EPA-600/2-76-283
December 1976
Environmental Protection Technology Series
MEASUREMENT TECHNIQUES FOR INORGANIC
TRACE MATERIALS IN CONTROL SYSTEM
STREAMS
Industrial Environmental Research Laboratory
Office of Research awl Development
U.S. Environmental PrWectien Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-283
December 1976 .
MEASUREMENT TECHNIQUES
FOR INORGANIC TRACE MATERIALS
IN CONTROL SYSTEM STREAMS
by
J.A. Starkovich, R.F. Maddalone, M. L. Kraft,
C.A. Zee, C. Lin, and C.A. Flegal
TRW Systems Group
One Space Park
Redondo Beach, California 90278
Contract No. 68-02-1393
ROAP No. 21AFC-004
Program Element No. 1AB013
EPA Project Officer: Robert M. Statnick
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
CONTENTS
INTRODUCTION "I
CHAPTER I - GENERAL INFORMATION 2
1.1 Sampling Test Planning 2
1.2 Flue Gas Sampling 2
1.3 Liquid and Slurry Sampling 3
1.4 Solid Sampling 3
1.5 Analytical Test Planning 3
1.6 Laboratory and Sample Preparation 3
1.7 Analytical Procedures 4
CHAPTER II -SAMPLING TEST PLANNING 5
2.1 Gathering Background Information . . . 5
2.2 Pre-Test Site Survey 8
2.2.1 Survey Team Logistics 8
2.2.2 Test Site Inspection 8
2.2.3 Sampling Point Selection ..." 9
2.2.4 Flow Measurements and Grain Loadings 11
2.3 Test Scheduling and Logistics 11
2.3.1 Test Matrix Development 12
2.3.2 Pre-Test Personnel Briefing 12
2.3.3 Sampling Team Organization and Equipment 13
2.3.4 Equipment Assembly and Checkout 17
CHAPTER III - FLUE GAS SAMPLING 20
3.1 Special Considerations 20
3.1.1 Contamination and Alteration of Sample by Sampling
Train 20
3.1.2 Multiphase Sampling Requirements 22
3.1.3 High Volume Sampling Requirement 22
3.1.4 Aerotherm High Volume Stack Sampler 23
3.1.5 HVSS Components -Selection and Design
Recommendations 29
3.2 Special Field Guidelines for Trace Element Source
Sampling 36
3.2.1 Work Area and Contamination Considerations .... 36
3.2.2 Filter and Impinger Solution Preparation 37
3.2.3 Probe Liner Preparation, Installation and
Removal 38
3.2.4 Handling and Storage of Impinger Samples 39
n
-------
CONTENTS (Continued)
CHAPTER IV - LIQUID AND SLURRY SAMPLING 44
4.1 Considerations for Liquid and Slurry Samples 45
4.2 Equipment Survey and Recommendations 47
4.3 Sample Collection Separation and Storage 48
4.3.1 Clear Water 50
4.3.2 Water-Solid Streams 51
4.3.3 Water-Organic Streams 51
4.3.4 Water-Organic Liquid-Solid Streams 51
4.3.5 Summary 52
CHAPTER V - SOLID SAMPLING 53
5.1 Sampling Methodology and Equipment Survey 53
5.2 Statistical Determination of Sample Size 57
5.3 Sample Collection and Storage 58
5.4 Other Considerations 59
CHAPTER VI - ANALYTICAL TEST PLANNING 61
6.1 Laboratory Preparations 61
6.2 Data Review Points 63
6.3 Precision and Accuracy 64
6.4 Calibration 65
6.4.1 Factor Method 65
6.4.2 Short Curve Method \ ] 66
6.4.3 Additions Method (Recommended) '.'.'. 66
CHAPTER VII -LABORATORY AND SAMPLE PREPARATION 67
7.1 Labware Preparation 57
7.2 Particulate and Impinger Solution Sample Preparation ... 68
7.2.1 Probe Liner Sample 6g
7.2.2 Cyclone Sample M
7.2.3 Filter Sample '.'.'*'' la
7.2.4 Impinger Solution . . TQ
7.3 Preparation of Solid Samples 7Q
7.3.1 Grinding 7,
7.3.2 Drying '!
7.3.3 Ashing '.'.!'.'.'. 12
7.4 Dissolution ?2
m
-------
CONTENTS (Continued)
CHAPTER VIII -ANALYTICAL PROCEDURES 74
8.1 Atomic Absorption Spectroscopy (AAS) 74
8.1.1 Types of Interferences 78
8.2 Ancillary Group 82
8.2.1 Arsenic Analysis 82
8.2.2 Boron Analysis 84
8.2.3 Fluoride Analysis 85
8.2.4 Mercury Analysis 87
8.2.5 Sulfate by the Gravimetric Method 91
8.2.6 Sulfate by the Turbidimetric Method 94
8.2.7 Cyanide Analysis 96
8.2.8 Chloride Analysis 106
8.2.9 Chloride-Silver Nitrate Potentiometric Method . . 110
8.2.10 Chloride-Colorimetric Method 1H
8.2.11 Nitrate-Brucine Method 117
8.2.12 Nitrate-Phenoldisulfonic Acid Method 120
8.2.13 Antimony Analysis 124
8.2.14 Selenium Analysis 125
8.2.15 Phosphate Analysis 129
REFERENCES 133
IV
-------
FIGURES
Number
1 Survey report sample 6
2 Sampling team task assignments 13
3 Aerotherm high volume stack sampler schematic 26
4 Schematic diagram of Kapton liner inside probe tube .... 27
5 Model CVE sampler schematic 48
6 Typical separation schemes for process liquids 50
7 Pneumatic line sampler schematic 60
8 Planning logic flow chart 63
9 Cyanide distillation apparatus 98
10 Example of differential titration curve (end point is
25.5 ml) 113
11 Schematic arrangement of equipment for determination of
arsenic and selenium 127
-------
TABLES
Number Page
1 Equipment Inventory for Typical Two-Point, 20 Test Gas
Sampling Task 14
2 Chemicals and Laboratory Equipment for Typical Two-Point
Gas Sampling Task Involving 20 Tests I8
3 Principal Functional Advantages and Disadvantages of
Aerotherm HVSS Unit for Trace Element Sampling and
General Source Particulate Testing
4 Probe Materials and Selection Criteria .......... 30
5 Filter Material for Trace Element Sampling ........ 33
6 Impingers for Trace Element Sampling ........... 36
7 Sample of Analytical Test Checklist ........... 62
8 Compilation of Accepted Standard Procedures by
Element ......................... 75
9 Compilation of References for Recommended Procedures ... 76
10 Atomic Absorption Operating Parameters .......... 77
11 Concentration Ranges for Color Measurement ........ 115
-------
INTRODUCTION
This manual has been prepared for the Industrial Environmental Research
Laboratory of EPA in partial fulfillment of contract number 68-02-1393.
The manual is written for professionals who are involved with the tasks
of collecting and measuring trace inorganic materials in process streams.
The procedures have been chosen to provide good, overall accuracy needed
for engineering evaluations of control process performance. Lower accuracy
procedures suitable for environmental assessment purposes are not addressed.
The first objective in the preparation of the manual was to present an
overview of approaches and procedures which have been used with success in
the past. It is intended that these procedures be of general nature to the
greatest extent possible. However, the procedures are based on experience
gained in evaluating control systems for coal fired utilities. Problem
areas which were identified during the course of the contract are elabo-
rated in some detail. Several of these problems are considered critical
areas in which professional judgment is still required in conducting a test.
It is beyond the scope of this manual to present all procedures for use in
every situation. Additional volumes of this manual will follow, broaden-
ing both the procedures base and the applications descriptions.
-------
CHAPTER I
GENERAL INFORMATION
The purpose of this manual is to present procedures and methods for
sampling and analysis of gas, liquid, slurry, and solid process streams for
trace inorganic materials. The materials addressed in this manual include
the following elemental forms and anions: As, Ba, B, Be, Ca, Cd, Cl , CN ,
Cr, Cu, F, Hg, Mn, Ni, NO^, Pb, PO^3, Sb, Se, SO^2, Sr, V, and Zn.
In sampling for these materials, the major area of concern both in the
field and in the analytical laboratory is that of contamination and cleanli-
ness. Every step of the sampling task is subject to contamination. Metal
particles can be introduced through abrasion of sampling trains; sample
containers can adsorb certain elements; volatile metals can be lost at sev-
eral different stages; unclean glassware or apparatus can introduce contam-
inants, and so on. Throughout this manual special attention is given to
these and other problems and requirements which pertain to the sampling of
trace inorganic materials.
The following sections briefly summarize the chapter contents and note
unique approaches to specific sampling and analytical problems.
1.1 SAMPLING TEST PLANNING
Source tests for trace elements must be planned in minute detail. The
constraints on pre-cleam'ng and packaging equipment, reagents, and sample
bottles are such that scrounging and borrowing at the test site cannot be
tolerated. In addition, the entire test crew must be briefed on the test
objectives, potential contamination problems, and proper procedures to
ensure that the test plan is carried out correctly.
1.2 FLUE GAS SAMPLING
Sampling gases is no doubt the most problematic and complex task.
Material compatibilities, contamination from abrasion, trace metal vola-
tiles, and inorganic background of reagents and filters are problems that
must be addressed in order to collect an accurate sample. The procedure
presented in this manual uses an Aerotherm high volume sampler, which is
-------
modified by lining the probe with a polymer film, using ultraclean filter
materials, and by selecting special sequential oxidative scrubbing solutions
for the impingers.
1.3 LIQUID AND SLURRY SAMPLING
Current technology is entirely adequate for this area of sampling.
Wastewaters have been analyzed for trace metals for years. The trains that
have been developed for this purpose take accurate samples and have been
designed to eliminate sources of contamination from within the train. This
chapter reviews the state-of-the-art methodology and selects and presents
the equipment and procedures applicable to trace elements.
1.4 SOLID SAMPLING
The variety of materials and sample sites that can be encountered in
sampling for solid materials is so diverse that it is impossible to consider
a single sampling procedure or train. Consequently, this chapter addresses
the advantages and disadvantages of various methods and equipment as they
pertain to trace element characterization.
1.5 ANALYTICAL TEST PLANNING
The accuracy of the final test data depends as much on the analytical
lab work as it does on correct field procedures.
Planning the organization and scheduling of laboratory work must be
based on test objectives and the relationship of the samples to each other.
Analytical data must be reviewed at several points during the laboratory
analysis to check accuracy and precision and to select the correct procedure
for the next step.
1.6 LABORATORY AND SAMPLE PREPARATION
Maintaining the cleanliness requirements throughout the sampling and
analysis scheme requires that the laboratory work area and all instruments
and glassware used be carefully prepared before beginning analytical work.
Procedures are presented for this preparation and for the preliminary sample
treatment and dissolution.
-------
1.7. ANALYTICAL PROCEDURES
In selecting analytical methods, special attention was given to those
capable of detecting 0.1 jag/ml, 1 yg/g, and/or 60 yg/M concentrations with
an accuracy and precision of ±10 percent. With few exceptions, all the
metals of interest can be analyzed by atomic absorption spectroscopy once
the samples are in solution. Those elements and radicals such as As,
SO^2, F, B, Sb, N0~, PO^3, CN", and Cl" that cannot be handled by AAS can
be analyzed by appropriately sensitive colorimetric methods. Atomic
Absorption Spectrometer is preferred for the bulk of the analyses since
this analytical procedure is common to most laboratories. However, due to
the complicated matrix effects that have been found in these sample solu-
tions, an AAS capable of background correction is absolutely necessary.
Therefore, only analytical laboratories which have this capability can be
employed.
The elements As, Se, B, and Sb can be analyzed by AAS when special tech-
niques are used to introduce the sample into the instrument However
very poor sensitivity is normally achieved for these elements when th*
direct aspiration procedure for sample introduction is used.
-------
CHAPTER II
SAMPLING TEST PLANNING
Before a sampling test is performed, detailed planning must be per-
formed to identify test requirements and anticipated problems. The source
test planning function can be divided into several overlapping steps; the
following sections discuss each step in sequential order. The scope of
each step depends on the magnitude, complexity, and familiarity of the test
program.
2.1 GATHERING BACKGROUND INFORMATION
Before traveling to a plant for the pre-test site survey, the survey
team members must become familiar with the chemical process they will sample.
This involves understanding the chemistry of the plant (chemical manufac-
turers, smelters, etc.), the type of fuel burned (gas, oil, coal, high or
low ash, high or low sulfur), and the pollution control process chemistry
employed (electrostatic precipitators, wet scrubbers or baghouse filters).
An understanding of all phases of the operation leads to initial choices of
possible sampling areas.
The location of all applicable process streams (gas, liquid, slurry)
should be determined. A schematic of the plant process that identifies
these streams should be obtained. Sampling areas can be defined from the
schematic and later can be translated into actual sampling points by the
pre-test survey team in the field.
Establishing a rapport with the plant personnel early ensures that
on-site problems can be solved rapidly and with minimum disruption of plant
activities. As an outgrowth of the pre-test survey planning, contacts with
plant personnel are established. It is extremely important, at an early
stage, to identify people or departments with whom the sampling teams will
be interfacing.
Figure 1 is a sample form for a pre-test survey report. This form
has been sectionalized for the logging of background information (prior to
survey trip) and field information (during survey trip).
-------
1
:IRM
LOCA1
BACKGROUND
PRE-TEST SITE SURVEY REPORT
DATE OF SURVFY
ION
PROCESS (PLANT)
(CONTROL)
OUTPUTS
SAMPLING AREAS
REMARKS
PLANT CONTACTS
NAME
1.
2.
3.,
A.
DATE OF TFST
POSITION
£
<
o
o
Ul
LL.
SAMPLING LOCATION (EPA METHOD 1)
SHOW DIMENSIONED SKETCH
DOWNSTREAM DIAMETERS - —
NEAREST FLOW x-^"
DISTURBANCE f
UPSTREAM DIAMETERS
TO NEAREST FLOW ,
DISTURBANCE \^
LIQUID FLOW RATE 1 /ft
RECTANGULAR
. •" CROSS SFfTTON.
^\ f EQUIVALENT DIAMETER
D^RLENGTH) (WIDTH) 1
L LENGTH + WIDTH J
7 1 TRAVERSE POINTS/PORT
RATIOS VFMTIIOI »D
SCRUBBER PH EQUIPMENT DOWNTIME
PROCESS MONITORS (TYPE, LOCATION, CALIBRATION STATUS, NUMBER OPERATING,
UNITS REPORTED*
Figure 1. Survey report sample.
-------
PRE-TEST SITE SURVEY REPORT
SAMPLING AREA WORKING ENVIRONMENT (NOXIOUS GASES; TEMPERATURE;
DUST LEVEL; RESPIRATORS REQUIRED; NOISE LEVEL; SAFETY; WALKWAYS;
PLATFORMS; SCAFFOLDING; LADDERS - OSHA STDS.)
WORKING AREA ACCESS
PULLEYS AND WINCHES REQ D DYES DNO FIXATION POINTS AVAILABLE DYES D NO
STORAGE AND LAB FACILITIES
ELECTRICAL OUTLETS
Q CIRCUIT BREAKER BOXES
td
EQUIPMENT CHECK LIST
D POLAROID CAMERA AND FILM D HARD HATS
D PITOT TUBE D SAFETY SHOES
D DRAFT GAUGE D RAIN GEAR
D THERMOCOUPLE D EARPLUGS
D VOLTMETER OR PORTABLE THERMOCOUPLE GAUGE D WATER JUGS
D ICE AND CONTAINER D SALT TABLETS
D BAROMETER D FIRST AID GEAR
D THERMOMETER D PORTABLE TAPE RECORDER
D GAS DETECTION TUBE KIT D U-TUBE MANOMETER
D PROBE - 13 MM
D TEE - 13 MM
D FITTING - 13 MM
D HOSE
D PUMP (SEVERAL LITERS PER MINUTE)
Figure 1. Survey report sample (continued).
7
-------
2.2 PRE-TEST SITE SURVEY
A decision to test a particular plant should be made soon enough to
allow a reasonable lead time before the actual test. It is recommended
that the pre-test site survey be completed at least two weeks before the
period of equipment assembly and checkout. This allows time for either
equipment modifications or the acquisition of special equipment.
2.2.1 Survey Team Logistics
The survey team should check on hotels, restaurants, and equipment
facilities available within an hour's ride of the sampling site. A vendor
for ice or dry ice should be located. Plane schedules for the nearest com-
mercial airport should be obtained. The equipment required for the survey
team is listed in Figure 1. (If an emergency requires additional equipment
to complete the test, the equipment can be shipped as airline baggage to
eliminate the restriction of freight office hours.)
2-2-2 Test Site Inspection
Upon arriving at the test site, the survey team should meet with the
plant engineer. Here, questions that arose from studying the schematic of
the plant can be addressed. All pre-selected sampling points should be
verified as areas where representative samples can be taken.
Process data such as fuel and air consumption, type of fuels, power
output, water consumption, system pressure, and temperatures at sampling
points can also be obtained from the plant engineer. For wet scrubber con-
trol processes, typical questions would concern liquid flowrate, L/G ratios,
AP in venturi, and pH of scrubber. Information concerning process schedules
and equipment downtime should be obtained. Another point to be investigated
with the plant engineer is the use of process monitors. What type of moni-
tors are used and where are they located? Have they been calibrated lately?
How many are operating and what are the units reported? Will this informa-
tion be available during the sampling test? This data will provide a record
of the particular process that will be operating during the time the sampling
test is performed. The availability and reliability of these monitors
directly affect the amount of monitoring equipment needed by the sampling
team. On a short-term basis, the availability of stack gas composition data
8
-------
will determine the need for the type of measurements to be made by the
survey team.
Under the guidance of plant personnel, a general tour of the plant
should be taken to clarify the plant's operational layout, as well as to
locate sampling points and their relationship to stairs, elevators, and
storage areas.
2.2.3 Sampling Point Selection
The selection of a sampling point for any type of sampling (gas,
liquid or solid) depends on four factors:
• Representative location
t Working environment
• Accessibility
• Sources of contamination
The most important issue of stack or duct sampling for gas or particu-
late matter is obtaining a representative sample; this is best done by
adhering to the specified test requirements. Therefore, the survey team's
prime consideration is to determine if a sampling point satisfies the
requirements of the specified test.^ ' In most cases, sampling ports are
not placed with EPA Method 1 requirements in mind, but are located for the
convenience of plant personnel. If a choice is possible, the requirements
of EPA Method 1 should be satisfied.
Typical liquid and slurry process streams (fossil fuel combustor or
coal processing plant) are closed piped systems, settling tanks, flyash
ponds, and slurry disposal areas. Many of the factors that apply in sam-
pling particulate-laden gas streams must also be considered in sampling
liquid and slurry streams, although their relative importance may vary.
Since a plant may have many liquid stream outlets from a given process, it
is important to determine the portion of the total outflow that the chosen
liquid stream represents.
Once the locations that will provide a representative sample are found,
the survey team should consider the working environment at each location.
Are noxious gases present? What is the temperature in the sampling area?
How heavy is the dust level? Will respirators be required? If so, for
-------
what gases? What is the noise level? Is the area safe (railing, lighting,
footing, etc.)? Is the area protected from the environment in the event of
rain or strong winds? All of these issues in some way affect the perform-
ance of the sampling team. The time to insist on safe scaffolding and lad-
ders is during the pre-test plant survey. All plant conditions should
meet OSHA standards.
Ease of access to a given source goes hand-in-hand with the test site
environment. Since sampling personnel must be able to carry their equip-
ment to the sampling point, elevators or stairways must be found that will
accommodate the team members carrying equipment. For heavy equipment, fix-
ation points for pulleys and winches must be located that are both safe to
plant personnel and equipment and close to the sampling point. If the sam-
pling location is not easily accessible, provisions should be made to
enlist help from plant personnel for the transport of the sampling equip-
ment (sometimes weighing as much as 200 to 300 pounds) from ground level
to the sampling location.
The selected sampling points must be fully characterized so equipment
needs and designs can be finalized before the sampling test. Polaroid snap-
shots can fix the relationship of beams, pipes, and obstructions to the sam-
pling point. Careful measurements should be taken so distances from obstruc-
tions, duct diameters, distances from duct obstructions, and port diameters
are well established. Notes should be taken on necessary equipment (lad-
ders, ropes, pulleys, etc.). If existing ports are to be used, the plugs
should be checked for rust and ease of removal. By removing the port caps,
the duct wall thickness can be checked and the type and quantity of gases
the sampling team will encounter can be determined.
Of extreme importance is the need for electrical outlets. Several
outlets on different circuits should be found that provide 115 volts and
20 amps. In addition, circuit breaker boxes should be located for the out-
lets. Plug connector requirements must be detailed.
Field sampling for trace materials requires extreme cleanliness in the
field as well as in the laboratory. In exposed sites, dust can be blown
into the impinger solution or filter while loading or unloading a sampling
train. The survey team must note the location of the sampling points with
10
-------
respect to plant laboratories or parking spaces adequate for an equipment
van so that provision can be made for sample handling in a closed environ-
ment. Because of the chance of contamination, the sample recovery area
(parking area for the van) should be located as close as possible to the
sampling points. If the plant has a laboratory, permission to use its
facilities should be obtained. A tour of the laboratory will clarify the
types and quantities of chemicals and equipment available. If equipment
is to be left overnight, a secure area must be found for its storage.
2.2.4 Flow Measurements and Grain Loadings
Besides characterizing the sampling point, the survey team should be
equipped to measure key stack parameters. However, the equipment required
is not elaborate, since the responsibility of the survey team is to provide
only approximate information about process conditions and pollutant con-
centrations. Stack temperature, gas composition, and grain loading should
be checked and grab samples of streams, settling ponds and solid wastes
should be obtained for survey level analyses.
The equipment necessary for these measurements should be kept to a
minimum. A pi tot tube and draft gauge are adequate to determine the pres-
sure head in the stack. A thermocouple can be attached to the pitot and
the readings can be obtained from a voltmeter or a portable thermocouple
gauge. Calibration of the thermocouple can be checked against an ice-water
bath (4°C or 39°F). If the pressure in the stack is less than 2.5 cm
(1 in.) of water, the draft gauge can be used to measure stack pressure by
attaching the static line to the gauge and leaving the other end of the
gauge open to the atmosphere. If the pressure in the stack exceeds this
amount, then a U-tube manometer is required. A barometer should be used to
measure the atmospheric pressure.
2.3 TEST SCHEDULING AND LOGISTICS
Several factors need to be considered in selecting a test date. The
choice of test conditions, personnel, and equipment will affect and deter-
mine scheduling. Time must be allowed for acquiring or modifying equipment
and developing special methods if the test program is unusual in scope or
complexity. Consideration must be given to the availability of other per-
sonnel (e.g., federal or state inspectors) who may wish to witness the
11
-------
sampling test. The test date must also be coordinated with plant manage-
ment to ensure that the process will be operating under the conditions
desired for the test. If the process operates continuously, the choice of
a sampling date can be a matter of convenience for all concerned. However,
if the process operates only a few hours a day or on an erratic basis, then
time also becomes a restraining factor in the experimental design. The
sample time for the selected method cannot exceed operation time of the
process. The testing dates must be planned with the plant personnel so as
not to conflict with construction and maintenance operations which would
produce nonrepresentative test conditions.
2.3.1 Test Matrix Development
Once the sample team has returned from the pre-test site survey, the
preparation of a test matrix should begin. Any samples the team has col-
lected should be analyzed for trace metal concentrations. These values,
together with the process information (was plant running at 50, 70, 100 per-
cent capacity), and the expected production schedules, will determine the
nominal sampling times necessary to obtain enough material for reliable
analytical results. Given the time budgeted for the test, a rough number
of time slots can be established. Rather than establishing rigid test
schedules, a flexible matrix of tests should be constructed. Developing
a test matrix allows the team coordinator to identify reasons for establish-
ing priorities among the various tests. The test matrix chart becomes a
"roster" and the team coordinator is the "manager" who, depending on the
situation, might wish to substitute a player.
Once the type and number of tests to be run are identified, a code
system should be established. Within the framework of the test matrix, all
chemical, bottle and filter requirements should be compiled and specific
material allocations and assignments made.
2.3.2 Pre-Test Personnel Briefing
As soon as the test matrix has been established, personnel and equip-
ment needs should be finalized at a pre-test briefing. This meeting should
be held with all the personnel associated with the sampling effort. An
alternate should attend this meeting, so that if a team member becomes sick
or is injured in the field, a replacement is available.
12
-------
The first meeting should brief the team on the expected test date, on
a projected daily schedule of time slots for sampling, on the objectives of
the test matrix, and on the physical appearance of the test points. The
Polaroid snapshots and rough sketches resulting from the pre-test site sur-
vey can be distributed for comment and recommendations. Even though speci-
fic assignments will be made at the meeting, each team member should realize
that he may be called upon to work any phase of the test program. Since all
facets of the test are interrelated, team members should be cautioned to
not become so involved in only their phase of the preparation that they are
unaware of or misunderstand activities and problems associated with other
phases of the test program.
2.3.3 Sampling Team Organization and Equipment
The normal breakdown of responsibilities among the team members is out-
lined in Figure 2. A checklist of items for field use must be compiled. A
typical checklist for a two-point, 20 test effort is shown in Table 1.
This list assumes that as much equipment assembly, such as lining the probes
with Kapton, as possible is done in the laboratory prior to leaving for the
field test. Additional items are needed in order to line probes in the
field; these items are listed in Section 3.2.3, along with directions for
insertion of the liners.
Another important equipment consideration is the fabrication of suit-
able shipment containers in order to ensure the safe transport of fragile
apparatus and to protect the collected samples from loss or contamination.
Individual "tote" boxes should be used to hold specific sampling train com-
ponents, and specially designed shipping boxes should be constructed to
TEST COORDINATOR
1 . GAS SAMPLERS
2. LIQUID SAMPLERS
3. SOLID SAMPLERS
1. TOOLS
2. SPECIAL EQUIPMENT
3. PROBE MANUFACTURE
1. PRE-TEST SITE SURVEY
2. TEST MATRIX
1. SOLUTIONS
2. LAB EQUIPMENT
3. LAB CHEMICALS
1. CLEANING SAMPLE
CONTAINERS
1. TARING FILTERS
EQUIPMENT CHECKOUT
• EQUIPMENT ASSEMBLY
CHEMICALS
BOTTLE WASH
FILTER TARE
Figure 2. Sampling team task assignments,
13
-------
TABLE 1. EQUIPMENT INVENTORY FOR TYPICAL TWO-POINT,
20 TEST GAS SAMPLING TASK
Quantity Description
Electrical Equipment
4 Eight-foot extension cords (12A)
1 Three-prong adapter
1 Six-outlet adapter
Miscellaneous electrical adapters
Box of fuses (20A)
1 Three-way plug
1 Two-way plug
3 Small variacs
2 Large variacs
8 Fifty-foot extension cords (12A)
1 Six-foot extension cord (12A)
1 Quad box with 20 feet of extension cord (30A)
1 Twenty-five-foot extension cord
1 Fifty-foot, 220 V extension cord
Electrical connectors
Electric tape
Spare light bulb for flood lamp
Light bulb for extension cord
3 1.5-watt walkie-talkies (3 channel)
Personnel Equipment (crew and spares)
Safety helmets
Pairs of coveralls
Ear protectors
Pairs of goggles
Dust respirators
Pairs of leather gloves
Pairs of asbestos gloves
Sets rain gear
First aid kit
(continued)
14
-------
TABLE 1. (CONTINUED)
Quantity Description
Aerotherm and Other Sampling Equipment
2 Aerotherm pumps
2 Controllers
2 Ovens in cabinet
2 Aerotherm horizontal mounting stands
20 Probes (0.5-inch stainless steel Kapton
lined) of appropriate lengths for test site
20 Bushing inserts
2 Standard pitot tubes (8-feet)
1 High velocity standard pitot tube plus
extensions to 8 feet
4 Caps for sampling with 1-1/4-inch hole in
center
10 Thermocouple connectors (spares)
1 Roll thermocouple wire (Chrome-Alumel)
(spare)
10 Heating tapes — 6-feet (spares)
1 CVE liquid and slurry sampler
Miscellaneous Equipment
1 200 feet of 1/2-inch rope
2 Six-gallon canvas buckets with 200 feet
nylon rope
12 4x4 wooden blocks
2 Styrofoam ice chests
1 Oxygen-propane torch
Sketch boards plus tablets
4 Boxes disposable towels
Assorted tarpaulins
1 Four-foot folding ladder
1 Camel hair brush
10 No. 7 stoppers (1/2-inch hole)
2 Paper notebooks
1 Suitcase
(continued)
15
-------
TABLE 1. (CONTINUED)
Quantity Description
Miscellaneous Equipment (continued)
1 Small tool box
Pliers
Tape measure
Epoxy glue
Assorted screwdrivers
Files
Allen wrenches
Calipers
Drill sets
Teflon tape
1 Large tool box
Assortment of wrenches (crescent, wrench
sets, etc.)
1 Solder gun
1 Voltmeter
1 Drill - 1/4-inch
1 roll Glass tape
1 roll Thermocouple wire Type K
1 roll Thermocouple wire Type T
Miscellaneous 1/2-inch fittings
1 Multi-speed Saber saw plus two blades
1 Saw
2 Lab jacks
1 Heat gun
1 Vacuum pressure gauge 9-30 psi absolute
1 Small portable vise
1 roll Black tape
1 roll Box tape
1 Stop watch
1 Pulley
C-clamps
WD-40 lubricant
1 Five-pound hammer
Socket sets
Sandpaper (various grades)
2 Twenty-four inch pipe wrenches
1 Thirty-six inch pipe wrenches
Sample Boxes
4 20 x 14 x 25-inch shipping boxes made of
3/4-inch plywood with aluminum reinforced
edges
2 16.5 x 16.5 x 18.5-inch shipping boxes made
of 3/4-inch plywood with aluminum reinforced
edges
20 11 x 6 x 16-inch module tote boxes
16
-------
protect the "tote" boxes and contents during shipment and to act as on-site
cabinets during field operations. Suitable tote and shipping boxes are
described in Table 1.
The chemicals task assignment assumes responsibility for collecting
assorted laboratory equipment (squeeze bottles, Nalgene graduates, etc.)
and chemicals (distilled water, acetone, etc.) to support all field activ-
ities. Reagents for the oxidative impinger solutions must be pre-weighed
or pre-mixed to correspond to the system established in the test matrix.
Table 2 lists the chemicals and lab equipment needed for a two-point,
20 test effort.
The bottle wash task is an important one. A special acid wash is
required for all bottles used to store samples collected in the field.
The chemical stores and bottle people must coordinate their efforts. A
count of the number of sample bottles needed plus reserves should be given
to the bottle person in time for the preparation of the proper number of
bottles.
The last responsibility to be delegated is the task of filter taring.
Normally, only one type of filter is used, but several could be needed
because of a possible variety in sampling environments. A code system
should be established for each type of filter. The filter person should
convey to the team coordinator a complete list of filter types and weights.
During a test, the filter used is noted on the data sheet and test matrix
form.
Throughout the pre-test period, it is important to brief the team on
the special requirements of trace metal sampling. Contamination in the
laboratory or contamination in the field both result in loss of data.
2.3.4 Equipment Assembly and Checkout
Once the individual tasks have been assigned to specific team members,
it is the responsibility of the test coordinator to monitor progress on
each task and prevent problems from developing through continual review of
work assignments and procedures. The test coordinator is also responsible
for initiating design changes and ordering special equipment to meet the
needs of the test matrix. For example, the sampling ports might require
17
-------
TABLE 2. CHEMICALS AND LABORATORY EQUIPMENT FOR TYPICAL
TWO-POINT GAS SAMPLING TASK INVOLVING 20 TESTS
Quantity
Description
8 1 (2 gal)
40 1 (10 gal)
450 g (1 Ib)
40 g (1.5 oz)
500 ml (1 pt)
500 ml (1 pt)
500 ml (1 pt)
25
100
12
2
2
2
4
2
4
Acetone, reagent grade
High purity, distilled water
(NH4)2S208
AgN03
30 percent HpO,,
Concentrated nitric acid
Concentrated hydrochloric acid
Sufficient tared filters in plastic bags for the
test matrix including a 30 percent reserve
Complete set of cleaned storage bottles for
both impinger solutions and liquid, slurry, and
solid samples (amount depends on test matrix)
Sheets, 46 cm x 61 cm (18" x 24") of Whatman
No. 1 filter paper to cover bench surfaces
500 ml (1 pt) precleaned Nalgene bottles
(spares)
Nalgene graduates, 50 ml
Nalgene graduates, 250 ml
Nalgene graduates, 500 ml
500 ml squeeze bottles
1000 ml squeeze bottles
Boxes disposable towels
special scaffolding to mount the sampling train for a vertical or horizontal
traverse. Since the test coordinator has firsthand knowledge of the site,
he is best suited to initiate the design of any special equipment.
The equipment and chemicals teams should assemble all required equip-
ment and chemicals specified by the team coordinator (Tables 1 and 2).
18
-------
Special attention should be paid to coordinating efforts between the bottle
wash and equipment checkout people and the chemicals and equipment assembly
people. Any additional equipment requirements (spare parts, special seals,
etc.) should be communicated to the equipment assembly or chemicals person-
nel and added to the master lists.
Is possible, a separate room should be set aside for the storage of
equipment and chemicals necessary for a sampling trip. This minimizes inter-
ference with other laboratory activities, and after the sampling trip an
inventory can be quickly completed.
With the equipment stored in one room, equipment checkout consists
chiefly of equipment set-up and appropriate testing to ensure the unit is
functioning properly. Following checkout, it can be packed for shipping.
Packaging of sample containers and solutions is a very important part
of equipment assembly. Care must be taken that every sample specified in
the test matrix has the appropriately labeled storage containers and
reserves. The module tote boxes can hold ten 500 ml (1 pint) Nalgene bot-
tles. These tote boxes are labeled on the outside with their test run num-
ber (normally one series to a tote box). These tote boxes are then inserted
into the shipping boxes which are foam padded and lockable. Each shipping
box is labeled with its contents. In the field the shipping boxes act as
cabinets to prevent contamination and promote organization of samples.
After the equipment has been assembled, checked out, and packed, one
last meeting should be held to verify test procedures and assignments.
This last meeting verifies that all team members understand the goals of
the test program and can perform the sampling test.
19
-------
CHAPTER III
FLUE GAS SAMPLING
Determining trace element levels in flue gas (FG) source streams is
similar to standard particulate sampling techniques, but requires several
special considerations. The problems encountered and the techniques used
for obtaining representative samples of source streams are common to both
particulate sampling and trace material sampling. The differences and spe-
cial considerations peculiar to trace element source sampling are concerned
with contamination of the sample, sample alteration, equipment selection,
and the properties of trace materials in sampling systems. The following
paragraphs discuss the peculiarities and problems associated with trace
element source sampling and present recommended procedures.
3.1 SPECIAL CONSIDERATIONS
3.1.1 Contamination and Alteration of Sample by Sampling Train
When sampling FG streams to determine particulate loading, care must
be taken that the samples do not become contaminated or altered. The same
problems exist when sampling for trace elements, but, in addition, close
attention must be given to material compatibilities with the sampling sys-
tem and the sample itself. An assessment of the compatibility of a trace
element sample begins at the sample probe and extends to the laboratory
environment where the sample is analyzed.
The sample is first exposed to the sampling nozzle and probe. The
conventional materials which have been used to construct these components
are stainless steel, glass, quartz, and Teflon or Teflon-coated steel.
Since the particulates are analyzed for Ni and Cr, stainless steel presents
a contamination problem for Cr, Ni, and other elements contained in stain-
less steel. Glass and quartz sampling trains have notorious breakage prob-
lems. Teflon or Teflon-coated components are excellent but have a temperature
limit of 230°C (450°F).
Although the standard sampling nozzles and probes are adequate for
many applications, a more universal approach using a removable probe liner
is recommended. The use of a liner prevents sample contamination and
20
-------
facilitates probe cleaning. The recommended liner material is a high tem-
perature, thermally stable polyimide, Kapton*, which is manufactured by
DuPont. The material is thermally stable in air to 450-500°C (842-932°F),
and has demonstrated stability up to 400°C (752°F) in combustion gas
streams. At present, there is no known organic solvent for the film;
strong alkali, however, will dissolve Kapton. It is infusible as well as
flame resistant. The results of a spark source mass spectroscopic analysis
of the film material indicate that Kapton does not represent a significant
source of contamination for trace element sampling.
As with the nozzle and probe, the cyclone and filter of the sampling
train present possible sample contamination or alteration problems. In the
case of cyclones constructed of stainless steel, the potential for sample
contamination by Ni, Cr, and other elements may be higher than in the case
of the probe due to the increased possibility of surface abrasion. Sample
contamination by filters is more subtle. Because of the high temperature
requirement, filters made of glass fiber or quartz materials are typically
chosen, However, these materials have relatively large concentrations of
several elements which are of interest in trace material sampling. A few
commercially available, high purity quartz filter materials have been
fabricated especially for trace element collection, and several other
materials appear to be suitable for trace element sampling. Further dis-
cussion of these filters and new alternative materials are presented in
Section 3.2.2.
The sampling train impingers and connecting hardware are another
source of sample contamination. In these cooler train components, areas
exist where condensed source materials or the impinger scrubbing solutions
can corrode metal parts leading to sample contamination. Loss of trace
materials due to adsorption and chemical reaction after condensation cannot
be totally eliminated by using glass impinger systems. Even the reagent
grade chemicals used in the impingers to scrub the volatile trace materials
can have blank values higher than the actual sample.
*Registered trademark.
21
-------
3.1.2 Multiphase Sampling Requirements
The requirement of collecting representative trace material samples
from multiphase source streams expands the scope of performance of a given
collection system. The sampling system must be able to collect materials
in both gaseous and condensed phases. This can be readily accomplished using
an EPA Method 5 type source sampling train. With this type of train, those
elements which exist as particulate matter in the stack are collected mainly
in the cyclone or on a filter, while those vaporous or submicron size ele-
ments, which behave as gases, are collected in the oxidative impinger sys-
tem. The distribution of elements within these components of the train has
been found to be different for various elements, and appears to depend on
the source from which trace materials have been sampled. The presence of
emission control equipment, such as electrostatic precipitators, sulfur
dioxide scrubbers, etc., significantly affects both the concentrations and
the physical and chemical forms of the trace materials found in source out-
let streams. Knowledge of the distribution of trace elements in the sam-
pling train can greatly simplify sample collection and analysis.
3.1.3 High Volume Sampling Requirement
Regardless of whether the trace elements are collected in the particu-
late collection section or in the vapor scrubbing impingers of a sampling
train, enough sample must be collected to ensure accurate and precise anal-
ysis. The exact amount of collected material required for elemental anal-
ysis depends on the particular analytical techniques used to determine that
element. Procedures recommended in this manual generally require 1-100 ug
of an element for analysis. Based on a need for determining source
3 *•> ?
stream concentrations of 60 yg/m (2.6 x 10~ /grains ft ) and a desire to
keep the sampling time to two hours, the sample collection rate must be
3 3
approximately 0.014 m /min (0.5 ft /min) to collect 100 yg of the element
of interest. This is roughly the maximum sampling rate attainable with
most commercially available Method 5 sampling units. Only one instrument
manufacturer, the Aerotherm Company,* currently has available a Method 5
type source sampler capable of operating at sampling rates up to
*The Aerotherm Company, a Division of the Acurex Corp., Mountain View,
California.
22
-------
3 3
0.16 m /min (5 ft /min). By using the Aerotherm high volume sampler, ade-
quate or excess amounts of sample can readily be collected. When analysis
procedures are developed that require less than 1-100 yg of element and
that are convenient and economical to use, other Method 5 type samplers
with lower sample acquisition rates can be employed for trace element
sampling.
3.1.4 Aerotherm High Volume Stack Sampler
A modified Aerotherm high volume stack sampler (HVSS) is the recom-
mended Method 5 type sampler for trace element source testing. This HVSS
unit operates at nearly ten times the sampling rate of other Method 5 sam-
pling systems while maintaining the capability of collecting particulate
samples for Method 5 particulate emission testing.
The Aerotherm sampling unit is shown schematically in Figure 3. The
principal components are: the control unit, probe, heated particulate col-
lection section, impinger system, vacuum pump, and umbilical line. A com-
plete and detailed description of each of these components is not presented
here, but can be obtained from the manufacturer. The following sections
discuss these components in sufficient detail to permit an understanding
of the modifications to the unit for trace element source sampling.
Table 3 summarizes the principal advantages and disadvantages of the
Aerotherm HVSS unit.
3.1.4.1 Control Unit —
The control unit contains all the instruments required for measuring
stack velocity, sampling flow rate, cumulative flow, and temperatures at
various points in the sampling system. All of the controls for the sam-
pling system are located in the control unit with the exception of the
valves for controlling sample flow rate. These values are mounted on the
vacuum pump, which is positioned adjacent to the control unit when the
sampling system is in operation. Thus, all controls and measurement dis-
plays are centered about the control unit and permit operation by one
individual.
23
-------
TABLE 3. PRINCIPAL FUNCTIONAL ADVANTAGES AND DISADVANTAGES OF AEROTHERM HVSS UNIT FOR
TRACE ELEMENT SAMPLING AND GENERAL SOURCE PARTICULATE TESTING
ro
Trace Element Sampling Advantages
Sample collection rate
Impinger assembly
Adaptability
Operating temperature
General Source Testing Advantages
Source access requirement
Electrical design
Mechanical strength
Flow measurement capability
Pump characteristics
Trace Element Sampling Disadvantages
Particulate collection system
Up to 0.17 m3/min (6 SCFM)
Rugged Lexan construction, easy to clean with 0-ring seals.
Separable from oven and capable of operating at sampling rates
of 0.23 m3/min (8 SCFM).
Original construction design permits use of probe liners and
nonstandard filter materials.
Up to 260°C (500°F) for probe and oven and even higher with
some modifications.
Rotatable probe capable of sampling horizontal and vertical
source streams.
Circuit breakers used instead of fuses and separate power wiring
for heaters and pump to assure required power. Digital display
of temperature.
Rugged and well-packaged for shipment.
Two magnehelic gauges for accurate readout over pi tot tube range
of 0-10 cm (0-4 in.) of water.
Oil-free vane type pump modified for low leakage.
Cyclone filter and connecting hardware possible source of Ni,
Cr, and other stainless steel elements. Teflon-coated filter
housing possible source of particulate sample alteration. Fil-
ter design distributes particulate nonuniformly across filter.
(continued)
-------
TABLE 3. (CONTINUED)
ro
en
Trace Element Sampling Disadvantages
Gas cooling coils
Impinger assembly
General Source Testing Disadvantages
Unit physical size
Total power requirement
Traversing hardware
Probe
Component connectors (gas flow)
Cyclone and filter hardware
Impinger assembly
Difficult and time-consuming to clean between sampling runs.
Shows potential for corrosion and impinger sample contamination.
Stainless steel connecting hardware subject to corrosion and not
compatible with strong oxidizing scrubbing solutions unless
coated inside and out with Teflon.
Larger and heavier than low volume Method 5 samplers.
Greater than 1.65 kilowatt (15 ampere, 110 V) which is a common
electrical power outlet rating.
Extremely heavy and somewhat difficult to assemble. Not adapt-
able enough to meet all source sampling conditions.
Heavier than necessary and requires 5 cm (2 in.) minimum port
access hole; "no breakdown" 3 m (10 ft) probe for easy shipping
and general sampling needs.
For the most part they are not quick disconnect type and are dif-
ficult to work with while hot.
Present Teflon gaskets and coated hardware not compatible with
high temperature (>260°C or 500°F) environments.
Not equipped with check valve to prevent backflush in case of
pump or electrical failure.
-------
°VEN
CYCLONE
FILTER
STACK TEMPERATURE T.C.
PROBE TEMPERATURE T.C.
IMPINGER
PITOT 4P
MAGNEHELIC
FINE ADJUSTMENT 1CE BATH
GAS METER BY PASS VALVE
TX' , I COARSE
' ADJUSTMENT
VALVE
VACUUM
LINE
VACUUM
GAGE
AIR TIGHT
VACUUM
PUMP
DRY TEST METER
ORIFICE A?
MAGNEHELIC GAGE
Figure 3. Aerotherm high volume stack sampler schematic.
(Cooling coils not shown.)
3.1.4.2 Probe and Nozzle --
The probe and nozzle components of the Aerotherm unit are similar in
design to those used in other Method 5 samplers. The stainless steel probe
is wrapped and heated with a fiberglass insulated strip heater. A liner
made of Kapton is placed in the probe and a specially fabricated bushing
is inserted at the tip to direct the incoming flue gas past the edge of the
film (see Figure 4 and Section 3.2.3). Swagelok tube fittings are used for
nozzle and cyclone-filter connections. The heated sampling tube, S-type
pitot, and thermocouples are sheathed in a stainless steel tube for mechan-
ical support and protection from the stack environment.
26
-------
UNION
FITTING
KAPTON LINING
STAINLESS STEEL PROBE
STAINLESS STEEL BUSHING
Figure 4. Schematic diagram of Kapton liner inside probe tube.
3.1.4.3 Particulate Collection Section —
This section consists of a cyclone, filter, and oven. The cyclone and
filter are constructed of stainless steel. Standard stainless steel
Swage!ok fittings are used to connect this hardware with the probe and
impinger assembly. The operation of this section is such that the filter
may be used alone for particulate collection, since the cyclone unit can
be removed and the probe connected directly to the filter housing. The
interior of the filter housing is Teflon coated and the sealing gasket for
the cyclone is also made of Teflon. The oven is capable of heating the
cyclone and filter to 260°C (500°F). The use of Teflon presents a problem
in handling gas streams at temperatures higher than 230 C (471 F).
3.1.4.4 Impinger System --
The impinger train for the Aerotherm sampler is unique. It consists
of gas cooling coils, a pre-cooler trap, and a set of four impingers. Gas
cooling coils are not needed in standard EPA Method 5 trains, but are
required for the HVSS due to its higher flow rate. The impinger assembly
for the HVSS is not constructed of glass, and is larger than the regular
Greenberg-Smith impinger (250 ml charge versus 100 ml charge) used in
Method 5 trains. The four impinger bottles are constructed of durable
Lexan (polycarbonate) plastic. The interconnections between the impingers
27
-------
are made of stainless steel pipe and Swagelok fittings. This connecting
hardware must be Teflon coated to prevent corrosion by the oxidative
impinger solutions and subsequent contamination of the sample. The
impingers are sealed with large diameter Lexan caps using 0-rings, and thus
the interior of the impinger is completely accessible for easy rinsing and
cleaning, which is important for trace element sampling. Unlike ground
joint systems for glass impingers, 0-ring seals do not require a layer of
grease (a source of trace element contamination) for sealing, and are not
subject to seizing. They are consequently easier to assemble and
disassemble.
Preceding the impingers are three meters (10 ft) of stainless steel
cooling coils and a small Lexan collector for water condensate (pre-cooler
trap). The pre-cooler trap, like the regular impingers, uses an 0-ring
seal. The cooling coils, pre-cooler trap, and set of four impingers can
be chilled with ice water or a mixture of dry ice and water.
3.1.4.5 Vacuum Pump —
A vane-type vacuum pump is used. This pump has a 3/4 hp motor, a flow
3 3
rate of 10 ft /min (0.28 m /min) at 0-inch Hg, and weighs approximately
27 kg (60 Ib) including all fittings. Some important features of this
pump include:
t Smooth, pulse-free flow
• High vacuum capacity
• Self-lubricating carbon vanes
• Special shaft seal
• Coarse and Fine flow control valves located on pump
• Carrying handle
• Aluminum filter and muffler jars
• Vacuum gauge to indicate filter condition
• Quick disconnect fittings.
3.1.4.6 Umbilical Line --
The control unit and pump are connected to the remainder of the
system through a pneumatic hose and an umbilical line consisting of a num-
ber of electrical lines, thermocouples, and pi tot tube connections. The
28
-------
umbilical provides all essential connections between the sampling train
and the control unit. Its significant features are:
t Quick disconnect fittings
• Smooth, kink-free, waterproof sheath for enclosing all lines
t Mating ends of connectors configured so that umbilical line
length can be increased.
3-1.5 HVSS Components —Selection and Design Recommendations
3.1.5.1 Probe Material Selection —
Standard HVSS probe designs consist of a heated tube equipped with an
outer protective sheath and a sampling nozzle. Construction materials are
typically carbon steel, aluminum, stainless steel (several types), glass,
quartz, and Teflon-coated materials (Section 3.1.1). The relative merits
of these materials as probe materials are tabulated in Table 4. As indi-
cated in the table, aluminum probes have the lowest operating temperature
limit while quartz has the highest limit. Above approximately 250°C
(482°F), aluminum and many of its alloys rapidly lose their physical
strength, but their light weight make them attractive for particulate sam-
pling in stacks having large diameters. Both glass and quartz exhibit
excellent chemical inertness except for possible reactivity with fluoride
as hydrogen fluoride. Glass or quartz probes would not be sources of
sample contamination in trace element sampling. For sampling source
streams that require probes longer than 5-6 ft (1.5-2 m), glass and quartz
become impractical because of breakage problems and use of a mechanical
support is recommended. The combined probe and support (usually made of
stainless steel) produce a heavy and cumbersome probe.
Carbon steel, while having a high temperature limit, is readily cor-
roded by many stack environments and is a source of Mn and other low per-
centage steel elements. It has not been widely used for stack sampling.
Teflon-coated probes are excellent from a contamination standpoint for
trace element sampling, but they do not quite have an adequate temperature
limit for many sampling situations. Stainless steel is perhaps the most
common material used for probe construction. It has a high operating tem-
perature limit and is usually readily available. However, particulate
29
-------
TABLE 4. PROBE MATERIALS AND SELECTION CRITERIA
Material
Aluminum
Carbon steel
Glass (Borosilicate)
Quartz
Kapton liner
Teflon coating
Stainless steel (316)
Titanium
Operating
Temperature,
OC (op)
250 (482)
>900 (>1652)
820 (1508)
>1500 (:>2732)
450-500 (842-932)
<260 (<500)
900 (>1652)
>1000 (>1832)
Cleaning
Difficult
Difficult
Difficult
Difficult
Very easy
Easy
Difficult
Unknown
Handling
Lightweight
Heavy
Fragile
Fragile
Somewhat
involved
Satisfactory
Heavy
Lightweight
Contamination/
Sample Alteration
Depends on alloy
Yes, possible Mn and other
elements
No contamination, possible
loss of F
No contamination, possible
loss of F
None
None
Cr, Ni, other stainless
elements
Depends on alloy
co
o
-------
material collected using stainless steel probes has shown Ni, Cr, and
other element contamination. Probes constructed of stainless steel are
not lightweight, especially if stainless steel sheath tubes are used as
protective outer covers. Particulate material from some sampling environ-
ments strongly adheres to the surface of stainless steel, thus presenting
a cleaning problem.
Titanium and its alloys are finding use as probe construction materials.
Their high temperature stability and light weight make them attractive sam-
pling probe materials. As with any probe, sample contamination, alteration
and recovery are still unresolved problems. Sample recovery is especially
important since, under certain sampling conditions, the amount of material
deposited on the probe walls can approach 50 percent of the total particu-
late material collected. This material must be cleaned from the probe and
added to the sample collected on the filter for source particulate determi-
nation according to Method 5 procedures. The same procedure also applies
to trace element source sampling to obtain accurate trace element emission
results. However, probe cleaning is difficult and time-consuming under
field test conditions. Rinsing and brush cleaning procedures are also
inadequate for strongly adherent particulates, resulting in biased particu-
late loading results. For trace element sampling, the problems are com-
pounded since the probe cleaning procedure exposes the collected sample and
washings to field contamination. It is possible at some industrial plants
to collect more sample material in a few minutes of "cleaning" the equip-
ment outside a stack than during an hour of source sampling. Leaded vehicle
exhaust, windblown dust, and scrubber and cooling tower mists, for example,
can contaminate carefully collected source samples and void trace element
test results.
For these reasons, a completely different approach to sample recovery
from the probe is recommended for trace element particulate sampling. This
approach utilizes an inert removable liner inside the normal sampling probe.
The recommended liner material is Kapton film (see Section 3.1.1), which is
an inert, clean, high temperature (400°C) polymeric film.
Figure 4 shows a schematic diagram of the liner location inside the
probe tube. Except for the short sections of the stainless steel nozzle
31
-------
and bushing insert, the stack gas sample is in contact with the Kapton
liner until the sample enters the particulate collection system and the
remainder of the sampling train. The liner can easily be removed after
sampling arid immediately placed in a sample storage bottle. A new liner is then
inserted for the next sampling task. The use of a liner greatly simplifies
recovery of probe material for analysis and reduces cross-contamination
between tests. It is also possible, though it has not been demonstrated,
that the liner approach may lead to more accurate and reproducible Method 5
particulate measurements.
3.1.5.2 Filter Selection --
Materials and filter types that have been typically used for general
particulate sampling have contained high and variable amounts of several
trace elements. Extraction procedures developed by many laboratories for
removing these contaminants have been only partially successful. Mate-
rials suitable for source particulate and trace element sampling recently
have been developed. The following paragraph describes filter materials
that have been successfully and unsuccessfully used for trace element
sampling.
Table 5 lists some of the filter materials that have been used for
trace element sampling and some key properties. Of the materials listed,
only two filter types have been field tested with the Aerotherm HVSS unit:
Kreha carbon fiber and Gelman Spectro Grade Type A filters. The trace ele-
ment content of both materials has been assessed and found acceptable.
Each package of filters purchased from Gelman is supplied with a trace con-
taminant assay for 25 common elements. Kreha carbon fiber material is not
analyzed by the manufacturer, but a spark source mass spectroscopic (SSMS)
analysis of samples of Kreha filter material has shown it to have a low
trace element content, except for fluoride. Trace contaminant analyses of
many of the other filter materials are presented in References 2, 3 and 4.
The relatively low operating temperature (<100°C or 212°F) for Gelman
Spectro Grade glass fiber filters is required to protect the surface treat-
ment used on the filter to prevent the oxidation of sulfur dioxide to par-
ticulate sulfate. The filter can be used at temperatures several hundred
32
-------
TABLE 5. FILTER MATERIAL FOR TRACE ELEMENT SAMPLING
Filter Material
Carbon fiber
(Kreha Corp.)
Cellulose paper
(Whatman 41)
Graphite
(Poco)
Glass fiber,
spectroquality
(Gelman Instr. Co.)
Microquartz fiber
(Under development)
(Arthur D. Little Co.)
Teflon membrane
Tissuquartz
(Pallflex Co.)
Sintered Silver
Puri ty
Good except
for F
Poor
Excellent
Excellent
Excellent
Excellent
Excellent
Excellent
Temperature
Limit,
QC (OF)
>500 (>932)
<200 (<392)
>1000 (>1832)
>100 (>212)
800 (1472)
<250 (<482)
800 (1472)
<900 (<1652)
Efficiency
Low
Medium
High
High
High
High
High
Undetermined
Utility
AP
Low
Medium
High
Medi urn
Medium
Extremely
High
Medi urn
Medium
Strength
Good
Good
Good
Good
Good
Good
Poor
Good but
corrodes
readily
co
co
-------
degrees higher if the protective surface coating is not required. The
Kreha material is stable to about 500°C (932°F) in air and other oxidizing
environments. The filtering efficiencies of the two materials are also
quite different. The Gelman product when tested with dioctyl phthalate
spheres gave a minimum of 99.5 percent retention for 0.3u particles. Data
is not available for the Kreha material, but single pieces of the material
used in the Aerotherm high volume sampler have been shown to pass large
amounts of particulate matter. Of these two materials, Gelman Spectro
Grade is recommended for trace element sampling.
3.1.5.3 Oxidative Impinger Solutions --
Oxidative impinger solutions are required for trace element sampling.
Several elements (As, Hg, Se, Sb, F", Cl"), are emitted in a vaporous state or
as very fine particulate and are not collectible on filters. Of these, Hg
is considered to exist in elemental form in many source emission streams,
while arsenic, selenium, and antimony are thought to exist as vaporous
oxides. Fluoride and chloride exist principally as hydrogen halide gases.
If cooled to near ambient temperature and passed through chilled water im-
pingers, trace quantities of As, Sb, F", and Cl" can be nearly quantitatively
collected. Selenium in source streams when cooled and contacted with aer-
ated water is apparently oxidized and retained in solution. Of these vapor-
ous elements, only mercury and cadmium are not trapped in a simple aqueous
impinger. To sample for these elements an oxidative agent is needed. Any
element or compound which passes through a filter because of its small size,
can be collected in simple aqueous impingers if adequate contacting with the
liquid is permitted.
In sampling for Hg in source streams, numerous oxidative scrubbers have
been used. The most common have been acid permanganate, iodine monochloride,
and hydrogen peroxide. Except for peroxide, these oxidative scrubbing solu-
tions contribute a residue which interferes with the analysis of these solu-
tions for other elements. In addition, the scrubbing efficiency of these
solutions, except for permanganate;5^ has not generally been quantitative
under source sampling conditions. This has been shown to be particularly
true in collecting vaporous Hg using a high volume sampler where gas con-
tacting times are very short. For these sampling conditions, oxidative
34
-------
scrubbers are needed with higher specific rates or a better method of con-
tacting the gas. The latter approach is impractical since the means used
to achieve high contact times greatly increases the gas flow resistance
through impingers. The alternative approach is to improve the reaction rate
of the oxidative scrubber systems through catalysis or selection of new
more potent systems.
An impinger system that is an effective scrubber for Hg in source streams
using high volume sampling rates is aqueous ammonium persulfate catalyzed
with silver ion. This system uses low concentrations of reagents so that
contamination from trace materials in the reagents does not interfere with
the analysis of trace elements collected during source sampling. This
solution also quantitatively collects other vapor-phase trace elements.
Catalyzed ammonium persulfate can theoretically oxidize chloride to chlor-
ate or perchlorate based on reaction thermodynamics and thus can be used to
sample for this element as well, if oxidation kinetics are satisfied.
All oxidative impinger solutions that are used for trace element
source sampling must be protected from stack gas reducing agents which are
capable of lowering impinger oxidizer concentrations to a point where the
impingers are not effective scrubbers for trace elements. The major stack
gas component which must be trapped before it reduces impinger oxident con-
centration is sulfur dioxide. A convenient and efficient means of accom-
plishing this is to precede the trace element scrubbing impinger(s) with a
simple hydrogen peroxide impinger; the HpOp concentration is determined by
the concentration of S02 in the stack effluent and the volume of stack gas
sample to be drawn through the impinger. The recommended impinger arrange-
ment for sampling trace elements using the Aerotherm high volume sampler is
shown in Table 6. The reagents used to prepare these solutions are readily
available from a number of sources in high purity grades compatible for
trace element sampling.
At present, acceptable alternatives to the catalyzed ammonium persul-
fate scrubbing impinger system cannot be recommended for use with the Aero-
therm high volume sampler. For lower flow rate stack samplers, there are
two acceptable oxidative systems. They are acidic permanganate( ' and
iodine monochloride^49^. If Hg is the only vaporous trace element to be
35
-------
TABLE 6. IMPINGERS FOR TRACE ELEMENT SAMPLING
Impinger
Condensate collector
No. 1
No. 2
No. 3
No. 4
Reagent
Empty
3M H?0?
0.2M (NH4)2S208 + 0.02M AgN03
0.2M (NH4)2S2Og + 0.02M AgN03
Drier ite or silica gel
Quantity
—
250 cc
250 cc
250 cc
225 g
examined,the dry gold amalgam tecnnique developed by TraDet, Inc. can be
used. These alternative oxidative scrubbers and amalgamation techniques
require impingers to reduce S02 content before the Hg collection impingers
or devices. Besides hydrogen peroxide and ammonium hydroxide, S02 scrubbers
include water and saturated alkali carbonate solutions. The latter alter-
native, while more effective in removing S02 from the sampled gas stream,
adds a considerable amount of solids to the impinger solutions which can
lead to difficulties in analysis for trace elements collected in this
impinger.
3.2 SPECIAL FIELD GUIDELINES FOR TRACE ELEMENT SOURCE SAMPLING
For the most part, the procedures and methods used in source sampling
for trace elements are similar to those used in sampling for particulate
matter. These are well described and documented' ' ' and are not
discussed here. This section presents those modifications to procedures
for particulate sampling and the special considerations that apply to trace
element source sampling.
3.2.1 Work Area and Contamination Considerations
Avoidance of contamination is of prime importance in trace element
source sampling. Contamination considerations extend from the equipment
preparative and packing stages to the analysis of the collected samples.
Bottles in which samples are to be stored and shipped and reagents that
are to be used must be free of trace element contamination. This requires
special cleaning of bottleware and the selection of pure reagents and clean
36
-------
materials. Bottleware used for packaging reagents for field use and stor-
age of samples should be dedicated for this purpose and scrupulously cleaned
before being put in storage for subsequent use. The bottle material should
be high density and/or crosslinked polyethylene or polypropylene. If Teflon
bottleware is available, it may also be used. Bottleware should be cleaned
according to the procedure outlined in Chapter VII, and stored in clean con-
tainers in a non-laboratory area. Bottleware made from low density and/or
linear polyethylene or polypropylene should not be used since these mate-
rials are made from recycled plasticware of varying composition and trace
contaminate levels. Laboratories, by their nature, tend to be reservoirs
of trace contaminants. Protective containers such as wood or laminated
boxes are important since they can prevent permeation of vaporous Hg from
such sources as broken thermometers from reaching plastic bottleware and
contaminating the collected samples.
The purity of the reagents should be the highest available. J.T. Baker*
Ultrex grade reagents are recommended for the preparation of the nitric acid,
water, and ammonium hydroxide impinger and wash solutions. Ammonium per-
sulfate and hydrogen peroxide are only available as analytical reagents, but
the trace element concentrations for these particular reagents has been
found to be acceptable.
The prevention of contamination should also be a factor in the selec-
tion of a base camp site for the field test. The base camp should be pro-
tected from wind and be away from vehicular traffic, process water sources,
cooling towers, and overhead conveyor systems. Even when these precautions
are followed, all solution transfers, impinger rinsings, and probe liner
manipulations should be performed in the mobile laboratory van or trailer.
3.2.2 Filter and Impinger Solution Preparation
Filters for trace element sampling and source particulate determina-
tions are prepared and handled in a similar manner. However, filters for
trace element sampling require extra care to prevent contamination, since the
filters will be chemically analyzed. Filters for trace element sampling should
be stored in clean petri dishes until ready for use during the field test,
*J.T. Baker Chemical Company, Phillipsburg, N.J.
37
-------
and extra filters should be taken to the field to be used later as blanks
in the analysis scheme.
The filters are conditioned for 3 hours at 287°C (550°F) and desic-
cated for 12 hours prior to weighing. For the best results the filters in
covered petri dishes should be stored in a desiccator prior to and after
use in the field. On return to the laboratory the filters are desiccated
for another 12 hours and then weighed.
Except for the ammonium persulfate impinger solution, all other
impinger reagents and solutions can be prepared in advance of the field
test and prepackaged in labeled bottles identified for a particular test
run. At the time of the test, these bottles are laid out in an orderly
fashion and the contents are then transferred to the appropriate impinger.
The empty bottle is saved and refilled with the sampling solution after
completion of the run. The ammonium persulfate impinger solution should
be prepared immediately prior to a sampling run. The stability of the dis-
solved reagent is such that it loses one half of its oxidizing strength in
8 to 10 hours after preparation. In lieu of preparing this solution, the
reagent can be weighed and prepackaged in dry labeled bottles for shipment
to the test site. The solution can then be prepared on site by adding
silver nitrate solution to the bottle and transferring the solution to the
appropriate impinger. This approach of using prepackaged solutions and
reagents saves a significant amount of time in the field.
3.2.3 Probe Liner Preparation, Installation, and Removal
At present, Kapton probe liners are made from a 0.002-inch thick film.
A strip of material 7.5 cm (3 in.) wide is cut from a roll of the film.
The length of the strip is the length of the probe that is to be lined.
The 7.5 cm (3 in.) width corresponds to approximately two circumferences
for the 1.3 cm (1/2 in.) diameter sampling probe for the Aerotherm unit.
If larger or smaller diameter probe tubing is used, the width of the liner
should be adjusted to give twice the probe tube circumference. A clean,
unoiled knife such as an X-acto type can be used to cut the liner strip.
Disposable, unpowdered polyethylene gloves should be worn during this cut-
ting stage to reduce contamination from handling, and a clean work area
should be used. After cutting, the film strip is dried overnight in a
38
-------
110 C (230 F) oven. After drying, the film strip is weighed and the weight
recorded. Insertion of the liner into the probe (shown in step-by-step
photographs on the following three pages) is begun by wrapping the film
around a 7 mm mandrel (usually a piece of stainless steel rod the length of
the probe) and placing the rolled film strip and mandrel into a 10 mm
(3/8 in.) OD Teflon or stainless steel tube (referred to as a straw tube).
The ends of this tube are then capped and loaded into a larger plastic con-
tainer tube for storage. In the field the straw tube containing the rolled
film strip is inserted into the sampling probe, and the film is held while
the straw tube is withdrawn, thus accomplishing the insertion of the film
liner in the sampling probe. The insertion of the film liner in the field
requires two people and less than five minutes. With the liner in the
probe, the bushing insert and sampling nozzle are installed on the probe.
The bushing insert conducts stack gases and particulate matter past the
end of the film, thus preventing material from getting between the film
and probe tube wall (see Figure 4).
Upon completion of a sampling task, the probe is disconnected from the
particulate collection system and the nozzle is also removed. With the end
capped, the probe is taken to a clean, wind-protected area and the liner is
removed with a pair of forceps and stored in a clean, labeled polyethylene
bottle or bag. A new liner can be installed and the next sampling effort
started. If a probe liner is not used, the probe is cleaned by rinsing
with high purity water and using a nylon or Teflon bristle brush if
required. The particulate material and washings are put into a bottle.
3.2.4 Handling and Storage of Impinger Samples
When a sampling task is completed, the impingers are disconnected
from the pump and particulate collection system and brought to the mobile
laboratory for emptying, rinsing, and refilling. The contents of each
impinger, including the pre-cooler trap and the silica or drierite filled
impinger, are returned to individually labeled solution bottles. Each of
the liquid impingers is thoroughly rinsed with three 50 cc portions of high
purity water and these rinsings combined with each catch. The gas cooling
39
-------
1. While wearing clean
gloves, cut a 5 cm (2 in.)
strip of Kapton adhesive
tape and wrap once around
one end of the mandrel
(7 mm SS rod). Place end
of Kapton liner half way
down width of tape.
2. Partially unroll
Kapton liner, holding the
lengthwise edge against
the mandrel. Have ready
a cone made of a 30 cm x
6 cm (12 in. x 2 in.)
piece of Kapton film
rolled and taped such that
the diameter at one end is
8-9 mm and at the other
end is 12-13 mm.
3. Wrap the remaining
tape, with the Kapton
liner inside, around the
mandrel. As the Kapton
is wrapped, slide the
cone, larger end first,
over the liner and
mandrel. The cone will
help curl the liner and
hold it in place.
-------
4. Roll the Kapton strip
around the mandrel, slid-
ing the cone over the
wrapped liner as you go.
When the cone clears the
end of the mandrel,
insert the tip into the
straw tube (10 mm 0.0
tubing) so that the liner
does not uncurl again.
5. Continue to simultane-
ously unroll the Kapton
strip, slide the cone
over the curling Kapton,
and insert the mandrel
into the straw tube.
The glove on the hand
holding the cone may be
removed at this time for
easier handling of the
cone.
6. Continue step 5 until
the entire rolled liner
and mandrel are in the
straw tube. This proce-
dure works best if per-
formed smoothly and
quickly. Cap both ends
of the straw tube with
Swagelok fittings.
-------
7. To transfer the liner
to the probe (in the
field), uncap the straw
tube ends and join the
tube to the probe with a
Swagelok union. Alter-
natively, the liner can
be loaded directly into
the probe in the
laboratory.
8. With the tube and
probe connected, push the
mandrel into the probe
with another piece of
stainless steel rod the
same length as the man-
drel or slightly longer.
When the mandrel is com-
pletely inside the probe,
disconnect the two.
9. Push the mandrel out
until it extends 7-10 cm
(3-4 in.) beyond the
probe end. With a clean,
degreased X-acto knife,
cut off the end of the
liner attached to the
tape. Pull out the man-
drel, insert the bushing,
push the bushing and
liner back into the
probe and attach the
nozzle.
42
-------
coil is rinsed with two 100 cc portions of high purity water, and these
rinsings combined with the condensate collector catch. High purity concen-
trated nitric acid is added to each bottle to reduce the pH to between 1
and 2. The addition of the nitric acid prevents the formation of precipi-
tates and reduces the absorption of sample trace elements on container bot-
tle surfaces. The bottles are sealed and returned to the shipping
containers.
43
-------
CHAPTER IV
LIQUID AND SLURRY SAMPLING
Wastewater from plants consists of both contaminated and relatively
clean effluent streams. In general, the contaminated wastewaters are those
taken from processes, while clean wastewaters are those used for indirect
heat exchange and general washing. The major sources of contaminated
wastewaters are:
• Brines from electrolysis and crystallization
• Filter cake washings (mining operations)
0 Waste acid and alkaline streams (wet scrubber equipment)
• Washing streams containing substantial amounts of suspended
particulate matter (coal gasifiers).
In general, these waters are characterized by suspended solids ranging
from tenths to tens of percent of the total weight.
Clean wastewaters are primarily composed of stream condensate and
cooling water. Normally, these are released into the environment with little
or no treatment. Due to process leaks, makeup water, or boiler blowdown,
these streams can become polluted and would be sources for trace metal
sampling.
The composition of liquid streams that might be sampled for their trace
metal content would fit in several broad categories:
• Water
t Water-solids (slurry)
• Water-organic liquids-solids (slurry)
• Organic liquids-solids (slurry)
• Organic liquids.
The amount of nonmiscible organic liquids in most outlet streams from
most plants will be low (<10 percent of total volume). For trace metal
sampling, major emphasis must be placed on streams in the water and water-
solids categories, along with the special category of water-organic liquid-
solids (slurry). Since the water and water-organic liquid categories fit
into general separative schemes at a lower level, they are not addressed
directly.
44
-------
4.1 CONSIDERATIONS FOR LIQUID AND SLURRY SAMPLING
The factors which must be considered in accurately sampling a fluid
stream for trace materials include:
t Stream homogeneity
• Stream flow rate and variations
• Prevention of sample loss
• Sources of contamination
• Sample size.
Of these, stream homogeneity is perhaps the most important factor.
Unlike stack effluent streams which are mixed fairly evenly due to higher
thermal agitation and lower fluid viscosities, liquid streams tend to be
more stratified and require more careful sampling. A flow-proportional,
composite sampling technique is required for sampling liquid and slurry
streams for trace materials. By comparison, obtaining a composite repre-
sentative gaseous sample from a stack, a traverse of the pipe or duct is
made. However, this is usually impractical in sampling liquid streams.
In liquid streams, a composite sample can be taken by using several
differently positioned probes, a single multiported probe, or a combination
of these. While either of these approaches is suitable, the single
multiported sampler is usually more convenient.
In the case of slurry sampling, it is also important to avoid segre-
gation of liquid and solid phases. This is similar to the requirement for
isokinetic sampling in particulate-laden gas streams. A recent EPA internal
study has evaluated 60 commercially available models of automatic sewer flow
samples ^9'. In this study, sampling velocity was determined to be the most
critical factor in sampling sewage slurries. Two units were found to perform
acceptably: Quality Control Equipment Company Model CVE and Testing Machines,
Inc. Fluid Stream Sampler. Both of these units are portable and completely
automated. The QCEC unit also has a built-in ice chamber for automatic
refrigeration of temperature-sensitive samples. Another common practice for
the preservation of liquid samples is to freeze them until they are ready for
analysis. However, this practice is now being questioned for trace material
analysis because of the tendency for metal ions to precipitate upon freezing.
The technique should only, therefore, be used for preserving samples for
bacteriological and dissolved gas analyses.
45
-------
Several studies have shown that trace materials in liquid phases may be
lost from a sample through adsorption on sampling line or reservoir surfaces
(10, 11, 12)^ Borosiiicate glass (Pyrex) surfaces appear to be particularly
effective in removing trace heavy metals, especially under alkaline conditions.
However, plastics such as polyethylene, polypropylene, and Teflon show little
or no tendency to adsorb inorganic materials. It is essential, therefore,
that the sampling lines and collection reservoirs used for sampling liquid
streams be made of plastic, preferably Teflon, because of its superior
chemical inertness toward strong acids, alkalies and other chemical reagents.
In addition to sample loss due to surface adsorption, a sample may also
be contaminated by elements from those surfaces. Surface wall material can
be deposited in a sample either by a chemical extraction of the wall materials
by reagents in the sample or by physical abrasion or erosion of the wall by
a sample. The latter case could be a significant problem for slurry systems
because the abrasive nature of the sample could expose unpassivated layers
of the wall to chemical interaction with the sample.
Another important factor which must be considered in sampling liquid
streams for trace material constituents is the sample size requirement.
Two principal requirements govern sample size. The first requirement is
that the amount of sample collected must be sufficient for the testing and
analysis procedures to furnish accurate and precise results. The second
requirement is based on the statistical sampling error that can be tolerated.
The minimum sample required for analysis varies between 1 and 1000 ug for
the trace materials of interest using the proposed analysis procedures.
For the lower ppm concentration levels of interest, this translates into
minimum sample volumes ranging between one ml and one liter. This range of
sample volumes is easily within the operating limits of presently available
liquid sampling equipment and presents no special difficulties.
Determining the minimum size liquid sample that must be collected to
reduce statistical sampling error to acceptable limits is considerably more
difficult. To meet a goal of a combined relative sampling and analysis
error of ±25 percent, the allowable error must be subdivided between sampling
and analysis errors. A relative analysis error of 5 to 10 percent is common
at the low ppm concentration levels. Allowing a maximum 15 percent error for
analysis, the sampling error can be as high as 20 percent and meet the 25
percent overall error (25 = /152 + 202 ).
46
-------
4.2 EQUIPMENT SURVEY AND RECOMMENDATIONS
The sampling of liquids and slurries in the categories defined above
requires equipment suitable for point sampling, as well as for sampling
from ponds, reservoirs, open tanks, open channel flows, and pipes which
have built-in sampling ports.
For pond and tank sampling, and in preliminary or point sampling,
point and depth-integrated samplers are commonly used. The Sirco Unicsoop,
which is made of No. 316 stainless steel, is recommended for point sampling.
The Uniscoop has a handle for collecting below-surface samples and is
convenient to use. A depth-integrated sampler consists of a weighted bottle
and is easily fabricated.
The recommended automatic samplers are Model CVE (Quality Control
Equipment Corporation) and Model 1940 (Instruments Specialty Corporation).
The schematic for the Model CVE is shown in Figure 5. All components in
both units which come in contact with the sample are composed of polypro-
pylene, polyethylene, or Tygon, and the sample never passes through any
valves or pumps. Both units can perform short-term or long-term sampling
at certain time intervals proportional to time or flow rate. The units also
have built-in ice cabinets to preserve the samples at lower temperatures.
The units offer a long-term stability without mechanical or electronic
malfunctions.
While the Model CVE sampler provides composite samples directly in the
field, and the Model 1940 takes sequential samples that are stored in separate
bottles, both models can perform time or flow proportional sampling depending
on the availability of a flow measuring device. The Model CVE was rated the
best unit in a study conducted by the EPA Regional Office at Kansas City,
Missouri ^. The Model 1940 was also highly recommended in this study and
is currently in use in the Los Angeles County District (250 units) and
Ontario Ministry of Environments in Canada.
For pipes having built-in sampling ports, the Model L-F (Quality Control
Equipment Corporation) is recommended. This sampler can be used for both
liquid and slurry samplings.
The equipment discussed in this chapter is capable of handling a wide
variety of process streams found in most industrial applications. For
47
-------
VACUUM SYSTEM
BLOW-DOWN
SOLENOID
VALVE
3=
115V INPUT
LIQUID SYSTEM
SYSTEM OPERATION
THE PATENTED VACUUM SYSTEM LIFTS LIQUID THROUGH A SUCTION LINE
INTO THE SAMPLING CHAMBER. WHEN FILLED, THE CHAMBER IS AUTO-
MATICALLY CLOSED TO THE VACUUM. THE PUMP THEN SHUTS OFF AND
THE SAMPLE IS FORCIBLY DRAWN INTO THE SAMPLE CONTAINER. A
SECONDARY FLOAT CHECK PREVENTS ANY LIQUID FROM REACHING THE
PUMP SHOULD THE PRIMARY SHUT-OFF PASS ANY MATERIAL. THE SUCTION
LINE DRAINS BY GRAVITY BACK TO THE SOURCE. NO POCKETS OF FLUID
REMAIN TO CONTAMINATE SUBSEQUENT SAMPLES. AUTOMATIC PRES-
SURIZED BLOW-DOWN OF SUCTION LINES AND THE ENTIRE LIQUID SYS-
TEM ASSURES THAT NO OLD MATERIAL REMAINS TO CONTAMINATE THE
CURRENT SAMPLE. THIS CLEARS THE LINE AND PROVIDES A FRESH AIR
PURGE OF THE PUMP AND THE ENTIRE SYSTEM.
Figure 5. Model CVE sampler schematic.
sampling streams having highly corrosive materials, the Teflon coating of all
metal parts should be considered in order to prevent contamination of the
samples and corrosion of the sampling equipment. However, for most appli-
cations, inherent durability and flexibility of the off-the-shelf samplers
are adequate.
4.3 SAMPLE COLLECTION SEPARATION AND STORAGE
Prior to sample collection, the liquid sampler must be cleaned in the
field to prevent any contamination. Flushing out the sampler with a liter
48
-------
(quart) of dilute nitric acid (0.1 N) followed by a liter of high purity water
eliminates any particles introduced during shipping and field storage. The
sampler is then placed near the process stream to be sampled. This site
should be free of windblown contamination. The sample probe or hose is
introduced into the stream and the timer set for the proper sampling period.
Following sample collection, the various phases present in a liquid
or slurry must be separated to prevent gross disruption of the trace metal
content of each phase. Allowing the phases to be in contact with each other
leads to a redistribution of the trace metal composition among the phases.
For these reasons a preliminary field phase separation procedure is required.
The equipment necessary for separating the phases of liquid and slurry
samples in the field consists of:
t Filters
• Nalgene Buchner funnel and filter flask
• Nalgene separatory funnel
t Small vacuum pump
• Acids, bases, methanol, and high purity distilled water.
The recommended filter is 0.5y Mi Hi pore Fluoropore (Teflon). Fluoropore
filters are both chemically clean and inert to most organic and corrosive
solvents. Furthermore, Teflon does not have a tendency to absorb metals on
its surface. If a slurry sample is found to contain a large amount of solids,
pre-filtration is necessary using another Teflon filter, Mi Hi pore Mitex.
This filter is available in a lOy-pore size and is designed to act both as a
membrane and depth-type filter. The larger pore size of the Mitex filter
allows for a higher solid content without clogging. Pre-filtration with
Mitex followed by filtration through 0.5y Fluoropore should produce a
solids-free solution. For purposes of this manual, any particle which passes
through a 0.5y membrane filter is considered in solution.
All the Nalgene equipment used in separation of the sample phases must
be pre-washed to prevent contamination. A solution of 0.1 N HN03 (high purity)
in a squeeze bottle must be used as a rinse between samples to prevent cross-
contamination. Enough clean replacements must be available, should the
liquids leave a film on the plasticware.
49
-------
The last item required is a small vacuum pump for the Buchner funnel.
Clean Tygon tubing and a spare clean filter flask can serve as a water trap
to protect the pump. All this equipment must be set up in a clean area in
the van or, if possible, in the plant's quality control lab. Contamination
from external sources should be prevented at all times.
Figure 6 summarizes the separation and stabilization scheme for several
categories of liquid and slurry samples. The following sections briefly
discuss the procedures associated with each stream category shown.
4.3.1 Clear Water
After sample collection, a clear stream is divided approximately in
half and placed in chemical bottles. Enough concentrated HN03 is added to
one bottle to reduce the pH to 1. The other bottle is treated with NaOH
to attain a pH of approximately 10. The addition of HN03 stabilizes the
trace metals in solution and prevents adsorption on the container walls.
- -3
The analysis of Cl , NO., , or P04 is not possible if any of the respective
acids are added. Sample streams containing relatively large percentages of
CN~ should not be acidified or volatilization of toxic HCN will occur.
WATER WATER-SOLID
FILTER HjOWASH
HNO,
NaOH
FILTRATE
HNO,
No OH
SOLIDS - STORE
IN CLEAN BOTTLE
WITH FILTER
WATER-ORGANIC LIQUID
SEPARATORY
FUNNEL
WATER
/ \
HNO,
NaOH
ORGANICS -
STORE IN GLASS
BOTTLES WITH
TEFLON LINERS
WATER-ORGANIC LIQUID-SOLIDS
FILTER-WASH WITH FILTRATE
LIQUIDS
SOLIDS - STORE
IN CLEAN BOTTLE
WITH FILTER
SEPARATORY FUNNEL
WATER " ORGANICS-
k / \ STORE IN GLASS
HN03 NaOH BOTTLES WITH
TEFLON LINERS
Mgure 6. Typical separation schemes for process liquids.
50
-------
Cyanide is a chelation agent which will keep most of the trace metals of
interest in solution, thus acidification is unnecessary.
The addition of NaOH stabilizes anions such as CN~ and NOl; however,
it is also highly likely that precipitation or the formation of a gelatine
mass of the hydroxides of metals like Fe+3, Al+3, and Ba+2 will occur.
These hydroxides can be redissolved back at the lab.
4.3.2 Water-Solid Streams
For water-solid streams, the slurry is first filtered using the equip-
ment and filters described above. Because the Teflon filters are hydrophobic,
they must be moistened with a small amount of methanol just prior to filtra-
tion to avoid any pressure buildup. The filter and solids are washed with
a small amount of H,,0 and sucked damp dry; both are placed in a clean storage
bottle. The solids are then dried (at 110°C) and weighed back at the lab.
The filtrate is treated in the same manner as the clear water stream described
above, except the amount of wash water added must be recorded. All volume
changes must be recorded so weight/volume percentages can be corrected to
the original sample.
4.3.3 Water-Organic Streams
Clear water-organic liquid is separated by placing the collected sample
in a clean Nalgene separatory funnel. The various liquid phases are allowed
to separate and are then drained into separate clear bottles. The water
phase is treated as described above, while the organic phase is stored in
glass bottles with Teflon seals. Note: If both inorganic and organic
analysis is going to be performed, then all glass separation apparatus should
be used. The aqueous and solid portions, however, are still stored in
polypropylene bottles.
4.3.4 Water-Organic Liquid-Solid Streams
The most complicated system is the water-organic liquid-solids stream.
In this system, the solids are first filtered and washed by cycling some of
the filtrate through the filter cake. The reason for this is to avoid dis-
rupting the three-phase (n-i-s) equilibrium by the addition of water. After
the filtrate is collected, it is placed in a separatory funnel, separated,
and stabilized as above.
51
-------
4.3.5 Summary
In general, there are three important requirements for proper collec-
tion and storage of liquid and slurry samples:
t Volume changes must be recorded because they affect the slurry
composition data.
• Cross-contamination must be avoided through cleanliness of
operations.
• All transfers should be made as quantitatively as possible.
By following the above procedures, liquid and slurry samples can be
properly stabilized for shipment to the analytical lab.
52
-------
CHAPTER V
SOLID SAMPLING
Solid sampling covers a broad spectrum of material sizes ranging from
large lumps to powders and dusts. There is an equally diverse assortment
of potential sample sites including railroad cars, large heaps, plant hoppers,
conveyor belts, and process stream pipes. Obviously no one method or piece
of equipment is suitable for all situations, but the advantages and disad-
vantages of each must be weighed in the light of individual field test
conditions. The following discussion provides an overview of common solid
sampling situations and summarizes the sampling approaches and alternatives
available to a field test team.
5.1 SAMPLING METHODOLOGY AND EQUIPMENT SURVEY
The sampling of solids for trace elements may include the use of three
manual grab sampling techniques: shovel sampling; boring techniques, which
include pipe or thief sampling; and auger sampling. Mechanical samplers,
both moving and stationary, can also be used to obtain solid samples.
Detailed treatments of each of these sampling techniques can be found in
several technical handbooks ' ' an '. The chief consideration of
solid sampling is the acquiring of representative samples.
Shovel sampling procedures include grab sampling, coning and quartering,
and fractional shoveling. Grab sampling consists of taking small, equal
portions at random or regular intervals, typically from railroad cars, large
heaps, or hoppers. The method is quick and inexpensive. However, grab
sampling makes no allowance for segregation of the sample by particle size
and also tends to give consistently high or consistently low results depending
on the person sampling. As such, grab sampling should be used for survey
sampling.
Coning and quartering consists of carefully piling the material into a
conical heap, with subsequent flattening of the cone into a circular cake.
The cake is then marked into quadrants; two opposite quadrants are taken as
the sample and the other two quadrants are discarded. The entire process
is repeated until the desired sample size is obtained. In general, this
53
-------
method is time-consuming and the symmetry of the intended vertical size
segregation is difficult to achieve in practice.
Fractional shoveling is applicable to materials being loaded, unloaded,
or moved from one place to another by shoveling. In fractional shoveling,
every third, fourth, fifth, or tenth shovelful is taken as the sample. This
method is inexpensive and relatively fast. If performed conscientiously,
fractional shoveling can be more reliable than coning and quartering. However,
its applicability is limited and errors are easily introduced by carelessness.
Pipe boring techniques represent another class of solid sampling method-
ology applicable to material stored in piles, silos or bins. The usual
method of pipe boring is to insert the pipe into the material to be sampled
at regular intervals. The method is fairly reliable provided that the pipe
is long enough to reach the bottom of the material. However, it is only
applicable to fine or powdered dry materials, because lumps or any stickiness
will jam or plug the pipe. Small pipe borers can be used to sample sacks
or cans of material. There are primarily two designs of pipe borers that
give best results. One is a simple pipe that is tapered so the end first
inserted is smaller in diameter than the handle end. A more sophisticated
design, known as a thief, makes the sample more representative vertically.
It consists of two close-fitting concentric pipes sealed at the base in a
conical point. Longitudinal slots are cut along the side of each pipe.
The thief is inserted with the slots turned away from each other and then,
when the sampler is in position, the outer pipe is rotated, lining up the
slots and allowing the inner pipe to fill the sample. For proper results
with any design of pipe borer, the opening through which the sample material
passes (slots or circular pipe ends) must be large relative to the maximum
particle size.
Auger samplers, a form of drill, pack the sample in the helical groove
of the auger and can be enclosed in a casing if the nature of the sample is
such that it will spill when the auger is removed from the hole. Like the
pipe borers, they are simple to use and have the further advantage of being
applicable to a greater variety of materials. For example, augers work well
for materials that are packed too hard for a pipe sampler to be forced in.
For very packed materials, machine-driven augers are available. However, a
54
-------
thief sampler would be the better choice if sample spillage is a possibil-
ity. Also, both pipe samplers and augers yield poor results if the material
being sampled is poorly mixed.
Mechanical samplers require that the sample material be in motion to
present it to the cutters as a thin ribbon or stream. Design considerations
for feeding these samplers and catching the sample and rejected material
generally necessitate the permanent installation of the sampler into the
flowing sample stream. Numerous mechanical samplers have been designed;
the most popular designs have been variously modified to satisfy specific
applications. However, all mechanical samplers fall into two general types:
those that take part of the stream all of the time (stationary samplers),
and those that take all of the stream part of the time (moving samplers).
In stationary mechanical samplers, the entire sample stream is fed
continuously through the device and stationary cutting edges divide out
and remove specific fractions. The two best-known designs of this type are
rifflers and whistle-pipes.
Rifflers take several slices of the stream by means of parallel chutes
alternately placed at 90 angles to each other, thereby cutting the stream
in half. Successive rifflers can be arranged in banks to cut the stream into
any desired fraction. The smaller the chute width, the greater the number
of increments in the sample. Therefore, the accuracy of riffler sampling
increases as the ratio of chute width to particle size decreases, to the
limiting condition where the chutes tend to clog. In general, chutes should
be at least three times the diameter of the largest size particle to avoid
clogging. Care must be taken to feed the riffler with a well-mixed, uniform
sheet of material since any compositional variations due to cross-sectional
segregation are multiplied by a bank of rifflers.
A whistle-pipe sampler consists of a vertical pipe with notched
openings cut halfway through the pipe, each spaced 90 horizontally from
the one above. Rectangular steel plates are placed in the notches at a
45° angle to the vertical so that the top edges coincide with a diameter of
the pipe. Thus each notch halves the sample and, with a series of five
openings, the sample obtained is 1/32 of the original volume. The same
fraction with improved accuracy can be obtained by using a cutter arrange-
ment that quarters the stream, rejecting opposite quarters, and spaces
55
-------
each cutter at 45° horizontally from the one above. In either design, a
hopper-shaped liner is placed above each notch to re-center the stream
before it reaches the next cutter.
Both rifflers and whistle pipes have irresolvable design problems
that reduce their reliability. Worn or bent cutting edges distort both the
volume and the particle size distribution of the sample. The housing nec-
essary for these samplers prevents examining them for clogged openings while
in operation. Material streams whose composition varies along the trans-
verse section are even further segregated by either of these samplers.
Moving samplers consist of cutters that move through the free-falling
sample stream taking all the stream for the duration of time they are moving
through it. There are two ways of effecting this. One is with rotating or
oscillating samplers whose cutters are set on the radii of an arc, and the
other is with straight-line samplers whose cutting edges are set parallel
to each other and perpendicular to the line of their path.
Among the well-known designs of rotating arc-path samplers are Vezins,
Synders, and Chas. Synders. They all consist generally of scoops with
vertical sides, set on an axis parallel to the stream flow. The best
oscillating samplers are known as Bruntons. The scoop travels back and
forth across the stream in a pendulum-type motion. The travel path must
be sufficiently long to minimize the bias created by taking more sample from
the sides of the stream than from the middle. All the arc-path samplers
have the advantage over stationary samplers in that they take an accurate
cut, are simply constructed, and are accessible for observation while in
operation. However, damp sample material may tend to clog the scoops and
care must be taken to maintain the cutting edges in good condition and to
keep them completely radial.
The straight-line samplers are generally considered to be the most
reliable and accurate of all available types of samplers. The design of
their cutters is such that the sampling scoop spends an equal amount of
time in every portion of the stream. Generally the travel is at right
angles to the stream. Though they provide increased reliability, these
samplers require more maintenance and attention because of their increased
mechanical complexity.
56
-------
5.2 STATISTICAL DETERMINATION OF SAMPLE SIZE
Whatever the sampling method, the amount of sample that should be taken
is a function of the desired accuracy, the material homogeneity, and the
reliability of the sampling method. A statistical means of determining the
sample size needed to yield results having a prescribed level of precision
based on the above factors has been theoretically derived by Welcher^16^
1963.
The general form of this equation is:
"•(I2)
2
where:
n = number of units to be taken for sample
a = advance estimate of the standard deviation
E = maximum allowable difference between the result to be obtained
from the sample and the result of testing the entire bulk of
material
t = a factor corresponding to the acceptable risk of exceeding E
The terms "E" and "t" are relatively easy to assign as they are the
parameters of the desired precision. The t is a statistical factor express-
ing the probability that, by chance, E will be exceeded. The following
tabulation lists several approximate probability values and the corre-
sponding values of t:
t. Probability that E will be exceeded
3 3 in 1000
2.58 1 in 100
2 45 in 1000,
1.96 1 in 20
1.64 1 in 10
Generally, a factor of 3 is used to minimize the possibility of the
sampling error exceeding E. Any degree of precision can be chosen for E,
although the required sample size increases as the square of the entire
57
-------
precision term. E can be expressed as a percentage or in units of measure-
ment; however.a must be expressed in the same way so that the resultant n
is unitless.
The result is a value of n units of sample. These units (e.g.,
shovelfuls, pounds, etc.) must be the same as those used to determine the
standard deviation, o, in order to relate the two sides of the equation.
The value of a can be determined in one of two ways. Either a preliminary
test must be run on at least 10 units of sample to calculate the standard
deviation between units; or a can be estimated, with the result that a
somewhat larger number of units is taken for the sample than the estimated
number necessary, and the sample size is readjusted after the actual g
has been determined.
The following example illustrates the use of this statistical equation.
Using some type of pipe sampler, a flowing stream of pulverized coal is
being sampled for percent ash. A preliminary test showed the average
deviation between samples taken by the pipe sampler to be 25 percent, and
a maximum sampling error of 10 percent is required. Then:
n 3 I v*/ v"/l = (7.5)2 = 56^25 s 56
and 57 samples must be taken to determine percent ash in the coal with a
maximum error of 10 percent.
5.3 SAMPLE COLLECTION AND STORAGE
It is always preferable to sample a moving stream either in pipes or
from conveyor belts, particularly if there is a large particle size range
in the material. Stored containers or heaped beds of material tend to
settle, creating segregation of particles according to size and density,
and it is difficult to compensate for this bias in the sampling. Further-
more, large masses of stored material are extremely difficult to handle.
The interior portions are relatively inaccessible and the amount of time
and space needed to move the material enough to take a representative sample
can quickly become prohibitive. However, such situations can generally be
avoided by a good sampling test plan.
58
-------
Typically, in a process test for trace elements, the solid materials
of interest are the feed materials and the residues from participate
scrubbers such as baghouses, high energy Venturis, and electrostatic pre-
cipitators. Raw feed stock as it passes through the process stream may pick
up other materials as contaminants and, therefore, differ greatly in compo-
sition from what is finally being fed to the process. Consequently, samples
should be taken at the last possible site before the stream is fed into the
process. This means that sampling will generally be conducted from a feed
hopper, if accessible, or from whatever pipes or conveyors feed the material
to the process. Similarly, scrubber residues can be sampled from whatever
collection hopper the device has or from pipes going to the hopper. Extra
handling steps only increase the chances of the sample becoming contaminated.
5.4 OTHER CONSIDERATIONS
As part of their own process control, many plants may have some type of
mechanical sampler already installed into their process stream. Whenever
possible, these devices should be used for taking samples. They are reliable,
take representative samples, and are fast and easy to use. Before being
used, however, the samplers' operation and cutting edges should be checked
to ensure accuracy. If reliable automatic samplers are an integrated part
of the plant and are available, no sampling equipment will be needed by the
field test personnel.
In cases where it is decided to take samples from moving conveyor belts,
the standard procedure is to stop the conveyor at regular intervals (e.g.,
every 10 to 15 minutes) and shovel off a section of the material. This is
continued until the desired sample size is obtained. Flat-nosed shovels
with straight perpendicular sides are best for these sampling purposes.
Another alternative is to sample process streams as they move through
pipes if there are appropriate ports. A variety of pipe samplers are
commercially available. The type most suitable for trace element sampling
is the pneumatic sampler, which eliminates the screw type or scraping action
of other types of samplers which grind the sample and abrade the sampler,
thereby introducing considerable contamination.
The best pneumatic sampler currently available is the Model RTA of
Quality Control Equipment Corporation. All parts in contact with the
59
-------
sample can be Teflon or nylon lined, which is a major advantage. It can be
used to sample solids with a particle size less than 0.64 cm (1/4 in. in
diameter), as well as slurries and liquids from pipes fitted with at least
2-inch diameter sampling ports. Figure 7 is a schematic for the Model
RTA sampler.
Samples collected by any of the techniques described in this chapter
must be stored in pre-washed and dried plastic bottles or plastic lined drums.
If plant personnel are to take the sample, they should be provided with
the appropriate bottles or containers.
SAMPLE INLET
MOUNTING FLANGE TO SAMPLE PORT
SAMPLING TUBE
25 1/2 IN.
SAMPLE OUTLET
AIR FITTINGS
"P" SAMPLING TUBE EXTENSION
4 IN. TO 6 IN. AS REQUIRED
SYSTEM OPERATION
UPON COMMAND FROM THE VARIABLE INTERVAL CONTROL, THE SAMPLING
TUBE IS EXTENDED INTO THE CONVEYING LINE, WHERE IT DWELLS FOR A
SHORT ADJUSTABLE PERIOD. THE SAMPLE IS TRAPPED IN A SUITABLY SIZED
CAVITY IN THE SAMPLING TUBE, WHICH IS THEN AUTOMATICALLY RE-
TRACTED AND THE SAMPLE EJECTED BY A BLAST OF AIR. THE PRESSURE AND
DURATION OF THE AIR BLAST ARE FULLY ADJUSTABLE TO MEET VARYING
CONDITIONS. SAMPLE IS COMPLETELY DISCHARGED. NO CARRY-OVER.
THE CONTROL PANEL IS A SEPARATE UNIT.
Figure 7. Pneumatic line sampler schematic.
60
-------
CHAPTER VI
ANALYTICAL TEST PLANNING
The success or failure of a field sampling test depends not only on
the correct field application of sampling and sample preservation techniques
but also on accurate and cost-effective sample preparation and analytical
procedures in the laboratory. For the analytical data to be relevant, sev-
eral factors must be considered. The laboratory work areas must be care-
fully cleaned and prepared to prevent contamination. The analytical work
must be planned and scheduled to fit the objectives of the field test.
Data review points at which to evaluate the fit of analytical procedures to
the test objectives must be predetermined. The procedures themselves must
be checked for precision and accuracy and calibration curves prepared.
The following paragraphs contain appropriate procedures and checklists
that can be followed in assembling a viable analytical test plan. A typi-
cal test data sheet (Table 7) and planning logic flowchart (Figure 8) are
supplied. Prior to beginning any work on the samples, the analyst must be
aware of the following:
• List of samples taken and any special notations from field
test crew
t Test objectives
• Elements to be analyzed, expected concentration ranges if
known, and degree of accuracy and precision needed
• Brief description of process sampled so that an estimate of
expected species can be made.
With the above information the analyst can begin to design an analytical
test plan.
6.1 LABORATORY PREPARATION
Samples received for trace element analysis must be stored and handled
in a clean work area. The laboratory area should be cleaned prior to open-
ing sample boxes by 1) removing any extraneous samples or equipment; 2)
washing the bench-tops with soap (Alconox) and water; and 3) and covering
them with clean paper mats. The exterior surface of each sample container
should be wiped with a clean towel moistened with deionized water. After
61
-------
TABLE 7. SAMPLE OF ANALYTICAL TEST CHECKLIST
Sample
Identification
(Cross ref-
erence to
sample
number)
Date
Rec ' vd
Prelim.
Wt./Vol.
Physical
Examina-
tion
©
Sample Work-Up
Drying
Grinding
Blending
Oxygen
Plasma
Ashing
•
Dissolu-
tion
©
Pre-
Concen-
tration
©
;
ro
NOTES: 1. Insert dates at completion of each increment and the analyst's initials.
2. Describe any anomalous behavior, i.e., formation of precipitates, discoloration, f
effervescence, etc.
3. (¥) denotes data and procedure review points.
-------
LIQUID AND/OR SOLID SAMPLE
VOLUME
COLOR
SOLIDS
ORGANIC FILM
t>H
SOLVENT EXTRACT
FOR ORGANICS
FILTRATION
FILTERED SOLIDS/
SOLIDS
ORGANIC
PHASE
AQUEOUS
PHASE
FILTRATE
DRY, WEIGH,
GRIND, AND
BLEND
DRY
ANALYSIS
DISSOLVE
XRD-j
XRF-j
SEM-I
EMISSION SPECTROSCOPY
ASH
~ AA SPECTROSCOPY
COLORIMETRY
- POLAROGRAPHY
A DATA REVIEW AND DECISION MILESTONES
Figure 8. Planning logic flow chart
samples have been removed from the boxes, wiped clean and stored in series,
they should be logged in on a master list (see Table 7). By comparing this
list with a list supplied by the field sampling team leader, it becomes
apparent whether any samples have been lost or misplaced.
Additional procedures for preparing and cleaning apparatus, instru-
ments, and reagents for the analytical work are provided in Chapter VII.
6.2 DATA REVIEW POINTS
A preliminary examination of the samples is the first step. All particulate
samples are weighed, all liquid volumes measured, and the appearance of any
precipitates, organic films or scums, or solution disco!orations are indicated.
The first data review point is at the conclusion of this task. At this
point the analyst must decide on the following: (1) whether an aliquot of
the solid samples can be taken or whether the entire sample must be used,
63
-------
and (2) for liquid samples, whether they need to be filtered and the resi-
due analyzed and,if so, by what means. A decision also must be made, based
on volume collected and information desired, whether to combine solutions
and/or perform ore-concentration procedures. All decisions must be made
with the overall test objectives in mind. The drying, grinding, blending,
and ashing steps discussed in Chapter VII can then be implemented. After
the preliminary sample preparation has been completed, decisions about the
most appropriate dissolution procedure to be followed and whether or not
further concentration steps are warranted must be made. When all the sam-
ples are prepared and in solution, they are ready for analysis.
At the conclusion of the chemical analysis (described in Chapter VIII),
the procedures for the analytical data must be reviewed. This data review
should include a comparison to standards, examination of standard addition
curves, and comparison of duplicates (if run). If discrepancies exist,
every effort must be made to identify the source of the problem. Appro-
priate changes in the procedures for sample handling, preparation, dissolu-
tion, and/or analysis must be incorporated into the analytical test plan
prior to rerunning the analytical tests.
6.3 PRECISION AND ACCURACY
Due to the diversity of matrices and the wide ranges of trace element
concentration expected from field samples, it is imperative that the analyst
use all means available to determine the accuracy and precision of the ana-
lytical technique employed for each element determined. The analytical
procedures proposed in this manual were selected to give at least ±15 per-
cent accuracy level at a concentration of 1 ppm. However, the accuracy
will improve significantly for any of the analyses with an increase in
concentration.
A clear distinction should be made between the terms "precision" and
"accuracy" when applied to methods of analysis. Precision refers to the
reproducibility of a method when repeated on a homogeneous sample under con-
trolled conditions, regardless of whether or not the observed values are
widely displaced from the true value as a result of systematic or constant
errors present throughout the measurements. Precision can be expressed by
the standard deviation. Accuracy refers to the agreement between the amount
64
-------
of a component measured by the test method and the amount actually present.
Relative error expresses the difference between the measured and the actual
amounts, as a percentage of the actual amount. A method may have very high
precision but recover only a part of the element being determined; or an
analysis, although precise, may be in error because of poorly standardized
solutions, inaccurate dilution techniques, inaccurate balance weights, or
improperly calibrated equipment.
There are two general techniques normally used for evaluating the
accuracy of analytical procedures. For coal-fired power plants, certified
standards which closely match sample matrix, such as NBS 1632 and 1633
trace elements in coal and fly ash, respectively, should be used whenever
available. When NBS standards are not available, the recommended procedure
is to use the standard addition technique. This entails the addition to
the sample aliquots of known concentrations of the element under analysis.
These values are then plotted and the calibration curve extrapolated through
the abscissa to the negative ordinate. The value of the negative ordinate
is now an accurate estimation of the elemental concentration. When this
value is compared with concentrations generated using pure elemental stan-
dards, the degree of chemical interference can be ascertained. The doped
samples and standards should be run in parallel in order to compare the two
for accuracy.
6.4 CALIBRATION
The selection of a calibration procedure depends on the degree of
accuracy required, which in turn depends on the degree and types of solu-
tion matrix interferences present. The following sections present three
techniques which are used primarily for AAS analysis. However, when
unknown solutions are encountered requiring ancillary techniques, the
method of standard additions should be used to ascertain the magnitude of
any interferences.
6.4.1 Factor Method
This method is the most rapid of the three but is inaccurate for cer-
tain elements due to interelement interference. The method involves the
analysis of a standard along with the unknown sample. The factor obtained
by dividing the standard concentration by its absorbance when multiplied by
65
-------
the absorbance of the unknown will give its concentration. The standard
must have approximately the same concentration as the unknown. This method
exhibits its greatest advantage where a large number of samples containing
similar concentrations are to be determined and where extreme accuracy is
not necessary.
6.4.2 Short Curve Method
The Short Curve Method is more accurate but requires slightly more
time than the Factor Method. This method involves the running of two stan-
dards, one slightly higher and one slightly lower than the unknown. The
absorbances of the standards are plotted against their concentrations, and
the concentrations of the unknowns are read from the curve. This method,
like the Factor Method, is ideally suited for running a large number of
routine samples containing similar concentrations where the ultimate in
accuracy is not required.
6.4.3 Additions Method (Recommended)
This is the most accurate and precise of the three methods. It incor-
porates the advantage of the short-curve method and also eliminates inter-
element interference. The method involves placing three identical aliquots
of the sample in volumetric flasks. None of the standard solution is placed
in the first flask. A quantity of standard equal to the approximate level
expected is placed in the second flask. A quantity of standard approxi-
mately equal to twice that amount is placed in the third flask. All three
flasks are diluted to the mark with water and their absorbances determined.
The absorbances are then plotted against concentration. This method is
applicable to special analyses where maximum accuracy is necessary and ana-
lytical time is not important.
66
-------
CHAPTER VII
LABORATORY AND SAMPLE PREPARATION
In determining the trace elements in control process streams, several
different types of samples must be collected, handled and manipulated.
These sample types are: particulate matter on filters, particulate matter
on liners, impinger solutions, bulk solids, liquids, and slurries. This
chapter presents the general procedures for preparing each of these samples
for trace element analysis, including the preparation of labware.
7.1 LABWARE PREPARATION
All labware which will come in contact with samples for trace element
analysis must be specially cleaned to prevent contamination and avoid mate-
rial losses. The procedures to be used are as follows:
• Remove all old labels and container/flask markings using
acetone or dry abrasive cleansers.
• Perform a preliminary but thorough wash of all labware.
A 2:1 mixture of Alconox or abrasive cleanser is satis-
factory. All detergent should be rinsed off with tap
water. Pipets and volumetric flasks should be rinsed
thoroughly.
• After a thorough scrubbing and rinsing, thoroughly rinse
each piece of labware in the specified acid wash (see
below). (CAUTION: All acid rinsing should be performed
in a hooded sink while wearing protective eyeware.)
- Aqua Regia (3:1 HCl-HNOj acid cleaning solution
is used on all labware dsed in Hg analysis.
H2S04-HN03 in 1:1 ratio is used on all remaining
glassware, except volumetric flasks.
- HN03 (20%) solution in high purity water is used
Tor all plastic labware.
' H2S04 Cone, warm (60°C or 140°F) is used for rinsing
glass volumetric flasks.
- Chromic Acid (100 g K2Cr207 per 3.5£ cone. H2S04)
is used for cleaning pipets. A 24-hour soak time
is required.
Note: All acid cleaning solutions except aqua
regia are reusable.
67
-------
• After acid rinsing, the labware is rinsed once with tap
water and immediately flushed three times with high
purity water.
• Following the final high purity water rinse, the labware
is air dried and stored ready for use. The following
storing procedures are recommended:
Pipets should be allowed to drain and dry in a vertical
position. Store in a drawer.
Open-mouth glassware should be turned upside down on
rack and allowed to dry in a quiet area free from
contaminating drafts from windows and hoods.
One to 5 ml high purity water should be poured into
storage bottles and volumetric flasks, the tops
replaced, and then the container inverted several
times to see that the water flows smoothly without
beading. Allow the water to remain during storing.
If beading occurs the container is not clean. Repeat
the above procedures.
Store all labware in clean, closed cabinets or drawers.
Contamination and loss of samples for trace element analysis are greatly
reduced if the above procedures are carefully followed.
7.2 PARTICULATE AND IMPINGER SOLUTION SAMPLE PREPARATION
The sampling and analysis system described here is designed to give
an accurate estimate of the trace element composition of a sampled gas
stream. To achieve this goal, the grain loading dry volume of the samples
must be determined. To calculate the grain loading and dry volume, the
total weight of particulate matter collected and the volume of moisture
condensed from the gas stream are needed.
The following sections present details of procedures for the prepara-
tion of particulates and impinger solution samples for analysis. A general
sample handling flow sheet which provides an overview of the analyses to be
performed was presented in Figure 8 (Chapter VI).
Various impurities present in water and reagents used in trace analysis
become serious sources of contamination, because reagents are used in rela-
tively large quantities compared with the sample itself. Commercially
available high purity reagents, such as J. T. Baker Ultrex Brand, should be
used whenever available. If high purity reagents are not available, the
68
-------
purification methods listed in Morrison(17) should be used to purify (high
purity) needed reagents. In all cases distilled deionized water should be
used to make up the solutions.
7.2.1 Probe Liner Sample
Remove each Kapton liner from the probe using plastic tweezers to
first pull it out far enough to get a firm grip by hand. Using plastic
polyethylene disposable gloves, roll the liner gently as it is being
removed. Place the liner in a pre-tared bottle for weighing back at the
lab. To remove particulate matter, first cut the liner into smaller, more
easily handled sections and rinse each one with Freon PCA (DuPont de
Nemours Co., Freon Products Division, Palo Alto, Ca.) into a pre-cleaned,
pre-tared 250 ml beaker. If a high portion of organic material is present,
then reagent grade acetone may be used to remove the particulate from the
liner. Determine the particulate weight after the Freon or acetone are
evaporated on a steam bath and the samples have been desiccated for 12 hours.
Repeat the same procedure for a Kapton liner returned unused from the
field. The unused liner will be a blank in the analysis scheme.
7.2.2 Cyclone Sample
There may be two cyclone samples, a dry particulate sample and a
Freon PCA rinse sample. Desiccate the powder for 12 hours prior to weighing.
Evaporate the Freon or acetone rinse samples to dryness on a steam bath,
then cool them in a desiccator (12 hours), and weigh. Retain all samples
in their respective beakers.
7.2.3 Filter Sample
Particulate matter collected on filters is desiccated for 12 hours and
then weighed. The particulate matter from the liners, cyclone, cyclone
wash, and particulates on the filter are collected together on the filter
pad after they have been individually weighed. The total collected sample
weight, if the field sampling parameters have been adjusted correctly,
should be between 0.1 and 2 g. Due to the filter housing design of the
HVSS, the collected particulate matter is not evenly distributed across the
surface of the filter. Thus, the filter cannot be divided and must be
taken as a whole for analysis.
69
-------
The composite participate samples on the filter are placed in a large
petri dish and oxygen plasma ashed for 4 hours. The samples are then
removed, transferred to pre-cleaned, covered 250 ml beakers, and digested
for 2 hours in 40 ml constant boiling aqua regia solution (8 ml 19% HCL +
32 ml 20% HN03). The solutions are filtered through No. 41 Whatman filter
paper into 100 ml Nalgene volumetrics. If appreciable residue remains, it
can be recombined with the original filtrates by ashing at 550 C (1022 F)
and fusing with a small amount of Na2C03 HO parts Na2C03 to 1 part resi-
due), followed by redissolution with 1:1 HC1. These solutions can now be
analyzed.
7.2.4 Impinger Solution
Using a Nalgene graduated cylinder, measure the volumes of each of the
impinger solutions and then return them to the original containers. It is
best to start with the last impinger solution in the train and work forward
to the pre-cooler trap where the elemental concentration is the highest.
The graduated cylinder used to measure the volume of the solutions should
be rinsed between each solution measurement and cleaned thoroughly with
high purity HNO_ solution between one series of solutions.
7.3 PREPARATION OF SOLID SAMPLES
Generalized procedures to be used for the preparation and dissolution
of solid samples include grinding, sieving, drying, and dissolution. The
dissolution procedures also include an oxygen plasma pretreatment to decom-
pose and remove organic material without the loss of volatile trace elements
whenever necessary. The final step is sample dissolution by addition of
the appropriate acids and fluxes to solubilize the sample for subsequent
analysis.
Solid samples received at the laboratory for trace element analysis
will range from large pieces of ore and coal to finely divided powders, fly
ash samples collected on filter pads, and filter cakes collected from slurry
sampling. Coal samples will require reduction to a workable mesh size
(e.g., 60 mesh) prior to drying and dissolution. Samples collected from
bag houses, electrostatic precipitators, and the filter cake from slurry
sampling should not require reduction.
70
-------
7.3.1 Grinding
Coal samples received at the laboratory typically fall into two dis-
tinct particle size groupings. "As received" coal is predominately 0.6 cm
(1/4 in.) in diameter. This coal may also contain a great deal of moisture
which will necessitate pre-drying at 50°C (122°F) overnight prior to grind-
ing. "As fired" coal is already ground to 100 mesh. The only further
treatment needed for this coal is drying.
To both ensure the homogeneity of the sample and expedite the decom-
position of the coal, the coal should be ground to a minimum of 100 mesh.
Place coal in a clean, one-quart (1.1-liter) ball mill. Add enough ceramic
balls to the mill until the mill is three-quarters full. Place the mill on
the rollers and allow enough grinding time to reach the required mesh size.
Two hours is usually sufficient time but this may vary depending on the
type of coal used. Next, remove the mill from the rollers. Assemble
14- and 100-mesh nylon screens and place a retaining pan on top of each
other. (In all instances, use nylon screens to minimize contaminating the
sample with small metal particles.) Empty the ball mill onto the large
mesh screen, to separate the ceramic balls from the sample. Shake to loosen
any sample that may adhere to the balls. Remove the 14-mesh screen and
return the balls to the mill. Place a top on the 100-mesh screen and place
both the screen and the pan on a shaker, or shake by hand, to sieve the
coal through the 100-mesh screen. After shaking for a short time (10-20 min-
utes), remove the top and observe for any particles that are too large to
go through the screen. If present, either regrind or use an agate mortar
and pestle to break up these particles. Repeat till 100 mesh is attained.
To clean the ball mill, replace the used balls in the mill and add
enough acetone to just cover the balls. Place the mill on the rollers for
10 minutes and then remove and empty the contents onto a large 14-mesh
screen and pan. Rinse the balls and mill with fresh acetone and let the
balls air dry; wipe the inside of the mill with paper towels to remove any
residue that may adhere to the side walls. Rinse the mesh screens with
fresh acetone until all coal is removed. Discard the acetone in an approved
waste receptacle for flammable solvents.
71
-------
7.3.2 Drying
Once the coal samples have been ground to the required 100-mesh size,
spread them evenly in large petri dishes and dry them overnight at 50 C
(122°F).(18) Coal powder samples collected from bag houses, electrostatic
precipitators, and filter cakes can be dried by spreading evenly in large
petri dishes and drying at 105UC (221UF) for 3 hours.
7.3.3 Ashing
Decomposition of the organic material in collected solid samples is
a necessary prerequisite to trace element analysis. This decomposition
should be performed using a low temperature oxygen plasma asher (Interna-
tional Plasma Corporation, Model 1001B or equivalent) to minimize the loss
of trace elements through volatilization. The procedure is as follows:
Weigh duplicate samples in acid cleaned petri dish covers. Place the petri
dish and contents into the plasma asher and begin the ashing cycle. Approx-
imately once every 4 hours, open the console and stir the coal sample to
expose fresh surface. Continue ashing 2 to 3 days or until no black par-
ticles remain.
7-4 DISSOLUTION
The dissolution of the sample is the last critical step prior to anal-
ysis. The dissolution procedure must completely solubilize all the ele-
ments of interest under conditions favorable to the retention of the more
volatile species. The following procedures, adapted from the literature
and modified slightly, have been proven effective for coal ash, fly ash,
electrostatic precipitator, bag house, and filter cake samples.
To dissolve the collected solid sample, transfer 0.5 g of the powder
material to the 24 ml Teflon acid digestion cups of a combustion bomb
(Parr Instrument Co. Model 4745 or equivalent) by tapping the edges of the
petri dishes and allowing the ash to flow through a wide,tip funnel into
the digestion bombs. Tapping the dish first will allow a minimum of ash
to escape into the room atmosphere. Once the bulk of the ash has been
removed from the dish, transfer the remaining fine particles of material
by repeated distilled water washings. To minimize the final volume, these
washings should be kept as small as possible. Next, add 6 ml ultra pure
concentrated HN03 (70 percent w/w) and 2.5 ml ultra pure concentrated HF
72
-------
7.3.1 Grinding
Coal samples received at the laboratory typically fall into two dis-
tinct particle size groupings. "As received" coal is predominately 0.6 cm
(1/4 in.) in diameter. This coal may also contain a great deal of moisture
which will necessitate pre-drying at 50°C (122°F) overnight prior to grind-
ing. "As fired" coal is already ground to 100 mesh. The only further
treatment needed for this coal is drying.
To both ensure the homogeneity of the sample and expedite the decom-
position of the coal, the coal should be ground to a minimum of 100 mesh.
Place coal in a clean, one-quart (1.1-liter) ball mill. Add enough ceramic
balls to the mill until the mill is three-quarters full. Place the mill on
the rollers and allow enough grinding time to reach the required mesh size.
Two hours is usually sufficient time but this may vary depending on the
type of coal used. Next, remove the mill from the rollers. Assemble
14- and 100-mesh nylon screens and place a retaining pan on top of each
other. (In all instances, use nylon screens to minimize contaminating the
sample with small metal particles.) Empty the ball mill onto the large
mesh screen, to separate the ceramic balls from the sample. Shake to loosen
any sample that may adhere to the balls. Remove the 14-mesh screen and
return the balls to the mill. Place a top on the 100-mesh screen and place
both the screen and the pan on a shaker, or shake by hand, to sieve the
coal through the 100-mesh screen. After shaking for a short time (10-20 min-
utes), remove the top and observe for any particles that are too large to
go through the screen. If present, either regrind or use an agate mortar
and pestle to break up these particles. Repeat till 100 mesh is attained.
To clean the ball mill, replace the used balls in the mill and add
enough acetone to just cover the balls. Place the mill on the rollers for
10 minutes and then remove and empty the contents onto a large 14-mesh
screen and pan. Rinse the balls and mill with fresh .acetone and let the
balls air dry; wipe the inside of the mill with paper towels to remove any
residue that may adhere to the side walls. Rinse the mesh screens with
fresh acetone until all coal is removed. Discard the acetone in an approved
waste receptacle for flammable solvents.
71
-------
7.3.2 Drying
Once the coal samples have been ground to the required 100-mesh size,
spread them evenly in large petri dishes and dry them overnight at 50 C
(122°F).^18) Coal powder samples collected from bag houses, electrostatic
precipitators, and filter cakes can be dried by spreading evenly in large
petri dishes and drying at 105UC (221°F) for 3 hours.
7.3.3 Ashing
Decomposition of the organic material in collected solid samples is
a necessary prerequisite to trace element analysis. This decomposition
should be performed using a low temperature oxygen plasma asher (Interna-
tional Plasma Corporation, Model 1001B or equivalent) to minimize the loss
of trace elements through volatilization. The procedure is as follows:
Weigh duplicate samples in acid cleaned petri dish covers. Place the petri
dish and contents into the plasma asher and begin the ashing cycle. Approx-
imately once every 4 hours, open the console and stir the coal sample to
expose fresh surface. Continue ashing 2 to 3 days or until no black par-
ticles remain.
7-4 DISSOLUTION
The dissolution of the sample is the last critical step prior to anal-
ysis. The dissolution procedure must completely solubilize all the ele-
ments of interest under conditions favorable to the retention of the more
volatile species. The following procedures, adapted from the literature
and modified slightly, have been proven effective for coal ash, fly ash,
electrostatic precipitator, bag house, and filter cake samples.
To dissolve the collected solid sample, transfer 0.5 g of the powder
material to the 24 ml Teflon acid digestion cups of a combustion bomb
(Parr Instrument Co. Model 4745 or equivalent) by tapping the edges of the
petri dishes and allowing the ash to flow through a wide,tip funnel into
the digestion bombs. Tapping the dish first will allow a minimum of ash
to escape into the room atmosphere. Once the bulk of the ash has been
removed from the dish, transfer the remaining fine particles of material
by repeated distilled water washings. To minimize the final volume, these
washings should be kept as small as possible. Next, add 6 ml ultra pure
concentrated HN03 (70 percent w/w) and 2.5 ml ultra pure concentrated HF
72
-------
(52 percent w/w) to the digestion cup. (Caution; HF attacks glass so
polyethylene pipets or graduated cylinders must be used.) Place the solu-
tion on an asbestos covered hot plate at 140 ±10°C (284 ±32°F) and evapo-
rate without boiling until the final volume is 50% of the original. Then,
place the sample cup in the bomb and heat the bomb in an oven at 130 ±5°C
(266 ±16°F) for a minimum of four hours.
Remove the sample from the oven and cool. After cooling, filter the
solutions through Whatman No. 41 filter paper into Nalgene polypropylene
volumetric flasks using a polypropylene funnel. Rinse with a small amount
of distilled water. With a small clean rubber policeman, scrape the Teflon
inner liner to remove any adhering ash and rinse into filter paper. When
filtering is complete, cap the volumetric flasks and transfer the filter
paper to platinum crucibles. Ignite the filter paper in a muffle furnace
at 800 ±5 C (1472 ±16 F) until no filter paper ash remains. Remove from
oven, allow to cool, then add two small scoops of ultra pure Na2C03 so that
the ratio of Na,,C03 to residue is -vlO/1. Fuse the ash and Na2C03 over a
burner flame until the crucible is cherry red and the fusion components are
in a molten state. Maintain this condition for 1 or 2 minutes or until
complete fusion has taken place.
Remove the fusion cake from the flame and allow to cool, then dissolve
using a 1:1 v/v HCl/water solution. Filter into the original volumetric
flask and repeat the washing with the 50 percent HC1 until the cake is
completely dissolved. Wash the filter paper with the same acid solution
and dilute to final 100 ml volume with distilled water. The solid sample
is now in solution and ready for trace element analysis.
73
-------
CHAPTER VIII
ANALYTICAL PROCEDURES
This chapter presents the recommended procedures for the trace element
analysis of all collected samples. These procedures are presently limited
to atomic absorption spectroscopy and other ancillary techniques such as
fluorometric, turbidimetric and specific ion electrode procedures. Table 8
lists general analytical methods which are presented in three standard ref-
erences *19* 20' and 21) on elemental analysis. However, since these meth-
ods do not completely meet all the requirements (e.g., sensitivity, accu-
racy and specificity) of trace material analysis, additional references
were consulted in order to develop the recommended procedures presented
below. These supplementary procedures, the applicable elements, and the
references consulted are tabulated in Table 9.
The recommended procedures presented in the following sections are
divided into two categories. The first category pertains to all elements
that are to be analyzed by direct Atomic Absorption Spectroscopy (AAS) and
includes Ba, Be, Cd, Ca, Cr, Cu, Pb, Mn, Hg, Ni, Sr, V, and Zn. The second
group includes the procedures applicable for the analysis of the remaining
elements and radicals. These include As, Se, SO^ , F", B, Sb, N0~, POT3, CN",
and Cl". The analyses for these remaining elements and radicals are primarily
performed using colorimetric procedures, or specialized AAS techniques.
8.1 ATOMIC ABSORPTION SPECTROSCOPY (AAS)
Metals in solution can be readily determined by atomic absorption
spectroscopy. The method is simple, rapid, and applicable to a large num-
ber of metals. Relative freedom from interference eliminates the need for
extensive sample preparation and separation techniques. When a suitable
instrument is available, the atomic absorption methods are preferable to
colorimetric procedures, although concentration of the sample by solvent
extraction may be required in order to achieve comparable sensitivity.
The solutions obtained as described in Chapter VII can be analyzed
directly by AAS for Mn, Cu, Cr, Ni, Sn, Sr, V, Pb, Cd, Zn, Ba, Cd, Ca, and
Be using the operating conditions specified in Table 10. Background cor-
rection must be used for all elements. In all cases, the standard employed
74
-------
TABLE 8. COMPILATION OF ACCEPTED STANDARD PROCEDURES
BY ELEMENT (References 19, 20, and 21)
El ement
Hg
Ba
Be
Cd
Cr
Cu
Mn
Ni
Pb
V
Zn
F"
As
B
Cl"
CN"
NO"
po;3
Sb
Se
Sr
SO'2
AOAC^19)
Col
NA
Col
AAS-Col
AAS
Col
AAS
Col
Col
Col
NA
NA
Col
Col
Col
Fluorometric
NA
Standard Methods ^20^
AAS
AAS
Pol -AAS-Col
Col -AAS
Pol -AAS-Col
Col -AAS
Pol -Col
Pol -AAS-Col
NA
Pol -AAS-Col
Specific ion
electrode-Col
Col
Col
NA
Col
Col
Col
Col
AE
Turbidimetric
ASTf^21)
NA
AAS
AAS-Col
Col
AAS-Col
AAS
AAS
AAS
Col
Col
Col
Col
Col
Col
LEGEND: AAS-Atomic Absorption Spectroscopy, AE-Atomic Emission
Col-Colorimetry, Pol-Polarography, NA-Reported Procedures
Not Applicable
75
-------
f
TABLE 9. COMPILATION OF REFERENCES FOR RECOMMENDED PROCEDURES
Element
or
Radical
Hg
Ba
Be
Ca
Cd
Cr
Cu
Mn
Ni
Pb
V
Zn
F"
As
B
Cl"
CN"
N03
po-3
Sb
Se
Sr
S042
Procedure
Flameless AAS
AAS
11
"
11
"
"
"
11
"
"
"
Specific Ion Electrode
Colorimetric
"
Ti trati on/Col orimetri c
Colorimetric
:
H2-Ar AAS
Colorimetric
Gravimetri c/Turbidimetri c
Page
Where
Found
8-14
8-6
8-6
8-6
8-6
8-7
8-7
8-7
8-9
8-7
8-8
8-8
8-12
8-9
8-11
8-33
8-23
8-44
8-56
8-51
8-52
8-7
8-18
References
22, 23, 24, 25, 6
20, 22, 24
20, 22, 24, 6
19, 21, 22, 24
20, 21, 22, 24, 26, 27,
12, 24, 28, 29
20, 21, 22, 24, 29, 30,
31
19, 20, 22, 24, 27, 28,
29, 31
19, 20, 21, 22, 24, 28,
29, 31, 32
21, 22, 24, 28, 31
20, 21, 22, 24, 28, 29,
31
22, 24, 33
19, 20, 21, 22, 24, 27,
28, 29, 31
19, 34, 35, 36
19, 20, 21, 37, 38, 39
19, 20, 39, 40, 41
21
20, 21
19, 20, 21, 42
19, 20, 21
19
48
20, 32
20
76
-------
TABLE 10. ATOMIC ABSORPTION OPERATING PARAMETERS
Element
Ba
Be
Cd
Ca
Cr
Cu
Pb
Mn
Hg
Ni
Se
Sr(AE)
y
Zn
Slit
Widths nm
0.4
1.0
0.4
1.0
0.2
1.0
0.4
0.4
1.0
0.2
1.0
0.2
1.0
1.0
Wavelengths nm
Analytical
553.6
234.9
228.8
422.7
357.9
324.7
283.3
279.5
253.7
232.0
196.0
460.7
318.4
318.4
Background
-
Ni-231.6 Non-
Absorbing
226.5
-
351.9
323.4
282.0
Pb 282.0 Non-
Absorbing
Si 252.5
231.6
Hg - 194.2
-
312.5
210.0
Gas Mixture
NpO-acetylene
NpO-acetylene
Air-acetylene
NpO-acetylene
NpO-acetylene
Air-acetylene
Air-acetylene
Air-acetylene
Flame! ess
Air-acetylene
hL - Argpn
-
N?0-acetylene
Air-acetylene
Detection
Limit ppm
0.02
0.005
0.003
0.005
0.005
0.003
0.03
0.003
0.001
0.01
0.003
0.005
0.01
0.003
Comments
Add O.U K as an ioni-
zation suppressant
—
-
Add 1% K as an ioniza-
tion suppressant
~
-
-
Reduction using SnCl?
Use standard additions
procedure to eliminate
interferences
-
-------
for calibration of the instrument should match the sample matrix as closely
as possible. If no chemical or matrix interferences are found after per-
forming accuracy checks, distilled water standards may be used. The imple-
mentation of either the factor method, short curve or standard addition
technique (see Section 6.4) for obtaining the required accuracy is sample
dependent and also depends on the skill of the analyst.
8.1.1 Types of Interferences
AAS as a general analytical tool is normally considered free of inter-
element interferences, and, because of the large dilutions employed, is
usually unresponsive to matrix changes. However, for trace elemental anal-
ysis of coal ash and other types of solid samples, the elements of interest
can be present in a very dilute form in a relatively concentrated matrix
consisting of the major inorganic components of the sample and the rela-
tively high concentrations of fluxes and acids needed for the dissolution.
High solids concentrations as well as complicated matrices make it manda-
tory for the analyst to be aware of and to investigate the presence of
interferences. The types of interferences commonly encountered are classi-
fied into the following three categories:
• Interelement or chemical interferences. For the most part,
these interferences when present can be eliminated by using
a high temperature N^O-acetylene flame, or by adding
suppressants.
• Matrix effects. These interferences are physical in nature
and stem from the large concentrations of acids and solids
in solution. These effects are compensated for by specially
preparing the standards to match the expected acid and salt
content of the sample, or by applying standard addition
techniques.
• Molecular absorptions. This type of spectral interference
can be particularly troublesome when determining trace
elements in solutions of high salt content. Molecular absorp-
tions predominately occur from species such' as CaOH or SrO
and result in a positive error in the absorption measurement.
The Jarrell-Ash 810 AAS or equivalent is especially suited
for the evaluation and elimination of this type of interference.
Since molecular absorptions are normally broad, the presence
of this interference can be ascertained by monitoring a non-
absorbing wavelength near the wavelength of interest on a
second channel. The molecular absorption, when present, is
visually recorded on a strip chart recorder concurrently with
the absorption of the desired element. The interference is
then subtracted from the combined signal.
78
-------
The following subsections discuss the known interferences for each
element determinate by AAS/22' 23' 24» 26 and 33)
8.1.1.1 Barium —
Only molecular absorbance by CaOH has been reported to interfere with
the analysis of barium using an air-acetylene flame. This effect is removed
when using the higher temperature NgO-acetylene flame, although the use of
this flame necessitates the addition of 1% K ion to suppress the ionization
of barium.
8.1.1.2 Beryllium —
No reported interferences. Beryllium can be determined directly using
an N20-acetylene flame. Additional sensitivity can be obtained by extract-
ing the 8-quinolinol complex at a pH of 8.0 ±0.5 using chloroform.
8.1.1.3 Cadmium —
Cadmium is one of the metals more sensitively determined by AAS using
the 228.8 nm resonant wavelength. The following elements and compounds
have been found not to interfere at the following cation weight ratios:
Cr/Cd = 1000, Ni/Cd = 1000, Mn/Cd = 1000, Co/Cd = 1000, Cu/Cd = 500,
Mo/Cd = 250, Ti/Cd = 250, V/Cd = 250, Pb/Cd = 250, Al/Cd = 250. A molec-
ular absorption interference by Fe on the Cd 228.8 nm line was found; how-
ever, it was corrected for by measuring the absorbance effect on the non-
resonance 232.12 nm line. The nonresonance 226.5 line can also be used to
correct for molecular absorption. Molecular absorption by a 0.01 M NaCl
solution was reported by Pulidlo^ who also reported a depression with
0.1 M H3P04> No interference was found in 100 mg/1 chloride salt solutions
of Ba, Ca, Co, Cr, Cu, Fe, Li, Mg, Mn, or Ni. Tenth molar solutions of
HN03 or H2S04 caused no interference, nor did 0.64 M HC104- Additional
sensitivity for cadmium can be obtained by extracting its APDC complex into
methyl isobutyl ketone (MIBK).
8.1.1.4 Calcium —
Interferences with calcium analysis have been reported from sulfate,
phosphate, alumina, and silica. These interferences are reported to be of
a chemical nature and can be eliminated by making the solutionJ percentjn
lanthanum. No interferences have been found using 1000 ppm CT, N03, N02,
HCO", EDTA, Fe, Ni, Zn, Mn, Cr, B, Pb, Mg, or Na.
79
-------
The following procedurecan be used for lanthanum addition: Wet
58.65 g lanthanum oxide (La203) with water. Add slowly 250 ml of HC1
(sp gr 1.19) to the mixture. When dissolved, dilute to 1 liter with water.
To 10 ml of sample, add 2.5 ml of above solution. This solution can also
be used for Sr analysis.
8.1.1.5 Chromium --
Ratios of Ni/Cr = 1000, Mn/Cr = 1000, Ca/Cr - 1000, Cu/Cr = 400,
and Mo/Cr = 200 in alloys and steel produced no interference on the absor-
bance of Cr; however, iron is a serious depressant. This effect can be
reduced by the addition of ammonium chloride or by determining chromium in
an N?0-acetylene flame. In an investigation of water samples, Platte and
Marcy' ' found no interference on 1 ppm Cr by 1000 ppm of SO^ , Cl",
P0~3, N03, NO];, HCO~, Si, EDTA, Fe, N1, Zn, Mn, B, Pb, Ca, Mg, or Na.
8.1.1.6 Copper —
No interference with copper absorption is found with 2000 ppm Ni, Cr
or Mn, or 1000 ppm Co or V. Also no interference has been found for HC1,
HN03, H2S04, or H3P04> 1000 ppm NO", HCO~, Si, EDTA, Fe, Ni, Mn, Zn, Cr,
B, Pb, Ca, Mg, or Na.
8.1.1.7 Lead —
In general, no chemical interferences have been found in lead analysis
using an air-acetylene flame. The lead resonance wavelength at 217.0 nm is
approximately twice as sensitive as the 283.3 nm line, but the latter is
preferred because of the lower flame absorption and noise level at this
wavelength. Lead is quantitatively removed over a wide pH range when an
APDC-MIBK extraction system is used.
8.1.1.8 Manganese —
Manganese has few interferences in an air-acetylene flame. One ppm of
manganese is unaffected by 1000 ppm Na, Mg, SO'2, N0~, NO" HCO" SiO ,
EDTA, Ni, Zn, Cr, B, or Pb.
8.1.1.9 Strontium --
Atomic emission spectroscopy is generally preferred for strontium, but
atomic absorption is also acceptable. The atomic emission method enables
80
-------
determination of strontium in the small concentrations prevalent in natural
water supplies. The strontium emission is measured at a wavelength of
460.7 nm. Because the background intensity at a wavelength of 454 nm equals
that at 460.7 nm and is unaffected by the variable strontium concentration,
the difference in readings obtained at these two wavelengths allows an esti-
mate of the light intensity emitted by strontium.
The emission intensity is a linear function of the strontium concen-
tration and also the concentration of the other constituents in the sample.
The standard addition technique distributes the same ions throughout the
standards and the unknown, thereby equalizing the radiation effect of pos-
sible interfering substances in the standards and the unknown.
Only limited work on strontium by AAS has been reported. No inter-
ferences have yet been reported in a O-acetylene flame, and by analogy
with calcium, interferences are unlikely. The use of this high temperature
flame necessitates the addition of 1 percent lanthanum as an ionization
suppressant.
8.1.1.10 Vanadium —
No interference has been reported for vanadium using a N20-acetylene
flame except a signal enhancement noted with H3P04- The three strong lines
at 318.34, 318.40 and 318.54 nm are normally used together and give a lin-
ear working curve to very high values of absorbance.
8.1.1.11 Zinc --
No interferences have been reported for zinc using a pre-mix air-
acetylene flame. Zinc is totally extracted into MIBK as the chelate of
APDC over a pH range of 2.5 to 5.
8.1.1.12 Mercury —
Trace quantities of mercury are most easily determined using a flame-
less atomic absorption spectrometric procedure. This technique depends on
reducing mercury to the elemental state and passing the vapor through a
quartz absorption cell of the spectrometer where its concentration is
measured. This technique is detailed in section 8.2.4.
81
-------
8.1.1.13 Nickel —
In a lean air-acetylene flame no interferences were found from solu-
tions containing 3000 ppm Cr, 2000 ppm Mn or W, 1000 ppm Cu and Co, or
5000 ppm V, Mo and Al. No interferences were found on a 1 ppm Ni sample
in the presence of 1000 ppm sulfate, phosphate, nitrate, nitrite, bicar-
bonate, silica, EDTA, Fe, Zn, Mn, Cr, B, Pb, Ca, Mg, or Na. Molecular
absorption by CaO has been reported as an interferent of the Ni 232.0 nm
line. Corrections for this absorption can be made on the 232.57 nm Ni line.
8.2 ANCILLARY TECHNIQUES
The following sections present the procedures for the quantification
of As, Se, SO"2, Hg, F, B, Sb, NO^, P0~3, CN~, and Cl" ions. For several
radicals and elements (e.g., SO^2, NOg, CM", and Cl"), more than one ana-
lytical procedure is specified. This permits a choice of methods if either
the concentration levels or presence of interfering elements will favor one
procedure or another. Interferents are specified where known and applica-
tions of the procedure to specific sample types are given where appropriate.
8.2.1 Arsenic Analysis
A sample of coal is mixed with MgO and combusted at 550°C (1022°F) in
a muffle furnace. The residue is transferred to a 125 ml Erlenmeyer flask
and treated with HC1 and KI. Stannous chloride and metallic Zn are added
and the arsenic is then volatilized as arsine and absorbed in a silver
diethyldithiocarbamate pyridine solution. The quantitative determination
is performed by comparing the absorbance of the developed color at 540 nm
to standards.
Fly ash samples are run using the same procedure except that a 0.1 g
sample is weighed directly into the Erlenmeyer flask. The MgO sintering
step is omitted.
8.2.1.1 Apparatus and Reagents --
• 20 percent w/v KI solution - 20 g KI, dissolve in 100 ml
deionized water
• SnCl2 solution - Dissolve 100 g SnCl2 in 100 ml concentrated
1 HC1; assist solution with application of heat
• Acidified water - Dissolve 5 ml cone. H2$04 in 500 ml water
82
-------
• MgO - reagent grade
• Zn - 40-mesh granular
• Lead acetate solution - saturated in water
0 Silver diethyldithiocarbamate, pyridine solution - Dissolve
5 g of reagent grade salt in one liter of pyridine. Allow
the solution to stand in a covered container for 48 hours
Filter the solution through a Whatman No. 40 filter and
store over molecular sieves in a brown bottle.
8.2.1.2 Analysis Procedure
To a procelain crucible, add 1.0 g sample and 1.0 g MgO, and mix the
contents. To another porcelain crucible, add 1.0 g MgO and no sample.
This will be used later for the blank. Place all crucibles into a muffle
furnace and heat slowly to 550°C (1022°F) and maintain at this temperature
for 1-1/2 hours. Remove from oven and allow to cool. Transfer the fused
mass to a wide-mouthed Erlenmeyer flask using three 5 ml rinsings of acidi-
fied water. Before transferring, wet the sample by slowly rinsing down
the sides of the crucible with the acidified water. Repeat until the sam-
ple is completely wetted. Wash crucible with the acid water solution until
an approximate volume of 50 ml is attained. Repeat following the same pro-
cedure with the blank. For fly ash samples, weigh 0.1 g directly into an
Erlenmeyer flask followed by 50 ml H,,0.
To all the flasks, add 5.0 ml cone. HC1, 2.0 ml of KI solution and
1.0 ml of the SnCl2 solution. Allow the solutions to stand for 15 minutes.
Fly ash samples are heated gently for 15 minutes. At the end of this time,
the reaction flasks are connected to a receiving flask by a tube containing
glass wool to which a few drops of a saturated lead acetate scrubbing solu-
tion have been added. Transfer 10 ml of silver diethyldithiocarbamate
pyridine solution to the receiving flask and a 3 g portion of granular zinc
to the reaction flask. Connect the reaction and receiving flasks together
in as short a time as possible to prevent any arsine gas loss. After allow-
ing 30 minutes for complete gas evolution, remove the receiver vessel and
mix the solution by bubbling nitrogen through the solution to remove any
residue that is adhering to the side wall. Transfer the absorbing solution
to 1-on quartz cells and measure its absorbance at 540 m against the blank
reagent using a spectrophotometer.
83
-------
8.2.1.3 Standardization Procedures —
Before running As determinations, prepare a 100 ppm As stock solution
(10 ml of 1000 ppm As diluted to 100 ml with distilled water). Once the
stock solution is prepared, take 1, 2, 5, 10, and 15 ml portions of the
100 ppm standard, transfer to five 100 ml volumetric flasks and dilute to
marks with distilled water. These 1, 2, 5, 10, and 15 ppm As solutions are
the working standards.
Place one gram of MgO in each of six ceramic crucibles and heat in
a 550°C (1022°F) muffle for 1-1/2 hours. Remove and cool; transfer to
a 125 ml Erlenmeyer flask with acidified water. Pipet one ml of each of
the five standards into its respective Erlenmeyer flask and proceed as
above. Note the following:
1) The pyridine-silver diethyldithiocarbamate solution will
deteriorate slightly and, if not filtered, will lead to
erratic values.
2) The type of mesh zinc used appears to have some bearing on
the arsine evolution. Therefore, only one bottle should
be designated for use and a new calibration curve should
be run when another bottle is employed.
3) Heating the reaction solution facilitates the evolution of
arsine and has proved helpful in improving the accuracy of
the analysis.
8.2.2 Boron Analysis
The coal is gently ashed at 550°C (1022°F), then fused with Na^O-.
After dissolving the fusion mixture in HC1, the boric acid is extracted
with 2-ethyl-l, 3-hexanediol and determined as the rosocyanine complex in
95 percent ethanol. This procedure is applicable for coals and fly ash
containing from 1-400 ppm B.
8.2.2.1 Apparatus and Reagents —
• 10 ppm standard boron solution. Prepare by appropriate
dilution of 1000 ppm stock boron solution, 2-ethyl-l,
3-hexanediol - 10 percent solution in chloroform.
• Curcumin reagent, 0.375% (w/v). Dissolve 0.375 g
curcumin in 100 ml glacial acetic acid; filter the solu-
tion and store it in a darkened polyethylene bottle.
• Ethanol - 95% reagent grade.
84
-------
• Sulfuric acid - high purity (Van Waters and Rogers ultrex
grade).
• Na2C03 - high purity (Van Waters and Rogers ultrex grade).
• IN HC1. Transfer 28 ml high purity concentrated HC1
to a 1-liter flask and fill to the mark with distilled water.
8.2.2.2 Analysis Procedure --
Weigh 1 g coal ±0.1 mg into a platinum crucible; ash at 550°C (1022°F)
for 1 hour. Fuse residue with 1 g of Na^CL, then dissolve the melt with
10 ml IN HC1 and dilute to 100 ml. For fly ash samples, weigh 0.1 g into
platinum crucible and fuse with 1 g Na?CO, and proceed as above. Pipet
2 ml of this solution into a 10-ml Nalgene centrifuge tube and extract by
shaking with 2 ml 2-ethyl-l ,3-hexandiol in CHC1-. Syringe off the liquid
phase and pipet 0.5 ml of the organic phase into a 50-ml Nalgene volumetric
flask. Add 1 ml of curcumin reagent followed by 0.3 ml of cone. H,,SO. and
allow to react for 15 minutes. Adjust the volume to 50 ml with reagent
grade 95 percent ethanol and read absorbance at 550 nm against 95 percent
ethanol. Run a reagent blank concurrently and subtract this absorbance
from the sample absorbance. The boron concentration of the sample is
calculated from a standard curve using the adjusted sample absorbance
reading.
8.2.2.3 Standardization Procedures —
Prepare standard solutions containing 0.1, 6.2, 0.5, 1.0, 2.0, and
3.0 ppm boron by successive dilution of the 10 ppm standard. Pipet 2 ml
of prepared standard into a Nalgene centrifuge and proceed as per general
procedure. Note that all apparatus is to be washed with 1:1 HNOg.
8.2.3 Fluoride Analysis
Coal is mixed with benzoic acid, pressed into a pellet and combusted
in a Parr bomb, and the combustion gases are scrubbed with a dilute caustic
solution. The pH of the solution is adjusted to 5.2-5.3 and C02 expelled
by gentle heating. The fluoride concentration is then determined using a
specific ion electrode procedure after readjusting the PH and addition of
a citrate, KNO- solution.
85
-------
8.2.3.1 Apparatus and Reagents —
• IN NaOH. Prepared from high purity reagents.
• 0.5N H?S04. Prepared from high purity reagents.
• Sodium citrate, KNO- buffer solution. Dissolve 294 g of
citric acid trisodiam salt dihydrate and 20 g of KN03 in
one liter of high purity water.
t Fluoride standard. Prepare a series of standard fluoride
solutions in the following molar concentrations: 0.0005,
0.001, 0.005, 0.01, 0.05, 0.10. Prepare by dissolving high
purity KF in buffer.
8.2.3.2 Procedure —
Mix a 1 g coal sample, ground to pass a 100-mesh screen, with about
0.25 g benzoic acid (primary standard) and place in a fused quartz sample
holder within a Parr combustion bomb that contains 10 ml of IN NaOH. Pres-
surize the bomb to about 28 atmospheres oxygen and ignite the contents.
Allow at least 15 minutes to elapse before depressurizing the bomb. Use
three approximate 5 ml aliquots of distilled water to rinse the bomb con-
tents into a 50 ml plastic beaker (plastic-ware is used from here on).
The beaker contents are magnetically stirred wtih a Teflon bar while
the pH is adjusted to 5.2-5.5 with 0.5N H2S04> (The initial pH before
adjustment will be about 7.0.) Place the beaker in a hot water bath for
about 10 minutes, remove it, and again stir to drive off most of the dis-
solved (XL- Add 5 ml of 1M sodium citrate-citric acid KNO~ buffer (pH 6.3)
to the beaker contents. Adjust the total volume to 50 ml with distilled
water and cool to room temperature. Read the potential of the solution
using a fluoride specific ion electrode vs a saturated calomel reference
electrode. In some cases, about 10 minutes are required for equilibrium
to be attained. Add 1 ml of 0.01M F to the solution, mix the solution well,
and read the potential of the solution again.
The pH is critical for the initial potential reading. At 5.0 to 5.5,
final results tend to be low because of F~ complexing with H+. Above 6.5,
final results tend to be high because of interference by OH" or HCO - at
1000 to one concentration over the F.
86
-------
8.2.3.3 Concentration Calculations —
The concentration of fluoride in the coal sample is calculated using
the following formulas:
F = AF
rS (exp [AE/S] - 1)
50 x F_ - (W1 x FB)
where
FS = fluoride content of the solution
AF = change in F concentration due to addition of spike = 3.8 ppm
AE = change in potential readings = E? - E,
S = slope of mv vs In (F ) concentration for the electrode,
S = -22.95
FC = fluoride concentration of coal
W.j = weight benzoic acid, g
FB = F content of benzoic acid
W2 = weight of coal taken for analysis, g
8.2.4 Mercury Analysis
The flameless cold vapor technique is used to analyze for mercury in
solution. This technique permits routine analysis at the ppb level; modi-
fications permit analysis of 0.1 ppb (or lower) solutions. Organic solid
samples (i.e., coal) can also be analyzed by burning the sample in a com-
bustion bomb, absorbing the mercury vapor in nitric acid, and analyzing
the solution.
8.2.4.1 Apparatus and Reagents --
• Mercury reduction apparatus. The usual desi"9n' ""s;s^h°f
a jar incorporating a two-hole rubber stopper through which
are passed a gas bubbler tube and a short gas outlet tube can
c starrer and
be used; the ontents are stirred using a
stirring bar. The design is essentially a U-tube with a glass
87
-------
frit in one side. The frit serves as the mixing device as
well as the gas bubbler, thus precluding the use of the
magnetic stirrer.
t Atomic absorption spectrophotometer. Use mercury hollow
cathode lamp and a wavelength of 253.7 nm.
• Absorption cell. A cylindrical tube approximately 25mm I.D.
x 125 mm long, with quartz windows, and incorporating inlet
and outlet sidearms to permit introduction and discharge of
carrier gas. This type of cell is available commercially from
several manufacturers of atomic absorption equipment, or it
may be constructed from readily available materials. In the
latter case, the cell should be tested carefully for possible
leakage after assembly. The cell is mounted in the optical
path of the AAS.
• Flowmeter. Capable of measuring gas flows in the range of
1.9 liters-min-1 (4 ft3-hr-l).
• Scavenging tube. This tube is filled with soda lime and is
connected between the gas outlet tube of the reduction vessel
and the inlet sidearm of the absorption cell. The soda lime
is replaced every 25 determinations; otherwise, a loss in
sensitivity occurs. Tygon tubing is used as connecting tub-
ing; no interferences have been noted.
In the case of solid samples, the following combustion apparatus and
reagents are required:
• Oxygen bomb. Standard 360 ml stainless steel combustion
bomb as used for coal calorimetry (45).
• Combustion crucible. Vycor or quartz crucible of proper
size to fit the bomb sample holder (A.H. Thomas No. 3879-C
or equivalent).
• Firing wire. No. 34 B & S gauge nickel-chromium alloy wire,
10 cm length.
• Firing circuit. As described in Reference 45.
• Stock mercury solution, approximately 1 gram/liter (1000 ppm).
Weigh one gram of pure, elemental mercury to the nearest
0.1 mg and dissolve in a solution consisting of 150 ml rea-
gent water and 50 ml concentrated HNOs (sp. gr. 1.42). Dilute
this solution to 1000 ml with reagent water. The final solu-
tion contains approximately 1000 ppm mercury (record exact
concentration) in a matrix of 5 percent v/v nitric acid.
• Standard mercury solutions. Prepare working standard solu-
tions of mercury down to 1 ppm by serial dilutions of the
1000 ppm Hg stock solution with 5 percent v/v HNOg. Such
-------
solutions can be assumed to be stable for up to one week
Below 1 ppm Hg, standard solutions should be prepared daily
and diluted with 5 percent HMOs and/or reagent water as
appropriate so the final solution matrix is approximately
1 percent v/v HN03- J
• Nitric Acid Solution. Dilute 500 ml cone, nitric acid to
1000 ml.
• Stannous Chloride Solution. Dissolve 20 g of SnCl2-2H20
in 20 ml cone. HC1 (warni the solution to accelerate the
dissolution process) and dilute to 100 mis.
• Argon Carrier Gas.
8.2.4.2 Standardization Procedures --
Standards in the range of 1 ppb-10 ppb are made. To the reduction
vessel, transfer 10 ml nitric acid solution, 5 ml of a standard solution,
and 5 ml of the stannous chloride solution. Close the system immediately.
For our system, initiate the argon flow at 1.9 liters-min" (4 ft hr" );
for the conventional system, stir for one minute, then initiate the argon
flow . Repeat the procedure for varying concentrations of mercury through-
out the specified range. For our system, a loss in sensitivity is noted
unless the glass frit is cleaned between analyses (flush with 1:1 nitric
acid, followed by deionized water).
Blanks should be run using a deionized water in place of the standard.
Plot absorption (peak height) against standard concentration to obtain a
working curve.
8.2.4.3 Solution Analysis Procedures —
Aqueous samples can be analyzed by a procedure identical to that used
for standardization. If a different sample size (e.g., 50 ml instead of
5 ml) is used, a new calibration curve must be constructed using the new
sample size.
'The optimum flow rate depends on the size of the absorptionj cell. Several
flow rates should be tried until maximum sensitivity is obtained.
89
-------
Strongly oxidative solutions require modifications to the procedure.
Larger amounts of stannous chloride reagent must be used, and a "reaction
time" (typically one minute) must be allowed after the system is closed but
prior to initiating the argon flow. As recovery of mercury from these
solutions is usually not quantitative, standard additions should be per-
formed on these samples. Of course, blanks must be run on these oxidative
solutions.
Silver ion appears to interfere by forming silver chloride which clogs
the glass frit.
8.2.4.4 Organic Solids Analysis Procedures --
Mix 1 g of coal and ^0.25 g of benzoic acid. Press into a pellet and
place in a fused quartz crucible. Transfer 10 ml of nitric acid solution
to the bomb; place the crucible in the electrode support of the bomb and
attach the fuse wire. Assemble the bomb and add oxygen to a pressure of 24
atmospheres (gauge). Place the bomb in the calorimeter (a cold water bath
in a large stainless steel beaker is also satisfactory) and ignite the sam-
ple using appropriate safety precautions ordinarily employed in bomb calo-
rimetry work.
After combustion, the bomb should be left undisturbed for 10 minutes
to allow temperature equilibration and the absorption of soluble vapors.
Release the pressure slowly and transfer the contents of the bomb (and
crucible) to the mercury reduction vessel by washing with nitric acid
2
solution.
Rinse the bomb, electrodes, and crucible thoroughly with several small
washings of nitric acid, then dilute the contents of the reduction vessel
with nitric acid to a total volume of 50 ml. Proceed with the determination
2
If there is any question as to whether the sample has undergone complete
oxidation during combustion, add 5 percent potassium permanganate solution
dropwise until a pink color persists.
90
-------
as described under section 8.2.4.2. Calculate the amount of mercury con-
tent as follows:
wi
Hg content (ppm) = rp-
W2
where
W, = weight of mercury, yg
I/L = weight of coal sample, g
As the bomb ages, there may be a tendency for mercury to become trapped
in the bomb wall fissures during combustion. In addition, if the same
bomb is used for normal calorimetry work, there may be a tendency for mer-
cury to accumulate in the bomb with time. Consequently, before a series of
mercury determinations is undertaken, several blank determinations should
be made by firing benzoic acid pellets (approximately 1 gram) in place of
the coal. Benzoic acid firings should be repeated until a stable, consis-
tently low blank value is obtained. This final blank value is then used
to correct the mercury values obtained for subsequent coal samples.
8.2.5 Sulfate by the Gravimetric Method
Sulfate is precipitated in a hydrochloric acid medium as barium sul-
fate by the addition of barium chloride. The precipitation is carried out
near the boiling temperature, and after a period of digestion the precipi-
tate is filtered, washed with water until free of chlorides, ignited or
dried, and weighed as BaSO..
The gravimetric determination of sulfate is subject to many errors,
both positive and negative. In potable waters where the mineral concen-
tration is low, these may be of minor importance. The analyst should be
familiar with the more common interferences, however, so that he can apply
corrective measures when necessary.
3The condition of the interior of the bomb should be nnservea
intervals. If evidence of significant pitting or corrosion is ooservea
(usually indicated by erratic mercury values for samples or benzoic acid
blanks), the bomb should be returned to the manufacturer for reconditioning.
91
-------
1. Interferences Leading to High Results - Suspended matter,
silica, barium chloride precipitant, nitrate, sulfite, and
water are the principal factors in positive errors. Suspended
matter may be present in both the sample and the precipitating
solution; soluble silicate may be rendered insoluble and sul-
fite may be oxidized to sulfate during processing of the sam-
ple. Barium nitrate, barium chloride, and water are occluded
to some extent with the barium sulfate, although water is
driven off if the ignition temperature is sufficiently high.
2. Interferences Leading to Low Results - Alkali metal sulfates
frequently yield low results. This is especially true of
alkali hydrogen sulfates. Occlusion of alkali sulfate with
barium sulfate causes the substitution of an element of lower
atomic weight than barium in the precipitate. Hydrogen sul-
fates of alkali metals act similarly and, in addition, decom-
pose on being heated. Heavy metals, such as chromium and
iron, cause low results by interfering with the complete pre-
cipitation of sulfate and by formation of heavy metal sulfates.
Barium sulfate has small but significant solubility, which
is increased in the presence of acid. Although an acid medium
is necessary to prevent precipitation of barium carbonate and
phosphate, it is important to limit its concentration to
minimize the solution effect.
8.2.5.1 Apparatus and Reagents —
• Steam bath.
• Drying oven, equipped with thermostatic control.
• Muffle furnace, with heat indicator.
• Desiccator, preferably containing a desiccant with color
indicator of the water content.
• Analytical balance, capable of weighing to 0.1 mg.
• Filters: either acid-washed, ashless hard-finish filter
paper sufficiently retentive for fine precipitates, or
porous-bottom silica or porcelain crucible with a maximum
pore size of 5 microns.
• Filtering apparatus, appropriate to the type of filter
selected.
• Methyl red indicator solution. Dissolve 100 mg methyl red
sodium salt in distilled water and dilute to 100 ml.
• Hydrochloric acid - 50 percent v/v.
92
-------
•
Barium chloride solution. Dissolve 100 g Bad? • 2H?0 in
1 liter distilled water. Filter through a membrane filter
or hard-finish filter paper before use; 1 ml of this reagent
is capable of precipitating approximately 40 mg $04.
• Silver nitrate-nitric acid reagent. Dissolve 8.5 g AgNO-s
and 0.5 ml reagent grade HN03 in 500 ml distilled water.
8.2.5.2 Analysis Procedure•--
1. Removal of cation interference - If the total cation concen-
tration in the sample is 250 mg/1 or more, or if the total
heavy metal ion concentration in the sample is 10 mg/1 or
more, pass the sample portion intended for sulfate precipi-
tation through a cation-removing ion-exchange column.
2. Removal of silica - If the silica concentration exceeds 25 mg/1,
evaporate the sample nearly to dryness in a platinum dish on
a steam bath. Add 1 ml HC1, tilt the dish, and rotate it
until the acid comes in contact with the residue on the sides;
continue the evaporation to dryness. Complete the drying in
an oven at 180°C (356°F) and if organic matter is present,
char over the flame of a burner. Moisten the residue with
2 ml distilled water and 1 ml HC1, and evaporate to dryness
on a steam bath. Add 2 ml HC1, take up the soluble residue
in hot water, and filter. Wash the insoluble silica with
several small portions of hot distilled water. Combine the
filtrate and washings.
3. Precipitation of barium sulfate - Adjust the clarified sample,
treated if necessary to remove interfering agents, to contain
approximately 50 mg sulfate ion in a 250 ml volume. Record
the volume of sample taken. Adjust the acidity with HC1 to
pH 4.5-5.0, using a pH meter or the orange color of methyl
red indicator. Then add an additional 1 to 2 ml HC1 to the
solution. Lower concentrations of sulfate ion can be toler-
ated if it is impracticable to concentrate the sample to the
optimum level, but in such cases it is better to fix the total
volume at 150 ml. Heat the solution to boiling and, while
stirring gently, add warm barium chloride s°l"tlon.Jlowly
until precipitation appears to be complete, then add about
2 ml in excess. If the amount of precipitate is small, add
a total of 5 ml barium chloride solution Digest the pre-
cipitate at 80-90°C (176-194<>F) preferably overnight but for
not less than 2 hours.
4. Preparation of .filters - If paper "I^'^^M
in the con1
preignite i
and weigh.
93
-------
5. Filtration and weighing - Mix a small amount of ashless filter
paper pulp witji the barium sulfate and filter at room tempera-
ture. The pulp aids filtration and reduces the tendency of
the precipitate to creep. Wash the precipitate with small por-
tions of warm distilled water until the washings are free of
chloride, as indicated by testing with silver nitrate-nitric
acid reagent. Dry the filter and the precipitate, and ignite
at 80QOC (1472QF) for 1 hour. Do not allow the filter paper
to flame. Cool in a desiccator and weigh the crucible and
contents.
6. Calculation
Sulfate content (mg/1) =
where
W = weight of barium sulfate preciptitate, mg
v = volume of sample taken for analysis, ml
8.2.6 Sulfate by Turbidimetric Method
Sulfate ion is precipitated in a hydrochloric acid medium with barium
chloride in such manner as to form barium sulfate crystals of uniform size.
The absorbance of the barium sulfate suspension is measured by a nephelom-
eter or transmission photometer and the sulfate ion concentration is deter-
mined by comparison of the reading with a standard curve.
Color or suspended matter in large amounts will interfere with this
method. Some suspended matter may be removed by filtration. If both are
small in comparison with the sulfate ion concentration, interference is
corrected for as indicated in Section 8.2.6.2? Silica in excess of 500 mg/1
will interfere, and in waters containing large quantities of organic mate-
rial it may not be possible to precipitate barium sulfate satisfactorily.
There are no ions other than sulfate in normal waters that will form
insoluble compounds with barium under strongly acidic conditions. Deter-
minations should be made at room temperature, which may vary over a range
of 10°C (18°F) without causing appreciable error.
The minimum detectable concentration is approximately 1 mg/1 sulfate.
94
-------
8.2.6.1 Apparatus and Reagents --
t Magnetic stirrer It is convenient to incorporate a timing
device to permit the magnetic stirrer to operate for exactly
1 minute. The stirring speed should not vary appreciably It
is also convenient to incorporate a fixed resistance in series
with the motor operating the magnetic stirrer to regulate
the speed of stirring. If more than one magnet is used, they
should be of identical shape and size. The exact speed of
stirring is not critical, but it should be constant for each
run of samples and standards and should be adjusted to about
the maximum at which no splashing occurs.
• Photometer. One of the following is required with preference
in the order given: nephelometer, such as Coleman Model #9;
spectrophotometer, for use at 420 nm and providing a light
path of 4-5 cm; filter photometer, equipped with a violet fil-
ter having maximum transmittance near 420 nm and providing a
light path of 4-5 cm.
• Stopwatch, if the magnetic stirrer is not equipped with an
accurate timer.
• Measuring spoon, capacity 0.2-0.3 ml.
• Conditioning reagent. Mix 50 ml glycerol with a solution
containing 30 ml concentrate HC1, 300 ml distilled water,
100 ml 95 percent ethyl or isopropyl alcohol, and 75 g
sodium chloride.
• Barium chloride, crystals, 20-30 mesh.
t Standard sulfate solution. Prepare a standard sulfate solu-
tion by diluting 10.41 ml of standard 0.0200N H2S04 titrant
to 100 ml with deionized water. Dissolve 147.9 mg anhydrous
sodium sulfate, Na?S04, in distilled water and dilute to
1000 ml.
8.2.6.2 Analysis Procedure --
1. Formation of barium sulfate turbidity - Measure 100 ml sample,
or a suitable aliquot made up to 100 ml, into a 250-ml Erlen-
meyer flask. Add exactly 5.00 ml conditioning reagent and
mix in the stirring apparatus. While the solution is being
stirred, add a spoonful of barium chloride crystals and begin
the timing immediately. Stir for exactly 1 minute at a con-
stant speed.
2. Measurement of barium sulfate turbidity - J™"^1^!,1^.^^,.
the stirring period has ended, pour some of the solution into
the absorption cell of the photometer and measure the turbidity
at 30-second intervals for 4 minutes. Since maximum
95
-------
usually occurs within 2 minutes and the readings remain con-
stant thereafter for 3 to 10 minutes, consider the turbidity
to be the maximum reading obtained in the 4-minute interval.
3. Preparation of calibration curve - Estimate the sulfate con-
centration in the sample by comparing the turbidity reading
with a calibration curve secured by carrying sulfate standards
through the entire procedure. Space the standards at 5 mg/1
increments in the 0 to 40 mg/1 sulfate range. Above 40 mg/1
the accuracy of the method decreases and the suspensions of
barium sulfate lose stability. Check reliability of the cali-
bration curve by running a standard with every three or four
unknown samples.
4. Correction for sample color and turbidity - Correct for the
color and turbidity present in the original sample by running
blanks from which the barium chloride is withheld.
5. Calculation
Su,fate content
8.2.7 Cyanide Analysis
These methods cover the determination of soluble and insoluble cyanides
in water by a modified Liebig titration when the cyanide level is known to
be greater than 1 mg/liter (ppm) as CN~, and by a colorimetric procedure
for lower concentrations. By an alkaline chlorination, the methods can be
used to determine cyanides amenable to chlorination. The methods do not
distinguish between CN" and CN complexes. Furthermore, they do not reveal
4 t>
the cyanates, the organo-cyanide complexes, or the cyanogen halides,
except for cyanogen chloride, which may be determined separately.
4
The cyanate complexes are decomposed when the sample is acidified in the
distillation procedure.
5
Only those organo-cyanic compounds will be revealed which hydrolyze in
water or an alkaline medium, or are decomposed by mineral acids to simple
cyanides.
6Cyanogen chloride is the most common of the cyanogen halide complexes as
it is a reaction product usually obtained when chlorinating cyanide-
containing industrial wastewater.
96
-------
The methods for the determination of cyanide require a distillation
procedure for concentrating and removing cyanides by refluxing the sample
with dilute sulfuric acid and cuprous chloride reagent. The liberated
hydrogen cyanide is collected in sodium hydroxide and its concentration
determined by either a colorimetric or titrimetric procedure.
The methods include procedures to remove the following interferences:
• Oxidizing agents, which can destroy the cyanide during manip-
ulation. If chlorine is present, add ascorbic acid as soon
as the sample is collected.
t Sulfides, which adversely affect the modified Liebig titration.
• Fatty acids, which form soaps under the alkaline titration
conditions, making the end point almost impossible to detect.
• Other interferences including substances that might contribute
color or turbidity, the cyanate or thiocyanate, and the
organic nitrogen compounds, particularly amino acids. In
most cases, the distillation procedure will remove these
interfering substances.
It is beyond the scope of these methods to describe procedures for
overcoming all the possible interferences that may be encountered. The
procedures used must sometimes be revised to meet the specific requirements
(see References 20 and 21 for alternate approaches).
8.2.7.1 Apparatus and Reagents --
• Buret, 25 ml, used for standardization of potassium cyanide
(KCN) solution only.
• Distillation apparatus. (See Figure 9.) The reaction vessel
is a 1-liter, two-neck distilling flask with 19/38 standard-
taper joints. The side neck is fitted with a joint-reduced
lower stem of 8-mm inside diameter, broken off to reach within
6 mm of the bottom of the flask. Fitted Into the other neck
is a coldfinger, separable-type condenser with 19/38 standard-
taper joints. A vacuum-type absorber with a me Jium-porosity,
heater.
97
-------
COOLING WATER
OUT
INLET TUBE
HEATER
SCREW CLAMP
/TO LOW VACUUM
3^ SOURCE
SEALS TO
BE MADE
AT THESE
POINTS \
-ABSORBER
CONDENSER
DISTILLING FLASK
Figure 9. Cyanide distillation apparatus.
98
-------
• Five-mil 111 Her microburet, used in the modified Liebig
i* i ti ct11 on •
• Spectrophotometer or filter photometer, suitable for measure-
ment at 620 nm, using 1.0-cm absorption cells. Filter
photometers and photometric practices used in this method
shall conform to ASTM Recommended Practice E60, Photometric
Methods for Chemical Analysis of Metals. Spectrophotometers
shall conform to ASTM Recommended Practice E275, for Describ-
ing and Measuring Performance of Spectrophotometers.
• Acetic acid (10 percent v/v). Mix one volume of glacial
acetic acid with nine volumes of water.
• Ascorbic acid, reagent grade.
t Bis-Pyrazolone. This reagent can be purchased commercially,
in 100 ml of 95 percent ethanol. Add 25 g of freshly dis-
tilled phenylhydrazine and reflux the mixture for at least
4 hours. Filter out the insoluble portion and wash with hot
alcohol. The product (melting point greater than 320°C
or 608°F) is stable indefinitely in dry form.
• Cadmium carbonate, powdered. This reagent is used if the
sample contains sulfides which would interfere with the
titration.
• Calcium hypochlorite solution (50 g/liter). Dissolve 5 g
calcium hypochlorite Ca(OCl)2 in 100 ml water. Store the
solution in an amber-covered glass bottle in the dark.
Prepare fresh monthly.
• Chloramine-T (1 percent w/v). Dissolve 1.0 g of the white-
colored, water-soluble grade powder in 100 ml of water.
Prepare fresh weekly.
• Cuprous chloride reagent. Transfer a weighed 20 g portion
of finely-powdered cuprous chloride (Cu2Cl2) into an 800-ml
beaker; wash twice, by decantation, with 250 ml portions of
dilute sulfuric acid (H2S04, 2 percent w/v) and then twice
with water. Add about 250 ml of water; then add concentrated
hydrochloric acid (HC1, sp gr 1.19) in 0.5 ml portions until
the salt dissolves.7 Dilute to 1 liter with water and store
in a tightly-stoppered bottle containing a few lengths of pure
upper wire or rod extending from the bottom of the mouth of
the bottle.8
~7The reagent should be clear; dark discoloration indicates the presence
of cupric salts.
If it is desired to use a reagent bottle of smaller volume it should be
kept completely filled and tightly stoppered. Refil it from
solution after each use.
99
8
-------
t Iso-octane, hexane, or chloroform (solvent preference in the
order named). The solvent is used for extraction if the sam-
ple contains fatty acids which would interfere with the
titration.
• 1-Phenyl-3-Methyl-5-Pyrazolone.
t Potassium cyanide, standard solution (1 ppm CN). Dilute 10 ml
of the stock KCN solution to 1 liter with water. Mix thor-
oughly, make a second dilution of 10 ml diluted to 100 ml,
1 ml of this solution containing 1 yg of CN". Prepare this
solution fresh daily and keep in a glass-stoppered bottle.
• Potassium cyanide, stock solution (1000 ppm CN). Dissolve
approximately 2 g potassium hydroxide (KOH) and 2.51 g potas-
sium cyanide (KCN) in 1 liter of water. (Caution - Because
KCN is highly toxic, avoid contact or inhalation.) Standardize
against the silver nitrate standard solution by the modified
Liebig titration using 25 ml of the KCN solution. Prepare
fresh weekly.
• Potassium iodide-starch test paper.
a Pyridine.
• Pyridine-pyrazolone reagent. Prepare daily.
Solution A. Add 0.25 g of l-phenyl-3-methyl-5-pyrazolone
to 50 ml of water. Heat the solution to about 60°C (14QOF)
with stirring. Cool to room temperature.
Solution B. Dissolve 0.01 g of bis-pyrazolone in 10 ml of
pyridine.
Mixed reagent. Filter solution A through coarse-grade filter
paper and collect the filtrate in a 100 ml beaker. Then pour
solution B through the same filter paper and collect the fil-
trate in the same beaker containing solution A. This mixed
reagent develops a pink color, but this does not affect the
color production with cyanide if used within 24 hours.
• Rhodanine indicator (0.2 g/liter). Dissolve 0.02 g of
(p-dimethylaminobenzylidene) rhodanine in 100 ml of acetone.
t Silver nitrate, standard solution (1 ml = 1 mg CN"). Crush
approximately 5 g silver nitrate (AgNOa) crystals and dry to
constant weight at 40°C (104°F). Dissolve 3.2647 g in water
and dilute to 1 liter with water.
t Sodium hydroxide solution (50 g/liter). Dissolve 50 g of
sodium hydroxide in water and dilute to 1 liter with water.
100
-------
concentr a°1d (5° P(?rcen* v/v)" Carefully add 1 volume of
8.2.7.2 Safety Precautions —
Because of the toxicity of cyanide, great care must be exercised in
its handling. Acidification of cyanide solutions produces HCN. All manip-
ulations should be done in the hood so that any HCN that might escape is
safely vented.
If a cyanide solution or a cyanide-containing waste is spilled on the
skin, wash it off with plenty of water. If splashed in the eyes, irrigate
with running water for 15 minutes and call a physician at once. If swal-
lowed, call a physician and give inhalations of amyl nitrite or ammonia
for 15 to 30 sec every 15 minutes for 1 hour; when the patient is conscious,
give emetics (warm salt water) until his vomit fluid is clear.
8.2.7.3 Pretreatment —
The following treatments for interference from sulfides, fatty acids,
or oxidizing agents are indicated. Care should be taken to reduce the time
for removal of interference to a minimum to avoid loss of cyanide.
Sulfides are removed by treating about 25 ml more of the alkaline
sample (pH > 11) than necessary for the cyanide determination with powdered
cadmium carbonate and mixing. Yellow cadmium sulfide precipitates in sam-
ples containing sulfides. Repeat this operation until a drop of the treated
sample solution does not darken a lead acetate test paper. Filter the
solution through a dry filter paper into a dry beaker and from the filtrate
measure the sample to be used for analysis. Avoid a large excess of cad-
mium and a long time of contact in order to minimize a loss of complexation
or occlusion of cyanide with the precipitated material.
Fatty acids are removed by extraction as suggested by Kruse and
Mellon.(47) Acidify the sample with acetic acid to pH 6.0 to 7.0. (Cau-
tion: This operation must be performed in the hood and the sample left
there until it can be made alkaline again after the extraction has been
performed.) Extract with iso-octane, hexane, or chloroform (preference in
order named) with a solvent volume equal to 20 percent of the sample volume.
One extraction is usually adequate to reduce the fatty acids below the
101
-------
interference level. Avoid multiple extractions or a long time at low pH in
order to keep loss of HCN at a minimum. When the extraction is completed,
raise the pH to greater than 12 with NaOH solution.
Oxidizing agents are treated with a reducing agent, ascorbic acid.
Test a drop of the sample with potassium iodine-starch test paper (Kl-starch
paper); a blue color indicates the need for this treatment. Add ascorbic
acid, a few crystals at a time, until a drop of sample produces no color on
the indicator paper. Then add an additional 0.5 g ascorbic acid.
If uncertain of cyanide concentration, distill a 500 ml sample, dilute
the absorption liquid and washings to 250 ml, and titrate a 200 ml aliquot.
If this titration indicates the cyanide concentration to be less than 1 mg/
liter (ppm), determine the cyanide concentration colorimetrically on the
remaining portion.
If cyanides amenable to chlorination are to be determined, add CafOCl^
dropwise to one part while agitating and maintaining the pH between 11
and 12 by the addition of NaOH solution. Test for chlorine by placing a
drop on a strip of Kl-starch paper. A distinct blue color will indicate
the presence of sufficient chlorine. Maintain the excess residual chlorine
for 1 h while agitating; if necessary, add additional Ca(OCl)2-
Add approximately 0.5 g of ascorbic acid to reduce the residual chlo-
rine. Test with Kl-starch paper; there should be no color. Again add
approximately 0.5 g of ascorbic acid to ensure the presence of excess reduc-
ing agent. Both parts are analyzed for cyanides by the following procedure.
8.2.7.4 Analysis Procedure --
If cyanides in the sample are known to be less than 1 mg/liter (ppm),
proceed as in the distillation procedure and the colorimetric method that
follows. Use a 500 ml sample in the distillation. If cyanides are known
to be more than 1 mg/liter (ppm) but less than 10 mg/liter (ppm), proceed
as in the following distillation procedure and the modified Liebig titration
using a 500 ml sample. If cyanides are suspected to be more than 10 mg/
liter (ppm), use a smaller aliquot so that no more than 5 mg of cyanide are
in the reaction vessel, and dilute to 500 ml. If cyanides are unknown,
proceed as in the preceding sections.
102
-------
1. Distillation
a) Set up the apparatus as shown in Figure 9.
b) Add 50 ml of NaOH solution to the absorber and dilute with
a measured amount of water, if necessary, to obtain an
adequate depth of liquid.
c) Introduce the 500-ml sample or diluted aliquot into the
reaction flask through the side neck. Replace the stem
and do not remove it until the distillation is entirely
completed.
d) Turn on the suction and adjust it so that approximately
one bubble of air per second enters the flask through the
air stem. It is very important that this air flow be
maintained throughout the reaction.
e) Add 50 ml of 50 percent v/v H2$04 through the air inlet.
f) Pour 10 ml of Cu2Cl2 reagent into the air inlet and
wash down with a stream of water.
g) Turn on cooling water and heat the contents of the flask
at such a rate that a slow refluxing action occurs. Too
rapid heating may release dissolved gases too fast and
force them up the air inlet.
h) Continue refluxing for 1 hour, watching both the air flow
and the reflux action. After 1 hour, turn off the heat
but maintain the air flow.
i) After cooling for 15 minutes, transfer the absorption
liquid to a separate container and carefully rinse the
absorber and its connecting tubes into this container.
(Caution. This liquid is highly toxic - avoid contact.)
This liquid may either be analyzed separately or saved
to combine with other portions. The former is recom-
mended for unfamiliar samples.
j) Refill the absorber as in 2 and repeat the reflux as in
4 and 7 to 9.
k) If the sample contains readily hydrolyzed cyanides, most
of these will be found in the absorber liquid from the
first reflux. More stable complex cyanides req uire more
time to hydrolyze. If these are present, there wi" fj
a siqnif icant yield from the second or even later refluxes,
depending on the stability of the complexes present.
1) When the optimum time of reflux for a certain type ol j sam-
ft
103
-------
2. Colorimetric Method: Cyanides less than 1 rag/liter (ppm)
a) From the standard KCN solution, prepare a series of 50 ml
standards containing from 1 to 10 yg of cyanide. Treat
the samples in accordance with steps c to f below.
b) Prepare a calibration curve by plotting the absorbances
of the standard samples against cyanide concentration in
micrograms per 50 ml of solution.
c) Judging from the calibration curve obtained in b above,
prepare as in 1 samples above containing less than 1 and
over 10 yg of cyanide to determine the limits of concen-
tration measurable with the particular photometer being
used.
d) Take an aliquot of the absorption liquid obtained in
Step 1-i above, so that its cyanide concentration falls
in the measurable range found in c above.
e) Place the aliquot of the absorption liquid in a 50 ml
beaker. Place the tip of a 50 ml buret well below the
level of the liquid and add the necessary volume of
10 percent v/v acetic acid as calculated from the for-
mula: ml of 10 percent v/v acetic acid required to
neutralize sample to pH 6.5 to 8.0.
f) Transfer immediately to a 50 ml volumetric flask; add
0.2 ml of chloramine-T solution; stopper and mix by
inversion two or three times. Allow 1 to 2 minutes for
the reaction.
g) Add 5 ml of the mixed pyridine-pyrazolone reagent; dilute
to the mark with water; stopper and mix well by inversion
and agitation. Allow 20 minutes for color development.
h) Measure the absorbance of the developed color with the
photometer at 620 nm.
i) Using the calibration curve and the formula in the fol-
lowing calculations section, determine the cyanide con-
centration in the original sample.
3. Modified Liebig Titration: Cyanides Greater than
1 mg/liter (ppm)
a) Obtain an aliquot of the absorption liquid calculated to
contain between 1 and 5 mg of cyanide and dilute to a
convenient volume for titration.
b) Add 0.5 ml of rhodanine indicator.
104
-------
rh - i Si1ver nitrate So1ution to the
change in color from canary yellow to salmon pink
d) Titrate a blank containing the same amount of alkali and
Wo LCI •
e) Record the results of the titrations and calculate the
cyanide concentration in the original sample as in the
modified Liebig titration.
4. Calculations
For the colorimetric procedure calculate the cyanide concen-
tration as follows:
CN~, mg/Hter (ppm) = (A x B)/(C x D)
where:
A = weight of cyanide, read from calibration curve, yg
B = volume of absorbing solution used in the distillation, ml
C = volume of original sample used in the distillation, ml
D = volume of aliquot of absorbing solution used, ml
For the modified Liebig titration calculate the cyanide concen-
tration as follows:
CN", ing/liter (ppm) = [(E - F) x 1000]/C x (B/D)
where:
E = volume of AgN03 solution required for titration of the
aliquot, ml
F = volume of AgN03 solution required for titration of the
blank, ml
C = volume of original sample used in the distillation, ml
B = total volume of absorbing solution used in the distilla-
tion, ml
D = volume of aliquot of absorbing solution used, ml
105
-------
Calculate the concentration of cyanides amenable to chlorination
as follows:
CN~, mg/liter (ppm) = 6 - H
where:
6 = concentration of cyanide found in the unchlorinated
portion of the sample, mg/ml (ppm)
H = concentration of cyanide found in the chlorinated
portion of the sample, mg/ml (ppm)
8.2.8 Chloride Analysis - Mercuric Nitrate Titration
This method can be used to determine all concentrations of chloride
ion in industrial water, provided intolerable interferences are absent.
It is particularly useful for analysis of boiler water, boiler feedwater,
distillate, condensate, and other relatively pure industrial waters where
low chloride concentrations must be determined accurately.
Dilute mercuric nitrate solution is added to an acidified sample in
the presence of mixed diphenylcarbazone-bromophenol blue indicator. The
end point of the titration is the formation of the blue-violet mercury
diphenylcarbazone complex.
The anions and cations generally found in industrial water offer no
interference. Zinc, lead, nickel, and ferrous and chromous ions affect
solution and end-point colors, but do not reduce the accuracy of the titra-
tion when present in concentrations up to 100 ppm. Copper is tolerable up
to 50 ppm. Titration in the presence of chromate ion requires indicator
with extra background color (alphazurine) and prior reduction for concen-
trations above 100 ppm. Ferric ion (at concentrations higher than 10 ppm)
must be reduced before titration, and sulfite ion must be oxidized. A part
of bromide ion and fluoride ion will be titrated with the chloride. Quan-
ternary ammonium salts also interfere if present in significant amounts
(1 to 2 ppm). Deep color also may interfere.
106
-------
8.2.8.1 Apparatus and Reagents —
• Microburet, 1-ml or 5-ml, with 0.01-ml graduation
intervals.
t Hydrogen peroxide (reagent grade) 30 percent w/w Hfy.
0 Hydroquinone solution (10 mg/ml). Dissolve 1 g of
purified hydroquinone in water and dilute to 100 ml.
• Mercuric nitrate, standard solution (0.025 N). Dissolve
4.2830 g mercuric nitrate (MNOs^-^O) in 50 ml water
acidified with 0.5 ml concentrated nitric acid (HN03,
sp gr 1.42). Dilute the acidified Hg(N03)2 solution
with water to 1 liter. Filter if necessary, and stand-
ardize against the standard NaCl solution, using the
procedure described in Section 8.2.8.1.9
o Mercuric nitrate, Standard Solution (0.0141N). Dissolve
2.4200 g of Hg(N03)2-H20 in 25 ml water acidified with
0.25 ml concentrated HMOs (sp gr 1-42). Dilute the acidi-
fied Hg(N03)2 solution with water to 1 liter. Filter the
solution, if necessary, and standardize against the stand-
ard NaCl solution. (Section 8.2.8.1)
• Mixed indicator solution. Dissolve 0.5 g of crystalline
diphenylcarbazone and 0.05 g of bromophenol blue powder
in 75 ml of ethyl alcohol (95% v/v), and dilute to 100 ml
with the alcohol.10 Store in a brown bottle and discard
after 6 months.11
9The end point, while sharp, can be improved somewhat for certain types
of water by adding to the titration sample several drops of a 0 05 g/
liter solution of xylene cyanole FF or alphazurine blue-green aye i
index 714). These chemicals can be mixed with the indicator
same proportions.
denatured alcohol is not suitable Methanol or isopropanol may be used
if pure ethyl alcohol is not available. ^ ^ ^
11Liquid indicator generally deteriorates to the P01" tenlperature
end-point color after 12 to 18 months of storage y ^^ gfi
(aboJe 37.8*C (100*F)) and exposure to ft$^ngreJlents Is stable
life. A dry powder mixture of the two ™^c (capsuie form) and the
for much longer periods. Both the P™°*
liquid indicator are available commercially.
107
-------
t Nitric acid (0.3% v/v). Mix volumes of concentrated nitric
acid (HN03 sp gr 1.42) with 997 volumes of water.
• pH indicating paper. Long-range type, covering a pH range
1 to 11.
• Sodium chloride, standard solution (0.025 N). Dry several
grams of sodium chloride (NaCl) for 1 hour at 600°C. Dis-
solve 1.4613 ±0.0002 g of the dry salt in water and dilute
to 1 liter at 20°C in a volumetric flask.
• Sodium hydroxide solution (10 g/liter). Dissolve 10 g of
sodium hydroxide (NaOH) in water and dilute to 1 liter.
8.2.8.2 Analysis Procedure —
Use a volume of sample that will contain not more than 20 mg of
chloride ion, diluting the sample with water to approximately 50 ml volume
if necessary. If the volume of sample contains less than 2.5 mg chloride
ion, make the final titration with 0.0141 N Hg(N03)2 solution, using a
1 or 5-ml microburet. In this latter case, determine an indicator blank
on 50 ml of chloride-free water, applying the same procedure followed for
the sample. If the sample contains less than 0.1 ppm chloride, concentrate
an appropriate volume of sample to 50 ml.
Add 5 to 10 drops of mixed indicator solution and shake or swirl the
flask. If a blue-violet or red color develops, add 0.3 percent HNO- drop-
O
wise until the color changes to yellow. Add 1 ml of excess acid. If a
yellow or orange color forms immediately on addition of the mixed indicator,
add NaOH solution dropwise until the color changes to yellow, then add 1 ml
excess of acid.^
12
"•The prescribed acidification provides a satisfactory pH range of 3.0 to
3.5. Acidified samples on which electrometric pH measurements have been
made shall not be used for chloride determinations, because the use of
the calomel reference electrode may introduce error due to chloride con-
tamination. For precise pH adjustment of samples having a low chloride
contamination, instrumental measurements may be made on one sample
aliquot to determine treatment needed for another to be used for the
chloride test.
108
-------
Titrate the solution with 0.025 N Hg(N03)2 solution until a blue-
violet color, as viewed by transmitted light, persists throughout the solu-
tion.13 Record the mill 11 Hers of Hg(N03)2 solution added.
If chromate ion is present in the absence of iron and in concentration
less than 100 ppm, use the alphazurine modified mixed indicator and acidify
the sample as described above but to pH 3 as indicated by pH indicating
paper. Titrate the solution as described above, but to an olive-purple end
point.
If chromate ion is present in the absence of iron and in concentration
greater than 100 ppm, add 2 ml fresh hydroquinone solution and titrate the
solution as described above.
If ferric ion is present in the absence of chromate ion, use a sample
of such volume as to contain no more than 2.5 mg ferric ion or ferric ion
plus chromate ion. Add 2 ml fresh hydroquinone solution, and acidify the
sample and titrate the solution as described above.
If sulfate ion is present, add 0.5 ml of H202 to 50 ml of the sample
in the Erlenmeyer flask and mix for 1 min. Then, acidify the sample and
titrate the solutions as described above.
8.2.8.3 Calculation —
Calculate the chloride ion concentration, in milligrams per liter, in
the original sample as follows:
Chloride, mg/liter (ppm) = [(V] - V2) x N x 35, 500]/S
13The use of indicator modifications and the Presence of heavy meta 1 uns
can change solution colors without affecting the accuracy of the deter
mination9 For example, solutions containing ajpjazurine my be b^g
blue when neutral, grayish P^P16"^^! SolutloSs containing about
and blue violet at the chloride end £"?*' J0^ pULle when neutral,
100 ppm nickel ion and ™™\*l**™§?^ point. Vn applying this
green when acid, and gray at the chlonae en*v require modified indi-
method to samples that contain colored ions or™"^ himself w1th the
cators, it is recommended that the operaw solutions prepared
specific color changes involved by experimenting
as standards for comparison of color effects.
109
-------
where:
V, = milliliters of standard Hg(N03)2 solution required for titration
1 of the sample
V9 = milliliters of standard Hg(NOj? solution required for titration
d of the blank J *
H = normality of the Hg(N03)2 solution
S = milliliters of sample used.
8.2.9 Chloride-Silver Nitrate Potentiometric Method
Chloride is determined by potentiometric titration with silver nitrate
solution using a glass and silver-silver chloride electrode system. During
titration an electronic voltmeter is used to detect the change in potential
between the two electrodes. The end point of the titration is that instru-
ment reading at which the greatest change in voltage has occurred for a
small and constant increment of silver nitrate added.
Iodide and bromide also are titrated as chloride. Ferricyanide causes
high results and must be removed. Chromate and dichromate interfere and
should be reduced to the chromic state or removed. Ferric iron interferes
if present in an amount substantially higher than the amount of chloride.
Chromic ion, ferrous iron, and phosphate do not interfere.
Grossly contaminated samples usually require pretreatment. Where
contamination is minor, some contaminants can be destroyed simply by the
addition of nitric acid.
8.2.9.1 Apparatus and Reagents —
t Glass and silver-silver chloride electrodes. The latter is
a silver electrode coated with silver chloride and may be
prepared in the laboratory if desired, but can be purchased
for use with specified instruments. Instructions on the
use and care of the electrodes are supplied by the
manufacturer.
t Electronic voltmeter, to measure the potential difference
between the electrodes. Many laboratories find it possi-
ble to convert a pH meter to this use by substituting the
appropriate electrode.
110
-------
Mechanical stirrer, with plastic-coated or glass impeller.
Standard sodium chloride solution, 0.014N Dissolve
8.243 g Nad, dried at 105°C (221°F) n dist led water
and dilute to exactly 500 ml/ Dilute to 50 Om of his
solution to exactly 1000 ml. The final solution conta ns
0.500 mg Cl per 1.00 ml (500 ppm).
Nitric acid, concentrated.
Silver nitrate titrant, 0.014N. Dissolve 2.38 g AgNOo
in distilled water and dilute to 1000 ml. Standardize
this solution by titrating exactly 10.0 ml standard Nad
solution using the standardization procedure below. Cal-
culate the normality of the AgN03 solution as follows:
Normality of AgNO, = 10-° x °-0141
*5 V
V = ml AgN03
• Special reagents for pretreatment:
1) Sulfuric acid, 1+1.
2) Hydrogen peroxide, 30 percent.
3) Sodium hydroxide, 1 N.
8.2.9.2 Standardization Procedure —
a) Inasmuch as the various instruments that can be used in
this determination differ in operating details, the manu-
facturer's instructions should be followed. Necessary
mechanical adjustments should be made. Then, after
allowing sufficient time for warm-up (10 min), the inter-
nal electrical components are balanced to give an instru-
ment setting of 0 mV or, if a pH meter is used, a pH
reading of 7.0.
b) Place 10.0 ml standard NaCl solution in a 250-rnl beaker,
dilute to about 100 ml, and add 2.0 ml cone HMOs-
Immerse the stirrer and the electrodes in the solution.
c) Set the instrument to the desired range of millivolts
or pH units. Start the stirrer.
d) Add AgN03 titrant, recording the scale rea£ng after each
addition At the start, large increments of AgN03 can be
111
-------
added; then, as the end point of the reaction is approached,
smaller and equal increments (0.1 or 0.2 ml) should be added
at longer intervals, so the exact end point can be deter-
mined. Determine the volume of AgNOs used at the point at
which there is the greatest change in instrument reading per
unit addition of AgNOs-
e) A differential titration curve should be plotted if the
exact end point cannot be determined by inspection of the
data. Plot the change in instrument reading for equal incre-
ments of AgN03 against the volume of AgNOa added, using the
average of the buret readings before and after each addition.
The procedure is illustrated in Figure 10. Calculate the
normality of the AgNOs using the following equation:
vci
where:
NCI = Normality of Nad" solution
Vp, = Volume of Nad aliquot titrated
V. = Volume of AgN03 used to titrate standard NaCl
Ag solution
8.2.9.3 Analysis Procedure --
a) Pi pet exactly 100.0 ml of sample, or an aliquot containing
not more than 10 mg chloride, into a 250-ml beaker. In
the absence of interfering substances, proceed with Step c
below.
b) In the presence of organic compounds, sulfite, or other
interferences (such as large amounts of ferric iron or
substantial amounts of cyanide or sulfide), acidify the
sample with H2S04, using litmus paper. Boil for 5 min-
utes to remove volatile compounds. Add more I^SO^, if
necessary, to keep the solution acidic. Add 3 ml ^02
and boil for 15 minutes adding chloride-free distilled
water to keep the volume above 50 ml. Dilute to 100 ml,
add NaOH solution dropwise until alkaline to litmus, then
10 drops in excess. Boil for 5 minutes, filter into a
250-ml beaker, and wash the precipitate and paper several
times with hot distilled water.
112
-------
60
50
40
30
20
10
25 26
VOLUME OF AgNO- SOLUTION - ml
27
EXPERIMENTAL DATA PLOTTED ABOVE
VOLUME, ml 23.50 24.50 25.00 25.25 25.50 25.75 26.00 26.50 27.50
CHANGE, mV/ml 18 36 48 52 52 40 32 18
Figure 10. Example of differential titration curve
(end point is 25.5 ml).
c) Add concentrated HMOs dropwise until acidic to litmus
paper, then 2.0 ml in excess. Cool and dilute to 100 ml
if necessary. Immerse the stirrer and the electrodes in
the sample and start the stirrer. After making the
necessary adjustments of the instrument according to the
manufacturer's instructions, set the selector switch to
the appropriate setting for measuring the difference of
potential between the electrodes.
d) Complete the determination by titrating according to Step d,
Section 8.2.9.1. If an end-point reading has been estab-
lished from previous determinations for similar samples
and conditions, this predetermined end point can be used.
For the most accurate work, a blank titration should be
made by carrying chloride-free distilled water through
the procedure.
113
-------
8.2.9.4 Calculation —
n (A - B x N x 35.45 x 1000)
ppm Cl = -* jj L
where A = ml AgN03, B = ml blank, N = normality of titrant, and D = ml
sample.
8.2.10 Chloride-Col orimetric Method
This method can be applied to waters containing chloride ion in con-
centrations from 0.02 to 10 ppm. It is particularly useful for analysis of
boiler water, boiler feedwaters, distillate, condensate, and other relatively
pure industrial waters where low chloride concentrations must be determined
accurately.
Solutions of ferric ammonium sulfate and mercuric thiocyanate are
added to the sample. The chloride ion reacts with the mercuric thiocyanate
to produce thiocyanate ion which in turn combines with ferric ion to form
red ferric thiocyanate. The intensity of the color, which is proportional
to the concentration of the chloride ion, is measured photometrically at a
wavelength of 463 nm, or by visual comparison with standard solutions.
Bromides, iodides, cyanides, thiosulfates, and nitrates interfere in
this method. Color, if present in the sample, will interfere with visual
comparison and, depending on its spectral absorbance, may interfere with
the photometric measurement.
8.2.10.1 Apparatus and Reagents --
• Nessler tubes or photometer. A set of 50-ml matched Nessler
tubes or a photometer suitable for measurements at a wave-
length of 463 nm may be used for evaluating the intensity
of the color produced. The optimum range of concentrations
for some typical methods of color measurement is shown in
Table 11.
114
-------
TABLE 11. CONCENTRATION RANGES FOR COLOR MEASUREMENT
Method of Color Measurement
—————___________
Nessler tubes, 300 mm
Filter photometer, blue filter,
425 run, 2.3-cm cell
Spectrophotometer, 463 nm:
1.0-cm cell
10-cm cell
Optimum Range,
mg of chloride
ion per 25-ml
water sample
— " •
0.005 to 0.25
0.003 to 0.25
0.005 to 0.25
0.0005 to 0.05
Ferric ammonium sulfate solution (5% w/v). Dissolve 5.0 g
of ferrous ammonium sulfate (Fe(NH4)2(S04J2-6H20) in
20 ml of water. Add 38 ml concentrated nitric acid (HN03
sp gr 1.42) and boil to oxidize the iron and remove the
oxides of nitrogen. Dilute to 100 ml with halide-free
water.
Mercuric thiocyanate, methanol solution (0.3% w/v). Dis-
solve 0.30 g mercuric thiocyanate (Hg(CNS)2) in 100 ml
methanol. Store in amber bottles. Allow to stand for
at least 24 hours before using. (Caution: Mercuric salts
are very poisonous. Due precautions should be observed
when using this material.) Do not use if more than 4 weeks
old.14
Sodium chloride, standard solution (10 ppm). Dry several
grams of sodium chloride (Nad) for 1 hour at 600°C
(1112°F). Prepare a stock solution by dissolving exactly
1.649 g of the dry salt in water and dilute to 1 liter.
Prepare the standard solution as needed by diluting 10 ml
of the stock solution to 1 liter with halide-free water.
The resulting standard contains 10 mg of chloride ion per
liter.
14A slight precipitate may form and settle °«* ^
be taken so this precipitate is not resuspended when
Only the clear, supernatant liquid must be used.
he oenf
the reagent.
115
-------
8.2.10.2 Analysis Procedure --
1. Sampling
Since chloride ion is a very common contaminant, extreme care must be
exercised in the collection and processing of the sample. Soak all new
glassware in hot nitric acid (5% v/v) for several hours. To be certain
that new glassware is conditioned for the test, run a chloride determina-
tion on halide-free water. After the run rinse the glassware thoroughly.
Soak the glassware in halide-free water between tests. Discard all glass-
ware that appears etched or scratched.
2. Calibration
Prepare a series of reference standards by diluting suitable volumes
of the standard chloride solution with halide-free water. The series should
cover the optimum range of the selected method of color measurement described
in Table 10. The temperature of the solutions used for calibration must be
the same as that of the sample to be tested.
Treat each reference standard as described in the procedures below.
Prepare a calibration curve by plotting the readings on the photometer
versus the concentration of chlorides. When the scale of the photometer
reads directly in absorbance, plot the curve on rectilinear paper. When
the scale reads in transmittance, it is convenient to plot the results on
semilog paper, using the single cycle log axis to plot transmittance and
the linear axis to plot the concentrations.
3. Methodology
Transfer 25 ml of sample to a glass-stoppered cylinder and add succes-
sively 5 ml ferric ammonium sulfate solution and 2.5 ml mercuric thiocyanate
solution. Mix thoroughly and allow to stand for 10 minutes.
Measure the intensity of the color formed either by comparison with
suitable reference standards in Nessler tubes or by a photometer chosen to
cover the desired range as indicated in Table 11. Adjust the zero setting
of the photometer by using 25 ml halide-free water.
116
-------
4. Calculation
Determine the ppm concentration of chloride ion in the sample either
by direct visual comparison with reference standards or by use of a suitable
calibration curve.
Chloride, mg/liter (ppm) = 40A
where A = milligrams of chloride ion in the sample.
8.2.11 Nitrate-Brucine Method
The reaction between nitrate and brucine produces a yellow color which
can be used for the colorimetric estimation of nitrate. The intensity of
the color is measured at 410 nm. The reaction rate between brucine and
nitrate ion is affected significantly by the amount of heat generated dur-
ing the test. Thus the procedure seeks heat control by reagent addition
sequence and incubation of the reaction mixture for a precise interval of
time at a known temperature. Acid concentration and reaction time have
been selected to yield optimum development and stability of color. The
method works well in waters of salinities varying from that of fresh water
to that of sea water. The method is recommended for the approximate range
of 0.1 to 2 ppm N03~ nitrogen.
All strong oxidizing or reducing agents interfere. The interference
by residual chlorine can be eliminated by the addition of sodium arsenite,
provided that the residual chlorine does not exceed 5 ppm. A slight excess
of sodium arsenite will not affect the determination. Ferrous and ferric
iron and quadrivalent manganese give slight positive interferences, but in
concentrations less than 1 ppm these are negligible. The interference due
to nitrite up to 0.5 ppm NO" nitrogen is eliminated by the use of sulfanilic
acid. Chloride interference is masked by the addition of excess NaCl.
High concentrations of organic matter such as in undiluted raw waste-
water will usually interfere.
117
-------
8.2.11.1 Apparatus and Reagents --
t Colorimetric equipment: One of the following is required:
Spectrophotometer, for use at 410 nm providing a light
path of 1 cm, or filter photometer, providing a light path
of 1 cm and equipped with a violet filter having maximum
transmittance between 400 and 425 nm.
t Safety pi pet.
• Wire racks, to hold tubes in which samples are to be incu-
bated (Van Waters and Rogers No. 60935 or equivalent).
t Stirred boiling water bath, with heating facility suffi-
cient to maintain a temperature of at least 95°C (203°F)
when cooled samples are introduced.
t Reaction tubes. Hard-glass test tubes, of approximate
dimensions 2.5 x 15 cm, in which reaction is performed.
[The 1-cm, colorimeter tubes (Van Waters and Rogers
No. 22366) used in conjunction with the Bausch & Lomb
Spectronic 20 or equivalent are convenient, since their
use avoids the necessity for a transfer, following reac-
tion, to determine transmittance.]
• Stock nitrate solution. Dissolve 721.8 mg anhydrous
potassium nitrate, KNOj, and dilute to 1000 ml with dis-
tilled water. This solution contains 100 mg/1 N,
(100 ppm).
• Standard nitrate solution. Dilute 10.00 ml stock nitrate
solution to 1000 ml with distilled water; 1.00 ml =
1.00 jjg N (1 ppm). Prepare immediately prior to using.
t Sodium arsenite solution (0.5% w/v). Dissolve 5.0 g
NaAs02 and dilute to 1 liter with distilled water.
(Caution: Toxic - take care to avoid ingestion.)
t Brucine-sulfanilic acid solution: Dissolve 1 g brucine
sulfate and 0.1 g sulfanilic acid in approximately 70 ml
hot distilled water. Add 3 ml cone HC1, cool, and make
up to 100 ml. This solution is stable for several months.
The pink color that develops slowly does not affect its
usefulness. (Caution: Brucine is toxic - take care to
avoid ingestion.)
• Sulfuric acid solution. Carefully add 500 ml cone ^$04
to 125 ml distilled water. Cool to room temperature
before using and keep tightly stoppered to prevent absorp-
tion of atmospheric moisture.
118
-------
Nan °n (30% W/V>' D1ssolve 300 g
Nad and dilute to 1000 ml with distilled water.
8.2.11.2 Analysis Procedure —
1. Preparation of nitrate standards - Prepare nitrate standards in
the range 0.1-1 ppm N by diluting 1.00, 2.00, 4.00, 7.00, and 10.0 ml
standard nitrate solution to 10.0 ml with distilled water.
2. Pretreatment of sample - If the sample contains residual chlorine,
remove by adding one drop (0.05 ml) sodium arsenite solution for each
0.10 mg Cl and mix. Add one drop in excess to a 50-ml portion.
3. Color development - Set up the required number of reaction tubes
in the wire rack, spacing them so each tube is surrounded by empty spaces.
Include a reaction tube for a reagent blank and reaction tubes for as many
standards as desired. To each tube add 10.0 ml sample or an aliquot diluted
to 10 ml so that the sample volume taken for analysis contains between 0.1
and 8 ug NO-" nitrogen. Place the rack in a cool water bath and add 2 ml
NaCl solution. Mix thoroughly, swirling by hand, and add 10 ml H^SO. solu-
tion. In no case use a "Vortex" mixer, since this type of mixing produces
inconsistent results in the analysis. Mix again thoroughly by swirling and
allow to cool. At this point, if any turbidity or color is present or if
optically unmatched colorimeter tubes are being used as reaction tubes, dry
the tubes and read a "sample blank" value against the reagent blank tube at
410 nm. Replace the rack of tubes in the cool water bath and add 0.5 ml
brucine-sulfanilic acid reagent. Swirl the tubes to mix thoroughly and then
place the rack of tubes in a well-stirred boiling water bath that maintains
a temperature of not less than 95°C (203°F). After exactly 20 minutes,
remove the samples and immerse them in a cold water bath. When thermal
equilibrium is reached (at approximately room temperature), dry off the
tubes with tissue and read the standards and samples against the reagent
blank at 410 nm in the spectrophotometer. Check the technique and the con-
stancy of reaction condition by running at least two standards with each
batch of samples.
119
-------
To obtain a standard curve, subtract the "sample blanks" from the final
absorbance readings and plot the resultant absorbance against ppm N03 nitro-
gen. Correct the absorbance readings of the samples by subtracting their
sample blank values from their final absorbance values. Read the concentra-
tions of NCL~ nitrogen directly from the standard curve.
4. Calculation
yg N0_" nitrogen
ppm nitrate N = m1 Satnp1e
ppm NO- = ppm nitrate N x 4.43
8.2.12 Nitrate-Phenoldisulfonic Acid Method
The yellow color produced by the reaction between nitrate and phenol-
disulfonic acid obeys Beer's law up to at least 12 ppm N at a wavelength
of 480 nm when a light path of 1 cm is used. At a wavelength of 410 nm, the
point of maximum absorption, determinations can be made up to 2 ppm with
the same cell path.
As even small concentrations of chloride result in nitrate losses using
this method, it is important that the chloride content be reduced to a mini-
mum, preferably below 10 ppm. However, the silver sulfate used for this
purpose presents problems with some water samples because of the incomplete
precipitation of silver ion, which produces an off color or turbidity when
the final color is developed. The preferred alkali for color development
in the final stage of the determination is ammonium hydroxide, particularly
where chloride removal must be practiced on the sample. Potassium hydroxide
should be used only if ammonia fumes must be reduced to a minimum in the
laboratory atmosphere (for example, when trace amounts of ammonia nitrogen
are being determined concurrently). A faint tinge of brown is imparted by
potassium hydroxide to the final color when a silver compound has been pre-
viously applied for chloride precipitation. Nitrite levels in excess of
0.2 ppm N erratically increase the apparent nitrate concentration. Colored
ions and materials physically modifying the color system should be absent.
120
-------
In the absence of interference, the phenoldisulfonic acid method is
sensitive to 1 yg nitrate nitrogen, which represents 10 ppb in a 100-ml
sample.
8.2.12.1 Apparatus and Reagents —
• Colorimetric Equipment - One of the following is required-
spectrophotometer, for use at 410 nm, providing a tight
path of 1 cm or longer; filter photometer, providing a
light path of 1 cm or longer and equipped with a violet
filter having a maximum transmittance near 410 nm; Nessler
tubes, matched,,50- or 100-ml.
Prepare all reagents from chemicals which are white in color and store
all solutions in pyrex containers.
• Standard silver sulfate solution. Dissolve 4.40 g Ag2S04,
free from nitrate, in distilled water and dilute to 1.0 liter;
1.00 ml is equivalent to 1.00 mg Cl.
• Phenoldisulfonic acid reagent. Dissolve 25 g pure white
phenol in 150 ml cone H2S04. Add 75 ml fuming ^$04
(15% free 803), stir well, and heat for 2 hours on a hot
water bath.
• Ammonium hydroxide. If this cannot be used, prepare
12 N potassium hydroxide solution by dissolving 673 g KOH
in distilled water and diluting to 1 liter.
• EDTA reagent. Rub 50 g disodium ethylenediamine tetra-
acetate di hydrate, also called (ethylenedinitrilo)-
tetraacetic acid sodium salt, with 20 ml distilled water
to form a thoroughly wetted paste. Add 60 ml cone NH4OH
and mix well to dissolve the paste.
• Stock nitrate solution. Dissolve 721.8 mg anhydrous potas-
sium nitrate, KN03, and dilute to 1000 ml with distilled
water. This solution contains 100 ppm N.
• Standard nitrate solution. Evaporate 50.0 ml stock nitrate
solution to dryness on a steam or hot water bath; dissolve
the residue by rubbing with 2.0 ml pheno disulfonic acid
reagent, and dilute to 500 ml with dTStilled water,
1.00 ml = 10.0 yg N = 44.3 yg N03.
121
-------
• Reagents for Treatment of Unusual Interference:
1) Aluminum hydroxide suspension. Dissolve 125 g aluminum
potassium or ammonium sulfate, A1K(S04)2--12 H20 or
A1NH4(S04)2.12 H20 in 1 liter distilled water. Warm
to 60°C (140°F) and add 55 ml cone NfyOH slowly, with
stirring. After permitting the mixture to stand about
1 hour, transfer to a large bottle and wash the precipi-
tate by successive additions (with thorough mixing)
and decantations of distilled water, until free from
ammonia, chloride, nitrate, and nitrite. Finally, after
settling, decant off as much clear liquid as possible,
leaving only the concentrated suspension.
2) Sulfuric Acid, 1 N.
3) Potassium permanganate, 0.1 N. Dissolve 316 mg KMn04
in distilled water and dilute to 100 ml.
4) Dilute hydrogen peroxide solution. Dilute 10 ml of
30% hydrogen peroxide (low in nitrate) to 100 ml with
distilled water.
5) Sodium hydroxide, 1 N.
8.2.12.2 Analysis Procedure —
a. Color Removal - If the sample is colored, decolorize by adding
3 ml aluminum hydroxide suspension to 150 ml of sample. Stir very thor-
oughly and allow to stand for a few minutes, then filter, discarding the
first portion of the filtrate.
b. Nitrite Conversion - To 100 ml of sample add 1 ml of HLSO. and stir.
Add dropwise, with stirring, either KMn04 or H202 solution. Let the treated
sample stand for 15 minutes to complete the conversion of nitrite to nitrate.
(A faint pink color persists for at least 15 minutes when sufficient KMnO,
4
is used.)
c. Chloride Removal - Determine the chloride content of the water and
treat 100 ml of sample with an equivalent amount of standard silver sulfate
solution. Remove the precipitated chloride either by centrifugation or by
filtration, coagulating the silver chloride by heat if necessary. (Excellent
removal of silver chloride can be achieved by allowing the treated sample
to stand overnight at laboratory temperature away from strong light. This
approach applies to samples free of contamination by nitrifying organism.)
122
-------
d. Evaporation and Color Development - Neutralize the clarified
sample to approximately pH 7, transfer to a casserole, and evaporate to
dryness over a hot water bath. Using a glass rod, rub the residue thor-
oughly with 2.0 ml phenoldisulfonic acid reagent to ensure dissolution of
all solids. If necessary, heat mildly on the hot water bath a short time
to dissolve the entire residue. Dilute with 20 ml distilled water and add,
with stirring, about 6 to 7 ml NH4OH-or about 5 to 6 ml KOH until maximum
color is developed. Remove any resulting flocculent hydroxides by passing
through a filter paper or filtering crucible, or add the EDTA reagent drop-
wise, with stirring, until the turbidity redissolves. Transfer the filtrate
or clear solution to a 50- or 100-ml volumetric flask or Nessler tube, dilute
to the mark, and mix.
e. Photometric Measurement - Make photometric readings in cells with
a 1-cm or longer light path at a wavelength of 410 nm, or with violet fil-
ters exhibiting maximum transmittance in the range from 400 to 425 nm. If
available, use a 5-cm light path for measurements in the nitrogen interval
from 5 to 50 yg, and a 1-cm light path in a proportionate range. Make read-
ings against a blank prepared from the same volumes of phenoldisulfonic acid
reagent and NhLOH or KOH as used for the samples.
f. Visual Comparison - In the case of 50-ml Nessler tubes, use the
following volumes of standard nitrate solution: 0, 0.1, 0.3, 0.5, 0.7,
1.0, 1.5, 2.0, 3.5, 6.0, 10, 15, 20, and 30 ml. Where it is more convenient
to use a total volume of 100 ml, double the volumes of standard solution.
To each of these standards add 2.0 ml phenoldisulfonic acid reagent and
the same volume of the same alkali as is used in preparation of the sample.
These standards can be kept several weeks without deterioration.
g. Calculation
uq nitrate N
ppm nitrate N = ml san)pie
ppm N03 = ppm nitrate N x 4.43
123
-------
8.2.13 Antimony Analysis
Pentavalent Sb in aqueous HC1 solution reacts with Rhodamine B to
form a colored complex extractable with organic solvents. Intensity of
extracted color is measured spectrophotometrically at 565 nm.
The following reagents are required:
• Hydrochloric acid solution. 6 N," dilute concentrated acid
with H20 (1 + 1).
• Dilute phosphoric acid. 3 N, dilute 70 ml H-PCL (85 percent)
to 1 L with H20.
• Rhodamine B solution. 0.02 percent w/v in H20.
• Antimony standard solutions. (1) Stock solution = 100 ppm.
Dissolve 0.1000 g pure Sb in 25 ml H2$04 with heat; cool,
and cautiously dilute to 1 L with H20. (2) Working solu-
tion = 1 ppm. Dilute 2.0 ml stock solution to 200 ml with
H20.
Cool hydrochloric acid, phosphoric acid, antimony reagents and approxi-
mately 100 ml benzene, and eight 125 ml separators with Teflon stopcocks
in refrigerator before use; maintain temperature of 5 to 10°C (41 to 50°F)
during extraction and color development. Work in subdued light.
8.2.13.1 Analysis Procedure —
Digest sample using H2S04-HN03 or HC1-HN03 (5 ml of each acid). Oxidiz-
ing conditions must be maintained.
Transfer digest or aliquot to 125 ml glass stoppered Erlenmeyer; add
enough H2S04 to make total of 5 ml H2S04> and evaporate until white fumes
of S03 are driven off. Cool flask, add 10 drops 70 percent HC104, and again
evaporate to white fumes.15 Cool digest in ice bath >30 minutes, then slowly
add 5 ml precooled 6N HC1 by pipet. Let stand in ice bath 15 minutes, then
add 8 ml precooled 3N H3P04. (Until color is extracted into benzene, perform
subsequent operations as quickly as possible. Color is stable in benzene
for several hours.) Immediately add 5 ml precooled Rhodamine B solution,
stopper, and shake vigorously. Transfer to precooled 125 ml separator.
15Under no circumstances should the sample be evaporated to dryness once the
HCO* has been added. Perform HC&04 digestion in a hood especially set
aside for HQO/i that is free of organic material in the vent system and
set aside for HC&04 titrations.
124
-------
Pi p. t 10 .1 precooled benzene Into separator, shake vigorously , minute
an d,scar aqueous layer. Transfer benzene layer (red if Sb ,s presen )
int. test tube and let H20 settle. R,nse 1 cm ce,, Wlth extract, fin J,l
and read at 565 m against benzene blank taken through entire detection
Refer readings to standard curve.
8.2.13.2 Standardization Procedure —
Pipet 0, 2, 4, 6, 8, and 10 ml Sb working standard solution into
125-ml glass stoppered Erlenmeyers; add 5 ml H2S04 to each, and proceed
as in determination. Plot absorbance against ug Sb.
8.2.14 Selenium Analysis
Selenium has a toxic effect on man and animals comparable with that
of arsenic, giving rise to similar symptoms. Selenium has also been sus-
pected of causing dental caries in man, and has been cited as a potential
carcinogenic agent.
The selenium concentration of most drinking waters falls below
10 yg/1. Concentrations exceeding 500 yg/1 are rare and limited to seep-
age from seleniferous soils. The sudden appearance of selenium in a water
supply might indicate industrial pollution. Little is known regarding the
valence state of selenium in natural waters, but because selenate and sele-
nite are both found in soils, it is reasonable to expect that both may be
present in seleniferous water. Water contaminated with wastes may contain
selenium in any of its four valence states. Many organic compounds of
selenium are known.
This procedure is based on the evolution of Se from the sample as its
hydride. The hydride is then passed into a H2-Ar flame of an AAS where the
Se is quantified.
8.2.14.1 Apparatus and Reagents —
, Flow meter, capable of measuring 1 1/mln, such as that used
for auxiliary argon.
. Medicine dropper, capable of delivering 1.5 ml, fitted into
a size "0" rubber stopper.
. Reaction flask, a pear-shaped vessel with side arm and 50 ml
capacity, both arms having 14/20 joint.
125
-------
• Special gas inlet-outlet tube, constructed from a micro
coldfinger condenser by cutting off the portion below
the 14/20 ground glass joint.
• Magnetic stirrer, strong enough to homogenize the zinc
slurry described in paragraph 8.2.14.2 below.
• Drying tube, 100-mm-long polyethylene tube filled with
glass wool to keep particulate matter out of the burner.
t Stannous chloride solution. Dissolve 100 g SnCl2 in
100 ml cone HC1.
• Zinc slurry. Add 50 g zinc metal dust (200 mesh.) to 100 ml
deionized distilled water.
• Diluent. Add 100 ml 18N H2S04 and 400 ml cone HC1 to
400 ml deionized distilled water in a 1-1 volumetric flask
and bring to volume with deionized distilled water.
• Stock selenium solutions. Dissolve 1.000 g selenium in
5 ml cone HMOs. Warm until the reaction is complete and
cautiously evaporate just to dryness. Dilute to 1,000 ml
with high purity water. This solution contains 1 mg
Se/ml.
t Intermediate selenium solution. Pipet 1 ml of the stock
selenium solution into a 100 ml volumetric flask and bring
to volume with high purity water containing 1.5 ml of con-
centrated HN03/£. This solution will contain 10 yg of Se
in each mi Hi liter.
• Standard selenium solution. Pipet 10 ml intermediate
selenium solution into a 100-ml volumetric flask and bring
to volume with high purity water containing 1.5 ml cone
HN03/j2. This solution contains 1 yg of Se in each milliter.
• Perchloric acid, 70 to 72% HC104-
8.2.14.2 Procedure ~
1. Apparatus setup. (Refer to Figure 11.) Connect the apparatus
with the burner of the spectrophotometer as shown in Fig-
ure 11. Connect the outlet of the reaction vessel to the
auxiliary oxidant input of the burner with Tygon tubing.
Connect the inlet of the reaction vessel to the outlet side
of the auxiliary oxidant (argon supply) control valve of the
instrument.
126
-------
ARGON
FLOW
METER
MEDICINE
DROPPER IN
SIZE "0"
RUBBER
STOPPER
DRYING
TUBE
(AUXILIARY AIR)
- ARGON
(NEBULIZER AIR)
HYDROGEN
(FUEL)
Figure 11. Schematic arrangement of equipment for determination
of arsenic and selenium (reference 48).
2. Instrument operation. Because of differences between makes
and models of satisfactory atomic absorption spectrophotom-
eters, it is not possible to formulate instructions appli-
cable to every instrument. In general, proceed as follows:
a) Install a hollow cathode lamp of the desired metal in
the instrument, set the wavelength at 196.0 mm and align
the lamp in accordance with the manufacturer's
instructions.
b) Set the slit width according to the manufacturer's sug-
gested setting for the element being measured.
c) Turn on the instrument and apply the amount of current
suggested by the manufacturer to the hollow cathode
lamp.
d) Allow the instrument to warm up until the energy source
stabilizes; this process usually requires 10 to 20 min.
e) Install a Boling burner head.
f) Turn on the argon and adjust to a flow rate of about
8 1/min, with the auxiliary argon flow at 1 1/irnn.
g) Turn on the hydrogen, adjust to a flow rate of about
7 1/min and ignite the flame. The flame 1S essentially
colorless. To determine whether the flame is ignited,
pass the hand about 30 cm (1 ft) above the burner to
detect the heat emitted.
127
-------
h) Atomize the standard solution (1.00 ml = 1.00 yg) of
the desired metal, and adjust the burner both sideways
and vertically in the light path until maximum response
is obtained.
i) The instrument is now ready to run standards and samples
by the arrangement of Figure 11.
3. Sample Preparation. To a 50-ml volumetric flask, add 25 ml
sample, 20 ml cone HC1, and 5 ml 18N H2$04.
4. Preparation of standards. Transfer 0, 0.5, 1.0, 1.5, and
2.0 ml standard selenium solution to 100-ml volumetric
flasks and bring to volume with diluent to obtain concen-
trations of 0, 5, 10, 15, and 20 yg/1 selenium.
5. Treatment of samples and standards.
a) Transfer a 25-ml portion of sample or standard to the
reaction vessel.
b) Add 0.5 ml SnCl2 solution. Allow at least 10 min for
the metal to be reduced to its lowest oxidation state.
Attach the reaction vessel to the special gas inlet-
outlet glassware.
c) Fill the medicine dropper with 1.50 ml zinc slurry that
has been kept in suspension with the magnetic stirrer.
d) Firmly insert the stopper containing the medicine dropper
into the side neck of the reaction vessel. Squeeze the
bulb to introduce the zinc slurry into the sample or
standard. The metal hydride will produce a peak almost
immediately. When the recorder pen returns part way to
the base line, remove the reaction vessel.
e) Record the peak height or if an integrator is available
the peak area.
8.2.14.3 Calculations —
1) Draw a standard curve by plotting peak heights or areas of
standards versus concentration of standards.
2) Measure the peak heights of the samples and read the concen-
tration from the curve.
3) Multiply these concentrations by two because the sample was
diluted 1 + 1 with acid.
128
-------
8.2.15 Phosphate Analysis
The ammonia phosphomolybdovanate colorimetric method used in this
analysis of phosphate samples was adopted from standard operating proce-
dures of the U.S. Industrial Chemical Company and the Association of Offi-
cial Analytical Chemists and as reported in atmospheric emissions from
thermal process phosphoric acid manufacture, cooperative study at the
Manufacturing Chemists Association and the Public Health Service, U.S.
Department of Health, Education and Welfare, October 1968. This method is
based on the spectrophotometric determination of the yellow ammonium
phosphomolybdovanadate complex formed when orthophosphate reacts with
ammonium molybdate-vanadate reagent in an acid medium. The method is
applicable to materials in which phosphorus compounds can be quantitatively
oxidized to the orthophosphate form.
Acid hydrolysis (HN03-HC1, 6 to 1) is used to destroy any organic
material present in the sample and to hydrolyze any phosphate in the meta
or pyrophosphate form to orthophosphate. The system obeys Beer's law to
about 2 milligrams of phosphorus pentoxide (Pg^) Per 10° mi Hi liters of
solution. Results of analyses are reported in terms of P20g.
Certain substances interfere with the ammonium phosphomolybdovanadate
color reaction:
t Certain ions such as ferrous, stannous, and iodine should
be absent because they reduce the color complex to molyb-
denum blue.
• Oxalates, tartrates, and citrates complex molybdenum and
tend to bleach the color.
t High concentrations of iron in the sample cause high results;
however, the iron salts can be converted to the perforate
complex ion which absorbs less light.
jnce OT trie i-uiui ui unc iv» "" >,in.
yellow complex ammonium phosphomolybdovanadate. If present
JheV?«n?Si nrocedure seems to be the volatilization^of the
an
HC10 fume hood is available.
4
129
-------
8.2.15.1 Apparatus and Reagents —
All reagents are prepared from ACS analytical reagent-grade chemicals
in phosphate-free distilled or deionized water.
• Nitric acid (concentrated).
• Hydrochloric acid (concentrated).
• Perchloric acid (70 percent).
• Ammonium molybdate solution (0.2 M). Dissolve 35.3 grams of
ammonium molybdate tetrahydare [(NH4)s M07024 • 4 H^O] in dis-
tilled water and dilute to 1 liter. The reagent is stable at
room temperature and can be stored in a glass stoppered bottle
for at least 3 months.
• Ammonium vanadate-perchloric acid solution (0.02 M
NH4V03 - 4 M HC104). Dissolve 1.17 grams of ammonium
metavanadate in 200 ml of distilled water and transfer
to a 500-ml volumetric flask. Acidify with 172 ml of
70 percent perchloric acid, and dilute with distilled
water to 500 ml. This reagent may be stored at room
temperature for several months.
• Standard phosphate solution. Dry several grams of potas-
sium dihydrogen phosphate (KH2P04) in an oven at 105°C
(221°F). Dissolve exactly 1.917 grams of dried KH2P04
in distilled water and dilute to 1 liter in a 1-liter
volumetric flask. One ml of this solution is equivalent
to 1 mg of Pp^5'
The following apparatus is required:
• Analytical balance.
• Volumetric flasks, 100-, 500-, 1000-ml.
• Erlenmeyer flasks, 250-ml.
t Hotplate.
• Spectrophotometer. This instrument should be capable of
measuring color intensity at 400 nm in 0.5-in. absorbance
cells or larger.
t Constant-temperature water bath. (20°C ±2°C or 68 ±4°F)
• Filter paper (Whatman No. 42).
t Filter funnels and rock.
• Pipets (1-, 2-, 5-, 10- and 20-ml).
130
-------
8.2.15.2 Precautions —
Use proper protective equipment and safety precautions when handling
perchloric acid. In case of contact, flush with plenty of water for
15 minutes.
Temperature and final acid strength play an important role in color
development and stability. A constant temperature bath (20°C ±2°C or 68
±4 F) should be used. Maximum color will develop in 15 minutes; absorbance
will remain constant for at least 2 hours. Final acid strength should be
constant at 0.4 M HC104 for each sample and blank. Slight increases in
absorbance are encountered when acid molarity is decreased from 0.40 to
0.20.
8.2.15.3 Analysis Procedure --
1. Transfer an aliquot of the sample to a 250-ml Erlenmeyer flask.
Simultaneously, prepare a blank (distilled water) and treat in
the same manner. Digest the sample and blank (distilled water)
with 30 ml nitric acid and 5 ml hydrochloric acid. Evaporate
until HC1 fumes are produced (i.e., almost to dryness) on a
hotplate.
2. Cool, dilute to 25 ml with distilled water, and filter
through Whatman No. 42 filter paper into a 100-ml volumetric
flask to remove any insoluble material. Wash filter and
flask several times with 5- to 10-ml portions of distilled
water, and dilute to 100 ml.
3. Pi pet 10 ml of the filtrate into another 100-ml volumetric
flask.
4. Add 10 ml of ammonium vanadate-perchloric acid solution and
20 ml of ammonium molybdate solution to the 100-ml volu-
metric flask and dilute to the mark with distilled water.
5. Place the samples in a water bath (20°C or 68°F). Allow
15 minutes for complete color development.
6. Measure the absorbance against the distilled J?ter-reagent
blank, prepared simultaneously, at a wavelength of 400 run,
using a spectrophotometer and 0.5-in. (l.J cmj cens.
7 Obtain the number of milligrams of PzOs present from a pre
vioSsly prepared calibration curve, where absorbance was
plotted versus milligrams of P205-
131
-------
8. If the amount of ?205 in the aliquot of the sample used is
greater than 2 mg, estimate the amount of P2®5 present by
extrapolating the calibration curve and calculate the proper
aliquot size needed. Take an aliquot from the prepared
filtrate (i.e., the remaining 90 ml) calculated to have an
amount of PgOs suitable for quantitative analysis (0.5 to
2 mg), and proceed with the analysis.
8.2.15.4 Calculation —
(mg P205 found)(volume of original solution)
Total mg P205 = aliquot volume
8.2.15.5 Preparation of Calibration Curve --
1. Pipet exactly 0, 0.5, 1.0, 1.5, and 2.0 ml of standard
P205 solution (1 ml = 1 mg P205) into 100-ml volumetric
flasks.
2. Add the color developing reagents as in the analysis and
dilute to the 100-ml mark. Place samples in a water bath
(20QC or 68QF) and allow 15 minutes for full color
development.
3. Measure the absorbance at 400 nm.
4. Plot absorbance versus milligrams of P20s on square grid
graph paper. The curve follows Beer's law up to 2 mg of
P205 per 100 ml of solution.
8.2.15.6 Comments --
This method is applicable to the determination of total phosphates
in the concentration range of from about 50 yg to 2 mg.
132
-------
REFERENCES
]* o;^iEn^1r°n^Jal Protect1on A9ency5 Federal Register 41 No. m
iijuol , June 1976. '
2* M1ef'-,'V,?nd R; SPe1?nts> "ultra Purity, Methods and Techniques,"
Marcel Dekker, Inc., New York, N.Y., 1972.
3. Benson, A.L., P.L. Levins, A. A. Massucco and J.R. Valentine, paper
presented at 67th Annual Meeting of the Air Pollution Control Asso-
ciation Meeting, Denver, Colorado, May 1974.
4. Gleit, C.E., P. Benson and W. Holland, Anal. Chem., 36, 2067 (1964).
5. Statnick, R. , Destrich and R. Steiber, paper presented at the Annual
ACS Meeting, Chicago, Illinois, 1973.
6. Baldeck, C,, and 6.W. Kalb, "The Determination of Mercury in Stack
Gases of High S02 Content by the Gold Amalgamation Technique," EPA-
R2-73-153, (PB 220-323), Tra Det Inc., Columbus, Ohio, January 1973,
111 pp.
7. Brenchley, D.L., D.C. Turley and R.G. Yaime, "Industrial Source Sam-
pling," Ann Arbor Science Publishers, Ann Arbor, Michigan, 1973.
439 pp.
8. Driscoll, J.N., "Flue Gas Monitoring Techniques," Ann Arbor Science
Publishers, Ann Arbor, Michigan, 1974.
9. Shelley, P.E., and G.A. Kirkpatrick, "An Assessment of Automatic Flow
Samplers," EPA-R2-73-261 , National Environmental Research Center,
Cincinnati, Ohio, June 1973.
10. Benes, P., and I. Rajman, Collect. Czeck. Chem. Commun., 34, 1375
(1969).
11. King, W.G., J.M. Rodriguez and C.M. Wai. Anal. Chem., 46(6), 771 (1974)
12. Robertson, O.E..Anal. Chim. Acta, 42, 533 (1968).
13. Peele, R. , "Mining Engineers Handbook," Vol. II, 3rd ed., J. Wiley
and Sons, New York, N.Y., 1966.
T H r H Chilton and S.D. Kirkpatrick, "Chemical Engineers
>4tt id!. IfcsS^H"! Publishing Co., New York, N.Y., 1969.
15 Taggart, A.F., "Handbook of Mineral Dressing," 2nd ed., J. Wiley and
Sons, New York, N.Y., 1945.
133
-------
16. Welcher, P.O. (ed.), "Standard Methods of Chemical Analysis," Vol. IIA,
6th ed., Van Nostrand Publishing Co., New York, N.Y., 1963.
17. Morrison, G.H., "Trace Analysis Physical Methods," Interscience Pub-
lishers, New York, N.Y., 1965.
18. ASTM Committee D-3 and D-5, "Gaseous Fuels; Coal and Coke," 1971 Annual
Book of ASTM Standards, Part 19, D2013-68, American Society for Testing
and Materials, Philadelphia, Pa., 1971, p. 323-336.
19. Horwitz, W. (ed.), "Official Methods of Analysis," llth ed., Associa-
tion of Official Analytical Chemists, Washington, D.C., 1970.
20. American Public Health Association (APHA), American Water Works Asso-
ciation, and Water Pollution Control Federation, "Standard Methods for
the Examination of Wastewater," 13th ed., Washington, D.C., 1971,
174 pp.
21. "1971 Annual Book of ASTM Standards," American Society for Testing
and Materials, Philadelphia, Pa., 1971.
22. Angino, E.E., and G.K. Billings, "Atomic Absorption Spectrometry,"
Vol. 7 in "Methods in Geochemistry and Geophysics," Elsevier Publish-
ing Co., New York, N.Y., 144 pp.
23. Rains, T.C., and 0. Menis, "Accurate Determination of Submicrogram
Amounts of Mercury in Standard Reference Materials by Flameless Atomic
Absorption Spectrometry," Analytical Chemistry Division National Bureau
of Standards, Washington, D.C., 1972.
24. Slavin, W., "Atomic Absorption Spectroscopy," Wiley Interscience Pub-
lishers, New York, N.Y., 1968, 307 pp.
25. Hatch, R.R., and W.L. Ott. Anal. Chem., 40(14), 2085 (December 1968).
26. Wilson, L., Anal. Chim. Acta.. 35., 123 (1966).
27. Ramakushna, T.V., J.W. Robinson and P.W. West, Anal. Chim. Acta., 37,
20 (1967). ~
28. Perhac, R.M., and C.J. Whelan, Journal of Geochemical Exploration, 1,
47 (1973).
29. "Instrumental Analysis of Chemical Pollutants, Training Manual,"
PB 214-504, Environmental Protection Agency Water Quality Office,
Washington, D.C., April 1971, 294 pp.
30. Kneip, T.J., R.S. Ajemian, J.R. Carlberg, J. Driscoll, L. Kornreich,
J.W. Love!and, J.L. Moyers and R.J. Thompson, Health Lab. Sci., 10(4),
357 (1973). —
134
-------
31 •
seruud.o23
Monsanto Research Corporation, Dayton, Ohio, December 1974, 49 pp.
32. Delgado, L.C., and D.C. Manning, Analyst, 92, 553 (September 1967).
pelgado, L.C., and D.C. Manning. Atomic Absorption Newsletter, 5,
1, (1966). " ~~ — ~
33.
34. Ruch, R.R., H.J. Gluskoter and N.F. Shimp, "Occurrence and Distribu-
tion of Potentially Volatile Trace Elements in Coal," Environmental
Geology Notes No. 72, Illinois State Geological Survey, August 1974,
96 pp.
35. McFarren, E.F. , B.J. Moorman, and J.H. Parker, "Water Fluoride Number
3, Study No. 33," Analytical Reference Service, Environmental Control
Administration, Cincinnati, Ohio, 1969, 71 pp.
36. Peters, E.T., J.E. Oberholtzer and J.R. Valentine, "Development of
Methods for Sampling and Analysis of Particulate and Gaseous Fluorides
from Stationary Sources," PB 213-313, EPA Contract 68-02-0099, A.D.
Little, Inc., Cambridge, Massachusetts, November 1972.
37. U.S. Dept. of the Interior, Bureau of Mines, "Colorimetric Method for
Arsenic in Coal," Report No. 7184, 1968.
38. Fisher Scientific, "Reagent of Choice for Arsenic," Technical Paper
TD 142, 1960.
39. Lishka, R.J., and E.F. McFarren, "Water Trace Elements No. 2,"
Analytical Reference Service, Environmental Control Administration,
Cincinnati, Ohio, 1966, 57 pp.
40. Peterson, H.P., and D. W. Zoromski . Anal. Chem. . 44(7), 1291 (1972).
41. Mair, J.W., Jr., and H.G. Day. Anal. Chem., 44(12), 2015 (1972).
42 Driscoll, J.N., and A.W. Berger, "Improved Chemical Methods for
Sampling and Analysis of Gaseous Pollutants from the Combustion of
Fossil Fuel," PB 209-268, Walden Research Corp., Cambridge,
Massachusetts, 1971.
43. Pulidlo P., K. Fuwa and B.L. Vallee, Anal. Biochem., 14. 393-404
(1966).
44 Platte, J.A., and V.M. Marcy, Atpmic_Abs^^ Perkins-
Elmer Corporation, 4, 289-292 (1965).
135
-------
45. ASTM Committee D-3 and D-5, "Gaseous Fuels; Coal and Coke," 1971
Annual Book of ASTM Standards, Part 19, D2015-68, American Society
for Testing and Materials, Philadelphia, Pa., 1971, p. 343-350.
46. ASTM Committee D-19 and D-22, "Water; Atmospheric Analysis," 1971
Annual Book of ASTM Standards, Part 23, E200-67, American Society
for Testing and Materials, Philadelphia, Pa., 1971, p. 870.
47. Kruse, J.M., and M.G. Mellon. Sewage and Ind. Wastes, 23_, 1402 (1951).
48. American Public Health Association (APHA), American Water Works
Association, and Water Pollution Control Federation, "Standard
Methods for the Examination of Wastewater," 14th ed., Washington,
D.C., 1976, p. 159.
49. U. S. Environmental Protection Agency, Federal Register 36 No. 234,
23248, 1971.
136
-------
REPORT NO.
EPA-600/2-76-283
3. RECIPIENT'S ACCESSION NO
4. TITLE AND SUBTITLE
/Plfa<* , .TECHNICAL REPORT DATA
(t'lease read Instructions on the reverse before completing!
Measurement Techniques for Inorganic Trace
Materials in Control System Streams
6. PERFORMING ORGANIZATION CODE
5. REPORT DATE
December 1976
J.A.Starkovich, R. F.Maddalone, M.L.Kraft,
C. A. Zee. C. Lin, and C.A. Flegal
8. PERFORMING ORGANIZATION REPORT ,\O.
D ADDRESS
TRW Systems Group
One Space Park
Redondo Beach, California 90278
10. PROGRAM ELEMENT NO.
1AB013; ROAP 21AFC-004
11. CONTRACT/GRANT NO.
68-02-1393
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 1/74-6/75
14. SPONSORING AGENCY CODE
EPA-ORD
^SUPPLEMENTARY NOTES JERL-RTP project officer for this report is R. M. Statnick,
919/549-8411 Ext 2557, Mail Drop 62.
16. ABSTRACT
repOrt gives results of B. study showing that inorganic materials in
control process streams at trace levels can be determined using modified, commer-
cially available sampling equipment and atomic absorption analysis procedures; how-
ever, special care must be taken to attain high accuracy. Contamination and alter-
ation of collected samples from sampling train components and laboratory analysis
procedures represent the largest errors in determining trace level materials in
process streams. A modified EPA/Aerotherm high-volume Method 5 sampling train
provides adequate size samples in a 1 to 2 hour sampling period for determining inor-
ganic elements present in gas source streams at 60 micrograms/cu m (1. 1 x 10 to the
minus 7th power gr/scf) and higher levels. The train's collection efficiency at sam-
pling rates from 0. 6 cu m/min (2 scfm) to 0.14 cu m/min (5 scfm) is greater than 95%
for all elements analyzed (e.g. , Hg, Li, Zn, Pb). Procedures and equipment curren-
tly in use for sampling liquids , slurries , and solids for major constituents are accu-
rate and reliable for sampling trace materials , if adequate care is taken to minimize
sample contamination or alteration. Atomic absorbtion methods employing dual chan-
nel instrumentation for background and sample matrix correction are applicable for
canbe determined accu-
KEY WORDS AND DOCUMENT ANALYSIS
Air Pollution
Measurement
Sampling
Inorganic Compounds
Industrial Processes
Colorimetry
Chemical Analysis
19 SECURITY CLASS (This Report/
3. DISTRIBUTION STATEMENT
Unlimited
EPA Form 2220-1 (9-73)
Air Pollution Control
Stationary Sources
Trace Materials
Atomic Absorption Anal-
ysis
20. SECURITY CLASS (This page I
Unclassified
13B
14B
07B
13H
22. PRICE
137
------- |