water Act requires EPA to publish water quality criteria
 that reflect the latest scientific knowledge about the physical fate (e.g., concentration and dispersal) of
 pollutants,  and their  effects  on ecological  and human  health,  and biological  community diversity,
 productivity, and stability.

 Section 303 of the Clean Water Act requires states to set a water quality standard for each body of water
 within their boundaries. A state water quality standard consists of a designated use or uses of a waterbody
 or a segment of a waterbody, the water quality criteria that are necessary to protect the designated use or
 uses, and an antidegradation policy. These water quality standards serve two purposes:  (1) they establish
 the water quality goals for a specific waterbody, and (2) they are the basis for establishing water quality-
 based treatment controls and strategies beyond the technology-based controls required by Sections 301(b)
 and 306 of the Clean Water Act.

 In defining water quality standards, the state may use  narrative criteria,  numeric  criteria,  or both.
 However, the 1987 amendments to the Clean Water Act required states to adopt numeric criteria for toxic
 pollutants  (designated in Section 307(a) of the Act) based  on EPA Section 304(a) criteria  or other
 scientific data, when the discharge or presence of those toxic pollutants could reasonably be expected to
 interfere with designated uses.

 In some cases, these water quality criteria are as much as 280 times lower than those achievable using
 existing EPA methods and those required to support technology-based permits. Therefore, EPA developed
 new sampling and analysis methods to specifically address state needs for measuring toxic metals at water
 quality criteria levels,  when such measurements are necessary to protect designated uses in state water
 quality standards.  The latest criteria published  by EPA are those listed in the National Toxics Rule (57
 FR 60848) and the Stay of Federal Water Quality Criteria for Metals (60 FR 22228). These rules include
 water quality criteria for 13 metals, on which the new sampling and analysis methods are based.  Method
 1639 was specifically developed  to provide reliable measurements of six of these metals at EPA WQC
 levels using stabilized temperature graphite furnace atomic absorption techniques.

 In developing these  methods, EPA found that one  of the greatest difficulties in measuring pollutants  at
 these levels was precluding sample contamination during collection, transport, and analysis.  The degree
 of difficulty, however, is highly  dependent on the metal  and site-specific conditions.   This analytical
 method, therefore, is designed to  provide the level of protection necessary to preclude contamination  in
 nearly all situations. It is also designed to provide the procedures necessary to produce reliable results
 at the lowest possible water quality criteria published by EPA. In recognition of the variety of situations
 to which this method may be applied, and in recognition of continuing technological advances, the method
 is performance based. Alternative procedures may be used, so long as those procedures are demonstrated
 to yield reliable results.

 Requests for additional copies should be directed to

 11029 Kenwood Road
 Cincinnati, OH 45242
January 1996

Method 1639
    Note: This method is intended to be performance based, and the laboratory is permitted to omit any;
    step or modify any procedure provided that all performance requirements set forth in this method
    are met. The laboratory is not allowed to omit any quality control analyses. The terms "must,"
    "may,"  and "should" are included throughout this  method  and are intended to illustrate the,
    importance of the procedures in producing verifiable data at water quality criteria levels. The term
    "must" is used to indicate that researchers in trace metals analysis have found certain procedures
    essential in successfully analyzing samples and avoiding contamination; however, these procedures!
    can be modified or omitted if the laboratory can demonstrate that data quality is not affected.
                                                                                     January 1996

                                                                          Method 1639
                             Method  1639

     Determination  of Trace  Eiements in  Ambient Waters
          by Stabilized Temperature Graphite Furnace     -
                             Atomic Absorption
 1.0   Scope and Application

 1.1    This method provides procedures to determine dissolved elements in ambient waters at EPA
       water quality criteria (WQC) levels using stabilized temperature graphite furnace atomic
       absorption (GFAA). It may also be used to determine total recoverable element concentrations
       in these waters. This method was developed by integrating the analytical procedures contained
       in EPA Method 200.9 with the stringent quality control (QC) and sample handling procedures
       necessary to avoid contamination and ensure the validity of analytical results during sampling
       and analysis for metals at EPA WQC levels. This method contains QC procedures that will
       ensure that contamination will be detected when blanks accompanying samples are analyzed.
       This method is accompanied by Method 1669:  Sampling Ambient Water for Determination of
       Trace Metals at EPA Water Quality Criteria Levels (the "Sampling Method").  The Sampling
       Method is necessary to ensure that contamination will not compromise trace metals
       determinations during the sampling process.

 1.2    This method is applicable to the following analytes:
Trivalent Chromium
Chemical Abstract Services
Registry Number (CASRN)
      Table 1 lists the EPA WQC levels, the Method Detection Limit (MDL) for each metal, and the
      Minimum Level (ML) set for each metal in this method.  Instrument operating conditions for
      the applicable elements are listed in Table 3, They are intended as a guide and are typical of a
      system optimized for the element employing commercial instrumentation.  However, actual
      linear working ranges will be dependent on the sample matrix, instrumentation, and selected
      operating conditions.
January 1996

Method 1639


This method is not intended to determine metals at concentrations normally found in treated
and untreated discharges from industrial facilities.  Existing regulations (40 CFR Parts 400-
500) typically limit concentrations in industrial discharges to the mid to high part-per-billion   •
(ppb) range, whereas ambient metals concentrations are normally in the low part-per-trillion   ;
(ppt) to low ppb range.

The ease of contaminating ambient water samples  with the metal(s) of interest and interfering
substances cannot be overemphasized. This method includes suggestions for improvements in
facilities and analytical techniques that should maximize the ability of the laboratory to make
reliable trace metals determinations and minimize contamination. These suggestions are given i
in Section 4.0, "Contamination and Interferences," and ;are based on findings of researchers
performing trace metals analyses (References 1-8). Additional suggestions for improving
existing facilities can be found in EPA's Guidance for Establishing Trace Metals Clean Rooms
in Existing Facilities, which is available from the National Center for Environmental
Publications and Information (NCEPI) at the address listed in the introduction to this

Clean and ultraclean—The terms "clean" and "ultracleaa" have been applied to the techniques
needed to reduce or eliminate contamination in trace metals determinations. These terms are
not used in this method, however, because they lack an exact definition. However, the
information provided in this method is consistent with and is copied from summary guidance
on clean and ultraclean techniques (Reference 9).

This method follows the EPA Environmental Methods Management Council's "Format for
Method Documentation" (Reference 10).

This method is "performance based"; i.e., an alternate procedure or technique may be used, as
long as the performance requirements in the method are met.  Section 9.1.2 gives details  of the
tests and documentation required to support equivalent performance.

For dissolved metal determinations, samples must  be filtered through a 0.45-um capsule filter
at the field site. The filtering procedures are described in the Sampling Method.  Except for
trivalent chromium, the filtered samples may be preserved in the field or transported to the
laboratory for preservation. Procedures for field preservation are detailed hi the Sampling
Method; procedures for laboratory preservation are provided in this method.  To determine
trivalent chromium, a field preparation step, which is described hi the Sampling Method, is
used to isolate the trivalent chromium.

To determine  total recoverable analytes in ambient water samples, a digestion/extraction is
required before analysis when the elements are not hi solution (e.g., aqueous samples that may
contain paniculate and suspended solids).

The sensitivity and limited linear dynamic range (LDR) of GFAA often implies the need to
dilute a sample prior to analysis.  The actual magnitude of the dilution as well as the
cleanliness of the labware used to perform the dilution can dramatically influence the quality
of the"analytical results.  Therefore, sample types  requiring large dilutions (>50:1) should be
analyzed by an another approved test procedure that has a larger LDR  or that is inherently less
sensitive than GFAA.                                                                   !
                                                                                   January 1996

                                                                                    Method 1639
 1.11   This method should be used by analysts experienced in the use of graphite furnace atomic
        absorption spectroscopy, the interpretation of spectral and matrix interferences, and procedures
        for their correction, and only by personnel thoroughly trained in handling and analyzing
        samples to determine metals at EPA WQC levels.  A minimum of six months' experience with
        commercial  instrumentation is recommended.

 1.12   This method is accompanied by a data verification and validation guidance document titled"
        Guidance on the Documentation and Evaluation of Trace Metals Data Collected for  CWA
        Compliance Monitoring.  Data users should state the data quality objectives (DQOs)  required
        for a project before using this method.

 2.0   Summary of Method

 2.1     An aliquot of a well-mixed, homogeneous aqueous sample is accurately measured for sample
        processing.  For total recoverable analysis of an aqueous sample containing undissolved
        material, analytes are first solubilized by gentle refluxing with nitric and hydrochloric acids.
        After cooling, the sample is made up to volume, mixed, and centrifuged or allowed to settle
        overnight prior to analysis.  To determine dissolved analytes in a filtered aqueous sample
        aliquot, the sample is made ready for analysis by the appropriate  addition of nitric acid, and
        then diluted  to a predetermined volume and mixed before analysis.

 2.2     The analytes listed in this method  are determined by stabilized temperature platform graphite
        furnace atomic absorption (STPGFAA). In STPGFAA, the sample and the matrix modifier are
        first pipeted  onto the platform or a device that provides delayed atomization.  The furnace
        chamber is then purged with a continuous flow  of a premixed gas (95% argon/5% hydrogen)
        and the sample is  dried at a relatively low temperature (about 120°C)  to avoid spattering.
        Once dried, the sample is pretreated in a char or ashing step, which is designed to minimize
        the interference effects caused by the concomitant sample matrix. After the char step, the
        furnace is allowed to cool before atomization. The atomization cycle  is characterized by rapid
        heating of the furnace to a temperature in which the metal (analyte) is atomized from the
        pyrolytic graphite  surface into a  stopped gas flow atmosphere of argon containing 5%
        hydrogen.  (Only selenium is determined in an atmosphere of high-purity argon.) The
        resulting atomic cloud absorbs the element-specific atomic emission produced by a hollow
        cathode lamp (HCL) or an electrodeless discharge lamp (EDL). Following analysis, the
        furnace is subjected to a cleanout period of high temperature and  continuous argon flow.
        Because the resulting absorbance usually has a nonspecific component associated with the
        actual analyte absorbance, an instrumental background correction device is required to subtract
        from the total signal the component that is nonspecific to the analyte.  In the absence of
        interferences, the background corrected absorbance is  directly related to the concentration of
        the analyte.  Interferences relating to  STPGFAA (Section 4.0) must be recognized and
        corrected.  Suppressions or enhancements of instrument response caused by the sample matrix
        must be corrected  by  the method of standard addition (Section 12.5).

3.0    Definitions
Apparatus—Throughout this method, the sample containers, sampling devices, instrumentation,
and all other materials and devices used in sample collection, sample processing, and sample
analysis activities will be referred to collectively as the Apparatus.
January 1996

Method 1639
3.2    Other definitions of terms are given in the glossary (Section 18) at the end of this method.

4.0   Contamination and  Interferences

4.1    Preventing contamination of ambient water samples during the sampling and analytical process
       constitutes one of the greatest difficulties encountered with trace metals determinations. Over
       the last two decades, marine chemists have recognized that much of the historical data
       regarding the concentrations of dissolved trace metals in seawater are erroneously high because
       the concentrations reflect contamination from sampling and analysis rather than ambient levels.
       More recently, historical trace metals data collected from freshwater rivers and streams have
       been shown to be similarly biased because of contamination during sampling and analysis
       (Reference 15). Therefore, it is imperative that extreme care be taken to avoid contamination
       when collecting and analyzing ambient water samples for trace metals.

4.2    Samples may become contaminated by numerous routes. Potential sources of trace metals   ;
       contamination during sampling include metallic or metal-containing labware (e.g., talc gloves
       that contain high levels of zinc), containers, sampling equipment, reagents, and reagent water;
       improperly cleaned and stored equipment, labware, and reagents; and atmospheric inputs such
       as dirt and dust. Even human contact can be a source of trace metals contamination.  For
       example, it has been demonstrated that dental work (e.g., mercury amalgam fillings) hi the
       mouths of laboratory personnel can contaminate samples that are directly exposed to exhalation
       (Reference 3).

4.3    Contamination Control

       4.3.1    Philosophy—The philosophy behind contaminaition control is to ensure that any object
               or substance that contacts the sample is metal free and free from any material that may
               contain metals.

      The integrity of the results cannot be compromised by contamination of
                      samples.  Requirements and suggestions for control of sample contamination
                      are given in this method and in the Sampling Method.

      Substances in a sample cannot be allowed to contaminate the laboratory work
                      area or instrumentation used for trace metals measurements.  Requirements and
                      suggestions for protecting the laboratory are given hi this method.

      While contamination control is essential, personnel health and safety remain
                      the highest priority.  Requirements and suggestions for personnel safety are
                      given in Section 5 of this method and hi the Sampling Method.

        4.3.2   Avoid contamination—The best way to control contamination is to completely avoid
               exposure of the sample to contamination in the first place. Avoiding exposure means
               performing operations hi an area known or thought to be free from contamination.
               Two of the most important factors hi avoiding or reducing sample contamination are
               (1) an awareness of potential sources of contamination and (2) strict attention to work
               being done.  Therefore it is imperative that the procedures described hi this method be
               carried out by well-trained, experienced personnel.
                                                                                  January 1996

                                                                                     Method 1639
        4.3.3   Use a clean environment—The ideal environment for processing samples is a class 100
                clean room (Section 6.1.1). If a clean room is not available, all sample preparation
                must be performed in a class 100 clean bench or a nonmetal glove box fed by particle-
                free air or nitrogen. Digestions must be performed in a nonmetal fume hood, ideally
                situated in the clean room.

        4.3.4   Minimize exposure—The Apparatus that will contact samples, blanks, or standard -
                solutions must only be opened or exposed in a clean room, clean bench, or glove box
                so that exposure to an uncontrolled atmosphere is minimized.  When not being used,
                the Apparatus should be covered with clean plastic wrap, stored in the clean bench or
                in a plastic box  or glove box, or bagged in clean zip-type bags. Minimizing the time
                between cleaning and use will also nrinimize contamination.

        4.3.5   Clean work surfaces—Before processing a given batch of samples, all work surfaces in
                the hood, clean bench, or glove box in which the samples will be processed should be
                cleaned by wiping with a lint-free cloth or wipe soaked with reagent water.

        4.3.6   Wear gloves—Sampling personnel must wear clean, nontalc gloves (Section 6.9.7)
                during all operations involving handling of the Apparatus, samples, and blanks. Only
                clean gloves may touch the Apparatus.  If another object or substance is touched, the
                glove(s) must be changed before again handling the Apparatus. If it is even  suspected
                that gloves have become contaminated, work must be halted, the contaminated gloves
                removed, and a new pair of clean gloves put on. Wearing multiple layers of clean
                gloves will allow the old pair to be quickly stripped with minimal disruption to the
                work activity.

        4.3.7   Use metal-free Apparatus—All Apparatus used for metals determinations  at ambient
                WQC levels must be nonmetallic and/or free of material that may contain metals.

        Construction materials—Only  the following materials should come in contact
                       with samples: fluoropolymer  (PEP, PTFE), conventional or linear
                      polyethylene, polycarbonate, polypropylene, polysulfone, or ultrapure quartz.
                      PTFE is less desirable than FEP because the sintered material in PTFE may
                      contain contaminates and is susceptible to serious memory contamination
                       (Reference 6). Fluoropolymer or glass containers should be used for samples
                      that will be analyzed for mercury because mercury vapors can diffuse in or out
                      of the other materials resulting either in contamination  or low-biased  results
                      (Reference 3). All materials, regardless of construction, that will directly or
                      indirectly contact the sample must be cleaned using the procedures described
                      in Section 11  and must be known to be clean and metal free before

      The following materials have been found to contain trace metals and must not
                      be used to hold liquids that come in contact with the sample or must  not
                      contact the sample itself, unless these materials have been shown to be free of
                      the metals of interest at the desired level: Pyrex,  Kimax, methacrylate,
                      polyyinylchloride, nylon, and Vycor (Reference 6).  In  addition, highly colored
                      plastics, paper cap liners, pigments used to mark increments  on plastics, and
                      rubber all contain trace levels  of metals and must be avoided (Reference 12).
January 1996

Method 1639
      Serialization—It is recommended that serial numbers be indelibly marked or
                      etched on each piece of Apparatus so that contamination can be traced, and
                      logbooks should be maintained to track the sample from the container through
                      the labware to injection into the instrument.  It may  be useful to dedicate
                      separate sets of labware to different sample types; e.g., receiving waters vs.
                      effluents.  However, the Apparatus used for processing blanks and standards
                      must be mixed with the Apparatus used to process samples so that
                      contamination of all labware can be detected.

      The laboratory or cleaning facility is responsible for cleaning the Apparatus
                      used by the sampling team.  If there are any indications that the Apparatus is  .
                      not clean when received by the sampling team (e.g., ripped storage bags), an
                      assessment of the likelihood of contamination must be made.  Sampling must
                      not proceed if it is possible that the Apparatus is contaminated. If the
                      Apparatus is contaminated, it must be returned to the laboratory or cleaning
                      facility for proper cleaning before any sampling activity resumes.

        4.3.8   Avoid sources of contamination—Avoid contamination by being aware of potential
               sources and routes of contamination.

      Contamination by carryover—Contamination may occur when a sample
                      containing low concentrations of metals  is processed immediately after a
                      sample containing relatively high concentrations of these metals.  To reduce
                      carryover, the sample introduction system may be rinsed between samples with
                      dilute acid and reagent water. When an unusually concentrated sample is
                      encountered, it is followed by analysis of a laboratory blank to check for
                      carryover.  For samples containing high levels of metals, it may be necessary
                      to acid clean or replace the connecting tubing or inlet system to ensure that
                      contamination will not affect subsequent measurements.  Samples known or
                      suspected to contain the lowest concentration of metals  should be analyzed
                      first followed  by samples containing higher levels. For instruments containing
                      autosamplers,  the laboratory should keep track of which station is used for a
                      given sample. When an unusually high: concentration of a metal is detected in
                      a sample, the station used for that sample should be cleaned more thoroughly  ,
                      to prevent contamination of subsequent samples, and the results for subsequent
                      samples should be checked for evidence of the metal(s) that occurred in  high
      Contamination by samples—Significant laboratory or instrument contamination
                      may result when untreated effluents, in-process waters, landfill leachates, and
                      other samples containing high concentrations of inorganic substances are     ;
                      processed and analyzed. As stated hi Section 1.0, this method is not intended;
                      for these samples, and samples containing high concentrations should not be
                      permitted into the clean room and laboratory dedicated for processing trace
                      metals samples.

      Contamination by indirect contact—Apparatus that may not dkectly come in
                      contact with the samples may still be a source of contamination.  For example,
                      clean tubing placed in a dirty plastic bag may pick up contamination from the
                                                                                    January 1996

                                                                                      Method 1639
                       bag and then subsequently transfer the contamination to the sample.
                       Therefore, it is imperative that every piece of the Apparatus that is directly or
                       indirectly used in collecting, processing, and analyzing ambient water samples
                       be cleaned as specified in Section 11.

       Contamination by airborne particulate matter—Less obvious substances capable
                       of contaminating samples include airborne particles.  Samples may be
                       contaminated by airborne dust, dirt, particles, or vapors from  unfiltered air
                       supplies; nearby corroded or rusted pipes, wires,  or other fixtures; or metal-
                       containing paint.  Whenever possible, samples  should be processed and
                       analyzed as far as possible from sources of airborne contamination.

        Interferences—Several interference sources may cause inaccuracies in determining trace
        elements by GFAA.  These interferences can be classified into three major subdivisions,
        namely spectral, matrix, and memory.
        4.4.1   Spectral interferences are caused by the resulting absorbance of light by a molecule or
               atom that is not the analyte of interest or emission from black body radiation.

       Spectral interferences caused by an element only occur if there is a spectral
                       overlap between the wavelength of the interfering element and the analyte of
                       interest. Fortunately, this type of interference is relatively uncommon in
                       STPGFAA because of the narrow atomic line widths associated with
                       STPGFAA.  In addition, the use of appropriate furnace temperature programs
                       and high spectral purity lamps as light sources can rninimize the possibility of
                       this type of interference. However, molecular absorbances can span several
                       hundred nanometers, producing broadband spectral interferences.  This type of
                       interference is far more common in STPGFAA.  The use of matrix modifiers,
                       selective volatilization, and background correctors are all attempts to eliminate
                       unwanted nonspecific absorbance. The nonspecific component of the total
                       absorbance can vary considerably from  sample type to sample type. Therefore,
                       the effectiveness of a particular background correction device may vary
                       depending on the actual analyte wavelength used as well as the nature and
                       magnitude of the interference. The background correction device to be used
                       with this method is not specified; however, it must provide an analytical
                       condition that is not subject to the occurring interelement  spectral interferences
                       of palladium on copper, iron on selenium, and aluminum on arsenic.

       Spectral interferences are also caused by the emissions from black body
                       radiation produced during the  atomization furnace cycle.  This black body
                       emission reaches the photomultiplier tube, producing erroneous results.  The
                       magnitude of this interference can be minimized by proper furnace tube
                       alignment and monochromator design. In addition, atomization temperatures
                       that adequately volatilize the analyte of interest without producing unnecessary
                       black body radiation can help  reduce unwanted background emission during

        4.4.2   Matrix interferences are caused by sample components that inhibit the formation of
               free atomic analyte atoms during the atomization cycle.
January 1996

Method 1639
      Matrix interferences can be chemical or physical.  In this method the use of a
                      delayed atomization device that provides  stabilized temperatures is required.
                      These devices provide an environment that is more conducive to the formation
                      of free analyte atoms and thereby nrininiize this type of interference. This
                      type of interference can be detected by analyzing the sample plus a sample
                      aliquot fortified with a known concentration of the analyte.  If the determined
                      concentration of the analyte addition is outside a designated range, a possible
                      matrix effect should be suspected (Section 9.3).

      The use of nitric acid is preferred for GFAA analyses to minimize vapor state
                      anionic chemical interferences; however,  in this method hydrochloric acid is
                      required to maintain stability in solutions containing antimony. When         <
                      hydrochloric acid is used, the chloride ion vapor state interferences must be
                      reduced using an appropriate matrix modifier.  In this method a combination   !
                      modifier of palladium, magnesium nitrate, and a hydrogen (5%)/argon (95%)
                      gas mixture is used for this purpose.  The effects and benefits of using this
                      modifier are discussed in detail in Reference 13 (Section 17.0).  Listed in
                      Section 4.4.3 are some typical observed effects.

       4.4.3  Specific element interferences

              Antimony: Antimony suffers from an interference produced by K2SO4 (Reference 14).
              In the absence of hydrogen in the char cycle (1300°C), K2SO4 produces a relatively
              high (1.2 abs) background absorbance, which can. produce a false signal, even with
              Zeeman background correction.  However, this background level can be dramatically
              reduced (0.1 abs) by the use of a hydrogen/argon gas mixture hi the char step.  This
              reduction in background is strongly influenced by the temperature of the char step.
              Because the actual furnace temperature may vary from instrument to instrument, it
              should be determined on an individual basis.

              Cadmium: The HC1 present from the digestion  procedure can influence the sensitivity
              for Cd.  The use of 20 uL of a 1% HC1 solution  with Pd as a modifier results in a
              80% loss in sensitivity relative to the analyte in a 1% HNO3 solution.  The use of
              Pd/Mg/H-j as a matrix modifier reduces this suppression to less than 10% (Reference

              Selenium:  Iron has been shown to suppress Se response with continuum background
              correction (Reference 14).  In addition, the use of hydrogen as  a purge gas during the
              dry and char steps can cause a suppression in Se response if not purged from the
              furnace before atomization.

        4.4.4  Memory interferences result from analyzing a s;ample containing a high concentration
               of an element (typically a high atomization temperature element) that cannot be
              removed quantitatively in one complete set of furnace steps. The analyte that remains
               in the furnace can produ ce false positive signals on subsequent sample(s).  Therefore,
             "  the analyst should establish the analyte concentration that can be injected into the
               furnace and adequately removed in one complete set of furnace cycles. If this
               concentration is exceeded, the sample should be  diluted and a blank analyzed to ensure
               the memory effect has been eliminated before reanalyzing the diluted sample.
                                                                                   January 1996

 5.0   Safety
 The toxicity or carcinogenicity of reagents used in this method have not been fully established.
 Each chemical should be regarded as a potential health hazard and exposure to these
 compounds should be as low as reasonably achievable.

 5.1.1   Each laboratory is responsible for maintaining a current awareness file of OSHA
        regulations regarding the safe handling of the chemicals specified in this method
        (References 15-18). A reference file of material safety data sheets should also be
        available to all personnel involved in the chemical analysis. It is also suggested that
        the laboratory perform personal hygiene monitoring of each analyst who uses this
        method and make the results available to the analyst. The references and bibliography
        at the end of Reference 18 are particularly comprehensive in dealing with the general
        subject of laboratory safety.

 5.1.2   Concentrated nitric and hydrochloric  acids present various hazards  and are moderately
        toxic and extremely irritating to skin and mucus membranes. These reagents should be
        used in a fume hood whenever possible.  If eyes or skin are touched, they must be
        flushed with a large volume of water. Protective clothing should always be worn
        along with safety glasses or  a shield for eye protection, and proper mixing of these
        reagents must be observed.

The acidification of samples containing reactive materials may result in the release of toxic
gases, such as cyanides or sulfides.  Acidification of samples should be done in  a fume hood.

All personnel handling environmental samples known to contain or to have been in contact
with human waste should be immunized against known disease-causing agents.

The graphite tube during atomization emits intense UV radiation.  Suitable precautions should
be taken to protect personnel from such a hazard.

The use of the argon/hydrogen gas mixture during the dry and char steps may  emit a
considerable amount of HC1 gas.  Therefore,  adequate ventilation is required.
6.0    Apparatus, Equipment,  and Supplies
       Disclaimer:  The mention of trade names or commercial products in this method is for
       illustrative purposes only and does not constitute endorsement or recommendation for
       use by the Environmental Protection Agency.  Equivalent performance may be
       achievable using apparatus and materials other than those suggested here.  The
       laboratory is responsible for demonstrating equivalent performance.

6.1    Facility
       Clean room—Class 100, 200-ft2 minimum, with down-flow, positive-pressure
       ventilation, air-lock entrances, and pass-through doors
January 1996

Method 1639
      Construction materials—Nonmetallic, preferably plastic sheeting attached
                      without metal fasteners.  If painted, paints that do not contain the metal(s) of
                      interest must be used.

      Adhesive mats, for use at entry points to control dust and dirt from shoes

       6.1.2   Fume hoods, nonmetallic, two minimum, with one installed internal to the clean ro"om

       6.1.3   Clean benches, class 100, one installed in the clean room; the other adjacent to the
               analytical instruments) for preparation of samples and standards                      ;

6.2 Graphite furnace atomic absorbance spectrophotometer

       6.2.1   The GFAA spectrometer must be capable of programmed heating of the graphite tube
               and the associated delayed atomization device.  The instrument must be equipped with
               an adequate background correction device capable of removing undesirable nonspecific
               absorbance over the spectral region of interest aad provide an analytical condition not
               subject to the occurrence of interelement spectral overlap interferences.  The furnace
               device must be capable of using an alternate gas supply during specified cycles of the
               analysis.  The capability to record relatively fast. (< 1 s) transient signals and evaluate
               data on a peak area basis is preferred.  In addition, a recirculating refrigeration bath is
               recommended for unproved reproducibility of furnace temperatures.

       6.2.2   Single-element HDLs or single-element EDLs along with the associated power supplies

       6.2.3   Argon gas supply (high-purity grade, 99.99%) for use during the atomization of       .
               selenium, for sheathing the furnace tube when in operation, and during furnace

       6.2.4   Alternate gas mixture (hydrogen 5%/argon 95%) for use as a continuous gas flow
               environment during the dry and char furnace cycles

       6.2.5   Autosampler capable of adding matrix modifier solutions to the furnace, a single
               addition of analyte, and completing methods of standard additions when required

6.3    Analytical balance, with capability to measure to 0.1 mg, to weigh solids and to prepare

6.4    A temperature-adjustable hot plate capable of maintaining a temperature of 95°C

6.5    A centrifuge with guard bowl, electric timer, and brake (optional)

6.6    A gravity convection drying oven with thermostatic control capable of maintaining 105°C (±

6.7    Alkaline detergent—Liquinox®, Alconox®, or equivalent

6.8    pH meter or pH  paper
January 1996

                                                                                     Method 1639
 6.9    Labware—To determine trace levels of elements, contamination and loss are of prime
        consideration. Potential contamination sources include improperly cleaned laboratory
        apparatus and general contamination within the laboratory environment from dust and other
        contaminants. A clean laboratory work area should be designated for trace element sample
        handling. Sample containers can introduce positive and negative errors in determining trace
        elements by contributing contaminants through surface desorption or leaching, and depleting
        element concentrations through adsorption processes.  All labware must be metal free.
        Suitable construction materials are fluoropolymer (FEP, PTFE), conventional or linear
        polyethylene, polycarbonate, and polypropylene.  Fluoropolymer should be used when samples
        are to be analyzed for mercury.  All labware should be cleaned according to the procedure in
        Section 11.4.  Gloves, plastic wrap, storage bags, and filters may all be used new without
        additional cleaning unless results of the equipment blank pinpoint any of these materials as a
        source of contamination. In this case, either an alternate supplier must  be obtained or the
        materials must be cleaned.
        NOTE: Chromic acid must not be used for cleaning glassware.

        6.9.1   Volumetric flasks, graduated cylinders, funnels and centrifuge tubes

        6.9.2   Assorted calibrated pipets

        6.9.3   PTFE (or other suitable material) beakers, 250 mL with PTFE covers

        6.9.4   Narrow-mouth storage bottles, FEP (fluorinated ethylene propylene) with ETFE
               (ethylene tetrafluorethylene) screw closure, 125-mL to 250-mL capacities

        6.9.5   One-piece stem FEP wash bottle with screw closure,  125-mL capacity

        6.9.6   Tongs—For removal of Apparatus from acid baths.  Coated metal tongs may not be

        6.9.7   Gloves—Clean, nontalc polyethylene, latex, or vinyl; various lengths.  Heavy gloves
               should be worn when working in acid baths as baths  will contain hot, strong acids.

        6.9.8   Buckets or basins—5-50 L capacity, for acid soaking of the Apparatus

        6.9.9   Nonmetallic brushes for scrubbing Apparatus

        6.9.10  Storage bags—Clean, zip-type, nonvented, colorless polyethylene (various sizes) for
               storage of Apparatus

        6.9.11  Plastic wrap—Clean, colorless polyethylene to store Apparatus

6.10    Sampling Equipment—The sampling team may contract with the laboratory or a cleaning
        facility that is responsible for cleaning, storing, and shipping  all sampling devices, sample
        bottles, filtration equipment, and all other Apparatus used to collect ambient water samples.
        Before shipping the equipment to the field site, the laboratory or facility must generate an
January 1996

Method 1639
       acceptable equipment blank (Section 9.5.3) to demonstrate that the sampling equipment is free
       from contamination.

       6.10.1  Sampling devices—Before ambient water samples are collected, consideration should
               be given to the type of sample to be collected and the devices to be used (grab,
               surface, or subsurface samplers).  The laboratory or cleaning facility must clean all
               devices used for sample collection.  Various types of samplers are described in the"
               Sampling Method.  Cleaned sampling devices should be stored in polyethylene bags or

       6.10.2  Sample bottles—Fluoropolymer (FEP, PTFE), conventional or linear polyethylene,
               polycarbonate, or polypropylene; 500 mL with Eds. Cleaned sample bottles should be
               filled with 0.1% HC1 (v/v) until use.
        NOTE: If mercury is a target analyte, fluoropolymer or glass bottles must be used.	

        6.10.3 Filtration apparatus

           Filters, Gelman Supor 0.45-um, 15-mm diameter filter capsules
                             (Gelman 12175), or equivalent

           Peristaltic pump—115-V a.c., 12-V d.c., internal battery, variable
                             speed, single-head (Cole-Parmer, portable, "Masterflex L/S," Catalog
                             No. H-07570-10 drive with Quick Load pump head, Catalog No. H-
                             07021-24, or equivalent).

           Tubing for use with peristaltic pump—Styrene/ethylene/butylene/
                             silicone (SEES)  resin, approx 3/8-in i.d. by approximately 3 ft (Cole-
                             Parmer size 18, Catalog No. G-06464-18, or approximately  1/4-in i.d.,
                             Cole-Parmer size 17, Catalog No. G-06464-17, or equivalent).  Tubing
                             is cleaned by soaking in 5-10% HC1 solution for 8-24 h, rinsing with
                             reagent water in a clean bench in a clean room, and drying  in the clean
                             bench by purging with metal-free air or nitrogen. After drying, the    .
                             tubing is double-bagged in clear polyethylene bags, serialized with a
                             unique number, and stored until use.

7.0    Reagents  and  Standards

        Reagents may contain elemental impurities that might affect analytical data.  Only high-purity
        reagents should be used.  If the purity of a reagent is in question, it should be analyzed for
        contamination.  All acids used for this method must be of ultra high-purity grade or
        equivalent  Suitable acids are available from a number of manufacturers.  Redistilled acids
        prepared by subboiling distillation are acceptable.

7.1     Reagents for cleaning Apparatus, sample bottle storage, and sample preservation and

        7.1.1   Nitric acid, concentrated (sp gr  1.41), Seastar or equivalent
January 1996

                                                                                    Method 1639
        7.1.2   Nitric acid (1+1)—Add 500 mL concentrated nitric acid to 400 mL of regent water
               and dilute to 1 L.                     ?

        7.1.3   Nitric acid (1+5)—Add 50 mL concentrated HNO3 to 250 mL reagent water.

        7.1.4   Nitric acid (1+9)—Add 100 mL concentrated nitric acid to 400 mL of reagent water
               and dilute to 1 L.

        7.1.5   Hydrochloric acid, concentrated (sp gr 1.19)

        7.1.6   Hydrochloric acid (1+1)—Add 500 mL concentrated hydrochloric acid to 400 mL of
               reagent water and dilute to 1 L.

        7.1.7   Hydrochloric acid (1+4)—Add 200 mL concentrated hydrochloric acid to 400 mL of
               reagent water and dilute to 1 L.

        7.1.8   Hydrochloric acid (HC1):  IN trace metal grade

        7.1.9   Hydrochloric acid (HC1):  10% wt, trace metal grade

        7.1.10  Hydrochloric acid (HC1):  1% wt, trace metal grade

        7.1.11  Hydrochloric acid (HC1):   0.5% (v/v), trace metal grade

        7.1.12  Hydrochloric acid (HC1):   0.1% (v/v) ultrapure grade

        7.1.13  Tartaric acid (CASRN 87-69-4)

7.2     Reagent water—Reagent water is demonstrably free of the metal(s) of interest and potentially
        interfering substances at  the MDL for that metal listed in Table 1. It is prepared by
        distillation, deionization, reverse osmosis, anodic/cathodic stripping voltammetry, or other
        technique that removes the metal(s) and potential interferent(s).

7.3     Matrix modifier—Dissolve 300 mg palladium (Pd) powder in concentrated HNO3 (1 mL of
        HNO3, adding 0.1 mL of concentrated HC1, if necessary).  Dissolve 200 mg of Mg(NO3)2 in
        reagent water. Pour the two solutions together and dilute to 100 mL with reagent water.

        NOTE: It is recommended that the matrix modifier be analyzed separately to assess
        the contribution of the modifier to the absorbance of calibration and reagent blank
Standard stock solutions—Stock standards may be purchased or prepared from ultra high-
purity grade chemicals (99.99% to 99.999% pure). All compounds must be dried for 1 h at
105°C, unless otherwise specified.  It is recommended that stock solutions be stored in FEP
bottles.  Replace stock standards when succeeding dilutions for preparation of calibration
standards can not be verified.
January 1996

Method 1639
       CAUTION: Many of these chemicals are extremely toxic if inhaled or swallowed
       (Section 5.1). Wash hands thoroughly after handling.   	

       Below are typical stock solution preparation procedures for 1 L quantities, but for the purpose
       of pollution prevention, the analyst is encouraged to prepare smaller quantities when possible.
       Concentrations are calculated based upon the weight of Ihe pure element or upon the weight of;
       the compound multiplied by the fraction of the analyte in the compound.
       From pure element,
                                 Concentration  =
                                                 volume (I)
       From pure compound,
                        -     ^_ .     weight (mg) x gravimetric factor
                        Concentration = —2—\ o/   o—  	.>	
                                                  volume (L)
             gravimetric factor = the weight fraction of the analyte in the compound
       7.4.1   Antimony solution, stock, 1 mL = 1000 ug Sb:  Dissolve 1.000 g of antimony powder,
              weighed accurately to at least four significant figures, in 20.0 mL (1+1) HNO3 and
              10.0 mL concentrated HC1. Add 100 tnL reagent water and 1.50 g tartaric acid.
              Warm solution slightly to effect complete dissolution.  Cool solution and add reagent
              water to volume in a 1-L volumetric flask.

       7.4.2  Cadmium solution, stock, 1 mL = 1000 ug  Cd:  Dissolve 1.000 g Cd metal, acid
              cleaned with (1+9) HNO3, weighed accurately to at least four significant figures, in 50
              mL (1+1) HNO3 with heating to effect dissolution.  Let solution cool and dilute with
              reagent water in a 1-L volumetric flask.

       7.4.3  Chromium solution, stock, 1 mL = 1000 ug Cr:  Dissolve 1.923 g  CrO3 (Cr fraction =
              0.5200), weighed accurately to at least four significant figures, in 120 mL (1+5)
              HNO3. When solution is complete, dilute to volume in a 1-L volumetric flask with
              reagent water.

       7.4.4  Nickel solution, stock, 1 mL = 1000 ug Ni: Dissolve 1.000 g of nickel metal, weighed
              accurately to at least four significant figures, in 20.0 mL hot concentrated HNO3, cool,
              and dilute to volume in a 1-L volumetric flask with reagent water.

       7.4.5  Selenium solution, stock, 1 mL = 1000 pg  Se:  Dissolve 1.405 g SeO2 (Se fraction =
              0.7116), weighed accurately to at least four significant figures, hi 200 mL reagent     ;
              water  and dilute to volume in a 1-L volumetric flask with reagent  water.
January 1996

                                                                                     Method 1639
         Zinc solution, stock 1 mL = 1000 ug Zn: Pickle zinc metal in (1+9) nitric acid to an
         exact weight of 0.100 g. Dissolve in  5 mL (1+1) nitric acid, heating to effect
         solution.  Cool and dilute to 100 mL with reagent water.
 Preparation of calibration standards—Fresh calibration standards should be prepared every 2
 weeks, or as needed. Dilute each of the stock standard solutions to levels appropriate to the
 operating range of the instrument using the appropriate acid diluent (see note). Calibration"
 standards should be prepared at a minimum of three concentrations, one of which must be at
 the ML (Table 1), and another of which must be near the upper end of the linear dynamic
 range.  Calibration standards should be initially verified using a quality control sample
 (Section 7.7).
        NOTE: The appropriate acid diluent for the determination of dissolved elements is 1%
        HNO3. For total recoverable elements in waters, the appropriate acid diluent is 2%
        HNO3 and 1% HCl.  The reason for these different diluents is to match the types of
        acids and the acid concentrations of the samples with the acid present in the standards
        and blanks.
 Blanks—The laboratory should prepare the following types of blanks.  A calibration blank is
 used to establish the analytical calibration curve; the laboratory (method) blank is used to
 assess possible contamination from the sample preparation procedure and to assess spectral
 background; and the rinse blank is used to flush the instrument autosampler uptake system.
 All diluent acids should be made from concentrated acids (Section 7.1) and reagent water
 (Section 7.2). In addition to these blanks, the laboratory may be required to analyze field
 blanks (Section 9.5.2) and equipment blanks (Section 9.5.3).
        Calibration blank—The calibration blank consists of the appropriate acid diluent
        (Section 7.5 note) (HC1/HNO3) in reagent water. The calibration blank should be
        stored in a FEP bottle.

        Laboratory blank—The laboratory blank must contain all the reagents in the same
        volumes as used in processing the samples.  The laboratory blank must be carried
        through the same entire preparation scheme as the samples including digestion, when
        applicable (Section 9.5.1).

        The rinse blank is prepared as needed by adding 1.0 mL of concentrated HNO3 and
        1.0 mL of concentrated HCl to 1 L of reagent water.
Quality control sample (QCS)—The QCS must be obtained from an outside source different
from the standard stock solutions and prepared in the same acid mixture as the calibration
standards (Section 7.5 note). The concentration of the analytes in the QCS solution should be
such that the resulting solution will provide an absorbance reading of approximately 0.1. The
QCS solution should be stored in a FEP bottle and analyzed as needed to meet data quality
needs.  A fresh solution should be prepared quarterly or more frequently as needed.

Ongoing precision and recovery (OPR) sample—The OPR should be prepared hi the same acid
mixture as the calibration standards (Section 7.5 note) by combining method analytes  at
January 1996

Method 1639
       appropriate concentrations.  The OPR must be carried through the same entire preparation
       scheme as the samples including sample digestion, when applicable (Section 9.6).

8.0   Sample Collection, Filtration, Preservation,  and Storage

8.1    Before an aqueous sample is collected, consideration should be given to the type of data
       required, (i.e., dissolved or total recoverable), so that appropriate preservation and pretreatment
       steps can be taken.  The pH of all aqueous samples must be tested immediately before
       aliquoting for processing or direct analysis to ensure the sample has been properly preserved.
       If properly acid-preserved, the sample can be held up to 6 months before analysis.

8.2    Sample collection—Samples are collected as described in the Sampling Method.

8.3    Sample filtration—For dissolved metals, samples and field blanks are filtered through a 0.45-
       um capsule filter at the field site.  Filtering procedures are described in the Sampling Method.
       For the determination of total recoverable elements, samples are not filtered but should be
       preserved according to the procedures in Section 8.4.

8.4    Sample preservation—Preservation of samples and field blanks for both dissolved and total
       recoverable elements may be performed in the field at time of collection or in the laboratory.
       However, to  avoid the hazards of strong acids in the field and transport restrictions, to
       minimize the potential for sample contamination, and to expedite field operations, the sampling
       team may prefer to ship the samples to the laboratory within 2 weeks of collection. Samples
       and field blanks should be preserved at the laboratory immediately  upon receipt. For all
       metals, preservation involves the addition of 10% HNO3 (Section 7.1.3) to bring the sample to
       pH < 2. For samples received at neutral pH, approximately 5 mL of 10% HNO3 per liter will
       be required.

        8.4.1   Wearing clean gloves, remove the cap from the sample bottle, add the volume of
               reagent grade acid that will bring the pH to < 2, and recap  the bottle immediately. If
               the bottle is full, withdraw the necessary volume using a clean pipet and then  add the
               acid. Record the volume withdrawn and the amount of acid used.
        NOTE: Do not dip pH paper or a pH meter into the sample; remove a small aliquot
        with a clean pipet and test the aliquot. When the nature of the sample is either
        unknown or known to be hazardous, acidification should be done in a fume hood
        (Section 5.2).	;

        8.4.2   Store the preserved sample for a minimum of 48 h at 0-4°C to allow the acid to
                completely dissolve the metal(s) adsorbed on the container walls. The sample should
                then verified to be pH < 2 just before withdrawing an aliquot for processing or direct
                analysis. If for some reason such as high alkalinity the sample pH is verified to be
                > 2, more acid must be added and the sample held for 16 h until verified to be pH < 2
                (Section 8.1).

        8.4.3   With each sample set, preserve a method blank and an OPR sample hi the same way
                as the sample(s).
                                                                                   January 1996

                                                                                    Method 1639
         8.4.4   Store sample bottles in polyethylene bags at 0-4°C until analysis.

 8.5     Samples collected to determine trivalent chromium

         8.5.1   Samples to determine trivalent chromium are prepared at the field site by isolating the
                trivalent chromium using a iron hydroxide coprecipitation technique. These
                procedures are described in the Sampling Method.

         8.5.2   The sampling team is also responsible for the preparation of laboratory blanks, OPRs,
                and MS/MSD samples for trivalent chromium as described in the Sampling Method.

 9.0    Quality Assurance/Quality Control
Each laboratory that uses this method is required to operate a formal quality assurance
program (Reference 19).  The minimum requirements of this program consist of an initial
demonstration of laboratory capability, analysis of samples spiked with metals of interest to
evaluate and document data quality, and analysis of standards and blanks as tests of continued
performance.  Laboratory performance is compared with established performance criteria to
determine that results of the analysis meet the performance characteristics of the method.
       The analyst will demonstrate the ability to generate acceptable accuracy and precision
       with this method.  This ability is established as described hi Section 9.2.
        9.1.2   In recognition of advances that are occurring hi analytical technology, the analyst is
               permitted to exercise certain options to eliminate interferences or lower the costs of
               measurements.  These options include alternate digestion, concentration, and cleanup
               procedures, and changes hi instrumentation. Alternate determinative techniques, such
               as substituting a colorimetric technique or changes that degrade method performance,
               are not allowed.  If an analytical technique other than the techniques specified hi the
               method is used, then that technique must have a specificity equal to or better than the
               specificity of the techniques hi the method for the analytes of interest.

      Each time a modification is made to the method, the analyst is required to
                      repeat the procedure hi Section 9.2. If the detection limit of the method will
                      be affected by the change, the laboratory is required to demonstrate that the
                      MDL (40 CFR Part 136, Appendix B) is lower than the MDL for that analyte
                      hi this method, or one-third the regulatory compliance level, whichever is
                      higher.  If calibration will be affected by the change, the analyst must
                      recalibrate the instrument according to Section 10.

      The laboratory is  required to maintain records of modifications made to this
                      method.  These records include the following, at a minimum:

                   The names, titles, addresses, and telephone numbers of the
                                    analyst(s) who performed the analyses and modification, and of
                                    the quality control officer who witnessed and will verify the
                                    analyses and modification
January 1996

Method 1639
                   A listing of metals measured, by name and CAS Registry

                   A narrative stating reasoiti(s)  for the modification(s)

                   Results from all quality control (QC) tests comparing the
                                     modified method with this method, including the following:

                                     (a)     Calibration
                                     (b)     Calibration verification
                                     (c)     Initial precision and recovery (Section 9.2)
                                     (d)     Analysis of blanks
                                     (e)     Accuracy assessment

                   Data that will allow an independent reviewer to validate each
                                     determination by tracing the  instrument output (peak height,
                                     area, or other signal) to the final result.  These data are to
                                     include, where possible,  the following:

                                     (a)     Sample numbers and other identifiers
                                     (b)     Digestion/preparation or extraction dates
                                     (c)     Analysis dates and times
                                     (d)     Analysis sequence/run chronology
                                     (e)     Sample weight or volume
                                     (f)     Volume before extraction/concentration step
                                     (g)     Volume after each extraction/concentration step        ;
                                     (h)     Final volume before  analysis
                                     (i)     Injection volume
                                     (j)     Dilution data, differentiating between dilution of a
                                             sample or extract
                                      (k)     Instrument and operating conditions (make, model,
                                             revision, modifications)                              ,
                                      (1)     Sample introduction  system (auto sampler, flow
                                             injection system, etc.)
                                      (m)    Operating conditions (background corrections,
                                             temperature program, flow rates, etc.)
                                      (n)     Detector (type, operating conditions, etc.)
                                      (o)     Mass spectra, printer tapes, and other recordings of raw
                                      (p)     Quantitation reports, data system outputs, and other
                                             data to link raw data to results reported

        9.1.3   Analyses of blanks are required to demonstrate freedom from contamination.  The
                required types, procedures,  and criteria for analysis of blanks are described in Section

        9.1.4   The laboratory will spike at least 10% of the samples with the metal(s) of interest to
                monitor method performance.  This test is described in Section 9.3 of this method.
                When results of these spikes indicate atypical method performance for samples, an
                                                                                      January 1996

                                                                                      Method 1639
                alternative extraction or cleanup technique must be used to bring method performance
                within acceptable limits.  If method performance for spikes cannot be brought within
                the limits given in this method, the result may not be reported for regulatory
                compliance purposes.

        9.1.5   The laboratory will, on an ongoing basis, demonstrate through calibration verification
                and through analysis of the ongoing precision and recovery aliquot that the analytical
                system is in control.  These procedures are described in Sections 10.7 and 9.6 of this

        9.1.6   The laboratory shall maintain records to define the quality of data that are generated.
                Development of accuracy statements is described in Section 9.3.4.

 9.2    Initial demonstration of laboratory capability
 Method detection limit—To establish the ability to detect the trace metals of interest,
 the analyst shall determine the MDL for each analyte acceding to the procedure in 40
 CFR 136, Appendix B using the apparatus, reagents, and standards that will be used in
 the practice of this method. The laboratory must produce an MDL that is less than or
 equal to the MDL listed in Table 1, or one-third the regulatory compliance limit,
 whichever is greater. MDLs should be determined when a new operator begins work
 or whenever, in the judgment of the analyst, a change in instrument hardware or
 operating conditions would dictate that they be redetermined.

 Initial precision and recovery (DPR)—To establish the ability to generate acceptable
 precision and recovery, the analyst will perform the following operations.  Analysis of four aliquots of reagent water spiked with the metal(s) of interest
        at two to three times the ML (Table 1), according to the procedures in Section
        12.  All digestion, extraction, and concentration steps, and the containers,
        labware, and reagents  that will be used with samples, must be used in this test.  Using results of the set of four analyses, computation of the average percent
        recovery (X) for the metal(s) in each  aliquot and the standard deviation of the
        recovery (s)  for each metal.  For each metal, comparison of s and X with the corresponding limits for initial
        precision and recovery in Table 2. If s and X for all metal(s) meet the
        acceptance criteria, system performance is acceptable and analysis of blanks
        and samples may begin.  If, however, any individual s exceeds the precision
        limit or any individual X falls outside the range for accuracy, system
        performance is unacceptable for that metal.  The problem must be corrected
        and the test repreated (Section

Linear dynamic range (LDR>—The upper limit of the LDR must be established for the
wavelength used for each analyte by determining  the signal responses from a minimum
of six different concentration standards across the range, two of which are close to the
upper limit of the LDR.  Determined LDRs must  be documented and kept on file.  The
linear calibration range that may be used for the analysis of samples should be judged
January 1996

Method 1639
              by the analyst from the resulting data.  The upper LDR limit should be an observed
              signal no more than 10% below the level extrapolated from the four lower standards.
              The LDRs should be verified whenever, in the judgment of the analyst, a change in
              analytical performance caused by either a change hi instrument hardware or operating
              conditions dictates that they be redetermined.
       NOTE: Multiple cleanout furnace cycles may be necessary to fully define or use the
       LDR for certain elements such as chromium.  For this reason the upper limit of the
       linear calibration range may not correspond to the upper LDR limit.    	

       Determined sample analyte concentrations that exceed the upper limit of the linear calibration
       range must either be diluted and reanalyzed with concern for memory effects (Section 4.4) or
       analyzed by another approved method.

       9.2.4   Quality control sample (QCS)—When beginning the use of this method, on a quarterly
               basis or as required to meet data quality needs, the calibration standards and acceptable
               instrument performance must be verified with the preparation and analyses of a QCS
               (Section 7.7).  To verify the  calibration standards, the determined mean concentration
               from three analyses of the QCS must be within ± 10% of the stated QCS value.  If the
               QCS is not within the required limits, an immediate second analysis of the QCS is     '•
               recommended to confirm unacceptable performance.  If the calibration standards and/or
               acceptable instrument performance cannot be verified, the source of the problem must
               be identified and corrected before proceeding with further analyses.

9.3    Method accuracy—To assess the performance of the method on a given sample matrix, the
       laboratory must perform matrix spike (MS) and matrix spike duplicate (MSD) sample analyses
       on 10% of the samples from each site being monitored, or at least one MS sample analysis
       and one MSD sample analysis must  be performed for each sample batch (samples collected
       from the same site at the same time, to a maximum of 10 samples), whichever is more
       frequent  Blanks (e.g., field blanks)  may not be used for MS/MSD analysis.

       9.3.1   Determine the concentration of the MS and MSD as follows:

       If, as in compliance  monitoring, the concentration of a specific metal in the
                      sample is being  checked against a regulatory concentration limit, the spike
                      must be at that limit or at one to five times the background concentration,
                      whichever is greater.

      If the concentration is not being  checked against a regulatory limit, the
                      concentration must be at one to five times the background concentration  or at  |
                      one to five times the ML in Table 1, whichever is greater.

        9.3.2   Assess spike recovery

      Determine the background concentration (B) of each metal by analyzing  one
                      sample aliquot according to the procedure in Section 12.
January 1996

                                                                                      Method 1639
       If necessary, prepare a QC check sample concentrate that will produce the
                       appropriate level (Section 9.3.1) in the sample when the concentrate is added.

       Spike a second sample aliquot with the QC check sample concentrate and
                       analyze it to determine the concentration after spiking (A) of each metal.

       Calculate each percent recovery (P) as 100(A - B)/T, where T is the known
                       true value of the spike.

        9.3.3   Compare the percent recovery (P) for each metal with the corresponding QC
                acceptance criteria found in Table 2. If any individual P falls outside the designated
                range for recovery, that metal has failed the acceptance criteria.

        For a metal that has  failed the acceptance criteria, analyze the ongoing
                       precision and recovery standard (Section 9.6).  If the OPR is within its
                       respective limit for the metal(s) that failed (Table 2), the analytical system is in
                       control and the problem is attributable to the sample matrix.

        For samples that exhibit matrix problems, further isolate the metal(s) from the
                       sample matrix using  dilution, chelation, extraction, concentration, hydride
                       generation, or other means and repeat the accuracy test (Section 9.3.2).

        If the recovery for the metal remains outside the acceptance criteria, the
                       analytical result for that metal in the unspiked sample is suspect and may not
                       be reported for regulatory compliance purposes.

        9.3.4   Assess recovery for samples  and maintain records.

        After the analysis of five samples of a given matrix  type (river water, lake
                       water, etc.) for which the metal(s) pass the tests in Section 9.3.3, compute the
                       average percent recovery (R) and the standard deviation of the percent
                       recovery (SR) for the metal(s).  Express the accuracy assessment as a percent
                       recovery interval from R - 2SR to R + 2SR for each matrix. For example, if
                       R = 90% and SR = 10% for five analyses of river water, the accuracy  interval
                       is expressed as 70-110%.

        Update the accuracy assessment for each metal in each matrix on a regular
                       basis (e.g., after each 5-10 new measurements).

9.4     Precision of MS and MSD
Calculate the relative percent difference (RPD) between the MS and MSD according to
the equation below using the concentrations found in the MS and MSD. Do not use
the recoveries calculated in Section for this calculation because the RPD is
inflated when the background concentration is near the spike concentration.
January 1996

Method 1639
                      Dl  = concentration of the analyte in the MS sample
                      D2  = concentration of the analyte in the MSD sample

       9.4.2   The relative percent difference between the MS aad the MSD must be less than 20
               percent If this criterion is not met, the analytical, system is judged to be out of
               control. Correct the problem and reanalyze all samples in the sample batch associated
               with the MS/MSD that failed the RPD test.

9.5    Blanks—Blanks are analyzed to demonstrate freedom from contamination.

       9.5.1   Laboratory (method) blank

      Prepare a method blank with each sample batch (samples of the same matrix
                      started through the sample preparation process (Section 12) on the same  12-
                      hour shift, to a maximum of 10 samples). To demonstrate freedom from
                      contamination, analyze the blank immediately after analysis of the OPR
                      (Section 9.6).

      If the metal of interest or any potentially interfering substance is found in the
                      blank at a concentration equal to or greater than the MDL (Table 1),  sample
                      analysis must be halted, the source of the contamination determined,  the
                      samples and a new method blank prepared, and the sample batch and fresh
                      method blank reanalyzed.

      Alternatively,  if a sufficient number of blanks (three minimum) are analyzed to
                      characterize the nature of a blank, the average concentration plus two standard
                      deviations must be less than the regulatory compliance level.

      If the result for a single blank remains above the MDL or if the result for the
                      average concentration plus two standard deviations of three or more blanks
                      exceeds the regulatory compliance level, results for samples associated with
                      those blanks may not be reported for regulatory compliance purposes. Stated
                      another way, results for all initial precision and recovery tests (Section 9.2)
                      and all samples must be associated with an uncontaminated method blank
                      before these results may be reported for regulatory compliance purposes.

        9.5.2   Field blank

      Analyze the field blank(s) shipped with each set of samples (samples collected
                      from the same site at the same time, to a maximum of 10 samples).  Analyze
                      the blank immediately before analyzing the samples in the batch.

      If the metal of interest or  any potentially interfering substance is found in the
                      field blank at a concentration equal to or greater than the ML (Table 1), or
                      greater than one-fifth the level  in the associated sample, whichever is greater,
January 1996 \

                                                                                     Method 1639
                      then results for associated samples may be the result of contamination and may
                      not be reported for regulatory cdinpliance purposes.

      Alternatively, if a sufficient number of field blanks (three niinirnum) are
                      analyzed to characterize the nature of the field blank, the average concentration
                      plus two standard deviations must be less than the regulatory compliance level
                      or less than one-half the level in the associated sample, whichever is greater.

      If contamination of the field blanks and associated samples is known or
                      suspected, the laboratory should communicate this to the sampling team so that
                      the source of contamination can be  identified and corrective measures taken
                      before the next sampling event.

        9.5.3   Equipment Blanks—Before using any sampling equipment at a given site,  the
               laboratory or cleaning facility is required to generate equipment blanks to demonstrate
               that the sampling equipment is free from  contamination. Two types of equipment
               blanks are required:  bottle blanks and sampler check blanks.

      Bottle blanks—After undergoing appropriate cleaning procedures (Section
                      11.4), bottles should be subjected to conditions  of use to verify the
                      effectiveness of the cleaning procedures.  A representative set of sample bottles
                      should be filled with reagent water acidified to pH < 2 and allowed to stand
                      for a minimum of 24 h. Ideally, the time that the bottles are allowed  to stand
                      should be as close as possible to the actual time that sample will be in contact
                      with the bottle.  After standing, the  water should be analyzed for any signs of
                      contamination. If any bottle shows  signs of contamination, the problem must
                      be identified, the cleaning procedures corrected or cleaning solutions changed,
                      and all affected bottles recleaned.

      Sampler check blanks—Sampler check blanks are generated in the laboratory
                      or at  the equipment cleaning contractor's facility by processing reagent water
                      through the  sampling devices using  the same procedures that are used in the
                      field  (see Sampling Method).  Therefore, the "clean hands/dirty hands"
                      technique used during field sampling should be followed when preparing
                      sampler check blanks at the  laboratory or cleaning facility.

                  Sampler check blanks are generated by filling a large carboy or
                                    other container with reagent water (Section 7.2) and processing
                                    the reagent water  through the equipment using the  same
                                    procedures that are used in the field (see Sampling Method).
                                    For example, manual grab sampler check blanks  are collected
                                    by directly submerging a sample bottle into the water, filling
                                    the bottle, and capping. Subsurface sampler check blanks  are
                                    collected by immersing the sampler into the water and
                                    pumping water into  a sample container.

                  The sampler check blank must be analyzed using the
                                    procedures given in  this method.  If any metal of interest or
                                    any potentially interfering substance is detected in the blank,
January 1996

Method 1639
                                    the source of contamination/ interference must be identified,
                                    and the problem corrected!. The equipment must be
                                    demonstrated to be free from the metal(s) of interest before the
                                    equipment may be used in the field.

                                    Sampler check blanks must be run on all equipment that will
                                    be used in the field.  If, for example, samples are to be
                                    collected using both a grab sampling device and a  subsurface
                                    sampling device, a sampler check blank must be run on both
                                    pieces of equipment.

9.6    Ongoing precision and recovery

       9.6.1  Prepare an ongoing precision and recovery sample (laboratory fortified method blank)
              identical to the initial precision and recovery aliquots (Section 9.2) with each sample
              batch (samples of the same matrix started through the sample preparation process
              (Section 12) on the same 12-hour shift, to a maximum of 10 samples) by spiking an
              aliquot of reagent water with the metal(s) of interest.

       9.6.2  Analyze the OPR sample before analyzing the method blank and samples from the
              same batch.

       9.6.3  Compute the percent recovery of each metal in the OPR sample.

       9.6.4  For each metal, compare the concentration with die limits for ongoing recovery in
              Table 2.  If all metals meet the acceptance criteria, system performance is acceptable
              and analysis of blanks and samples  may proceed. If, however, any individual recovery
              falls outside of the range given, the analytical processes are not being performed
              properly for that metal.  Correct the problem, prepare the sample batch again, and
              repeat the ongoing precision and recovery test (Section 9.6).

       9.6.5  Add results that pass the specifications in Section 9.6.4 to initial and previous ongoing
              data for  each metal in each matrix.  Update QC charts to form a graphic representation
              of continued laboratory performance. Develop a statement of laboratory accuracy for
              each metal in each matrix type by calculating the average percent recovery (R) and the
              standard deviation of percent recovery (SR). Expiress the accuracy as a recovery
              interval from R - 2SR to R + 2SR.  For example, if R = 95% and SR = 5%, the
              accuracy is 85—105%.

9.7    The specifications contained in this method can be met if the instrument used is calibrated
       properly and then maintained in a calibrated state. A given instrument will provide the most
       reproducible results if dedicated to the settings and conditions required for the analyses of
       metals by this method.

9.8    Depending on specific program requirements, the laboratory may be required to analyze field
       duplicates collected to determine the precision of the sampling technique.  The RPD between
       field duplicates should be less than 20%.  If the RPD of the field duplicates exceeds 20%, the
       laboratory should communicate it to the sampling team so that the source  of error can be
       identified and corrective measures  taken before the next sampling event.
                                                                                    January 1996

                                                                                   Method 1639
 10.0  Calibration  and  Standardization

 10.1    Recommended wavelengths and instrument operating conditions are listed in Table 3.
        However, because of differences between makes and models of spectrophotometers and
        electrothermal furnace devices, the actual instrument conditions selected may vary from those
        listed.                                                          .                   "

 10.2    Before using this method, optimize the instrument operating conditions.  The analyst should
        follow the instructions provided by the manufacturer while using the conditions listed in Table
        3 as a guide.  Of particular importance is the determination of the charring temperature limit
        for each analyte.  This limit is the furnace temperature setting where a loss hi analyte will
        occur before atomization. This limit should be determined by  conducting char temperature
        profiles for each analyte and when necessary, in the matrix of question.  The charring
        temperature selected should rninimize background absorbance while providing some furnace
        temperature variation without loss of analyte.  For routine analytical operation the charring
        temperature is usually set at least 100°C below this limit.  The optimum conditions selected
        should provide the lowest reliable MDLs and be similar to those listed in Table 3.  Once the
        optimum operating conditions are determined, they should be recorded and available for daily

 10.3    Before an initial calibration, determine the linear dynamic range of the analyte (Section 9.2.3)
        using the optimized instrument  operating conditions (Section 10.2).  For all determinations,
        allow an instrument and hollow cathode lamp warm-up period  of not less than 15 minutes.  If
        an EDL is to be used, allow 30 minutes for warm-up.

 10.4    Before daily calibrating  the instrument, inspect the graphite furnace, the sample uptake system,
        and the autosampler injector for any change in the system that  would affect instrument
        performance.  Clean the system and replace the graphite tube and/or platform when needed or

 10.5    After the warm-up period but before calibration, demonstrate instrument stability by analyzing
        a standard solution with a concentration three times the ML a minimum of five times.  The
        resulting relative standard deviation  (RSD) of absorbance signals must be < 5%. If the RSD is.
        > 5%, determine and correct the cause before calibrating the instrument.

 10.6    For initial and daily operation, calibrate the instrument according to the instrument
        manufacturer's recommended procedures using the calibration blank (Section 7.6.1) and
        calibration standards  (Section 7.5) prepared at three or more concentrations, one of which must
        be at the ML (Table  1),  and another that must be near the upper end of the linear dynamic
January 1996

Method 1639
        10.6.1  Calculate the response factor (RF) for each metal in each CAL solution using the
               following equation and the height or area produced by the metal.
                         Rx = height or area of the signal for the metal
                         C  ~ concentration of compound injected (v-g/L)
        10.6.2  For each metal, calculate the mean RF (M), the standard deviation (SD) of the RF, and
               the relative standard deviation (RSD) of the mean, where RSD = 100 x SD/M.

        10.6.3  Linearity-If the RSD of the mean RF for any metal is less than 25% over the
               calibration range, an averaged RF may be used for that analyte.  Otherwise, a
               calibration curve for that metal must be used over the calibration range.

10.7.    Calibration verification—Immediately following calibration, perform an initial calibration
        verification.  Adjustment of the instrument is performed until verification criteria are met.
        Only after these criteria are met may blanks and samples be analyzed.

        10.7.1  Analyze the midpoint calibration standard (Section 10.6).

        10.7.2  Compute the percent recovery of each metal using the average RF or the calibration
               curve obtained in the initial calibration.

        10.7.3  For each metal, compare the recovery with the co:rresponding limit for calibration
               verification hi Table 2. If all metals meet the acceptance criteria, system performance
               is acceptable and analysis of blanks and samples may continue using the response from
               the initial calibration.  If any individual value falls outside the range given, system
               performance is unacceptable for that compound. Locate and correct the problem
               and/or prepare a new calibration check standard and repeat the test (Sections
               10.7.1-10.7.3), or recalibrate the system according to Section 10.6.

        10.7.4  Verify calibration following every 10 samples by analyzing  the midpoint calibration
               standard.  If the recovery does not meet the acceptance criteria specified in Table 2,
               analysis must be halted, the problem corrected, and the instrument recalibrated.  All
               samples after the last acceptable calibration verification must be reanalyzed.

10.8    Analyze a calibration blank following every calibration verification to demonstrate that there is
        no carryover of the analytes of interest and that the analytical system is free from
        contamination.  If the concentration of an analyte in the blank result exceeds the MDL, correct
        the problem, verify the calibration (Section 10.7), and repeat the analysis of the calibration
January 1996

                                                                                     Method 1639
 11.0  Procedures for Cleaning the Apparatus
All sampling equipment, sample containers, and labware should be cleaned in a designated
cleaning area that has been demonstrated to be free of trace element contaminants. Such areas
may include class 100 clean rooms as described by Moody (Reference 20), labware cleaning
areas as described by Patterson and Settle (Reference 6), or clean benches.

Materials, such as gloves (Section 6.9.7), storage bags (Section 6.9.10), and plastic wrap
(Section 6.9.11) may be used new without additional cleaning unless the results of the
equipment blank pinpoint any of these materials as a source of contamination. In this case,
either an alternate supplier must be obtained or the materials must be cleaned.

Cleaning procedures—Proper cleaning of the Apparatus is extremely important, because the
Apparatus may not only contaminate the samples but may also remove the analytes of interest
by adsorption onto the container surface.
        NOTE: If laboratory, field, and equipment blanks (Section 9.5) from Apparatus
        cleaned with fewer cleaning steps than those detailed below show no levels of analytes
        above the MDL, than those cleaning steps that do not eliminate these artifacts may be
        omitted provided all performance criteria outlined in Section 9 are met.

        11.3.1  Bottles, labware, and sampling equipment

             Fill a clean basin (Section 6.9.8) with a sufficient quantity of a 0.5%
                              solution of liquid detergent (Section 6.7), and completely immerse each
                              piece of ware.  Allow to soak in the detergent for at least 30 minutes.

             Using a pair of clean gloves (Section 6.9.7) and clean nonmetallic
                              brushes (Section 6.9.9), thoroughly scrub down all materials with the

             Place the scrubbed materials in a clean basin. Change gloves.

             Thoroughly rinse the inside and outside of each piece with reagent
                              water until there is no sign of detergent residue (e.g., until all soap
                              bubbles disappear).

             Change gloves, immerse the rinsed equipment hi a hot (50-60°C) bath
                              of concentrated reagent grade HNO3 (Section 7.1.1) and allow to soak
                              for at least 2 h.

             After soaking, use clean gloves and tongs to remove the Apparatus and
                              thoroughly rinse with distilled, deionized water (Section 7.2).

             Change gloves and immerse the Apparatus hi a hot (50-60°C) bath of
                              IN trace metal grade HC1 (Section 7.1.8), and allow to soak for at
                              least 48 h.
January 1996

Method 1639
             Thoroughly rinse all equipment and bottles with reagent water.
                              Proceed to Section 11.3.2 for labware and sampling equipment.
                              Proceed to Section 11.3.3 for sample bottles.

        11.3.2  Labware and sampling equipment

             After cleaning, air-dry in a class 100 clean air.bench.

             After drying, wrap each piece of ware/equipment in two layers of
                              polyethylene film.

        11.3.3  Fluoropolymer sample bottles—These bottles should be used if mercury is a target

             After cleaning, fill sample bottles with 0.1% (v/v) ultrapure HC1
                              (Section 7.1.11)  and cap tightly. It may be necessary to use a strap
                              wrench to ensure a tight seal.

             After capping, double-bag each bottle in polyethylene zip-type bags.
                              Store at room temperature until siample collection.

        11.3.4  Bottles, labware, and sampling equipment (polyethylene or material other than

             Apply  the steps  outlined above in Section to all
                              bottles, labware, and sampling equipment.  Proceed to Section
                              for bottles or Section for labware and sampling equipment.

             After cleaning, fill each bottle with 0.1% (v/v) ultrapure HC1 (Section
                              7.1.12). Double-bag each bottle in a polyethylene bag to prevent
                              contamination of the surfaces with dust and dirt. Store at room
                              temperature until sample collection.

             After rinsing labware and sampling equipment, air-dry in a class 100
                              clean air bench.   After drying, wrap each piece of ware/equipment hi
                              two layers of polyethylene film.
        NOTE: Polyethylene bottles cannot be used to collect samples that will be analyzed
        for mercury at trace (e.g., 0.012 ug/L) levels because of the potential of vapors
        diffusing through the polyethylene.                     	

             Polyethylene bags—If polyethylene bags need to be cleaned, clean
                              according to the following procedure:

                   Partially fill with cold, (1+1) HNO3 (Section 7.1.2) and rinse
                                      with distilled deionized water (Section 7.2).
January 1996

                                                                                   Method 1639
           Dry by hanging upside down from a plastic line with a plastic

11.3.5 Silicone tubing, fluoropolymer tubing, and other sampling apparatus—Clean any
       silicone, fluoropolymer, or other tubing used to collect samples by rinsing with 10%
       HC1 (Section 7.1.9) and flushing with water from the site before sample collection.

11.3.6 Extension pole—Submerging the 2-m polyethylene extension pole (used in with the
       optional grab sampling device) in acid solutions  as described above is unpractical
       because of its length.  If such an extension pole  is used, a nonmetallic brush (Section
       6.9.9) should be used to scrub the pole with reagent water and the pole wiped down
       with acids as described hi Section 11.3.4 above.  After cleaning, the pole should be
       wrapped hi polyethylene film.

Storage—Store each piece or assembly of the Apparatus in a clean, single polyethylene zip-
type bag.  If shipment is required, place the bagged apparatus hi a second polyethylene zip-
type bag.

All cleaning solutions and  acid baths should be periodically monitored for accumulation of
metals that could lead to contamination. When levels of metals hi the solutions become too
high, the solutions and baths should be changed and the  old solutions neutralized and
discarded in compliance with state and federal regulations.
 12.0  Procedures for Sample Preparation and Analysis

 12.1    Aqueous sample preparation—dissolved analytes (except trivalent chromium)

        12.1.1  To determine dissolved analytes hi ground and surface waters, pipet an aliquot (> 20
               mL) of the filtered, acid-preserved sample into a clean 50-mL polypropylene centrifuge
               tube. Add an appropriate volume of (1+1) nitric acid to adjust the acid concentration
               of the aliquot to approximate a 1% (v/v) nitric acid solution (e.g., add 0.4 mL (1+1)
               HNO3 to a 20-mL aliquot of sample).  Cap the tube and mix. The sample is now
               ready for analysis.  Allowance for sample dilution should be made in the calculations.

 12.2    Aqueous sample preparation—total recoverable analytes (except trivalent chromium)
        NOTE: To preclude contamination during sample digestion, it may be necessary to
       perform the open-beaker, total-recoverable digestion procedure described in Sections
        12.2.1-12.2.6 in a fume hood that is located in a clean room. An alternate digestion
       procedure is provided in Section 12.2.7; however, this procedure has not undergone
        interlaboratory testing.

        12.2.1 To determine total recoverable analytes hi ambient water samples, transfer a 100-mL
              (± 1 mL) aliquot from a well-mixed, acid-preserved sample to a 250-mL Griffin
              beaker (Section 6.9.3).  If appropriate, a  smaller sample volume may be used.
January 1996

Method 1639
       12.2.2  Add 2 mL (1+1) nitric acid and 1.0 mL of (1+1) hydrochloric acid to the beaker and
               place the beaker on the hot plate for digestion.  The hot plate should be located hi a
               fume hood and previously adjusted to provide evaporation at a temperature of
               approximately but not higher than 85°C. (See the following note.) Cover the beaker
               or take other necessary steps to prevent sample contamination from the fume hood
       NOTE: For proper heating, adjust the temperature control of the hotplate so that an
       uncovered Griffin beaker containing 50 mL of water placed in the center of the hot
       plate can be maintained at a temperature of approximately but not higher than 85°C.
       (Once the beaker is covered with a watch glass, the temperature of the water will rise
       to approximately 95°C.)

       12.2.3 Reduce the volume of the sample aliquot to about 20 mL by gentle heating at 85°C.
              Do not boil.  This step takes about 2 h for a 100--mL aliquot with the rate of
              evaporation rapidly increasing as the sample volume approaches 20 mL.  (A spare
              beaker containing 20 mL of water can be used as; a gauge.)

       12.2.4 Gently reflux the sample for 30 minutes.  (Slight boiling may occur, but vigorous
              boiling must be avoided to prevent loss of the HC1-H2O azeotrope.)

       12.2.5 Allow the beaker to cool.  Quantitatively transfer the sample solution to a 50-mL
              volumetric flask or 50-mL class A stoppered graduated cylinder, make to volume with
              reagent water, stopper, and mix.

       12.2.6 Allow any undissolved material to settle overnight, or centrifuge a portion of the
              prepared sample until clear.  (If, after centrifuging or standing overnight, the sample
              contains suspended solids that would clog the nebulizer, a portion of the sample may
              be filtered to remove the solids before analysis.  However, care should be exercised to
              avoid potential contamination from filtration.)  The sample is now ready for analysis.
              Because the effects of various matrices on the stability of diluted samples cannot be
              characterized, all analyses should be performed as soon as possible after the completed

       12.2.7 Alternate total recoverable digestion procedure

            Open the preserved sample under clean conditions. Add ultrapure
                             nitric and hydrochloric acid at the rate of 10 mL/L and 5 mL/L,
                             respectively. Remove the cap from the original container only long
                             enough to add  each aliquot of acid. The sample container should not
                             be filled to the lip by the addition of the acids. However, only
                             minrmal headspace is needed to avoid leakage during heating.

            Tightly recap the container and shake thoroughly. Place the container
                             in an oven preheated to 85°C.  The container should be placed on an
                             insulating piece of material such as wood rather than directly on the
                             typical metal grating.  After the isamples have reached 85°C, heat for
                             two hours.  (Total time will be  2.5-3 h depending on the sample size).
January 1996

                                                                                      Method 1639
                              Temperature can be monitored using an identical sample container with
                              distilled water and a thermocouple to standardize heating time.

             Allow the sample to cool.  The sample is now ready for analysis.
                              Remove aliquots for analysis under clean conditions.

 12.3   Sample preparation for trivalent chromium

        12.3.1  When the samples are received at the laboratory, shake the vial until the precipitate on
                the filter dissolves.

        12.3.2  After the precipitate dissolves, add 4 mL of reagent water.

        12.3.3  Samples are now ready for analysis according to Section 12.4.

 12.4   Sample Analysis

        12.4.1  Initiate instrument operating configuration. Tune and calibrate the instrument for the
                analytes of interest (Section 10.0).

        12.4.2  Use an autosampler to introduce all solutions into the graphite furnace.  Once the
                standard, sample or QC solution plus the matrix  modifier is injected, the furnace
                controller completes furnace cycles  and cleanout period as programmed.  Analyte
                signals must be integrated and collected as peak  area measurements.  Background
                absorbances, background corrected analyte signals, and determined analyte
                concentrations on all solutions must be able to be displayed on a CRT for immediate
                review by the analyst and be available as hard copy for documentation to be kept on
                file. Hush the autosampler solution uptake system with the rinse blank (Section 7.6.3)
                between each solution injected.

        12.4.3  Analyze samples in the same operational manner used in the calibration routine.

        12.4.4  During the analysis of samples, the laboratory must comply with the required quality
                control described hi Sections 9 and  10.

        12.4.5 For every new or unusual matrix, when practical, it is highly recommended that an
               inductively coupled plasma atomic emission spectrometer be used to screen for high
               element concentration. Information gained from  this screening may be used to prevent
               potential damage to the instrument and to better estimate which elements may require
               .analysis by graphite furnace.

        12.4.6 Determined sample analyte concentrations that are 90% or more of the upper limit of
               calibration must either be diluted with acidified reagent water and reanalyzed with
               concern for memory effects (Section 4.4.4), or determined by another approved test

        12.4.7 When it is necessary to assess an operative matrix interference (e.g., signal reduction
               due to high dissolved solids), the MSA described hi Section 12.5 is recommended.
January 1996

Method 1639
        12.4.8  Report data as directed in Section 13.

12.5    Standard Additions—If the MSA is required, the following procedure is recommended:

        12.5.1  The MSA (Reference 21)  involves preparing new standards in the sample matrix by
               adding known amounts of standard to one or more aliquots of the processed sample
               solution. This technique compensates for a sample constituent that enhances or
               depresses the analyte signal, thus producing a different slope from that of the
               calibration standards. It will not correct for additive interference, which causes a
               baseline shift. The simplest version of this technique is the single-addition method.
               The procedure is as follows:  Two identical aliquots of the sample solution, each of
               volume Vx, are taken.  To the first (labeled A) is added a small volume Vs of a
               standard analyte solution of concentration  Cs. To the second (labeled B) is added the
               same volume Vs of the solvent.  The analytical signals of A and B are measured and
               corrected for nonanalyte signals. The unknown sample concentration Cx is calculated:
               where SA and SB are the analytical signals (corrected for the blank) of solutions A and
               B, respectively. Vs and Cs should be chosen so that SA is roughly twice SB on the
               average.  It is best if Vs is made much less than Vx, and thus Cs is much greater than
               Q, to avoid excess dilution of the sample matrix. If a separation or concentration step
               is used, the additions are best made first and carried through the entire procedure. For
               the results from this technique to be valid, the following limitations must be taken into

               1.      The response vs amount must be linear.

               2.      The chemical form of the analyte  added must respond in the same manner as
                      the analyte in the sample.

               3.      The interference effect must be constant over the working range of concern.

               4.      The signal must be corrected for any additive interference.

13.0  Data Analysis and Calculations

13.1    Sample data should be reported in units of ug/L (parts-per-bilh'on; ppb). Report results at or
        above the ML for metals found in samples and determin<}d in standards. Report all results for
        metals found in blanks, regardless of level.

13.2    Compute the concentration of each analyte in the  sample using the averaged RF determined
        from the calibration data (Section 10.6) according to the following equation:
January 1996

                                                                                   Method 1639
                          •where the terms are defined in Section 10.6.1.

 13.3    For total recoverable aqueous analytes (Sections 12.2.1-12.2.6), multiply solution analyte "
        concentrations by the dilution factor 0.5, when 100-mL aliquot is used to produce the 50-mL
        final solution. If a different aliquot volume other than 100 mL is used for sample preparation,
        adjust the dilution factor accordingly. Also, account for any additional dilution of the prepared
        sample solution needed to complete the determination of analytes exceeding the upper limit of
        the calibration curve.  Do not report data below the determined analyte MDL concentration or
        below an adjusted detection limit reflecting smaller sample aliquots used in processing or
        additional dilutions required to complete the analysis.

 13.4    For data values less than the ML, two significant figures should be used for reporting element
        concentrations. For data values greater than or equal to the ML, three significant figures
        should be used.

 13.5    The QC data obtained during the analyses provide an indication of the quality of the sample
        data and should be provided with the sample results.

 14.0  Method  Performance

 14.1    The MDLs listed in Table 1 and the quality control acceptance criteria listed in Table 2 were
        validated in three laboratories (Reference 22) for all dissolved analytes except trivalent
15.0  Pollution  Prevention

15.1    Pollution prevention encompasses any technique that reduces or eliminates the quantity or
        toxicity of waste at the point of generation. Numerous opportunities for pollution prevention
        exist in laboratory operation. EPA has established a preferred hierarchy of environmental
        management techniques that place pollution prevention as the management option of first
        choice. Whenever feasible, laboratory personnel should use pollution prevention techniques to
        address their waste generation.  When wastes cannot be feasibly reduced at the source, EPA
        recommends recycling as the next best option.  The acids used hi this method should be reused
        as practicable  by purifying by electrochemical techniques.  The only other chemicals used in
        this method are the neat materials used in preparing standards.  These standards are used hi
        extremely small amounts and pose little threat to the environment when managed properly.
        Standards should be prepared hi volumes consistent with laboratory use to niinimize the
        volume of expired standards to be disposed.

15.2    For information about pollution prevention that may be applicable to laboratories and research
        institutions, consult Less is Better:  Laboratory Chemical Management for Waste Reduction,
        available from the American Chemical Society's Department of Government Relations and
        Science Policy, 1155 16th Street NW, Washington DC 20036, 202/872-4477.
January 1996

Method 1639
16.0  Waste Management

16.1   EPA requires that laboratory waste management practices be conducted consistent with all
       applicable rules and regulations. The Agency urges laboratories to protect the ah", water, and
       land by minimizing and controlling all releases  from hoods and bench operations, complying
       with the letter and spirit of any sewer discharge permits ;and regulations, and complying with
       all solid and hazardous waste regulations, particularly the hazardous waste identification rules
       and land disposal restrictions.  For further information on waste management, consult The
       Waste Management Manual for Laboratory Personnel, available from the American Chemical
       Society at the address listed in Section 15.2.
17.0  References

1      Adeloju, S.B.; Bond, A.M. "Influence of Laboratory Environment on the Precision and
       Accuracy of Trace Element Analysis," Anal Chem. 19&5, 57, 1728.

2   '   Berman, S.S.; Yeats, P.A. "Sampling of Seawater for Tra,ce Metals," CRC Reviews in
       Analytical Chemistry 1985,16, 1.

3      Bloom, N.S. "Ultra-Clean Sampling, Storage, and Analytical Strategies for the Accurate
       Determination of Trace Metals in Natural Waters," presented at the  16th Annual EPA
       Conference on the Analysis of Pollutants in the Environment, Norfolk, Virginia, May 5, 1993.

4      Bruland, K.W. 'Trace Elements hi Seawater," Chemical Oceanography 1983, 8, 157.

5      Nriagu, J.O.; Larson, G.; Wong, H.K.T.; Azcue, J.M.  "A Protocol for Minimizing
       Contamination in the Analysis of Trace Metals in Great Lakes Waters," J. Great Lakes
       Research 1993,19, 175.

6      Patterson, C.C.; Settle, D.M.  "Accuracy hi Trace Analysis," hi National Bureau of Standards
       Special Publication 422;  LaFleur, P.D., Ed., U.S.  Government Printing Office: Washington,
       DC, 1976.

7      Fitzgerald, W.F.; Watras, C.J. Science of the Total Environment 1989, 87/88, 223.

8      Gill, G.A.; Fitzgerald, W.F. Deep Sea Res. 1985, 32, 287.

9      Prothro, M.G. "Office of Water Policy and Technical Guidance on Interpretation and
       Implementation of Aquatic Life Metals Criteria," EPA Memorandum to Regional Water
       Management and Environmental  Services Division Directors, Oct. 1, 1993.

10     "Format for Method Documentation," distributed by the 13PA Environmental Monitoring
       Management Council, Washington, DC, Nov. 18,  1993.

11     Windom, H.L; Byrd, J.T.; Smith, R.G., Jr.; Huan, F.  "Inadequacy of NASQAN Data for
       Assessing Metal Trends in the Nation's Rivers," Environ, Sci.  Technol. 1991, 25, 1137.
January 1996

                                                                                 Method 1639
12     Zief, M.; Mitchell, J.W.  "Contamination Control in Trace Metals Analysis," Chemical
       Analysis 1976, 47, Chapter 6.

13     Creed, J.T.; Martin, T.D.; Lobring, L.B.; O'Dell, J.W. Environ. Sci. TechnoL, 1992, 26, 102-

14     Waltz, B.; Schlemmar, G.; Mudakavi, J.R. JAAS, 1988, 3, 695.

15     Carcinogens—Working with Carcinogens; Department of Health, Education, and Welfare,
       Public Health Service, Center for Disease Control, National Institute for Occupational Safety
       and Health, Publication No. 77-206, August 1977.  Available from the National Technical
       Information Service (NTIS) as PB-277256.

16     OSHA Safety and Health Standards, General Industry, (29 CFR 1910); Occupational Safety
       and Health Administration, OSHA 2206 (revised January 1976).

17     Safety in Academic Chemistry Laboratories; American Chemical Society Publication,
       Committee on Chemical Safety, 3rd ed., 1979.

18     Proposed OSHA Safety and Health Standards; Laboratories, Occupational Safety and Health
       Administration, Federal Register, July 24, 1986.

19     Handbook of Analytical Quality Control in Water and Wastewater Laboratories; U.S.
       Environmental Protection Agency. EMSL-Cincinnati, OH, March 1979; EPA-600/4-79-019.

20     Moody, J.R. "NBS Clean Laboratories for Trace Element Analysis," Anal. Chem. 1982, 54,

21     Winefordner, J.D. Trace Analysis:  Spectroscopic Methods for Elements; in Chemical Analysis,
       Vol. 46, pp. 41-42.

22     "Results of the Validation Study for Determination of Trace Metals at EPA Water Quality
       Criteria Levels," April 1995. Available from the Sample Control Center (operated by
       DynCorp), 300 N. Lee Street, Alexandria, VA 22314, 703/519-1140.
January 1996

Method 1639
18.0 Glossary

       Many of the terms and definitions listed below are used in the EPA 1600-series methods, but
       terms have been cross-referenced to terms commonly used in other methods where possible.

18.1   Ambient Water—Waters in the natural environment (e.g., rivers, lakes, streams, and other
       receiving waters), as opposed to effluent discharges.

18.2   Analyte—A metal tested for by the methods referenced in this method. The analytes are
       listed in Table 1.

18.3   Apparatus—The sample container and other containers, filters, filter holders, labware, tubing,
       pipets, and other materials and devices used for sample collection or sample preparation, and
       that will contact samples, blanks, or analytical standards.

18.4   Calibration Blank—A volume of reagent water acidified with the same acid matrix as in the
       calibration standards. The calibration blank is a zero standard and is used to calibrate the ICP
       instrument (Section 7.6.1).

18.5   Calibration Standard (CAL)—A solution prepared from a dilute mixed standard and/or stock
       solutions and used to calibrate the response of the instrument with respect to analyte

18.6   Dissolved Analyte—The concentration of analyte in an aqueous sample that will pass through
       a 0.45-um membrane filter assembly before sample acidification (Section 8.3).

18.7   Equipment Blank—An aliquot of reagent water that is subjected in the laboratory to all
       aspects of sample collection and analysis, including contact with all sampling devices and
       apparatus.  The purpose of the equipment blank is to determine if the sampling devices and
       apparatus for sample collection have been adequately cleaned before shipment to the field site.
       An acceptable equipment blank must be achieved before the sampling devices and apparatus
       are used for sample collection.  In addition, equipment blanks should be run on random,
       representative sets of gloves, storage bags, and plastic wrap for each lot to determine if these
       materials are free from contamination before use.

18.8   Field Blank—An aliquot of reagent water that is placed in a sample container in the
       laboratory, shipped to the field, and treated as a sample in all respects, including contact with
       the sampling devices and exposure to sampling site conditions, storage, preservation, and all
       analytical procedures, which may include filtration. The purpose of the field blank is to
       determine if the field or sample transporting procedures and environments have contaminated
       the sample.

18.9   Field Duplicates (FD1 and FD2)—Two separate samples collected in separate sample bottles
       at the same time and place under identical circumstances and treated exactly the same
       throughout field and laboratory procedures.  Analyses of FD1 and FD2 give  a measure of the
       precision associated  with sample collection, preservation, and storage, as well as with
       laboratory procedures.
January 1996

                                                                                   Method 1639
 18.10   Initial Precision and Recovery (IPR)—Four, aliquots of the OPR standard analyzed to
        establish the ability to generate acceptable precision and accuracy. IPRs are performed before
        the first time a method is used and any time the method or instrumentation is modified.

 18.11   Instrument Detection Limit (IDL)—The concentration equivalent to the analyte signal which
        is equal to three times the standard deviation of a series of 10 replicate measurements of the
        calibration blank signal at the selected analytical wavelength.

 18.12   Laboratory Blank—An aliquot of reagent water that is treated exactly as a sample including
        exposure to all glassware, equipment, solvents, reagents, internal standards, and surrogates that
        are used with samples.  The laboratory blank is used to determine if method analytes or
        interferences are present in the laboratory environment, the reagents, or the apparatus.
        (Sections 7.6.2 and 9.5.1).

 18.13   Laboratory Control Sample (LCS)—See Ongoing Precision and Recovery (OPR) Standard.

 18.14   Laboratory Duplicates (LD1 and LD2)—Two aliquots of the same sample taken in the
        laboratory and analyzed separately with identical procedures.  Analyses of LD1 and LD2
        indicates precision associated with laboratory procedures, but not with sample collection,
        preservation, or storage procedures.

 18.15   Laboratory Fortified Blank (LFB)—See Ongoing Precision and Recovery (OPR) Standard.

 18.16   Laboratory Fortified Sample Matrix (LFM)—See Matrix Spike and Matrix Spike Duplicate.

 18.17   Laboratory Reagent Blank (LRB)—See Laboratory Blank.

 18.18   Linear Dynamic Range (LDR)—The concentration range over which the instrument response
        to an analyte is linear (Section 9.2.3).

 18.19   Matrix Modifier—A  substance added to the graphite  furnace along with the sample to
        minimize the interference effects by selective volatilization of either analyte or matrix

 18.20   Matrix Spike  (MS) and Matrix Spike Duplicate (MSD)—Aliquots of an environmental
        sample to which known quantities of the method analytes are added hi the laboratory.  The
        MS and MSD  are analyzed exactly like a sample. Their purpose is to quantify the bias and
        precision caused by the sample matrix.  The background concentrations of the analytes in the
        sample matrix must be determined hi a separate aliquot and the measured values in the MS
        and MSD corrected for background concentrations (Section 9.3).

 18.21   May—This action, activity, or procedural step is optional.

 18.22   May Not—This action, activity, or procedural step is prohibited.

 18.23   Method Blank—See Laboratory Blank.
January 1996

Method 1639
18.24  Method Detection Limit (MDL)—The minimum concentration of an analyte that can be
       identified, measured, and reported with 99% confidence that the analyte concentration is
       greater than zero (Section 9.2.1 and Table 1).

18.25  Minimum Level (ML)—The lowest level at which the entire analytical system gives a
       recognizable signal and acceptable calibration point (Reference 9).

18.26  Must—This action, activity, or procedural step is required.

18.27  Ongoing Precision and Recovery (OPR) Standard—A laboratory blank spiked with known
       quantities of the method analytes.  The OPR is analyzed exactly like a sample.  Its purpose is
       to determine whether the methodology is in control and to ensure that the results produced by
       the laboratory remain within the method-specified limits for precision and accuracy.

18.28  Preparation Blank—See Laboratory Blank.

18.29  Primary Dilution Standard—A solution containing the analytes that is purchased or prepared
       from stock solutions and diluted as needed to prepare calibration solutions and other solutions.

18.30  Quality Control Sample (QCS)—A sample containing all or a subset of the method analytes
       at known concentrations. The QCS is obtained from a source outside of the laboratory or is
       prepared from a source of standards different from the source of calibration standards.  It is
       used to check laboratory performance with test materials prepared outside of the normal
       preparation process.

18.31  Reagent Water—Water demonstrated to be free from the method analytes and potentially
       interfering substances at the MDL for that metal in the method.

18.32  Should—This action, activity, or procedural step is suggested but not required.

18.33  Standard Addition—The addition of a known amount of analyte to the sample to determine
       the relative response of the detector to an analyte within the sample matrix.  The relative
       response is then used to assess either an operative matrix effect or the sample analyte

18.34  Stock Standard Solution—A solution containing one or more method analytes that is
       prepared using a reference material traceable to EPA, the National Institute of Science and
       Technology (NIST), or a source that will attest to the purity and authenticity of the reference

18.35  Total Recoverable Analyte—The concentration of analyte  determined by analysis of the
       solution extract of an unfiltered aqueous sample following digestion by refiuxing with hot
       dilute mineral acid(s) as specified in the method (Section 12.2).
January 1996

                                                                                                         Method 1639
                                                       Table 1
        List of Analytes Amenable to Analysis Using Method 1639:  Lowest Water Quality Criterion
                    for Each Metal Species, Method Detection Limits, and Minimum Levels
Chromium (ffi)
Lowest EPA
Method Detection
Limit (MDL) and
Minimum Level (ML);

Lowest of the freshwater, marine, or human health WQC at 40 CFR Part 131 (57 FR 60848 for human health criteria and 60 FR 22228
for aquatic criteria).  Hardness-dependent freshwater aquatic life criteria also calculated to reflect a hardness of 25 mg/L CaCO3, and
all aquatic life criteria, except chronic criteria for Se, have been adjusted to reflect dissolved levels in accordance with the equations
provided in 60 FR 22228.  Hardness dependent dissolved criteria conversion factors for Cd were also calculated at a hardness of 25
mg/L per 60 FR 22228.

Method Detection Limit (MDL) as determined by 40 CFR Part 136, Appendix B.

Minimum Level (ML) calculated by multiplying laboratory-determined MDL by 3.18 and rounding result to nearest multiple of 1, 2,
5,10,20,50, etc. in accordance with procedures used by HAD and described in the EPA Draft National Guidance for the Permitting,
Monitoring, and Enforcement of Water Quality-Based Effluent Limitations Set Below Analytical Detection/Quantitation Levels, March
22, 1994.
January 1996

Method 1639
                                             Table 2

                   Quality Control Acceptance Criteria for Performance Tests1
Initial Precisian and
Recovery (Section
s X
60 24-144
11 67-142
26 76-129
36 69-141
31 60-128
19 67-142
(Section 10.7)
Ongoing Precision
Recovery (Section
(Section 9.3)
       All specifications expressed as percent
January 1996 \

                                                                                  Method 1639
                                            TABLE 3

                                      MATRIX MODIFIER1-3
Temperature °C4
Char Atom
1100 2000
800 1600
1650 2600s
1400 2500
1000 2000
1   Matrix Modifier = 0.015 mg Pd + 0.01 mg Mg(NO3)2.

2   A 5% H2 in Ar gas mix is used during the dry and char steps at 300 mL/min for all elements.

3   A cool-down step between the char and atomization is recommended.

4   Actual char and atomization temperatures may vary from instrument to instrument and are best determined on
   an individual basis.  The actual drying temperature may vary depending on the temperature of the water used
   to cool the furnace.

5   A 7-s atomization is necessary to quantitatively remove the analyte from the graphite furnace.

6   An electrodeless discharge lamp was used for this element.
     January 1996