DEVELOPMENT AND EVALUATION OF AN
ANALYTICAL METHOD FOR THE DETERMINATION
OF TOTAL ATMOSPHERIC MERCURY
FINAL REPORT
Covering the Period
November 1970 Through March 1972
Contract No. EHSD 71-32
Prepared for: Environmental Protection Agency
Division of Atmospheric Surveillance
Project Director: W. M. Henry
Authors: D. L. Chase
D. L. Sgontz
E. R. Blosser
W. M. Henry
. June 16, 1972
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
MANAGEMENT SUMMARY
Contract No. EHSD 71-32
"Development and Evaluation of an Analytical Method to Determine Total
Atmospheric Mercury".
Objective of Research Program
This program was initiated to develop a method for determin!ng mercury,
both organic and inorganic, in ambient air.
It was directed first to systems
using iodine monochloride as a collecting medium and later was broadened to include
other mercury collection and measuring systems.
Significance of Research Results
The research data show that the chemical absorption method involving
iodine monochloride, while capable of collecting various forms of mercury and
therefore valuable in certain respects, presents severe contamination and
analytical problems.
A pyrolysis-amalgamation technique appears to be quite
useful and reliable for many, if not all, forms of mercury and is amenable to
straightforward chemical analysis as well.
Two commercial instruments were
evaluated; one appeared valid for elemental mercury vapor, but the other was
unreliable.
How Sponsor Can Use Results
The Environmental Protection Agency will be able to use the information
generated in this research program to guide it in the development of an analytical
system for monitoring the air for mercury content.
Future Effort
Additional research effort is outlined that could provide more basic
information about the iodine monochloride system.
Also additional research
effort could be useful in providing more precise operating conditions for the
pyrolyzer-s~lver wool system.
-------
ABSTRACT
Total mercury in ambient air can.be collected in iodine monochloride,
but the subsequent analysis is relatively complex and tedious, and contamination
from reagents and containers is a problem.
A silver wool collector, preceded by
a catalytic pyrolysis furnace, gives good recovery of mercury and simplifies the
analytical step.
An instrumental method based on particle counting proved
unreliable, but another instrument using the 253.7 nm Hg optical absorption line proved
to be quite accurate for the determination of elemental mercury in air.
-------
TABLE OF CONTENTS
SUMMARY AND RECOMMENDATIONS.
.
.
INTRODUCTION. .
.
EXPERIMENTAL WORK.
Preparation of Working Mixtures,
Solutions, and Standards
Direct Addition.
Aeration
Vaporization.
Permeation Tube.
.
.
.
.
.
.
.
o
.
ICI Experiments
.
Electroplating.
Dithizone Extraction.
Analysis of Reagents.
Preparation of Purified Reagents.
Coprecipitation of Mercury on CdS.
Analysis of Air Samples in ICI
Check Analyses Using the Dithizone Method
Preparation and Analysis of Synthetic Air Samples.
Double Aeration. 0
Mercury Vapor Co11ec tion . .
Flow Rate Experiments. .
o
.
.
.
.
Evaluation of a Commercial Mercurv Vapor Detection Svstem.. .
Principle of Operation.
Preliminary Evaluation.
Evaluation Using Permeation Tubes.
.
Evaluation of a Mercury Collection Instrument.
PHASE IV .
.
Introduction.
.
.
.
Principle of Operation.
.
.
EXPERIMENTAL WORK.
Construction of Equipment.
Calibration and Efficiency Experiments.
o
Calibration by Aeration.
Direct Loading of Collectors From Permeation Tube Assembly.
i
.
Pa~e
1
2
3
. 3
o 3
4
4
7
7
7
9
10
.11
12
. 14
. . . 15
15
18
19
21
. 24
. 25
. 25
. 26
. 30
. . 31
. 31
. 31
. 32
. 32
. 35
. 35
. 38
.
-------
TABLE OF CONTENTS (Continued)
Mercury Vapor Loading Through the Catalyst Furnace. .
. . . .
. . . .
Interference Evaluation. .
. . . .
......
. . . . . . . . 0 . .
Studies of Possible Interference From S02 . . . . . . . . .
Alteration of Mercury Permeation Tube Assembly.
Standard Curve Based on Peak Area. . . . . . . . . . .
Experiments Using Redesigned Equipment. . . . . . . .
.....
. . . . .
. . . . .
.....
S~udies of Possible Interference from H2S . .
. . .
.....
. . . e
Experiments Using Redesigned Equipment. .
......
......
Studies of Possible Interference From N02 .
Mercury Collection From Dimethyl Mercury Vapor.
. . . .
......
. . . .
........
Collection of Mercury From HgC12 Particulate Using Pyro1yzer
Collection of Mercury From (CH3)2Hg Using Pyro1yzer at Several
Tempera tures . . . 0 . . . . . 0 . . . . . . . . . . . . . . . .
. . . . . . .
.....
Construction Details of the Pyrolysis Tube and Furnace
.....
. . . . .
DISCUSSION AND CONCLUSIONS.
. . 0 . . . 0
. . . .
.....
. . . . . . .
FUTURE WORK. . . . . . . . .
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.
Figure 10.
Figure 11.
Figure 12.
Figure 13.
Figure 14.
. . . . . . .
. . . .
. . . . . .
.....
LIST OF FIGURES
Mercury in Air Calibration Assembly. . . . . . . . . . . . .
Vaporizer Assembly. . . . . . . . . . . . . . . . . . .
Permeation Tube Assembly. . . . . . . . . . . . . . . . . . . . . .
Response Plots for CdS Coprecipitation Experiments. . . . . .
Optical Absorption by Mercury at 253.7 run Versus Mercury Content. .
Results of Calibration Runs. . . . . . . . . . . . . . . . . . . .
Collector. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pyrolysis Tube. . . . . . . . . . . . . . . . . . . . . . . . . . .
Calibration Curve For Silver Collectors. . . . . . . . . . . . . .
Modified Permeation Tube Assembly. . . . . . . . . . . . . .
Calibration Curve For Silver Collectors, Based on Peak Areas.
Vaporizer Assembly. . . . . . . . . . . . . . . . . . . . . .
Vaporizer-Pyrolyzer Assembly. . . . . . . . . . . . . . . . . . . .
Pyrolysis Tube. . . . . . . . . . . . . . . . . . . . .
ii
Page
38
40
40
43
43
46
46
46
50
50
57
57
59
59
63
5
6
8
13
20
29
33
34
37
44
45
55
60
62
-------
Table l.
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9.
Table 10.
Table ll.
Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.
TABLE OF CONTENTS (Continued)
LIST OF TABLES
Recovery of Dimethyl Mercury. . . . . . . . . . . . . . . . .
Recovery of Mercury Vapor. . . . . . . . . . . . . . . . . . . . . .
Recovery of Vaporized Methyl Mercuric Chloride. . . . . . . . . . .
Recovery of Vaporized Dimethyl Mercury. . . . . .
Results of Evaluation. . . . . . . . . . . . . . . . . . . . . . . .
Results of Evaluation Runs. . . . . . . . . . . . . . . . . .
Standardization of Collectors by Aeration. . . . . . . .
Results of Direct Permeation Tube Loading. . . . . . . . . . .
Results of Mercury Vapor Loading Through the Catalyst Furnace.
Mercury Recovery From Gas Stream Containing SOZ' .
Recovery of Mercury From Gas Stream Containing S02 . . . . . . . . .
Recovery of Mercury From Gas Stream Containing H2S . . . . . . . . .
Recovery of Mercury From Gas Stream Containing H2S . . . . . . . . .
Recovery of Mercury From Gas Stream Containing N02 . . . . . . . . .
Mercury Collection From Dimethyl Mercury Vapor. . . . . . . . . . .
Collection of Mercury From HgC1Z Particulate Using Pyrolyzer . . . .
Collection of Mercury From (CH3)ZHg Using Pyrolyzer. . . . . . . . .
Hi
PaRe
17
21
23
23
28
30
36
39
41
42
47
49
51
53
56
58
61
-------
DEVELOPMENT AND EVALUATION OF AN ANALYTICAL METHOD FOR THE
DETERMINATION OF TOTAL ATMOSPHERIC MERCURY
by
D. L. Chase, D. L. Sgontz, E. R. Blosser, and W. M. Henry
SUMMARY AND RECOMMENDATIONS
Iodine monochloride is an effective collection medium for the several
chemical forms of mercury, providing the rate of air flow is kept sufficiently
low to permit an adequate reaction time between the ICl and organometallic mercury.
The actual determination of the mercury so collected may be made by a dithizone
extraction or by aeration techniques.
Double aeration is preferable to eliminate
possible interferences from the ICI solution.
Electrodeposition or coprecipitation
of mercury from ICI solutions are not satisfactory.
The ICI collection and sub-
sequent dithizone extraction and/or double aeration methods are too tedious for the
rapid analysis of large numbers of samples and may not provide satisfactory accuracy
for mercury contents of air in the nanogram range.
Of two commercial instruments designed to determine ambient air mercury,
one proved unreliable while the other (primarily a collection system), under limited
evaluation gave satisfactory data on elemental mercury for several simulated samples.
It is recommended that the ICl absorber system not be adopted in its
present state of development for universal use in mercury collections owing to the
somewhat difficult (dithizone) and unreliable (reduction-aeration) analytical
procedures required, plus lack of sensitivity and high blank.
It is a suitable
collection medium if necessary precautions are observed and if above-average analysis
3
time and if a detection limit of about 1 ng Hg/m can be tolerated.
The adoption of one commercial instrument for the. determination of mercury
is not recommended at least in its present form.
Insufficient data were available
-------
2
on this program to permit a thorough judgment of a second commercial instrument,
but the data available are favorable for its application to determining
elemental mercury in air.
The pyrolysis-silver wool system appears to be the best collection
method yet devised for total atmospheric mercury.
It has shown no interferences
in work to date; it is capable of collecting elemental, particulate, and organo-
metallic forms of mercury; and it collects the mercury in a form suitable for
easy, rapid analysis.
Pending further evaluation and field testing, it is
recommended that this system receive favorable attention and research effort.
INTRODUCTION
In the mounting concern with environmental quality many formerly
unsuspected hazards have emerged, usually with widespread publicity and
accompanied by alarmist overstatements of the danger.
Mercury, widely used by
industry and by consumers in relatively small amounts, finds its way into the
air, the soil, and the water by many routes.
Natural losses such as evaporation
and leaching also contribute to the mercury cycle.
One of the first requirements in evaluating the extent of man's impact
upon the mercury cycle (or any other environmental stress, for that matter) is
accurate analytical data.
The Environmental Protection Agency, charged by law
and executive order with gathering these data, setting standards, and overseeing
enforcement of regulations, is studying means to assure that the data upon which
decisions must be made are valid and accepted by those concerned.
Mercury is a heavy and toxic element that exists in the atmosphere
.as an element, as a gaseous compound, and as a particulate.
Its chemistry is
such that many of the usual means used to collect samples for analysis
fail to collect all the mercury present.
Furthermore, the volatility and
instability of some forms of mercury make the retention of mercury difficult unless
-------
3
special precautions are taken.
For these reasons several systems have been
devised that will catch and retain the important known forms of mercury.
One such method, using a solution of iodine monochloride (ICI), has
been reported to be effective in removing different forms or chemical spectes
(20)*
of mercury from air.
The present program had as one of its primary
objectives the development of a reliable method to analyze ICI solutions in
which had been collected the mercury from air samples.
As the program progressed,
other collection schemes and other analytical methods were also explored.
During the course of this program the emphasis of the research effort
was altered several times when difficulties were encountered, when new develop-
ments elsewhere came to the attention of EPA, and when expediency dictated that
a particular technique be studied in a short time.
In this report the experimental
work and results are presented topically as much as possible, rather than chronologi-
cally.
Reference is made to the monthly report(s) in which the work was reported;
however, the technical details are described fully in this report.
EXPERIMENTAL WORK
Preparation of Working Mixtures. Solutions. and Standards
Working dilutions (solutions, air mixtures, or standards) of mercury
were prepared in two basic ways:
(1) by the direct addition of a known volume of
a solution of a mercury compound, and (2) by the addition of a known quantity of
mercury (elemental or compound) by aeration, volatilization, or permeation into an
inlet air stream, a collector, or the atomic absorption cell.
The two approaches
are described in the following sections.
Direct Addition
Several series of standards containing known amounts of mercury were pre-
pared daily by adding suitable aliquots of aqueous HgClz in 0.1 M HCI to ICI solutions. (1)
*
References are listed at the end of the report.
-------
4
This direct method, carefully carried out, should be quite free from any bias errors
except the usual inaccuracies of weighing and measuring, and was used as a means of
preparing reference solutions for several experiments.
These solutions are only
stable at ppm or higher concentrations and should not be stored in plastic ware.
Aeration
Aeration is only slightly less direct than the preceding direct addition
method, and it has the advantage that a known amount of mercury vapor (elemental)
is produced and can be absorbed in an impinger or introduced directly into a
readout chamber such as the absorption cell in the atomic absorption unit.
The
(2 5 6 17)
aeration method, employed to check out several of the analytical procedures' " ,
++
is based on the reduction of Hg to HgO in solution and subsequently sweeping the
*
HgO vapor from the solution with a stream of air (or inert carrier gas, if desired) .
Vaporization
This technique lends itself to a somewhat easier time-quantity control
of the released mercury. A known quantity of a mercury compound is placed, in
solution, into a vaporizing chamber such as shown in Figure 1(4) and Figure 2(13,19).
In practice, the desired amount of the mercury compound is pipetted into the vapor-
izing chamber which is promptly attached to the system and isolated.
Proper mani-
pulation of the stopcocks and vacuum pump gives a steady flow of air through the
vaporizer, sweeping the mercury compound into an impinger or other collection or
readout device.
Vaporization, in common with direct introduction and aeration
methods, provides good control over the total quantity of mercury introduced, but
poor simulation of a steady-state mercury loading as is encountered in nearly all
ambient air sampling conditions.
Also, the small amount of residual solvent for
the mercury compound may interfere with the subsequent analysis as, for example,
the toluene used to dissolve dimethyl mercury interferes with the direct flameless
atomic absorption determination of mercury.
* We can neither confirm nor refute the reported variation between
Hg vapor pressure calibration techniques described by Muscat, et
~, No.2, p 218 (1972). .
aeration and
al (Anal. Chern.
-------
!S-I
A
FIGURE 1.
D
5
G
To
--
vacuum
H
MERCURY IN AIR CALIBRATION ASSEMBLY
E
F
(A)
(B-1 and B- 2)
(C)
(D)
(E)
(F)
(G)
(H)
Vaporizer
Teflon stopcocks
Ball joint
Impinger
Thiosulfate scrubber
Charcoal trap
Particulate filter
Manifold containing
limiting orifice
-------
Rotameter
Silver wool
plug
FIGURE 2.
6
I
Itrfu
Vaporizer
VAPORIZER ASSEMBLY
L
-. To exhaust
Impinger
-------
7
Permeation Tube
Permeation tube techniques are used to introduce known amounts of
a substance into a gas stream at a constant rate. A tube is fabricated by
placing the desired material in a semipermeable tube and sealing the ends.(2l)
The rate of permeation is a function of temperature as well as of the physical
dimensions and characteristics of the tubing walls, but for reasonable time
periods is not dependent upon the total amount of the material remaining in
the tube.
In this program some difficulty was encountered in constructing
tubes with permeation rates suitable for use in preparing simulated gas samples.
For two experiments in this program permeation tubes were used:
mercury vapor
was permeated through a silicone rubber tube for evaluation of a mercury sensing
instrument (10) and for evaluating a collector system(17). The design of the
assembly is shown in Figure 3. (10,17)
ICI Experiments
Iodine monochloride (0.1 M, 0.5 N HCl) was chosen by EPA as a
collection media for total atmospheric mercury using liquid impingers; the
development of an analytical method to determine the mercury so collected was
the initial purpose of this program.
The solution could be analyzed directly
using a flame-atomic absorption technique, but severe problems would arise owing
to the corrosive nature of the solution, the (relative) insensitivity of the
technique, and probable interference from free-iodine vapor.
Therefore, several
alternative methods were explored.
Electroplating
A series of solutions containing 0, 5, 10, and 15 nanograms of mercury
as HgClz in 40 ml portions of 0.1 M ICI was prepared.
electrolyzed at 3 volts for 20 minutes with a copper coil as the cathode and a
These solutions were
platinum wire as the anode.
After the plating step was completed the coils were
washed in water, ethanol, and acetone, then dried and positioned in an absorption
-------
8
Circulating water
"
.Ll...=
To
collector
or readout
Il
Thermometer
Rotameter
n
Permeation tube
Nitrogen -----
Constant
temperature
bath
FIGURE 3.
PERMEATION TUBE ASSEMBLY
-------
9
cell mounted in the light path of the Perkin Elmer Model 303 atomic absorption
spectrophotometer (AAS).
A 5-volt current was passed through each coil to heat
the wire and release the monoatomic mercury vapor for measurement by ultraviolet
light absorption at 253.7 nm. (22)
All of the coils, including the blank, gave off a small cloud of
smoke when heated, making accurate measurement of the mercury vapor impossible.
On close inspection the coils were found to be covered with a thin deposit of a
salt-like material (possibly a copper-iodine compound) that could not be washed
off.
Plating from an ICI solution which had been reduced by sodium thiosulfate
produced the same type of deposit. These experiments indicated that the direct
plating of mercury from ICI solutions onto inexpensive copper coils is impractical. (1)
Dithizone Extraction
One of the proposed methods to determine the mercury collected in ICI
is to extract this mercury with dithizone using the procedure described by
Linch, et al(20).
To evaluate this technique, several preliminary experiments
were performed.
A simple flow-through aeration system was constructed and attached to
the absorption cell placed in the light path of the Perkin-Elmer 303 atomic
absorption spectrophotometer (AAS). (2) A series of standards containing 0, 5,
10, 15, and 20 nanograms of mercury as HgC12 in 50 ml of 2 N HCl was prepared.
These standards were analyzed for mercury content by placing them individually
in the aeration cell, reducing with stannous chloride solution, and pulling air
through the system at about 1.0 liter per minute.
Results were very encouraging.
The response was good, blanks were relatively low, and readings were reproducible.
Having established that HgO could be aerated from an HCI solution into
the AAS successfully, the extraction by dithizone of mercury from HCI, followed
by aeration, was attempted.
The method is based on a dithizone extraction from an
HCl solution, back extraction into another HCl solution followed by aeration and
-------
10
measurement.
However, when the method was applied to a series of standards
identical to those used to check out the flow-through aeration system, the
recoveries were low and erratic.
A study of the system showed that longer
contact time was needed in both the dithizone extraction and back-extraction
steps.
At least 5 minutes shaking time was required in both steps to obtain
reasonable recoveries of mercury.
When these precautions were taken, the
results with the HCI system were promising enough to tryout the method on
standards containing ICI.
A series of standards containing 0, 5, 10, 15, and 20 nanograms of
mercury in 50 ml of the ICI collecting solution were prepared.
The dithizone
extraction and back extraction into 5 N HCl were carried out, using the
5 minute shaking time indicated by the previous experiments.
When the final
HCl solutions were reduced and aerated, the response obtained with all of the
,
solutions was erratic.
The blank was quite high. and the readings obtained
from standards containing different levels of mercury were all high and rough1y the
same.
This suggested the possibility that the KI and/or KI03 used to make up
the absorbing solution contained mercury, or that some vapor other than mercury
vapor was being swept into the absorption cell and absorbing ultraviolet light.
Analysis of Reagents.
Samples of reagent grade KI and KI03 used to
make up the ICl, as used as atmospheric sample collecting solutions in this study,
were analyzed for mercury by a neutron activation technique. (3) The results of
this analysis showed that the KI contained 0.1 ppm Hg and the KI03' 0.3 ppm Hg.
approximately 0.555 g KI and 0.375 g KI03 were present in the diluted atmospheric
Since
sample absorbing solution, a blank of about 170 nanograms of Hg was being intro-
duced by these reagents.
This was unacceptable for the purpose of this collecting
system and necessitated efforts to reduce the blanks to a more reasonable
concentration.
-------
11
Preparation of Purified Reagents. Initial efforts to purify the
(23)
reagents followed a technique described by Monkman, et al. CdS pads were
prepared by treating a glass filter pad with 10 percent cadmium acetate and
10 percent sodium sulfide solutions alternately, beginning and ending with
the cadmium acetate solution.
The pad was then washed with water, pumped
dry, and placed in an oven at 1000 C for one hour.
Finally the pad was heated
in a muffle furnace at 375°C for one hour and stored in a desiccator until
used.
Small disks to fit the filtering apparatus were then cut out of the
pad with a cork borer.
A solution containing 11.1 g KI in 200 ml of water and another
solution containing 7.5 g KI03 in 200 ml of water were prepared.
Each
solution was filtered through a CdS pad.
By adding 45 ml HCl to the filtered
KI solution, adding the filtered KI03 solution, and diluting to 1 liter with
water the diluted form of the ICI collecting solution was obtained.
The
mercury content of this solution was checked by adding an excess of NaOH to a
50 ml aliquot, reducing with hydroxylamine hydrochloride. and aerating through
the absorption cell of the AAS.
A comparison of the responses from the treated
and untreated collecting solutions indicated that the mercury level had been
lowered from about 150 nanograms to 10-15 nanograms in a 50-ml aliquot.
Unfortunately, the overall dithizone extraction procedure is lengthy
and ill-suited to a large number of samples.
By mutual agreement with the
Project Monitor work on the extraction method was suspended indefinitely, even
though it remains a method of demonstrated validity.
As mentioned below, the
dithizone extraction was employed to check some surprisingly high mercury values
obtained in the analysis of submitted air samples and to check the glass impinger-
ICI collection system.
-------
12
~
Coprecipitation of Mercury on CdS
The coprecipitation of HgS and CdS by filtering a sample solution adjusted
.
to pH 5-6 through a CdS pad forms the basis for a sensitive method for the determi-
(23)
nation of mercury.
The CdS pad containing the separated mercury is dried and
placed in a tube furnace heated to 5500 C to release mercury vapor.
Air is drawn
through the combustion tube to sweep the mercury vapor into an optical cell where
the amount of ultraviolet light absorbed by the mercury vapor can be measured.
This
method was evaluated as a technique to be used for the determination of mercury in
.
atmospheric sample absorbing solutions containing ICI.
An electric combustion furnace was fitted with a 25-mm-diameter Vycor
combustion tube which tapers to a 7-mm diameter. at one end.
The constricted end
was connected to the absorption cell of the AAS with rubber tubing.
A pump was
connected to the exhaust end of the absorption tube in order to draw air through
the whole assembly.
A series of standards containing 0, 10, 20, and 40 nanograms
of mercury were prepared by adding 0, 10, 20, and 40 microliter aliquots of a
standard l~gHg/ml solution of mercuric chloride in acetone to CdS pads.
A micro-
syringe was used for measuring the standard solution.
These standards were dried
and then placed, one by one, in a quartz boat which was inserted into the combustion
tube heated to 5500 C.
Air was drawn through the system at about 1 liter per
minute.
The response of the atomic absorption spectrophotometer versus mercury
concentration was linear and a satisfactory calibration was achieved as shown in
Figure 4, Curve A.
The coprecipitation method was then applied to a series of standard
solutions containing 0, 10, 20, and 40 nanograms of mercury in dilute HCl.
These
solutions were adjusted to pH 6 and filtered individually through CdS pads.
The
pads were dried and processed in the same manner as described above.
Again the
response was linear and a smooth calibration curve, closely matching curve A in
Figure 4, was constructed for this set of conditions.
-------
13
.
..
50
40
A
~ HgC'2 in acetone
o dried on CdS
u
CI)
x 30
CI)
c
0
CI) B
>
a 20 X""HgCI2 in IC I
+- fi Itered through
:" '-
0 CdS (pH6)
.£:.
u
[ . HgCI2 in IC I
filtered through
CdS (pHS.5)
00
10
20 30
Mercury Added, nanograms
40
50
FIGURE 4. RESPONSE PLOTS FOR CdS COPRECIPITATION EXPERIMENTS
.~
[ .
-------
14
Standard solutions containing 50 ml of the purified ICI absorbing
solution and 0, 10, 20, and 40 nanograms of mercury were prepared.
These
solutions were neutralized with NH40H to pH 6 and filtered through CdS filter
pads.
The pads were dried, transferred to small quartz boats, and heated
individually in the furnace assembly described above.
The blank at pH 6 was
appreciable, equivalent to about ~ 15 ng Hg.
Also, the slope of the curve
(Figure 4, Curve B) was lower, indicating a possible suppressant effect or
a partial (proportional) failure to achieve the complete precipitation of
mercury by the CdS.
This experiment was repeated with all conditions the
same except the solutions were adjusted to pH 8.5 before filtering through
the CdS pads.
The response from the blank and standards was essentially the
same as shown in Curve C, Figure 4.
Another possible explanation of the
failure to recover mercury is that it is tied up in a strong covalent iodide
2-
complex (HgI4 ) at pH 6 and is not available to react completely with the sulfide
ion.
At pH 8.5 the mercury is released from the complex but does not form a
sulfide in this strongly alkaline medium.
Of course these observations are
based on very limited data.
On the basis of these limited experiments the coprecipitation method,
while applicable to HCI solutions, does not appear to be suitable for the
determination of mercury in ICI solutions.
Analysis of Air Samples in ICI
Four samples from the group of air samples collected and supplied by
the Sponsor were selected for mercury analysis.
The solutions from the plastic
collecting tubes were transferred to volumetric flasks and diluted to 100 ml
with water.
Aliquots of 25 ml were taken from each flask, made alkaline with
NaOH,and reduced with hydroxylamine hydrochloride.
Each solution was then
aerated,and the air stream was directed through the UV absorption cell of the
-------
15
atomic aDsorption spectrophotometer.
All readings were off scale, indicating
an unusually large amount of mercury was present.
Check Analyses Using the Dithizone Method.
To verify the unexpected
high level of mercury, additional aliquots of 25 ml were analyzed by the
dithizone extraction method.
The total mercury content of the collecting
solutions was found to be in the range of 120-140 micrograms.
This level of
mercury strongly indicated a severe blank problem attributable to
mercury in plastic collection tubes being leached out by the ICI collecting
solution.
Mercury compounds are used during the formulation of some
plastic materials.
Preparation and Analysis of Synthetic
Air Samples
The results obtained on the actual air samples suggested that the
ICI system be examined anew, beginning with the materials of construction of
the collectors. (4) For that reason, a vaporizer air calibration assembly
(Figure 1) was employed. (19)
The manifold containing the hypodermic needles,
the particulate filter, and the charcoal trap of the NASN sampler were retained
(Figure 1).
However, the glass intake manifold was isolated from the system,
the plastic collector tube was converted into a scrubber containing alkaline
thiosulfate solutio~ and a direct connection was made between it and the charcoal
trap.
The glass impinger with the vaporizer attached was then connected to the
thiosulfate scrubber.
To operate, stopcock B-2 was closed, 0.1 m1 of a standard mercury
solution was added to the vaporizer, and the stopper with stopcock B-1 in closed
position was put in place.
The vacuum pump was starte~ and after 5 to 10 seconds
stopcock B-2 was opened.
Then after a few seconds the upper stopcock B-1 was
opened slowly.
Air was then passed through the system for .the required length of
time.
A blank run was made at the same time to monitor the laboratory air being
-------
16
drawn through the system.
Air flow calibration runs were made to establish
conditions yielding flow rates of approximately 0.2 and 2.0 liters per minute.
A standard solution of dimethyl mercury was prepared by dissolving
84.1 mg dimethyl mercury (equivalent to 73.2 mg Hg) in 100 ml of toluene, or
732 micrograms Hg per mI.
Successive 1:10 dilutions of this solution provided
standard solutions A, B, and C containing 0.732, 7.32, and 73.2 micrograms Hg
per ml, respectively.
A microsyringe was used to measure 0.1 ml of solution A and place
it in the vaporizer.
The vacuum pump was started, the stopcocks on the
vaporizer were opened, and the run continued for 1.5 hours.
At a flow rate
of 0.2 l/min, this equals 73.2 ng Hg/0.018 m3, or 4Q70 ng/m3 average.
This
procedure was repeated with 0.1 ml of solution B (average Hg concentration
equals 40.7 ~g/m3)and again with solution C (average Hg
407 ~g/m3), a freshly prepared impinger containing 25 ml
concentration equals
of ICI collecting
solution being used for each.
A 97 percent recovery was obtained by the
dithizone-AAS method on the 7.32 ~g Hg addition.
The lower concentrations of
Hg additions were recovered with considerably less efficiency as judged by the
atomic absorption results.
However, these apparently low recoveries may be
due to variable blank, possible spectral interference of the iodine monochloride
encountered in the atomic absorption procedure, or adsorption, of Hg on the container walls.
The same volatilization apparatus was used to test the efficiency
of the ICI - glass impinger collecting system under conditions of high loading
(5-15 micrograms Hg) and high air flow (1.7 l/min).
Three calibration runs were made at the 1.7 l/min air flow rate
with 5.32, 10.6, and 14.9 micrograms of mercury as dimethyl mercury in toluene
added respectively via the vaporizer.
Glass microimpingers containing 25 ml
of ICI solution were used for collecting the vapor.
Three standards were
prepared by adding the same amounts of the standard dimethyl mercury - toluene
-------
17
solution to 25 ml portions of the ICI solution.
The mercury content of the
three standards and the three test solutions were determined by a dithizone
method.
The results are shown in Table 1.
TABLE 1.
RECOVERY OF DIMETHYL MERCURY
(Flow rate 1.7 liters per minute)
Hg Added, Hg Found, Absolute Recovery, Recovery Based on
Sample micrograms micrograms percent Standards, percent
Std. (a) 5.32 4.7 88
Std.2 (a) 10.6 9.9 93
Std'3 (a) 14.9 13.2 89
Run 1 (b) 5.32 3.6 76
Run 2(b) 10.6 7.2 73
Run 3(b) 14.9 9.6 73
*
average 90, 0' = 3 74, 0' = 2
(a)
(b)
Toluene solution of dimethyl mercury added directly to ICI solution.
Toluene solution of dimethyl mercury vaporized and collected in ICI
solution.
These data, although not showing as quantitative recovery at the high
flow rate as could be desired, nevertheless served to indicate that the system,
as used with glass impingers in these experiments, does not give rise to high
mercury values.
By implication, therefore, it can be postulated that the very
high mercury values obtained in the analysis of the ambient air samples were
caused by contamination from the reagents or more importantly from the plastic
impinger bottles.
By mutual agreement between BCL and the EPA Project Officer it was
agreed that further work on Phase III (sample analysis) be suspended indefinitely
owing to the high and undoubtedly incorrect mercury values obtained for the few
samples analyzed.
The available time and funds were directed to additional studies
under Phases I and II and, as an extension of the effort, into Phase IV.
* In this report, 0' refers
= . /r,(R - R) 2
equation 0' V~ n - 1
to the precision, not accuracy, as calculated by the
, where Rand R are the average and individual percent
-------
18
Double Aeration
Erratic responses often were observed when solutions containing ICI
were reduced and aerated directly through the optical cell of the atomic
absorption spectrophotometer.
This was most apparent when the 3x or
lOx expansion scales were used to provide more sensitivity.
The source
of this erratic behavior was not identified positively, but it appeared to be
caused by some interfering substance other than mercury vapor being evolved
from the solution.
This substance then absorbed ultraviolet energy in the
optical cell and caused spurious results.
Residual toluene may have been the
.interfering agent in synthetic samples prepared by vaporizing mercury compounds
dissolved in toluene.
An analysis scheme involving double aeration was investigated.
By
reducing the mercury in the ICI collecting solution, driving the mercury vapor
out by aeration, and collecting the mercury vapor in a second absorbing solution
which does not contain ICl, it was hoped that the interfering substance would
not be absorbed.
The second absorbing solution could then be reduced and aerated
directly into the optical cell of the AAS.
Five standards were prepared, each containing 25 ml of the ICI
collecting solution and 25, 50, 75, 100, and 150 nanograms of mercury as HgC12'
respectively.
Each standard was made basic with sodium hydroxide, reduced with
hydroxylamine hydrochloride, and aerated for five minutes into a scrubbing bottle
containing 50 ml of a 20 percent KBr - 5 percent HCl solution.
Each KBr-HCl
solution was then made basic with sodium hydroxide, reduced with hydroxylamine
hydro~hloride, and aerated directly into the optical cell of the AAS operated
with the 3x scale expansion.
Absorption was measured in chart divisions and
plotted against the mercury content of the standards.
-------
19
Three standards containing 50 ml of the 20 percent KBr - 5 percent HCl
solution and 50, 100, and 150 nanograms of mercury were prepared.
Each of these
standards was made basic with sodium hydroxide, reduced with hydroxylamine
hydrochloride, and aerated directly into the measuring system.
Absorption was
measured in chart divisions and plotted against the mercury content.
Results
are shown in Figure 5.
The results of this preliminary investigation were quite promising.
There appeared to be no serious loss of mercury in the transfer from one collect-
ing system to another and a linear relationship between ultraviolet absorption
and mercury content was indicated.
However, work on the double aeration technique
was suspended so that studies of other aspects of ICI solution collection and
analysis could be expedited.
Mercury Vapor Collection
The apparatus shown in Figure 1 was modified by substituting a mercury
vapor generator for the vaporizer and was used to test the efficiency of the
ICl-glass impinger collecting system under conditions of low loading (20 to
100 nanograms Hg) and high air flow (1.7 l/min).
Three calibration runs were made at the high flow rate of 1.7 l/min
with 20, 50, and 100 nanograms of mercury as mercury vapor added by passing the
air stream through an aeration cell containing dilute HCl, stannous chlorid~ and
measured amounts of standard mercury HgC12 solution.
aeration cell was reduced to the elemental for~ and mercury vapor was carried by
The mercury present in the
the air stream through a glass impinger containing 25 ml of ICI collecting
solution.
Mercury in the ICI solution was determined by making the solution
basic, reducing with hydroxylamine hydrochloride, and aerating directly into the
ultraviolet absorption cell mounted in the optical system of the atomic absorption
spectrophotometer.
Results are shown in Table 2.
The results shown in Table 2,
-------
100
90 0
x
80
70
en 60
c
.2
en
> 50
a
.....
L..
0
.c 40
U
N
o
30 0 Double aeration- IC I to KBr to cell
x Direct aeration of Hg from KBr
20
o
o
10
20
30
40
50
60 70
Mercury
160
FIGUIq1:- 5.
OPTICAL ABSORPTION BY MERCURY AT 253.7 nm VERSUS MERCURY CONTENT
-------
21
TABLE 2.
RECOVERY OF MERCURY VAPOR
(Flow rate 1.7 liters per minute)
Hg Added,
nanograms
Hg Recovered,
nanograms
Recovery,
percent
20
50
100
17
48
98
85
96
98
average 93,
(j = 7
namely the recovery of HgO in ICI under conditions of high air flow rate
(1.7 l/min) followed by analysis using aeration directly into the AAS, may
be compared with those given in Table 1.
The latter data (Table 1) show
an overall recovery of about 75 percent and are for the absorption at high
flow rates (1.7 l/min) of dimethyl mercury in ICI followed by analysis using
the dithizone method.
Two variables are involved:
the chemical form of
mercury presented to the ICI collector and the method of analysis.
To check
which variable caused the difference in recovery (75 percent for dimethyl
mercury - dithizone versus 90+ percent for elemental mercury - direct AAS),
portions of the experiments were modified and performed as described in the
following section. (12)
Flow Rate Experiments
An aqueous stock solution of CH3HgCl was prepared by dissolving
8.675 mg of the compound in water and diluting to 1 liter. (13) A similar
stock solution of (CH3)2Hg was prepared by dissolving 6.760 mg of the compound
in 25 ml of ethyl alcohol, then diluting to 1 liter with water. Aliquots of
these stock solutions were subsequently diluted to give working solutions for
use in the experimental runs.
-------
22
The mercury content of the prepared solutions was checked by adding
measured amounts to 25 ml of the iodine monochloride solution, then making the
solution basic, reducing with hydroxylamine hydrochlorid~and aerating through
the AAS.
Initial recoveries were quite low and it was discovered that the
,
organic mercury compounds must react with the iodine monochloride for at least
20 minutes before maximum recoveries were obtained.
A procedure including a
30-minute initial reaction period was adopted, and four runs were made on aliquots
of each of the stock solutions.
Results indicated 85 percent recovery of the
(CH3)2Hg and 96 percent recovery of the CH3HgCI.
The (CH3)2Hg stock solution
was considered to contain 5.75 mg mercury per liter, and the CH3HgCI stock
solution was considered to contain 6.65 mg mercury per liter.
These values
were used to calculate collecting efficiency in the vaporization experiments.
The air calibration assembly shown in Figure I was modified to provide
for a positive air flow rather than the vacuum system previously used.
A diagram
of the modified assembly is shown in Figure 2.
Air was supplied from a compressed
air cylinder, and any mercury vapor which might be present was removed by filtering
through a silver wool plug.
Flow rates were monitored by a rotameter.
To operate, 25 ml of the ICI collecting solution was placed in the
impinger which was then clamped into place.
The lower stopcock on the vaporizer
was closed, the vaporizer cap was removed, and a measured portion of the standard
working solution was transferred to the vaporizer.
The vaporizer cap was then
quickly replaced.
Both vaporizer stopcocks were opened,and the air flow was started
and adjusted to the required flow rate.
The run was continued for about 45 minutes,
after which time the bottom of the vaporizer was heated gently to assure complete
vaporization, and the flow continued for an additional 5 minutes.
Mercury in the
collecting solution was then determined in the usual manner by making it basic with
. .
NaOH, reducing it with hydroxylamine hydrochloride, and aerating it through the AAS.
-------
23
Runs were made with three mercury levels and at two flow rates for
each of the compounds tested.
Results of these runs are reported in
Tables 3 and 4.
TABLE 3.
RECOVERY OF VAPORIZED METHYL MERCURIC CHLORIDE
Mercury
Hg Added, Hg Found, Recovered,
Run Flow Rate nanograms nanograms percent
1 0.2 liter /min 20 16 80
2 0.2 liter/min 50 37 74
3 0.2 li ter /min .50 38 76
4 0.2 li ter /min 100 78 78
5 0.2 liter/min 100 78 78
average 77,cr = 2
6 1.0 li ter /min 20 11 55
7 1.0 liter/min 50 38 76
8 1.0 li ter /min 100 73 73
average 68,cr == 11
TABLE 4.
RECOVERY OF VAPORIZED DIMETHYL MERCURY
Mercury
Hg Added, Hg Found, Recovered,
Run Flow Rate nanograms nanograms percent
1 0.2 liter /min 17 16 94
2 0.2 li ter /min 17 12 71
3 0.2 liter/min 43 33 77
4 0.2 li ter /min 43 37 86
5 0.2 liter /min 129 117 91
6 0.2 li ter /min 129 114 88
li ter /min 17 11 average 85,cr = 9
7 1.0 65
8 1.0 liter/min 17 11 65
9 1.0 li ter /min 43 27 63
10 1.0 liter/min 43 29 67
11 1.0 li ter /min 129 89 69
12 1.0 liter/min 129 91 71
average 67,'3 = 3
-------
24
These experiments clearly show that collection efficiency
may be function of the flow rate.
This dependency may be especially important
when the volatilization method is used to introduce the mercury compound into
the gas stream, because the volatilization has a tendency to occur over a
small time span, not necessarily uniformily throughout the total air flow period.
Furthermore, Tables 3 and 4 show that the flow rate dependency (at
least with the unknown variable of volatilization rate) is not the same for
different compounds.
Methyl mercuric chloride showed no consistent rate
dependency) while dimethyl mercury showed better recovery at the lower air flow
rate.
Thus, on the basis of the data available, it appears that the collection
efficiency for mercury also depends to some extent upon the form in which the
mercury is present in the air, given a constant set of sampling conditions.
Returning briefly to the data presented in Tables land 2, it seems probable
that the difference in recovery can be attributed to the difference of chemical
form, rather than to the method of ana~ysis.
The need for more definitive
experiments to test this conclusion are apparent, and such experiments are
suggested in the recommendations for Future Work section.
As noted earlier, no further work was performed on the analysis of
EPA-collected samples because excessively high values had been obtained for
mercury in those few samples analyzed.
Instead, at the request of the Project
Officer efforts were directed to the evaluation of other systems, as described
in the following sections of this report.
. (7 10 16)
Evaluation of a Commercial Mercury Vapor Detect10n System) ) .
About midway in this program, EPA desired an evaluation of an instru-
ment designed to determine mercury vapor (Rgo) in ambient air.
These studies
are reported in this section.
-------
25
Principle of Operation
The subject of this evaluation was an instrument for atmospheric
mercury vapor detection purchased by EPA and delivered to Battelle's Columbus
Laboratories for evaluation.
This instrument consists of two units, a mercury
converter and a condensation nuclei monitor.
Samples are collected by drawing
air through a sampling cartridge containing a plug of silver wire.
The
cartridge is then placed in the mercury converter and connected to a source of
compressed air.
The collected mercury is released by heating, and the resulting
mercury vapor is irradiated by ultraviolet light forming submicroscopic particles
of HgO.
These nuclei are carried by the air stream into the condensation nuclei
monitor where they are passed through a cloud chamber.
Water vapor condenses on
the nuclei, and droplets are formed which may be counted by optical and electronic
circuits.
Preliminary Evaluation
The instrument as received at Battelle's Columbus Laboratories was
first used to measure mercury collected in laboratory air.
Consistent dupli-
cation of results on samples taken simultaneously could not be achieved; after
two weeks, the instrument failed to give a response to any samples introduced
into the system.
A representative of the manufacturer was informed of the difficulty,
and he came to Battelle to inspect the instrument and to try to resolve the
difficulties.
He found that the ultraviolet lamp was coated by a thin film.
This film was removed, but during the operation the aperture regulating the
amount of effective ultraviolet energy in the system was altered slightly.
When the background count remained high, he also removed a small humidifying
unit from the converter unit, explaining that too much water vapor at this
point in the system sometimes led to high counts.
-------
26
The mercury vapor detection system, altered as explained above, was
used for the tests described in this report.
The instrument's sensitivity was
drastically lower than indicated by the operating manual.
Standards were not
reproducible consistent1~and over a five-day period of testing there were
three occasions when the instrument failed to operate.
Evaluation Using Permeation Tubes
An apparatus employing a permeation tube was constructed to supply a
constant amount of mercury vapor in a gas stream.
A diagram of the apparatus
is shown in Figure 3.
The permeation tube was constructed by placing about 10 grams of
mercury in silicone rubber tubing (1.25-cm ID, 0.48-cm wall thickness [1/2-inch
ID, 3/16-inch wall thickness]) plugged on both ends with Teflon.
The length of
tubing between the two end plugs was approximately 2.2 cm (7/8-inch).
The permeation tube and a thermometer were placed in the water-jacketed
condenser.
The circulating water in the condenser jacket was maintained at 3f c.
Nitrogen gas was passed through the constant temperature bath and then through
the condenser at a rate of 220 m1/min.
After allowing the system to equilibrate for 24 hours, it was calibrated
by passing the nitrogen gas into impingers containing 25 m1 of iodine monoch1oride
solution, timing each collection with a stopwatch.
The collected mercury in each
impinger was then determined by making the solution basic with NaOH, reducing with
hydroxylamine hydrochloride, and aerating through the ultraviolet absorption system.
Samples were collected for 5, 7, and 10 minutes. An average of three runs indicated
3
a mercury vapor flow of 15 I 1 nanograms per minute (~ 70 ug/m ) from the permeation
tube assembly.
Four silver-plug sampling cartridges, designated 103, 106, 110, and 114,
were selected for the initial phase of the study.
Samples of mercury were collected
by connecting a cartridge directly to the gas stream passing through the permeation
-------
27
tube assembly.
Each sampling period was timed with a stopwatch.
Since the
working range of the instrument was expected to be approximately 0.2 to 1.0
nanograms of mercury, the first samples were taken for intervals of 5 seconds
to give approximately 1.3 nanograms of mercury.
These samples gave no response
on the instrument.
Subsequent samples collected for 10, 15, 20, 30, and 60
seconds, giving a range of approximately 2.5 to 15 nanograms of mercury, also
failed to produce a response on the instrument.
Sampling times were increased;
an~ finall~ two samples collected for 5 minutes, representing approximately
75 nanograms of mercury, produced responses above the background count.
("Normal"
ambient air contains perhaps l/lOOth of that amount; the exact figure is still
uncertain.)
Cartridges 110 and 114 were selected for subsequent tests and the
instrument was operated for a week.
Results of the test runs are reported in
Table 5.
It should be noted that the response of the instrument, registered as
the number of nuclei per milliliter, is a logarithmic function of mercury con-
centration.
Background readings were between 800 and 1400 counts.
During the first three days of evaluation samples containing the
same amount of mercury (300 nanograms) were prepared and measured on the instru-
men t .
Results varied widely, and on two occasions the program was interrupted
by failure of the instrument to give any response.
In each instance the instru-
ment was turned off; when reactivated later, it registered counts once more.
On the fourth day of testing, samples were collected for four different
time intervals, and an attempt was made to draw a calibration curve.
As shown in
Figure 6 the results obtained with cartridge 110 produce a smooth calibration
curve.
However, cartridge 114 produced scattered data, especially at the 225
nanograms level, and a calibration curve could not be drawn.
-------
I -
I
28
TABLE 5.
RESULTS OF EVALUATION
Sampling Mercury
Cartridge Time, Added,
Date Number minutes nanograms Count
September 20, 1971 110 20 300 7,500
110 20 300 16,500
110 20 300 9,000
September 21, 1971 110 20 300 68,000
110 20 300 2,600
110 20 300 2,400
110 20 300 No response
September 22, 1971 110 20 300 400,000
114 20 300 25,500
110 20 300 180,000
114 20 300 15,500
110 20 300 145,000
114 20 300 No response
110 20 300 215,000
114 20 300 40,000
110 20 300 275,000
114 28 420 150,000
110 22 330 400,000
September 23, 1971 114 23.5 352 110,000
110 20 300 220,000
114 15 225 10,000
110 15 225 96,000
114 10 150 4,600
110 10 150 29,000
110 5 75 3,900
114 5 75 1,100
114 15 225 40,000
114 15 225 120,000
September 24, 1971 110 20 300 1,300
114 20 300 No response
I.
-------
1000
100
ro
I
0
)(
U)
-
C
:3
0
U
~
C'
C
"0
0
Q)
a::
-
c
Q)
E
:3
~ 10
-
U)
c
10
29
x
Cartridge 110
Cartridge 114
o
o
~
o
100
200
Mercury Added I nanograms
FIGURE 6.
RESULTS OF CALIBRATION RUNS
0/
o
300
-------
30
On the following day two samples were run, each containing 300 nano-
grams of mercury.
The registered count on the first sample was barely above
background, and no counts were registered with the second sample.
At this point,
evaluation was halted.
The mercury vapor detection system was returned to the manufacturer
on October 19, 1971, at its request.
Evaluation of a Mercury Collection Instrument
This work, in which Battelle participated only to the extent of
supplying known samples, is reported here to follow up the mention of this
instrument in a monthly letter report(12) .
Nine gold screens, supplied by the manufacturer, were loaded with
known amounts of mercury using the permeation tube assembly described above
(Figure 3).
These screens were loaded using a battery-powered air suction
pump (sniffer).
Via EPA, these screens were returned to the manufacturer and
analyzed by using their instrument.
The manufacturer's results via EPA are
given in Table 6.
TABLE 6.
RESULTS OF EVALUATION RUNS(a)
Hg Added, Hg Found, Recovery,
nanograms nanograms percent
25 23.4 94
25 22.5 90
25 23.6 94
average 93
50 52 104
50 50 100
50 49.5 99
average 101
100 104 104
100 115 115
100 105 105
average 108
overall average 101
(a) Data via Dr. R. J. Thompson, EPA.
-------
31 .
PHASE IV
Introduction
This phase was initiated late in the program period as part of a
joint concentrated effort with EPA to evaluate a system designed by EPA as an
alternative to the IC1 technique for collecting mercury from ambient air.
As discussed previously, IC1 does collect total mercury but presents analytical
difficulties; other collection devices are not effective for all forms of
mercury.
The operating principle of the method is described in the next section.
Principle of Operation
Mercury is collected (amalgamated) on many metals.
For this ama1gam-
ation to occur, the collecting metal must be clean, i.e., present a surface with
which the mercury can alloy.
The mercury itself must be in an elemental form.
Thus, compounds such as (CH3)2Hg are not collected on metals which form an
amalgam with elemental mercury.
The present co~cept utilizes a pyrolysis furnace in which mercury
compounds are converted into HgO by virtue of the thermal instability of almost
-------
32
all mercury compounds.
(HgO, for example, is decomposed as HgO~ Hg + [0]; at
500°C the partial pressure of oxygen over HgO is about 980 rom.)
The products of
pyrolysis, including HgO vapor, are passed through a ,collector maintained at
a temperature near ambient.
The collector is a Pyrex, or preferably Vycor,
tube packed with silver wool with which the mercury amalgamates.
The experi-
ments designed to evaluate this concept for collection efficiency and for inter-
ferences are described in the next sections.
~XPERIMENTAL WORK
Construction of EQuipment(17)
Twelve collector tubes were constructed of 7-mm Pyrex tubing with 12/5
male and female ball joints on each end, respectively.
These tubes duplicate,
as nearly as possible, the collector tube used by EPA in concurrent experiments
on this concept.
Each was wound with 100 cm (3 feet) of 22-gage Chromel A wire
and loaded with 1 gram of silver wool cleaned at 700° C before use.
A drawing of
the collector tube is shown in Figure 7.
Two catalyst pyrolysis tubes were constructed of 7-mm Vycor tubing with
12/5 male and female ball joints on each end, respectively.
These tubes are 20 cm
(8 inches) in length with identations at 5 em (2 inches) and 12.5 em (5 inches)
from the inlet end to aid ~n holding the catalyst material and provide a 8 cm
(3 inch) heating zone.
The tubes were packed alternately with CuO, silver wool,
silver tungstate/MgO, and CuO.
A small plug of quartz wool was added at each
end of the catalyst material to hold it in place.
The heating zone was
wrapped with 150 cm (5 feet) of 22-gage Chromel A wire.
A thermocouple was
placed next to the tube in the middle of the heating zone and the heating zone
was covered with asbestos.
A drawing of the catalyst tube is shown in Figure 8.
-------
33
~ 10 em ---1
I rs em-1 I
: j: wr" ~ i ~
Heating zone
"-
1.0 g silver wool
FIGURE
7.
COLLECTOR
Tube - 6-mm ID Pyrex with 12/5 ball joints.
Heating zone wrapped with 100 em of 22 gage
Chrome1 A wire.
-------
34
7.5cm
5
6
7
FIGURE
8.
PYROLYSIS TUBE
Tube - 7-mm ID Vycor with 12/5 ball joints.
Tube Packing
(1)
(2)
(3)
(4)
(5 )
(6)
(7)
Quartz wool
Copper oxide
Silver tungstate/MgO
0.5 g silver wool
Silver tungstate/MgO
Copper oxide
Quartz wool
Heating zone wrapped with 150 cm of 22-gage
Chromel A wire and covered with asbestos.
-------
35
Calibration and Efficiency Experiments
Calibration by Aeration
The collector tubes were conditioned by passing a stream of nitrogen
through them at a rate of 300 m1 per minute and heating them electrically at
about 3000 C for 30 seconds.
They were then cooled and capped.
The collection and detection systems were standardized by the aeration
technique. Mercury was vaporized from mercuric chloride solutions by bubbling
air through the solution at a rate of 300 m1 per minute, after reduction of Hg2+
to HgO by stannous chloride. This air flow, containing the mercury vapor, was
passed through the collector tubes for 2 minutes.
The collector tubes, containing the absorbed mercury, were then placed
in a 300 m1/minute nitrogen stream leading to the 15-cm (6-inch) absorption tube
mounted on the Perkin-Elmer Model 303 atomic absorption spectrophotometer.
This
absorption tube was centered on the isolated 253.7-nm mercury line generated by
a hollow cathode mercury tube.
The mercury contained in the collector tube was
released by heating to about 3000 C for 30 seconds.
The optical absorbance of
the released mercury vapor was measured by the spectrophotometer and recorded on
a strip chart.
Results of the standardization runs are shown in Table 7.
Replication of results was very good.
At the 20-nanogram mercury level, seven
measurements showed an average deviation of 5.0 percent.
The 13 measurements
made at the 100-nanogram level had an average deviation of 2.6 percent.
Averages of all peak height readings taken at the five concentrations
of mercury (20, 50, 100, 150, 200 nanograms) were converted to absorption units
to provide a more linear curve and were used to construct the standard curve
shown in Figure 9.
This curve was used to check recoveries of mercury in the
experiments that follow.
-------
36
TABLE 7.
STANDARDIZATION OF COLLECTORS BY AERATION
Mercury Added, Peak Height, Absorption
Collector No. Nanograms Chart Division Value
Blank 0.0 none 0.0000
1 20.0 14.0 0.0655
1 20.0 13.0 0.0605
1 50.0 28.0 0.1427
1 50.0 28.5 0.1457
1 100.0 47.5 0.2798
1 100.0 48.0 0.2840
1 100.0 48.5 0.2882
1 150.0 60.0 0.3979
1 150.0 59.0 0.3872
1 200.0 68.5 0.5017
1 200 . 0 67.5 0.4881
Blank 0.0 1.0 0.0044
2 20.0 12.5 0.0580
2 50.0 27.5 0.1397
2 150.0 61.5 0.4145
3 20.0 12.5 0.0580
3 100.0 51.5 0.3143
3 100.0 51.5 0.3143
3 200. 0 68.5 0.5017
4 20.0 13.5 0.0605
4 100.0 48.0 0.2840
5 20.0 12.0 0.0555
5 100.0 50.0 0.3010
6 20.0 12.0 0.0555
6 100.0 48.0 0 . 2 840
7 20.0 12.5 0.0580
7 100.0 47.0 0.2757
8 100.0 4910 0.2924
9 100.0 48.5 0 . 2 882
10 100.0 51.0 0.3098
11 100.0 50.0 0.3010
-------
c
o
-
e- 0.3
o
(/)
.D
~
0.5
0.4
0.2
0.1
FIGURE
w
.......
50
lOO 150
Mercury Added, nanograms
200
9. CALIBRATION CURVE FOR SILVER COLLECTORS
-------
38
Direct Loading of Collectors From Permeation Tube Assembly
A mercury vapor permeation tube assembly was constructed as shown in
Figure 4.
The permeation tube contained about 10 grams of mercury in silicone
rubber tubing (1.25-cm ID, 0.48-cm wall thickness [1/2-inch ID, 3/16-inch wall
thickness)) plugged on both ends with Teflon.
The length of tubing between the
two end plugs was approximately 2.2 cm (7/8 inch).
For this work nitrogen gas
was passed through the system at 100 m1 per minute, and the circulating water in
the condenser jacket was maintained at 26° C.
The calculated permeation rate at
this temperature is 10.7 nanograms of Hg per minute, based on previous calibrations
at 23.5° C and 31° C.
(This is equivalent to 107 ~g/m3. or about 80 ppb Hg.)
After allowing the system to equilibrate, a series of runs was made in
which the collector tubes were connected directly to the outlet and loaded by
the mercury in the gas stream.
Each run was carefully timed with a stopwatch.
The mercury retained by each collector was then released by heating and m~asured
as described in the preceding section.
Results of these runs are shown in Table 8.
Replication again was very good.
Peak height readings on the 3, 6, and 15-
minute runs (32.1, 64.2, and 160.5 ng Hg, respectively) show an average deviation
of 4.4, 3.0, and 1.8 percent, respectively.
The accuracy of the system is indicated
by the 105, 100, and 98 percent average recovery of added mercury on the 3, 6,
and 15 minute runs, respectively.
Mercury Vapor Loading Through the Catalyst Furnace
The first two catalytic pyrolysis furnaces cracked when brought up to
temperature.
Although the cause of these mishaps is not certain, the catalytic
material probably had been packed too tight~y.
A third furnace was constructed as
described above (Figure 2), baked out slowly, and finally brought up to 600°C and
conditioned at this temperature for 3 hours.
The temperature was monitored with
-------
39
TABLE 8. RESULTS OF DIRECT PERMEATION TUBE LOADING
Peak
Time of Run, Height, Absorption Mercury Added, Mercury Found,
Collector Number minutes chart division Value nanograms nanograms
1 3 20.5 0.0996 32.1 31
2 3 21.5 0.1051 32.1 33
3 3 22.5 0.1107 32.1 35
5 3 22.0 0.1079 32.1 35
6 3 20.5 0.0996 32.1 31
7 3 21.0 0.1024 32.1 32
8 3 20.5 0.0996 32.1 31
9 3 23.0 0.1135 32.1 36
10 3 23.0 0.1135 32.1 36
11 3 23.0 0.1135 32.1 36
1 6 35.0 0.1871 64.2 63
2 6 36.5 0.1972 64.2 66
4 6 37.0 0.2007 64.2 66
6 6 37.0 0.2007 64.2 66
7 6 34.0 0.1805 64.2 60
8 6 34.5 0.1838 64.2 61
10 6 35.5 0.1904 64.2 64
11 6 37.0 0.2007 64.2 66
1 15 61.5 0.4145 160.5 157
3 15 61.5 0.4145 160.5 157
5 15 63.0 0.4318 160.5 165
7 15 60.0 0.3979 160.5 150
9 15 59.5 0.3979 160.5 150
11 15 63.0 0.4318 160.5 165
-------
40
a thermocouple-potentiometer and controlled with a Variac.
The furnace was the
attached to the exit_end of the permeation tube assembly, and a collector tube (a
separate one for each test run) was attached to the outlet of the catalytic tube.
The first three runs showed good recovery of added elemental mercury, but the-
recoveries on the next three runs dropped considerably.
A close examination
of the catalyst assembly revealed the presence of fine dust particl~s adhering
to the inside walls of the outlet and of the tube.
This portion of the tube
was cleaned carefully and heated briefly with a flame to drive off any mercury
which might be present.
The system was then reassembled,and the runs were
resumed with the results reported in Table 9 .
The average recoveries were 90
percent for the 3-minute runs, 98 percent for the 6 minute runs, and 95 percent
for the l5-minute runs.
Interference Evaluation
Studies of Possible Interference From S02
A new catalyst unit was constructe~and all catalyst material was
loaded from the entrance end.
This was done to reduce the risk of contaminating
the exit end of the tube with fine powders which might act as mercury vapor
collectors.
A Wheelco furnace controller was substituted for the Variac to
control the temperature of the catalyst zone more closely.
The catalyst unit was attached to the exit of the mercury permeation
tube assembly as before.
An S02 permeation tube with a calibrated weight loss of 0.37
microgram per minute was placed, together with the mercury permeation tube, in
the water jacketed section of the mercury vapor permeation tube assembly (Figure 4 ).
This provided a gas stream with an 802 concentration nearly 35 times that of the
mercury vapor concentration.
Runs of 3, 6, and 15 minutes were made.
The results
obtained are shown in Table 10.
The average recoveries of 102, 96, and 98 percent
-------
41
TABLE
9.
RESULTS OF MERCURY VAPOR LOADING THROUGH
THE CATALYST FURNACE
Peak
Time of Run, Height, Absorption Mercury Added, Mercury Found,
Collector Number minutes chart division Value nanograms nanograms
6 6 35.0 0.1871 64.2 62
7 6 34.0 0.1805 64.2 60
9 6 33.5 0.1772 64.2 58
10 6 35.0 0.1871 64.2 62
11 6 35.5 0.1904 64.2 63
1 6 37.0 0.2007 64.2 66 .
2 6 36.0 0.1938 64.2 64
3 6 37.0 0.2007 64.2 66
4 3 20.5 0.0996 32.1 32
5 3 18.0 0.0862 32.1 27
7 3 19.5 0.0942 32.1 30
1 15 61.0 0.4089 160.5 155
3 15 62.0 0.4200 160.5 160
5 15 59.5 0.3925 160.5 147
9 15 59.Q 0.3872 160.5 145
-------
42
TABLE 10. MERCURY RECOVERY FROM GAS STREAM CONTAINING S02
(S02 Rate = 0.37 Microgram Per Minute)
Peak
Time of Run, Height, Absorption Mercury Added, Mercury Found,
Collector Number minutes chart division Value nanograms nanograms
1 3 20.5 0.0996 32.1 32
2 3 22.5 0.1107 32.1 35
3 3 21.5 0.1051 32.1 33
4 3 20.5 0.0996 32.1 32
5 3 20.0 0.0969 32.1 31
6 3 20.0 0.0969 32.1 31
9 3 20.5 0.0996 32.1 32
10 3 23.0 0.1135 32.1 36
11 3 20.0 0.0969 32.1 31
3 6 35.5 0.1904 64.2 63
4 6 35.0 0.1871 64.2 62
5 6 34.8 0.1860 64.2 61
6 6 34.0 0.1805 64.2 60
9 6 35.5 0.1904 64.2 63
10 15 63.0 0.4318 160.5 165
11 15 62.5 0.4260 160.5 163
1 15 60.0 0.3980 160.5 150
2 15 60.0 0.3980 160.5 150
-------
43
of the added mercury on the 3, 6, and 15 minute runs indicate that 802 at a
concentration of 3.7 micrograms per liter, much higher than would be expected
in ambient air sampling, has little or no effect on the efficiency of the system.
Alteration of Mercury Permeation Tube Assembly.
The mercury vapor
permeation tube assembly described above(17)was modified by
(18)
adding a seco~d temperature controlled chamber and an air stream splitter.
These alterations were made in order to add small measured amounts of 802' H2S,
and NOZ' respectively, to the air stream containing mercury vapor.
assembly is shown in Figure 10.
The modified
Compressed ai~, controlled with a needle valve and monitored by
Rotameter A, flows through the chamber containing the S02' H2S, or NOZ permeation
tube.
A pump draws air, controlled with a needle valve and monitored by Rota-
meter B, through the chamber containing the mercury permeation tube.
By varying
the flow of compressed air through Rotameter A and bleeding off measured amounts
of this flow through Rotameter C, small measured amounts of SOZ' HZS, or NOZ may
be introduced into the air stream carrying mercury vapor.
Standard Curve Based on Peak Area.
During this phase of the experimental
program it became evident that peak areas rather than peak heights provided more
accurate measurements of mercury by atomic absorption.
Consequently, all the
calibration data reported in Table 7 were recalculated as peak areas and were used
to construct the calibration curve shown in Figure 11.
Ideally, of course, the
integrated absorbance (a function of peak area) would be plotted; instead, peak
areas were plotted for expediency.
As noted in the text accompanying the tables
that follow, some recovery data were calculated using the peak height (absorbance)
calibration plot (Figure 9), and some using the peak area calibration plot
(Figure 11) .
The latter, as mentioned above, are believed to be more valid because
some peak broadening was observed, especially upon the addition of "interfering"
-------
Mixing bUlb\
To hood --:::-Jc
-AI -=l r~
I I
Rotameter
(C)
nl
Consta nt
temperature
bath
Circulating water
Permeation tube
Silver-
wool plugs
Thermometer
12/5 ball
joints
Mercury permeation tube
Rotameter
(S)
Air
..
FIGURE 10.
MODIFIED PERMEATION TUBE ASSEMBLY
Rotameter
(A)
~
Compressed
air
+:-
+:-
-------
14
12
10
N
I
0
)(
N 8
E
E
~
0
<1>
L.. 6
~ ~
.:tt: Vt
o
<1>
a..
4
2
50
100 150
Mercury Added, nanograms
200
00
FI CURE 11.
CALIBRATION CURVE FOR SILVER COLLECTORS, BASED ON PEAK AREAS
-------
46
gases - S02' H2S, and N02 - to the gas stream.
Although some of the broadening
of peaks observed during these experiments may have been caused by the gas
additives, it is also possible that small changes in the geometry and/or the
heating efficiency of the collector tubes, caused by repeated heating and cooling,
contributed to this altered response.
Experiments Using Redesigned Equipment.
An S02 permeation tube was
placed in the permeation tube assembly (Figure 10), and mercury collections were
3
made at S02 levels of 180, 120, and 37 ~g/m. Results obtained for the three
sets of runs calculated from the peak area calibration curve (Figurell), are
reported in Table 11. A slight negative bias is indicated with an S02 concentration
of 180 ~g/m3. Concentrations of 120 and 37 ~g S02/m3 appear to have no effect on.
the collection and measurement of mercury vapor.
Studies of Possible Interference From H2S
The equipment setup for the first two experiments was as shown in
Figure, with one of two H2S permeation tubes and the mercury tube together
in the permeation tube section (17). Two runs were made, one with an H2S permeation
tube with a permeation rate of 1.31 micrograms per minute and the other with an
H2S permeation tube with a permeation rate of 1.60 micrograms per minute.
provided gas streams containing 13.1 and 16.0 micrograms of H2S per liter, a
This
concentration about two orders of magnitude higher than would be expected in
ambient air.
Results of the two series of runs are shown in TableU.
These data
were obtained from the absorbance calibration curve, Figure 9.
Even at this high
concentration of H2S the recovery of mercury is about 90 percent.
Experiments Using Redesigned Equipment. An H2S permeation tube was
placed in the temperature controlled chamber shown in Figure 10 (18). Air flows
were adjusted to provide an H2S concentration of 650 ~g/m3 and a mercury concentra-
tion of 107 ~g/m3, and a series of runs was made.
In subsequent runs the H2S
-------
47
TABLE 11. RECOVERY OF MERCURY FROM GAS STREAM CONTAINING S02
Run No.1
3
S02 Concentration = 180 ~g/m
3
Hg Concentration = 107 ~g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collector mm Nanograms Nanograms Percent
1 2.31 32 24
2 2.84 32 30
3 2.74 32 29
4 2.74 32 29
5 2.96 32 31
6 3.20 32 34
9 2.79 32 30
10 2.84 32 30
11 5.61 75 61
1 3.26 43 34
2 3.79 43 40
3 2.79 32 30
6 2.93 32 31 90.0
Run No.2
= 120 3
S02 Concentration ~g/m
3
Hg Concentration = 107 ~g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery
Collec tor mm Nanograms Nanograms Percent
9 3.06 32 32
10 3.42 32 36
11 4.18 43 45
1 2.83 32 30
11 3.14 32 33
2 1.16 11 12
1 3.32 32 35
11 3.00 32 32
1 2.71 32 29 102.0
-------
48
TABLE 11. (Continued)
Run No.3
3
802 Concentration = 37 ~g/m
3
Hg Concentration = 107 ~g/m
Peak Area, Average
nnn2 x 10-2 Hg Added, Hg Found, Recovery,
Co11ec tor Nanograms Nanograms Percent
2 5.26 59 57
3 3.62 32 38
4 3.22 32 34
5 3.55 32 37
6 3.4'7 32 37
9 3.47 64 67
10 6.10 64 68
11 8.83 107 106
11 3.07 32 32
10 3.49 32 37
9 3.21 32 34
6 3.17 32 33
5 5.96 64 64
4 5.98 64 64 100.0
-------
49
TABLE 12. RECOVERY OF MERCURY FROM GAS STREAM CONTAINING H2S
Run No. 1
H2S Concentration = 13.1 3
IJ.g/m
3
Hg Concentration = 107 IJ.g/m
Time of Run, Peak Height, Absorption Hg Added, Hg Found,
CoHec tor Minutes Chart Divisions Value Nanograms Nanograms
1 6 31.5 0.1643 64.2 54
2 6 33.0 0.1739 64.2 57
3 6 35.0 0.1871 64.2 62
4 6 31.5 0.1643 64.2 54
1 6 30.5 0.1580 64.2 52
2 6 32.5 0.1707 64.2 56
4 6 32.5 0.1707 64.2 56
5 6 31.0 0.1612 64.2 52
1 3 19.5 0.0942 32.1 30
2 3 19.0 0.0915 32.1 29
4 3 18.5 0.0888 32.1 29
5 3 18.5 0.0888 32.1 29
6 15 58.5 0.3820 160.5 142
9 15 58.0 0 . 3768 160.5 140
10 15 58.5 0.3820 160 . 5 142
Run No.2
H2S Concentration = 16 IJ.g/m3
3
Hg Concentration = 107 IJ.g/m
Time of Run, Peak Height, Absorption Hg Added, Hg Found,
Co11ec tor Minutes Chart Divisions Value Nanograms Nanograms
1 3 20.5 0.0996 32.1 31
3 3 20.5 0.1221 32.1 31
4 3 17.5 0.0835 32.1 26
5 3 19.0 0.0915 32.1 29
6 3 18.0 0.0862 32.1 27
10 6 34.0 0.1805 64.2 59
11 6 32.5 0.1707 64.2 56
1 6 31.5 0.1643 64.2 54
2 6 32.0 0.1675 64.2 55
3 6 36.5 0.1972 64.2 65
4 15 61.5 0.4145 160.5 157
5 15 57.0 0.3665 160.5 135
9 15 57.5 0.3716 160.5 138
10 15 59.0 0.3872 160.5 143
-------
50
concentration was adjusted to 130, 43, and 17 ~g/m3 for each run, respectively.
Results obtained from the four runs are report in Table 13 and are calculated
from the peak area calibration curve, Figure11. A slight positive bias is
indicated at HZS concentrations of 650 'and 130 ~g/m3. However, HZS concentra-
3
tions of 43 and 17 ~g/m seem to have no effect on the collection and measurement
of mercury vapor.
Studies of Possible Interference From NOZ
An NOZ permeation tube was placed in the proper controlled temperature
chamber (Figure 10), and sets of collection runs were made at N02 concentrations
3
of 100, 50, and 25 ~g/m .
Results of the three sets of runs are reported in
Table: 14 and are calculated from the peak area calibration curve, Figure 11.
No .
interference was noted in the collection and measurement of mercury vapor at any
of these NOZ concentrations.
Mercury Collection From Dimethyl Mercury Vapor
A standard stock solution of dimethyl mercury was prepared by
dissolving 114 mg of (CH3)ZHg in 500 ml of ethanol to give a solution containing
198.3 ~g Hg per mI.
A standard working solution was prepared by diluting 1.0 ml
of the stock solution to 100 ml with ethanol, giving a solution containing
1.983 ~g Hg per mI.
Four runs were made using the vaporizer assembly shown in Figure 12.
In each run 50 microliters of the working solution, containing 99.Z nanograms
of mercury, were introduced into the vaporizer.
The flow rate was controlled
at 100 ml per minute and continued for 30 minutes.
Results of the four runs
are reported in Table 15. No indication of mercury collec tion was noted.
-------
51
TABLE 13.
RECOVERY OF MERCURY FROM GAS STREAM CONTAINING ~ S
Run No. 1
= 650 3
H2S Concentration ~g/m
3
Hg Concentration = 107 ~g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collector rom Nanograms Nanograms Percent
1 3.95 43 42
2 3.22 32 34
4 3.58 32 38
5 3.60 32 38
9 3.63 32 38
10 3.43 32 36
11 3.29 32 35 111.0
Run No.2
H2S Concentration = 130 3
~g/m
3
Hg Concentration = 107 ~g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collec tor mm Nanograms Nanograms Percent
1 3.75 32 40
2 3.09 32 33
3 3.46 32 36
4 3.84 32 40
5 3.37 32 36
6 3.19 32 34
9 3.35 32 35
10 3.51 32 37
11 6.05 64 67
1 11. 79 161 169 110.0
Run No.3
43 3
H2S Concentration = ~g/m
3
Hg Concentration = 107 ~g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collec tor rom Nanograms Nanograms Percent
1 2.89 32 31
2 2.98 32 32
3 3.59 32 38
4 3.03 32 32
5 3.16 32 34
6 5.52 64 60
9 5.82 64 64
10 5.78 64 64
11 11.47 161 161 101. 00
-------
52
TABLE 13.
(Continued)
Run No.4
H2S Concentration = 17 ~g/m3
Hg Concentration = 107 ~g/m3
Peak Area, Hg Added, Hg Found,
2 -2
CoHec tor mm x. 10 Nanograms Nanograms
3 5.76 64 63
4 4.03 43 43
5 3.00 32 32
6 8.64 107 104
Average
Recovery,
Percent
98.8
-------
53
TABLE 14.
RECOVERY OF MERCURY FROM GAS STREAM CONTAINING N02
Run No.1
3
N02 Concentration = 100 IJ.g/m
3
Hg Concentration = 107 IJ.g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collector rom Nanograms Nanograms Percent
1 2.79 32 30
2 3.27 32 34
3 3.83 32 40
4 3.51 32 37
5 3.05 32 32
6 3.33 32 35
9 9.61 107 120
10 4.20 43 45
11 2.87 32 30
10 3.01 32 32
9 3.05 32 32
6 5.83 64 64
6 5.73 64 63
5 5.93 64 65 105.0
Run No.2
3
N02 Concentration = 50 IJ.g/m
3
Hg Concentration = 107 IJ.g/m
Peak Area, Average
2 x 10-2 Hg Added, Hg Found Recovery,
Co11ec tor rom Nanograms Nanograms Percent
11 3.52 32 37
10 3.23 32 34
9 4.19 43 45
6 11.53 161 162
4 3.81 32 40
3 3.82 32 40
2 6.54 64 73
1 8.32 118 99
11 5.91 64 65 103.0
-------
54
TABLE 14.
(Continued)
Run No.3
. N02 Concentration = 25 IJ.g/m3
3
Hg Concentration = 107 IJ.g/m
Peak Area, Hg Added, Hg Found,
Co11ec tor mrn2 x 10-2 Nanograms Nanograms
6 3.24 32 34
5 3.50 32 37
4 3.19 32 34
3 3.25 32 34
1 5.33 64 58
11 6.02 64 66
10 6.32 64 70
Average
Recovery,
Percent
104.0
-------
Rotameter
Silver wool
55
Teflon stopcocks
Vaporizer
12/5 boll joints
Silver wool
collector
FIGURE 12. VAPORIZER ASSEMBLY
To pump
..
-------
56
TABLE 15.
MERCURY COLLECTION FROM DIMETHYL MERCURY VAPOR
(Silver Wool Collectors Without Pyro1yzer)
Average
Time of Run, Hg Added, Hg Found, Recovery,
Collec tor Minutes Nanograms Nanograms Percent
1 15 99 None
2 15 99 None
3 15 99 None
4 15 99 None 0.0
-------
57
. Collection of Mercury From HgCl2 Particulate Using pyrolyzer(19)
A stock solution of mercuric chloride was prepared by dissolving
135.4 mg of HgC12 in 500 ml of ethyl alcohol, giving a concentration of
200 micrograms of Hg per mI.
A working solution containing 2 micrograms of
Hg per ml was prepared by diluting 1 ml of the stock solution with ethyl
alcohol to a volume of 100 mI.
An empty Pyrex collector tube was modified into a vaporizing chamber
by heating and blowing out a small buldge in the tubing wall to form a small
well to hold the standard HgC12 solution.
A thermocouple was taped to the
wall, and the whole tube was wrapped with heating tape and was placed at the
entrance end of the pyrolyzer unit.
A silver wool collector was attached to
the exit.
Using a microsyringe, a measured amount of the working solution was
placed in the well of the vaporizing chamber.
A silver wool plug was inserted
in the entrance end of the assembly and air was drawn through at a rate of
100 ml per minute.
After 10 minutes the heating tape was activated and adjusted
to give a temperature of 150 C to sublime the (dried) HgC12 remaining in the tube.
The run was continued for an additional 20 minutes after which the collected
mercury was determined in the usual manner.
The results obtained from seven
separate runs are shown in Table 16.
These data indicate that the pyrolyzer is
effective in releasing mercury vapor from HgCl2 particulates.
Collection of Mercury From (CH3)2Hg Using Pyrolyzer at Several Temperatures
A standard stock solution of dimethyl mercury was prepared by dissolving
114 mg of (CH3)2Hg in 500 m1 of alcohol to give a solution containing 198.3 ~g Hg
per mI.
A standard working solution was prepared by diluting 1.0 ml of the stock
solution to 100 ml with alcohol, giving a solution containi~g 1.983 ~g Hg per mI.
-------
58
TABLE 16.
COLLECTION OF MERCURY FROM HgC12 PARTICULATE USING PYROLYZER
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collector mm Nanograms Nanograms Percent
9 7.77 100 90
9 8.98 100 109
9 7.98 100 93
9 8.25 100 97
9 4.35 50 47
9 4.69 50 50
9 1.93 20 20 97.0
-------
59
A series of runs were made using the vaporizer-pyro1yzer assembly
shown in Figure 13~
In each run a measured amount of the working solution was
placed in .the vaporizer by means of a microsyringe.
The flow rate was controlled
at 100 m1 per minute and continued for 30 minutes.
At the end of each run the
collector was removed and the collected mercury was determined in the usual
manner.
The. results of runs made with pyro1yzer temperatures of 500°C, 550°C,
and GOOoC are shown in Table 17.
The average recovery of mercury at each pyro1yzer
temperature was more than 100 percent.
These data indicate no significant
differences in pyro1yzer efficiency over the temperature range of 500°C to GOO°C.
Construction Details of the Pyrolysis Tube and Furnace
Details of the pyrolysis tube and heating unit used in the studies
reported above are shown in Figure 14.
Since the data indicate that close tempera-
ture control may not be necessary to obtain reliable results, a simple step-down
transformer might be used as a power supply for a field-test unit.
However, the
varying heat losses which may be encountered in field testing situations have not
been determined.
Some changes in the amount and type of insulation may be
necessary to accommodate all conditions.
DISCUSSION AND CONCLUSION
This research program dealt primarily with two systems for the collection
and analysis of mercury collected from ambient air.
The first method utilizes a
solution of IC1
and was chosen by EPA for study because at the initiation of the
program it appeared to be the best of the existing methods.
This judgment was
confirmed by the current research program.
With proper care, principally in main-
taining relatively low air flows, the collection of mercury and (at least some)
of its compounds is quantitative.
This fact alone makes the IC1 solution system
unique and valuable.
-------
Rotameter
Silver
wool
60
Vaporizer
FIGURE 13.
Teflon stopcocks
VAPORIZER-PYROLYZER ASSEMBLY
To pump
: IW/A :--.
~COliector
Pyrolyzer
-------
61
TABLE 17.
COLLECTION OF MERCURY FROM (CH3)2Hg USING PYROLYZER
Run No. 1 - Pyro1yzer Temperature = 600°C
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Collector nun Nanograms Nanograms Percent
6 9.37 99.2 115
9 8.39 99.2 100
10 9.09 99.2 111
11 8.74 99.2 105 109.0
9 4.76 49.6 51
10 4.76 49.6 51
9 4.73 49.6 51
6 4.93 49.6 53 104.0
10 2.02 19.8 21
9 2.03 19.8 21
10 2.00 19.8 21 106.0
Run No. 2 - Pyro1yzer "Temperature = 550 C
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Recovery,
Co11ec tor nun Nanograms Nanograms Percent
9 4.88 49.6 53
4 9.44 99.2 114
9 2.03 19.8 21 111.0
Run No. 3 Pyro1yzer Temperature = 500 C
Peak Area, Average
2 x 10-2 Hg Added, Hg Found, Rec ove ry ,
Co11ec tor rom Nanograms Nanograms Percent
4 2.01 19.8 21
9 8.64 99.2 103
4 4.86 49.6 52
11 5.04 49.6 54
4 8.83 99.2 106 106.0
-------
62
2
3
4
5
6
7
FIGURE 14.
PYROLYSIS TUBE
Tube:
7-mm ID Vycor with 12/5 ball joints.
Tube Packing
(1) Quartz wool - small plug
(2) Copper oxide - 0.5 g
(3) Silver tungstate/MgO - 0.4 g
(4) Silver wool - 0.5 g
(5) Silver tungstate/MgO - 0.4 g
(6) Copper oxide - 0.5 g
(7) Quartz wool - small plug
Heating zone wrapped with 150 cm of 22-gage Chromel A
wire and covered with asbestos to provide 1.2-cm
insulation. With this configuration a current of about
5 A (20 VAC input, 100 watts) through the Chrome1 A wire
heats the tube packing to 600 I 25 C under laboratory
conditions.
-------
63
- Analytical procedures to determine the mercury collected by ICI have
been developed and are reasonably reliable.
A dithizone extraction method is
not recommended from the standpoint of accuracy and reliability at low mercury
levels; it also suffers from the disadvantage of being slow and tedious.
Double
aeration, if carefully practiced, gives good results; but poor recovery is noted
occasionally, with no apparent explanation.
Apart from the extra time required
for two aerations, the method is generally satisfactory.
Were it not for the method discussed below, the ICI solution collection
scheme would be the logical candidate for additional study, and perhaps for
adoption as the preferred means to collect total atmospheric mercury.
The method that, in experiments described in this report, has proven
superior to the ICI collection system is based upon the collection of mercury
o
(Hg ) vapor on silver wool.
For total atmospheric mercury collection, the
collector is preceded by a catalytic pyrolysis furnace that converts inorganic
particulates (HgCI2) and organometallics [(CH3)2Hg] to Hgo.
reported here, using these two mercury compounds, the system is well-suited for
Based on the studies
its intended purpose, total mercury collection.
Furthermore, common gaseous
pollutants that might be suspected poisons for the silver wool collector have
been shown to produce no appreciable lessening in the collection efficiency.
Finally, the mercury is collected in a form (silver amalgam) readily adaptable
to the rapid flame less atomic absorption analysis for mercury.
Considering these
facts, the pyrolysis-silver wool system appears, at this time, to be the method
of choice for total mercury collection.
FUTURE WORK
It is recommended, that several aspects of the two major systems receive
additional attention.
-------
64
(1)
Iodine monochloride system.
A fundamental study is needed to
clarify the chemistry involved in the collection of mercury compounds and the
subsequent release of mercury.
The role of complex ions should be understood.
The kinetics of absorption (complex formation?) appeared in this program to be
a probable cause of incomplete recovery, but the data were not definitive.
If
ICI is chosen as a primary or back up (referee) method for total mercury
collection, these basic studies should be made.
The actual analysis appears satisfactory and improvements should be
in time and cost efficiency, with perhaps some work being directed toward
making the aeration method completely reliable.
(2)
Pyrolysis and silver wool collection.
This collection method
offers the most promise at this time.
A proposal has been submitted to EPA
in reply to RFP DU-72-B406 to develop a prototype instrument for field collection
of atmospheric mercury.
This proposal is based on the pyrolysis-silver wool
system
and acknowledges the source (Drs. Thompson and Scott) of the concept.
If accepted by EPA, the work proposed should answer most of the remaining
questions about the validity of the pyrolysis-silver wool method.
-------
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
65
REFERENCES
(1)
Chase, Dan L., Huber, Frank E., and Henry, William M., "Development and
Evaluation of an Analytical Method for the Determination of Total
Atmospheric Mercury", Contract No. ERSD 71-32, Environmental Protection
Agency.
- First Monthly Report, December 28, 1970.
(2)
(3)
- Second Monthly Report, January 26, 1971.
- Third Monthly Report, February 26, 1971.
(4)
(5)
- Fourth Monthly Report, March 26, 1971.
- Fifth Monthly Report, April 27, 1971.
(6)
(7)
- Sixth Monthly Report, May 26, 1971.
- Seventh Monthly Report, June 15, 1971.
(8)
(9)
- Eighth Monthly Report, July 22, 1971.
- Ninth Monthly Report, August 26, 1971.
- Tenth Monthly Report, September 29, 1971.
- Eleventh Monthly Report, October 25, 1971.
- Twelfth Monthly Report, November 30, 1971.
- Thirteenth Monthly Report, December 22, 1971.
- Fourteenth Monthly Report, February 2, 1972.
- Fifteenth Monthly Report, February 25, 1972.
- Summary Report, 'September 2, 1971.
Chase, Dan L., Henry, William M., and Sgontz, D. L., "Development and
Evaluation of an Analytical Method for the Determination of Total
Atmospheric Mercury", Contract No. EHSD 71-32, Environmental Protection
Agency.
- First Progress Report, Phase IV, March 20, 1972.
- Second Progress Report, ~hase IV, March 24, 1972.
- Third Progress Report, Phase IV, April 4, 1972.
Linch, A. L., Stalzer, R. F., and Lefferts, D. T., "Methyl and Ethyl
Compounds - Recovery From Air And Analysis", Am. Ind. Hygiene Assoc. J.
29 (No.1), 79 (1968).
-------
(21)
(22)
(23)
66
REFERENCES (Continued)
O'Keefe, A. E. and Ortman, G. C., "Primary Standards for Trace Gas
Analysis", Anal. Chern. 32, p 760 (1966).
Hatch, W. Ronald and Ott, Welland L., "Determination of Sub-Microgram
Quantities of Mercury by Atomic Absorption", Anal. Chern. 40 (No. 14),
p 2085 (1968). .
Monkman, J. L., Mafett, Patricia A., and Doherty, T. F., "The
Determination of Mercury in Air Samples and Biological Materials",
Am. Ind. Hygiene Assoc. Quart. J. !l (No.4), p 418 (1956).
------- |