EPA-R2-73-153
January 1973
Environmental Protection Technology
The Determination of Mercury
in Stack Gases of High SC>2
Content by the Gold
Amalgamation Technique
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
EPA-R2-73-153
The Determination of Mercury
in Stack Gases of High SC>2 Content
by the Gold
Amalgamation Technique
by
Charles Baldeck and G. William Kalb
. TraDet, Inc.
P.O. Box 5093
930 Kinnear Road
Columbus, Ohio 43212
Contract No. 68-02-0697
Program Element No. 1H1326
Project Officer: Howard L. Crist
Quality Assurance and Environmental Monitoring Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND M3NITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
January 1973
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
Elemental mercury, present in the reducing atmosphere of exhaust flue
gases from a zinc smelter may be quantitatively determined by the gold
amalgamation technique. This method avoids interferences by strongly
reducing substances, such as SO , encountered in the direct application
of the normally used wet oxidation techniques (e.g., IC1 or KMnO,
scrubbers) to these sources. The gas sample may be taken isokinetically
using a standard isokinetic stack sampling apparatus in which some of the
impingers are replaced by a series of amalgamators, each containing 30
grams of gold chips. After sampling, these amalgamators are removed from
the sampling unit and the trapped mercury is fired by an induction
furnace into a nitrogen stream which carries the revolatilized mercury
into a solution of 3% KMnO, in 10$ HWO , where it is oxidized and
retained. The resulting solution is then analyzed for mercury by
reduction with hydroxylamine hydrochloride and stannous chloride
followed by direct aeration through a "mercury vapor monitor" which
measures the absorbance at 2^3.7 nm. Mercury collected with the part-
iculate portion of the sample may be determined by nitric acid diges-
tion of the filter followed by reduction with stannous chloride and
aeration.
Several combinations of impingers and amalgamators were investigated
to determine the optimum train configuration. Collection efficiency
of the optimized train was found to approach 98-100£ and to be indepen-
dent of the sampling rate in the range 0.3 to 0.8 CM. Equations were
iii
-------
derived for estimating the collection efficiency of the train from
the relative distribution of mercury found on successive amalgama-
tors. The most crucial parameter affecting the collection efficiency
was found to be the cleanness of the gold used. Sources of error and
possible gold contamination are discussed. Analytical procedures for
determining the mercury concentration were studied, including (1) KMnOi
and IC1 as oxidizing solutions, (2) direct aeration and reamalgamation,
(3) air and nitrogen as tne carrier gases, (h) the use of magnesium
perchlorate as a drying reagent, (5) the use of mixing chambers, and (6)
the utilization of a mercury vapor monitor as compared to a modified
atomic absorption spectrophotometer. The recommended procedure for the
determination of mercury in a stack gas using the method optimized in this
study is presented in the appendix.
IV
-------
TABLE OF CONTENTS
INTRODUCTION 1
LABORATORY INVESTIGATIONS ^
1 - Introduction
2 - Instrumental Methods
3 - Oxidizing Solutions
a - KMnO^
b - Id
U - Water Background
£ - Carrier Gases
6 - Direct Aeration
7 - Amalgamation
8 - Recommended Analytical Procedure
9 - Mercury Capacity of the Gold
10- Filter Analysis
FIELD INVESTIGATIONS 33
1 - The Sampling Site
2 - Sampling Equipment
3 - Field Sampling Procedure
k - Laboratory Analytical Procedure
RESULTS AND DISCUSSION US
1 - The Data Table
2 - Sampling Train Configuration
3 - Initial Scrubber Solution
U - The Use of KMnOi Solutions
5 - Critical Parameters
6 - Collection Efficiency of the Amalgamators
a - Theory
b - 20 grams of Gold per Amalgamator
c - 30 grams of Gold per Amalgamator
7 - Sources of Error
a - Sampling Errors
b - Analytical i Errors
8 - Application to Isokinetic Sampling
CONCLUSIONS 73
APPENDIX I Data Table 77
APPENDIX II Recommended Procedure 99
APPENDIX in Isokinetic Data Sheet, Run 72 108
-------
LIST OF FIGURES
Figure
1. Relative Response of the IDC Mercury Monitor and
an Atomic Absorption Spectrophotometer with a
$g Inch Quartz Optical Cell using the Direct
Aeration Technique. 7
2. Relative Response of the LOG Mercury Monitor and
an Atomic Absorption Spectrophotometer with a
$g Inch Quartz Optical Cell using the Amalgamation
Technique. 8
3» Effect of Mixing Chambers on Analytical Curves with
the Direct Aeration Procedure. 10
I;. Effect of Mixing Chambers on Analytical Curves with
Amalgamation. 11
£. Analytical Curves obtained from Direct Aeration of
IC1 Standard Solutions. 19
6. Effect of Magnesium Perchlorate on Water Absorption
under High Humidity Conditions. 21
7» Analytical Curves obtained by Direct Aeration illustrat-
ing the Effect of Water on Magnesium Perchlorate
Deterioration. 23
8. Water Absorption at 2$U nm as a Function of Water
Temperature. 25
9. Analytical Curves obtained by Direct Aeration utilizing
Magnesium Perchlorate with KMinD^ and ^0 Standard Solutions
and Nitrogen and Air as the Carrier Gases. 26
10. Analytical Curves obtained by Direct Aeration of KMnO> ,
IC1 and H20 Standard Solutions. u 27
11. Analytical Curves obtained from KMnOj. and HpO Standard
Mercury Solutions by Amalgamation. 29
12. Mercury Collection Efficiency of Gold Amalgamators:
Mercury Bypass as a Function of the Quantity of Gold. 31
13. Flow Chart of the Initial Smelting Process at the
ASARCO Columbus, Ohio Zinc Smelter. 3U
vi
-------
1U. Illustration of the 15 mm Amalgamator. 35
1$» Apparatus used for Firing the Amalgamators. ^3
16. Collection Efficiency vs. Sampling Rate for 20 (^
Grams of Gold per Amalgamator.
17. Collection Efficiency vs. Total Mercury Collected (£
for 20 Grams of Gold per Amalgamator.
18. Collection Efficiency vs. Sampling Rate for 30 Qg
Grams of Gold per Amalgamator.
19. Collection Efficiency vs. Total Mercury Collected ^
for 30 Grams of Gold per Amalgamator.
20. Configuration of the Recommended Sampling Train. 103
21. Data Form 108
vii
-------
LIST OF TABLES
Table
1. Half-Reactions Involved in the Oxidation-
Reduction of Mercury. lU
2. Sampling Train Configuration. i;7
3. Distilled Water as the Scrubber. 1*8
U. Stannous Chloride as the Scrubber. i#
5. The Effect of Firing the Amalgamators in the
Presence of the Quartz Wool Plug 55
6. Collection Efficiencies 58
7. Theoretical Distribution of Mercury in a Series
of Amalgamators. ' 59
8. Maximum Escape Fraction for 95$ Recovery. 62
viii
-------
INTRODUCTION
The discovery of unacceptable levels of mercury in water and certain
foods has led to an increased concern over the possible pollution of
the environment with mercury. The persistence of mercury and its
tendency to accumulate in some parts of the ecological system are im-
portant aspects of the problem. These developments have stimulated
interest in the analytical techniques for mercury and numerous improved
methods have been published for its determination in foodstuffs, animal
tissues, blood, urine, water, geological samples, sedimoit, pulp, soil
and rocks. For many of these applications the flameless atomic absorp-
tion method (often combined with an amalgamation step to improve
selectivity and sensitivity-'-) has been found faster, less cumbersome
and more sensitive than the classical dithizone extraction method.
Although the U.S. Government has established maximum permissible levels
for mercury in various foodstuffs and for plant effluent to streams,
little data is available on the quantity of airborne mercury emitted
from such sources as coal-fired power plants or smelting operations.
Most coal has been shown to contain 0.0$ to 0.50 ppm mercury and a
cross section of copper, zinc, and lead sulfide ores from the United
.W. Kalb, The Determination of Mercury in Water and Sediment
Samples by Flameless Atomic Absorption, Atomic Absorption Newsletter
9(U), PP. 81-37(1970).
^Personal Communication, ASTM Committee D£,21, Trace Element
Task Group.
-------
States has been analyzed for mercury showing a range of 0.0£ to
300 ppm. This naturally occurring mercury is volatilized during
combustion and could result in the release of a significant quantity
of mercury into the atmosphere.
The two methods ordinarily used for the determination of mercury in
air are subject to massive interferences from the other components
normally found in stack gases, particularly SO . Efforts to determine
the mercury in stack gases by drawing the gas through a sampling train
containing a liquid oxidizing agent (acidic KMnO^ or IC1) have not been
successful. The high S02 concentration in the sample reduces the
oxidizer almost immediately, eliminating its ability to oxidize the
mercury to the mercuric state. The direct measurement of mercury in
air using a "mercury vapor monitor" is another widely used technique.
The air sample is drawn between an ultraviolet source which emits the
2$3.7 nm mercury vapor resonance line and a photocell detector. The
absorbance is measured and can be converted directly to mercury con-
centration; but the response is not specific since SO , most organic
substances, smokes, and aerosols also absorb at this wavelength. An
alternate method of mercury collection is needed for stack sampling;
one which does not depend on the oxidation of mercury for entrapment.
Such a method would enable longer sampling times to be used, even in
^Personal Communication, David Patrick, Environmental Protection
Agency, Research Triangle Park, North Carolina.
-------
the presence of high SO concentrations. Previous work by the authors
has shown that elemental mercury can be quantitatively collected from a
stack gas sample by direct amalgamation onto gold. The mercury can then
be revolatilized by heating and determined by any of several methods. Good
recovery of mercury was obtained from the effluent gas of a coal-fired
power plant and a zinc roaster where SO- concentrations averaged around
7-8$. Although recovery of mercury in this earlier work was generally good,
the analytical method used (firing the gold in an air stream which carried
the revolatilized mercury through a quartz cell positioned in the beam of
an atomic absorption spectrophotometer) was too sensitive, limiting the
feasible sampling time to 1-3 minutes, depending on the mercury concentra-
tion, and subject to interferences from any moisture, sulfuric acid mist or
other substances which condensed on the gold during sampling. These problems
limited the accuracy and reproducibility of the results obtained. The
present work was undertaken to solve these problems and generally improve
the practicality of the gold amalgamation method as applied to stack
sampling. The goal of this work was threefold:
1. To develop a procedure which would allow a sampling time of
at least 15 minutes at isokinetic rates (i.e., a sampling rate
of 0.5 - 0.8 CFM, or under actual isokinetic conditions).
2. To obtain and show a collection efficiency of at least 9$%
for the mercury.
3. To perfect the method of firing the gold into an acidic
KMnOi solution with subsequent analysis of aliquots of this
solution by flameless atomic absorption.
These goals were successfully attained in this study.
^G.W. Kalb and C. Baldeck, The Development of the Gold Amalgamation
Sampling and Analytical Procedure for Investigation of Mercury in Stack
Oases, Environmental Protection Agency Contract No. 68-02-03l4l (1972).
%.W. Kalb, The Adaptation of the Gold Amalgamation Sampling and
Analytical Procedure for Investigation of Mercury in Stack Gases to High S02
Environments Observed in Smelters, Environmental Protection Agency Contract
No. 68-02-03U1 (1972).
-------
LABORATORY INVESTIGATIONS
1 - Introduction
A previous study has shown that volatile mercury in smelter gases may
be quantitatively collected on gold by an amalgamation reaction. (At
the high temperatures (500°F) and the reducing atmospheres observed in
smelter gases the mercury present will be in the elemental state.) The
mercury was analyzed by firing the amalgam in an induction furnace and
then monitoring the volatile mercury concentration by flameless atomic
absorption* Standards were run after each sample to calibrate the
instrument. Although the smelter investigated used an ore of relatively
low mercury concentration, the mercury concentration was sufficiently
high to limit the sampling time to one minute. Samples collected iso-
kinetically for longer periods of time contained more mercury than could
be analyzed by the system.
In order to obtain a representative sample a longer sampling period was
required resulting in the otherwise arbitrary choice of a desired
15-minute sampling period. Sample splitting procedures and redesign
of the optical system could not adequately desensitize the system,
especially when it is realized that ores used in some smelters contain
300 times the mercury concentration observed in the smelter studied*
As a result of this it was decided to revolatilize the mercury collected
on the amalgam, absorbing the revolatilized mercury in an acidic perman-
ganate or iodine monochloride solution. Aliquots of these solutions
could then be diluted for analysis. It was the objective of this phase
of the contract to study various methods for analyzing these solutions.
-------
The mercury-gold amalgam obtained from sampling the stack gas is fired
in a resistance or an induction furnace with a compressed gas stream
carrying the mercury through a liquid absorption cell containing an
iodine monochloride or acidic permanganate solution that quantitatively
removes the mercury from the gas stream. An aliquot of this solution
is then diluted to a satisfactory range for analysis. Procedures in-
vestigated for the analyses of these solutions included: (1) direct
aeration of the reduced KMnOi or IC1 solution into a flameless atomic
absorption spectrophotometer, and (2) secondary amalgamation accom-
panied by the direct firing of the amalgam into a flameless atomic
absorption spectrophotometer. These procedures were studied using
both air and nitrogen as the carrier gases and with and without magnes-
ium perchlorate as a drying reagent in the gas stream. Various mixing
chambers were also investigated as an additional means of desensitizing
the analytical method.
The second objective of this study was to determine the ultimate cap-
acity of the gold for mercury. The results from this are to be used to
determine the size and shape of the final amalgamator to be used in
conjunction with the isokinetic sampler in obtaining the sample from
i
the stack.
The final objective of the laboratory study was to develop a method
for analyzing the filters containing the participates collected during
sampling.
-------
2- Instrumental Methods
The instrumental methods were investigated by comparing their sensitivity
and response with standard mercury in water solutions. $0 ml aliquots
of the various mercury standards were reduced with 2 ml of a solution
of 20$ SnCl in 50£ HCl. These standards were then aerated, quickly
volatilizing the reduced mercury. The air stream carried the mercury
either directly into a quartz optical cell located in the path of an
atomic absorption spectrophotometer operated at 25U nm (direct aeration)
or onto gold (secondary amalgamation). In the latter use, the gold
amalgamator was then fired in an induction furnace revolatilizing the
mercury which was carried by the air stream into the optical path of
the spectrophotometer. A model 303 Parkin-Elmer atomic absorption
spectrophotometer and a Laboratory Data Control (IDC) Mercury Monitor
were utilized for the study.
Figure 1 shows the relative response of the LOG Mercury Monitor witn
a 30 cm path lengtn and the atomic absorption spectrophotometer with
a 6-^5 incn optical cell, using the direct aeration technique. The two
units were operated at the same air flow (l.U liters/minute). The
IDC unit, operated at a 0.614. range (least sensitive available),
shows an absorbance of 80 with a 1 tig mercury standard. The atomic
absorption unit has a considerably lower sensitivity. Figure 2
illustrates a similar comparison, but with the secondary amalgamation
metnod. The gold amalgams were fired at different temperatures (%
variac setting) in the induction furnace. The secondary amalgamation
-------
6O-
Figure 1. Relative Response of the IDC Mercury Monitor and an Atomic Absorption
Spectrophotometer with a &$ Inch Quartz Optical Cell using the Direct
Aeration Technique.
-------
Figure 2. Relative Response of the IDC Mercury Monitor and an Atomic Absorption
Spectrophotometer with a 62g Inch Quartz Optical Cell using the
Amalgamation Technique*
-------
procedure, originally designed to eliminate interferences and increase
sensitivity, is considerably more sensitive than the direct aeration
method.
The LDC unit is, as would be expected, more sensitive than the atomic
absorption spectrophotometer. The LDC unit in conjunction with amal-
gamation approaches the maximum limit of sensitivity. At the low
mercury concentrations required to remain on scale there is considerable
fluctuation in the absorbance values. The sensitivity of this method
is such that it is useful only under optimum conditions. This appears
to be because of temperature fluctuations in the optical cell due to
the firing of the amalgam, resulting in an unstable equilibrium between
the volatile materials and the cell walls. This, in conjunction with
the high sensitivity due to the long cell, severely limits its usage.
Figures 3 and k illustrate the effect of mixing chambers on the sensi-
tivity of the direct aeration and secondary amalgamation procedures.
The two mixing chambers investigated, l£0 ml and 270 ml gas collection
tubes, did not significantly decrease the sensitivity with either the
direct aeration or amalgamation techniques. They did improve the repro-
ducibility of the peaks by decreasing the sharpness of the peak tip with
both procedures. It is likely that utilizing the mixing chambers
with the secondary amalgamation procedure will decrease the sudden
thermal expansion of the gas, which could be responsible for the lack
of reproducibility with the highly sensitive LDC Mercury Monitor.
-------
10
Absorbance
80-
40-
no mixing chamber
150ml
27Oml
Figure 3* Effect of Mixing Chambers on Analytical Curves with the Direct
Aeration Procedure.
-------
SO-
rng Hg
Figure 1*. Effect of Mixing Chambers on Analytical Curves with Amalgamation.
-------
12
The results of the instrumental analysis comparison has suggested that
the secondary amalgamation procedure is unnecessary. The amalgamation
procedure is designed to eliminate interferences and increase sensitivity.
The primary amalgamation step utilized in the original collection of the
stack gas sample and the desire to decrease the sensitivity, limits its
usefulness unless an interference is found in the oxidizing solutions.
The amalgamation procedure can be partially desensitized by lowering
the firing temperature or the air flow rate. If the amalgamation pro-
cedure is adopted, it is recommended that the mixing cells should be
used in conjunction with the 6-^g inch optical cell in the atomic
absorption spectrophotometer. If the IDC Mercury Monitor is to be used,
a shorter optical cell would be advantageous ($500.00). Considering the
high mercury concentrations of the oxidizing solutions to be analyzed,
the direct aeration procedure, utilizing either the LDC Monitor or the
Perkin-Elmer atomic absorption spectrophotometer is recommended. With
this procedure mixing chambers would not be required.
3- Oxidizing Solutions
The oxidizing solutions to be utilized in collecting the mercury at the
field site are to be a 3% KMnOi solution in 1.0% concentrated nitric
acid or a 0.1 N IC1 solution. In practice these solutions would be
diluted for final analyses. Due to the various dilutions that will be
performed on the actual samples and as a check on any possible inter-
ferences, the laboratory study was performed with the above strength
solutions. These solutions were then spiked with known amounts of
-------
13
mercury before analyses. Analysis of mercury in water standards have
been thoroughly investigated by this laboratory and were utilized for
comparison with the oxidizing solutions, as well as to compare the
various instrumental procedures.
a.
The reactions responsible for the oxidation and reduction of mercury
in the permanganate system are shown in Table 1, In acidic solutions
the permanganate ion is a strong oxidizing reagent. The half reaction
is
MnOr + 8H+ + 5e~ * - - Mn++ + ltf_0 E° = 1.51
U 7 2
This half reaction only occurs in strongly acidic solutions (0.1 N or
greater) . In less acidic permanganate solutions
Mrfl" + l|H* + 3e" ' , Mifl2 + 21^0 E° - 1.695
In alkaline solutions
MnO~ + 2H00 + 3e~ -—7 MnD. + UOH* E°
In strongly alkaline solutions
Mrf)~ + e" ^ , MnO" E°
k h
Permanganate is unstable in the presence of Mn , but the reaction is
very slow in acidic media
2MnO + 3Mn + 2H 0
k 2
-------
Ill
TABLE 1
Half-Reactions Involved in the Oxidation-Reduction of Mercury
Potential
1. Oxidation of Mercury:
Hg° ±^ Hg4
(Strong Acid)
2e'
-0.852
(Weak Acid)
+ 2H20
2. Reduction of Excess Mn07 :
k
(Strong Acid)
Mn + liH 0
(Weak Acid)
MnO,
+ 2H20
Mn + 2H 0
2
HgNOH'HCl
3« Reduction of Mercury:
°
Hg
Sn*"*" + 6C1~
NO, + 2H00
2 h 2
HNO« + UH 0
^ 2
N02 + 2H20
MnOr + 8H+
~ -1.51
3e~ -1.695
MnO~
U
8H*
MnO- + UH* + 3e- -1.695
u
i4H+ + 2e~ -1.23
MrtD
(N0, N0, HN02)
Hg++ + 2e~
SnCl. + 2e~
6
2NO- + UH*
3
NO" + 3H*
NO" + UK*
-0.852
-0.15
+ 2e- -0.9
+ 2e- -C.9U
+ 3e~ -0.96
H.A. Laitinen, "Chemical Analysis," McGraw-Hill Book Company,
New York, N.T., I960.
-------
This reaction would affect the length of storage of KMnOr solutions
after the mercury has been collected. The overall reaction would be
5Hg° + 2MnOj + 16H+ * £Hg++ + 2Mn++ + 8H20
The Mn would eventually lead to the deterioration of the permanganate.
catalyzes the decomposition of permanganate under acidic conditions
resulting in the necessity of filtering fresh permanganate solutions.
Any organic material,, present in the storage vessel will reduce the
permanganate to MnO^ resulting in autodecomposition. Acidic and alka-
line solutions of permanganate are less stable than neutral solutions.
MnD2 will also form in the initial oxidation of the volatile mercury if
the solution is not sufficiently acidic (see Table 1).
With the utilization of the permanganate system the excess permanganate
(that not required to oxidize the mercury) is reduced with hydroxylamine
hydrochloride, The half reactions are shown in Section 2 of Table 1,
(The oxidation potential of the hydroxylamine hydrochloride is unknown,
but would be between 0.9 and 1.23 volts.) The formation of Mn + or
is dependent upon the acidity of the solution. It should be noted
that the by-products from the oxidation of the hydroxylamine hydro-
chloride are volatile NOX compounds.
The final revolatilization of the mercury is accomplished by reducing
the mercury with a stannous chloride solution. The half reactions are
shown in Section 3 of Table 1. The stannous chloride will also reduce
the nitric acid (used to acidify the permanganate) forming volatile
NO compounds.
-------
16
In this study 3$ w/v aqueous permanganate solutions were prepared from
reagent grade potassium permanganate and were filtered to remove any
o present in the original crystalline material. Acidic solutions
ml) were then prepared by adding concentrated nitric acid to achieve
a 10/5 nitric acid concentration. Known concentrations of mercury in
water (slightly acidified) were then added to these prior to analysis.
A 10J6 hydroxylamine hydrochloride solution (5 ml) was used to reduce
the excess permanganate and a 20% SnCl2 in 5Q£ HC1 (3 ml) was used to
reduce the bivalent mercury to elemental mercury. These solutions
were then aerated, either directly through the spectrophotometer or
amalgamated onto gold.
The acidification and reduction of the excess permanganate are both
exothermic reactions. Optimization of the addition of the reducing
agents showed that more SnCl was required with the permanganate than
with a mercury-water standard, even when the excess permanganate has
been completely reduced. The results showed that 3 ml of the SnCl2
solution was required compared to one ml with a similar mercury-water
standard sample. (Additional hydroxylamine hydrochloride had no effect
on the stannous chloride requirements.) This probably represents the
reduction of the nitric acid which has a slightly higher oxidation
potential than the elemental mercury (Table 1) . NOX compounds are
released during both reduction steps. The final reduction of the mer-
cury occurs in a closed system resulting in the inclusion of the
volatile NO compounds in the sensing cell with the volatilized mercury,
jC
unless gold amalgamation (secondary amalgamation) is utilized. Several
-------
17
investigators have suggested that NO compounds interfere with the
analysis. Although some questionable results have been obtained, NO.,
has not been correlated to problems with either the direct aeration
or amalgamation procedures.
The reduced permanganate standards produced higher peaks than observed
with water standards. The higher peaks are a result of the blank mercury
concentration in the permanganate as well as a slightly quicker release
of the mercury from the solution probably due to the excess stannous
chloride.
Initial studies showed an increase in sensitivity with the age of the
acidified permanganate. Approximately 3 hours was required after
acidification before maximum sensitivity was obtained. More recent
studies with both the 10JK and 25? acidified permanganate samples showed
no differences in sensitivities between samples stored for 15 minutes or
3 hours before analysis. Both series were performed by identical pro-
cedures and no explanations are offered for the observed differences.
Mn02 is sometimes formed during the reduction of the excess permanganate.
The brown precipitate disappears with additional stirring of the sample.
The occurrence of the MnD has not been correlated to any procedural
differences and no analytical differences have been observed because of
it. Mr. Joseph DeGarmo of American Electric Power has correlated this
with deterioration of the magnesium perchlorate leading to absorption
of the volatile mercury by the perchlorate.
"Personal Communication.
-------
16
The stability of the spiked permanganate solutions reduced with
hydroxylamine hydrochloride was investigated. Permanganate solutions
spiked with 0.25> and O.£0 ug mercury were reduced with hydroxylamine
hydrochloride and then stored for 10 rain., 30 min., 1 hour, h hours,
and 22 hours before final reduction and analysis. No significant
differences were observed between the various intervals. It is the
opinion of the authors that since the HIJO-, is not reduced by the
hydroxylamine hydrochloride there should not be a significant loss of
mercury with storage. This is in contrast to the results reported by
Mr. S.T. Hirozawa of the Wyandotte Chemicals Corporation.'
Although some problems have occurred with permanganate solutions, the
authors feel that when collecting inorganic mercury, accurate, reliable
data can be obtained from permanganate solutions utilizing either direct
aeration or amalgamation. The purpose of the oxidizing solutions in the
stack sampling procedure is to split or dilute the sample. This will
decrease any peculiar effects of the solution.
b. IC1
The IC1 method utilized in this study was the Determination of Mercury
in Particulate and Gaseous Emissions from Stationary Sources developed
/Q\
by EPA. The results obtained from this procedure are presented in
7
S.T. Hirozawa and J.K. Rottschafer, "Trip Report - Mercury
Emission Via Hydrogen Gas and Fume Headers at Port Edwards," memo to
C.V. Francis, 1/15/71.
8 Federal Register 36(234), Dec. 7, 1971.
-------
19
Absorbance
60-
a N2-Mg(CIO4)2
• N2-No Mg(CI04)2
O Air- Mg(CIO4)2
• Air-No Mg
-------
20
Figure 5- The technique is extremely sensitive to any procedural
variations and the presence of any water or water vapor in the system.
This resulted in extensive practice being required for each technician
involved in the study. Even with experienced personnel different indi-
viduals could not obtain acceptable agreement in results. Because of
this, the utilisation of this procedure for routine analyses is ques-
tioned by the authors.
li- Water Background
Water vapor will absorb 2^U nn radiation resulting in erroneous peak
heights when determining mercury concentrations. This interference is
dependent upon the relative humidity r;p_:i nay be eliminated by utiliza-
tion of the gold amalgamation technique or a desicant, such as magnesium
perchlorate, between the aeration cell and the optical cell. The amal-
gamation procedure separates the water vapor and the mercury by collecting
the mercury on the gold and passing the water vapor. With the direct
aeration procedure the water vapor is eliminated by absorption onto
magnesium porchlorate.
Figure 6 illustrates the effect of utilizing magnesium perchlorate with
the direct aeration technique under high humidity conditions. These
results were obtained under high humidity conditions using nitrogen as
the carrier gas. Under dry humidity conditions, essentially no back-
ground water vapor is observed when using the direct aeration technique.
Extensive analyses of these parameters has shown that a high background
water vapor value is obtained under high humidity, even when the air or
-------
o
u
c
o
JO
I
Water Background
N2 Flow 3
No Mg(CIO4)2
O.5 H9 Hg
Direct Aeration
H2O blank no
O.5 jjg Hg no Mg(CIO4)2
O.5 pg Hg Mg(CIO4)2
Time
Figure 6. Effect of Magnesium Perchlorate on Water Absorption under High Humidity Conditions.
-------
22
nitrogen used as the carrier gas is dried before the aeration cell.
Since the air in the dead space above the liquid phase in the aeration
cell is quickly purged with the carrier gas, it is concluded that high
humidity conditions have a chemical effect on the liquid sample result-
ing in the vaporization of more water under these circumstances.
With the introduction of magnesium perchlorate as a desiccant between
the aeration cell and the optical cell this background is eliminated
(see Figure 6). Magnesium perchlorate has been observed by the authors
and reported by other investigators to undergo deterioration and even-
tually absorb the volatile mercury. To determine the effect of this
deterioration in order to know when the magnesium perchlorate should be
replaced, water was added dropwise to the magnesium perchlorate and
standard analytical curves were obtained. The results of this study are
shown in Figure ?• There were no significant losses of mercury observed.
Additional water was added until it had completely saturated the magnes-
ium perchlorate. There was no loss of mercury until the tubing below
the desiccant became clogged with the saturated suspension. Temperature
effects have been observed to deteriorate the magnesium perchlorate
resulting in the absorption of mercury. If, when using the secondary
amalgamation technique, magnesium perchlorate is positioned between the
amalgamator and the optical cell, a significant loss of volatile mercury
is observed after the first few runs. Apparently, this is related to the
effect of the hot air, obtained when the amalgam is fired in the induction
furnace, interacting with the magnesium perchlorate.
-------
23
Absorbance
80-
6O-
40-
2O-
O Dry Mg(CIO4)2
• 5 Drops H2O
O 1O Drops H2O
• 15 Drops H2O
O.4
0.8
HS
Figure 7. Analytical Curves obtained by Direct Aeration illustrating the
Effect of Water on Magnesium Perchlorate Deterioration.
-------
2U
Without magnesium perchlorate the observed water background is related
to the temperature of the aqueous solution to be aerated as well as the
relative humidity. Figure 8 illustrates this relationship. This rela-
tionship is critical when a potassium permanganate solution is used
since the reduction with the hydroxylamine hydrochloride and the stannous
chloride is exothermic.
5- Carrier Gases
Compressed air and nitrogen were investigated as possible carrier gases
in the system. With the use of magnesium perchlorate there were no dif-
ferences observed between the two gases. Figure 9 contains the analytical
curves obtained with magnesium perchlorate with KMnOi and H^O standards
using both nitrogen and air as the carrier gases. Without magnesium
perchlorate, compressed air volatilized more water than nitrogen, result-
ing in a higher water vapor background.
6- Direct Aeration
Figure 10 illustrates the differences observed in the standard curves
obtained by direct aeration from KMnO, , IC1 and E^O standard solutions.
The KMnO, curve shows the blank mercury concentration as would be
expected. The parallel lines of the KMnOi and H_0 standard curves show
that the rate of release of the mercury from the two solutions is very
similar. The different slope of the IC1 curve illustrates a slightly
different rate of release of the mercury. This partially explains the
lower curve obtained with IC1, but it still appears that not all the
mercury is being released during the reduction and aeration.
-------
Absorbance
80-
60-
40-
20-
60°C
Water Background
Air Flow 3
Chart 2 cm/m
29°C
9°C
Time
Figure 8. Water Absorption at 25U nm as a Function of Water Temperature.
-------
26
Absorbance
6O-
4O-
2O-
• • Nitrogen
O O Air
(X2
O!4
Hg
Figure 9o Analytical Curves obtained by Direct Aeration utilizing Magnesium
Perchlorate with KMnO. and H20 Standard Solutions and Nitrogen and
Air as the Carrier Gases,
-------
Absorbance
6O-
KMnO,
Mg Hg
Figure 10. Analytical Curves obtained by Direct Aeration of
and H_0 Standard Solutions*
Id,
-------
28
The difficulties with the reproducibility of the analytical results
using IC1 and the slower release rate of the mercury from the IC1
solution results in the recommendation to adopt the KMnOr solution
when studying volatile elemental mercury as it is observed in smelters.
7- Amalgamate on
Figure 11 illustrates the results obtained from amalgamating mer:-_-
aerated from KMnCr and water standard solutions. The blank mercury
concentrations are similar to those observed with direct aeration from
permanganate solutions. The amalgamation procedure, firing the gold-
mercury amalgam in a Leco induction furnace at a 90% variac setting in
conjunction with an LDC Mercury Monitor (range 0.6U), results in an
absorbance of 70 with a 0.2 ng sample of mercury. As discussed pre-
viously, this is extremely sensitive and greatly limits the working
range oi" the procedure.
8- Recommended Analytical Procedure
The laboratory study has resulted in the adoption of the direct aeration
procedure in conjunction with KMnO. solutions. Compressed air or nit-
rogen may be used as the carrier gas. It is recommended that a magnesium
perchlorate drying tube be incorporated into the system. Either the LDC
Mercury Monitor or a standard atomic absorption spectrophotometer may be
used. The analyses of the samples collected in the field portion of this
study utilized compressed air as the carrier gas and an LDC Mercury Monitor.
The final recommended analytical procedure is presented in Appendix II.
-------
Absorbance
29
60-
20 -I
0.1
Hg
0.2
Figure 11. Analytical Curves obtained from KMnO. and IUO Standard Mercury
Solutions by Amalgamation,, **
-------
30
9- Mercury Capacity of the Gold
The efficiency of mercury retention by the gold in the amalgamators
waa determined by measuring the amount of mercury bypassing the gold.
This was studied in the laboratory by reducing and volatilizing a
standard mercury solution which was carried by an air stream (l.U
liters/minute) through a gold amalgamator. The mercury bypassing
the amalgamator was absorbed in a KMnO. solution using the bubbler
assembly. This solution was then analyzed by the direct aeration
method. Determinations were obtained from solutions containing
quantities of mercury up to 70 |ig. The results of this study utilizing
a standard 15 mm diameter amalgamator are illustrated in Figure 12»
25 mm diameter amalgamators (shorter height of gold) showed a greater
bypass of mercury than the 15 mm amalgamators with comparable amounts
of gold. The results of this study are dependent upon the air flow
rate and thus does not represent a true maximum mercury capacity of
the gold. (Since amalgamation represents the mercury dissolving the
gold at the point of contact there is no ultimate mercury capacity of
the gold.) The results of this study are used to obtain a rough
indication of the amount of gold required when obtaining a stack gas
sample.
10- Filter Analysis
The filters obtained during the isokinetic sampling of stack gases
must be analyzed for mercury to determine the total mercury concen-
tration of the gas stream. Two analytical methods for dissolving the
particulates were investigated:
(1) One half of the filter was boiled in 10 ml of concentrated
-------
15 mm Amalgamator
O 10 gms Gold
A 20 gms Gold
B 3O gms Gold
5H
4H
. 3-
&
51-1
i_
'i
A
A n n . -O-
| V ' T~ ' '
2O 40 60
lig Hg Delivered
Figure 12<, Mercury Collection Efficiency of Gold Amalgamators: Mercury Bypass as a Function of the Quantity of
Gold.
-------
32
nitric acid for 10 minutes. The filter was disintegrated
with a high pressure stream of distilled water and the
mixture was diluted to 100 ml.
(2) The other half of the filter was placed in a Bethge
apparatus with 2 mg ammonium meta-vanadate (catalyst),
5 ml nitric acid, and 10 ml ?0£ perchloric acid. The
mixture was heated, collecting and withdrawing the
nitric acid. The filter was then digested in the reflux-
ing perchloric acid for 10 minutes. The nitric and per-
chloric acid solutions were combined and diluted to 100
ml.9
After cooling, the solutions were analyzed for mercury by placing $0
ml aliquots in the interchangeable sample holders, reducing the mer-
cury with 2 ml of the stannous chloride solution and aerating the
mercury through the LOG Mercury Monitor at a flow rate of l.U liters
per minute.
Both methods dissolved nil visible particulates and produced comparable
data within experimental error. Because of the simplicity of the
procedure, the nitric acid method is recommended and was used by this
laboratory for the field phase of this study.
9
"The Wet Chemical Oxidation of Organic Compositions Employing
Perchloric Acid." The G. Frederick Smith Chemical Co., Inc., 1965.
-------
33
FIELD INVESTIGATIONS
1- The Sampling Site
The samples for this study were taken at the American Smelting and
Refining Company (AbARCO) zinc smelter in Columbus, Ohio. A flow
chart of the initial smelting process at this plant is given in
Figure 13. The ore is roasted at 900°C in a fluid bed roaster
volatilizing the sulfur as S0?, and the mercury, presumably, as
elemental mercury (the high temperature of the roasting operation
and the fact that there is essentially no organic material present in
the ore precludes the formation of volatile organo-mercury compounds).
The particulates are removed from the gas stream by a waste heat boiler,
a cyclone and two electrostatic precipitators and the gas is then carried
through a 3-5 foot diameter horizontal steel pipe (the cross-over duct)
running about 60 feet above ground level to the acid recovery plant,
where the S02 is converted to sulfuric acid. This duct has a 150 foot
straight section, without constriction or bends, and a single U-inch
sampling port was located in this duct about 30 feet from the down-
stream end. The flow in this duct is controlled by ID and FD fans,
resulting in a positive-negative pressure interface- which migrates back
and forth along the duct in the area of the sampling port. Under normal
runrdng conditions, static pressure at the port was usually within
+ 0.£ inches of water.
-------
electrostatic pretfip.
cross.over duel
cyclone
air
• sampling site
heat exchanger
stack
Figure 13. Flow Chart of the Initial Smelting Process at the ASARCO Columbus,
Ohio Zinc Smelter*
-------
The gas stream in tne duct contains about 7 - 8£ S0? and flows at a
rate of 12,000 - 15,000 CFM. The average molecular weight of the gas
has been calculated to be 31.k and contains about $% water (figures
provided by ASARCO and EPA).
2- Sampling Equipment
The samples for this study were taken with a standard Model 23U3 RAC
"Staksampler" portable stack gas sampling unit utilizing a 5-foot
glass probe heated to l50°F. T'he sampling train was a standard EPA
isokinetic sampling train with some of the wet impingers replaced by
amalgamators. It consisted of a probe mounted on a sample box and
connected to a cyclone and filter which were enclosed in a heated
compartment, then a series of impingers (and/or amalgamators) in an
ice bath, followed by an impinger containing silica gel. The sample
box was connected to the console containing the dry gas meter and pump
by a 30-foot umbilical cord.
Modified Greenburg-Smith impingers (without the tip) were altered as
shown in Figure lit by forming three indentations in the center (to
support a quartz wool plup and the gold) and attaching a 28/15 pyrex
ball joint to the bottom of the straight vertical tube. These units,
called amalgamators, can be inserted into a standard hPA isokinetic
sampling train in place of the standard impingers. The interchange-
ability of impingers and amalgamators permits the use of a combination
of amalgamation and wet absorption techniques to verify the collection
efficiency of the various components. This style of amalgamator also
-------
36
LJ>
Figure lU. Illustration of the 15 mm Amalgamator.
-------
37
allows the standard isokinetic sampling train to be used for collection
of mercury without major alterations to the equipment.
For the purpose of this study, a larger number of impingers/amaiganators
were used in the sampling train than are ordinarily employed for iso-
kinetic sampling. The sampling box was originally designed to hold k
impingers in the ice bath, including the silica gel. For this work,
as many as nine impingers (including the silica gel) were used in the
sampling train. A maximum of six. could normally be placed in the
sample box, and two more could be taped to the outside of the box. By
compressing the insulation somewhat on one of the two sample boxes used,
it was possible to fit seven inpingers inside the box and two more could
be taped to the outside. As a result the box was severely crowded and
some misalignment of the connecting tubes was unavoidable. In addition,
some of the cooling capacity reserve of the box was lost as there was
less room for the ice/water mixture and the impingers touching the in-
side edge of the box were not surrounded on all sides by the coolant.
Since the ambient temperature during this work was fairly low (mostly
in the 30's) and the runs were comparatively short ( 5 - 15 minutes),
no troubles were experienced keeping the impingers well cooled. It
was observed, however, that when the ambient temperature rose occasion-
ally to the UO's or 50's, the ice was somewhat depleted after about
10 to 1$ mJnutes running time.
Luring this work, a considerable quantify °f SC0 wao drawn through the
-------
38
console unit. In this procedure, the SO is not removed at the sample
box but continues through the pump and dry gas meter. After each run
about 1 CF of ambient air was drawn slowly through the intake of the
silica gel impinger in an effort to rinse out some of the SOo in the
console unit. Some difficulty was experienced with the pumps in the
consoles. The pumps were periodically found to pull erratically or
not at all. Disassembly of the pump according to instructions furnished
by the manufacturer showed that the oil which normally lubricates the
sliding fiber vanes had congealed so that the vanes were sticking in
their slots in the rotor. The ptmp8 were cleaned with solvent (acetone)
sid reassembled. The oil in the reservoir was changed to a mixture of
1%% SAE 10 and 2$% kerosene as recommended by the manufacturer for
operation at temperatures below freezing. After this treatment, the
pumps ran smoothly for two or three runs and then started to operate
erratically as before. It was then necessary to disassemble and clean
the pumps again. A similar difficulty was experienced with the check
valve in the metal adapter which goes from the last impinger to the
umbilical cord. Toward the end of this work the check valve on both
adapters became corroded and stuck in a partially closed position,
causing a high flow resistance at that point. The check valve assembly
was removed from the adapter.
3- Field Sampling Procedure
Since glassware stored openly in the laboratory may absorb traces of
-------
39
mercury, all glassware was rinsed before use with the sequence: 1
SnCl2 in 2.5/5 HC1, 1:3 HMOy-H 0, distilled water, acetone, and then
dried.
The amalgamators were prepared for each run as follows. A small plug
of quartz wool was inserted from the top and pushed into place against
the supporting indentations. A length of ^-inch dowel rod and a
stiff (about #10) wire are handy to help wedge the wool into position.
In earlier work, the authors used a small wad of gold wire for this
purpose, which sometimes became loose in the tube from handling and
allowed some of the finer gold particles to fall out the bottom of the
tube. No trouble of this sort was ever encountered with the quartz
wool plug. The gold chips were prepared by cutting up a 0.007-inch
thick sheet of the metal into small 1/16-inch squares. These gold
chips were placed in small crucibles and fired overnight in a refrac-
tory oven at about 600-?00°C. At the start of each day's sampling,
the gold was removed from the oven and allowed to cool. The gold
chips were then weighed out and poured into the amalgamator on top of
the plug, using a plastic funnel. The amalgamator was held in a vertical
position and tapped gently to help settle the chips. The amalgamator
was then fired in the induction furnace to insure that any
10 J.D. Brooks and W.E. Wolfram, American Laboratory 3(5U),
(1971).
-------
mercury picked up during the handling procedure from glassware, plug,
etc., was driven off before the amalgamator was assembled into the
train. The train sequence was then assembled for each run as shown
in Part B of Appendix II.11
After positioning the sample box at the port, the heater was turned on,
the ice compartment filled with ice and water and a leak check was per-
formed. The probe was then connected and inserted with the tip facing
upstream (with the exceptions of Runs 61 and 62 where the tip was
positioned facing downstream). The 0.2£-inch diameter tip was used
for Runs 1-60 and 71 - 72; the 0.50-inch diameter tip was used for
Runs 61 - 70. After the probe and sample box had been allowed to come
to the proper temperature (probe heater setting was 100£ (l50°F)j
sample box 250°F), the pump was turned on and the flow adjusted to give
the desired sampling rate. A stopwatch was used to time the run* In
a five-minute run the dry gas temperature, stack temperature, etc., were
read at 2-3g minutes. On longer runs the readings were made every five
minutes.
After obtaining the sample, the probe and sample box were taken to the
on-site mobile laboratory for the cleanup procedure. Wide-mouthed jars,
holding a pint or six ounces and equipped with plastic caps and liners
were used to store the samples for transport to the laboratory. These
jars were cleaned before use with the rinse sequence described above.
See the explanation of the Data Table given on page
-------
A stock solution of 3% KMnO in 10$ HNO. was freshly prepared every
other day. This acidic permanganate was used to stabilize the mercury
in the samples taken from the train.^2
The following cleanup procedure was adopted to account for »11 mercury
deposited in any part of the train ahead of the silica gel. Distilled
water was used for all rinses unless otherwise noted.
1. The probe, cyclone,, and the glass parts of the filter
assembly were washed into a 1-pint jar containing
25 ml of the 3% KMnO, solution. The total volume of
this solution was measured in the laboratory prior to
analysis.
2. The filter (previously weighed) was placed in a dis-
posable plastic petri dish and marked with the run
number.
3» The inside portion of the first impinger (A position)
and the right angle connector leading into it were
rinsed into the contents of the impinger "shell". When
this impinger had originally contained distilled water,
KMnOi was added to oxidize the S02 and mercury in
solution. This was found to be necessary after initial
attempts to analyze the SO^ saturated water for mercury
gave highly erratic results from the evolution of S02
when the mercury was aerated after reduction. The
volume of this solution was measured with a graduate
and it was transferred to a one-pint bottle.
U. Each "empty" impinger in the train (and the connector
leading to it) was rinsed into a six-ounce jar con-
taining 25 ml of the KMnOi solution. The total volume
of this rinse was measured and recorded. This step
was found to be necessary when it was discovered that
moisture condensed in an empty impinger often contained
appreciable amounts of mercury, particularly if that
impinger was ahead of the first amalgamator.
12R.V. Coyne and J.A. Collins, Anal. Chem. 14*, p. 1093 (1972).
-------
U2
5. Each amalgamator case and its leading connector were
rinsed into separate jars, each containing 25 ml of
KMnO, solution as outlined in Step lu
6. Each amalgamator was fired into 50 ml of the 3$ KMnOi
solution. The apparatus used for firing the amalgam-
ators into permanganate is shown schematically in
Figure 15. The lower portion of the bubbler consists
of a closed tube of about 100 ml capacity with a stand-
ard taper fitting which matches a female taper on the
body of the bubbler assembly. These tubes were used
as interchangeable sample holders for the KMnOu
solutions. The amalgamator was centered in the coil
of the induction furnace and connected to the nitrogen
supply and the bubbler with two female ball-joint adapters
and two clamps. The nitrogen flow was set at 0.5 liters
per minute. The induction furnace was equipped with a
variac to control the energy transmitted to the coil.
Firing was commenced at a setting of 60J6 and increased
5$ each minute until the gold was glowing. This was done
to avoid a large "spike" of mercury into the bubbler, and
to insure complete firing of the gold. Laboratory tests
on mercury aerated from aqueous solutions into UO ml of
3% KMnO, showed no bypass of the permanganate for 100 ng
of mercury aerated at an air-flow rate of one liter per
minute. After the firing of each amalgamator, the
sample tube was detached, the drops of permanganate
clinging to the bubbler tube were rinsed into it and
then the contents of the tube was rinsed quantitatively
into a properly labelled six-ounce wide-mouthed jar.
The amalgamators were fired in reverse order (i.e., A£,
A, , A,., A , AO to minimize contamination of successive
samples by drop carry-over. After firing, the amalgama-
tors were assembled into the train to be used for the
next run of the day. After the series of amalgamators
had been fired, the bubbler apparatus and all of the
sample tubes were cleaned, using the rinse procedure
described above. It was found necessary to replace the
tygon tubing connecting the amalgamator to the bubbler
for each run in order to avoid contamination of succeed-
ing samples by mercury adsorbed and then desorbed by the
tygon. This tygon tubing was kept short (2-?g inches)
to minimize mercury loss.
Upon receipt in the laboratory, each sample (from an
amalgamator) was- diluted to 100 ml in a volumetric flask
just prior to analysis.
-------
Ampoule Stopper
3-way Stopcock
28/15 Ball-joint
Amalgamator
Induction Furnace
To Vent
Bubbler
\S KMnO4 Solution
Nitrogen
Figure l£. Apparatus used for Firing the Amalgamators*
-------
7. The volume(s) of the KMnO, backup solution(s) (where these
were used to back up the gold train) was measured with a
graduate and the solution was transferred to a one-pint
jar. Any permanganate stain remaining on the impinger was
removed with a few drops of 10% hydroxylamine hydrochloride
followed by a rinse with distilled water. The volume of these
rinsings was measured and they were added to the perman-
ganate in the sample jar.
8. A $0 ml blank of the permanganate solution was taken each
day, placed in a six-ounce jar, and sent to the laboratory
with the samples.
A field, record was kept of all data recorded during the run and of each
sample taken for analysis from the train. At the end of each day the
sample jars were packed in cartons and transported to the laboratory
for analysis.
li- Laboratory Analytical Procedure
Upon receipt of the samples by the laboratory, the volume of the probe
and cyclone washings were measured and recorded along with the total
volume of each of the other solutions, with the exception of those
obtained from firing the amalgamators. The latter were each diluted to
100 ml in a volumetric flask. A suitable aliquot of each of these
samples was then withdrawn by pipette and analyzed for mercury by the
direct aeration procedure. The filters were analyzed by digestion
with nitric acid, followed by analysis using the direct aeration tech-
nique. The procedures used are those developed under "Laboratory
Investigations" and are described in detail in Appendix II.
-------
RESULTS AM) DISCUSSION
1- The Data Table
The data for each of the samples taken during this study is listed in
Appendix I. Because of the large number of items associated with each
run, the table is divided into Parts A and B on separate pages. In
general, Part A contains the data taken in the field and Part B con-
tains the data resulting from the laboratory analysis of the samples.
Where there is more than one entry for a run in a column of Part A,
the entry represents data taken every five minutes of a 10- or l£-
mimite run. In Part B, under the general heading of "Train Configu-
ration" there are four horizontal lines of data across Columns A
through H. The letter of the column heading represents the position
of the impinger or amalgamator in the train sequence, counting from
the filter (position "A" is the one immediately following the filter,
"B" is next, etc.).
The topmost of the four horizontal data lines for each run is a code
describing the contents of the impinger or amalgamator in that position:
H - Distilled water, 2$0 ml
E = Empty impinger
A = 20 grams of gold chips in each amalgamator of
the series unless another amount is indicated.
K = 250 ml of saturated KMnOi solution unless
some other quantity is indicated.
ml 12 SnClg in 2-l£ HC1
-------
U6
HS = 250 ml distilled water used for the run, then
13 ml 2.0$ SnCl2 in $0% HC1 were added to the
impinger and ambient air drawn through at that
point for about 3-5 minutes in an attempt to
aerate the mercury in that impinger onto the
following amalgamator.
The second of the four horizontal lines for each run gives the total
number of micrograms of mercury found in that impinger or amalgamator.
Under this is given in parentheses the micrograms of mercury found
in the amalgamator case and connector washings. On the fourth line
the distribution of mercury among the amalgamators is given on a
percentage basis. This figure is useful for comparison of the perform-
ances of the amalgamator trains used in the various runs.
Under the column headed "Total Hg", the total mercury found in the entire
train is given. Under this value is listed the percentage of that total
found in all parts of the train except the permanganate backup solution.
This gives a figure which corresponds to an experimentally determined
percentage of total mercury recovery for the system, as it would be used
for longer runs, without KMnOi backups. It also gives a useful indication
of the efficiency of the amalgamator train when comparing runs. Under
the column headed "Filter", the topmost figure for each run gives the
number of micrograms of mercury found in the filter for that run. Under
this figure the total weight increase of the filter (weight of particul-
ates collected on the filter) is given in grams.
2- Sampling Train Configuration
The initial runs (1 - 16) of the field study were performed to determine
the optimum combination of amalgamators, empty impingers, and "scrubber"
(a liquid-filled impinger ahead of the amalgamator train to help remove
-------
U7
moisture, sulfuric acid mist, etc,). The train configuration was
varied for eacti run and each combination was generally done in
duplicate and in random order. Sampling rate was ^0.5 CFM.
Runs 7, 11, 12, and 13 were performed without any scrubber solution
ahead of the amalgamator train. The configuration used and the
relative distribution of the mercury found in the amalgamators is
given below in Table 2.
TABLE 2
Sampling Train Configuration
Run Configuration Hg Distribution on
13 AAAEEK 28.1 32.2
12 EAAAEK 81.5 10.9
7 EEAAAK U5.5 U8.2
11 '* 86.9 7.5
Amalgamators
39.8
7.6
6.2
5.6
The results suggested that maximum mercury pickup by the first amal-
gamator would be obtained in position "B" or nCn.
3- Initial Scrubber Solution
Runs 1, 2, U, 5, and 6 were performed using 250 ml of distilled water
for a scrubber in position "A", with the amalgamator train starting at
position "Brt. Runs 8 and 1U were performed in the same manner but with
-------
an empty impinger in position "B" and the amalgamator train starting
at Position "C". The results from these runs are given in Table 3.
TABLE 3
Distilled Water as the Scrubber
Run
1*
2*
u*
5*
6
8
1U
C onf igur ation
HAAAK
it
n
II
II
HEAAAK
it
Hg Distribution on
57.U
96.7
2li.li
U7.0
72.6
68.1
3U.1
16.9
2.U
61.7
33.0
8.1
31.3
1U.6
Amalgamators
25.6
1.0
13.0
20.0
19.2
0.5
51.2
*
The pump ran very erratically during these runs.
These results suggested that, within the limits of the reproducibility
obtained there was no obvious advantage for either of these configura-
tions over the other in terms of train efficiency; both produced at
least one very good run (e.g., Runs 2,8) and some poor runs (5, lU).
The cause of the inconsistent results obtained with these and other of
the earlier runs was later discovered to be contamination of the gold
chips caused by some substance evolved from the quartz wool plugs
during the pre-firing of the amalgamators. This problem and its solu-
tion is discussed in more detail in Section £, Critical Parameters.
It was noted that a significant amount of condensate was collected, in
the "Brt position, while the "C" position showed very little condensate.
This was observed whether Position A contained a scrubber or an empty
impinger. It was therefore decided to adopt the practice of starting
-------
the amalgamator train at the "C» position to minimize condensation in
the amalgamators. The scrubber solution in the "A" position was found
to collect a significant amount of mercury. Some mercury was also
found in the condensate collected in the empty "B" impinger. It was
decided to try a 1$ solution of SnCl2 as the scrubber solution to see
if this would minimize the retention of mercury in the scrubber. Two
methods were tried to accomplish this. In the first method, the im-
pinger in the "A" position was sijnply filled with 250 ml of 1% SnCl2
in 2-5$ HC1. In the second method, 250 ml of distilled water were
used in the "A" impinger and at the end of the run about 13 ml of a
2C$ SnCl2 in 50$ HC1 solution were added to the inlet of the "A" im-
pinger and about 3-5 CF of air was drawn through the train at that
point to aerate the reduced mercury onto the first amalgamator. The
results of these runs are given in Table U.
TABLE U
Stannous Chloride as the Scrubber
Run Configuration
3 SAAAK
9 SEAAAK
15
10 (HS)AAAK
16 »
Hg Distribution on
' - 85.0
67.6
23.6
U.8
21.6
(*)
10.2
Hi. 7
U3.5
9.6
12.8
Amalgamators
U.8
17.6
33.1
85.6
65.5
When the SnCl2 solutions from the "A" position were taken for analysis,
they were found to require a large addition of KMnOi for oxidation of
the contents, increasing the blank correction required. After addition
-------
50
of the required amount of permanganate, some precipitate was noted in
the solution and the analysis of these solutions showed poorer reprodu-
cibility than analyses obtained from the distilled water scrubbers.
The aeration attempt did not prove to be successful; the "A" solutions
of Runs 10 and 16 were found to contain mercury even after the aeration
step*
On the basis of these results it was decided that the use of SnClg as
the scrubber offers no advantage and may present a disadvantage over
the use of distilled water. The results obtained on addition of the
followed by aeration suggested the possibility of mercury being
washed through the train by SnCl vapors or droplets carried from the
first impinger. It was therefore decided to use only distilled water
as the scrubber for the remainder of this work.
U- The Use of KMnO^ Solutions
For most of the runs in this study a solution of KMnO. in 2% HNCU was
used in one or two impingers as a back-up system to catch any mercury
going through the amalgamator train. At the high levels of SO- en-
countered, a saturated solution of KMnOi was completely decolorized in
U - 6 minutes, depending on the sampling rate, SO- concentration, etc.
It was found to be most practical to prepare the backup impingers
I
using KMnOi as follows. The HNO^ and distilled water was added to the
impinger "shell" (250 ml for sampling rates of 0.5 CFM or less; 150 ml
for sampling rates above 0.5 CUM) and then sufficient solid KMnOj was
-------
51
weighed out and added to the water (.about 3-5 grams per CF of sample
to be taken). The resulting solution contained excess KMnQ. crystals,
but these dissolved as the solution was reduced during sampling.
Several problems were encountered with the use of KMnO. for back-up
solutions.
1. It was apparent that more than two of the permanganate
backups would be required to catch all of the mercury
in those cases where a substantial quantity of mercury
was passing through the amalgamator train. For example,
Runs 1*7-62 were made using two KMnOi back-up solutions
in series. In these runs some mercury was usually
found in the second permanganate solution, although
normally less than the amount found in the first one.
Eight impingers and amalgamator units were already being
used in the train and the pressure drop in the train
was great enough so that the pump was only drawing about
0.7 CYK at maximum effort (pump vacuum at 25 inch Hg).
2. Since the permanganate backups were just before the
silica gel, carryover from these solutions into the
silica gel often completely ruined it in about £-10
minutes. At flow rates above 0.5 CFM the volume in the
permanganate impingers had to be reduced to 150 ml in
order to avoid serious loss from violent "bumping".
The heat of reaction as the KMrD. was reduced by the S02
(the permanganate impingers became quite warm to the
touch) aggravated the carryover problem and caused im-
pinger temperatures to go as high as ll£°F after 15
minutes of sampling.
3» When longer sampling times of 10rl5 minutes and/or
higher sampling rates called for large amounts of KMnQ^
(50-80 grams) in 150 ml of water, the solution resembled
a thick paste after the run. The efficiency of such a
"solution" is very questionable.
In spite of these problems encountered with the permanganate solutions,
it was felt that they could be made to yield some useful information
for our purposes. The use of more than two permanganate backups
-------
52
was not felt to be practical at this point. Instead, it was decided
to use the data obtained from one or two permanganate backups as a
check on the performance of the amalgamator train, with an understand-
ing of its limitation as pointed out above. In other words, it was
decided that the highest priority should be to improve the performance
of the amalgamator train. If the amalgamation train could be made
efficient enough, one or two permanganate backups would prove the point;
if the amalgamators could not be made efficient enough, neither system
would be practical. Subsequent results vindicated this approach to
the problem.
5- Critical Parameters
Using the optimum train configuration, HEAAAK (determined as described
above), a series of five-minute runs was performed at various sampling
rates in the range of 0.211 to 0.833 CM to see if the sampling rate
had any effect on the collection efficiency of the amalgamators (Runs
17-30). The results were inconsistent. Some runs where the amalgam-
ator sequence showed an orderly progression in the percentage distri-
bution of mercury also showed a significant percentage of the mercury
passing through the amalgamator train, as shown by the amount of mer-
cury found in the permanganate backup solution (e.g., Runs 1? and 214.).
A useful way to compare this aspect of total train efficiency
between runs is the figure given in the Data Table, Part B, under the
column headed "Total Kg". Under the figure for the total mercury found
in the train is a figure which represents the percentage of that total
found ahead of the permanganate solutions.
-------
$3
Some runs showed a recovery ahead of the permanganate of 9%% or more,
but the distribution of mercury among the amalgamators appeared to be
nearly randomn ( e.g., Runs 25, 27). Neither the distribution of mercury
in the amalgamators nor the percentage of mercury found ahead of the
permanganate backups showed a clear relationship to sampling rate.
Runs 31-36 were performed using the same train configuration as above,
the sampling times being varied from one to twenty minutes at a
sampling rate of approximately 0.5 CM. The results of this series
showed the same kind of inconsistency in results as above. For example,
the percentage of mercury found ahead of the permanganate averaged less
than that found in Runs 17-30, although both shorter and longer sampling
times were used. In this series of runs, only about 60% of the mercury
was recovered ahead of the permanganate solutions. In an attempt to
improve on this figure several variations in the system were tried.
Runs 37-liO were made using 33 grams of gold in each amalgamator (an
increase of 6$% in the total amount of gold in the train for Runs 37
and 38). In Runs 39 and 1*0, the amalgamator train was moved back one
positionj in Run UO and extra scrubber was inserted to give the sequence:
Run 39 HEEAAK (66 gm gold total)
Run UO HHEAAK (66 gm gold total)
None of these variations produced any improvement in the performance
of the amalgamator train.
-------
5U
Runs Ul-£l were performed using five amalgamators in series, starting
from the filter, each containing 20 grams of gold. It was hoped that
distribution data from the increased amount of total gold divided
among the five amalgamators might provide an insight into the cause of
the inconsistent results obtained up to this point. Sampling times of
five to thirty minutes and sampling rates of about 0.3 to 0.7 GEM were
employed. Inspection of mercury distribution data for these runs
showed that the expected orderly progression of mercury concentration
through the amalgamator train was largely absent. The percentage of
mercury recovered ahead of the permanganate ranged from 3356 to 93%•
Since previous work had given better results than this, using less
gold, it was tentatively assumed that the gold was being effectively
"poisoned", either by some substance from the stack, or by some step
in the handling procedure. The lack of correlation between sampling
time or rate, and the efficiency of the train discounted the former
possibility, so we turned our attention to the handling procedure being
used. To see if the trouble was coming from the pre-firing step,
Run U9 was performed using amalgamators prepared in the same manner as
before,, except that the pre-firing step was eliminated. The gold was
placed in the amalgamators directly from the firing crucibles after
cooling. This simple modification proved to be the key to the 'problem:
the first amalgamator of Run U9 picked up 92$ of the mercury found in
the amalgamator train. This was a much higher figure than any ob-
tained in a previous run. Runs $0 and !?1 were performed in the manner
previously used for the second and third runs of any day} the
-------
55
amalgamators were placed in the train after the normal firing procedure
used for the previous run (first run of the day). The results showed
that firing the amalgamator with the quartz wool plug in place causes
a progressive decrease in that amalgamator's ability to remove mercury
from the sample stream. To make sure of this point, Runs 52 and 53
were performed using gold fired overnight in tne oven and not pre-fired.
Run 5U was performed using the same gold, amalgamators, and quartz wool
plug as Run 53. The pertinent data from Runs U9-5U is shown in Table 5.
TABLE 5
The Effect of Firing the Amalgamators in the Presence of the
Quartz Wool Plug
Run
k9
50
51
52
53
5k
Fired with
Quartz Wool
Plug
in Place
No
Yes
Yes
No
No
Yes
Percentage of Total
Amalgamator Mercury
Found on First
Amalgamator
92.2
6U.O
U7.9
9U.U
9U.8
Ul.6
Percentage of
Total Mercury
Found Ahead of
Permanganate
52.6
61u2
93.6
95.8
1*5.7
-------
From these results and the results of subsequent work, the following
conclusions were drawn:
1. The heating of the quartz wool plug during firing of the
gold causes the plug to give off some substance which
collects on the surface of the gold (presumably while the
gold is cooling) partially coating it and thus decreasing
the effective surface area available for amalgamation of
mercury from a flowing gas stream.
2. In the laboratory, a piece of the quartz wool was weighed,
fired for two days in a crucible in the refractory oven
and then reweighed. The results were:
weight before firing 0.2l;22 g
weight after firing 0.2U08 g
weight loss 0.001U g
percentage weight loss = 0.58$
Only a small amount of this volatile material is necessary
to desensitize the gold.
3» The condition of the gold, (cleanness of the surface) is
probably the single most important factor affecting the
efficiency of the amalgamators.
Firing in an oven seems to be the best way to clean the
gold before use and between uses.
U. Since the condition of the gold was found to be such a
large factor in the performance of the amalgamation train,
it was obvious that the previous study of the dependence
(if any) of collection efficiency on sampling rate and
sampling time should be at least partially repeated using
only freshly (oven) fired clean gold*
-------
57
6- Collection Efficiency of the Amalgamators
The collection efficiency of tne amalgamator train was studied as a
function of sampling rate, sampling time, and the total amount of mercury
collected.. This was done for a train of three amalgamators containing
20 grams of gold each, and for trains of four and five amalgamators,
each containing 30 grams of gold. Clean gold, not fired in the presence
of the quartz wool plug, was used for these studies. The collection effic-
iencies of Euns 55-72 are presented in Table 6.
a. Theory
Consider a series of identical amalgamators (tne same size and shape,
each containing the same amount of gold, the same distribution of
particle size, void fraction, etc.). If we assume that the amount of
gold is much larger than the amount of mercury to be collected (i.e.,
the accumulation of mercury on an amalgamator during sampling does not
alter its collection properties; that it is operating well below
capacity), a relationship can be derived between the ratio of mercury
found in any two adjacent amalgamators and the collection efficiency
i
of one or any total number of amalgamators.
Let r = the fraction of mercury entering an amalgamator
which is trapped by the gold. This can be called
the "trapped fraction" of a single amalgamator.
then (1 - r) = the fraction of mercury passing through that amal-
gamator which is not trapped by the gold.. This
can be called the "escape fraction" of a single
amalgamator.
If n amalgamators are connected in series, the fraction of the total
mercury passing through the nth amalgamator, fe n, is given by:
-------
TABLE 6
58
Collection Efficiencies
Run
No.
55
56
57
58
59
60
61
62
63
6U
65
66
67
68
69
70
71
72
Metered
Gas Volume
CF
1.533
3.609
3.1U6
- U.899
7.2UO
6.861
10.738
6.UOU
3.156
U.897
2.3H
7.830
8.838
10.U06
3.816
U.7U2
U.215
6.923
Percent of Amalgamated Mercury
in each Position
C
88.6
92.3
90.7
93.3
91.3
67.6
59.3
7U.U
76.1
86.7
82.1
98.7
97.1
89.0
8U.8
79.1
61.5
79.U
D
5.7
5.7
7.7
5.U
5.9
25
31.8
21.U
1.1
3.2
0.0
0.2
1.0
10.3
13.8
16.5
9.5
12.3
E
5.7
2.0
1.6
1.2
2.8
7.U
8.9
U.2
0.7
3.5
0.0
0.1
0.6
o.h
0.6
2.U
10.5
2.8
F
K
K
K
K
K
K
K
K
o.h
3.U
3.1
0.0
0.6
0.3
0.9
1.9
10.7
3.0
G
K
K
K
K
K
K
K
K
21.6
3.2
m.8
1.0
0.7-
SG
K
K
7.8
2.5
Percent of Mercury
Recovered before
Permanganate Backup
96.9
96.6
99.0
97.6
9U.5
95.1
91.1
89.1
99.7
#
98.8
•«•
*
*
100.0
98.3
*
#
K - Acidic Permanganate Scrubber
SG - Silica Gel
•«• - Train did not contain an acidic
permanganate scrubber
-------
f = (1 - r)n (1)
S jli
The fraction of the total mercury trapped by the nth amalgamator, f. ,
is equal to the fraction passing through the (n-l)th amalgamator times r:
f - r(l - r)11"1 (2)
T»f U
The expansion of terms (1) and (2) for each amalgamator is given in
Table 7.
TABLE 7
Theoretical Distribution of Mercury in a Series of Amalgamators
AmalgamatorFraction of TotalFraction of Total
No. Trapped by Amal. Through Amal.
f f
t,n e,n
1 r 1-r
2 r-r2 l-2r + r2
3 r-2r2 +r3 l-3r +3r2 -r3
U r -3r2 +3r3 -r^ 1 -Ur +6r2 -Ur3 +r4
5 r -Ur2 •|-6r3 -Ur^ -t-r^ 1 -^r +10r2 -10r3
Similarly to Equation (2), the fraction of the total mercury trapped
by the (n -I)**1 amalgamator, f+ n_i> is equal to the fraction passing
through the (n -2)nd amalgamator times r:
f. = r(l-r)n-2 (3)
t,n-l
If T = the total n-g of mercury passing through the train, the ^ig of
-------
6o
mercury trapped by the n1 amalgamator, t , is given by
Tf
11-1
t = T r(l -r)11- (5)
n
and the ng of mercury trapped by the (n -l)th amalgamator is given by
W T ft,n-l
Vi = T r^ -r>n~2
Dividing (5) by (7) we obtain:
tn T r(l -r)11-1
T
" (1 -r) (8)
Thus, the "escape fraction", (1 -r), can theoretically be obtained
directly from the ratio of total mercury found in any two adjacent
amalgamators. In actual practice, however, where the "escape fraction"
is low, only the first and second amalgamators in the series show
enough mercury to give a value of t^ and t? with sufficient relative
accuracy to calculate a meaningful value for (1 -r) . For this reason,
calculations of tjj/tj^ in this report have been confined to the first
two amalgamators.
The fraction of the total mercury trapped by the whole series of n
amalgamators, F. , is given by the sum
Ft,n = ft,l +ft,2 +ft,3
-------
61
Since F + f si (10)
t,n e,n v '
Substituting (1) into (10) we obtain:
Ft,n - 1 - (1 -r)n (11)
Equation (11) can be used to calculate:
A. The maximum "escape fraction" permissible to
achieve a given percentage of total mercury
recovery for any number of amalgamators.
B. The percentage of total mercury recovery for an
experimentally obtained "escape fraction" when
using any given number of amalgamators*
C. The number of amalgamators necessary to obtain
a given percentage total recovery if the
"escape fraction" for the amalgamators is
known.
For example, if we wish to obtain a 95% total recovery
using three amalgamators, from equation (11),
We must obtain a value for t2A1 of less than 0.368
in order to achieve 95% recovery using three amalgamat-
ors in the train.
The maximum t2/t. permissible to achieve a 95% total mercury recovery
for any number of amalgamators is given in Table 8.
-------
62
TABLE 8
»
Maximum Escape Fraction for 95£ Recovery
No. of Amalgamators Maximum t/t^ for 9$% Recovery
2 0.22U
3 0.368
h O.U73
5 0.5U9
b. 20 grains of Gold per Amalgamator
For Runs 55-62, the train sequence was HEAAAKK with 20 grams of gold in
each amalgamator. Sampling times of 5, 10, and 15 minutes, and sampling
rates from 0.30? to 0.722 CFM were used. Since Runs U9, 52, and 53 were
also performed with 20 grams of gold not fired in the amalgamator before
use, the data from these runs is also included here. All of these runs
were found to give much better and more consistently good results than
those obtained previously. The amount of mercury found showed a pro-
gressive and orderly decrease through the series of amalgamators (see
Table 6)5 the first one retaining about 90/5 of the total "amalgamator
mercury11 and the last one about 1-6^. The percentage of mercury recovered
ahead of the permanganate backup solution showed a corresponding improve-
ment (the median value was
In order to discover whether the collection efficiency of the amalgam-
ators showed any dependence on the sampling fate, the ratio of t2A1
was calculated for each of these runs and plotted against the sampling
-------
63
rate, as shown in Figure 16. The resulting curve showed almost no
dependence of tne collection efficiency on sampling rate witnin the
range of sampling rates studied.
The ratio tg/t was also plotted against t« + t, as shown in Figure 17,
to see if tnere was any decrease in collection efficiency as the amount
of mercury collected increased (saturation effect). This plot showed
no dependence of collection efficiency on the micrograms of mercury
collected. The dotted lines shown on Figures 16 and 17 represent a
tp/t-, value of 0.368, corresponding to a total amalgamator train
efficiency of 9$% for 3 amalgamators.
Figures 16 and 17 both show all values of t2/t-, clustered below
t,,/t = 0.1, except for Runs 60-62, which are noticeably higher. The
discrepancy of these runs may perhaps be explained by reference to
some observations made while these samples were being taken and then
fired. During the cleanup of Run 60, a considerable quantity of what
appeared to be a black, tarry substance was found in the cyclone and
filter assembly. This substance was not soluble in water but was sol-
uble in acetone. In firing each of the amalgamators from this run, a
quantity of smoke was evolved which condensed on the portion of the
amalgamator tube above the gold as an oily film. On Run 61, a consid-
erable quantity of the black tarry substance was found condensed in
the first amalgamator tube after it had been fired and the quartz wool
plug was quite dark looking. The filters from Runs 60-62 were also
much darker in appearance than usual. These observations suggest that
-------
0.6-
0.4-
Tram:HEAAAKK
20g. of Gold
o 5min.
A lOmin.
0 15m in.
"run 6(J
o
run 61
(tar on 1)
0.2-
(oily smoke)
run 62
0.2
0.4
Sampling Rate (CFM)
0.6
Figure 16. Collection Efficiency vs. Sampling Rate for 20 Grains of Gold per Amalgamator,,
-------
Train: HEAAAKK
to/t 20 g. of Gold
o 5min.
AlOmin.
A DlSmin.
O.o-
D
run 61
(tar on 1)
0.4-
run 60
4 (oily smoke)
run 62
0.2-
' '—S l !
Total ug Hg on 1st & 2nd amalgamator
Figure 17. Collection Efficiency vs. Total Mercury Collected for 20 Grains of Gold per Amalgamatorc
ON
vn
-------
66
some abnormal variation in the ore being roasted, or a malfunction in
some part of the equipment at the plant may have been responsible for
releasing some high boiling organic substance into the gas stream where
it was collected by the sampling unit and a portion of it was deposited
on the gold. This seems the most likely explanation, as the first
amalgamator in each series seems to be the one most affected, the ty^
ratio in each case being less than t2/t, for the same run. In this
connection it is interesting to note that the percentage of mercury
recovered ahead of tne permanganate backups did not decrease very much
for these three runs, the figures for Runs 60, 61, and 62 being 95*1.%,
91.1$, and 89.1$, respectively.
These results point out once again the fundamental importance of the
condition of the gold, and suggest that it would be wise to employ
enough gold in the amalgamator train to provide a generous reserve
capacity, as the deposition of any oil or tar on the gold (which may
be inadvertent or unavoidable in an actual sampling situation) will
greatly decrease its collection efficiency. Concluding that this
approach would prove to be the best one to the problem of obtaining
consistently high collection efficiency while sampling under actual
field conditions, we made a series (63-72) using what was considered
to be our "optimum train."
c. 30 grams of Gold per Amalgamator
Thirty grams of the gold chips used in this study make a column
3.2 - 3.7 cm high in the amalgamator tube. This was about the tallest
-------
67
column of chips which could be uniformly heated by the present coil
in the induction furnace (other sized coils can be used in this furnace
if desired). Since a total of only l6u grams of gold was available for
this study, we made a series of runs (63 - 72) using a train of either
four or five amalgamators with thirty grams of gold in each one. Of
this series, Runs 63, 6£, 69, and 70 also contained a permanganate
backup solution to check the efficiency of the amalgamator train.
Sampling times of S, 10, aJ«i 15 minutes were used and the sampling rate
was varied from 0.281 to 0.763 CFM. Run 72 was a 15 minute run taken
under isokinetic conditions. These runs showed a large percentage of
the total "amalgamator mercury" on the first amalgamator and an orderly
decrease in mercury through the train. The total percentage of mercury
collected ahead of the permanganate backups was very high, ranging from
98 to 100/5. The ratio t2/tn vs. sampling rate is shown in Figure 18.
From this experimental data, there does not seem to be any dependence
of collection efficiency on the flow rate. The ratio t2/t^ is also
plotted against t2 + t-j_ in Figure 19. Again, there does not appear to
1
be any dependence of collection efficiency on the total amount of mer-
cury collected, at least within the range studied. The collection effi-
ciency also appears to be independent of the sampling time used, and
our experience with these and earlier runs (e.g., Runs 36, 38, U6) does
not suggest that 15" minutes is the maximum feasible sampling time.
The data from Runs 63-72 shown in Figures 18 and 19 show the same
kind of clustering referred to above. The values of t2/t-L for
-------
0.4
Runs 63-72
30g. of Gold
o 5min.
AlOmin.
DlSmin.
0.2-
Q
isokinetic
run
D ©
1 ] v/ , w | ( |
0.4 0.6 0.8
Sampling Rate (CFM)
Figure 180 Collection Efficiency vs. Sampling Rate for 30 Grams of Gold per Amalgamator,!
-------
°'4-l Runs 63-72
3Og. of Gold
O 5min.
AlOmin.
0.3
0.2
0.1
QlSmin.
D w D
isokinetic
run
A
C
T " 1 r
10 20 30 40
Total yg Hg on 1st & 2nd amalgamators
Figure 19. Collection Efficiency vs0 Total Mercury Collected for 30 Grams of Gold per Amalgamator.
-------
70
Runs 63-67 are clustered below 0.05>, while t2/t-L for Runs 68-72 are all
in the range of 0.11 to 0.21. All values were, however, well under
the to/t^ ratio of O.lj.7 corresponding to 95% efficiency for the train
(see Table 8).
Run 68 was taken at about 1:30 pm on December 6, 1972. The ASARCO
plant had been having some trouble with their acid recovery plant,
which had started up around noon after having been shut down in the
morning. The roaster operator on duty informed us at that time that
one of the two electrostatic precipitators had not been working for
some time and that the other one was not working properly. This could
explain the slightly lower collection efficiency obtained on this and
the following runs.
7- Sources of Error
The errors which could be encountered in the application of this method
to an actual sampling situation might be broadly classified as sampling
errors or analytical errors.
a. Sampling Errors
The most likely sampling error, in our opinion, is the failure to col-
lect all of the mercury in the sample stream due to insufficiently clean
gold. Contamination of the gold from the gas stream being sampled is,
of course, minimized by the use of the normal cyclone and filter assemb-
ly. The wet scrubber solution next to the filter followed by an
empty impinger is also an aid here. In our experience these items
-------
71
undoubtedly help but do not eliminate the problem of gold contamina-
tion during sampling. Our most successful approach has been to use
enough gold in enough amalgamators to provide a reserve capacity in
case such contamination does occur. Examination of the data from the
amalgamator train should enable one to tell if such contamination has
indeed occurred, and allow at least a rough estimate of its severity
to be made. 3y using five amalgamators in the series, each containing
30 grams of gold, we were able to obtain very high train efficiencies,
in spite of a certain amount of gold contamination (at least for 1$
minutes of sampling time). We do not know how our sample source com-
pares to others with respect to such contamination problems.
b. Analytical Errors
Special precautions should be taken to avoid mercury contamination of
the glassware used. The normal precautions of good analytical tech-
niques apply here. Reagent blanks and permanganate blanks should be
periodically checked and appropriate records kept. Vie used one lot of
KMnO. for our work. Careful determination of the mercury blank for the
lot used in this study gave a value of 0.012' tig of mercury per gram of
KMnO. and this value was used as a correction factor for the permanganate
backup solution analyses. In addition, daily blanks were taken from the
working solution of 3% KMnO, , 10£ HNO used for preserving the samples,
and these blank corrections applied to the analytical results obtained.
For earlier runs of this study, a single piece of tygon tubing, about
18 inches long, was used to connect the amalgamator to the bubbler
-------
72
while firing the amalgamators. This piece of tubing was washed out
along with the bubbler between runs. On Runs 63 and 6£ some contam-
ination of the last amalgamator was suspected due to the unexpected and
otherwise unexplainable amount of mercury found in the last position.
It was suspected that the tygon tubing was adsorbing mercury when high
concentrations were present in the nitrogen stream (from firing the
first amalgamator of the previous run) and desorbing it again during
subsequent firings when heated nitrogen (containing little or no
mercury) was flowing through it. The washing procedure was not clean-
ing it thoroughly. On Runs 67-72, the tygon was shortened to 2-5g
inches (to minimize loss by adsorption on the tygon) and replaced with
a new piece for each run (to eliminate contamination of the solution
obtained by firing the last amalgamator in the train, which was also
the first one fired in the series). There was no further evidence of
cross contamination between subsequent runs.
8- Application to Isokinetic Sampling
Isokinetic sampling is usually performed at a sampling rate of 0.5 to
0.8 CFM. Since quantitative recovery of mercury from the gas sample
was obtained at these rates, and in view of the independence of recovery
on sampling rate, there is no reason why this procedure could not be
used for isokinetic sampling. Run 72 was, in fact, taken isokinetically,
and the data sheets for this run are included in Appendix III.
-------
CONCLUSIONS
The results of this study show that the gold amalgamation technique, for
the collection of mercury from a gas stream containing a high percentage
of S02, can achieve quantitative collection efficiencies at the sampling
rates normally used for isokinetic sampling. Figures 16 and 18 show
quantitative collection independent of flow rate in tne range of 0.3 to
0.8 CFM. Isokinetic sampling is carried out in the range of 0.5 to 0.8
CFM showing that the procedure developed in this study can be used
isokinetically. One run was made under actual isokinetic conditions,
and the collection efficiency achieved was the same as the efficiencies
obtained with the same equipment and procedure under non-isokinetic
conditions. Since some mercury was always found in the probe, cyclone
and filter assembly, a sample taken for the purpose of establishing
total emission of mercury from a source should be taken isokinetically
to insure representative sampling. The standard isokinetic sampling
procedure is to sample for about five minmtes at each traverse point so
that total sampling time is about one hour. Although sampling times of
15> minutes or less were used with the train and procedure as finally
optimized in this study, none of the results obtained suggest that longer
sampling times could not have been used with equally good results.
Some of the data obtained suggests that collection efficiencies are
lowered by volatile materials which get through the cyclone and filter
and one initial wet scrubber and coat the gold. In this study a series
of five amalgamators, each containing 30 grams of gold chips, provided a
73
-------
7U
sufficient reserve capacity to compensate for this problem as it was
encountered in this stack. Since other stacks on other plants could be
significantly different with respect to the amount of such volatiles, it
might be advisable (as least initially) to have some kind of feedback on
the sampling efficiency obtained so that additional samples could be
taken if necessary before equipment and personnel have vacated the
sampling site. This could be accomplished by on-site analysis of a
portion of the samples obtained.
Some suggestions are also offered here for certain improvements in the
equipment which the authors feel may make the sampling and analysis
operations easier and more efficient, For best results from the
laboratory analyses, an all glass system for the mercury analysis is
recommended. The use of tygon tubing in the system should be minimized
or avoided: tygon has been found by the authors to absorb (and desorb)
mercury from an air stream. Other workers have also found that both
tygon and teflon absorb mercury. ;
A simpler (less expensive and easier to clean) bubbler could be designed
for use in firing the amalgamators. The presently used bubbler assembly
should be reserved for the laboratory analysis procedure, which it was
originally for.
I
The levels of SO^ encountered and the procedure of drawing it through the
pump is hard on the sampling equipment, especially the pump. It may be
Subcommittee D-£.21, Trace Element Task Group, Methodology
Subgroup, meeting of May 25>, 1972, private communication.
-------
necessary to disr.iantle and clean it after each run to insure dependable
operation. It may also be necessary to make sone parts of the equipment
from corrosion-proof materials.
A larger sampling box which could hold 8-10 impingers in the ice bath
without excessive crowding would be useful and would have the advantage
that an extra scrubber or empty impinger could be inserted ahead of the
gold train if such a change was found necessary to insure cleanness of
the gold during sampling on an especially dirty stack.
-------
APPENDIX I
TABLE 9
Part A. Field Parameters:
RUN #
DATE
TIME
1972
1
10/17
4:16
pm
2
10/19
11:38
am
3
10/19
4:04
pm
4
10/20
10:S3
am
5
10/20
2:35
pm
6
10/20
4:05
pm
7
10/23
2:18
pm
SAMPLING
TIME
minutes
5.0
7.0
7.0
10.0
12.0
15.0
5.0
BAROMETRIC
PRESSURE
in. Hg
29.48
29.48
29.48
28.84
AMBIENT
TEMP.
OF
60
41
41
41
50
48.2
73.4
STACK
TEMP.
(Ts)
op
440
460
480
460
460
460
430
£so2
STATIC
PRESS.
(Ps)
in. H20
* 0.05
* 0.32
* 0.32
+ 0.04
- 0.13
-0.06
-0.41
VELOCITY
HEAD
Caps)
in. H20
ORFICE
PRESS.
(AH)
in. H20
WINGER
TEMP.
op
60
50
55
<50
67.5
56.0
60.0
DRY GAS
TEMP.
<*m)
OF
58.5
44.0
45.5
48.0
57.5
58.0
67.5
METERED
GAS VOL
(vm)
Cu. ft.
2.152
2.856
3.014
4.136
6.379
6.022
2.585
SAMPLING
RATE
CFM
.430
.407
.430
.414
.532
.401
.517
-------
RUN #
DATE
TIKE
1972
8
10/23
4:03
Dm
9
10/24
10:44
am
10
10/24
12:03
pm
11
10/24
2:20
pm
12
10/24
3.47
pm
13
10/30
10}16
ain
14
10/30
11:49
am
SAMPLING
TIME
minutes
5.0
5.0
5.0
5.0
5.0
5.0
5.0
BAROMETRIC
PRESSURE
in. Hg
28.84
29.12
29.12
29.12
29.12
29.30
29.30
AMBIENT
TEMP.
OF
62.6
50.0
44.6
45.5
45.5
46.4
50.0
STACK
TEMP.
(T8>
op
430
460
470
480
480
480
495
%so2
6.0
6.4
STATIC
PRESS.
(Ps)
in. H20
- 0.26
- 0.67
0
+ 0.36
+ 0.20
+ 0.12
- 0.10
VELOCITY
HEAD
CflPB)
in. H20
ORFICE
PRESS.
(AH)
in. H20
0.92
0.86
MPINGER
TEMP-
Op
60
50
<50
50
50
50
50
DRY GAS
TEMP.
(Tra)
op
68.0
51.5
50.5
49.0
50.0
52.0
50.0
METERED
GAS VOL.
(vm)
Cu. ft.
2.391
2.477
2.355
2.549
2.532
2.529
2.485
SAMPLING
RATE
CFM
.478
.495
,471
.510
.506
.506
.497
-------
RUN #
DATE
TIME
1972
15
10/30
2:15
cm
16
10/30
3:45
pm
17
10/31
10:17
am
18
10/31
11:39
am
19
10/31
1:50
pm
20
10/31
3:35
pm
21
11/1
1:10
pm
SAMPLING
TIME
minutes
5.0
5.0
5.0
5.0
5.0
5.0
5.0
BAROMETRIC
PRESSURE
in. Hg
29.30
29.30
29.30
29.30
29.30
29.30
29. 04
AMBIENT
TEMP.
op
52.7
50.0
46.4
48.2
50.0
46.4
56.3
STACK
TEMP.
(Ts)
°F
490
495
520
520
370
470
500
%so2
6.4
6.6
6.5
STATIC
PRESS.
(PS)
in. H^
- 0.17
0
•* 0.77
+ 0.91
* 2.75
+ 1.50
+ 0.14
VELOCITY
HEAD
(APS)
in. H20
ORFICE
PRESS.
(*H)
in. H20
0.95
0.93
0.21
0.30
1.30
1.75
2.20
IMPINGER
TEMP*
Op
50
50
50
50
50
50
50
DRY GAS
TEMP.
(Tm)
OF
53.5
54.0
46.0
49.0
51.5
49.5
60.5
METERED
GAS VOL.
(vm)
Cu. ft.
2.409
2.561
1.292
1.867
3.073
3.526
4.033
SAMPLING
RATE
CFM
.482
.512
.258
.373
.615
.705
.807
-------
RUN #
DATE
TIME
1972
22
11/1
2:42
pm
23
11/1
4:35
cm
24
11/2
10:32
am
25
11/2
12:04
pm
26
11/2
3:35
pm
27
11/2
4:45
pm
28
11/3
11:53
am
SAMPLING
TIME
minutes
5.0
5.0
5.0
5.0
5.0
5.0
5.0
BAROMETRIC
PRESSURE
in. Hg
29.04
29.04
28.85
28.85
28.85
28.85
29.06
AMBIENT
TEMP.
op
57.2
57.2
60.8
62.6
55.4
59.9
60.8
STACK
TEMP.
(Ts)
op
500
500
490
480
510
510
500
Sso2
6.4
6.8
7.2
7.0
8.1
8.3
8.0
STATIC
PRESS.
(Ps)
in. H20
+ 0.28
0
- 0.16
- 0.16
•* 1.01
+ 1.20
- 0.36
VELOCITY
HEAD
(4PS)
in. H20
ORFICE
PRESS.
(4H)
in. H20
0.24
0.80
1.20
1.70
0.17
0.29
0.89
IMPINGER
TEMP.
Op
50
50
55
55
55
55
60
DRY GAS
TEMP.
(TJ
op
62
62
63.5
64
60
61
65
METERED
GAS VOL.
(vm)
Cu. ft.
1.465
2.495
2.982
3.499
1.054
1.551
2.620
SAMPLING
RATE
CFM
.293
.499
.596
.700
.211
.310
.524
CO
o
-------
RUN #
DATE
TIME
1972
29
11/3
2:05
pm
30
11/3
3:22
pm
31
11/8
11:10
am
32
11/8
1:36
pm
33
11/8
3:15
Dm
34
11/8
4:34
pm
35
11/9
10:25
am
SAMPLING
TIME
minutes
5.0
5.0
1.0
3.0
5.0
10.0
15.0
BAROMETRIC
PRESSURE
in. Hg
29.06
29.06
29.06
29.06
29.06
29.06
29.26
AMBIENT
TEMP.
op
57.2
53.6
44.6
47.3
46.4
47.3
44.6
STACK
TEMP.
(Ts)
op
500
510
500
500
500
510
520
#30
7.9
7.7
7.4
8.4
8.4
8.0
7.6
STATIC
PRESS.
(Ps)
in. H20
- 0.36
- 0.27
- 0.56
- 0.47
- 0.41
- 0.34
- 0.27
- 0.26
- 0.27
VELOCITY
HEAD
(4P5)
in. H20
ORFICE
PRESS.
(4H)
in. 1^0
0.56
2.60
0.90
1.00
0.71
0.83
0.78
0.90
0.88
WINGER
TEMP*
OF
60
60
<50
<50
50
50
55
72.5
72.5
DRT GAS
TEMP.
dm)
op
60.0
60.0
46.5
48.0
49.0
50.5
46.0
48.5
51.0
METERED
GAS VOL.
(vm)
Cu. ft.
2.075
4.166
0.474
1.632
2.306
5.232
7.416
SAMPLING
RATE
CFM
.415
.833
.474
.544
.461
.523
. 494
-------
RUN #
DATE
TIME
1972
36
11/9
1:00
pm
37
11/9
3:05
pm
38
11/9
4:30
pm
39
11/10
2:42
pm
40
11/10
4:42
pm
i 41
11/13
11:33
am
42
11:13
2:36
pm
SAMPLING
TIME
minutes
20.0
15.0
20.0
20.0
20.0
15.0
10.0
BAROMETRIC
PRESSURE
in. Hg
29.26
29.26
29.26
28.90
28.90
29.04
29.04
AMBIENT
TEMP.
OF
45.5
46.4
44.6
53.6
59.0
50.0
56.3
STACK
TEMP.
(Ts)
op
525
520
520
510
520
510
510
*so2
7.6
7.6
7.7
7.5
7.5
6.6
7.2
STATIC
PRESS.
(Ps)
in. H20
- 0.14
- 0.16
- 0.24
- 0.28
- 0.33
- 0.34
- 0.36
- 0.42
- 0.41
- 0.41
- 0.43
- 0.54
- 0.52
- 0.53
- 0.55
- 0.50
- 0.52
- 0.54
- 0.55
- 0.12
- 0.06
0
- 0.34
- 0.46
VELOCITY
HEAD
UPS)
in. H20
ORFICE
PRESS.
(*H)
in. H20
0.82
0.97
0.88
0.86
0.79
0.89
0.86
0.91
0.96
0.93
0.91
0.58
0.81
0.90
0.92
0.89
0.88
0.87
0.85
0.85
0.92
0.91
1.10
1.00
IMPINGER
TEMP,
op
50
75
87.5
85
65
75
85
50
67.5
80
75
65
105
100
85
<50
50
57.5
65
<50
65
75
<50
55
ERT GAS
TEMP.
(Tm)
OF
48
52
54.5
56
50.5
54.5
57
50
53
56.5
59
55
65
72.5
77
52
54.5
56.5
58.5
44
47
50.5
45.5
47.5
METERED
GAS VOL.
(vra)
Cu. ft.
10.021
7.428
10.293
9.647
10.165
7.272
6.137
SAMPLING
RATE
CFM
.501
.495
.514
.482
.508
.485
.614
oo
ro
-------
RUN #
DATE
TIME
1972
43
11/13
4:40
jjm
44
11/14
10:36
am
45
11/14
1:16
pm
46
11/14
3:25
pm
47
11/15
12:22
pm
48
11/15
3:55
pm
SAMPLING
TIME
minutes
5.0
15.0
15.0
30.0
4.38
5.0
BAROMETRIC
PRESSURE
in. Hg
29.04
28.70
28.70
28.70
29.24
29.24
AMBIENT
TEMP.
op
57.2
44.6
42.8
—
39.2
35.6
35.6
STACK
TEMP.
(Ts)
op
510
520
520
540
575
560
£SO?
7.1
7.7
7.6
8.2
8.5
8.6
STATIC
PRESS.
(P5)
in. H20
- 0.47
- 0.38
- 0.37
- 0.36
- 0.39
- 0.34
- 0.35
- 0.36
- 0.36
- 0.36
- 0.36
- 0.32
- 0.31
- 0.25
- 0.25
VELOCITY
HEAD
CflF8)
in. H20
ORFICE
PRESS.
(AH)
in. H20
.95-1.5
0.36
0.28
0.29
1.90
1.90
1.90
0.59
0.54
0.57
0.58
0.50
0.67
0.98
.85-. 90
IMPINGER
TEMP,
op
<50
<50
55
60
-no
100
85
50
75
75
65
55
55
<50
<50
DRY GAS
TEMP.
(TJ
OF
43.5
45.0
46.0
46.5
48.0
55.0
59.5
37.5
38.5
39
40
40.5
41.5
36
37.5
METERED
GAS VOL.
(vm)
Cu. ft.
2.617
4.579
10.652
11.592
2.140
2.376
SAMPLING
RATE
CFM
.523
.306
.710
.386
.488
.475
-------
RUN #
DATE
TIME
1972
49
11/16
10:45
am
50
11/16
1:45
pm
51
11/16
3:48
pm
52
11/17
10: 20
am
53
11/17
1:34
pm
54
11/17
3:35
pm
55
11/20
11:21
am
SAMPLING
TIME
minutes
5.0
5.0
5.0
5.0
5.0
5.0
5.0
BAROMETRIC
PRESSURE
in. Hg
29.23
29.23
29.23
29.16
29.16
29.16
29.25
AMBIENT
TEMP.
op
37.4
39.2
39.2
33.8
39.2
39.2
41.0
STACK
TEMP.
(Ts)
op
550
560
560
550
560
560
570
£so2
8.7
8.5
8.8
7.1
8.5
8.8
8.2
STATIC
PRESS.
(Ps)
in. H20
0
0
* 0.02
- 0.05
- 0.15
- 0.28
- 0.19
VELOCITY
HEAD
WP8)
in. H20
ORFICE
PRESS.
(4H)
in. #2®
0.85
0.9-1.5
1.0-1.5
0.98
>.87-0.95
1.1
0.43
3MPINGER
TEMP.
OF
<50
<50
<50
<50
<50
<50
<50
DRY GAS
TEMP.
(*»>
OF
36.5
41.0
44.5
36.0
38.0
40.0
39.5
METERED
GAS VOL.
Cm)
Cu. ft.
2.166
2.394
2.462
2.327
2.288
2.495
1.533
SAMPLING
RATE
CFM
.433
.479
.492
.465
.458
.499
.307
-------
RUN #
DATE
TIME
1972
56
11/20
2:15
pm
57
11/20
4:15
pm
58
11/21
10:27
am
59
11/21
1:30
pm
60
11/21
3:47
pm
61
11/22
10145
am
62*
11/22
2:30
pm
SAMPLING
TIME
minutes
5.0
5.0
10.0
15.0
15.0
15.0
10.0
BAROMETRIC
PRESSURE
in. Hg '
29.25
29.25
29.33
29.33
29.33
29.22
29.22
AMBIENT
TEMP.
op
41.0
41.0
38.3
37.4
39.2
37.4
36.5
STACK
TEMP.
(Ts)
op
570
570
565
560
560
560
560
570
570
570
550
550
550
560
560
%SQ
8.5
8.5
8.1
7.8
7.9
7.5
7.5
STATIC
PRESS.
(Ps)
in. H20
- 0.22
- 0.11
- 0.13
0
- 0.25
- 0.22
- 0.16
- 0.03
0
- 0.07
- 0.11
- 0.21
- 0.15
- 0.08
- 0.20
VELOCITY
HEAD
(dPs)
in. H20
ORFICE
PRESS.
(4H)
in. H20
2.40
1.40
0.86
0.85
0.85
0.85
O.c(4
0.85
0.85
0.63
1.90
1.90
1.90
1.30
IMPINGER
TEMP*
OF
85
<50
<50
<50
<50
<50
60
<50
50
75
<50
115
85
<50
115
DRY GAS
TEMP.
(Tm)
OF
41.0
42.5
45.0
48.0
42.5
44.0
44.5
42.5
46.5
51.0
40.5
48.0
56.0
43.5
45.5
METERED
GAS VOL.
(vm)
Cu. ft.
3.609
3.146
4.899
7.240
6.861
10.738
6.404
SAMPLING
RATE
CFM
.722
.629
.490
.482
.457
.716
.572
•Pump stopped during run and had to be repaired
-------
RUN #
DATE
TIME
1972
63
12/1
4:00
pm
64
12/4
10:46
am
65
12/4
4jl5
pm
66"
12/5
11:00
am
67
12/5
4:23
pm
68
12/6
1:23
pm
69
12/7
11:00
am
SAMPLING
TIME
minutes
5.0
10.0
5.0
15.0
15.0
15.0
5.0
BAROMETRIC
PRESSURE
in. Hg
28.80
28.96
28.96
29.08
29.08
29.30
29.68
AMBIENT
TEMP.
op
41.9
42.8
42.8
50.0
50.0
28.4
23
STACK
TEMP.
(Ts)
op
520
540
540
560
540
540
540
540
540
540
475
480
485
555
?so2
7.4
7.8
8.2
8.1
8.0
9.0
-8.2
8.4
STATIC
PRESS.
(Pa)
in. H20
«• 0.11
0.32
0.19
- 0.31
- 0.37
•* 1.40
+ 0.41
VELOCITY
HEAD
WP8)
in. H20
0.53
0.53
0.54
0.57
0.59
0.58
0.57
0.57
0.57
0.57
0.49
0.49
0.64
0.50
ORFICE
PRESS.
(AH)
in. H20
1.60
0.98
1.05
0.94
1.20
1.20
1.30
1.55
1.60
1.60
2.20
2.20
2.25
2.3-2.8
IMPINGER
TEMP.
Of
50
<50
60
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
55
DRY GAS
TEMP.
(T»)
OF
48.5
41.0
44.5
44.0
51.5
55.5
58.0
62.5
67.0
70.0
31.5
34.0
35.0
39.0
METERED
GAS VOL.
(Vm>
Cu. ft.
3,156
4.897
2.311
7.830
8.838
10.406
3.816
SAMPLING
RATE
CFM
.631
.490
.462
.522
.589
.694
.763
-------
RUN #
DATE
TIME
1972
70
12/7
4:21
pm
71
12/11
11:38
am
72
12/11
4:03
pm
SAMPLING
TIME
minutes
15.0
15.0
15.0
BAROMETRIC
PRESSURE
in. Hg
29.55
29.61
29.60
AMBIENT
TEMP.
op
28.4
23.0
25.7
STACK
TEMP.
(TS)
op
570
570
570
580
580
580
580
580
580
£S02
7.4
8.2
8.2
STATIC
PRESS.
(Ps)
in. H20
+ 0.32
+ 0.13
- 0.58
VELOCITY
HEAD
WPB)
in. H20
0.54
0.55
0.54
0.54
0.53
0.53
0.51
0.51
0.52
ORFICE
PRESS.
(*H)
in. H20
0.42
0.41
0.41
0.31
0.29
0.29
1.04
1.05
1.05
BIPINGER
TEMP.
op
<50
<50
<50
<50
. <50
<50
<50
<50
<50
DRY GAS
TEMP.
(*m)
op
33.0
36.0
36.5
29.0
32.5
35.5
31.5
35.0
38.5
METERED
GAS VOL.
(vj
Cu. ft.
4.742
4.215
6.923
SAMPLING
RATE
CFM
.316
.281
.462
OD
-------
Part B. Laboratory Parameters:
RUN
#
1
2
3
4
5
6
7
PARTICULATES
PROBE AND
CYCLONE
.»g
1.56
0.28
0.09
0.15
0.24
FILTER
jag
TOTAL
-»g
A
H
H
2.8
S
3.0
H
H
H
E
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
A
2.24
57.4%
A
4.08
96.7%
A
5.02
85%
A
1.7
24.4%
A
1.6
47.0%
A
5.0
72.6%
B
C
A
0.66
16.9%
A
0.10
2.4%
A
0.60
10.2%
A
4.30
61.7%
A
1.12
33%
A
0.56
8.1%
A
2.04
45.5%
D
A
1.0
25.6%
A
<.04
<1%
A
0.28
4.8%
A
0.96
13.8%
A
0.68
20.0%
A
1.32
19.2%
A
2.16
48.23!
E
K
K
<.28
K
K
K
K
0.10
A
0.28
6.2%
F
SG
SG
SG
K
0.25
K
1.3
K
0.50
K
0.70
G
SG
SG
SG
SG
H
TOTAL
Hg
Jig
3.90
8.86
96.8%
9.18
7.30
4.85
7.72
92.2%
5.18
86.5%
Hg
CONCENTRATION
Jig/CF
1.81
3.10
3.04
1.76
0.76
1.28
2.00
^g/1
0.064
0.110
0.108
0.062
0.027
0.045
0.071
CD
CO
-------
RUN
#
8
9
10
11
12
13
14
PARTICULATES
PROBE AND
CYCLONE
Jag
0.56
0.53
FILTER
-»g
TOTAL
-»g
L
H
2.08
S
0.44
H-S
0.48
E
E
A
1.24
28.1%
H
0.77
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IK EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
E
E
E
E
A
3.0
81.5%
A
1.42
32.2%
E
C
A
2.52
68.1%
A
0.92
67.6%
A
0.28
4.8%
A
3.70
86.9%
A
0.40
10.9%
A
1.76
39.8%
A
1.40
34.1%
D
A
1.16
31.3%
A
0.20
14.7%
A
0.56
9.6%
A
0.32
7.5%
A
0.28
7.6%
E
A
0.60
14.63
E
A
0.02
0.5%
A
0.24
17.6%
A
5.00
85.6%
A
0.24
5.6%
E
E
A
2.10
51.2%
F
K
0.37
K
0.55
K
0.20
K
0.59
K
1.73
K
0.42
K
0.67
G
SG
SG
SG
SG
SG
SG
SG
H
TOTAL
Hg
jug
6.15
94.0%
2.35
76.6%
6.52
96.9%
4.85
87.8%
5.41
68.0%
5.40
92.2%
6.07
89.0%
Hg
CONCENTRATION
Jig/CF
2.57
0.95
2.77
1.90
2.14
2.14
2.44
^g/1
0.091
0.034
0.098
0.067
0.075
0.075
0.086
00
-------
RON
#
15
16
17
18
19
20
21
PARTICULATES
PROBE AND
CYCLONE
J*g
0.95
0.55
0.25
0.31
0.09
0.43
0.81
FILTER
.Mg
TOTAL
-»g
A
S
0.25
H-S
0.31
H
0.25
H
1.50
H
0.30
H
1.03
H
1.86
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
B
B
E
E
E
0,60
E
0.34
E
0.19
C
A
1.70
23.6%
A
1.52
21.6%
A
3.70
90.9%
A
0.44
30.6%
A
0.18
58.0%
A
0.80
45.5%
A
2.75
44.1%
D
A
3.15
43.5%
A
0.90
12.8%
A
0.29
7.1%
A
0.84
58.3%
A
0.07
22.6%
A
0.68
38.6%
A
3.10
49.7?!
E
A
2.40
33.1%
A
4.60
65.5%
A
0.08
2.0%
A
0.16
11.1%
A
0.06
19.4%
A
0.28
15.9%
A
0.38
6.0%
F
K
0.12
K
0.15
K
0.48
K
1.20
K
0.18
K
0.68
K
1.29
G
SG
SG
SG
SG
SG
SG
SG
H
TOTAL
Hg
Jig
8.57
98 . 6%
8.03
98.1%
5.05
90.5%
4.45
73.0%
1.48
87.8%
4.24
84.0%
10.38
87.6%
Hg
CONCENTRATION
^Jig/CF
3.56
3.14
3.91
2.38
0.48
1.20
2.57
^g/1
0.125
0.111
0.138
0.084
0.017
0.042
0.091
-------
RUN
*
22
23
24
25
26
27
28
PARTICULATES
PROBE AHD
CYCLONE
Jig
0.30
0.46
0.43
0.31
0.09
0.05
1.49
FILTER
-»g
TOTAL
Jig
A
H
0.61
H
1.39
H
2.20
H
2.08
H
0.52
H
1.58
H
2.57
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
E
0.30
E
0.36
E
0.69
B
0.36
B
0.20
E
0.42
E
1.01
C
A
1.70
78%
A
1.71
60.0%
A
3.65
93.8%
A
1.85
33.1%
A
1.28
51.6%
A
0.88
24.7%
A
0.84
46.1%
D
A
0.28
12.8%
A
0.59
20.7%
A
0.19
4.9%
A
3.50
62.5%
A
0.44
17.7%
A
1.40
39.3%
A
0.88
48.4%
E
A
0.20
9.2%
A
0.55
19.3%
A
0.05
1.3%
A
0.24
4.3%
A
0.76
30.6%
A
1.28
36.0%
A
0.10
5.5%
F
K
0.48
K
1.63
K
1.16
K
0.44
K
0.69
K
0.17
K
1.57
G
SG
SG
SG
SG
SG
SG
SG
H
TOTAL
Hg
jag
3.87
87 . 6%
6.69
75.6%
8.37
86.1%
8.78
95.0%
3.98
82.7%
5.78
97.0%
8.46
81.4%
Hg
CONCENTRATION
Jig/CF
2.64
2.68
2.81
2.51
3.78
3.73
3.23
^g/1
0.093
0.095
0.099
0.089
0.133
0.132
0.114
-------
RUN
#
29
30
31
32
33
34
35
PARTICULATES
PROBE AND
CYCLONE
.Mg
1.65
1.77
0.08
0.13
0.70
1.45
2.75
FILTER
Jag
TOTAL
-»g
A
H
1.29
H
3.09
H
0.20
H
0.10
H
1.10
H
0.05
H
1.18
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
E
0.19
E
0.34
E
0.05
E
0.17
B
0.28
E
0.15
B
0.51
C
A
3.50
85.4%
A
2.06
62.4%
A
0.10
40.0%
A
0.26
57.7%
A
0.40
50.0%
A
0.22
57.9%
A
3.55
66.8%
D
A
0.36
8.8%
A
0.74
22.4%
A
0.08
32.0%
A
0.10
22.2%
A
0.22
27.5%
A
0.10
26.3%
A
0.88
16.6%
E
A
0.24
5.8%
A
0.50
15.1%
A
0.07
28.0%
A
0.09
20.0%
A
0.18
22.5%
A
0.06
15.8%
A
0.88
16.6%
F
K
0.25
K
0.47
K
1.41
K
1.25
K
3.90
K(15g)
2.00
K(20g)
3.54
G
SG
SG
SG
SG
SG
SG
SG
H
TOTAL
Hg
•»g
7.48
96.7%
8.97
94.8%
1.99
29.1%
2.10
40.5%
6.78
42. S%
4.03
50.4%
13.29
73.4%
Hg
CONCENTRATION
.Jig/CF
3.60
2.15
4.20
1.29
2.94
0.77
1.79
>»gA
0.127
0.076
0.148
0.045
0.104
0.027
0.063
ro
-------
RUH
#
36
37
38
39
40
41
42
PARTICULATES
PROBE AND
CYCLONE
J»g
3.00
2.03
1.88
1.56
1.51
0.45
0.70
FILTER
J*g
TOTAL
Jig
A
H
2.92
H
2.63
H
4.46
H
4.58
H
2.24
A
1.36
(1.04)
15.3%
A
1.91
(0.97)
24.3%
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
E
0.33
E
0.26
E E
0.18
E
0.48
H
0.19
;GS TIP)
A
3.16
(0.04)
35 . 6%
A
1.86
(0.16)
23.6%
C
A
3.85
55.0%
A(33g)
2.25
44.2%
A(33g)
1.74
34.0%,
E(gwf
0.11
E
0.07
A
2.46
(0.02)
27.7%
A
1.91
(0.04)
24.3%
D
A
1.20
17.2%
A(33g
1.48
29.1%
A(33g)
2.04
39.8%
A(33g
2.13
49.7%
A(33g)
0.74
59.6%
A
0.96
(0.03)
10.8%
A
1.04
(0)
13.2%
E
A
1.94
27.8%
A(33g)
1.36
26.7%
A(33g)
1.34
26.2%
A(33g)
2.15
50.2%
A(33g)
0.50
40.3%
A
0.94
(0.03)
10.6%
A
1.14
(0.02)
14.5%
F
K(33g)
8.44
K(25g)
4.61
K(33g>
5.50
K(34g)
8.49
K(34g)
9.13
K(27g)
4.16
K(16g)
0.72
G
SG
SG
SG
SG
SG
SG
SG
H
TOTAL
Hg
Jttg
21.68
61.1%
14.62
68.5%
17.14
67.9%
19.50
56.5%
14.38
3fi.S*
14.65
71.6%
10.47
93.1%
Hg
CONCENTRATION
^ig/CF
2.16
1.97
1.66
2.02
1.41
2.01
1.71
^gA
0.076
0.069
0.059
0.071
0.050
0.071
0.060
*A 4 cm column of glass wool was packed in "C." This increased pressure drop in the
that the pump was at maximum vacuum throughout the run.
train so much
-------
RUN
t
43
44
45
46
47
48
PARTICULATES
PROBE AND
CYCLONE
jag
0.44
0.35
1.60
1.68
0.74
0.43
FILTER
jag
TOTAL
-Ug
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
A
A
2.04
68.9%
A
6.96
(1.55)
63.8%
A
1.13
(1.93)
11.2%
A
7.18
(3.59)
57.5%
A
3.39
(0.51)
62.1%
A
1.63
(0.04)
32.0%
B
A
0.44
14.9%
A
2.06
(0)
18.9%
A
2.48
(0.27)
24.5%
A
2.08
(0.11)
16.7%
A
0.62
(0)
11.4%
A
1.82
(0.41)
35.7%
C
A
0.20
6.8%
A
0.50
(0)
4.6%
A
3.68
(0.04)
36.4%
A
1.18
(0.05)
9.4%
A
0.66
(0)
12.1%
A
0.68
(0.05)
13.4%
D
A
0.12
4.0%
A
0.88
(0)
8.1%
A
1.38
(0)
13.6%
A
1.14
(0)
9.1%
A
0.36
(0)
6.6%
A
0.44
(0.05)
8.7%
E
A
0.16
5.4%
A
0.51
(0)
4.7%
A
1.44
(0)
14.2%
A
0.90
(0.09)
7.2%
A
0.42
(0)
7.7%
A
0.50
(0)
10.2%
F
K
2.76
K(25g)
2.80
K(55g)
7.04
K(63g)
8.88
K
9.93
K(15.5g
1.58
Q
SG
SG
SG
SG
K(3%)
3.64
K(3%)
2.98
H
SG
SG
TOTAL
Hg
jag
6.16
55.2%
15.61
82.1%
20.99
66.5%
26.88
67.0%
20.27
33.0%
10.63
57.1%
Hg
CONCENTRATION
^ig/CF
2.35
3.41
1.97
2.32
9.47
4.47
^g/1
0.083
0.120
0.070
0.082
0.334
0.158
-------
RUN
#
49
50
51
52
53
54
55
PARTICULATES
PROBE AND
CYCLONE
Jig
0.50
0.45
0.24
0.44
0.42
0.28
0.56
FILTER
Jig
0.62
(0.0364g
0.72
(0.0330i
0.82
(0.0229{
0.60
0.0639g
0.32
;0.0477g
0.58
0.0289g
0.40
0.0165g
TOTAL
Jig
1.12
)
1.17
)
1.06
)
1.04
0.74
0.86
0.96
&
A(24g)
6.10
(0.35)
92.2%
A(24g)
3.48
(0.45)
64.0%
A(24g)
3.06
(1.02)
47.9%
H
2.41
H
2.04
H
1.51
H
1.52
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IK AMALGAMATORS
B
A(24g)
0.35
(0)
5.3%
A(24g)
0.50
(0.02)
9.2%
A(24g)
1.40
(0.02)
21.9%
E
0.38
E
0.05
E
0.04
E
0.03
C
A(24g)
0.04
(0)
0.6%
A(24g)
1.20
(0)
22.0%
A(24g)
0.16
(0.02)
2.5%
A
6.40
(.02)
94.4%
A
5.20
(0)
94.8%
A
0.60
(.02)
41.6%
A
4.98
(0)
88.6%
D
A(24g)
0.08
(0)
1.2%
A(24g)
0.10
(0)
1.8%
A(24g)
1.44
(0)
22.6%
A
0.22
(0)
3.2%
A
0.12
(0)
2.2%
A
0.48
(0.09)
33.3%
A
0.32
(0.045
5.7%
E
A(24g)
0.04
(0)
0.6%
A(24g)
0.16
(0)
2.9%
A(24g)
0.32
(0)
5.0%
A
0.16
(0.32)
2.4%
A
0.16
(0)
2.9%
A
0.36
(0.01)
25%
A
0.32
(0)
5.7%
F
K(15g)
10.45
K(20g)
3.99
K(25g)
3.45
K
0.15
K(15g)
0.28
K(15g)
2.99
K
0.23
G
K(10g)
0.46
K(10g)
2.40
K(3%)
1.30
K
0.60
K
0.08
K
1.73
K
0.03
H
SG
SG
SG
SG
SG
SG
SG
TOTAL
Hg
Jig
18.99
42.5%
13.47
52.6%
13.25
64.2%
11.70
93.6%
8.67
95.8%
8.69
45.7%
8.43
96.9%
Hg
CONCENTRATION
Jig/CF
8.77
5.63
5.38
5.03
3.79
3.48
5.50
jigA
0.310
0.199
0.190
0.178
0.134
0.123
0.194
\A
-------
RUN
#
56
57
58
59
60
61
62
PARTICULATES
PROBE AND
CYCLONE
Jig
0.05
0.77
0.88
0.68
1.15
2.80
0.45
FILTER
Jig
0.56
(0.0231g
0.48
(0.0393s
0.92
(0.0972g
0.96
(0.1360g
2.48
(0.1421g
1.14
0.1590g)
0.94
(0.1756§
TOTAL
Jig
0.61
1.25
1.80
1.64
3.63
3.94
1.39
A
H
3.36
H
2.40
H
5.40
H
5.32
H
3.15
H
1.58
H
2.89
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
£
0.12
B
0.45
B
0.07
B
0.10
B
0.11
B
0.10
B
0.12
C
A
11.98
(0)
92.3%
A
10.18
(0)
90.7%
A
12.00
(0)
93.3%
A
16.80
(0.03)
91.3%
A
13.80
(0)
67.6%
A
6.63
(0)
59.3%
A
10.75
(0.03)
74.4%
D
A
0.74
(0)
5.7%
A
0.86
(0)
7.7%
A
0.70
(0)
5.4%
A
1.08
(0)
5.9%
A
5.10
(0)
25.0%
A
3.55
(0.05)
31.8%
A
3.1
(0)
21.4%
E
A
0.26
(0)
2.0%
A
0.18
(0.42)
1.6%
A
0.16
(0)
1.2%
A
0.52
(0)
2.8%
A
1.5
(0)
7.4%
A
1.00
(0.04)
8.9%
A
0.60
(0)
4.2%
F
K(25g)
0.22
K(25g)
0.13
C(30g)
0.44
K(45g)
1.15
K(50g)
1.30
K(80g)
1.08
K(68g)
1.58
G
K
0.38
K
0.03
K
0.06
K(10g)
0.33
K(15g)
D.10
K(23g)
0.57
K(20g)
0.72
H
SG
SG
SG
SG
SG
SG
SG
TOTAL
Hg
Jig
17.67
96.6%
15.90
99.0%
20.63
97.6%
26.97
94.5%
28.69
95.1%
18.54
91.1%
21.18
89.1%
Hg
CONCENTRATION
Jig/CF
4.90
5.05
4.21
3.72
4.18
1.73
3.31
^gA
0.173
0.178
0.149
0.132
0.148
0.061
0.117
-------
RUN
#
63
64
65
66
67
68
69
PARTICUIATES
PROBE AMD
CYCLONE
^g
0.33
0.34
0.09
0.32
0.66
2.17
0.22
FILTER
jag
0.24
(0.0757E
0.38
(0.0813g
3.46
(0.0469g
1.06
(0.0517g
0.52
0.0481g
0.91
0.0865g
0.30
0.0903g
TOTAL
-»g
0.57
)
0.72
i
3.55 ~
)
1.38
)
1.18
3.08
0.52
&
H
3.51
H
8.90
H
0.63
H
14.49
H
16.45
H
3.97
H
5.52
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
E
0.13
E
0.11
E
0.19
E
0.09
E
0.17
E
0.06
E
0.11
C
A(30g)
6.70
(0.23)
76.1%
A(30g)
5.16
(0.02)
86.7%
A(30g)
3.16
(0.01)
82.1%
A(30g)
22.47
(0.02)
98.7%
A(30g)
21.98
(0.04)
97.1%
A(30g)
32.73
(0.01)
89.0%
A(30g)
5.78
(0.01)
84.8%
D
A(30g)
0.10
(0)
1.1%
A(30g)
0.19
(0.01)
3.2%
A(30g)
0
(0)
0%
A(30g)
0.05
(0.01)
0.2%
A(30g)
0.22
(0.03)
1.0%
A(30g)
3.78
(0.01)
10.3%
A(30g)
0.94
(0.02)
13.8%
£
A(30g)
0.06
(0.04)
0.7%
A(30g)
0.21
(0)
3.5%
A(30g)
0
(0)
0%
A(30g)
0.03
(0.02)
0.1%
A(30g)
0.14
(0.03)
0.6%
A(30g)
0.13
(0.01)
0.4%
A(30g)
0.04
(0.02)
0.6%
F
A(30g)
0.04
(0)
0.4%
A(30g)
0.20
(0.26)
3.4%
A(30g)
0.12
(0.0.1)
3.1%
A(30g)
0
(0.02)
C%
A(30g)
0.14
(0.03)
0.6%
A(30g)
0.12
(0.01)
0.3%
A(30g)
0.06
(0.02)
0.9%
G
A(30g)
1.90
(0)
21.6%
A(30g)
0.19
(0)
3.2%
A(30g)
0.57
(0.02)
14.8%
A(30g)
0.22
(0.04)
1.0%
A(30g)
0.16
(0.03)
0.7%
SG
K(20g)
0
H
K(20g)
<0.04
SG
K(20g)
0.10
SG
SG
SG
TOTAL
Hg
-»g
13.32
99.7%
15.97
8.36
98.8%
38.84
40.60
43.91
13.04
1001
Hg
CONCENTRATION
^ig/CF
4.22
3.26
3.62
4.96
4.59
4.22
3.42
^g/1
0.149
0.115
0.128
0.175
0.162
0.149
0.121
-------
RUN
#
70
71
72
PARTICULATES
PROBE AND
CTCLONE
J*g
0.15
0.08
0.07
FILTER
Jig
7.76*
(O.OSOOg
0.64
(0.0359g
0.54
(0.0493g
TOTAL
-ug
7.91
)
0.72
i
0.61
>
&
H
1.69
H
11.14
H
13.78
TRAIN CONFIGURATION
MICROGRAMS OF Hg FOUND IN EACH POSITION
PERCENTAGE DISTRIBUTION OF Hg IN AMALGAMATORS
B
E
0.04
E
0.14
B
0.09
c
A(30g)
1.63
(0.02)
79.17.
A(30g)
4.99
(0.04)
61.5%
A(30g)
9.24
(0.03)
79.4%
D
A(30g)
0.34
(0.02)
16.5%
A(30g)
0.77
(0.03)
9.5%
A(30g)
1.43
(0.02)
12.3%
E
A(30g)
0.05
(0.02)
2.4%
A(30g)
0.85
(0.02)
10.5%
A(30g)
0.33
(0.01)
2.8%
F
A(30g)
0.04
(0.02)
1.9%
A(30g)
0.87
(0.01)
10.7%
A(30g)
0.35
(0.02)
3.0%
Q
K(25g)
0.20
A(30g)
0.63
(0.02)
7.8%
A(30g)
0.29
(0.05)
2.5%
H
SG
SG
SG
TOTAL
Hg
Jig
11.98
98.3%
20.23
26.25
Hg
CONCENTRATION
Jig/CF
2.53
4.80
3.79
^g/1
0.089
0.169
0.134
*The filter became wet during this run.
-------
APPENDIX II
RECOMMENDED PROCEDURE
The following is an outline of the recommended procedure as determined
in this study, for the analysis of mercury in stack gases of high SOg
content.
1- Apparatus and Reagents
In addition to the equipment normally used for isokinetic particulate
sampling (glass lined probe, weighed filter, filter holder, cyclone,
sample box, console with pump, etc.) the following items are required.
a- For Sampling and Cleanup Procedure:
1- Apparatus
Leco induction furnace equipped with a variac
Tank of nitrogen gas equipped with a pressure reducing
valve, flow meter and a drying tube packed with magnesium
perchlorate.
Bubbler apparatus with at least five interchangeable
sample holders
Two female ball joint adapters to fit amalgamators
Two ball joint clamps
Three impimgers - two modified and one Greenburg-Smith
Five amalgamators and "shells"
Gold chips, about 1/16 inch square by 0.007 inch thick,,
1$0 grams required for each run. These should have
been fired in a refractory oven at about 600-700°C for a
few hours (or overnight) before each run.
Sample containers - 6 oz. and 16 oz. wide mouth jars
with lids.
99
-------
100
Sample containers for filters - plastic Petri dishes
Graduated cyclinders - 2£, %
-------
101
Laboratory Data Control Mercury Monitor (or a standard
atomic absorption unit equipped for flameless determinations)
0-10 nrv recorder
Bubbler apparatus with interchangeable sample holders
Drying tube
Tygon tubing
Syringes - 5 and 10 ml.
2- Reagents
Standard mercury solutions
(A) Stock 1000 ppm mercury standard containing ~L%
concentrated nitric acid
(B) 1 ppm mercury standard containing 1/6 concentrated
nitric acid prepared fresh weekly
(C) f>0 ppb mercury standard containing 1% concentrated
nitric acid prepared fresh daily
Stannous chloride solution: mix 200 gram SnCl2*2 HoO with
£00 ml HC1 and dilute to 1 liter with distilled water.
Hydroxylamine hydrochloride solution: dissolve $0 gram
NH2OH-HC1 into £00 ml of distilled water.
Magnesium perchlorate
Distilled water
2- Preparation of the Amalgamators
The amalgamators are prepared as follows. A small plug of quartz wool is
inserted from the top of the amalgamator tube and pushed into place against
the supporting indentations. A length of \ inch dowel rod and a piece of
stiff wire can be used to help wedge the wool into position. The gold chips
are removed from the oven and allowed to cool for at least £ minutes. The
gold chips are then weighed out and about 30 grains of chips poured into each
amalgamator using the plastic funnel. The amalgamator may be tapped gently
-------
102
to help settle the chips against the plug. After filling the amalgamator
tubes with the gold chips, the ground glass tapers should be greased
and the amalgamator tubes placed in the "shells." If the amalgamators
are not to be assembled into the train immediately, the ends should be
stoppered as a precaution against contamination.
3- Assembly of the Sampling Train
All glassware should be rinsed before use with the following sequence:
1£ SnCl2 in 2.$% HC1, 1:3 HNO-j-HgO, distilled water and acetone. The
glassware should be allowed to dry before placing it into the train.
The probe, cyclone and filter are assembled into the sampling box in the
same configuration ordinarily used for taking an isokinetic particulate
sample. The impinger and amalgamator sequence is then assembled as
follows, starting with the filter:
a) A Greenburg-Smith impinger containing 2^0 ml of distilled
water.
b) An empty impinger
c) Five amalgamators in series, each one containing 30 grams
of gold chips.
d) An impinger containing about 2|?0 grams of silica gel.
Each impinger and amalgamator should be labeled with the run number and
the position of that unit in the train (e.g. 13-C). The configuration
of the recommended sampling train (excluding cyclone) is illustrated in
Figure 20.
It- Sampling
After assembling the train a leak check should be performed and the
sample box filled with ice and water. From this point on the sampling
-------
103
pilot Tube
Heated, Filter
Silica Gel
Thermometer
.Check Valve
Vacuum Line
Greenburg-Smith Impinger /
Scrubber Solution
/
Modified Impinger
Empty
\\^
s\\\\\
Modified Implngers
Gold Amalgamators
Figure 20. Configuration of the Recommended Sampling Train (excluding cyclone).
-------
ioU
procedure and the sample data recorded is the same as for the normal
isokinetic procedure as used for particulate sampling. The sample is
taken isokinetically for 15 minutes.
5- Cleanup Procedure
After obtaining the sample, the probe and sample box are taken to a
suitable area for the cleanup procedure. Distilled water is used for
all rinses. Samples are taken from each part of the train ahead of
the silica gel and placed in appropriately labeled sample containers
as follows:
a) The probe, cyclone and the glass parts of the filter
assembly are rinsed into a 16 oz. jar containing 25 ml
of the 3/6 KMnO, solution.
b) The filter (previously weighed) is placed in a suitable
container which is marked with the run number.
c) The inside part of the first impinger (Position A) and
the right angle connector leading into it are rinsed
into the contents of the irapinger "shell". About 7 grams
of solid KMnO, are then added slowly, with stirring, to
the liquid inuthe "shell" until the violet color of the
permanganate persists. The solution is then transferred
to a 16 oz. sample container and the grams of added
potassium permanganate are recorded.
d) The empty impinger- and the connector leading into it
are rinsed into a 6 oz. sample jar containing 25 ml of
the 3$ KMnOr solution.
e) Each of the amalgamator "shells" and the connector leading
into that amalgamator are rinsed into separate 6 oz. sample
jars, each containing 25 ml of the Jk KMnO^ solution.
f) Each amalgamator is fired into 50 ml of the 3? KMnOr solution
using the apparatus shown in Figure l5« The amalgamators
are fired in reverse of their order in the train (i.e. the
last one is fired first, the next to the last one is fired
second, etc.) in order to minimize contamination of suc-
cessive samples. The amalgamator is centered in the coil
of the induction furnace and connected to the nitrogen
-------
105
supply and the bubbler with two female ball-joint adapters
and two clamps. The nitrogen flow is set at 0.5 liters per
minute. The firing is commenced with the variac at a
setting of 60/6 and increased by 5? each minute until the
gold is glowing over its entire length. After firing each
amalgamator, the sample tube is detached, the drops of
permanganate clinging to the bubbler tube are rinsed into
it and then the contents of the sample tube are rinsed
quantitatively into a 6 oz. sample container. After
firing the series of amalgamators, the bubbler apparatus
and all the sample tubes are rinsed using tfte rinse
sequence described previously. If any tygon tubing is
used between tne amalgamator and tne bubbler, it should
be kept as short as possible and then replaced for each run.
g) A 50 ml blank of the 3$ permanganate solution is taken
for each run, placed in a 6 oz. sample container and
sent to the laboratory along with the samples.
A field record should be kept of all data recorded during the run
and of each sample taken for analysis from the train. Upon receipt
in the laboratory, each of the samples from the train is diluted to
the appropriate volume in a volumetric flask just prior to analysis.
6- Analysis of the Samples
a- Permanganate Solutions
A standard curve is prepared in duplicate for 0.05, 0.10, 0.25, 0.50,
and 0.75 M-g mercury standards by diluting 1, 2, 5, 10, and 15 ml aliquots
of the standard 50 ppb mercury solution to 50 ml. Each sjtandard is
placed in an interchangeable sample tube and attached to the bubbler.
Three ml of the stannous chloride solution is added with a syringe
through the ampoule stopper using sufficient force to mix the solution
with the standard. The sample is then aerated ( l.U liter/minute air),
volatilizing the mercury which is carried through a drying tube filled
with magnesium perchlorate and then through the mercury monitor operated
at a 0.6k range setting (least sensitive).
-------
106
The permanganate samples which contain the mercury from the amalgam-
ators are diluted to 100 ml with distilled water and returned to their
containers. (Great care must be taken to rinse the volumetrics with
stannous chloride solution between dilutions or cross contamination
is observed.) The wash and scrubber solutions are analyzed at the strength
at which they arrive from the sampling site. Before each aliquot of
sample solution is removed from a jar, the container is shaken thoroughly
until all solids are evenly dispersed in the solution. An appropriate
sized sample is quickly pipetted from the jar and placed in an inter-
changeable sample tube where it is diluted to approximately J>0 ml with
distilled water. Three ml of hydroxylamine hydrochloride solution is
added and the tube is swirled until the permanganate color disappears.
The tube is then attached to the bubbler and the solution is reduced
with three ml of the stannous chloride solution. The mercury is then
volatilized by aerating the solution (l.li liter/minute air). The revol-
atilized mercury is carried by the air stream through the mercury monitor.
All samples should be run in duplicate. The concentrations are calculated
from the standard curve and the amount of mercury in each sample is calcul-
ated from the dilution factor and the size of the aliquot.
b- Filters
One half of the filter is boiled in 10 ml of concentrated nitric acid
for 10 minutes. The filter is disintegrated with a high pressure stream
of distilled water and the mixture is diluted to 100 ml. After cooling,
i
the solutions are analyzed by placing 50 ml aliquots in the interchange-
able sample tubes and following the sample procedure as is used for
-------
107
standard solutions.
7- Analysis of the Data
The total amount of mercury found and its distribution in the sample
train is obtained from the data determined in the laboratory. This
information may conveniently be organized on a data form such as the
one shown in Figure 21.
The ratios of t2/tj_, t./tp, etc. may be calculated for those amalgam-
ators showing sufficient mercury to give a valid value for the ratio.
These values may then be compared to give an estimate of the extent of
gold contamination occuring during the sampling process. In the absence
of such contamination the ratio tjj/t^^ should be about 0.1 or less.
In addition the percentage of the total "amalgamator mercury" found on
each amalgamator should show an orderly decrease through the train and
the percentage found on the last one or two amalgamators should be only
one or two percent of the whole (or about the same value as the blank).
If these criteria are satisfied, then a collection efficiency of at least
may be assumed.
-------
Figure 21. Laboratory Data Form
SAMPLE
NUMBER
HILLILITERS
TAKEN FOR
ANALYSIS
TOTAL VOLUME
MICROGRAMS
MERCURI FOUND
TOTAL
MICROGRAMS
OF MERCURY
BLANK CORE.
NET TOTAL
MICROGRAMS
OF MERCURY
COMMENTS
LJ
b
-------
109
APPENDIX III
Isokinetic Data Sheets, Run 72
-------
TIELD DATA
PLANT.
DATE_
PROBE LENGTH AND TYPE
NOZZLE I.D
SAWPUNG LOCATION C^jSo
SAMPLE TYPE /Merc fry
RUN NUMBER "7^
OPERATOR Ra-ldvik J- -S
AMBIENT TEMPERATURE '.
BAROMETRIC PRESSURE 7=1
STATIC PRESSURE, (P ) ni
FILTER NUMBER (s)
ASSUMED MOISTURE,»,.
SAMPLE BOX NUMBER.
METER BOX NUMBER _
METER AHg
C FACTOR • 97
SCHEMATIC OF TRAVERSE POINT LAYOUT
PROBE HEATER SETTING too %
HEATER BOX SETTING, 3OQ °F
REFERENCE ap_
READ AND RECORD ALL DATA EVERY.
MINUTES
TPllfrPCn
1 l\nVC.IVJL
DHtUT^
nUlM
fciminfa
'IIUIIIULU
3amp|jna
Time.0
fS~ /w i n
\. CLOCK TIME
"T^XrX
TIME, mm ^v
"
s-n^r V:^>3
VIcJ?
v : 12
5r0p V:/*
/
GAS METER READING
-------
Ill
NOMOGRAPH DATA
PLANT.
DATE_
SAMPLING LOCATION.
72_
r
CALIBRATED PRESSURE DIFFERENTIAL ACROSS
ORIFICE, in. H20
AVERAGE METER TEMPERATURE (AMBIENT + 20 °F),°F
PERCENT MOISTURE IN GAS STREAM BY VOLUME
BAROMETRIC PRESSURE AT METER, in. Hg
STATIC PRESSURE IN STACK, in. Hg
(Pm±0.073 x STACK GAUGE PRESSURE in in. H20)
RATIO OF STATIC PRESSURE TO METER PRESSURE
AVERAGE STACK TEMPERATURE, °F
AVERAGE VELOCITY HEAD, in. H20
MAXIMUM VELOCITY HEAD, in. H20
C FACTOR
CALCULATED NOZZLE DIAMETER, in.
ACTUAL NOZZLE DIAMETER, in.
REFERENCE Ap, in. H20
AH@
Tmavg.
Bwo
pm
ps
PVPm
savg.
APavg.
APmax.
I.&?
LfS0
J"
T-^to
-.or
.^3%
J~?o
.S~l
, <77
. zr
EPA (Dur) 234
4/72
U. =. GOVERNMENT PRINTING OFFICE: 1973 746769/4162
------- |