HANDBOOK FOR ANALYTICAL QUALITY CONTROL
  IN WATER AND WASTEWATER LABORATORIES
                   For
  U.S. ENVIRONMENTAL PROTECTION AGENCY
             Technology Transfer
                   By
 ANALYTICAL QUALITY CONTROL LABORATORY
      National Environmental Research Center
               Cincinnati, Ohio
                 June 1972

-------
The mention of trade names or commercial products in this handbook is for illustration
purposes, and does not constitute endorsement or recommendation for use by the U. S.
Enviromental Protection Agency,

-------
                                      ABSTRACT


One of the fundamental responsibilities of management is the establishment of a continuing
program to insure the reliability and validity of analytical laboratory and field data gathered
in water treatment and wastewater pollution control activities.

This handbook is addressed to laboratory directors, leaders of field investigations, and other
personnel who bear responsibility for water and wastewater data. Subject matter of the
handbook is  concerned primarily with quality control for chemical and physical tests and
measurements.  Sufficient  information is offered to  allow the reader to inaugurate, or to
reinforce, a program  of analytical quality control which will emphasize early recognition,
prevention and correction  of factors leading to breakdowns in the validity of data.
                                         in

-------
                                ACKNOWLEDGEMENT
This handbook was prepared by the Analytical Quality Control Laboratory of the United
States Environmental Protection Agency. The contributions of the following individuals in
preparing the handbook are gratefully acknowledged.

                                D. G. Ballinger
                                R. L. Booth
                                M. R. Midgett
                                R. C. Kroner
                                J. F. Kopp
                                J. J. Lichtenberg
                                J. A. Winter
                                R. C. Dressman
                                J. W. Eichelberger
                                J. E. Longbottom

Inquiries regarding material contained in the handbook should be made to Environmental
Protection Agency, National Environmental Research Center, Analytical Quality Control
Laboratory, Cincinnati, Ohio 45268.

-------
                           TABLE OF CONTENTS
Chapter                                                               Page

           ABSTRACT                                                  iii
           ACKNOWLEDGEMENT                                        v
           TABLE OF CONTENTS                                       vii
           LIST OF FIGURES                                            ix
           LIST OF TABLES                                             xi

   1        IMPORTANCE OF QUALITY CONTROL                       1-1
           1.1    General                                               1-1
           1.2    Quality Control Program                                1-1
           1.3    Analytical Methods                                    1-2
           1.4    References                                            1-3

   2        LABORATORY SERVICES                                   2-1
           2.1    General                                               2-1
           2.2    Distilled Water                                        2-1
           2.3    Ammonia-Free Water                                   2-4
           2.4    Carbon Dioxide-Free Water                              2-4
           2.5    Ion-Free Water                                        2-4
           2.6    Compressed Air                                        2-4
           2.7    Electrical  Services                                      2-5
           2.8    References                                            2-5


   3        INSTRUMENTAL QUALITY CONTROL                        3-1
           3.1    Introduction                                          3-1
           3.2    Analytical Balances                                    3-2
           3.3    pH Meters                                            3-4
           3.4    Conductivity Meters                                    3-7
           3.5    Turbidimeters                                         3-11
           3.6    Spectrophotometers                                    3-11
           3.7    Organic Carbon Analyzer                                3-23
           3.8    Selective Ion Electrodes                                 3-28
           3.9    References                                            3-29
           GLASSWARE                                                4-1
           4.1     General                                               4-1
           4.2     Types of Glassware                                     4-1
           4.3     Volumetric Analyses                                    4-2
           4.4     Federal Specifications for Volumetric Glassware            4-4
           4.5     Cleaning of Glass and Porcelain                           4-6
           4.6     Special Cleaning Requirements                           4-7
           4.7     Disposable Glassware                                   4-8
           4.8     Specialized Glassware                                   4-8
           4.9     Fritted Ware                                           4-9
           4.10   References                                            4-11
                                    vn

-------
REAGENTS, SOLVENTS AND GASES                          5-1
5.1     Introduction                                           5-1
5.2     Reagent Quality                                        5-1
5.3     Elimination of Determinate Errors                        5-4
5.4     References                                            5-7

CONTROL OF ANALYTICAL PERFORMANCE                 6-1
6.1     Introduction                                           6-1
6.2     Precision and Accuracy                                 6-1
6.3     Evaluation of Daily Performance                          6-4
6.4     Quality Control Charts                                  6-5
6.5     References                                            6-18

DATA HANDLING AND REPORTING                          7-1
7.1     Introduction                                           7-1
7.2     The Analytical Value                                   7-1
7.3     Report Forms                                         7-5
7.4     STORET-Computerized Storage and Retrieval
         of Water Quality Data                                7-9
7.5     SHAVES-A Consolidated Data Reporting
         and Evaluation System                                7-11
7.6     References                                            7-11

SPECIAL REQUIREMENTS FOR TRACE ORGANIC ANALYSIS   8-1
8.1     Introduction                                           8-1
8.2     Discrete Bottled Samples                                8-1
8.3     Carbon Adsorption Samples                             8-2
8.4     Glassware                                             8-2
8.5     Reagents and Chemicals                                 8-2
8.6     Common Analytical Operations                          8-3
8.7     Gas-Liquid Chromatography                             8-4
8.8     Qualitative Analysis                                    8-7
8.9     Quantitative Analysis                                   8-8
8.10   Thin-Layer Chromatography                             8-9
8.11   Column Chromatography                                8-10
8.12   References                                            8-10

SKILLS AND TRAINING                                     9-1
9.1     General                                               9-1
9.2     Skills                                                 9-2
9.3     Training                                              9-4
                         vm

-------
                              LIST OF FIGURES
Figure No.                                                                Page

  3-1      Analytical Balance                                             3-3
  3-2      pH Meter                                                     3-5
  3-3      Conductivity Meter                                            3-8
  3-4      Turbidimeter                                                  3-12
  3-5      Spectrophotometer                                            3-13
  3-6      Atomic Absorption Unit                                        3-19
  3-7      Device for Reproducible Positioning of Burner Height             3-22
  3-8      Organic Carbon Analyzer                                       3-27
  3-9      Selective Ion Meter                                            3-30

  4-1      Titration Bench                                                4-3
  4-2      Example of Markings on Glassware                              4-6

  6-1      Essentials of Control Chart                                     6-5
  6-2      Effects of a and 0 Levels on Standard Control Chart              6-8
  6-3      Laboratory Quality Control Charts                              6-10

  7-1      Example of Bench Sheet                                        7-7
  7-2      Example of Summary Review Sheet                             7-8
  7-3      Example of STORE! Report  Form                              7-10
                                    IX

-------
                              LIST OF TABLES
Table No.                                                                 Page

  2-1      Water Purity                                                   2-1
  2-2      Comparison of Distillates from Glass and Metal Stills               2-2
  3-1      Instruments Commonly Used in Water and Wastewater Analysis     3-1
  3-2      pH Values of NBS Standards from 0 to 30° C                      3-6
  3-3      Performance Characteristics of Typical pH Meter                  3-7
  3-4      Electrical Conductivity of Potassium Chloride
                  Reference Solutions                                     3-10
  3-5      Design Features of Some Common Spectrophotometers             3-15
  3-6      Design Features of Some Common Atomic
                  Absorption Instruments                                  3-25
  4-1      Tolerances for Volumetric Glassware                             4-5
  4-2      Fritted Ware Porosity                                           4-9
  4-3      Cleaning of Filters                                              4-10
  6-1      Precision Data on River Water Samples for Phosphorus
                  AutoAnalyzer Method                                   6-2
  6-2      Accuracy Data on River Water Samples for Phosphorus
                  AutoAnalyzer Method                                   6-3
  6-3      Factors for Computing Control Chart Lines                       6-12
  9-1      Skill-Time Rating of Standard Analytical Operations               9-3
                                    XI

-------
                                       Chapter 1

                          IMPORTANCE OF QUALITY CONTROL
1.1  General

The role of the analytical laboratory is to provide qualitative and quantitative data to be
used in decision making. To be valuable, the data must accurately describe the character-
istics or  the concentration of constituents in the sample  submitted to the laboratory. In
many cases, an approximate answer  or incorrect result  is worse than no answer at all,
because it will lead to faulty interpretations.

Decisions made using water and wastewater data  are far-reaching. Water quality standards
are  set to establish satisfactory conditions  for a given water use. The laboratory data define
whether  that condition is being met, and whether the water can be used for its intended
purpose. If the laboratory results indicate a violation of the standard, action is required on
the  part of pollution control authorities.  With the  present emphasis on legal  action and
social pressures  to  abate pollution,  the analyst should be aware of his responsibility to
provide laboratory results that are a reliable description of the sample. Furthermore, the
analyst must be aware that his professional  competence, the procedures he has used, and the
reported values may be used and challenged in court. To satisfactorily meet this challenge,
the  laboratory data  must be backed up by an adequate program to document the proper
control and application of all of the factors which affect the final result.

In wastewater analyses,  the laboratory data define the treatment plant influent, the status
of the steps in the treatment process, and  the final load imposed upon the water resources.
Decisions on process changes, plant modification, or even the construction of a new facility
may be based upon the results of laboratory analyses. The financial implications alone are
significant reasons for extreme care in analysis.

Research investigations in water pollution control rest upon a firm base of laboratory data.
The final result sought can usually be described  in numerical terms. The progress of the
research  and the alternative pathways available are generally  evaluated  on the basis of
laboratory data. The value of the  research effort  will depend upon  the validity of the
laboratory results.

1.2  Quality Control Program

Because  of  the  importance  of  laboratory analyses and the resulting actions which  they
produce, a program to insure the reliability of the data is  essential.  It is recognized that all
analysts practice quality control to varying degrees,  depending somewhat upon their train-
ing, professional pride, and awareness of the importance of the work they are doing. How-
ever,  under the  pressure of daily workload,  analytical  quality  control may be easily
neglected. Therefore, an established, routine control program applied to every analytical test
is important in assuring the reliability of the final results.

The quality control program in the laboratory has two primary functions. First, the program
should monitor the reliability (truth) of the results reported. It should continually provide
an  answer to  "How good (true)  are the  results submitted?"  This  phase may be termed
"measurement of quality." The second function is the control of quality in order to meet
                                         1-1

-------
the program  requirements for reliability. For example, the processing of spiked samples is
the measurement of quality, while the use of analytical grade reagents is a control measure.
Just as each  analytical method has a rigid protocol, so the quality control associated with
that test must  also involve definite required steps to monitor and assure  that the result is
correct. The  steps in quality control will vary with the type of analysis. For example, in a
titration, standardization of the titrant on a frequent basis is an element of quality control.
In an instrumental method, the check-out of instrument response and the calibration of the
instrument in concentration  units is also a quality control  function. Ideally, all of the
variables which can affect the final answer should be considered, evaluated, and controlled.

This handbook  considers the factors which go into creating an analytical result, and provides
recommendations for the control of these factors in order to  insure that  the best possible
answer is obtained. A program based upon these recommendations will give the analyst and
his supervisor confidence in  the  reliability and the representative nature of the sample
charactieristics being reported.

Without  exception, the final responsibility  for the reliability  of the analytical results sub-
mitted rests with the Laboratory Director.

1.3 Analytical  Methods

In general, the widespread use of an analytical method indicates that it is a reliable means of
analysis, and this fact tends to support the validity of the test result reported. Conversely,
the use of a  little-known technique forces the data user to place  faith in  the judgement of
the analyst. When the analyst uses a "private"  method, or one not commonly accepted in
the field, he must  stand alone in defining  both his  choice of the method and the result
obtained.

The need  for  standardization of  methods within a single laboratory is  readily apparent.
Uniform methods between cooperating laboratories are also important in order to remove
the  methodology as  a variable in comparison or joint use of data between laboratories.
Uniformity of methods is particularly important when laboratories are providing data to a
common data bank, such as STORET*, or when several laboratories are cooperating in joint
field  surveys. A lack  of standardization of methods raises doubts as to the validity of the
results reported. If the same  constituent is measured by different analytical procedures
within a  single laboratory, or in  several laboratories, the  question is raised as to which
procedure is superior, and why the superior method is not used throughout.

The physical and chemical methods used should be selected by  the following criteria:

    a.  The method should measure  the desired constituent  with  precision and accuracy
       sufficient to meet the data needs  in  the presence of the interferences normally
       encountered in polluted waters.

    b. The procedure  should utilize  the  equipment and skills  normally available in the
       average water pollution control laboratory.
 *STORET is the acronym used  to  identify the computer-oriented  U.S.  Environmental
 Protection Agency Water Quality Control Information System for STOrage and RETrieval
 of data and information.
                                         1-2

-------
    c.  The selected methods should be in use in many laboratories or have been sufficiently
       tested to establish their validity.

    d.  The method should be sufficiently rapid to permit routine use for the  examination
       of large numbers of samples.

The use of EPA methods in all EPA laboratories provides a common base for combined data
between Agency  programs. Uniformity throughout EPA lends considerable support to the
validity of the results reported by the Agency.

Regardless of the analytical method used in the laboratory,the specific methodology should
be  carefully  documented. In some water pollution  reports it  is customary to state that
Standard Methods (1) have been used throughout. Close examination indicates, however, that
this is not strictly true. In  many laboratories, the standard method has been modified
because of recent research or personal preferences of the laboratory staff. In other cases the
standard method  has been replaced with a better one. Statements concerning the methods
used in arriving at laboratory data should be clearly and honestly stated. The methods used
should be adequately referenced and the procedures applied exactly as directed.

Knowing the specific method which has been used, the reviewer can apply the associated
precision  and accuracy of the method  when interpreting the laboratory results.  If  the
analytical methodology is in doubt, the data user may honestly inquire as to the reliability
of the result he is  to interpret.

The advantages of strict adherence to accepted methods should not stifle investigations
leading to improvements in analytical procedures. In spite of  the value of accepted and
documented methods,  occasions do arise when a procedure must be modified to eliminate
unusual interference, or to yield increased sensitivity. When modification is necessary,  the
revision should be carefully worked out to accomplish the desired result. It is  advisable to
assemble data using both the regular and the modified procedure to show the superiority of
the latter. This useful  information can be brought to the attention of the individuals and
groups responsible for methods standardization.  For maximum benefit,  the  modified
procedure should be rewritten in the standard format so that the substituted procedure may
be  used throughout the laboratory for routine examination of samples.  Responsibility for
the use of a non-standard procedure rests with the analyst and his supervisor, since such use
represents a departure from accepted practice.

In field operations, the problem of transport of samples to the laboratory, or  the need to
examine a large number of samples to arrive at gross values will sometimes require the use of
rapid field methods yielding  approximate  answers.  Such methods  should  be used with
caution, and with a clear understanding that the results obtained do not compare in relia-
bility with those  obtained using standard laboratory  methods. The  fact that  "quick and
dirty" methods have been used should be noted, and the results should not be reported along
with more reliable  laboratory-derived analytical information. The  data  user is entitled to
know that approximate values have been obtained  for screening purposes only, and that the
results do not represent the customary precision and accuracy obtained in the laboratory.

1.4 References

1.  Standard Methods for the Examination of Water and Wastewater, 13th Edition, Amer-
    ican Public Health Association, New York (1971).
                                         1-3

-------
                                      CHAPTER 2

                                LABORATORY SERVICES
2.1  General

Quality control of laboratory analyses involves consideration and control of the many
variables which affect the production of reliable data. The quality of the laboratory services
available to the analyst must be included among these variables. An abundant supply of
distilled water, free  from interferences and other undesirable contaminants, is an  absolute
necessity. An adequate source of clean, dry, compressed air is needed. Electrical power for
routine laboratory use and voltage-regulated sources for delicate electronic instrumentation
must  be  provided.  This  chapter,  therefore, will  be devoted to  describing methods of
maintaining the quality of these services, as used in laboratory operations.

2.2  Distilled Water

Distilled or demineralized water is used in the laboratory for dilution, preparation of reagent
solutions, and final rinsing of glassware. Ordinary distilled water is usually not pure. It may
be contaminated by dissolved gases and by materials leached from the container in which it
has been  stored. Volatile  organics distilled over from the feed water may be present, and
non-volatile impurities may occasionally  be carried over by  the steam, in the form of a
spray. The concentration of these  contaminants is usually quite small, and distilled water is
used for many analyses without further purification. However, it is highly  important that
the still, storage tank, and any associated piping be carefully  selected, installed, and main-
tained in such a way as to assure minimum contamination.

Water purity has been defined in many different ways, but one generally accepted definition
states that high purity water is water that  has been distilled and/or deionized so that it will
have a specific resistance  of 500,000  ohms (2.0 micromhos conductivity) or greater. This
definition is satisfactory as a base to  work from, but for more critical requirements, the
breakdown shown in Table 2-1 has been suggested to express degrees of purity (1).
                                      Table 2-1


                                  WATER PURITY

Degree of Purity                  Maximum                 Approximate Concentration
                                 Conductivity                     of Electrolyte
                              (micromhos/cm)                         (mg/1)
Pure                                10                                 2-5
Very Pure                            1                               0.2-0.5
Ultrapure                          0.1                             0.01-0.02
Theoretically Pure                0.055                                0.00
                                         2-1

-------
Properly designed metal stills from reputable manufacturers offer a convenient and reliable
source of distilled water. These stills are usually constructed of copper, brass, and bronze.
All surfaces that contact  the  distillate  should be heavily coated  with pure tin to prevent
metallic contamination. The metal storage tank  should be of sturdy construction, have a
tight-fitting cover, and a filter in the air vent to remove airborne dust, gases, and fumes.

For special purposes, an all-glass distillation unit may be preferable to the metal still. These
stills are usually smaller, and  of more limited capacity than the metal  stills. An actual
comparison in which the  distillates from an  all-glass still and a metal still were analyzed
spectrographically for certain trace metal contaminants is shown below in Table 2-2.

                                      Table 2-2
        COMPARISON OF DISTILLATES FROM GLASS AND METAL STILLS

                      Element and Concentration  (Micrograms/1)	

                                                                       Ni       Pb
Source
All-Glass Still
Metal Still
Zn
<1
9
B
12
13
Fe
1
2
Mn Al
<1 4
<1 <5
Cu
5
11
                                                                       <1       <2
                                                                       <2        26

It can be  seen  that  the all-glass still produced  a product which had substantially lower
contamination from zinc, copper, and lead. Other analyses have indicated the same general
relationship, except that a boron concentration of 100 micrograms/liter was found in water
from the all-glass still on one occasion. This was probably related to the length of time the
distillate had remained in the glass storage reservoir.

All stills require periodic cleaning to remove solids which have been deposited from the feed
water. Hard water and high dissolved solids content promote scale formation in the evapor-
ator, and cleaning frequency will thus depend on the quality of the feed water. The boiler of
an all-glass still  should be drained daily and refilled with clean water. Build-up of scale is
easily detected,  and the  boiler and condenser coils should be cleaned  at frequent intervals.
Metal stills  usually incorporate a constant bleeder device which retards scale formation to
some extent. However, these units should still be dismantled and cleaned at regular intervals.
Cleaning should  always be in accordance with the manufacturer's instructions.

Pre-treatment of the incoming feed  water will often improve still performance and raise the
quality of the distillate.  For  example, preliminary softening of hard water removes calcium
and  magnesium  prior to distillation. This reduces scale formation in the boiler  and conden-
ser,  thereby reducing  maintenance  service.  These  softeners  employ the  ion  exchange
principle using a sodium chloride cycle, and are relatively inexpensive to operate. A carbon
filtration system, installed at the feed water intake, will remove organic materials which
might subsequently be carried over in the distillate. If trace concentrations of ions are a
major concern, the distillate may  be passed through a mixed-bed ion exchanger.

At least two commercially manufactured systems are available for production of high purity
water by ion exchange. The  Millipore Super-Q System (Millipore Corp., Bedford, Mass.),
consists essentially of disposable  cartridges for prefiltration, organic  absorption, deioniza-
tion  and  Millipore  filtration. The  company  claims it  can produce  10 megohm  water,
containing no particulate matter  larger than 0.45 micron in size, from tap water, at the rate
of 20 gallons per hour. Continental  Water Conditioning  Corp., El Paso, Texas, advertises a
                                         2-2

-------
system  which  can be tailored  to the needs of the customer. Performance specifications
include minimum flow rates of 45 gallons per hour and total dissolved solids of less than 0.1
mg/1 when required.

Specific conductance  is a  rapid  and simple measurement for determining the inorganic
quality of distilled water. Stills of the types previously discussed are capable of producing a
distillate with a specific conductance of less than 2.0 micromhos at 25°C. This is equivalent
to 0.5-1.0 mg/1  of ionized material.  Frequent checks should  be made to determine that
optimum performance is being maintained. A purity meter installed between the still and
the storage reservoir will monitor the conductivity of the distillate, in  terms of the equiva-
lent in mg/1  of sodium chloride. If the reading on the meter begins to rise above the present
limit of conductivity,  effective action should be taken to eliminate the source of contamin-
ation. Organic  quality is more difficult to monitor, but the total organic carbon  determin-
ation is a simple and rapid test for organic contaminants.

A piping system for delivering distilled  water to the area of use within the laboratory is a
convenient and desirable feature.  In this case, special care should be taken that the quality
of the water is not degraded  between the still and the point of use. Piping may  be of tin,
tin-lined brass, stainless steel, plastic, or chemically resistant glass, depending on the quality
of the water desired  and on  available funds. Tin is best,  but  is also very expensive. As a
compromise, plastic pipe, or glass pipe with  Teflon* O-rings at  all  connecting joints is
satisfactory for most  purposes. The glass pipe has an obvious advantage when freedom from
trace amounts of organic materials is important.

When there  is no piped-in supply,  distilled water  will probably  be   transported to the
laboratory and stored in polyethylene or glass bottles of about 5-gallon capacity. If stored in
glass containers, distilled water will gradually leach the more soluble  materials from the glass
and  cause  an increase in dissolved solids.  Therefore, only borosilicate-free glass containers
should  be used.  Polyethylene bottles  contain organic plasticizers, and  traces of  these
materials may be leached from the container walls. These are of little consequence, except
in some organic analyses. Rubber stoppers often used in storage containers contain leachable
materials, including significant quantities of zinc. This is usually  no problem, since the water
is not in  direct  contact with the stopper.  However, the analyst should  be aware of the
potential for contamination, especially when the supply is not replenished by frequent use.

The  delivery  tube may consist of a piece of glass tubing which extends almost to the bottom
of the bottle, and which is bent downward above the bottle neck, with a three-to-four-foot
piece of flexible tubing attached for mobility. Vinyl tubing is preferable to latex rubber,
because it  is less leachable.  However,  a short piece of latex tubing may be required at the
outlet for better control of the pinchcock. The vent tube in the stopper should be  protected
against the entrance of dust.

Ordinary distilled water is quite adequate for many analyses, including the determination  of
major cations  and anions.  Certain needs may require the use  of  double- or even triple-
distilled water. Redistillation  from an alkaline permanganate solution can be used to obtain
a water with low organic background. When determining trace organics by solvent extrac-
tion and gas chromatography,  distilled water  with sufficiently low background may be
extremely  difficult to  obtain.  In this case, pre-extraction of the water with the solvent used
in the respective  analysis may be helpful in eliminating undesirable peaks in the blank.
*Trademark of E.I. duPont de Nemours & Co.
                                          2-3

-------
Certain analyses require special treatment or conditioning of the distilled water, and these
will now be discussed.

2.3 Ammonia-Free Water

Removal of ammonia can be accomplished by shaking ordinary distilled water with a strong
cation exchanger, or by  passing distilled  water through a column  of such material.  For
limited volumes of ammonia-free water, use of the Quikpure (Box 254, Chicago, 111.) 500-ml
bottle is highly recommended. The ion-free water described below is also suitable for use in
the determination of ammonia.

2.4 Carbon-Dioxide-Free Water

Carbon-dioxide-free water may be  prepared by boiling distilled water for 15  minutes and
cooling to room  temperature. As an alternative, distilled water  may be vigorously aerated
with a stream of inert gas for a period sufficient to achieve saturation and CO2 removal.
Nitrogen is most  frequently used. The  final pH of the water should lie between 6.2 and 7.2.
It is not advisable to store CO2-free water for extended periods.

2.5 I on-Free Water

A multi-purpose  high purity water, free from trace amounts of  the common ions, may be
conveniently  prepared by slowly passing  distilled  water through an ion-exchange column
containing one part of a strongly acidic cation-exchange resin in the hydroxyl  form. Resins
of a  quality  suitable for analytical work must  be used. Ion-exchange  cartridges of the
research grade, available from scientific supply houses,  have been found satisfactory. By
using a fresh column and high quality  distilled water, a water corresponding to the ASTM
designation for referee reagent water (2)  (maximum 0.1 mg/1 total matter and maximum
conductivity  of 0.1  micromho)  can be obtained. This water  is suitable  for use in the
determination  of ammonia, trace  metals,  and low concentrations of most  cations  and
anions. It  is  not suited to some organic analyses, however, because  this treatment adds
organic contaminants to the water by contact with the ion-exchange materials.

2.6 Compressed Air

The  quality of compressed  air required in the laboratory is usually very high, and special
attention should  be given to producing and maintaining clean air until it reaches the outlet.
Oil, water, and dirt are undesirable  contaminants in compressed air, and it is important to
install equipment which generates  dry, oil-free air. When pressures of less than 50 psi are
required, a rotary-type compressor, using  a water seal and no oil, eliminates any addition of
oil which would subsequently have to  be  removed from the  system. Large, horizontal,
water-cooled  compressors will usually be used when higher pressures are required.

Compression  heats air, thus increasing its tendency to retain moisture. An aftercooler  is
therefore necessary to remove water. Absorption filters should be used at the compressor to
prevent  moisture from entering  the piping system. Galvanized steel  pipe  with  threaded,
malleable-iron fittings, or solder-joint copper tubing should be used for piping the air to the
laboratory.

When the compressed air entering the  laboratory is of low quality, an efficient filter should
be installed  between the outlet and  the point  of use to trap oil,  moisture, and other
contaminants. As an alternative, high quality compressed air of the dry grade  is commercially
available in cylinders when no other source exists.

                                         2-4

-------
2.7 Electrical Services

An adequate  electrical system  is indispensable  to  the modern laboratory. This involves
having both 115- and 230-volt sources in sufficient capacity for the type of work that must
be done. Requirements for satisfactory lighting, proper functioning of sensitive instruments,
and operation of high current devices must be considered. Any specialized equipment may
present unusual demands on the  electrical supply.

Due to the special type of work, requirements  for a laboratory lighting system are quite
different from those in other areas. Accurate readings of glassware graduations,  balance
verniers, and other measuring lines must be made. Titration endpoints, sometimes involving
subtle changes in  color or shading,  must be  observed. Levels of illumination,  brightness,
glare, and  location of light sources should be controlled to facilitate ease in making these
measurements and to provide maximum comfort for the employees.

Spectrophotometers, flame  photometers, atomic absorption equipment,  emission spectro-
graphs, gas  chromatographs, etc. have complicated  electronic  circuits  which  require
relatively constant voltage to maintain stable, drift-free instrument operation. If the voltage
to these circuits varies, there is a  resulting change in  resistance, temperature, current,
efficiency,  light output, and component life. These characteristics are interrelated, and one
cannot  be  changed without  affecting the others. Voltage regulation is therefore necessary to
eliminate these conditions.

Many instruments have built-in voltage regulators  which perform this function satisfactorily.
In the absence of these, a small, portable, constant-voltage transformer should be placed in
the circuit between the electrical outlet and the  instrument. Such units are available from
Sola Basic  Industries, Elk Grove Village, 111., and are capable of supplying a constant output
of 118  volts from an input which varies between 95 and 130 volts. When requirements are
more stringent, special transformer-regulated circuits can be used to supply constant voltage.
Only  the instrument receiving the regulated voltage should be operated from such a circuit
at any  given time.  These lines are in addition to, and separate from the ordinary circuits
used for operation of equipment with less critical requirements.

Electrical heating  devices provide desirable heat sources, and  should  offer continuously
variable temperature control. Hot plates and muffle furnaces wired for 230-volt current will
probably give  better service  than those which operate on 115 volts, especially if the lower
voltage  circuit  is  only marginally  adequate. Water  baths  and  laboratory  ovens  with
maximum operating temperatures of about 200° C perform well at 115 volts. Care must be
taken to ground all equipment  which could constitute a shock hazard. The three-pronged
plugs which incorporate the ground are best for this purpose.
2.8  References

1.   Applebaum, S. B., and Crits, G. J., "Producing High Purity Water." Industrial Water
    Engineering, Sept./Oct. 1964.

2.   1968  Book of ASTM Standards, Part 23, Atmospheric Analysis, pp. 225-6, American
    Society for Testing and Materials, 1916 Race Street, Philadelphia, Pa. 19103.
                                         2-5

-------
                                    CHAPTER 3

                         INSTRUMENTAL QUALITY CONTROL
3.1  Introduction

The modern analytical  laboratory depends very  heavily upon instrumentation.  This
statement may  be  completely  obvious, but it should be remembered that the exceptional
emphasis on electronic equipment has really begun since the development of the transistor
and  the  computer. Analytical instrumentation, to a certain extent,  is always  in  the
development  stage,  with  manufacturers  continually redesigning and  upgrading their
products,  striving  for miniaturization, better durability  and  sensitivity, and improved
automation. The net result to laboratory supervisors and staff members is a bewildering
stream  of advertising  brochures,  announcements, and  catalogues of  newly  available
equipment. Consequently, the selection and  purchase of  analytical equipment is, at all
times, beset with uncertainty.

Table 3-1 lists  the instruments most commonly used for water and waste water analysis.
These represent basic equipment used in routine work and should be the subject of careful
consideration before purchase. Further, operation and maintenance of these devices ought
to be a primary consideration in production of satisfactory data. Obviously, a fundamental
understanding of instrument  design will  assist the analyst  in  the  correct use  of the
instrument and  in some cases aid in detecting instrumental failures.

In the pages that follow an attempt is made to discuss basic instrument design and to offer
some remarks whenever possible about desirable instrumental features.
                                      Table 3-1

                            INSTRUMENTS COMMONLY USED
                         IN WATER AND WASTEWATER ANALYSIS

                             Analytical Balance
                             Potentiometer (pH meter)
                             Conductivity meter
                             Turbidimeter
                             Spectrophotometers
                               a. Visual
                               b. Ultraviolet
                               c. Infrared
                               d. Atomic absorption
                             Total Carbon Analyzer
                             Gas Chromatographs
                             Miscellaneous
                               a. Temperature devices (ovens, water
                                 baths, etc.)
                               b. Recorders
                               c. Selective Ion Electrodes
                                        3-1

-------
3.2 Analytical Balances

The most  important piece of equipment in any  analytical  laboratory is the analytical
balance  (See Figure 3-1).  It  bears  the same relationship  to accuracy of measurements
produced  by   a laboratory  as  the  Greenwich  standard  clock  has to international
time-keeping. If the balance is not accurate all data related to weight-prepared standards will
contain the same degree of error. The balance, therefore, should be the most protected and
cared-for instrument  in the laboratory. Unfortunately,  care  of the balance is frequently
overlooked.

There are many fine  balances on  the market designed to meet a variety of needs such as
sensitivity,  speed weighing, batch weighing, etc.  Types of balances include general purpose,
micro-, electro-, semi-analytical, analytical and other special purpose instruments. Each type
of balance  has  its  own place in  the  scheme  of laboratory  operation but the analytical
balance is by far the most important in the production of reliable data.

Most analytical balances in use today in well equipped laboratories are of the "single pan"
variety. Single-pan capacities range from 80 grams to the popular 200-gram models with
sensitivities from 0.01 to 1 mg. Features of single-pan balances include mechanical  lifting
and substitution of weights, digital readout of weights, and mechanical zeroing of the empty
balance.  The advantage of the  single-pan balance over  the  old "two-pan" balance is in
greatly increased weighing speed and improved weighing accuracy because of mechanical
weight handling. With all the design  improvements, however, the modern analytical balance
is  still  a  fragile instrument, subject  to  shock,  temperature and humidity  changes,
mishandling and various  other  insults. Some  of the  precautions to be  observed in
maintaining and prolonging the dependable life of a balance are as follows:

    a.   Analytical balances should be mounted on a heavy shock-proof table, preferably
        one with adequate working surface  and a  suitable drawer for storage of balance
        accessories;  balance level should  be  checked  frequently  and  adjusted  when
        necessary.

    b.   Balances should be located  away from laboratory  traffic, protected from sudden
        drafts and humidity changes.

    c.   Balance  temperatures  should  be equilibrated  with  room  temperature;  this  is
        especially important if  building heat is shut  off or  reduced during non-working
        hours.
    d.  When not  in  use, the beam should be raised  from the knife edges, the weights
        returned to the beam, objects such as weighing dish removed from the pan, and the
        slide door closed.

    e.  Special precautions should be taken to avoid spillage of corrosive chemicals on the
        pan  or  inside the balance case; the  interior of balance housing should be kept
        scrupulously clean.

    f.  Balances should be checked and adjusted periodically by  a company service man or
        balance consultant; if service is not available  locally, follow  the manufacturer's
        instructions as closely as possible.
                                         3-2

-------
                                                              w
                                                              u
                                                              2
                                                              U
                                                              HH
                                                              OJ
                                                              S-i
3-3

-------
    g.   The  balance should  be operated  at  all times according to the  manufacturer's
        instructions.

Standardized weights to be used in checking balance accuracy, and that meet U.S. Bureau of
Standards specifications, may be purchased from various supply houses. A very complete set
of directions for checking the performance of a balance is contained in Part 30 of ASTM
Standards (1).

Since all analytical balances of the 200-gram capacity have about the same specifications
with reference to sensitivity, precision, convenience, and price, and since these specifications
are suitable for normal weighing requirements in water and wastewater laboratories, it is safe
to assume that there is no clear preference for a certain model, and selection will probably
be made on the basis of service availability.

3.3 pH Meters

The concept  of pH as  a means of expressing the degree of effective acidity or alkalinity as
contrasted with total acidity or alkalinity was  developed by Sorenson in  1909. It was not
until  about  1940  that  commercial instruments were developed  for routine  laboratory
measurement of pH.

A basic pH  meter (See Figure  3-2) consists of a voltage source, amplifier, and readout
device, either scale or digital. Certain additional  refinements produce varying performance
characteristics  between models.  Some models  incorporate expanded scales for increased
readability and solid  state  circuitry for operating stability  and  extreme accuracy.  All
instruments of recent design also include temperature adjustment and slope adjustment to
correct for asymmetric potential of glass  electrodes. Other features are scales that facilitate
use of selective ion electrodes, recorder output, and interfacing with complex data handling
systems.

In routine analytical work,  the  glass electrode  is used as the indicator  and the calomel
electrode as the reference.  Glass electrodes have a very fast  response time in highly buffered
solutions.  However, accurate readings are obtained slowly in poorly buffered samples, and
particularly when changing from buffered to unbuffered samples, as after standardization.
Electrodes, both  glass and calomel, should be well rinsed with distilled  water after each
reading, and  should be rinsed or dipped several times into  the  next test sample before the
final reading  is taken. Weakly buffered samples should  be stirred during measurement. Glass
electrodes should not be allowed to become dry during periods of inactivity. When not in
use they should be immersed in distilled water.

The  first  step  in standardization of the instrument is done by immersing the glass and
calomel electrodes into a buffer of known pH, setting the meter scale or needle to the pH of
the buffer and adjusting  the proper controls  to bring  the  circuit into balance.  The
temperature  compensating dial should be set at the sample temperature. The  pH of the
standard buffer should be within about two pH units of the sample. For best accuracy, the
instrument should be calibrated against two buffers that bracket the pH of the samples.

The presence of  a  faulty electrode is indicated  by failure  to  obtain  a reasonably correct
value  for the  pH  of the second  reference buffer solution   after  the  meter has been
standardized  with the first. A cracked glass electrode will often yield pH  readings that are
essentially the same for both standards. The response of electrodes may also be impaired by
                                         3-4

-------
                                                       w
                                                       H
                                                       W
                                                        s
                                                        3
                                                        so
                                                       (Jn
3-5

-------
failure to maintain  the KC1 level in the calomel electrode, or by certain specific materials
such as oily substances and precipitates that may coat the surfaces. A faulty condition can
be recognized from the check with the two buffer solutions. If either of these conditions
should occur the electrode can probably be restored to normal by an appropriate cleaning
procedure. Complete and detailed cleaning methods are given in Part 23 of ASTM Standards
(2).

Because of the asymmetric potential of the glass electrode most pH meters are built with a
"slope adjustment" which enables  the analyst to  correct for slight  electrode errors that
occur when standardization is performed at two different pH levels. Exact details of slope
adjustment  and  slope check may vary with different models of instruments.  The slope
adjustment must be made whenever electrodes are changed, subjected to vigorous cleaning,
or refilled with  fresh electrolyte. The  slope adjustment  feature is  highly desirable  and
recommended for consideration when purchasing a new meter.

Most pH meters now available are built with transistorized circuits rather than vacuum tubes
which  greatly reduces warm-up  time  and increases stability of the meter.  Also,  many
instruments are designed  with a switching circuit so that the conventional 0-14  scale may be
used to  read  a  single pH unit.  The  "expanded-scale" feature allows for  more accurate
readings  and may be of definite value when the meter is used for potentiometric titrations.
It is of dubious value, however, in routine analytical work, since readings more percise than
±0.1 pH are seldom required.

Solid  state circuitry has  led to improved design of compact instruments suitable for field
work. Field-type instruments are generally battery-powered,  and require more  maintenance
and more frequent standardization than laboratory instruments.

Standard  buffer solutions, covering a range of pH, may be purchased from almost  any
chemical supply house and are completely satisfactory for routine use. Table 3-2 below gives
a list of NBS buffers (easily made in the laboratory) and the resulting pH at several different
temperatures.
Temp.
°C
0
10
15
20
25
30
0.05M
Potassium
Tetroxalate
1.67
1.67
1.67
1.68
1.68
1.69
                                      Table 3-2

                  pH VALUES OF NBS STANDARDS FROM 0 - 30°C
                                                         0.025 M
                     Potassium      0.05 M              Potassium            0.01 M
                     Acid Tar-     Potassium      Dihydrogen Phosphate     Sodium
                     trate (Sat.       Acid          + 0.025 M Sodium        Tetra-
                     at 25°C)     Phthalate      Dihydrogen Phosphate      borate

                                     4.01                  6.98                9.46
                                     4.00                 6.92                9.33
                                     4.00                 6.90                9.27
                                     4.00                 6.88                9.22
                       3.56          4.01                  6.86                9.18
                       3.55          4.01                  6.85                9.14
                                         3-6

-------
Some idea of the effect of temperature on pH may be obtained by observing temperature vs
pH  of various buffers  shown in the table. Theoretically, the  potential response of the
electrode system changes 0.20 mV per pH unit per degree centigrade. Since all pH meters
measure potential but read out in pH, a variable compensation is used. A rough rule of
thumb is that temperature  compensation is about 0.05 pH units  per 5 degree increase in
temperature.

Typical performance data of a conventional expanded scale pH meter is shown in Table 3-3
below.

                                     Table 3-3

           PERFORMANCE CHARACTERISTICS OF TYPICAL pH METER
Range


Smallest scale division


Accuracy


Reproducibility


Temp,  compensation


Input impedance

Power  requirements

Dimensions
Normal Scale

0 to 14 pH
±1400 mv

0.1 pH
10 mv

±0.05 pH
±5 mv

±0.02 pH
±2 mv

0 to 100°C  (manual
or automatic)

>1014

115/220V, 50/60 Hz

12V2" w x 11" d x 8y4" h
Expanded Scale

  1 pH
  ±100 mv

  0.005 pH
  0.5 mv

  ±0.002 pH
  ±2% of reading

  ±0.002 pH
  ±0.2 mv
      |13
3.4  Conductivity Meters

Solutions of electrolytes conduct  an  electric current by  the migration of ions under the
influence of an electric field. For a constant  applied EMF, the current flowing  between
opposing electrodes immersed in the electrolyte will vary inversely with the resistance of the
solution. The reciprocal of the resistance is called conductance and is expressed in reciprocal
ohms or mhos. For natural  water samples where the resistance is high, the usual reporting
unit is in micromhos. Figure 3-3 shows a typical conductivity meter.

The passage of direct current through an electrolyte causes changes in the electrolyte, hence,
to prevent polarization, it is necessary  to use an alternating current or current pulses of
short  duration when measuring conductivity. Originally, conductivity meters were built
using  a  rapidly alternating current of low intensity in the audio range, and were equipped
                                       3-7

-------
                                                   w
                                                   H
                                                   W
                                                   I— I
                                                   H
                                                   U

                                                   Q
                                                   Z
                                                   O
                                                   U

                                                   ro
                                                   ro
                                                   p-
3-8

-------
with balance. For obvious reasons, the earpiece has been replaced by other devices, such as
the cathode ray tube commonly  known as  the "magic eye". Practically all  conductivity
meters on the present market use  some variation of the "magic eye" for indicating solution
conductivity, and include a  stepping switch for varying resistances in steps of  10X. The
instruments are therefore capable of reading conductivities from about 0.1 p-  mhos to about
250,000 F mhos.

The sensing element  for a conductivity measurement is  the  conductivity  cell,  which
normally consists of two thin plates of platinized metal, rigidly supported with a very
precise  parallel  spacing.  For protection, the plates are mounted inside  a glass tube, with
openings in the side  walls and submersible end for access of sample. Variations in designs
have included use of hard rubber and plastics for protection of the cell plates. Glass may be
preferable, in that the plates may  be visually observed for cleanliness and possible damage,
but the more durable encasements have the advantage of greater protection and reduced cell
breakage. One manufacturer offers a cell containing  circular  carbon plates embedded in an
epoxy-type plastic. Reversing the usual procedures, the sample is poured into the cell. The
cell is particularly attractive because of its ruggedness and the fact that it can be cleaned
without changing the cell factor.

Special precaution is taken with the arrangement of lead wires from the plates. Since stray
electrolytic and  capacitive  current  will  pass between them, thereby  distorting the bridge
balance, the wires must be properly separated, usually by use of a non-conducting bead or
collar on one of the lead  wires.

In routine use, cells should be  frequently examined to insure that (a) platinized coating of
plates is intact, (b) plates are  not coated with suspended matter, (c) plates are  not  bent,
distorted, or misaligned, and (d) lead wires are properly spaced.

Temperature has  a  pronounced effect  on  the conductance  of  solutions, and  must  be
corrected when  results are reported. The specified temperature for reporting data used by
most  analytical groups  (and  all  EPA laboratories) is 25°C. Data  correction may  be
accomplished  by adjusting  sample temperatures to  25°C.,or  by use of mathematical or
electronic adjustment. Adjustment of sample temperature is the preferred system, because
of the empirical nature of the mathematical correction. However, investigative work in the
Analytical Quality Control Laboratory, and at the upper Ohio Basin Office, EPA, has shown
that  acceptable  data is  obtained  if conductivity  readings taken  at stream-side  are
electronically temperature-corrected to 25°C.

Instrumental troubles  are  seldom encountered with conductivity meters because of the
design  simplicity. When  troubles occur they  are usually in the cell, and  for most accurate
work the following procedures should be used:

    a.   Standardize the cell and establish a cell factor by measuring the  conductivity of a
        standard potassium chloride solution  (standard conductivity tables may be found in
        various handbooks).

    b.   Rinse the cell by repeated immersion  in distilled water.

    c.   Again, immerse the cell in the sample several times before obtaining a reading.

    d.   If the meter is equipped with a "magic eye", determine the maximum width of the
                                         3-9

-------
       shadow at least twice, by approaching the endpoint from a low reading upward, and
       from a high reading downward.

Because the cell constants are subject  to slow  change, even under ideal conditions, and
sometimes to more rapid change under adverse conditions, it is recommended that the cell
constant be periodically established. Table 3-4, below, can be used for this operation.


                                    Table 3-4
                          ELECTRICAL CONDUCTIVITY
               OF POTASSIUM CHLORIDE REFERENCE SOLUTIONS

  Solution Normality  Method of Preparation  Temp. (°C)    Conductivity Q mhos)

    A      0.1 •      7.4365 g KCI/1               0             7,138
                        at 20°C                  18            11,167
                                                 25            12,856
    B      0.01      0.7440 g KCI/1               0               773
                        at 20° C                  18             1,220
                                                 25             1,408
    C      0.001     Dilute 100 mlofB
                        to 1 1 at20°C            25               147
                      K1+K2
  For instruments reading in mhos, calculate the cell constant as follows:


                L   =
                       1,000,000 xKx
  where
                L   =  cell constant
                Kj  =  Conductivity, in n mhos/cm, of theKC1 solution
                       at the temperature of measurement
                K2  =  conductivity, in ^ mhos, at the same temperature,
                       of the distilled water used to prepare the reference
                       solution
                Kv  =  measured conductance, in mhos
                 A

Conductivity equipment  which has generally been found to be reliable for laboratory work
includes the YSI #31 (Yellow Springs Instrument Co., Box  279, Yellow Springs, Ohio
45387), the Lab-Line Mark IV #1100 and MC3 #11025 (Lab-Line Instruments Inc., 15th &
Bloomingdale Aves., Melrose Park, 111.  60160), and Industrial Instruments #RC 16B2 and
RC-18 (Industrial Instruments Inc., 89 Commerce Road, Cedar Grove, N.J. 07009).

The YSI #31 is particularly suitable for routine lab work because of the brilliance of the
magic  eye.  The Lab-Line  meters  provide the sturdy cell previously mentioned.  The
Industrial  Instruments #RC-18 is designed for extreme accuracy but the  number of dial
settings required to obtain a single reading do not recommend it for routine analysis. A
recent instrument  survey conducted  by  the  Methods and Performance Activity  of the
Analytical Quality  Control Laboratory showed that a preponderance of EPA  Laboratories
used the Industrial Instruments RC 16B2.
                                       3-10

-------
3.5 Turbidimeters

Instruments for the measurement  of turbidity  have traditionally employed principles of
design related to transmission or reflectance of light. The lack of a primary standard for
turbidity, however, has resulted in a complete absence of uniformity among the available
instruments. Further, the Jackson  Candle Turbidimeter, which does not depend upon the
use of a primary standard, is a primitive instrument, subject to many interferences, and the
measurements generally are not reproducible.

Recent investigations (3) have resulted in  the  design of an instrument which has  been
adopted by the Environmental Protection Agency as a standard. The specifications for the
instrument are described elsewhere (4). Presently,  the Hach Turbidimeter, Model 2100,
(Box 907,  Ames, Iowa 50010) is the only turbidity measuring device manufactured which
meets these specifications. Figure 3-4 shows this instrument.

The Hach Turbidimeter Model 2100 employs, for standardization, a suspension of formazin,
especially used because of its stability and uniform particle size. For calibration purposes,
the formazin is permanently embedded  in a cylinder of lucite, the cylinder duplicating the
size and shape of the sample cuvette. Although the instrument is designed with a series of
scales ranging 0-1, 0-10, 0-100 and 0-1000 JCU,  it  has been recommended that turbidity
readings in excess of 40 JCU be rejected  (3). Correct use of the meter therefore requires that
samples containing turbidities in excess  of 40 JCU be diluted to a value below this level and
the results multiplied by the proper  dilution factor.

For production of data with maximum accuracy and precision the following precautions
should be observed:

    a.   Protect the lucite standard from  nicks, scratches and fingerprints.

    b.  Use a  constant orientation of the lucite standard while  calibrating the instrument.

    c.   Use a  well mixed sample in the sample  cuvette; do not take readings until finely
        dispersed bubbles have disappeared.

    d.  Dilute samples  containing  excess turbidity  to some value below 40 JCU;  take
        reading, and multiply results by correct dilution factor.
3.6 Spectrophotometers

Since a large portion of routine quantitative measurements are performed colorimetrically,
the  spectrophotometer (See  Figure 3-5)  is  usually the  workhorse  of  any  analytical
laboratory. Indeed, the versatility of the instrument, and the number of demands imposed
upon it, have resulted in a large variety of designs and price ranges. A systematic listing and
detailed discussion of all instrumental types would be beyond the scope of this chapter. As a
matter  of  convenience   and  practicality, therefore,  Spectrophotometers  are  discussed
separately as visible, ultraviolet, infrared, and atomic absorption instruments.

A spectrophotometer is an instrument for measuring an amount of light or radiant energy
transmitted through a solution, as a function of wave-length. A spectrophotometer differs
                                        3-11

-------
                                            w
                                            H
                                            W
                                            Q
                                            KH

                                            03
3-12

-------
                                                   w
                                                   S
                                                   o
                                                   H
                                                   O
                                                   ffi
                                                   OH
                                                   O
                                                   CJ
                                                   W
                                                   OH
                                                   00
                                                    o>


                                                    eo

                                                   PH
3-13

-------
from  a  filter  photometer  in  that  it  uses  continuously variable, and more  nearly
monochromatic, bands of light.  Filter photometers are relatively insensitive, and lack the
versatility of spectrophotometers. They are used most profitably where a single method can
be designed to fit the instrument.

The essential parts of a spectrophotometer are:

    a.   A source of radiant energy, usually a tungsten filament bulb,

    b.   The Monochromator, a device for isolating narrow bands of light,

    c.   Cells (cuvettes), for holding the colored solution under investigation, and

    d.   The photodetector, a device to detect and  measure the  radiant  energy passing
        through the sample solution.

Each of the essential features listed, especially the monochromator and the photodetector
system,  vary in design principles from one instrument to another. Table 3-5 shows some of
the features of the more commonly used spectrophotometers.

3.6.1. Visible Range

Desirable features on a visible-range spectrophotometer are determined by  the anticipated
use O'L the instrument. Simple, limited programs requiring use of only a few parameters at
gross concentrations, can probably be supported by an inexpensive, but reliable instrument,
such as  the B&L Spectronic 20. (See Table 3-5.) On the other hand, if a laboratory program
requires a wide variety of measurements on  diverse samples at very  low concentrations, a
more  versatile  instrument may be  needed.  One  of  the  prime considerations would be
adaptability to various cell sizes, at least from 1.0 to 5.0 cm. Many spectrophotometers now
available are satisfactory for water quality analyses.

As shown in Table  3-5, which lists  only  a few of the  available spectrophotometers, higher
priced instruments strive for versatility,  including interchangeable sources, detectors, and
cells. Complete information on instrument specifications can be found in the publication
Industrial Research (Beverly Shores, Ind. 46301), Nov. 20,  1969.

3.6.2. Ultraviolet Range

An ultraviolet  spectrophotometer  is similar  in design to  a visual range  instrument,
differences being in the light source  and the optics. The  light source  is a hydrogen or
deuterium  discharge lamp which emits  radiations in the UV portion of the spectrum,*
generally from about 200 m/* to the low visible region. The optical system, if of the prism
type, must be constructed of UV-transparant material, usually quartz.  Sample cells must
also be constructed of quartz or other UV-transparent material. If a grating system is used in
an UV  system,  the  grating may be  specially cut  (blazed) in the  UV region for greater
sensitivity. A number of UV spectrophotometers are available:  the Beckman  DU, the
Hitachi, the Bausch and Lomb, the Gary, the Hilger and the Leeds & Northrup.

3.6.3.  Infrared Range

A number of instrumental changes are required in the construction of spectrophotometers
                                        3-14

-------
for measurements in the infrared  (IR)  region. Modifications  are needed, because optical
materials, such as  glass and  quartz, absorb radiant energy in the IR region and ordinary
photocells do not respond.

Most  IR  spectrophotometers use  front-surfaced mirrors to eliminate  the necessity for
radiant energy to pass through quartz, glass, or other lens materials. The mirrors are usually
parabolic  to facilitate gathering the diffuse IR energy.  Instruments must be protected from
high humidities  and water vapor to avoid deterioration  of the optical system, and also to
avoid extraneous absorption bands in the  IR.

The energy or light source  for an IR instrument may be a Nernst glower or a Globar. Either
source has  certain  characteristics  that  recommend it for use,  but the Globar is  more
commonly used because it has a more stable emission and it is more rugged.  Receiving or
detection  units may be a thermocouple,  a bolometer, thermistor, or a photoconductor cell.

Of all the IR instruments available, two  are in relatively  common use in EPA laboratories.
The Perkin-Elmer Infracord is a low cost model designed for routine work with simplicity of
operation. The Perkin Elmer #621 (723-G Main Ave., Norwalk, Conn., 06852) is a far more
sophisticated instrument, designed for basic research.

3.6.4 Proper Use of Spectrophotometers

The  manufacturer's instructions  for proper use should  be followed in  all cases.  Several
safeguards against misuse of the instruments, however, are mandatory.

Instruments should be checked for  wavelength  alignment. If a particular colored solution is
to be  used  at  a  closely  specified wavelength, considerable  loss  of  sensitivity  can  be
encountered  if  the  wavelength  control  is misaligned. In visual instruments,  an excellent
reference  point is the maximum absorption for a dilute solution of potassium permanganate,
which has a dual peak at 526 mn and 546m/x.  On inexpensive grating instruments, which
possess less resolution than the prism instruments, the permanganate peak appears at 525 to
550 rri|t( as a single flat-topped spike.

For both UV and IR instruments,  standard absorption  curves for many  organic materials
have  been published  so  that reference material  for standard  peaks is easily  available.
Standard  films  of styrene  and other transparent plastics are available for  IR wavelength
checks.

Although most instruments contain built-in transformers for stabilization  of the electronic
circuits, an exterior, high capacity, constant-voltage transformer is recommended for general
laboratory control. A number of controlled-voltage outlets in the  laboratory are especially
desirable in industrial areas or in buildings containing heavy, electrically operated equipment
where voltage surges on adjacent power lines are apt to be frequent. An unstable voltage is
frequently indicated by a flickering needle on  the meter. The flickering  behavior may be
intermittent or it may occur at certain times of the day  when heavy machinery in the area
may be starting or stopping.

Too  much emphasis cannot be placed on care of absorption cells. All cells should be kept
scrupulously  clean, free of scratches, fingerprints,  smudges and  evaporated film residues.
Matched cells should be checked to see that they are equivalent by placing portions of the
                                         3-17

-------
same solution in both cells and taking several readings of the %T or OD values. If a cell is
mismatched it should be discarded or reserved for rough work. (Directions for cleaning cells
are detailed in Chapter 4).

Generally speaking,  trained  technicians  may  operate   any  of the spectrophotometers
successfully.  However, interpretation of  data  from  both the UV and IR instruments
becomes increasingly complex, and requires more training and specialization. IR interpreta-
tion requires special training, and because of the special techniques of sample preparation,
instrument operation, and  interpretation of absorption curves,  mere compliance with the
operations manual is not sufficient.

3.6.5. Atomic Absorption

There are a number of differences in the basic design and accessories for atomic absorption
equipment that  require consideration before purchase and during subsequent use. These
choices  concern  the  light  source, nebulizer burners,  optical systems, readout devices, and
mode conversion. Some  of these choices are not readily obvious, and require that  the
purchaser or user be  familiar with the types and numbers of samples to be analyzed and the
specific  elements to be measured  before a  choice is made. For a program analyzing a wide
variety of samples for a number of elements at varying  concentrations, an instrument of
maximum versatility  would be required. A typical atomic absorption unit is shown in Figure
3-6.

3.6.5.1 Lamp Mounts

A basic  design feature of atomic absorption spectrometers is the convenience of the hollow
cathode (HC) lamp changeover system. Some instruments provide for as many as six lamps
in a  rotating turret,  all electronically stabilized and  ready for use by simply rotating the
lamp turret.  Other instruments provide for  use of only  one  lamp at a time in the lamp
housing, and require  manual removal and replacement whenever more than one element is
to be measured. A quick  changeover system is desirable, especially if  a number  of lamp
changes are needed during a period of operation. Conversely, if lamp changes are infrequent,
multi-lamp mounts do not represent a great convenience.

For optimal use  of the instrument, certain precautions should be observed. After the proper
lamp has been selected the hollow  cathode current  should be  adjusted according to the
manufacturer's recommendations and allowed to electronically  stabilize (warm up) before
use. This requires approximately  15  minutes. During  this  period, the monochromator may
be positioned at the correct  wavelength  and the proper slit  width selected.  For  those
instruments employing a multi-lamp turret,  a warm-up current  is provided to those lamps
not  in  use, thereby  minimizing  the  warm-up  period when the turret is  rotated.  In  a
single-lamp instrument, the instability exhibited during warm-up is minimized by the use of
a double-beam optical system.

3.6.5.1.1 Single Element and Multi-Element Lamps

As an adjunct to single-lamp mounts, HC lamps using from two to as many as six elements
in combination are available, thereby increasing the versatility of the AA spectrometer. For
instance, a single-lamp instrument such as  the Perkin-Elmer may analyze for elements with
only one lamp substitution, whereas a six-lamp  turret such as the Jarrell-Ash (590 Lincoln
Street, Waltham, Mass. 02154), using multi-element lamps  could run 18 or more elements.
                                        3-18

-------
                                                  o
                                                  H
                                                  (X
                                                  Pi
                                                  O
                                                  y

                                                  §
                                                  O
                                                  H
                                                  rn

                                                   OJ
                                                   tH
3-19

-------
Multi-element lamps are  considerably cheaper per element than single-element lamps, but
the savings may not realized if the lamps are not used strategically, because all the elements
in the cathode deteriorate when the lamp is used, regardless of which element is measured.
The  deterioration phenomena results  from the different volatilities of metals used in the
cathode. One metal volatilizes (sputters) more rapidly than the others and redeposits upon
the cathode causing an increase in surface area of that metal, and decreasing the exposed
area of  the  other cathode metals.  Thus, with continual use, a drift in signal will be noted
with at  least one metal increasing and  the other (or others) decreasing. If one can ignore the
dubious cost savings of multi-element lamps, use of single-element lamps would result in
more precise and accurate data.

The  individual line intensities of an element in a multi-element HC lamp will usually be less
than that of a lamp containing a  pure cathode of the same element. This is because  each
element must now share  the discharge energy with all other elements present. However, this
reduction should not affect the output by a factor of more than 1/2 to 1/6, depending on
the combination and number of elements combined. The output can be even greater in some
multi-element lamps because alloying may permit  a higher operating current than the pure
cathode itself. All HC lamps have life  expectancies which are related to the volatility of the
cathode metal, and for this reason, the manufacturer's recommendations for amperage at
which the lamp is operated should be closely followed.

A recent advance in HC lamp design, the high-intensity lamp, promises increased sensitivity
for some elements. It is also predicted that the newly designed lamps will be used in atomic
fluorescence techniques, with significant gains in sensitivity for metals analysis.

Recent improvements in  design and manufacture of hollow cathode lamps have  resulted in
lamps with more constant output and a longer life. Under normal conditions a HC lamp may
be expected to operate satisfactorily for several years. At  one time, hollow  cathode lamps
were guaranteed for a minimum amp-hour period.  This has been  changed, however, to a
90-day warranty. It is good practice to date newly purchased lamps and inspect immediately
upon receipt. Operating current and voltage will be indicated on the lamp and should not be
exceeded during use. An increase in background noise and/or a loss of sensitivity are signs of
lamp deterioration.

3.6.5.2 Burner Types

The most difficult and inefficient step in the AA process is converting  the metal in the
sample  from an ion or a molecule to the  neutral atomic stats. It is the function of the
atomizer and the burner to produce the desired neutral  atomic  condition of the elements.
With minor modifications burners are the same as those used for flame photometry.

Basically there are two  different  types  of burners. They are the total-consumption or
surface-mix burner, and  the laminar-flow or pre-mix burner. There are many variations of
these two basic types,  such as the  Boling, the high-solids,  the turbulent-flow, the tri-flame,
nitrous-oxide  burner  and many others. As one might expect, there  are  many  similarities
among the various burners, the different names resulting from the different  manufacturers.
The element being determined and the type of sample solution dictate the  type of burner to
be used.


Generally,  all types  and makes  of  burners can be adjusted  laterally,  rotationally and
                                         3-20

-------
vertically  for selection of the most sensitive absorbing area of the flame for the specific
element sought. The vertical adjustment is probably the most important since the position
of greatest sensitivity varies from element to element.

Burner height is of utmost importance in changing from one element to another. Certain
instruments are  provided with a vernier adjustment for reproducing burner-height settings,
but many are not. Figure 3-7 shows a simply designed device used in the Analytical Quality
Control Laboratory for reproducing exact burner height. The gauge is  positioned on the
burner and  the  height determined  from the light beam striking the calibrated scale. The
point at which the beam strikes the gauge is recorded for future use.

3.6.5.3 Single-Beam and Double- (Split) Beam Instruments

There is a great deal  of existing uncertainty  among  instrument users about  the relative
merits of single-beam and double-beam instruments. Neither system is the final answer.

With  a single-beam instrument the  light beam from  the source passes directly through the
flame to the detector. In a double-beam system the light from the source is  divided by a
beam splitter into two paths. One path, the reference beam, goes directly to the detector.
The  second path,  the sample  beam, goes through the flame to  the detector. The chopper
alternately reflects and passes each  beam, creating two equal beams falling alternately upon
the detector. If  the beams are equal they cancel the alternate impulses reaching the detector
and no  signal is generated.  If the beams are different, the resulting imbalance causes the
detector to generate an a.c. signal which is amplified  and measured. Any difference between
the reference  and sample  beam is measured as  a direct function  of  absorbed light. The
advantage of the double-beam design, therefore, is that any variations  in the source  are of
reduced importance, and smaller dependence  is  placed upon the stability of the  power
supply.  However, stabilization of the power supply can eliminate the apparent need for the
split-beam  system.  Further, a beam splitter requires use of additional mirrors or  optical
accessories that  cause some loss  of radiant energy. Neither system, however,  compensates
for variation in flame intensity.

A single-beam  system  does  not  monitor  source  variations  but offers  certain   other
advantages.  It allows use of low-intensity  lamps, smaller slit settings and smaller gain. As a
consequence, the single-beam  instrument,  properly  designed, is capable of operating with
lower noise,  better signal-to-noise ratio  and therefore  better precision  and  improved
sensitivity. Because the simplified optical  system  conserves radiant energy, especially in the
shorter  wavelengths,  it  facilitates  operation  in the  low  wavelength range. With this
advantage, it should be possible to  obtain better  sensitivity  for those elements with  strong
resonance lines below 350 m/i  and even those slightly below  300  m^.

3.6.5.4. Readout Devices

Early models  of the AA  instrument offered  only  a meter,  calibrated  in  percentage
absorption.  In  the surge of competitive design,  more sophisticated readout devices were
built  into  or  offered as accessories  to  various models. At the present time  any  desired
readout method may be obtained with almost  any instrument. Less expensive designs still
provide  meters  with conventional  needle indicators.  More costly instruments offer any
combination of  built-in digital sealers,  calibrated  in concentration, external digital printout
in concentration,  typewriter printout or  typewriter with  punch  tape. Even  inexpensive
instruments are built with recorder interfacing.
                                         3-21

-------
        SMALL CENTIMETER GAUGE
             CIRCULAR
               BASE
BEAM FROM HOLLOW CATHODE LAMP

                                  >>
                                               o
  Figure 3-7 DEVICE FOR REPRODUCIBLE POSITIONING OF BURNER HEIGHT
                         3-22

-------
Choice of a readout system is predicated largely upon laboratory needs and availability of
budget. In general, any step toward complete  automation is desirable but the degree of
automation should be compatible with the laboratory program.

3.6.5.5 Miscellaneous Accessories

A number of instruments contain a mode selector, making an instrument usable for either
flame absorption or flame emission. The  conversion to flame emission is a desirable feature
since certain elements are more amenable to analysis by this method.

Automatic sample changers are offered for almost all instruments on the market, and as has
been previously stated, any automation  feature is desirable. However, unless a laboratory
program  performs a large number of repetitious measurements daily, an  automatic  sample
changer would not  be required.  As a practical measure, other commonly used sample-
changing devices not expressly designed for AA use, can easily be interfaced with almost any
AA instrument.

3.6.5.6  Instrument Choice

Table  3-6  summarizes  some  of  the design features of  various  commercially available
instruments.  Since  many of the  features are common to most models, the basic choice
appears to be  between double-beam and single-beam instruments. On  this basis,  at this
writing, it is probable that  the most widely used instrument in the EPA is the Perkin-Elmer
303, one of two double-beam instruments on the market.

The  Instrumentation Laboratories Model  153 (113 Hartwell Ave., Lexington, Mass. 02173)
has a number of desirable  features. In addition to features  listed in the table, it includes
push-button ignition, push-button wavelength scan, fail-safe solenoids to  prevent improper
flame settings,  fail-safe flame monitoring, absorption integrator at selected time intervals,
visual "peaking" meters, curve correction and several other features. The  new Perkin-Elmer
Model 403 also offers most of these features in addition to very sophisticated data handling
accessories.

3.7 Organic Carbon Analyzer

A number of devices designed to measure the organic content of  aqueous samples  have
appeared on the market within the last five years. The oldest, or first of these instruments, is
the Dow Beckman  Carbonaceous Analyzer (Figure 3-8).  The apparatus  measures organic
carbon as carbon, by oxidizing a very small sample at a temperature of 900° , in a stream of
oxygen, converting  all organically bound  carbon to carbon dioxide, which is then measured
by a Golay-type thermal detector. The instrument is able to detect about  20 mg.  of carbon.
A recent modification of  the  original Dow Beckman Carbonaceous Analyzer employs a
dual-combustion-tube system operating  at  different temperatures to distinguish between
total carbon and inorganic carbon. Organic carbon is found by difference.

Another  such instrument, also developed by Dow Chemical Company, is marketed  by the
Fisher Scientific Company (Instrument Division, 711 Forbes Ave., Pittsburgh, Pa. 15219) as
the "Aquarator". The principle of operation is similar to the Dow Beckman Carbonaceous
Analyzer, except that carbon dioxide fed through the combustion tube is reduced to carbon
monoxide and measured in an infrared detector. The reaction does not measure carbon, per
j>e, but reducing materials. The results can be equated with the COD test, but the instrument
                                        3-23

-------
is best suited for  industrial wastes and sewage. It lacks the  sensitivity needed for COD
measurements on relatively clean waters. The lower limit of measurement is about 40 mg/1
COD.

A third instrument developed by Dow researchers is franchised to Ionics Corporation (165
Grove Street, Watertown, Mass. 02172) and sold as the "TOD Analyzer". This instrument is
also similar in principle to the  carbon analyzer, except that the device measures oxygen in
and out of the combustion tube. Depletion of oxygen is correlated with oxygen demand of
the sample. However, the reaction of oxygen with sulfur and nitrogen is not stoichiometric,
and results for some samples may be questionable. A malfunction of the sample injection
system appears to  be the main problem with the instrument to date. An evaluation study
conducted by the Hudson Delaware Basins Office, EPA, indicates a very good potential use
for the instrument.

Another  device, designed  by Union  Carbide  Corporation (Ionics Inc.,  65  Grove  St.,
Watertown, Mass. 02172) is similar to  the original Dow Beckman Analyzer except that the
combustion tube contains a coil of heated palladium wire. An aqueous sample injected into
the combustion tube is decomposed to  hydrogen and oxygen through the catalytic action of
the palladium. Carbon  from organic matter combines with the water-produced oxygen to
form carbon  dioxide, with a final infrared measurement.  Results obtained are theoretically
identical  to the Dow Beckman Analyzer,  but the repeated failure of the sample injection
system has prevented evaluation of the apparatus  as a laboratory  instrument. It is being
routinely used, however, as a continuous monitoring system for total carbon.

To date the only instrument which has been successfully demonstrated and wmch meets the
needs of EPA laboratories is the Dow Beckman Carbonaceous Analyzer or its successor, the
Model 915. Operating instructions for this device are detailed elsewhere (4).

Precautions to be observed in use of the Dow Beckman Carbonaceous Analyzer or the DB
Model 915 are detailed below:

    a.   All inorganic carbon present in the  sample as carbonate or bicarbonate must be
        removed prior to analysis, or accounted for, when using the Model 915.

    b.   Particulate  matter in  the  sample should not be larger than the  opening of the
        hypodermic needle used to take the  sample. If necessary the sample should be
        homogenized by some means designed to reduce the particle  size.

    c.   The  instrument should be preconditioned with repeated injections of distilled water
        to obtain stable operating conditions  before actual sample injection.

    d.   Furnace temperature  and  oxygen flow should  be maintained at  proper settings
        during period of operation.

    e.   Perform periodic maintenance  of instrument as follows:

        1.  repack combustion tube with fresh asbestos,
        2.  clean and dry micro filter,
        3.  clean and dry infrared cell,
        4.  recharge reference cell with nitrogen.

Need for replacement  of asbestos packing is indicated  by a  loss of sensitivity and wide
                                        3-24

-------
                                                     OS!
                                                     .-1
                                                     z
                                                     o
                                                     CQ
                                                     U
                                                     O

                                                     oi

                                                     O


                                                     op

                                                     CO
                                                      oo
3-27

-------
peaks.  Unsteady baseline and excessive noise is caused by a dirty or partly clogged micro
filter. An accumulation of moisture in the infrared cell is indicated by loss of sensitivity and
excessive noise.

3.8 Selective Ion Electrodes

In recent years a variety of ion-selective electrodes have been commercially available, and
show great promise as fast and efficient tools  for m situ monitoring and for laboratory
analysis of all types of samples. A list of ions susceptible to analysis is indicative of the
interest and progress being made in this field. Electrodes, or "probes" as they are popularly
called, are available for measurement of monovalent cations, sulfate, nitrate, perchlorate,
and  a  number of  others. Dissolved  oxygen  probes  should also be included  in this list,
although they are not  technically selective ion probes. There have been a number of reviews
concerning the theory and application of probes, most recently a concise article by Rechnitz
(5).

Selective ion probes generally  measure what  they claim to measure—specific ion activity.
They do not measure concentrations  of un-ionized  materials.  For example, the probe
designed to measure divalent cations promises  to measure total hardness as a  function of
calcium and magnesium. However, since the probe does not respond to un-ionized calcium
and  magnesium, it  does not accurately measure total hardness.  As a consequence of this
deficiency, much investigation is being carried out to devise means of determining total
concentration of the constituent sought.

"Methods  for Chemical Analysis of Water and Wastes" (4) outlined procedures for use of
two probes, fluoride and DO. The chloride probe is also used on the automatic monitoring
devices and will be listed as an approved procedure in future editions  of EPA Methods.
Evaluation of additional ion selective electrodes for cyanide, ammonium, and sulfide ions is
now underway, and will be reported in the near future. Various techniques for use of the
probes are reviewed by Riseman (6).

Personnel in  water  pollution laboratories are encouraged to investigate the use of selective
ion  probes as a means of reducing analytical work and  improving data quality. At the
present time, however, only the fluoride, DO, and chloride  probes are  recommended for
routine use in data collection.

A basic question relating to the use of selective ion probes is the number of standards
required to prepare a standard curve. It is generally agreed  that the more standards used for
the preparation of  a colorimetric curve, the more reliable the data resulting from use of the
curve.  On  the other hand, only one  or two points are normally used in standardization of a
meter for measuring pH, conductivity or DO. In the conversion of a colorimetric procedure
such as  fluoride to a  probe-type measurement, the tendency is to prepare a  millivolt vs
concentration curve using the usual  six to ten standards. Probe manufacturers insist that
only one or two points are  needed since the linearity of response has been established and
only the slope of the line must be known. The alternatives are:

     a.   Take readings  in distilled water at  0  and at  some concentration approximating
        concentration of sample; establish slope of line.

     b.  Take readings  in distilled water at 0 and two  concentrations bracketing expected
        concentration of unknowns to establish curve or slope.
                                         3-28

-------
    c.   Take readings at 0  and at decade concentrations as 1,  10, 100, etc., for standard
        curve.

    d.   Take reading of sample, add known increment of measured constituent and read
        again; establish proportionality factor.

The method for the use of the fluoride electrode specifies use of multiple standards in the
range between  0 to 2 mg/liter, because this system has supplied very precise data when
compared to the colorimetric methods using the  same set of standards. The system of
incremental addition  appears  to have considerable  merit since the  electrode response  is
established in the presence of possible interferences. At the present time no single procedure
for standardizing probe response has been adopted by a majority of users.

When a selective ion electrode appears to be malfunctioning, the same check system may be
used as for a faulty glass pH  electrode. It is unlikely, however, that the electrode  will be
cracked; it will probably be dry, or insufficiently filled with  the necessary  solution. The
piobe assembly and instructions for refilling customarily  accompany the item  when shipped
by the manufacturer and said instructions should be followed by the user.

Selective ion probes are available from several manufacturers including Beekman, Corning,
Coleman and Orion. The Orion organization (11 Blackstone St., Cambridge, Mass. 02139) is
the largest producer  in the field and offers  not only  selective ion probes  but a  sizable
complement of electronic equipment for use with the probes. Figure 3-9 illustrates an Orion
selective ion meter.

Dissolved oxygen probes and meters  of various designs have been offered by a large number
of manufacturers including Weston & Stack, Beekman, Jarrell-Ash, Union Carbide, Yellow
Springs Instrument, Delta Scientific, and others.
3.9  References

1.  1968  Book of ASTM Standards, Part 30; Testing  Single Arm Balances, pp.  1071-84:
    American Society for Testing and Materials, 1916 Race Street, Philadelphia, Pa.  19103.

2.  1968  Book of ASTM Standards, Part 23; Test for pH of Industrial Waste Water, pp.
    292-3; American Society for Testing and Materials, 1916 Race Street, Philadelphia, Pa.
    19103.

3.  Black, A. P., and Hannah, S. A., "Measurement of Low Turbidities," JAWWA, jT7, 901
    (1965).

4.  Methods for Chemical Analysis of Water and Wastes,  EPA, Analytical Quality Control
    Laboratory, 1971.

5.  Rechnitz, G. A., "Ion Selective Electrodes," Chemical Engineering News, p. 146, June
    12,  1967.

6.  Riseman,  Jean M.,  "Measurement  of Inorganic Water  Pollutants by Specific  Ion
    Electrode," American Laboratory, p. 32, July 1969.
                                        3-29

-------
                                                    W
                                                    I-H

                                                    U
                                                    w
                                                    _l
                                                    w
                                                    1/5

                                                    cK
                                                    CO
3-30

-------
                                     CHAPTER 4

                                     GLASSWARE

4.1  General

The measurement of trace constituents in water  demands methods capable of maximum
sensitivity.  This is especially true for metals and trace organics such as pesticides, as well as
for the  determination  of ammonia  and phosphorus.  In  addition to sensitive  methods,
however, there are other areas that require special consideration. One such area is that of the
cleanliness  of laboratory glassware. Obviously, the very sensitive analytical systems are more
sensitive to errors resulting from  the improper use or  choice  of apparatus, as well as to
contamination effects due to an improper method of cleaning the apparatus. The purpose of
this chapter is  to discuss the kinds of glassware available, the use of volumetric ware, and
various cleaning requirements.

4.2 Types of Glassware

Laboratory vessels serve  three  functions: storage of reagents, measurement  of solution
volumes and confinement of reactions. For special purposes, vessels made from materials
such as  porcelain, nickel, iron,  aluminum,  platinum,  stainless steel, and plastic may be
employed  to advantage. Glass, however, is the most widely used material of construction.
There are many grades and types  of glassware from which to choose, ranging from student
grade to others possessing specific properties such as resistance to thermal shock, alkali, low
boron content, and super strength. Soft-glass containers are  usually relatively soluble, and
therefore are not recommended  for general use, especially for storage of reagents.  The
mainstay of the modern analytical laboratory is a highly resistant borosilicate glass, such as
that manufactured by Corning Glass Works under the name "Pyrex" or by Kimble Glass Co.
as  "Kimax". This glassware  is  satisfactory  for  all analyses included in  "Methods for
Chemical Analysis of Water and Wastes (1)."

Depending on the particular manufacturer, various trade names are used for specific brands
possessing  special properties such as  resistance  to heat, shock, alkalies, etc. Examples of
some of these follow:

    a.  Kimax- or Pyrex-brand glass is a relatively inert all-purpose borosilicate glass.

    b.  Vycor-brand glass is a silica glass (96%) made to withstand continuous temperatures
        up to 900°C and can be down-shocked in ice water without breakage.

    c.  Corning  brand  glass is  claimed to be  50 times more resistant to alkalies than
        conventional ware and practically boron free (max. 0.2%).

    d.  Ray-Sorb- or Low-Actinic-brand glass is for use with light-sensitive material.

    e.  Corex-brand labware is harder than conventional borosilicates, better able to resist
        clouding and scratching.

The use of plastic vessels, containers and other apparatus made of Teflon, polyethylene,
polystyrene and  polyproplylene has increased markedly over recent years. Some of these
materials,  such  as Teflon, are  quite expensive; however,  Teflon stopcock  plugs have
                                         4-1

-------
practically replaced glass plugs in burets, separately funnels, etc.  because lubrication to
avoid sticking or "freezing" is not required. Polyprolylene, a methylpentene polymer,  is
available as laboratory  bottles, graduates, beakers and even volumetric flasks. It is crystal
clear, shatter-proof, autoclavable and chemically resistant.

Some points to consider in choosing glassware and/or plasticware are:

    a.  Generally,  the  special types  of glass listed above are not required to perform the
        analyses given in "Methods for Chemical Analysis of Water and Wastes" (1).

    b.  Unless instructed otherwise, borosilicate or polyethylene bottles are to be used for
        the storage of reagents and standard solutions.

    c.  Certain dilute  metal solutions  may plate out on  glass container  walls over long
        periods of storage.  Thus,  dilute metal standard solutions are prepared fresh at the
        time of analysis.

    d.  For some operations, disposable glassware is entirely satisfactory.  One example  is
        the use of disposable test tubes as  sample containers for use  with the Technicon
        Automatic Sampler.

    e.  Plastic bottles of polyethylene  and/or Teflon have been found satisfactory for the
        shipment of water samples. Strong mineral acids (such as sulfuric acid) and organic
        solvents will readily attack polyethylene and are to be avoided.

    f.  Borosilicate glassware is  not completely  inert,  particularly to alkalies; therefore,
        standard solutions  of silica, boron and the alkali metals  are usually stored in
        polyethylene bottles.

For additional information the reader  is referred  to the catalogs of the various glass and
plastic  manufacturers.  These catalogs   contain  a wealth  of information as to specific
properties, uses, sizes, etc.

4.3  Volumetric Analyses

By common usage, accurately calibrated glassware for precise measurements of volume has
become known as volumetric glassware.  This group includes volumetric flasks, volumetric
pipets and accurately calibrated burets.  Less accurate types of glassware including graduated
cylinders,  serological  and  measuring  pipets  also have specific uses in the  analytical
laboratory,  when exact volumes are  unnecessary.  A  typical laboratory glassware setup  is
shown in Figure 4-1.

The precision  of volumetric work  depends in part upon the accuracy with which volumes of
solutions  can  be measured  and there are certain sources of error which must be carefully
considered.  The volumetric apparatus must  be read  correctly; that  is, the bottom of the
meniscus  should be tangent  to the  calibration mark. There are other sources of error,
however,  such as changes in temperature which result in changes  in  the actual capacity of
glass  apparatus and in the volume of the solutions.The capacity of an ordinary glass flask of
1000 ml  volume increases  0.025 ml  per  degree rise  in  temperature,  but if made of
borosilicate glass the increase is much  less.  One  thousand ml of water or of most 0.1  N
solutions  increases  in  volume by  approximately 0.20 ml per  1°C increase  at room
                                         4-2

-------
                                                        ffi
                                                        u
                                                        z
                                                        w
                                                        PQ


                                                        O
                                                        HH
                                                        H
                                                        H

                                                        H
                                                        u.


                                                        60
4-3

-------
temperature. Thus solutions must be measured at the temperature at which the apparatus
was calibrated. This temperature (usually 20° C) will be indicated on all volumetric ware.
There may also be errors of calibration of the apparatus; that is, the volume marked on the
apparatus may not be the true volume. Such  errors can be eliminated only by recalibrating
the apparatus or by replacing it.

Volumetric apparatus is  calibrated "to contain" or "to deliver" a definite volume of liquid.
This will  be  indicated on the  apparatus with the letters "TC" (to contain) or "TD" (to
deliver). Volumetric flasks  are calibrated to contain  a given volume. They are available in
various shapes and sizes ranging from 1-to 2000-ml capacity.

Volumetric  pipets are calibrated to deliver a fixed volume. The usual capacities are 1 thru
100 ml although micro-pipets are also available. In emptying volumetric pipets, they should
be held in a vertical position and the outflow should  be unrestricted. The tip of the pipet is
kept in contact with the wall of the receiving vessel for  a second or two after the free flow
has stopped. The liquid remaining in the tip is not removed; this is most important.

Measuring and serological pipets should also  be held in  a vertical position  for dispensing
liquids; however, the tip of the pipet is  only touched to the wet surface of the receiving
vessel after the outflow has ceased. For those pipets where the  small amount of liquid
remaining in the tip  is to be blown out and added, indication  is made by a frosted band near
the top.  The band is  usually located far enough  down so that it will not touch the
technician's lips when liquid is being drawn up or blown out.

Burets are used to deliver definite volumes. The more common types are usually of 25- or
50-ml capacity,  graduated to tenths  of a milliliter, and  are provided with stopcocks. For
precise  analytical methods in microchemistry, micro-burets are  also used. ^Micro-burets
generally  are  of 5-  or   10-ml  capacity,  graduated in hundredths  of a milliliter division.
Automatic burets with reservoirs are also  available ranging in capacity from  10 to  100 ml.
Reservoir capacity ranges from 100 to 4000 ml.

General rules in regard to the manipulation of a buret are as follows: Do not attempt to dry
a buret which has been cleaned for use, but rinse it two or three times with a small volume
of the solution  with which it  is to be filled.  Do not allow alkaline solutions to stand in a
buret, because the glass  will be attacked, and the stopcock, unless made of Teflon, will tend
to freeze. A 50-ml buret should not be emptied faster than 0.7 ml per second, otherwise too
much liquid will adhere to the walls and as  the solution drains down, the meniscus will
gradually  rise, giving a  high false  reading. It  should  be emphasized that improper use of
and/or reading of burets  can result  in serious calculation errors.

In the case of all apparatus for delivering liquids, the glass must be absolutely clean so that
the film of liquid never  breaks at  any point. Careful  attention must be paid  to this fact or
the required amount of  solution will not  be delivered. The various cleaning agents and their
use are described later.

4.4 Federal Specifications for Volumetric Glassware

Circular 602 of the National Bureau of Standards, "Testing of Glass  Volumetric Apparatus",
describes  the Federal  Specifications  for volumetric glassware. The National Bureau  of
Standards no longer accepts stock  quantities of volumetric apparatus from manufacturers or
dealers for certification and return for future sale to  consumers. This certification service is
                                         4-4

-------
still available, but apparatus will be tested only when submitted by the ultimate user, and
then only after an agreement has been reached with the Bureau concerning the work to be
done.

Consequently, the various glass manufacturers have discontinued the listing of NBS-certified
ware. In its place catalogue listings of volumetric glass apparatus which meet the Federal
Specifications are designated as Class A and all such glassware is permanently marked with a
large "A". These NBS specifications are listed in Table  4-1. The ware in question includes
the usual burets, volumetric flasks and volumetric pipets.

                                      Table 4-1

                  TOLERANCES FOR VOLUMETRIC GLASSWARE
                (Abridged from National Bureau of Standards Data, 1941)

                 Capacity (ml)
            less than and including                Limit of error (ml)

                                  Graduated Flasks

                      25                              0.03
                      50                              0.05
                     100                              0.08
                     200                              0.10
                     250                              0.11
                     300                              0.12
                     500                              0.15
                   1,000                              0.30
                   2,000                              0.50

                                   Transfer pipets

                       2                              0.006
                       5                              0.01
                      10                              0.02
                      25                              0.025
                      30                              0.03
                      50                              0.05
                     100                              0.08
                     200                              0.10

                                      Burets1

                       5                              0.01
                      10                              0.02
                      30                              0.03
                      50                              0.05
                     100                              0.10

 Limits of error are of total or partial capacity. Customary practice is to test the capacity at 5 intervals.
                                        4-5

-------
In addition to the "A" marking found on calibrated glassware and the temperature at which
the calibration was made, other markings also appear. These include the type of glass, such
as Pyrex, Corex, Kimax,  etc., the stock number of the particular item,  and the capacity of
the vessel. If the vessel contains a ground-glass connection, this will also be included along
with the TD or TC symbol. An example of the markings usually found on volumetric ware is
shown in Figure 4-2.
PYREX
USA *

A <
ft *•
Sflf) ml ~*
TC 20°C
Mn Rfiftl


") *-
-J 16

"0.20ml <-

GLASS CO.
TYPE
STANDARD
	 TAPER 	
SIZE
CLASS
VO LU ME
TO
CONTAIN
STOCK
KIMAX
' USA
* 19 \
> A
* ft
> snn r
* JUU '
» TP °fl
' 1 U LU
-. MH 9


nl
°C
ftftU
               Figure 4-2. EXAMPLE OF MARKINGS ON GLASSWARE

Class  A  glassware need not be  recalibrated before use. However,  if it should become
necessary to calibrate a particular piece of glassware, directions may be found in texts (2) on
quantitative analysis.

4.5  Cleaning of Glass and Porcelain

The method of cleaning should be adapted to both the substances that are to be removed,
and the  determination to be performed. Water-soluble  substances are simply washed out
with hot or cold water, and the  vessel is finally rinsed with successive small amounts of
distilled  water.  Other substances more difficult  to remove may require  the use  of a
detergent,  organic solvent,  dichromate  cleaning solution, nitric acid or aqua  regia (25
percent v/v cone. HNO3  in cone.  HC1). In all cases it is good practice to rinse a vessel with
tap water as soon as possible after use. Material allowed to dry on glassware is much more
difficult  to remove.

Volumetric glassware, especially burets, may be thoroughly cleaned by a mixture  containing
the following:  30  g sodium hydroxide, 4 g  sodium  hexametaphosphate  (trade name,
Calgon), 8 g trisodium phosphate, and  1 liter water. A gram or two of sodium lauryl sulfate
or other surfactant will improve its action in some cases. This solution  should be used with a
buret brush.

Dichromate cleaning solution (chromic acid) is a powerful cleaning agent; however, due to
its  destructive nature upon  clothing and upon laboratory furniture, extreme care must be
taken when using this  mixture.  If any of the solution is spilled, it  must be  cleaned up
                                         4-6

-------
immediately. Chromic acid solution may be prepared in the laboratory by adding 1 liter of
concentrated sulfuric  acid slowly,  with stirring, to  35  ml saturated sodium dichromate
solution. This mixture must be allowed to stand for approximately 15 minutes in the vessel
which is  being cleaned  and may  then be returned to a  storage bottle.  Following the
chromic-acid wash, the  vessels are rinsed  thoroughly with  tap  water,  then with  small
successive  portions of  distilled  water. Fuming  nitric acid  acts more  rapidly, but  is
disagreeable to handle. In either case, when the acid becomes dilute, the cleaning mixture is
no longer effective. A mixture of concentrated sulfuric and fuming nitric acids is even more
efficient but is also hazardous to  use. A persistent greasy layer or spot may be removed by
acetone or by allowing a warm solution of sodium hydroxide, about 1  g per 50 ml of water,
to stand in the vessel for 10-15 minutes; after rinsing with water, dilute hydrochloric acid,
and water again, the vessel is usually clean. Alcoholic potassium hydroxide is also effective
in removing grease. To dry glass apparatus, rinse with acetone and blow or draw air through
it.

4.6  Special Cleaning Requirements

Absorption  cells, used in  spectrophotometers, should be  kept scrupulously clean, free of
scratches, fingerprints, smudges and evaporated film residues. The cells may be  cleaned with
detergent  solutions for removal of organic residues, but should not be soaked for prolonged
periods in caustic solutions  because of the possibility of etching. Organic solvents may be
used  to rinse  cells in  which  organic materials have  been  used. Nitric acid rinses are
permissible,  but  dichromate solutions  are not  recommended because of  the  adsorptive
properties of dichromate on glass.  Rinsing and drying  of cells with alcohol or acetone before
storage is  a preferred  practice. Matched cells  should  be checked  to see that they are
equivalent by placing portions of  the same solution in both cells and taking several readings
of the transmittance (%T) or optical density (OD) values. If a cell is mismatched it should be
discarded or reserved for rough work.

Hor certain determinations, especially trace metals, the glassware should also be rinsed with
a 1:1  nitric acid-water mixture. This operation is followed by thoroughly rinsing with tap
water and successive portions of distilled water.  This may require as many as  12-15 rinses,
especially if chromium is being determined. The nitric acid rinse is also especially important
if lead is being determined.

Glassware to be used for phosphate determinations should not be washed with detergents
containing phosphates.  This glassware must  be  thoroughly  rinsed  with tap  water and
distilled water. For ammonia and  Kjeldahl nitrogen, the glassware must be rinsed with
ammonia-free water (See Chapter 2).

Glassware to be used in  the  determination of trace  organic constituents in water, such as
chlorinated pesticides, should be as free as possible of organic contaminants. A chromic acid
wash  of at least 15 minutes is necessary to destroy these organic residues. Rinse thoroughly
with tap water, and finally with distilled water. Glassware may be dried for immediate use
by rinsing with redistilled acetone.  Otherwise glassware may  be oven dried or drip dried.
Glassware should be stored immediately after drying to prevent any accumulation of dust.
Store inverted or with mouth of glassware covered with foil.

Bottles to be used for  the collection  of samples for organic analyses should be rinsed
successively with chromic acid cleaning solution, tap water,  distilled water, and finally
several times with  redistilled solvent (e.g., acetone, hexane, petroleum ether, chloroform).
                                          4-7

-------
Caps are washed with detergent, rinsed with tap water, distilled water and solvent. Liners are
treated in the same way as the bottles and are stored in a sealed container.

4.7 Disposable Glassware

When  the risk  of washing a pipet for reuse becomes too great, as in the case of use with
toxic materials, or when  the  cost  of washing glassware becomes prohibitive, disposable
pipets may be the answer, provided they meet the necessary specification. Various types are
available including bacteriological, serological and micro-dilution  pipets. Disposable glass-
ware generally is made of soft glass.

4.8 Specialized Glassware

The use  of vessels and glassware fitted with standard-taper, ground-glass, and ball-and-socket
joints has increased because of certain advantages such as less leakage and fewer freezeups.
Standard-taper, interchangeable ground joints save time and trouble in assembling apparatus.
They are precision-ground  with tested abrasives to insure an accurate fit and freedom from
leakage.  Ball and  socket joints increase flexibility of operation  and eliminate the need for
exact alignments of apparatus. Symbols  and their meaning as  applied to standard joints,
stoppers and stopcocks are shown below.

4,8.1. Standard Taper (J)

| is the  symbol used  to designate interchangeable joints, stoppers and stopcocks, complying
with the requirements of Commercial Standard CS-21, published by the National Bureau of
Standards. All mating parts are finished to a 1:10 taper.

The size of a particular piece appears after the appropriate symbol. Due primarily to the
greater variety of apparatus equipped with J fittings,  a number of different types of
identifications are used, as follows:

     a.   For joints—a two-part number, as J 24/40, with 24 being the approximate diameter
         in mm at the large end of the taper, and 40 the axial length of taper, also in mm.

     b,   For  stopcocks—a single number, as J  2, with 2 mm being the approximate diameter
         of the hole or holes through the plug.

     c.   For  bottles—a single number, as J  19, with 19 mm being the appropriate diameter
         at top of neck. However, there are differences in dimensions between the bottle and
         flask stoppers.

     d.   For  flasks, etc. — a single number, as J  19, with 19 mm being the appropriate
        diameter of the opening at top of neck.

4.8.2 Spherical Joints (f)

5 is the designation  for spherical (semi-ball) joints  complying  with  CS-21. The complete
designation of a spherical joint also  consists of a two-part number, as 12/2, with 12 being
the Approximate  diameter  of the ball and 2 the bore of the ball  and the socket, also in mm.
                                         4-8

-------
4.8.3 Product Standard (g)

s is a new symbol. It will appear in a forthcoming NBS Product Standard for stopcocks with
Teflon plugs, with the mating surfaces being finished to a 1:5 taper. As with J stopcocks, a
single number is used. Thus, £  2 means a Teflon  stopcock with a hole of approximately
2-mm diameter in the plug.

4.9  Fritted Ware

For certain laboratory operations the use of Fritted Ware for filtration (as in total dissolved
solids and suspended solids determinations), gas  dispersion, absorption, and/or extractions
may be of an advantage.

There are six different porosities of Fritted Ware available, so that precipitates varying in
size can be filtered at maximum speed with no sacrifice or retentivity. Porosity is controlled
in manufacture, and discs are individually tested and graded into  these classifications. The
extra-coarse and coarse porosities are held toward the maximum pore diameter as listed. The
medium, fine, very fine, and ultra-fine are held toward the minimum pore diameter as listed'
in Table 4-2.
                                      Table 4-2

                             FRITTED WARE POROSITY
     Porosity
     Grade

   Extra Coarse
   Coarse



   Medium

   Fine

   Very Fine

   Ultra-Fine
Designation

     EC
     M

     F

     VF

     UF
Pore Size
(Microns)          Principal Uses

170-220      Coarse filtration. Gas
              dispersion, washing,
              absorption.

 40-60        Coarse filtration. Gas
              dispersion, washing,
              absorption.

 10-15        Filtration and extraction.

 4-5.5        Filtration and extractioni

 2-2.5        General bacterial filtration.

0.9-1.4        General bacterial filtration^
Pore sizes are determined by the method specified  in  ASTM  E  128,  "Maximum
Diameter and Permeability of Rigid Porous Filters for Laboratory Use." (3)
                                                        Pore
                                         4-9

-------
4.9.1 Recommended Procedures for Maximum Filter Life

    a.   New Filters. Wash new filters by suction with hot hydrochloric acid, followed by a
        water rinse.

    b.   Pressure Limits. The maximum safe differential pressure on a disc is 15 pounds per
        square inch.

    c.   Thermal Shock. Fritted ware has less resistance to thermal shock than non-porous
        glassware.  Hence, excessive, rapid  temperature changes and  direct exposure to a
        flame should be avoided. Heating  in a furnace to  500°C  may  be  done safely,
        provided the heating and cooling are gradual. Dry ware may be brought to constant
        weight by heating at 105-110°C.

Never subject a damp  filter of ultra-fine porosity to a sudden temperature change. Steam
produced in the interior may cause cracking.

4.9.2 Cleaning of Used Filters

In many cases, precipitates  can be removed by rinsing with water, passed through from the
underside, with the pressure not exceeding  15 pounds per square inch. The suggestions that
follow in Table 4-3 will be  helpful in dealing with material that will not be removed by the
reverse water-wash.  The use of strong alkalies, strong hydrofluoric  acid and phosphoric acid
should be avoided. Also, scratching  of the surfaces will weaken the discs.
                                      Table 4-3

                              CLEANING OF FILTERS

        Material                                       Removal Agent
     Albumen                           Hot ammonia or hydrochloric acid

     Aluminous and                      2%  Hydrofluoric  acid  followed  by con-
     siliceous residues                    centrated sulfuric acid. Rinse immediately
                                        with water until  no trace of acid can be
                                        detected.

     Copper or iron oxides                Hot  hydrochloric  acid  plus  potassium
                                        chlorate.

     Fatty materials                      Carbon tetrachloride

     Mercuric sulfide                     Hot aqua regia

     Organic matter                      Hot concentrated cleaning solution, or hot
                                        concentrated sulfuric acid with a few drops
                                        of sodium nitrite.

     Silver chloride                      Ammonium or sodium hyposulfite.
                                        4-10

-------
4.10 References

1.   "Methods for Chemical Analysis of Water and Wastes," EPA, Analytical Quality Control
    Laboratory, 1971.

2.   Willare, H.  H., and Furman, N.  H., Elementary Quantitative Analysis—Theory and
    Practice, D. Van Nostrand Co., Inc., New York (1947).

3.   1968 Book of ASTM Standards, Part 30, American Society for Testing and Materials,
    1916 Race Street, Philadelphia, Pa. 19103.
                                      4-11

-------
                                       Chapter 5

                           REAGENTS, SOLVENTS, AND GASES
5.1  Introduction
The objective of this chapter is to provide general information and suggestions  that will
serve to  keep the analyst conscious of his responsibilities in analytical quality control, as
they relate to reagents, solvents and gases. While the material presented here will assist the
analyst in producing high quality data, it is by no means complete. It is incumbent on the
analyst to  obtain details  of special precautions required to insure proper selection,
preparation, and storage  of reagents, solvents and gases from the descriptions of individual
methods.

5.2 Reagent Quality

Chemical reagents, solvents, and  gases are  available in a wide variety of grades of purity,
ranging from technical grade to various "ultra pure" grades. The purity of these materials
required in analytical chemistry  varies with  the  type  of analysis. The parameter being
measured and the sensitivity and specificity of the detection system are important factors in
determining the purity of the reagents required. For many analyses, e.g., most inorganic
analyses, analytical reagent grade is  satisfactory. Other analyses, e.g., trace  organic and
radiological, frequently require special "ultra pure" reagents, solvents, and gases. In methods
where the purity of reagents is not specified it is intended that analytical reagent grade be
used. Reagents of lesser purity than that specified by the method should not be used. The
labels on the  container should be checked and  the  contents examined to verify that the
purity of the  reagents meets  the needs of the particular method involved. The quality of
reagents, solvents, and gases required  for the various classes  of analyses: inorganic, metals,
radiological, and organic,  are discussed below.

Reagents must always be prepared and standardized with the utmost of care and technique,
against reliable primary standards. They must be restandardized or prepared fresh as often as
required by their stability. Stock  and  working standard solutions must be checked regularly
for signs of deterioration, e.g., discoloration, formation of precipitates, and concentration.
Standard solutions should be properly labeled as to compound, concentration, solvent, date,
and preparer.

Primary  standards must  be  obtained  from  a reliable source, pretreated, e.g., dried, under
specified conditions, accurately prepared in calibrated volumetric glassware, and stored in
containers that will not alter the reagent. A large number of primary standards are available
from the National Bureau of Standards (NBS). A complete listing of available  standards is
given in  NBS  Special Publication 260 (1).  Primary standards may  also be obtained from
many chemical supply companies. Suppliers for special quality reagents, solvents, and gases
are noted in later discussions of the various classes of analyses. Reagents and solvents of all
grades are available from many chemical supply houses.

There is some confusion among chemists as to the definition of the terms ANALYTICAL
REAGENT GRADE, REAGENT GRADE, and ACS ANALYTICAL REAGENT GRADE. A
review of the literature  and  chemical supply catalogs indicates that the  three terms are
synonymous. Hereafter,  in this document, the term ANALYTICAL REAGENT  GRADE
                                         5-1

-------
(AR) will  be used.  It is intended that AR chemicals and solvents shall conform to the
current specifications of the Committee on  Analytical Reagents of the American Chemical
Society (2).

The  ASTM Manual on Industrial Water and Industrial Waste Water (3), Part 23  of ASTM
Standards (4), and "Standard Methods for the Examination of Water and Wastewater" (5)
devote separate chapters to problems related to preparation, standardization, and storage of
reagents.  The  information  provided  therein  is  particularly  appropriate  to  inorganic
determinations. The  type of volumetric glassware to be used, the effect of certain reagents
on glassware, the effect  of temperature on volumetric measurements, purity of reagents,
absorption of gases and water vapor from the air, standardization of solutions, instability,
and need for frequent standardization of  certain reagents are among the topics discussed. It
is recommended that the analyst become thoroughly familiar with these publications.

5.2.1 General Inorganic Analyses

In general, AR-grade reagents and solvents are satisfactory for inorganic analyses. Primary
standard reagents must,  of  course,  be  used  for  standardizing all volumetric  solutions.
Commercially prepared reagents and standard solutions are very  convenient  and may be
used  when it  is demonstrated that  they meet the method requirements. All prepared
reagents must be checked for accuracy.

The individual methods specify the reagents that require frequent standardization, or other
special treatment, and the analyst  must follow through with these essential operations. To
avoid waste,  the analyst should prepare  a limited volume of such reagents, depending on the
quantity required over a given period of time. Examples and brief discussions of the kind of
problems that occur are given under Paragraph 5.3, "Elimination of Determinate Errors".

As far as possible, distilled water used for preparation of reagent solutions must  be free of
measurable amounts of the constituent  to be determined. Special requirements for distilled
water are given in Chapter 3 of this manual and in individual method descriptions.

Compressed gases, such as oxygen and nitrogen, used for total organic carbon determination
may be of commercial grade.

5.2.2 Metals Analyses

All  standards  used   for  atomic  absorption and  emission spectroscopy  should  be of
spectroquality.  It   is  recommended   that  other  reagents  and  solvents  also be of
spectroquality, although AR grade is sometimes satisfactory. Standards may be prepared by
the analyst in the laboratory or prepared, spectrographically standardized materials may be
purchased  commercially. Standards required for determination of metals in water are not
generally available from the National Bureau of Standards.

Analytical reagent grade  nitric and hydrochloric acids  must be  specially  prepared by
distillation in borosilicate glass and diluted with deionized distilled water. All other reagents
and standards are also prepared in deionized  water.

In general, fuel and  oxidant gases  used  for atomic  absorption can be of commercial grade.
Air supplied by an ordinary laboratory compressor is quite satisfactory, if adequate pressure
is maintained and necessary  precautions  are taken to  filter oil, water,  and possible trace
                                         5-2

-------
metals  from the line.  For certain  determinations, e.g., aluminum, reagent-grade nitrous
oxide is required.

5.2.3 Radiological Analyses

The great sensitivity of radioactive counting instruments requires that scintillation grade
reagents and solvents, or equivalent, be used for all radioactivity determinations. Some of
the reagents,  for  example, strontium  carbonate and yttrium oxide carriers used for the
determination of strontium 90 and yttrium 90, must be stable, that is  free of radioactivity.
Barium sulfate, used for coprecipitation of radium must be free  from all traces of radium.
These reagents and solvents are commercially available from chemical supply houses.

Calibrated standard sources of specific  radioactive materials with known count and date of
counting are available from various suppliers.  No single company supplies all standards.

Gases used for radioactive counting must be of high purity  and extra dry. Gases such as
helium  and air are aged for about  30  days  to allow radioactive background to decay. All
gases are checked for background before use. Some cylinders contain inherent radioactivity
which is imparted to the gas.  When  this background is  above normal, the gas should not be
used for radioactivity determinations.

5.2.4 Organic Chemical Analyses

The minimum purity of reagents and solvents that can be used for organic analyses is AR
grade. Reference grade standards should be used whenever available. Special note should be
taken of  the  assay  of standard materials. Owing  to the great sensitivity (nanogram and
subnanogram quantities) of gas chromatography (GC), which is often used to quantitate
organic results, much greater purity  is frequently required (6). The specificity of some GC
detectors requires that reagents and solvents be free of certain classes  of compounds. For
example,  analyses  by electron capture require that  reagents  and solvents  be  free  of
electronegative materials that would  interfere with the determination of specific compounds
in the  sample. Similarly, use of the  flame  photometric detector  requires that reagents and
solvents be free from  sulfur and/or  phosphorus interference. Pesticide  quality solvents are
available  from  several  sources.  These  are  often  satisfactory for  many organic  GC
determinations.  However,  the contents of each container must be checked to assure  its
suitability  for the analyses. Similarly, all analytical reagents and  other chemicals must also
be checked routinely.

The quality of  gases  required for GC determinations varies somewhat with the type of
detector. In general, the compressed gases are a prepurified dry grade. Grade A helium from
the U.S. Bureau of Mines has always been satisfactory. The  Dohrmann nitrogen-detection
system  requires  the use  of  ultra-pure  hydrogen for satisfactory  results. The use  of
molecular-sieve,  carrier-gas filters and drying tubes is required on combustion gases. They
are recommended for use on all other gases.  It is recommended that the analyst familiarize
himself with an article by Burke (7) on practical aspects of gas chromatography.

All reagents, solvents, and adsorbents used for thin-layer chromatography must be checked
to be certain  that  there  are  no impurities present that will react with the chromogenic
reagent  or otherwise interfere with  subsequent qualitative or quantitative determinations.
Glass-backed  layers prepared  in  the  laboratory  or precoated  layers  supplied by a
manufacturer  may  be used.  However, precoated  layers are more  difficult  to scrape.
                                         5-3

-------
Therefore, it is recommended that layers prepared in the laboratory be used when zones are
to be scraped in order  to recover isolated compounds. Plastic-backed layers are generally
unsatisfactory for this type of analysis.

Adsorbents most commonly used for column chromatographic clean-up of sample extracts
are Florisil, silica gel, and alumina. These must be pre-activated according to the method
specifications and checked for interfering constituents.

5.3 Elimination o. Determinate Errors

In order to produce high quality analytical data, determinate errors must be eliminated or at
least minimized. For purposes of this discussion, we assume that a competent analyst and
reliable  equipment, in optimum operating condition, are available. Thus, determinate errors
that might   result  from  an inexperienced or careless analyst  and poor equipment  are
eliminated. The remaining sources of error are the reagents, solvents, and gases that are used
throughout the analyses. The quality of these materials, even though they are AR grade or
better, may vary from one source to another, from one lot to another, and even within the
same lot.  Therefore, the analyst must predetermine that all of these materials are free of
interfering substances under the  conditions of the analyses. To do  this he must have a
regular  check program. Materials that do not meet requirements are replaced or purified so
that they can be used.

5.3.1 Reagent Blank

The first step the analyst  must take is to determine the background or blank of each of the
reagents and solvents used in a given method of analysis. The conditions for determining the
blank must be  identical  to those used throughout the analysis, including the  detection
system. If the reagents and solvents contain  substances  that interfere with a  particular
determination, satisfactory  reagents  and  solvents must  be found.  Where  possible and
practical, they should be treated so that they can be used.

5.3.2 Method Blank

After determining the individual reagent or solvent blanks, the analyst must determine the
method blank to see if the cumulative blank interferes with the analyses. The method blank
is determined by following the procedure step by step, including  all of the reagents and
solvents, in the quantity required by the method. If the cumulative blank interferes with the
determination, steps must be taken to  eliminate or reduce the interference to a level that
will permit this combination of solvents and  reagents to  be used.  If the blank  cannot be
eliminated,  the magnitude  of  the interference must be  considered when calculating  the
concentration of specific constitutents in the samples being analyzed.

A method blank should be determined whenever an analysis is made. The number of blanks
to be run is determined by the method of analysis and the number of samples being analyzed
at a given time. In  some methods, such as the AutoAnalyzer procedures, the method blank
is automatically and continuously compensated for since a continuous flow of the reagents
passes  through the  detector.  In other  procedures,  such  as  the gas  chromatographic
determination of pesticides, a  method  blank  is run with each series of samples analyzed.
Usually this is one blank for every nine samples.
                                         5-4

-------
5.3.3  Elimination of Interferences and Other Sources of Error

Procedures  for eliminating or  at  least  minimizing  impurities  that  produce  specific
interferences or high general background, vary with the reagent and method involved. These
procedures  may  include:  recrystallization,  precipitation,  distillation, washing with  an
appropriate solvent, or a combination of these. Examples of procedures used for various
types of analyses are given below. For complete information, the analyst should consult the
individual methods.

5.3.3.1 General Inorganic Analyses

Analytical reagent grade chemicals and solvents usually present no interference problems in
inorganic analyses. However, some  reagents do not always meet methods requirements. An
example is  potassium persulfate used in phosphorus and  nitrogen determinations.  This
reagent is frequently  contaminated with ammonia. Therefore,  it is routinely purified by
passing air through a heated water solution of the reagent. The purified potassium persulfate
is recovered by recrystallization.

A problem more commonly encountered in inorganic  analyses is the rapid deterioration of
the standard reagents and other ingredients.  To minimize or eliminate this problem some
reagents, for example, ferrous ammonium sulfate, must be standardized daily. Others, such
as sodium thiosulfate used  for  dissolved oxygen  determination, may require  a substitute
reagent, e.g.,  phenyl  arsene oxide. Solid phenol which readily oxidizes  and acquires  a
reddish color can be purified by distillation. Starch  indicator used for idiometric titrations
may be prepared for  each use or preserved by refrigeration,  or by addition of zinc chloride
or other suitable compounds.

5.3.3.2 Metals Analyses

In general, spectrograde chemicals, solvents, and gases present no interference  problems in
atomic absorption or emission spectrographic determinations. However, standards which do
not meet the requirements of the method are sometimes obtained. Ordinarily, no effort is
made to purify them. They are  simply replaced by new  reagents of sufficient purity. Some
reagents may form  precipitates  on standing. Such reagents will reduce the  accuracy  of
quantitative analyses and should not be used.

5.3.3.3 Radiological Analyses

In  general,  reagents  that do  not  meet  the  purity requirements  for  radiological
determinations are replaced with reagents that are satisfactory. However, in some instances
(for example, barium sulfate used  for coprecipitation of radium), it may be necessary  to
carry   out  repeated  recrystallization  to remove all  forms  of  radium,  and  reduce the
background count to  a  useable  level. In some instances, solvents that  do  not meet
requirements may be distilled to  produce adequate  purity. In some cases, gases  having
background counts may be useable after aging as described  earlier. If not,  they should be
replaced with gases that are satisfactory.

5.3.3.4 Organic Analyses

Many AR-grade chemicals and solvents, and at times pesticide quality solvents, do not meet
the specifications required for the determination of specific organic compounds. Impurities
                                         5-5

-------
that are  considered trace, or insignificant, for many analytical uses, are often present in
greater  quantities  than  the organic  constituents  being measured.  Coupled  with  the
several-hundred-fold concentration  of the sample  extract that is usually  required, such
impurities  can cause very significant interferences in trace organic analyses. Reagents and
solvents found to be unsatisfactory, under the conditions of the analyses, must be replaced
or cleaned  up so that they are useable. Some useful clean-up procedures are:

    a.   Washing the inorganic reagents with each solvent that the reagent contacts during
        the analysis,

    b.   Washing the adsorbents, such as silica gel G and Florisil, with the solvents that are
        used for a specified column  or thin-layer chromatographic procedure,

    c.   Pre-extracting distilled water with solvents used for the particular analysis involved,

    d.   Pre-extracting aqueous reagent solutions with the solvents involved,

    e.   Redistilling solvents in all-glass systems using an efficient fractionating column,

    f.   Recrystallizing reagents and dyes  used in colorimetric or thin-layer determinations,

If the  reagents and solvents thus produced are not of sufficient purity, they should  be
replaced.

Dirty gases (quality less than specified) are particularly troublesome in gas chromatographic
analyses. They  may reduce the sensitivity of the detector, and produce a high  or noisy
baseline. If this occurs, the cylinder should be replaced immediately. Similarly, if cylinders
of compressed gases are completely emptied in use, the end volumes of the gas may produce
a similar and often more severe  effect. Oils and water may get into the system and  foul the
detector. When this occurs  the system must be dismantled and cleaned. Overhaul of the
detector may be required. To  reduce chances of this, it is recommended that  all  gas
cylinders be replaced when the pressure falls to 100-200 psi. Filter driers are of little help in
coping with this type of contamination.

5.3.4 Storing and Maintaining Quality of Reagents and Solvents

Having carried out the tasks of selecting, preparing,  and verifying the suitability of reagents,
solvents, and gases, the analyst must  properly  store them  to prevent contamination and
deterioration prior  to their use.  Borosilicate  glass  bottles with ground  glass stoppers are
recommended  for  most  standard  solutions  and  solvents.  Plastic   containers,  e.g.,
polyethylene, are recommended for alkaline solutions. Plastic  containers must not be used
for reagents or solvents intended for organic analyses. However, plastic containers may be
used for reagents not involved with organic analyses if they maintain a constant volume, and
it is demonstrated that they do not produce interferences and do not absorb constitutents
of interest. It is important that all containers be properly cleaned and stored prior to use.
(Refer to Chapter 4 for details).

Standard reagents  and solvents must always be stored according  to the  manufacturer's
directions.  Reagents or solvents that are sensitive  to the light should  be stored in dark
bottles and/or stored  in  a cool, dark  place.  It  is particularly important  to store materials
used for radiological determinations in dark bottles, since photoluminescence will produce
                                          5-6

-------
high background  if light sensitive detectors are used for counting. Some reagents require
refrigeration.

Adsorbents for thin-layer and column chromatography are stored in the containers that they
are supplied in, or according to the  requirements of individual methods. Activated carbon,
used for collection of samples for organic analyses, must be stored and processed in areas
protected from atmospheric and other sources of contamination (8).

The analyst  should pay  particular attention  to the stability of the standard reagents.
Standards should not be kept longer than recommended by the  manufacturer, or in the
method. Some standards are susceptible to changes in normality due to absorption of gases
or water vapor from  the air. Provisions for minimizing this effect are given in Part 23 of
ASTM Standards (4).

The concentration of  the standards  will change as a result of evaporation of solvent. This is
especially true of standards prepared in volatile organic solvents. Therefore, the reagent
bottles should be kept stoppered, except when actually in use. The chemical composition of
certain  standards  may change  on standing. Certain pesticides,  for instance, will  degrade if
prepared in acetone that contains small quantities of water. Thus, it is essential that working
standards be frequently  checked to determine  changes in concentration or composition.
Stock solutions should be checked before preparing new working standards from them.
5.4  References

1.   National Bureau of Standards, Special Publication 260, "Standard Reference Materials",
    July 1969.

2.   "Reagent Chemicals, American Chemical Society Specifications", American Chemical
    Society, Washington, D.C.

3.   ASTM Special Technical Publication No.  148-H,  "Manual  on Industrial Water and
    Industrial Waste Water", 2nd Edition, p. 869 (1965).

4.   1968  Book of ASTM Standards,  Part 23;  p. 897: American Society for Testing and
    Materials,  1916 Race Street, Philadelphia, Pa. 19103.
5.  Standard Methods  for  the  Examination  of Water  and Wastewater,  13th Edition,
    American Public Health Association, New York (1971).

6.  "FWPCA Method for Chlorinated Hydrocarbon Pesticides in Water and Waste Water",
    Federal   Water  Pollution   Control   Administration,  Analytical  Quality  Control
    Laboratory, November 1969.

7.  Burke, J., JAOAC, 48, 1037 (1965).

8.  Breidenbach, A.  W., et  al,  "The Identification and  Measurement  of  Chlorinated
    Hydrocarbon Pesticides in Surface Waters", Publication WP-22, Federal Water Pollution
    Control Administration, Washington, D.C. (November 1966).
                                        5-7

-------
                                       Chapter 6

                       CONTROL OF ANALYTICAL PERFORMANCE
6.1  Introduction
This chapter is limited  to  a discussion of the control of analytical performance in the
laboratory.  It  is assumed  that  a  valid sample  has been properly taken, preserved,  and
delivered to the laboratory for analyses; that the laboratory analyses were done according to
currently-recognized  methods;  and  that  the  recording and reporting of  subsequent
laboratory results were done in a systematic, uniform, and permanent fashion (See Chapter
7).  It must  be recognized  (and practiced!), however, that quality control begins with the
sample  collection and does not end until the resulting data are reported. The laboratory
control of analytical performance is but one vital link in obtaining valid data. A continuous
rapport and  conscientious use of quality  control  between  field  sampling, laboratory
analyses, and management decisions are necessary to insure this validity.

Earlier  chapters have discussed such key  elements as laboratory services,  instrumentation,
glassware, reagents,  solvents, and  gases;  the reader should refer to these  sections to
determine the necessary specifications and requirements required for quality control. On the
assumption  that these variables are under  control, that a  single method is being used, and
that the complete system is initially under control, what should be done in  the evaluation of
daily performance to document that valid data are being produced? First, valid precision
and accuracy  data should be available on the method and analyst. Thereafter, systematic
daily checks are required to show that reproducible results are being obtained, and that the
methodology is actually  measuring what is in the sample. These items are discussed in detail
in the following sections.

6.2  Precision and Accuracy

Precision refers to the reproducibility among replicate observations. In  an Analytical Quality
Control Program, it  is determined, not on reference standards, but  by the  use of actual
water samples which cover a range  of concentrations and  a variety  of interfering materials
usually  encountered  by the  analyst. Obviously, such data should not be collected until the
analyst  is thoroughly familiar with the method, and  has obtained a reproducible standard
curve. For colorimetric analyses, the initial standard curve should include a blank and a
series of at  least eight standards encompassing the full concentration range to  be used for
routine  sample analyses.  Subsequently, at least two standards (a high  and  a low) should be
analyzed  to  verify  the  original standard curve.  For  other measurements, such as  pH,
conductivity,   turbidity,  etc.,  instruments  should  be   standardized  according  to
manufacturer's instructions (See Chapter 3) and sound, scientific practice.

There are  a  number  of different methods available for the determination of precision. One
method that has been successfully employed by experienced AutoAnalyzer users, and can
be adapted to many  other analytical instrumentation  and chemical procedures,  is described
as follows:

    a.   Four separate concentration levels  should be studied,  including a low concentration
        near the sensitivity level of the  method,  two intermediate concentrations, and a
        concentration near the upper limit of application of the method.
                                         6-1

-------
    b.   Seven replicate determinations should be made at each of the concentrations tested.

    c.   To allow for changes in instrument conditions, the precision study should cover at
        least two hours of normal laboratory operation.

    d.   In  order to permit the maximum  interferences in sequential  operation,  it is
        suggested that the samples be run in the following order: high, low, intermediate,
        intermediate. This series is then repeated seven times to obtain the desired replica-
        tion.

    e.   The precision statement should  include a range of standard deviations over the
        tested range of concentration. Thus, four standard deviations will be obtained over
        a range of four concentrations, but the statement should contain only the extremes
        of standard deviations and concentrations studied.

An example of data generated from such an approach is shown in Table 6-1.
                                     Table 6-1

         PRECISION DATA ON RIVER WATER SAMPLES FOR PHOSPHORUS
                             AUTOANALYZER METHOD
                                  Cone., (mgP/1)

Sample        Kanawha        Klamath        Arkansas        Big Souix
1
2
3
4
5
6
7
Avg.
s
0.05
0.06
0.06
0.06
0.06
0.06
0.06
0.059
0.004
0.10
0.10
0.10
0.11
0.11
0.11
-
0.105
0.005
0.48
0.48
0.49
0.48
0.48
0.48
-
0.482
0.004
0.62
0.62
0.62
0.63
0.62
0.62
0.62
0.621
0.004
The resulting precision statement would read as follows:

    "In a single laboratory, using surface water samples at concentrations of 0.06 and 0.62
    mg P/l, the standard deviation was ±0.004 (Analytical Quality Control Laboratory)."

Thus, the statement contains the number of laboratories involved, the type of samples, the
concentrations used, the resulting standard deviation (s) and the reference source.
                                        6-2

-------
Accuracy refers to a degree of difference between observed and known, or actual, values.
Again, accuracy  should  be  determined on actual water samples routinely analyzed, and
preferably,  on the  same  series as those used in the precision determinations. The method
employed by experienced Auto Analyzer users consists of the following key steps:

    a.  Known amounts of the particular constituent should be added to actual samples at
       concentrations where the precision of the  method  is satisfactory. It is suggested
       that amounts be added to the low-concentration sample, sufficient to double that
       concentration,  and  that  an  amount  be  added to one of  the  intermediate
       concentrations,  sufficient to  bring  the  final concentration  in  the  sample to
       approximately 75% of the upper limit of application of the method.

    b.  Seven replicate determinations at each concentration should be made.

    c.  Accuracy should be reported as the percent recovery at the final concentration of
       the spiked sample. Percent recovery at each concentration should be the mean of
      . the seven replicate results.

Data were obtained with this approach by using two of the water samples previously used in
the precision study  reported in Table 6-1 (Kanawha and Arkansas Rivers). They are reported
in Table 6-2.
                                     Table 6-2

         ACCURACY DATA ON RIVER WATER SAMPLES FOR PHOSPHORUS
                           AUTOANALYZER METHOD


                                 Cone., (mgP/1)

                          Kanawha                         Arkansas
   Sample            (Added 0.06 mg/lP)               (Added 0.3 mg/1 P)

     1                   0.105                              0.74
     2                   0.105                              0.75
     3                   0.105                              0.75
     4                   0.110                              0.73
     5                   0.110                              0.74
     6                   0.110                              0.75
     7                   0.105                              0.75

    Avg.                 0.107                              0.74
% Recovery
                                   100 = 90                    x 100 = 95
                                                  f     °™   1
                                                  [ 0.48 + 0.30 J
                                       6-3

-------
Again, in order to contain the key elements, the accuracy statement would read as follows:

    "In a single laboratory,  using surface water samples at concentrations of 0.11 and 0.74
    mg P/l,  recoveries  were  90% and  95%,  respectively  (Analytical  Quality Control
    Laboratory)".

Once collected and documented, these precision and accuracy data may be used in a number
of ways. Two important examples are: (1) They present clearcut evidence that the analyst in
question is indeed capable of analyzing the water samples for that particular parameter.
That is, he has the standard method under control, and is capable of generating valid data;
and (2) the data can be used in the evaluation of daily performance in reference to replicate
samples, spiked standards and samples, and in the preparation of quality control charts.

As observed, the above methods can be adapted to other chemical procedures and analytical
instruments. They have  been used  on manual titration methods for such parameters as
alkalinity,  chloride,  and   hardness;  on  general  inorganic  instruments  such as  pH,
conductivity,  selective ion,  and turbidity  meters; and on  the  Beckman  Carbonaceous
Analyzer.  Other   instruments,  such  as  atomic   absorption   and   flame  emission
spectrophotometers, could  also  be evaluated by  these  methods;  however, radiological
instrumentation  and  gas   chromatography systems  (See  Chapter  8)  require  special
techniques.

6.3  Evaluation of Daily Performance

Once valid  precision and  accuracy data are available on the method and  the analyst,
systematic daily checks are necessary to insure that valid data are being generated. First of
all, verification  of the originally-constructed standard curve  is mandatory. As previously
noted, at least two standards (a high and a low) should be analyzed  routinely along with a
blank  to determine  that  comparable operating conditions  exist.  If the data do not
substantiate such control, the analyst must systematically trouble-shoot his system until the
problem is corrected.

In order to document that  reproducible results are being obtained  (i.e., precision  of the
method),  it is necessary to run replicate samples. Although  frequency of such replicate
analyses is, by nature, dependent on such factors as the original precision of the method, the
reliability of  the instrumentation involved,  and  the  experience of the  analyst,  good
laboratory technique is  to  run duplicate  analyses at least ten percent  of the  time. The
resulting data should agree favorably with the known precision of the method. If they do
not, the system is not under  control, and results are subject to question.

Concurrently,  quality  control  should  include assurance that  the daily system is actually
measuring what is in the sample (i.e., accuracy of the method). Although it is far preferable
to have obtained values  check with known or actual values,  it should be recognized that
inaccuracy does not destroy the value of data if the degree  and precision of the error is
known and taken into account.  In order to  account for background  contamination and/or
sample interferences, and as a matter of routine practice, spiked samples should be used in
addition to standards.  As  in  the  case  of duplicate  sample analyses,  good laboratory
technique dictates that spiked samples be run at least ten percent of the time.

Thus,  daily control of analytical performance in  the  laboratory requires approximately
15-20 percent of the analyst's time. Considering the elapsed time and combined efforts of
                                         6-4

-------
skilled personnel that are represented in a final laboratory result, this is a comparatively
small price to pay for, not a "number", but a valid concentration value.

A most convenient way of recording the obtained precision and accuracy data is through the
preparation of  quality control charts.  Plotting of said  data systematically  answers the
question as to whether the laboratory analyses are under  control, and is useful in observing
developing trends of positive or negative bias. Because of its importance in documenting the
quality control being practiced daily in the laboratory, the construction and uses of quality
control charts are treated as a separate topic in the next section.

A broader and somewhat  different form of evaluation of daily performance may be made
through routine participation in  interlaboratory round-robin  studies. Samples analyzed in
such a cooperative program  should be treated  as part of the routine sample load. In so
doing,  the analyst is  able to compare his individual performance against other laboratory
personnel, and  to  have a reliable measure of the particular method's capabilities.  In many
respects such samples can be regarded as reputable  "blind samples"; a necessary ingredient
in the quality control of laboratory results.

6.4  Quality Control Charts

Quality control charts were  originally developed for the control of production processes
where  large numbers  of items were  being manufactured and inspected  on an  essentially
continuous basis. As shown in Figure 6-1, a control chart consists of a graphical chart with the
vertical scale plotted in units of the test result and  the horizontal scale in units  of time or
sequence  of results. The upper and lower control  limits  shown on the chart are used as
criteria for action,  or for judging the significance of variations between duplicate samples.
The central line represents the  average or the standard value of the statistical measure being
plotted.
                    D
                    13
                                 Upper Control Limit
                                  Central (Average)
                                 Lower Control Limit
                               Time or Order of Results

                    Figure 6-1  ESSENTIALS OF CONTROL CHART
As observed in the previous section on the evaluation of daily performance, daily precision
and accuracy data can be plotted by means of these quality control charts to determine if
valid, questionable, or invalid data are being generated from day to day. There are several
techniques available for actually constructing quality control charts and plotting subsequent
data. Two  currently in use are the Shewhart technique (1,2) and the CuSum technique (3).
In both techniques,  precision control  charts  are  constructed  from  duplicate  sample
                                         6-5

-------
analyses,  whereas, accuracy  control charts  are  constructed from  spiked  samples  or
standards data generated in monitoring  recovery  efficiencies. At  least 15 to 20 sets of
duplicate and 15 to 20 sets of spiked sample data from an in-control process are necessary
for the initial construction. A system is  initially said  to be in control when the standard
deviation  and recovery  efficiency data for a given parameter are comparable  to  those
obtained  by other experienced  laboratories. It is also  necessary that the  initial and
subsequent sets of data be obtained under normal laboratory operation conditions, that the
same analyst or group of analysts run the analysis, and that the same analytical method is
used.

6.4.1 Cumulative-Summation (CuSum) Quality Control Charts

There  are various systems currently  available  for plotting  data in the form of cumulative
sum charts (4). One system that has been in  continuous use within EPA Region VI  is that of
Harkins  and Crowe  (3). It  has proved most useful in monitoring the validity of data
generated by a contracting laboratory and is currently being used routinely to daily record
intra-laboratory  performance in  technical  operations.  The  following material  has  been
excerpted from their manual (3), in order to accurately describe the construction and use of
these charts:

6.4.1.1  Construction of CuSum Quality Control Charts

The control charts are derived from three basic calculations:

    a.   Standard  deviations  (S.) of the  differences between  duplicates or, in  the case of
        spiked  or standard  samples,  between  the  known  quantity and the  quantity
        obtained.

    b.   The upper control limit (UL)

    c.   The lower control limit (LL)

Prior to these calculations, two decisions must be  made:

    a.   The a and 0 levels

    b.   The allowable variability levels

                              Mathematical Equations
                     n
           n        *--i
               di2
      S2 =	= Variance of the differences
       d         N — 1
      S  = ~~\   S  =. Standard deviation of the differences
      'd  =   V  d   ^ianaara deviation 01 ine umerences                 /a\

      S2  =  (.8Sd)2 = .64 S^ (estimates  a 2Q )


      S2  =  (1.2Sd)2  -1.44 S2 (estimates a \ )
                                         6-6

-------
                         2 loge
                UL(M) =       L  "  J  + M
•M
 L  «  J  +;
                                                         1
                             s2   s2             s2        s2
                              01              01
                LL(M) =
Where: UL(M)   = upper limit at M sets of samples

       LL(M)   = lower limit at M sets of samples

           di   = the difference between the itn set of duplicates or spiked
                  samples

           N   = the total number of sets of duplicates or spiked samples
                  used to construct the control charts

          SQ    = minimum amount of variation allowed in the system

          S2    = maximum amount of variation allowed in the system

           a   = percent (decimal fraction) of time you are willing to judge
                  the procedure out of control when it is in control

           0    = percent (decimal fraction) of time you are willing to judge
                  the procedure in control when it is out of control

           M   = number of sets of  duplicates  or spiked samples used in
                  calculating the value to be plotted on the chart

By definition, a is the probability of judging the process to be out of control when in fact, it
is in control.  It is recommended that a be chosen to lie between the boundaries of .05 and
.15, that is, the laboratory personnel are willing to stop the laboratory process somewhere
between  5 and 15% of the time, judging it to be out of control, when in fact, it is in control.
If  the  cost of examining a process to determine the reason or reasons for being out of
control is considerable, then it may be desirable  to choose a low a. Likewise, if the cost is
negligible, it may be desirable to  choose a larger a value, and thus stop the process more
frequently. (See Figure 6-2)
                                       6-7

-------
   On the other hand, /3 is defined as the probability of judging the process to be in control
   when it is not.  Again, it is recommended that 0 be chosen to lie between the values of .05
   and .15; thus, the laboratory personnel are willing to accept out of control data somewhere
   between 5 and  15% of the time. The economic considerations used for choosing a are also
   applicable to the choice of B. (See Figure 6-2.)

   It is also  essential to set maximum and minimum allowable variability levels. It is necessary
   to specify a value for the  minimum and  maximum  amount of variation that will be
                                                                             2
   allowable in the system. These minimum and maximum amounts are referred to as  a 0
          2
    and
respectively. The values used should be based on a knowledge of the variation in
    the procedure under consideration. However, if such knowledge is not available, the values
                          2            22            2
    may be arbitrarily set at  a   = (a —  .20a)  and a  = (a +  .20 o )  .
 ^f^
w
           LABORATORY IDENTITY CONTROL CHART
           PARAMETER - METHOD
           DATE
           RANGE OF CONCENTRATION
           a and 0 LEVELS
           STANDARD DEVIATION
           UPPER CONTROL LIMIT EQUATION
           LOWER CONTROL LIMIT EQUATION
                                 Sample Set No. (M)
       Figure 6-2. EFFECT OF a  AND p  LEVELS ON STANDARD CONTROL CHART
                                         6-8

-------
6.4.1.2 Use of CuSum Control Charts

Once the control charts are constructed, and prior to their use, consideration must be given
to the number of duplicate analyses to be conducted during a series of samples; likewise, the
same decision must be made on spiked or standard samples.

In considering the number of duplicate and spiked sample  analyses to be conducted in a
series of samples, it is necessary  to weight the consequences when the data go out of
control. The consequences of this situation are reanalyzing a series of samples or discarding
the questionable  data obtained. The  samples to be reanalyzed are those lying between the
last in-control point and the present  out-of-control point.  A  realistic frequency for running
duplicate and spiked samples would be every fifth sample; however, economic consideration
and experience may require more or less frequent duplicate and spiked sample analyses.

Once the frequency  of duplicate and spiked samples has been  determined, it is then
necessary  to  prepare  spiked  or  standard  samples  in  concentrations  relative to  the
concentration of  the control charts, which should be similar to those of the environmental
samples. These spiked or standard samples must be intermittently dispersed among the series
of samples to be  analyzed and without the analyst's knowledge of concentration. Similarly,
duplicate samples must be intermittently dispersed throughout the series of samples to be
analyzed,  and ideally, without the analyst's knowledge; however, this is sometimes very
difficult to accomplish.

The  results  of the duplicate and spiked sample analyses should be calculated immediately
upon analyzing the samples to allow  for early detection of problems that may exist  in the
laboratory. An example of these calculations  follows:
     Duplicate
    Sample No.
        M           No.  1     No.  2   Difference (di)     di2    2 (di2)
         1              5.4       5.2         .2            .04       .04
         2              4.8       4.7         .1            .01       .05
         3              6.1       5.8         .3            .09       .14
Upon plotting the summation or S(di2),  one of three possibilities can occur (See Figure
6-3):

    a.   Out of control on the upper limit
        When data goes  out of control on the upper limit  the following steps should be
        taken:
            1.  Stop work immediately
            2.  Determine problems
               (a)  Precision control chart
                   (1) The analyst
                   (2) Nature of the sample
                   (3) Glassware contamination
                                        6-9

-------

-------
                (b)  Accuracy control chart
                    (1)  The analyst
                    (2)  Glassware contamination
                    (3)  Contaminated reagents
                    (4)  Instrument problems
                    (5)  Sample interference with the spiked material
            3.   Rerun samples represented by that sample set number, including additional
                duplicate and spiked samples.
            4.   Begin plotting at sample No. 1 on chart.


    b.  In control within the upper and lower limit lines

        When data continuously fall  in between the upper and  lower  control limits, the
        analyses should be continued until an out-of-control trend is detected.

    c.  Out of control on the lower limit

        When data fall out of control  on the  lower limit, the following steps should be
        taken:
            1.   Continue analyses unless trend changes
            2.   Construct new control charts on recent data
            3.   Check analyst's reporting of data
6.4.2 Shewhart Quality Control Charts

Dr. Walter A.  Shewhart of Bell Telephone Laboratories developed the basic theory of
control charts in the  1920's. His book on statistical quality control (1) grew out of this
original work. Since then, industrial acceptance of these control chart concepts and  other
statistical techniques have refined and quantitated the quest for quality in manufacturing.
Although originally developed for control of production processes when large numbers of
articles were being manufactured and inspected on  an essentially continuous basis,  these
same  concepts have  been readily  adapted  to laboratory operations where the analyst
produces comparatively fewer results on an intermittent basis.

As in the CuSum approach, precision control charts are prepared from data resulting from
duplicate sample analyses and accuracy control charts from duplicate spiked standards or
samples. Once  the  control charts are constructed, however, data are plotted as individual
values rather than cumulative sums.

Certain constants (factors) are also  involved  in the  preparation  of Shewhart Charts.
Depending upon  how  the data are  grouped, what the size of each grouping is, and what
control limit formulation is being calculated,  Table 6-3 will serve as a basic  reference
point:
                                        6-11

-------
                                     Table 6-3


            FACTORS FOR COMPUTING CONTROL CHART LINES (5, 6)
                     Observations in    Factor        Factor
                        Subgroup CQ       A2           D4
                            2           1.88          3.27
                            3           1.02          2.58
                            4           0.73          2.28
                            5           0.58          2.12
                            6           0.48          2.00
                            7           0.42          1.92
                            8           0.37          1.86
Inherent in the Shewhart  approach is recognition of the basic assumption that variations
exist in every method. That is, no procedure is so perfect, so unaffected by its environment,
that it will always give exactly the  same assay value or product. Where such situations seem
to exist, either the device used to measure the process is not sensitive enough or the person
making the  measurements is not  performing properly. For our purposes, the  recorded
difference between  paired  samples  should never be  less  than one-half  the minimum
detectable limit of  the parameter under  consideration. In  the  following  outlines  for
preparing  precision  and accuracy  control charts, nitrate data were used to develop  the
examples. Therefore  the minimum  values  shown are 0.05 (one-half the observed minimum
detectable limit of 0.1 mg/1 as N).
6.4.3 Precision Control Charts

These charts are developed by collecting data for many samples, a minimum of 15 to 20,
run in duplicate under assumed controlled conditions. Once these data have been generated,
preferably  over an extended period of  laboratory time,  the  following steps  should  be
followed to construct the control chart:
   a.  List the range  (R) for each set  of samples. That  is,  the  absolute value  of the
       difference between each set of duplicate samples.'
                                        6-12

-------
        Note: The following ranges were observed in the nitrate data:
0.1
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.1
0.05
0.05
0.05
0.30
0.10
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05

b.  Calculate the average range (R) by summing the list of R values and dividing by the
    number of sets of duplicates:
                                   SR
                              R =  *l
                                    35

c.   Calculate the Upper Control Limit (UCL) on the range according to the formula:

                            UCLR = D4R,

    where D, is a constant dependent on the number of units in the subgroup. In this
    case, since two observations are in the subgroup, d4 = 3.27 (see Table 6-3).

                         UCLR = 3.27 R = 0.20

d.   Calculate the Upper Warning  Limit (UWL) on the range according to the formula:

                     UWLR =  2/3 (D4R-R)+R,

    which for duplicate samples reduces to

                            UWLR = 2.51  R

                            UWLR = 0.15

    This UWL corresponds to the  95% confidence level.
                                    6-13

-------
e.   Now graph R, UWLR, and UCLR in the following manner:
 R
        0.3
        0.2
        0.1
                                                        UCLR = 0.20
                                            UWLR = 0.15
                                                             R - 0.06
             12345
                              Order of Results
                          (e.g., duplicate sample sets)

f.   The above precision control chart for nitrates is now complete, and can be used to
    plot  R values on subsequent duplicate samples to determine if the system is in
    control, out of control (plotted R value beyond the UCL), and/or to detect any
    trends developing within the system.

    1.   In  this example, a trend has developed between duplicate sample sets 4 thru 7.
        Although the system  is not out of control, all variables in the procedure should
        be  checked in an attempt to stop this obvious trend before the UCL is reached.
 R
       0.3
       0.2
        0.1
         0
X
                                                      UCLR = 0.20
                                                      UWLR = 0.15
                                    X
X
                                                          R - 0.06
     XXX
              12    3456789

                             Order of Results
                                             10
    2.   In this example, the system has clearly gone out of control between duplicate
        samples 3 and 4. At this point, the system can be stopped and all variables in
        the system  checked, or another set of duplicate samples can be run to verify
        the  observed difference.  Once  the system has been  corrected,  all samples
        between set 3 and 4 should be rerun to insure the validity of the data.
                                    6-14

-------
          0.3

          0.2

          0.1
                  X
                            UCLR = 0.20
UWLR = 0.15
X
R = 0.06
X X
                 1234567
                                   Order of Results
6.4.4 Accuracy Control Charts

As in the above system, these charts are developed by collecting data for many samples, a
minimum  of 15 to  20,  but on spiked samples (preferably) or standards under assumed
controlled conditions. Again, these data should be generated over an extended period of
laboratory time, and be  representative of normal  operating conditions (7). The following
steps should be followed to construct accuracy control charts:
    a.   List the range (R) and the average (X)* of each subgroup of data.
        Note: In  the  following  example  of nitrate data, subgroups of  monthly  data
        involving four observations were used:
                               *X  =  Ed Found  - Actual
       Month
       Sept.
       Nov.
       Dec.
       Jan.
Actual
   1.1
   1.1
   1.1
   1.1
   1.1
   1.1
   1.2
   1.2
   1.0
   1.0
   1.0
   1.0
   1.0
   1.0
   1.0
   1.0

Found
1.1
1.1
1.1
1.1
1.1
1.1
1.2
1.1
1.1
1.0
1.0
1.0
0.9
1.0
1.1
1.0
n
R

0



0.1



0.1



0.2


X
                                                          -0.025
                                                          +0.025
                                                          0
                                        6-15

-------
b.  Calculate the average range (R) by summing the list of R values and dividing by the
    number of subgroups:
                       R = 0,4 = o.lO
c.   Calculate the Upper Control Limit (UCL) on the range according to the formula:


                       UCLR = D4R,

    where D4 is a constant dependent on the number of units in the subgroup. In this
    case, since four observations are in the subgroup, D4 = 2.28 (see Table 6-3).

                    UCLR = 2.28 R = 0.23

d.   Calculate the Upper Warning Limit (UWL) on the range according to the formula:

                     UWLR = 2/3(D4R-R)+R

                     UWLR = 2/3 [2.28 (0.1>0.1 J+0.1 = 0.19

e.   Now graph R, UWLR, and UCLR in the following manner:
    R
0.3

0.2

0.1
                                                      UCLR = 0.23
                                                      UWLR = 0.19
                                                          R=0.1
                              Order of Results
f.   Turning now to the X values, calculate the UCL^ by the formula:
                                =  A2R
     where A2 is a constant dependent on the number of units in the subgroup.  In this
     case, since four observations are in the subgroup, A2 = 0.73 (see Table 6-3).
                           = 0.73R = 0.07
                                   6-16

-------
g.  Calculate the UWL^ by the formula:
                              = 2/3 A2(R)

                              = 2/3 [ 0.73 (0.1)] = 0.05
    Note: Lower Warning Limit (LWL^) and Lower Control Limit (LCL^) are simply
    the negative values of UWL^ and UCL^, respectively.
h.   Now graph the standard Nominal Value  (set equal to zero),
    LWLy, and LCLv in the following manner:
                                                                   , and
               .07
               .05
               .03
               .01
               .01

               .03
               .05
               .07
                                             UCLx = °-07
                                                 v = 0.05
                                             Std Nominal Value
                                         order of subgroups
                                                   = 0.05
                                             LCLx = 0.07
    In order to detail any trends forming within each subgroup, individual differences
    may be plotted by preparing the following graph:
+.2
+.1
0
-.1
-.2


1
1
1




























i
1









1 1
i 1
Nominal

subgroup
                    Individual value
j.   Thus, as in the precision control charts, once all of the above accuracy control
    charts  have  been constructed,  all  future  data  can  be plotted on  each set of
    duplicate spiked samples or standards, to determine if the system is in control, out
    of control, and/or to detect any trends developing within the system.
                                   6-17

-------
6.5  References

1.   Shewhart, W. A., Economic Control of Quality of Manufactured Product, 1931.

2.   Anon., "Statistical Method-Evaluation and Quality Control for the Laboratory",
    DHEW Training Course Manual in Computational Analysis, August 1968.

3.   Anon.,  "Laboratory  Quality Control  Manual",  Federal Water  Pollution Control
    Administration, Robert S. Kerr Water Research Center, 1969.

4.   Griffin,  D. F., "Systems Control  by Cumulative Sum Method", Amer. J.  Med. Tech.,
    34,644(1968).

5.   Duncan, A. J., Quality Control and Industrial Statistics,  3rd Ed., R. D.  Irwin, Inc.,
    Homewood, III, Chap. 18 (1965).

6.   ASTM Special Technical Publication  No.  15-C,  "Manual  on Quality  Control of
    Materials", pp. 59-64, January 1951.

7.   Grant, E. G., Statistical Quality Control, 3rd Ed., McGraw-Hill, New York (1964).
                                       6-18

-------
                                      CHAPTER 7

                           DATA HANDLING AND REPORTING
7.1 Introduction

To  obtain  meaningful data  on  water  quality,  the  laboratory  must  first  collect  a
representative sample and deliver it unchanged for analysis. The analyst must then complete
the proper analysis in  the prescribed fashion. Having accomplished these steps, one other
important  step  must be  completed  before  the data are of use. This step includes the
permanent recording  of the analytical data in meaningful, exact terms, and reporting it in
proper form to some storage facility for future interpretation and use.

The brief sections that  follow discuss the data value itself, recording and reporting the value
in the proper way, means of quality control of data, and storage and retrieval.

7.2 The Analytical Value
7.2.1 Significant Figures

The term significant figure is used rather loosely to describe some judgment of the number
of reportable digits in a result. Often the judgment is not soundly based and meaningful
digits are lost or meaningless digits are accepted.

Proper  use  of significant  figures gives  an  indication of the  reliability  of the  analytical
method used. The following  definitions  and rules are suggested for retention of significant
figures:

A number is an expression of quantity. A figure or digit is any of the characters  0, 1,2, 3, 4,
5, 6, 7, 8, 9, which, alone or in combination, serves to express a number. A significant figure
is a digit that denotes the amount of the quantity in the place in which it stands.

Reported  values should contain only significant figures. A value is made up of significant
figures when it contains all digits known to be true and one last  digit in doubt. For example,
if a value is reported as  18.8 mg/1, the "18" must be firm values while the  "0.8" is
somewhat uncertain and may be "7" or "9".

The number zero may or may not be a significant figure:

    a.   Final zeros after a decimal point are  always  significant figures. For example, 9.8
        grams to the nearest mg is reported as 9.800 grams.

    b.   Zeros before a decimal point with other preceding digits are significant. With no
        other preceding digit, a zero before the decimal point is not significant.

    c.   If there are no digits preceding  a decimal point, the zeros after the decimal point
        but preceding other digits are not significant. These zeros only indicate  the position
        of the decimal point.
                                         7-1

-------
    d.  Final zeros in a whole number may or may not be significant. In a conductivity
        measurement of 1000 /Ltmhos/cm, there is no implication that the conductivity is
        1000 + 1 /zmho. Rather, the zeros only indicate the magnitude of the number.

A good measure of the significance of one or more zeros before or after another digit is to
determine whether the zeros can be dropped by expressing the number in exponential form.
If they can, the zeros are not significant. For example, no zeros can be dropped  when
expressing a weight of 100.08 grams in exponential form; therefore the zeros are significant.
However, a weight of 0.0008 grams can be expressed in exponential form as 8 x 10"4 grams,
and the zeros are not significant. Significant figures reflect the limits of the  particular
method of analysis. It must be decided beforehand whether this number of  significant digits
is sufficient for  interpretation purposes. If not, there is little that can  be  done within  the
limits of normal  laboratory operations to improve these values. If more significant figures
are needed, a further improvement in method or selection  of another  method will be
required to produce an increase in significant figures.

Once  the number of  significant  figures is established  for a type of analysis, data resulting
from such analyses are reduced according to set rules for rounding off.

7.2.2 Rounding Off Numbers

Rounding off of numbers is a necessary operation in all analytical areas. It is automatically
applied by the limits of measurement of every instrument and all glassware. However, it is
often applied in chemical calculations incorrectly by blind rule or prematurely, and in these
instances,  can seriously affect the final results. Rounding off should normally be applied
only as follows:

7.2.2.1  Rounding-Off Rules

    a.   If the figure following those to be retained is less than 5, the figure is dropped, and
        the retained figures are kept unchanged. As an example:  11.443 is rounded off to
        11.44.

    b.  If the figure following those to be retained is greater than 5, the figure is dropped,
        and the last retained figure is raised by 1. As an example: 11.446  is rounded off to
        11.45.

    c.  When the figure following those to be retained  is 5, and there  are no figures other
        than zeros beyond the 5, the figure is dropped, and the last place  figure retained is
        increased by  1 if it is an odd number, or it is kept unchanged if an  even number. As
        an example:  11.435 is rounded off to 11.44, while 11.425 is rounded off to 11.42.

 7.2.2.2 Rounding Off Single Arithmetic Operations

    a.  Addition: When adding a series of numbers, the sum should be rounded off to the
        same numbers of decimal places as the addend with the smallest number of places.
        However, the operation is completed with all decimal places intact and rounding off
        is done afterward. As an example:
                                           11.1
                                           11.12
                                           11.13
                                           33.35 The sum is rounded off to 33.4.
                                          7-2

-------
    b.  Subtraction: When subtracting one number from another, rounding off should be
        completed before the subtraction operation,  to avoid invalidation of the whole
        operation.

    c.  Multiplication: When two numbers of unequal digits are to be multiplied, all digits
        are carried through the operation, then the product is rounded off to the number of
        significant digits of the less accurate number.

    d.  Division:  When two numbers of unequal digits  are to be divided, the  division is
        carried out on the two numbers using all digits. Then the quotient is rounded off to
        the number of digits of the less accurate of the divisor or dividend.

    e.  Powers  and Roots: When a  number  contains n  significant digits, its  root  can be
        relied on for n digits, but its power can rarely be relied on for n digits.

7.2.2.3 Rounding Off the Results of a Series of Arithmetic Operations

The rules for rounding off are reasonable for simple calculations, however, when dealing
with two nearly equal numbers, there is a danger of loss of all significance when applied to a
series of computations which rely on a relatively small difference in two values. Examples
are calculation of variance and standard deviation. The recommended procedure is to carry
several extra figures through the calculation and then to  round off the final answer to the
proper number of significant figures.

7.2.3  Glossary of Terms

To clarify the meanings of reports and evaluations of data, the following terms are defined.
They are derived in part from American Chemical Society  and American Society for Quality
Control usage (1, 2).

7.2.3.1 Accuracy Data

Measurements which relate to the difference between the average test results and the true
result when the latter is known or assumed. The following  measures apply:

        Bias is defined as error in a method which systematically  distorts results. The term
        is used interchangeably with accuracy in that bias  is a measure of inaccuracy.

        Relative  error is the mean error of a series of test results as a percentage of the true
        result.

7.2.3.2 Average

In ordinary usage, the arithmetic mean. The arithmetic mean of a set on_n_ values is the sum
of the values divided by n.

7.2.3.3 Characteristic

A property  that can serve to differentiate between items.  The differentiation may be either
quantitative (by variables), or qualitative (by attributes).
                                         7-3

-------
7.2.3.4 Error

The difference between an observed value and its true value.

7.2.3.5 Mean

The sum of a_series of test results divided by the number in the series. Arithmetic mean is
understood (X).

7.2.3.6 Population

Same as Universe. (See subparagraph 7.2.3.13).

7.2.3.7 Precision

Degree of mutual agreement among individual measurements. Relative to a method of test,
precision is the  degree of mutual agreement among individual measurements made under
prescribed, like conditions.

7.2.3.8 Precision Data
ft
Measurements which relate to the variation  among  the test results themselves, i.e., the
scatter or dispersion of a series of test results, without  assumption of any prior information.
The following measures apply:

    a.  Standard Deviation (a). The square root of the variance.
                                              n

    b.  Standard Deviation, estimate of universe (s).
                                                 n- 1
    c.   Coefficient of Variance  (V)^_ The ratio  of the  standard  deviation (s) of a set of
        numbers, ji, to their average, X, expressed as a percentage:
                                             v.i.

    d.   Range. The difference between the largest and smallest values in a set.

    e.   95% Confidence Limits. The interval within which one estimates a given population
        parameter to lie, 95% of the time.
                                         7-4

-------
7.2.3.9 Sample

A group of units, or portion of material, taken from a larger collection of units, or quantity
of material, which serves to provide information that can be used as a basis for judging the
quality  of the  larger quantity as a basis for action  on the  larger  quantity  or on  the
production process. Also used in the sense of a "sample of observations."

7.2.3.10  Series

A number of test results which possess common properties that identify them uniquely.

7.2.3.11  Skewness(k)

A measure of the lopsidedness or asymmetry of a frequency distribution defined by the
expression:

                                     (X.  -  X)3
                                         na3


This measure is a pure signed number. If the data are perfectly symmetrical, the skewness is
zero. If k is negative, the long tail of the distribution is to the left. If k is positive, the long
tail extends to the right.

7.2.3.12 Unit

An object on which a measurement or observation may be made.

7.2.3.13 Universe

The totality of the set of items, units, measurements, etc., real or conceptual, that is under
consideration.

7.2.3.14 Variable

A term used to designate a method of testing, whereby units are measured to determine, and
to record for each unit, the numerical magnitude of the characteristic under consideration.
This involves reading a scale of some  kind.

7.3 Report Forms

The analytical information reported  should include the parameter, the details of the analysis
such as burette readings, absorbance, wavelength, normalities of reagents, correction factors,
blanks, and finally, the reported value.

To reduce errors in manipulation of  numbers, a good  general  rule is  to keep  data
transposition to an absolute minimum. If this were pursued, the ideal report form would
include all preliminary information of the analysis, yet it  would be possible to use the same
form through to the final reporting  of data into  a computer or  other storage  device.
However, the ideal report form is not usually in use. Rather, a variety of methods are used to
record data. They are:
                                         7-5

-------
7.3.1 Loose Sheets

Reporting of data onto loose or ring-binder forms is an older, but much used means of
recording data.  It  does allow easy  addition of  new sheets,  removal of older data,  or
collection of specific data segments. However, the easy facility for addition or removal also
permits easy loss or misplacement of  sheets, mix-ups as to date sequence, and questionable
status in formal display, or for presentation as evidence.

7.3.2 Bound Books

An improvement in data recording is  use of bound books which force the sequence of data
insertion. Modification beyond a simple lined book  improves  its effectiveness with little
additional  effort. Numbering  of  pages  encourages  use in  sequence and aids also  in
referencing data, through a  table of contents, according  to time, type of analysis, kind of
sample, analyst, etc.

Validation can be easily accomplished by requiring the analyst to date and sign each analysis
on  the  day completed. This validation can be strengthened further by providing space for
the laboratory supervisor to  sign off as to the date and acceptability of the analysis.

A further development of  the  bound  notebook is the commercially available version
designed for research-type  work. These note  books  are preprinted  with  book and page
numbers and spaces for title of project, project number, analyst signature, witness signature
and dates. Each report sheet has  its detachable duplicate sheet  which allows for up-to-date
review by management without disruption of the book in the laboratory. The cost is about
four times that of ordinary notebooks.

Use of bound  notebooks is  essentially limited to research and development work where an
analysis is part of a relatively long  project, and where the recording in the notebook is the
prime disposition of the data until a status or final report is written.

7.3.3 Pre-Printed Report Forms

Most field laboratories or other installations doing repetitive analyses for many parameters
day in and day out, develop their own system of recording and tabulating laboratory data.
This may include bound notebooks;  but  a vehicle for forwarding data is also required. In
many instances, laboratory units tailor a form to fit a specific group of analyses, or to  report
a  single type of analysis  for series  of samples, with  as much  information  as  possible
preprinted to simplify use of the form. With loose-sheet multicopy forms (use of carbon or
NCR paper) information  can be  forwarded  daily,  weekly, or on whatever schedule  is
necessary, while  allowing retention of all data  in the laboratory. Still, the most common
record is an internal bench  sheet, or  bound book, for recording of all data in rough form.
The bench sheet or book never leaves the laboratory but serves as the source of information
for all subsequent report forms (See Figure 7-1).

In most instances the supervisor and analyst wish to look at the data from a sample point in
relation to other sample points on  the river or lake. This review of data by the supervisor,
prior to release, is a very  important part  of the laboratory's quality control  program;
however, it is  not easily accomplished with bench sheets. For this purpose, a summary sheet
can be prepared which compares a related group of analyses from a number of stations. An
example is shown in Figure  7-2. Since the form contains all of the information necessary for
                                         7-6

-------
W
W
K
on
pa
PL,
o
tu
X
w
 Ul

 ao

PH
                O
                ' 3
-^
C
o
U
c
3
Qi
4-*
C
o
u






























N














"O
O














<














pa














PH





1























«














c/5
                                                                       7-7

-------
w
w
K
oo
W
HH

W
D
oo
W
K-l
PH
S

X
W

fN
 3
 60
E
u
o
u
oo"
w
PH
W
HH
&
O
S
o
«"
w

O
N
IX,
O
00
W
oo
Z

oo


W
HH
S

CM'

3



00
Q
3
00























n,
3
00

GO
Q

03
"o
H
•c
o
3
E
1)
1
3
00
•o
•e
0
2
u
C/3
O
e
T3
E
'c
"o3
Q ~
+•* S
Z
O
H
<
H
00
1
1







CN •— i ON O ro O O
m ^^ ro *—< in C — O\
— i o >— i c~~ <— i O O
O O O ^H O ro O
O O O oo O oo oo
rN 04 rN ro ts ro ro



•*-»
03
S C .1
Q 5 JJ | S
e ° § •? s 2 a
! i i ! S ! -
C g -^ 3 >> 2 -g
s S 1 i 1 s 1
-4_, ^ H_» CS +_, ^_, |g
nj a3 n3 Q cd o3
.2 .2 .2 o .9 .2 ^
4^ 43 ^5 '3 5 4^ .-S
O O O oo O O J
r-~rorN^Hf\l-^ r-- in
rocJc--r-4(NO ^ --<
Omominro O O
OOrooo-3" ro ro
OOOOOOOI^ OO 00
co fN ro rs CN •— < ro ro

•o
03
0
pX i
-M S c M ^
H )_,
.„ £ t; a ~ •§ I ^
!S«21|gS22l
•§ S g 3 2 -| .2 .2 .2c§
.s^^3 rtl i s s2
U 13 S S >> S .3 .2 .2 "«
+j bB-^-uo-^SSlS^
03 c=« rt -4 OS^^^I-
— -^ -^ -t-» -<-> "3
OlCOOtiOcSoSas.fi
2rS2Sg22££fe
O^JOO^OOOO
^
0
o
ro
00
ro



£3
a
Great Miami at Americ;
Materials Bridge









G
O
Whitewater at Suspensi

c-
o
ro
oo
ro
60
j— (
o

o
Great Miami at Lawren
(Lost Bridge)
                                                                                                       it

-------
reporting  data  it is used also to complete  the  data forms forwarded to the storage and
retrieval system.

The forms used to report data to data storage systems require a clear identification of the
sample point, the parameter code, the type of analysis used, and the reporting terminology.
Failure to provide the correct information can result in rejection of the data, or insertion in
an incorrect parameter. As a group of analyses  is completed on one or more  samples, the
values are reported in floating decimal form, along with the code numbers, for identifying
the parameter and the sampling point (station). Figure 7-3 shows an example of a preprinted
report form for forwarding data to keypunch.

7.3.4 Digital Read-out

Instrumental analyses, including automated, wet-chemistry instruments, such as Technicon
AutoAnalyzer,  atomic absorption spectrophotometer,  pH  meter, selective electrode meter,
etc., now  can provide direct digital readout of concentration, which can be recorded directly
onto report sheets without further calculation. Electronics manufacturers now produce
computer-calculators that will construct best-fit curves, integrate curves, and/or perform a
pre-set series of calculations required to obtain the final reported value for recording by the
analyst.

7.3.5 Key Punch  Cards and Paper Tape

Since much of the analytical  data generated in laboratories is recorded on bench sheets,
transferred to  data  report  forms,  key-punched, then  manipulated on small terminal
computers, or  manipulated and stored  in a larger data storage  system, there is a built-in
danger of transfer error. This increases with each transposition of data. It is suggested that
the analyst can reduce this error by recording data onto punch cards directly from bench
sheets. The cards can be retained, or forwarded  immediately to the data storage system as
desired. IBM now offers a small hand-operated key-punch for this purpose.

It is anticipated that in future water quality systems, the intermediate report sheets will be
eliminated and  the  data will be punched automatically by the analytical instrument system
onto key-punch cards and/or paper tapes for  direct use as computer input.

7.4  STORET-Computerized Storage and Retrieval of Water Quality Data

The  use of computers with their almost unlimited ability to record, store, retrieve, and
manipulate huge  amounts of data  is a natural  outgrowth of  demands for meaningful
interpretation of the great masses of data generated in almost any technical activity.

In August 1961,  an informal  conference was held in  the Basic  Data Branch, Division of
Water Supply  and Pollution Control, U.S. Public Health  Service. A number of ideas were
brought together  in the basic design of a system for storage and retrieval  of data for water
pollution  control, called STORET. In 1966, the STORET system was transferred, with the
Division, into the Federal Water Pollution Control Administration, U.S. Department of the
Interior.  A refinement of this system is now operated  by  the  Technical  Data and
Information Branch, Division of Applied Technology, EPA.

If properly stored, the data can be retrieved according to the point of sampling, the date,
the specific parameters stored, etc., or all data at a sample point or series of points can be
                                         7-9

-------
         Figure 7-3. EXAMPLE OF STORET REPORT FORM
WATER QUALITY DATA
LABORATORY BENCH DATA
ST AT ION
Di *



DATi
YR.
MO.
•LI
DAY
"^r ^^OMPosAe^AMPLE" L*ST
TEM
n
a
i
"T1
t
ITEM


8
i


ITEM
1


t


ITFM
1

1


ITEM


* 7
ITEM
,

,


ITEM

_-
1

T,
Jecal Coliform
,
[J
b
< i i i
Li_


4*3 2 I
UNTT MF/100
0 0 0 O
L_



000




Fecal Streptococci t
i

5
till



0000
1
4321
NH3-N + Org-N
t
,

1


tilt
-J_



NH3-N
I
-J

1


1 t 1 t





NO2-N + NOj-N

e

5




4321
P, Total
i




ITEM
t

1

i

ITEM
1
r
ire
i

r-'-,

1
i

1


1111





P, Soluble

TOC
i






1111







JMIT MF/100
000




UNIT mg/1
0000




000




MN.T mg/1

r~
L__







UNIT mg/1








UNIT mg/1










UNIT mg/1





0000




Phenol
i

1111




Cyanide
,
4-
t

1111



-J
i







u
0
1





attnmgjl
000



UNIT Ug/1



UNIT mg/1



COMPUTER CODED DATA
STATION CODE SERIAL YR. MO.

l-« 7-2
1
13-18
PARAMETER CODE VALUE EXPON
31616 [ |
1 9-23 24-27 28

3 , 6 73d n
31-38 36-39 40

o o 6 3311 n
43-47 46-61 82

0 ° 6 ' JLJ d LJ
It-It tO-tl <4
DAY

[—1—

ENT PMK-S
DD
29 ~3
nn
41 tf.
nn
B3 *4
DD
COLUMN
00630 ||
•7-71 72-71 71
NEXT CARD -REPEAT COLUMNS 1-11 ABOVE
0 0 6 6 5 | | [|
11-11 24-27 21

DC
77 7
' (BLANK)
CHO.
]Q
7ft
nn
2B 'C

ooeeTin nnn

0 0 6 8 oj | ||

3273 3C D


0072ol| II
•7-71 72-78 71 ,
nn
nn

CHG
nnn
77 '"• ""-1
                          7-10

-------
extracted as a unit.

There is a  State/Federal cooperative activity which provides State water pollution control
agencies with direct, rapid access into a central computer system for the storage, retrieval,
and analysis of water quality control information.

Full details on use  of the STORET system  are given in the STORET handbook recently
revised (3).

7.4 SHAVES-A Consolidated Data Reporting and Evaluation System

Information systems have been developed to bridge the gap between the analyst and his raw
data, and  a complex  data storage and control  system. These systems include  preprinted
report forms, computerized verification,  and  evaluation  of data and data storage. An
example is  the SHAVES system.

The term, SHAVES, is an acronym  for "Sample Handling and Verification System," which
originated  at  the Great Lakes-Illinois River Basin Comprehensive Project Laboratory at
Grosse  Isle, Michigan. Although the system's original purpose was  verification of the
calculations following laboratory  analyses, it now  includes data  storage,  checks for
completeness and consistency of data, procedures for submitting analytical requests, a set of
forms for recording sampling and analytical information, and a clerical procedure  to account
for  analyses  completed  and  pending.  The  primary  purposes  of  SHAVES  are  the
standardization,  automation and control of reporting analyses. All samples received at the
Pacific Northwest Water Laboratory for routine analysis are processed through the system.

Although SHAVES uses a computer to perform its operations, it is not primarily a computer
program. It is intended for use as an intra-laboratory quality control tool, and as such
compliments the STORET system. It is described in detail elsewhere (4).
7.6  References

1.  "Guide for Measure of Precision and Accuracy," Anal. Chem., Vol. 33, p. 480, (1961).
    p. 480.

2.  "Glossary of General Terms Used in Quality Control," Quality  Progress,  Standard
    Group of the Standards Committee, ASQC, II, (7), pp. 21-2, (1969).

3.  Water Quality Control Information System (STORET), EPA, Washington, D.C. 20460,
    Nov. 15, 1971.

4.  Byram, K.  V.  and Krawczyk,  D.  F., "An Evaluation of SHAVES: A Water Quality
    Sample Handling System," Environmental Protection Agency, Pacific Northwest Water
    Laboratory, 1969.
                                       7-11

-------
                                      CHAPTER 8

                SPECIAL REQUIREMENTS FOR TRACE ORGANIC ANALYSIS
8.1  Introduction

The high sensitivity of the instrumentation used in trace organic chemical analysis, and the
low concentration of compounds being investigated, dictate that special attention be given
this field  of analytical endeavor. Contamination of the sample  from any possible source
must  be  diligently guarded against, and interferences in the  sample must be carefully
controlled. Finally, strict  attention to method and  highly refined technique are required  to
produce valid quantitative results.


8.2  Discrete Bottled Samples

Sample collection  should  be done with wide-mouth glass bottles, equipped with screw caps
fitted  with Teflon  liners.  The  use of a screw  cap without  a Teflon liner may  cause
contamination of  the sample  by the liner or adhesive used in sealing the liner to the cap.
Plastic bottles (polyethylene) are not used because traces of plasticizer may be leached from
the plastic by the  water, and can be a source of analytical interference. Moreover, organics
from  the  sample may be adsorbed  on the  plastic. It has  been suggested  that  high grade
Teflon bottles may be satisfactory for this use; however, the cost is prohibitive at present.
Many  investigators avoid the use of glass  sample bottles, because breakage in shipment
frequently causes  loss of sample. This is overcome  by the  use of relatively inexpensive,
expanded  polystyrene foam shipping containers molded to fit the bottle.  These shipping
containers  can be  purchased  from  Preferred  Plastics Corp.,  North   Grosvenordale,
Connecticut.

To  insure freedom from organic contaminants, bottles are rinsed  successively with chromate
cleaning solution, running tap water, distilled water, and finally several times with redistilled
solvent (e.g., acetone, hexane, petroleum ether, chloroform). Caps are washed with
detergent, rinsed with tap water, distilled water, and  solvent. Liners are treated in the same
way as the bottles and are stored in a sealed container.

Each  method designates a recommended sample size for surface water analysis. Duplicate
samples are recommended. If analysis by more than one method is to be requested on the
same  sample, sufficient sample must be simultaneously obtained to supply the needs of each
analysis.

It is also recommended that when requesting a non-specific analysis, any information that
could help direct  the analytical approach, or aid in  interpretation of results, be supplied.
Such  information  could  include industrial or agricultural activities in the area from which
the sample was obtained,  spills,  or other accidents that may have occurred in the area. Also
mention of similar upstream activity could provide valuable  assistance.

Samples should be stored in a cool, dark place, and analyzed  as soon as possible. If the
sample cannot be analyzed immediately, reporting the holding time can help in interpreting
results where die-off rates  are known.
                                          8-1

-------
8.3  Carbon Adsorption Samples

The affinity of carbon for organic substances requires that supplies of carbon be protected
from  extraneous  sources  of contamination. For example,  carbon  can  adsorb  organic
substances such as paint vehicles and insecticides from the air. Therefore,  the carbon is
stored and processed in an area adequately protected from such sources of contamination.
As  an additional precaution, the ventilating, heating, and air-conditioning systems for the
laboratories in which carbon adsorption samples are processed are completely isolated from
all other laboratories. All carbon  is obtained from the manufacturer in sealed metal drums.
Obviously, spraying with  pest-control chemicals  is not permitted in these areas.  Carbon
blank determinations supplement  these precautions.

8.4  Glassware

Proper calibration of volumetric glassware is essential to valid analytical  results, because
quantitation is performed by comparison to measured amounts of standard compounds, and
by accurate measurement of sample volume.

Individual  concentrator  tubes, used to measure final concentrate volumes,  must  first be
calibrated at the working volume. This is particularly important for volumes less than 1 ml.
Calibration should be made by noting the number of microliters of solvent required to bring
the liquid level  (lower miniscus) up to a particular graduation mark. A precision 100 //I
syringe should be used to measure calibration volume.

It is also very important in trace  organic analysis that glassware be as free as possible of any
organic contaminants. A chromic  acid cleaning solution is required for removing all traces of
organic material from glassware.

8.5  Reagents and Chemicals

The  minimum  purity  of  reagents and chemicals  should  be analytical reagent  grade.
Analytical standards  should be reference grade, when available. The analyst should take
special note of  the assay of less pure materials (most often pesticides). All reagents and
chemicals should be stored according to manufacturer's instructions to prevent degradation.
Proper storage is especially important if the chemical is to be used in preparing an analytical
standard. Refrigerated chemicals should be allowed to come to room temperature before
exposing them to the atmosphere.

When preparing stock solutions, it is recommended that at least 0.100 gram  of material be
used  for greater accuracy in weighing.  Solutions should be carefully stored so as to preserve
their  concentration, and to protect  them  from  ultraviolet radiation.  Usually storage  in
ground-glass stoppered bottles, either amber-colored, or out of the line of direct lighting, is
sufficient.

Standard solutions should  be prepared using precision syringes, preferably  equipped with a
Cheney adapter, to measure the volume of stock solution to be diluted. The syringe barrel
should be pre-wetted with solvent and air bubbles expelled. Dilution should be done in a
Class  A volumetric flask to insure accurate measurement. If these solutions are to  be used
frequently, they are best stored in a screw-cap, septum, sealed vial. These vials allow instant
access to the solution  and  offer good  protection against concentration changes  of the
standard solution. Evaporation of the solvent caused by repeated removal of the  cap is a
                                          8-2

-------
serious problem with other containers. If septum vials are not available it is advisable to
prepare standard solutions in a volumetric container of 100 ml or more and transfer a small
portion to a separate container for daily use, then discard that portion at the end of the day.

All stock or  standard  solutions  should  be  carefully watched for  signs of changes in
concentration or deterioration. As an aid to monitoring these solutions, it is wise to label
them as to compound, concentration, solvent used, date,  and preparer. Also, in the case of
GC solutions, it is necessary to retain some evidence of its chromatographic behavior as a
fresh solution for comparison at a later date.

Distilled water  used as dosed, control samples must be free of organic interferences. A very
effective way of  removing organic interferences from distilled water  is to pre-extract the
water with the  solvent that is to be used in the analysis, then boil the water to remove the
residual solvent.

Organic solvents used in pesticide analysis  should be pesticide quality, and demonstrated to
be free of interferences in a manner compatible with whatever analytical operation is to be
performed.  Solvents can be checked  by  analyzing a volume equivalent to  that used in the
analysis and concentrated to the minimum final volume. Possible interferences are noted in
terms of factors such as relative retention times, peak geometry, peak intensity, and width
of  solvent  response.  Interferences  noted under these  conditions  can  be  considered
maximum. If necessary, a solvent must be redistilled  in glass using a 60-cm column packed
with 1/8" glass  helices, or  an equivalent system.

Hexane - ethyl ether and  benzene  are commonly  used in the  extraction of water and
wastewater  in  conjunction  with analysis by  electron-capture gas-liquid chromatography.
Because  electron-capture detection  methods are  extremely sensitive  to interferences
normally found in these  solvents,  the cleanest possible reagent grade or pesticide-quality
solvents must be used.  Redistillation in the lab in an all-glass system  is usually necessary.
Experience  in  the Analytical Quality Control  Laboratory  has  shown Burdick-Jackson
hexane and  Baker Chemical Co. ethyl ether to be satisfactory. All solvents vary from lot to
lot. Therefore,  when a good lot is found, subsequent order should specify that lot number.
Solvents in the  same lot may also vary and therefore each container should be checked.

8.6 Common  Analytical Operations

Adequate steps must be  taken to  eliminate or minimize  interferences from solvents and
other materials. A blank should be run simultaneously under the same analytical conditions
as any block of samples analyzed. A block of samples is defined as any group of one or more
samples analyzed  using a common batch of  analytical supplies. Should  any one of the
supplies be changed (i.e. solvent, silica gel, Florisil, etc.) a new blank is required.

Quantitation of micro amounts of organic materials requires extremely careful technique to
avoid loss of sample. Quantitative transfers  are essential to obtain accurate and precise
results. Practice in these manipulations is recommended for the inexperienced analyst.

Concentration  of sample  extracts to  very small volumes for  micro analysis requires great
care to avoid loss of constituents. AKuderna-Danish evaporator is a very useful apparatus to
accomplish  this operation.  Instructions  in  the use of this  evaporator  must be strictly
followed to  avoid  loss of  desired sample (1). Final concentration in an ampule or calibrated
tube is accomplished in a warm  water bath with a gentle stream  of  clean, dry air, if air
                                         8-3

-------
oxidation  is  not  a problem;  otherwise, nitrogen  should  be  used.  During  the final
concentration,  the  inside walls  should  be rinsed repeatedly  with the working solvent  to
insure  the  total sample is contained in the  bottom of the tube.  Complete evaporation  of
solvent must  be avoided  to prevent loss  of sample  constituents.  The step should  be
accomplished within 10 or 12 minutes for best results.

8.7  Gas-Liquid Chromatography

To  obtain reproducible results, it is necessary to have very accurate control of the column
oven temperature. The temperature should be reproducible within ±1°C, and have minimum
gradients of 2°C throughout the oven. For temperature-programmed operation, a low mass
oven is required to allow rapid heating and cooling of the  column. Most manufacturers
produce  gas  chromatographs  which meet these  requirements. Many organic compounds
decompose when they come in contact with hot stainless steel. For this reason the injection
block should be capable of accepting a quartz or glass tube to prevent the compounds from
contacting  hot metal.  Memory  peaks  observed  using  direct,  aqueous-injection, gas
chromatography  can be eliminated or greatly reduced by  employing direct on-column
injection. A Bio-med injection kit manufactured by Tracor Instrument Co. is very  suitable
for  this purpose.

The septum should be changed at the end of each day's use. Changing at the end of each day
allows  overnight purging of the system of any bleed-off of contaminants from the  septum.
To  avoid most of this bleed-off  the septa can be  preconditioned by heating  at 250°C in a
vacuum oven for two hours.

The gas chromatograph should be equipped with accurate needle-valve, gas-flow controls. If
these flow  controls are not previously calibrated this can easily be done using a soap-bubble
flow meter and a stopwatch.

The nature of constituents  to be measured dictates the  type of detector  to be employed.
The electron capture detector is extremely sensitive to electronegative functional groups and
substituents,  such as: halogens, conjugated carbonyls, nitrites, nitrates, and organometals. It
is virtually insensitive  to hydrocarbons,  alcohols, and ketones. The selective sensitivity  to
halides makes this detector particularly valuable for the analysis of many pesticides.

Electron capture detectors employing two sources of ionization are available:  tritium (H3)
and radioactive nickel (Ni63). Each has its advantages. Most H3  detectors possess greater
sensitivity  than Ni63 detectors.  However, the H3  detector is limited to 225°C because  of
radioactive leakage.  This temperature limitation  makes  the  H3 detector susceptible to a
buildup of high-boiling contaminants which reduce its sensitivity. The Ni6 3 detector can  be
operated up to 400°C to prevent this buildup  of contamination.

The microcoulometric detector  is  specific  for  halogen-, sulfur-, or nitrogen-containing
compounds depending on the conditions used. Although the sensitivity  of this detection
system is not as great as others available, the extreme specificity  makes it a very  valuable
device  in indentification, minimizing the need for cleanup of the extract.

Another selective detector,  with about the same sensitivity for chlorinated  compounds as
microcoulometric titration, is the Coulson electrolytic conductivity detector. The specificity
of this detector makes it very useful in the identification of pesticides, and sample cleanup
and pre-treatment are less critical. However, if chlorinated compounds are  being detected, a
                                         8-4

-------
scrubber tube  must  be used  with this instrument to  remove  any SCL  produced by
sulfur-containing compounds in the sample.

The flame ionization detector (FID) responds to virtually  all compounds. Some notable
exceptions are  air,  water, carbon  disulfide, and the fixed gases. This makes the FID very
useful for  direct- aqueous-injection  gas chromatography. However, since the detector is
sensitive to  such  a wide  range of compounds,  it is  also subject to interference from
extraneous material. This means that the extract must be cleaned up  considerably before
analysis. The flame photometric detector (FPD) is used to analyze residues of phosphorus-
and sulfur-containing pesticides and their metabolities.  Little  or no cleanup of the sample
extract is required, and extraneous material causes no appreciable interference.

The alkali flame detector (Thermionic)  can also be used in the analysis of organophosphate
pesticides.  Even though this detector gives an enhanced  response to phosphorus, a large
amount  of extraneous  material  demands  that the extract  be  cleaned  up  before  the
phosphorus-containing constituents are quantitated. An  enhanced response to sulfur is  not
obtained with the alkali flame.

A one-millivolt, one-second,  full-scale-response, strip-chart recorder  should be  used to
maintain a permanent record of the results.

The type of detector employed dictates the type of  carrier gas that must be used. Nitrogen
can  be used for the  FID, electron-capture,  flame-photometric, and microcoulometric
titration detectors.  The Analytical Quality Control  Laboratory  has used nitrogen from
various suppliers, and has found the J. T. Baker Chemical Company's extra dry grade to be
satisfactory. However, to avoid the risk of system or detector contamination from materials
possibly present in the gas cylinder, the cylinder should  be replaced when the tank pressure
reaches 200 psi. It is also recommended that some type of gas purifier which is composed of
a molecular sieve, desiccant, and filter be used on all combustion gases and carrier gases used
with electron capture detectors.

A precision, gas-tight,  microliter syringe, which can  be accurately  filled,  will  deliver
reproducible injections,  and  can be easily and thoroughly  cleaned,  is recommended for gas
chromatographic use.  A Teflon plunger  seal to  prevent backwash, and no dead volume are
desirable features.  These types  of syringes are available from various suppliers,  but  the
Glencoe syringes have additional design features to prevent  bending plungers.

It is not possible to enumerate here the many  types and applications of liquid phases  and
solid  supports.  The literature (2,  3) will provide helpful guidance in the selection of these
materials.

The type of column tubing required to  perform a specific GC  analysis  is prescribed by the
method  for  that analysis.  Aluminum  has  been  found suitable  for chlorinated  organic
pesticides and stainless  steel for phenols. Generally  1/8" or 1/4" OD tubing is used. The
1/4" column is capable  of accepting larger injections but the 1/8" column is more efficient
at chromatographing small injections. It should be noted that  the use of glass columns will
prevent the degradation  of material often associated with metal  columns.

Silanized glass  wool should be used to minimize degradation  or absorption of organic
compounds  by the glass-wool  plugs. Glass wool  can be silanized by treating with  10%
                                         3-5

-------
dimethyldichlorosilane in toluene for 10 minutes, or the prepared material can be purchased
from Applied Science Laboratories.

Column  packings  can be  prepared  by the analyst, or  purchased already prepared  and
conditioned.  If prepared in the laboratory a highly refined technique should be developed
by  the  analyst.  A  well prepared  packing  is essential  to an acceptable chromatographic
analysis.  Particular  care  should  be taken  to accurately measure  loadings,  uniformly
distribute the liquid phase, and preserve the structure of the fragile solid support.  One
method used  successfully is the beaker slurry method in which  the  proper amounts of
stationary phase and solid support are mixed together in a solvent  in a beaker. The solvent
can be evaporated by immersing the beaker in a warm  water bath  or filtering it through a
Buchner funnel. The mixture is then dried at 110° C to remove residual solvent. A fluidized
dryer (4) may also be used to remove the remaining solvent. Details on techniques for
preparation  of  Teflon packings may  be  obtained  from Applied Science  Laboratories
(Bulletin FF124).

An important consideration in packing a column is obtaining  a uniform density, not so
compact as  to restrict gas  flow,  and not so  loose  as to create  voids during use. Also
important is the need to exercise care not to crush particles during packing.

It cannot be  overemphasized that column  conditioning is essential  to obtain acceptable GC
analysis.  Proper  conditioning further distributes the liquid phase over available active sites,
and removes excess liquid phase that may bleed off the column and  impair GC performance.
Unstable performance  in the form of baseline  drift, varying sensitivity, and wide solvent
peaks are usually  the result of column bleed due  to improper  conditioning. This  also
contaminates the detector and makes frequent cleaning necessary.

A good general procedure for conditioning columns is as follows: install the column, leaving
it disconnected from the detector; heat the column in the GC oven to just above maximum
recommended operating temperature for the liquid phase, without gas flow, for  2 hours;
follow with a half-hour equilibration period at a temperature at least 40 degrees below the
maximum recommended temperature, still without flow; raise  the temperature 20 degrees
above  operating  conditions  (do not  exceed the  maximum recommended operating
temperature  for the liquid phase), adjust  gas flow  to 50 ml/min,  and maintain these
conditions 24 to 48 hours.

There  are  many  factors  that  must be considered  in any attempt  to obtain  optimum
operating conditions. As indicated, optimum conditions begin with the proper selection of
column materials including tubing, glass wool, liquid phase, and  solid support. Each method
dictates the selection of these  items including  the percent liquid  phase. Liquid load and
operating temperature  are  critical  to  optimum resolution and elution  rate. Uniformly
coating the  solid  support, properly packing the column, and properly conditioning the
column all contribute greatly to attaining optimum conditions. Improper attention to these
factors contributes to either column  bleed, which fouls the detector, or tailing peaks, which
prevent resolution of closely eluting compounds.

The system must be free of gas leaks since these affect sensitivity  and reproducibility. The
proper selection of a carrier gas is important.  Injection blocks should be of proper  flow
design, and should be kept clean. Temperature  control of injection port, column oven, and
detector must be accurate and  constant, and must be at equilibrium to obtain reproducible
results.
                                         8-6

-------
Considerable attention must be given to those factors affecting detector response, such as
gas flows and regulated current. Also, standard solutions and condition of electrodes for
MCT detectors, condition of radioactive source for EC detectors, and ratio of combustion
gases in flame ionization detectors must be carefully attended to.

The gas chromatograph must either be operated within the linear range of the detector, or a
suitable  calibration  curve  must be employed.  The  linear range can be  determined by
chromatographing different  amounts of a compound and observing the response. The peak
area of the responses should  be  proportional  to  the amount  of compound  injected
throughout  the  full range of the recorder.  If these conditions are not fulfilled  at  the
attenuation  investigated, a more sensitive attenuation should be selected until the linear
range is found.

Once having obtained optimum operating conditions,  the analyst must assure sustained
optimum performance through a  routine  maintenance  program,  as  prescribed in  the
instrument manual. Such common practices as inspecting for gas leaks, changing the septum
after daily use, replacing gas tanks before they run too  low, and keeping a close check on
injection block, oven, and detector temperatures must be performed frequently.

Column  performance can be monitored by observing daily response to a selected group
standard, and comparing it to response to that same  standard under previously determined
optimum conditions. Changes in elution pattern, relative proportions of peaks, and peak
geometry are signs of  a deteriorating column if the rest of the system has been  properly
maintained.  A column should be replaced as soon as deterioration is observed.

A good syringe-handling technique is important when doing GC analysis. Before measuring
the volume  to be injected into the gas chromatograph, first wet the barrel of the syringe and
expel all air  bubbles. The volume injected is determined by drawing a selected portion of the
extract entirely into the glass,  calibrated, syringe  barrel, noting the volume, injecting the
sample, partially withdrawing the plunger, and measuring the liquid remaining in the syringe
barrel. Injections  are made rapidly after the needle is in the gas chromatographic injection
port. The needle  is immediately removed from the system to prevent volatilization of any
sample  in the needle.  This method improves the analytical accuracy since any  absolute
volume is injected and  reproducible technique is not required. It also eliminates the built-in
bias of injecting slightly more than expected each time, when direct barrel readings are used
exclusively.

The standard solutions can  be more accurately used if prepared at a concentration  that
allows injection volumes similar to those of the sample.

8.8  Qualitative Analysis

The retention time  (R() of a component on a given column,  under given conditions, is
characteristic of that particular component, and is used for qualitative identification. The
standard way of reporting retention data is to give the relative retention (RRt) defined as R
(component) -r  R( (reference compound). The retention time may be measured as the time
elapsed from an unrestrained solute peak elution, to the apogee of the peak of interest. This
works  well when a flame ionization or electron-capture  detector are used. However, when
the microcoulometric titration  or the flame  photometric detector are used (the  flame is
extinguished by  the injection), no unrestrained solute peak is observed. In  such cases,  the
injection point  is manually  or  electrically marked, and used as the point of reference for
                                         8-7

-------
retention times.  Caution must be exercised when manual marking is practiced, so that an
accurate reference point is provided.

Tentative identification is made by matching the relative retention time of the unknown
component  with that  obtained from  a known compound analyzed  under  identical
conditions, provided  that  peak geometry  is also similar.  It must be pointed  out  that
similarity  of  retention  time  and  peak geometry  to a  known compound  does  not
unequivocally  identify an unknown. Additional gas chromatographic columns of different
polarity and  other  detectors may be used for confirmation. If  retention time and peak
geometry match on two or more columns, the identity is corroborated. However, further
corroboration  using analytical tools such as infrared  spectroscopy, mass spectrometry, or
thin-layer chromatography should be used whenever possible.

Proof of identity of compounds which produce a multi-peak response should be evidenced,
not only by relative  retention and peak geometry, but also  by  the  correct number and
relative proportion of each peak in the  total chromatogram. This is called a "fingerprint"
comparison of the known standard chromatogram with that of the  unknown constituent.


8.9  Quantitative Analysis

The quantitative interpretation of a gas chromatogram is based either on the peak height or
the peak area. The area measurement is generally preferred because peak height is extremely
sensitive  to  small  changes  in  the  operating  conditions,  particularly  in  the  column
temperature. However, in chromatograms where the peaks are extremely sharp and narrow,
the error involved in the area measurement makes height measurements more reliable.

Peak  area measurement  should  be  carried out using height  x width  at half height, disc
integrator, or electronic digital integrator. These techniques are rapid and  simple, and give
good  results with symmetric peaks of  reasonable width. The use of a planimeter, although
less precise than other methods, is presently found to be the best method for measuring the
area of unsymmetric peaks which do not originate at the baseline. Precision is improved by
tracing each peak several times and taking an average value. The electronic integrator is also
recommended for those unsymmetric peaks which originate at the  original baseline.
Concentrations of constituents are determined  from standard  calibration curves obtained
under identical conditions. An absolute calibration curve is obtained from peak areas or
peak  heights  plotted against  known weights  of  a compound chromatographed under
identical conditions. These  standard injections must be made during the sample run to
detect any  change in instrument  conditions  or  response  which would invalidate  the
calibration.

The use of an internal standard is the most accurate method of quantitating constituents in
a sample. A calibration  curve can be obtained by simultaneously chromatographing  the
previously  identified  sample  constituent  and  a standard,  in  known weight ratios, and
plotting the weight ratios versus area ratios. An accurately known amount of the standard is
then added to the unknown sample and the mixture chromatographed. The area ratios are
calculated, and the weight ratio of the sample constituent to the standard is read from the
curve. Since the amount of standard added is  known, the amount of the sample constituent
can be calculated. Using this method, injection volumes need not be accurately measured

-------
and detector response need not remain constant since any change in response will not alter
the ratio.

If the following requirements can be met, the internal standard is the preferred method of
calibration. The internal standard must be well resolved from other peaks, must elute close
to peaks of interest, should approximate the concentration of unknown, and should have
structural similarity to unknown.

8.10 Thin-Layer Chromatography

Special equipment for thin-layer chromatography as supplied by the Brinkman Instrument
Company has been found by the Analytical Quality Control Laboratory to be particularly
convenient to use. Adsorbents supplied by Warner-Chilcott Laboratories, particularly Silica
Gel G and Alumina G, are found to be free of impurities. In any case, it is best to check  the
adsorbent for impurities, and discard it if necessary. Plates precoated by the manufacturer
can often be used when quantitation is visual; however, trials of plastic-backed plates in  the
Analytical Quality Control Laboratory have shown that they often develop unevenly and
very slowly. The use of precoated glass-backed plates is therefore recommended.

Prepared plates should be carefully inspected for flaws that might interfere with proper
chromatographic development. Plates should be marked in such a way that any grain lines
would be perpendicular to the direction of development when used. Prepared plates are best
activated by heating at 110°C in a convection oven for 30 minutes. Phosphorous pentoxide
is recommended as the desiccant for use when storing plates.

Two important points to remember before using the developing chambers are to give  the
chamber atmosphere time to become saturated with solvent vapor, and to keep the chamber
in constant temperature surroundings shielded from drafts. Failure to observe these points
could result in an erratic chromatographic development, generally off line.

When spotting  a plate, care must be taken not to overload the spot. Overloading results in
poor chromatographic efficiency. Also, when applying a gentle stream of air or other gas to
evaporate the solvent,  care must  be taken not to blow away any of the layer or sample
droplets. The analyst should also be aware of whether or not the use of air to evaporate  the
solvent could cause any oxidation of the sample material, thereby causing erroneous results.
If oxidation is a problem an inert gas should be substituted.

When spraying a plate with aqueous  or other non-volatile sprays, care should be taken not to
soak the plate. Soaking  with such sprays  may  cause  spots  to run as  the  plate  stands
vertically. When using any spray it is better  to use a stronger solution, then apply increased
amounts to develop the spots. Overspraying can produce a background color which impairs
the visibility of the developed spot.

Quantitation by visual  comparison  of sample  response  to a  series  of  standards is
semi-quantitative at best. However, by making  the spot size the same for samples and
standards,by using the same  solvent for both, and by using careful development techniques,
a fair degree of reliability can be obtained, when comparing the factors of  spot size and
color intensity.

When zone  scraping  and collecting  the  layer material with the aid of a  vacuum,  the
collection apparatus is usually plugged with  glass  wool to trap the adsorbent.  Since poorly
                                         8-9

-------
packed glass wool can cause high loss of material by failure to trap the dust, the scraping of
the periphery of the zone should be done first. A clear tygon vacuum line, or a glass section
near the end of the tube makes it easy to monitor for untrapped adsorbent.
8.11  Column Chromatography

The features of a chromatographic column which define its utility are shape, liquid capacity,
and elution rate. The liquid capacity is principally a matter of convenience and should be
weighed against the increased labor required for cleaning equipment. A flow-control device
is  critical to an efficient separation because increased flow  decreases resolution between
emerging components. A Teflon, flow-control stopcock  is required because lubricant will
produce an analytical interference.

Most adsorbents used in column chromatography are preactivated by the manufacturer and
shipped in air-tight  containers. Storage in the laboratory  should also be in an air-tight
container because moisture will greatly affect the activation state of an adsorbent.  If a
second activation step is required, the material should be heated at 130°C for at least five
hours.  The  activity of the  adsorbent  can  be monitored  by  eluting  a  mixture  of
chromatographically pure  dyes (5). The  elution rate and  degree of separation  of the
individual dyes  is a function  of the activation  state  of the adsorbent. This enables the
analyst to accurately attain the same  activation for different batches of adsorbent.

When the adsorbent  is added to the  column,  gentle tapping or a vibrator should be used to
settle the material. This minimizes the space between particles and prevents channeling of
the eluting solvent through the adsorbent which reduces separation efficiency.

Liquids should be added slowly, down the inside wall of the column, to avoid disturbing the
packing surface. Mixing of the  solution above the adsorbent  with the fresh eluting solvent
can be minimized  by introducing the new solvent just as the last of the solution reaches the
packing surface. The column top must not go dry  or air may be introduced which will lower
the separation efficiency of the  system.

Before addition of the sample,  columns should be pre-eluted with 50-75 ml of the solvent
prescribed by the procedure. This is done to  remove trapped air and to clean the column
material  of trace  contaminants.  During this  pre-elution  it is often necessary to tap the
column to free all trapped air, especially if a volatile solvent is  used in the pre-elution.


8.12  References

1.   Gunther, F. A., et al, Anal.  Chem., 23, No. 2, p. 1835 (1951).

2.   Lynn, T. R., et al, Guide to Stationary Phases for Gas Chromatography. Analabs, Inc.,
    1968.

3.   McNair, H. M. and Bonelli, E. J., Basic Gas Chromatography. Varian Aerograph, 1969.

4.   Kruppa, R. F., et al, Anal. Chem., 39, 851  (June 1967).

5.   Brockman, H. and Schodder, H., Chem. Ber., 74, p. 73 (1941).
                                        8-10

-------
                                      CHAPTERS

                                 SKILLS AND TRAINING
9.1  General

Analytical  operations in  the  laboratory can  be  graded according  to the  degree  of
complexity. Some analyses require no sample treatment, with the measurement performed
in  minutes on  a  simple  instrument.  Other  determinations  require  extensive  sample
preparation prior to complex instrumental examination. Consequently, work assignments in
the laboratory should be clearly defined. Each analyst should  be completely trained and
fully understand all the assignments of his job before being given new responsibilities. In this
regard,  all  analysts, sub-professional or professional,  should  be thoroughly instructed in
basic laboratory operations, according to the degree  of professional maturity. Some of the
basic operations that should be reviewed periodically with laboratory personnel follow:

    a.   Sample Logging

        Emphasize the routine procedure for recording of samples entering the laboratory,
        and assign primary responsibility. Establish what information is required, and how
        sample is  routed to analyst. Discuss stability of samples, and how they should  be
        stored prior to analysis.

    b.   Sample Handling

        The  analyst  should  understand  thoroughly when the sample is  to be settled,
        agitated, poured, pipetted, etc.,  before removal from the  container.

    c.   Measuring

        The analysts, especially new employees and  sub-professionals, should be  instructed
        in  the use of volumetric glassware. The correct use of pipettes and graduates should
        be emphasized as discussed in Chapter 4.

    d.   Weighing

        Because almost every measuring operation in the analytical laboratory is ultimately
        related to a weighing operation, the  proper use of the analytical balance should be
        strongly   emphasized.   Maintenance   of   the  balance,   including   periodic
        standardization,  should be reiterated to all personnel.

    e.   Glassware

        All glassware  should  be washed and  rinsed  following  the requirements  of the
        analysis  to be performed. Not  only must the personnel assigned to this task  be
        instructed, but all lab personnel  should know the routine  for washing glassware, and
        also special requirements for particular uses. In addition, the precision tools of the
        laboratory such as  pipets, burets, graduates, Nessler tubes, etc., should be inspected
        before use for cleanliness, broken delivery tips, and clarity of marking. Defective
        glassware should be discarded or segregated.
                                         9-1

-------
In summary, quality control begins with basic laboratory techniques. Individual operator
error  and  laboratory error can  be minimized if  approved techniques are consistently
practiced. To insure the continued use of good technique, laboratory supervisors should
periodically review  the basic techniques with each analyst and point out, when necessary,
areas of needed improvement.
Continuing improvement of technical competence for all laboratory personnel is, of course,
the final responsibility of the laboratory supervisor. In a well organized laboratory, however,
a big brother attitude of higher ranking to lower grade personnel should be encouraged; each
person should be eager to share  experience, tricks-of-the-trade, special skills,  and special
knowledge with subordinates. Obviously, improved efficiency and improved data  quality
will result.
9.2 Skills

The  cost of data production  in the analytical laboratory is  based largely upon two
factors—the pay scale of the analyst, and the number of data units produced per unit of
time. However, estimates of the number of measurements that can be made per unit of time
are difficult, because of the variety of factors involved. If the analyst is pushed to produce
data at a rate beyond his  capabilities, unreliable results may be produced. On the other
hand, the  analyst should be under some compulsion to produce a minimum number of
measurements per unit of time, lest the cost of data production become prohibitive. In the
following table,  estimates  are given for the  number of determinations that an analyst
should be expected to perform on a routine basis. The degree of skill required for reliable
performance is also indicated. The arbitrary rating numbers for the degree of skill required
are footnoted in the tables, but are explained more fully below:

    a.   Rating  1—indicates an  operation  that can be  performed  by  a  semi-skilled
        sub-professional with limited background; comparable to GS-3 through GS-5.

    b.   Rating  2—operation requires  an  experienced  aide  (sub-professional)   with
        background in general laboratory technique and some knowledge of chemistry, or a
        professional  with modest training  and experience; comparable to GS-4 through
        GS-7.

    c.   Rating 3—indicates  a complex procedure requiring a good background in analytical
        techniques; comparable to GS-7 through GS-11.

    d.   Rating  4—a  highly involved  procedure  requiring  experience   on  complex
        instruments; determination requires specialization by analyst who interprets results;
        comparable to GS-9 through GS-13.

The time limits presented in the table are based on use of EPA methods.

A tacit  assumption  has been  made that multiple  analytical  units  are  available for
measurements requiring special equipment, as for cyanides, phenols, ammonia, nitrogen and
COD. For some of the simple instrumental or simple volumetric measurements, it is assumed
that other operations such as filtration, dilution or duplicate readings are required; in such
cases the number of measurements performed per day may appear to be fewer than one
would  normally anticipate.
                                        9-2

-------
                                    Table 9-1
         SKILL-TIME RATING OF STANDARD ANALYTICAL OPERATIONS
        Measurement
       pH
       Conductivity
       Turbidity (HACK 2100)
       Color
       DO (Probe)
       Fluoride (Probe)
Skill Required (Rating No.)

  (Simple Instrumental)
         1
         1
         1
         1
         1,2
         1,2
       Alkalinity (Potentiometric)
       Acidity (Potentiometric)
       Chloride
       Hardness
       DO (Winkler)
       Solids, Suspended
       Solids, Dissolved
       Solids, Total
       Solids, Volatile
       Nitrite N (Manual)
       Nitrate N (Manual)
       Sulfate (Turbidimetric)
       Silica
       Arsenic
   (Simple Volumetric)
         1
         1
         1
         1
         1,2

   (Simple Gravimetric)
         1,2
         1,2
         1,2
         1,2

   (Simple Colorimetric)
         2
         2
         2
         2
         2,3
No./Day
100-125
100-125
 75-100
 60-75
100-125
100-125
 50-75
 50-75
100-125
100-125
 75-100
 20-25
 20-25
 25-30
 25-30
 75-100
 40-50
100-125
100-125
 20-30
SKILL REQUIRED
1 - aide with minimum training, comparable to GS-3 through GS-5
2 - aide with special training or professional with minimum training,
    comparable to GS-5 through GS-7.
3 - experienced analyst, professional, comparable to GS-9 through GS-12.
                                      9-3

-------
                                Table 9-1 (continued)
          SKILL-TIME RATING OF STANDARD ANALYTICAL OPERATIONS
      Measurement
       BOD
       COD
       TKN
       Phosphorus, Total
       Phenol (Dist'n only)
       Oil & Grease (Soxhlet)
       Fluoride (Dist'n)
       Cyanide
    Skill Required (Rating No.)
(Complex, Volumetric or Colorimetric)
              2,3
              2,3
              2,3
              2,3
              2,3
              2,3
              2,3
              2,3
        (Special Instrumental)
              2,3
No./Day

  30-40*
  25-30
  25-30
  50-60
  20-30
  15-20
  25-30
  10-15

    150

  60-80

    3-5

    2-4
       Metals by AA
       (No preliminary treatment)
       Metals by AA                    2, 3
       (With preliminary treatment)
       Pesticides by GC                  3,4
       (Without cleanup)
       Pesticides by GC                  3,4
       (With cleanup)

       SKILL REQUIRED
        2 - aide  with  special  training  or  professional  with  minimum  training,
            comparable  to GS-5 through GS-7.
        3 - experienced analyst, professional, comparable to GS-9 through GS-12.
        4 - experienced analyst, professional, comparable to GS-11 through GS-13.
        * - depends on type of sample.

9.3 Training
For more experienced, higher grade  personnel, formal training  in special fields, possibly
leading to specialization, should be  almost  mandatory. Such  training can be fostered
through local institutions and through the training courses provided by the Environmental
Protection Agency. Regional policies on after-hours, government-supported training should
be properly publicized.

Formalized, training for lower grade personnel,  comparable  to GS-3 to GS-5, is relatively
scarce, How.ever, skills can be most efficiently improved at the bench level on a personal,
informal  basis by more  experienced analysts  working in  the  same  area.  Exposure to
pertinent literature should also be a definite program policy.
                                        9-4
                                                U.S, Government Printing Office: 1975—626-593/217

-------