EPA-660/2-74-086
DECEMBER 1974
Environmental Protection Technology Series
Mercury Recovery From Contaminated
Waste Water and Sludges
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. Environmental Protection Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY STUDIES series. This series describes research
performed to develop and demonstrate instrumentation, equipment
and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This report has been reviewed by the Office of Research and
Development, EPA, and approved for publication. Approval does
not signify that the contents necessarily reflect the views and
policies of the Environmental Protection Agency, nor does mention
of trade names or commercial products constitute endorsement or
recommendation for use.
-------
EPA-660/2-74-086
December 1974
MERCURY RECOVERY FROM CONTAMINATED WASTE
WATER AND SLUDGES
By
Richard Perry
Project 12040 HDU
Program Element 1BB037
ROAP/TASK No. 21 AZX/022
Project Officer
Ralph H. Scott
Pacific Northwest Environmental Research Laboratory
National Environmental Research Center
Corvallis, Oregon 97330
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH & DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
For sale by the Superintendent of Documents. U.S Government Printing Office
Washington, D.C. 20402 - Stock No. 5501-00972
-------
ABSTRACT
A system was designed, installed and operated to
recover mercury (Hg) from waste water and sludge produced by
a mercury cell chlor-alkali plant. Hg content of the waste
water ranged from 300 - 18,000 ppb while Hg content of the
brine sludge ranged from 150 - 1500 ppm Hg. Deposits from
the waterway near the plant outfall were also processed.
From a variety of removal techniques evaluated,
sulfide precipitation was selected for process water treat-
ment and high temperature roasting for sludge treatment.
The sulfide precipitation system steps include
collecting the process water streams, adjusting the pH to 5-
8 with spent sulfuric acid, settling the large particles in
a surge tank, adding sodium sulfide to a 1 - 3 ppm excess,
adding diatomaceous earth at the rate of 0.7 gpl (0.62 lb/
1000 gal), and filtering through an R. P. Adams pressure
filter. The effluent Hg levels range from 10 - 125 ppb and
average 50 ppb Hg, an 87 - 99% removal, averaging 97%• The
44.8 m2 (169 ft2) filter processes up to 380 1/min (100 gpm)
with an approximate 48-hour cycle time between backwashings.
Capital costs totaled $143,900 and operating costs average
50 /3785 1 (1000 gal).
The sludge treatment system includes a collection
system, a 3.7 m (12 ft) diameter thickener, a 1.8 m (6 ft)
diameter rotary vacuum filter, a 1.37 m (4.5 ft) i.d. multi-
ple hearth furnace, and 3 stainless steel condensers 21 m2
(224 ft2 ) each. Processing rate for the sludge is 140 -
320 kg/hr (300 - 700 Ib/hr), dry basis. At present, ap-
proximately 18 m tons (20 s tons) of sludge per month are
processed. Operating temperatures range from 540°C - 760°C
(1000°F - 1400°F), feed Hg content ranges from 290 - 440
ppm Hg (dry basis), and clinker Hg content after treatment
varies from 0.5 - 7.2 ppm Hg, for a removal rate of 98.3 -
99-8$ Waterway sediments containing 12.8 ppm were roasted
at 750°C (1350°F) and the clinker contained 0.95 -1.7 ppm
Hg, for an 87 - 92$ removal. Capital costs totaled $364,500
and operating costs are $32/m ton ($35/s ton) of dry sludge
treated.
This report was submitted in fulfillment of
Project Number 12040 HDU by the Georgia-Pacific Corporation,
under the partial sponsorship of the Environmental Protec-
tion Agency. Work was completed in April, 1974.
111
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
300 S.W. 35TH ST.
CORVALLIS, OREGON 97330
February 7, 1975
Dr. Ho L. Young, Chemist
U.S. Environmental Protection Agency
6ZO Central Avenue
Alameda, CA 94501
Dear Dr. Young:
In response to your recent request, I am enclosing a copy of
our publication entitled "Mercury Recovery From Contaminated Waste
Water and Sludges," hPA-660/2-74-086.
Thank you for your interest and please call us if we can be of
further assistance.
Sincerely,
Chris L. West, Director
Public Affairs Office
Enclosure
-------
CONTENTS
Sections Page
I Conclusions 1
II Recommendations 2
III Introduction 3
IV Process Design 7
V Construction 44
VI Operation and Evaluation 48
VII Discussion 55
VIII References 76
IX Patents and Publications 81
X Glossary 82
XI Appendices 84
-------
FIGURES
No. Page
1 The Bellingham Chlor-Alkali Plant and Brine 4
Sludge Pond
2 The Bellingham Chlor-Alkali Plant with Sludge 5
Pond and Hg Recovery Structure
3 Lab Kiln Test Assembly for Roasting Sludge 10
4 Hg Contents of Untreated Brine Sludge After 12
Roasting in Lab and Pilot Kilns
1 R
5 Cut Away Drawing of the Internals of a Multiple ±0
Hearth Furnace Showing the Rake Arms and Hearth
Assembly
6 Acid Treatment of Brine Sludge Before Roasting 20
in Lab and Pilot Kiln
7 Eimco Rotary Vacuum Pilot Filtration of Brine 2^
Sludge
8 Proposed Brine Sludge Handling System '
9 Solubility of HgS in Excess S= 33
10 Schematic of the Reduction Method of Hg Re- ^5
moval from Water
•?£
11 Schematic of the Ion exchange Process for Hg J
Removal from Water
12 Schematic of Activated Carbon Hg Removal from 3'
Waste Water
13 Schematic of the Sulfide Precipitation Process 3
for Waste Water
14 Proposed Sulfide Precipitation System 39
IIP
15 Data from Pilot Tests with 11 Sq. Ft. Niagara
Filter
16 Pilot Filter Tests to Determine Cycle Length ^
with and without Precoat and Body Feed
-------
FIGURES (cont.)
No. Page
17 The Model of the Hg Recovery System ^6
18 Installed Brine Sludge Handling System 51
19 Lab Tests on Mercury Removal from Water Using 62
Sulfide ppt., Activated Carbon, Ion Exchange
Resins and Reduction Chemicals
20 Installed Sulfide Precipitation System for Water 64
Treatment
21 Installed 6' x 6' Eimco Rotary Vacuum Filter 65
for Brine Sludge Dewatering ,
22 Installed 54" i.d. BSP Multiple Hearth Furnace 66
23 Installed R. P. Adams Filter for HgS Removal 6?
24 The 12' x 6' Sludge Thickener Prior to the 68
Rotary Vacuum Filter in the Sludge Treatment
System
25 The Full Scale Hg Recovery System as Installed 70
at the Bellingham Chlor-Alkali Plant
26 Mix Tank in the Sulfide Precipitation System 71
where the D.E. and Sulfide are Added
27 Particle Distribution in Brine Sludge which was 73
Washed and Screened to Remove Particles 0.007"
diameter (Experiment 34)
28 Bench Test Set-ups for Chemical Oxidation of 87
Sludge
29 Effect of Staging on Mercury Recovery by 89
Tokawa
VI
-------
TABLES
No. Page
1 Kiln Treatment of Brine Sludge & Graphite . 8
2 Kiln Roasting of Sludge, No Chemical Treatment H
3 Rotary Calciner Roasting of Brine Sludge 13
4 Multiple Hearth Furnace Roasting of Brine Sludge 1^
BSP 76 cm (30 in) Pilot Furnace
5 Multiple Hearth Furnace Roasting of Brine Sludge 15
BSP 76 cm (30 in) Pilot Furnace
6 Multiple Hearth Furnace Roasting of Brine Sludge 17
BSP 76 cm (30 in) Pilot Furnace, ICI Data
7 Kiln Roasting of Chemically Treated Brine Sludge 19
8 BSP Test on Batch Kiln 21
9 Test Results from Barrett Centrifuge on De- 23
watering of Brine Sludge by Centrifuging
10 Data from the Eimco Pilot Rotary Vacuum Filter 24
11 Lab Test Data on Sulfide Precipitation for Hg 30
Removal from Water
12 Hg Levels in Water After Exposure to Sulfide Ion 31
for 30 Seconds to 10 Minutes
13 Partial List of Data from Start-up of the Full 50
Scale Sulfide Precipitation System
14 Data from Start-up of the Full Scale Brine Sludge 53
Treatment System
15 Names and Addresses of the Companies Contacted 56
for Information by Direct Communication During
the Project
16 Data from Literature on Ion Exchange Resins ^
17 Lab Test Data on Ion Exchange Resin for Hg 59
Removal from Water
vn
-------
TABLE'S, (cont. )
No. Page
18 Lab Test Data on Activated. Carbon for Hg Removal
from Water
19 Lab Test Data on Hg Removal from Water by Reduc-
tion
20 Cost Estimate - Water Treatment System ^
21 Cost Estimate - Sludge System ^5
22 Oxidation of Brine Sludge Using Sodium Hypo- ^°
chlorite
23 Oxidation of Brine Sludge Using Sodium Hypo- 88
chlorite "Work Performed at University of
British Columbia
24 Oxidation of Brine Sludge Using Chlorine Gas 91
25 Oxidation of Brine Sludge Using Combinations 92
of Hypo, Chlorine, Electrolytic Acid Treatment
and Roasting
26 Hg Analysis of Brine Sludge Size Fractions 94
27 Hg Removal Rates Necessary for Various Size
Chlorine Plants to Achieve 45 gm (0.1 Ib)
Per Day Mercury in the Effluent
28 Comparison of Substances Used or Considered
for Reducing Mercury Ion in Solution
Vlll
-------
ACKNOWLEDGEMENTS
The research, process selection, pilot tests,
analytical work and report preparation were performed by
a team of chemists, engineers and technicians at Georgia-
Pacific Corporation, Bellingham Division, consisting of
Dr. Scott Briggs, Ed Dahlgren, Luther Dunn, Karen Hulford,
Dick McLeod, Dick Perry and Don Rachor. Research assist-
ance was also provided by Dr. Bill Groves of Vancouver,
B. C., Canada.
Engineering design and construction supervision
of the full-scale plant were provided by Lynn Baker, Ivan
Campbell and Hal Henkel of G-P.
The start-up team for the full-scale system con-
sisted of Steve Baklund, Steve Earp and Bruce Swanson of
G-P. Don Elliot and Don Wines coordinated operational
aspects.
The support of the Project Officer, Ralph H.
Scott, and Director, N. A. Jaworski, EPA Pacific Northwest
Environmental Research Laboratory in Corvallis, Oregon is
gratefully acknowledged.
IX
-------
SECTION I
CONCLUSIONS
1. Sulfide precipitation offers several advantages over
other methods of Hg removal from water: (a) fewer
process steps, (b) pH range compatible with total
plant effluent, (c) concentrated Hg products, (d) in-
expensive chemicals used, and (e) minimal environ-
mental stress.
2. In the laboratory, sulfide treatment achieved 99.9$
removal of Hg from solutions containing 10 - 100 ppm Hg.
3. In the plant, the sulfide process achieved 87 - 99-2%
removal from solutions containing 0.3-6 ppm Hg. The
average effluent Hg content was 50 ppb.
4. The major problem experienced with the sulfide process
was pH control. With concentrated sulfuric acid, a
two-stage addition system was needed.
5. Sulfide system capacity is 380 1/min (100 gpm). Capital
costs totaled $1*13,900, and operating costs of the
sulfide system are 13
-------
SECTION II
RECOMMENDATIONS
1. Sludge from other chlor-alkali plants should be
roasted to determine the efficiency of this process
on various wastes. Also, sludges from other indus-
tries and municipal sewage plants which contain Hg
should be tested.
2. The Hg recovery from the air leaving the furnace re-
quires further work to solve the dust removal problem.
3. To offset the operating cost of the sludge process,
potential uses for the calcium and magnesium oxide in
the clinker should be investigated.
*J. A further step in the water treatment process would
be the design of a polishing filter to remove most of
the remaining Hg in the filter effluent.
-------
SECTION III
INTRODUCTION
The industrial hygiene problems associated with
Hg vapor, and both inorganic and organic Hg compounds have
long been recognized and safeguards have been developed to
avoid harmful exposures. The situation changed dramatic-
ally with the publication in 1968 of the biological con-
version of inorganic Hg to methyl Hg and similar compounds.
The toxicity, persistence, and concentration of methyl Hg
in food chains caused concern for any discharge of Hg into
the environment.
Extensive analyses in North America indicated
high Hg levels in fish and sediments associated with cer-
tain Hg cell chlor-alkali facilities. All facilities in
North America rapidly took steps to reduce total Hg dis-
charges to less than 0.23 kg/day (6.5 Ib/day) to the re-
ceiving waters at each installation. In most instances,
this involved stockpiling of Hg-containing materials such
as process sludges (Figure 1).
An objective of these studies was to develop a
system to reduce the Hg content of brine process sludge and
other Hg-containing solids and liquids to a level suffic-
iently low that they may be disposed of without significant
hazard to the environment. A further objective was the re-
covery of Hg without significant loss into the atmosphere.
The Bellingham Chlor-Alkali plant, Figure 2,
went into production in 1965 with a capacity of 122 m tons
(135 s tons) per day of chlorine. Brine sludge averaging
1.4 m tons (1.5 s tons) per day resulted from the precipi-
tation of calcium and magnesium compounds from the incom-
ing solar salt, and erosion of graphite from cell anodes.
The sludge had been stockpiled in an impoundment basin
pending the development of a Hg recovery system. Hg con-
taminated water is generated at the rate of 110 - 190 1/min
(30 - 50 gpm).
The major sources of the brine sludge are: (1)
the brine clarifier, (2) the brine filters, and (3) the
salt saturator residue. Other Hg-containing solids in-
clude: caustic filter backwash, cell residue, and caustic
storage tank residues.
-------
Figure 1. The Bellingham Chlor-Alkali Plant
and brine sludge pond
-------
Figure 2. The Bellingham Chlor-Alkali Plant with sludge
pond and Hg recovery structure
-------
The major sources of Hg-bearing waste water are:
(1) floor washings from the cell room, (2) purge streams
from the cell end-box wash water recycle systems,
(3) purge streams from the brine system, (4) drainage from
the caustic filtration area, and (5) water from tank
cleaning.
An extensive literature search was conducted
before and during the project; the results are included
in the reference section.
This is the final report on the project and the
work performed from June 1, 1971 to April 30, 197^.
-------
SECTION IV
PROCESS DESIGN
SLUDGE TREATMENT FOR Hg REMOVAL
Lab or at or y Me thod s
The chemical oxidation tests performed in the
lab were on the scale of 250 - 1000 ml of brine sludge
treated in beakers (Appendix A). The sludge roasting
tests were carried out in lab furnaces with volumes of 3
liters and 100 liters. The temperature, heating time, and
air purge were controlled as described below. The labor-
atory phase of the project lasted 9-10 months before the
single process to be used was selected.
Chemical and Electrolytic Oxidation
An extensive investigation was devoted to de-
veloping a chemical means of removing Hg from the brine
sludge (Appendix A). The alternatives tried involved ad-
dition of sodium hypochlorite, chlorine, or electrolysis
of brine to generate small bubbles of chlorine gas. These
treatments are reportedly used to treat Hg ores as well as
chlor-alkali cell wastes. Removal rates of over 99% are
claimed for concentrated ores and residual Hg levels of less
than 0.1 ppm for chlor-alkali sludge (5, 6, 9).
In this study, the treatments not only dissolved
Hg, but significant quantities of other components of the
sludge as well, so that Hg separation was not effective.
The maximum Hg removal was less than 88$ with a minimum
Hg in sludge after treatment of 47 ppm (Appendix A).
Due to problems of (1) dissolving components
other than Hg, (2) multi-staging to achieve desired percent
recovery, and (3) difficulty of the filtration and wash
steps between the stages, the chemical alternatives to
sludge treatment were abandoned.
Roasting at High Temperatures
The roasting of Hg-bearing solids has been used
since ancient times to separate Hg from other material
(1). In preliminary tests in a small lab muffle furnace,
crucibles of brine sludge were heated to several tempera-
ture levels for various lengths of time to determine the
approximate temperature and time parameters (Table 1). The
-------
Table 1. KILN TREATMENT OF BRINE SLUDGE & GRAPHITE
(ppm Hg)
Brine
sludge
Cell
graphite
Time.
hr.
Start
1
8
16
24
Start
5
16
Temperature
°C 121 427 538
°F 250 800 1000
140 140
4.6
4.5
48
4.2
4200
20
140
3.7
3.1
2.5
:::
649
1200
140
0. 19
0.06
0.04
4200
6.2
8
-------
tests yielded residues ranging from 0.3 - 1.7 ppm Hg. These
initial results were 20 - 100 times lower than the lowest
residuals achieved by chemical treatment.
Following the preliminary tests, a series of trials
were conducted in a large kiln on samples ranging, in size
from 100 g to over 30 kg (Figure 3). The air rate through
the kiln was carefully controlled to remove the vaporized
Hg to keep from saturating the vapor phase with Hg. Re-
siduals as low as 0.02 ppm Hg were achieved (Table 2 and
Figure 4). Temperatures, in the range of 800°C - 900°C
(1450°F - 1750°F) were required to achieve Hg residuals
below 0.2 ppm.
Furnace Selection
Following these successful lab runs, kiln manu-
facturers were contacted to verify the data on a pilot
scale. Tests were conducted at Bartlett-Snow, Cleveland,
and BSP Division of Envirotech, Brisbane, California.
At Bartlett-Snow, a 15 cm (6 in) diameter rotary
calciner was operated at 800°C (1475°F) with a residence
time of 30 minutes. The minimum Hg level achieved in the
tests was 25 ppm Hg, which was significantly higher than
the batch kiln test at the same temperature (Table 3).
The tests were shifted to a multiple hearth furnace to
gain better control over residence time and eliminate
short-circuiting.
Two multiple hearth furnace manufacturers were
contacted; the BSP Division of Envirotech was selected to
test the dewatered brine sludge. Tests run in April and
June 1972 in a 33 cm diameter (13 In) batch kiln yielded
clinker Hg contents of 0.32 ppm.
From these data, a pilot run was scheduled in
July to test the procedure on a 76 cm (30 in) furnace at a
higher solids feed rate. At temperatures of 730°C - 760°C
(1350°F - 1400°F), the Hg level in the clinker was 3.2 ppm
(Table 4). This was not as low as desired; however, in a
second test at 870°C - 955°C (1600°F - 1750°F), residuals
of .12 - .14 ppm Hg were obtained (Table 5).
From these data, the furnace hearth loading was
found to be a maximum of 39 kg/m2 (8 lb/ft2) per hour so
that a wet solids feed rate of 224 kg/hr (600 Ib/hr)
would require a 7 m2 (75 ft2) furnace. This corresponds
to a standard 1.37 m (4.5 ft) i.d. 6-hearth unit with
7-9 m2 (84 ft2) hearth area.
-------
Figure 3. Lab kiln test assembly for roasting sludge
w///////////////,
MANUAL AIR
CONTROL VALVE
TEMPERATURE
CONTROLS FOR
ELECTRIC
HEATING
ELEMENTS IN
KILN TEMPERA-
TURE INDICATOR
COMPRESSED AIR
10
-------
Table 2. KILN ROASTING OF SLUDGE, NO CHEMICAL TREATMENT
Exp.
no.
31
32
M
33
38
42
51
tl
1 1
It
57
61
61
63
65
65
70
72
72
73
76
70
80
87
89
90
81
91
91
91
Sample, treatment
Brine sludge, rotated
it
tf
Brine sludge
tr
ir
ir
it
it
M
it
Metal anode sludge
Our brine sludge
Pond sludge
Brine sludge
it
ii
ti
ii
it
it
it
it
M
It
tt
II
II
Brine sludge from
filter tests
Residence Temperature Agitation, Hg content, ppm
time, hr. 'C T min. Start End
8
8
8
8
8
8
5
6
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
7
7
7
7
649
677
677
649
824-38
774-793
149-260
232-343
649-760
649-760
649-760
649-760
649-760
649-760
649-760
649-760
649-760
649-774
649-774
649-760
593-649
593-649
677-718
649-774
649-663
538-571
802-830
941-9G9
802-830
941-969
1200
1250
1250
1200
1525-100
1425-1460
300-500
450-650
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1200-1425
1200-1425
1200-1400
1100-1200
1100-1200
1250-1325
1200-1425
1200-1225
1000-1060
1475-1525
1725-1775
1475-1525
1725-1775
0
0
0
3
4
4
9
9
9
9
2
2
2
2
i
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
...
...
500
1340
880
1340
1340
...
158
250
...
246
822
1100
822
822
822
822
1735
1735
1735
1735
1735
2250
2250
4.4
0.4
1.3
1.7
0.18
0.12
880
50
0.5
4.8
0.53
0.02
0.95
0.69
0.97
1.7
0.37
0.47
0.75
0.07
5.21
5.6
2.0
1.7 )
5.6 )
14 )
0.08 }
0.03 )
0.07
0.02
Comments
6 ceramic balls (fine powder)
n ii
NaCl appeared to fuse
Still slightly wet
Still slightly wet
Top 1" of 9" depth *
Bottom of 9" depth *
pH 10
Excess of air
Sludge shipped to Envirotech
7-10-72, 5 barrels for 30"
kiln test.
* Volume decreased to 2/3 upon roasting.
-------
to
a
E
P.
a.
figure 4. Hg contents of untreated brine sludge after roasting
in lab and pilot kilns
10
° r
01 L
[TI • ~
:
H
J
O
CK
0 1
h
0
O
§
—
-
KEY
• Lab
1 1 ° ub
g 1 (5 Lab
0.1 1— 3 Lab
E • Lab
r A Lab
\y
%
•
kiln, agitated
kiln 1/2" layer
kiln 2" layer
kiln 4" layer
kiln 9" Layer
kiln, metal anode sludge
L Q Bartlett-Snow pilot rotary
.01
V BSP
A BSP
A BSP
r id
i •*• i
batch kiln, 13"
30" kiln, 200 Ibs/hr.
30" kiln, approx. 300 Ibs/hr.
Australia Ltd. 30" kiln tests
i i i 1 1 L_
9
V 1
A
A
O
AGO 600 800 1000 1200 1400
KILN TEMPERATURE, 'F
1600
1800
12
-------
Table 3. ROTARY CALCINER ROASTING OF BRINE SLUDGE
GO
Hg content of starting material, ppm dry basis
Maximum temperature at steady state, °C.
°F.
Retention time, min.
Water content of feed, %
Solids feed rate, gm/min.
Ib. /min.
3
Purge air rate, standard m /hr.
scfh
Screen analysis of solids collected
in first 2 hours, % +8 mesh
-8 +60 mesh
-60 mesh
Hg content of screened fractions, ppm
+8 mesh
-8 +60 mesh
-60 mesh
Hg content of steady state
5 minute sample, ppm
Hg removed, %
Run no. 1
1540
768
1415
30
30
57
0.125
7.4
80
40
31
29
14
92
284
50
96. 8
Run no. 2
(drying only)
1540
357
675
30
30
57
0.125
7.4
80
----
89
94.2
Run no. 3
89
802
1475
30
0
57
0.125
7.4
80
59
33
8
6.6
19
93
25
98. 4 Over;
72 This pass
-------
Table 4. MULTIPLE HEARTH FURNACE ROASTING OF BRINE SLUDGE
BSP 76 cm (30 in^ PILOT FURNACE
Test
no.
1
2
3
Sample
number
1
2
3
4
Hearth
sampled
3
4
5
6
3
4
5
6
3
4
5
6
Hg, Wet feed rate, Retention time,
ppm kg/hr Ib/hr min.
5.7, 5.4 91 200 30
11.4, 5.8
4. 8, 4. 7
3.2, 3.4
5.6 91 200 45
4.7
4.1
3.2
4.1 136 300 30
3.7
4.3
4.4
Test date: 7-24-72
Feed moisture content: 37.6%
Feed Hg content: 1735 ppm, dry basis
Furnace temperature: 760° C. (1400° F.)
Average gas consumption per pound of feed: 0.068 m (2.64 ft. )
-------
Table 5. MULTIPLE HEARTH FURNACE ROASTING OF BRINE SLUDGE
BSP 76 cm (30 in) PILOT FURNACE
Test
no.
1
2
3
4
Hearth
sampled
3
4
5
6
3
4
5
6
3
4
5
6
3
4
5
6
Retention
Hg, Wet feed rate, Temperature, time,
ppm kg/hr Ib/hr °C. ° F. min.
1.8 91 200 870 1600 30
1.2
0.31
0.14
1.5 91 200 955 1750 20
0.69
0,74
0,12
5.6 136 300 870 1600 20
5.1
2.5
1.3
182 400 870 1600 20
10.7
4.7
2.3
Test date: 8-10-72
Feed moisture content: about 47%
Feed Hg content: 1, 735 ppm, dry basis
-------
Since the tests at BSP were run, contacts were
made with investigators from an Australian chlor-alkali
plant who were also searching for a brine sludge treating
method for removing Hg. They also tested the multiple
hearth furnace at our suggestion after experiencing un-
favorable results from chemical treatment methods. The
data from their runs in the 76 cm (30 in) pilot kiln fur-
nace show results similar to ours (Table 6 and Figure 4).
Minimum values of 0.1 ppm Hg in clinker were achieved at
gas temperatures >800°C
The multiple hearth furnace is shown in Figure
5. The feed material is conveyed into the top and is
carried across the top hearth slowly by the rabble arm
plows, then falls to the next hearth. This continues from
hearth to hearth until the clinker falls out the bottom
of the furnace to be cooled and/or discarded. The heat
for the furnace is supplied by gas jets on 2 - 4 hearths
and the temperature is controlled by thermocouples and gas
flow control valves. Smooth furnace operation with minimum
attention is dependent on a constant feed of uniform mois-
ture sludge from the filtering step.
The roasting method should involve the least op-
erator attention of any of the methods considered. The
chemical methods studied required many more processing
steps with more equipment and more .critical control points.
Roasting with Acid Treatment
During the roasting tests a number of variations
were tried, including reducing the volume of the sludge so
that a smaller kiln could be used to treat the sludge. To
reduce the basic sludge, acids were tried successfully.
Surprisingly, when acid-treated sludges were roasted, even
lower final Hg levels were achieved than for untreated
sludge at the same temperature: 0.02 ppm Hg was achieved
-in the clinker below 760° C (1^00°F) (Table 7 and Figure
6). The mechanism is not known although the phenomenon
was observed in 31 separate tests.
A patent application has been submitted on this
process to the EPA Office of the General Council.
In pilot tests at BSP Division of Envirotech on
June 21, 1972, Hg residuals as low as 0.10 ppm were found
after 30 minutes at 730°C (1350°F) in their 33 cm (13 in)
batch kiln (Table 8).
16
-------
Table 6. MULTIPLE HEARTH FURNACE ROASTING OF BRINE
SLUDGE BSP 76 cm (30 in.) PILOT FURNACE, Id DATA
Test
no.
1
2a
2b
2c
3
4
5
Furnace temperature
Solids Gas
°C. °F. °C. °F.
699
610
538
610
599
599
599
1290
1130
1000
1130
1110
1110
1110
___
840
849
840
921
799
799
— _-
1544
1560
1544
1690
1470
1470
Hg,
Feed
640
640-1200
it
it
640
667
450
ppm
Clinker
3.8 to 0.5
1.0 to 0.3
8.6 to 0.3
2.2 to 0.6
4.0 to 0.1
5.7 to 0.3
0.1
Test date: 8-29-72
Conditions:
Feed moisture content: 23-38%
Feed rate, dry material discharged: 100-300 Ib/hr.
Estimated retention time: 10-30 minutes
Data from ICI Australia Limited Plant Pilot test in Australia.
17
-------
Figure 5. Cut away drawing of the internals of a multiple
hearth furnace showing the rake arms and hearth assembly
FEED
FURNACE SHELL OF
SHEET METAL LINED
WITH 6" - 9" OF FIRE
BRICK
DISCHARGE
STEP OL'
DXVC-. NB
18
-------
Table 7. KILN ROASTING OF CHEMICALLY TREATED BRINE SLUDGE
CO
Exp. Residence Temperature
no. Treatment time. hr. • C • F
35
53
61
63
63
65
65
70
70
70
70
72
72
72
72
73
73
73
73
73
73
74
74
74
74
74
74
74
76
76
79
7!)
79
73
79
80
*
**
Chlorine pretreated *
HCl treated
HCl treated in crucible
C12 treated
HCl treated
HCl treated
HCl treated
Acetic acid to pH 2
H.SO to pH 2
HCl fo pH 2
HCl to pH 2; NaOH to pH 10
H,SO.
Acetic acid
HCl
HCl then NaOH to pH 10
Ar:ntic acid pll 2-8
H2S04pH 2
HCl pH 2
HCl pH 0
HCl pH 3
HCl pH 0 (80- C.)
H2SO4
M
M
11
M
II
t|
HCl treated
H2S04 treated +*
HjSO4 treated
tl (I
It II
M It
II M
«2S04 **
Drum rotated 1st hour
Sent to Envirotech 5/24/72
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
677
649-746
649-760
649-760
649-760
649-760
649-760
649-774
649-774
649-774
649-774
649-774
649-774
649-774
649-774
640-760
649-760
649-760
649-760
649-760
649-760
663-732
663-732
603-732
663-732
663-732
663-732
663-732
593-649
593-649
593-649
593-649
593-649
593-649
593-649
677-719
12SO
1200-1375
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1200-1425
1200-1425
1200-1425
1200-1425
1200-1425
1200-1425
1200-1425
1200-1425
1200-1-1UO
1200-1400
1200-1400
1200-1400
1200-1400
1200-1400
1225-1350
1225-1350
1225-1350
1225-1350
1225-1350
1225-1350
1225-1350
1100-1200
1100-1200
1100-1200
1100-1200
1100-1200
1100-1200
1100-1200
1250-1325
Agitation,
min.
0
2
2
2
2
i
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Hg contentt ppm
Start End
113
230
3660
495
3660
2280
2280
246
246
246
246
58
855
563
764
862
206
634
284
669
120
586
379
418
503
364
364
312
585
585
120
503
364
379
418
...
244.
0.065
0.21
0.48
0.25
0.26
0.36
0.26
0.05
0.14
0.37
0.02
0.17
0.58
0.10
0. 08
0.02
0.03
0.16
0.15
0.06
0.06
0.02
0. 01
0.05
0.10
0. 002
0.22
0.64
4.84
0.46
0.10
-.
0.13
0.04
0.09
-------
U
Figure 6. Acid treatment of brine sludge before roasting in
'lab and pilot kiln
10
60
t 1.0
P*
8
H
CO
w
I
0.1
.01
1 1
-
™
-
—
.
-
-
-
—
"
1 1 1 1 1 1 1 -
KEY
Lab, HC1 O
Lab, H2S04 Q
Lab, Acetic A
Lab, Cl ^ "
2
BSP 13", •
o 2 k —_
0 o. :
D »J
8 ;
A ^h
W •
D ^
B ° ~
D
O
i
i i i t i i i
800
1000 1200 1400
KILN TEMPERATURE, °F
1600
1800
20
-------
Table 8. BSP TEST ON BATCH KILN
Test
no.
1
2
Sample
Date treatment
4-16-72 Untreated
t>
it
it
6-21-72 Acid
treated
M
ti
ii
it
ii
Temperature Retention
0 C. ° F. time, min.
760 1400 30
60
120
180
732 1350 15
30
45
60
75
90
Hg content,
Start
874
874
874
874
620
620
620
620
620
620
ppm
End
1.40
.41
.42
.32
.55
.10
.17
.45
.15
.11
21
-------
Acid pretreatment was not included in the final
process design because heating the untreated sludge an
additional 1650° (300F°) achieves the same residual Hg
levels at less capital and operating expense. However,
the acid pretreatment may be incorporated into the process
at a later date if it appears to be necessary- due to in-
creased sludge volume or higher than expected residual Hg
levels.
Solids Dewatering
The starting material for the sludge processing
system is sludge as it comes from the sludge pit.
Typically, this sludge is only 5 - 10$ total suspended
solids as it is pumped from the sludge pit; a 45 - 60/E
solids feed to the furnace is desirable for economic op-
eration. The dewatering methods tried were: (1) gravity
settling, (2) centrifuging, and (3) filtration.
Gravity settling was not satisfactory for this
sludge since the maximum concentration achieved was ap-
proximately 30% solids.
In lab tests at Barrett Centrifuge, the sludge
was dewatered to 12% solids in the first stage and the
liquid from the second stage contained 0.2 - 1% insoluble
solids (Table 9). However, there may be disadvantages to
centrifuging over filtration for this application.
Possible disadvantages are: (1) higher capital
cost for equipment for a given capacity, (2) difficulty of
obtaining corrosion-resistant material for wetted parts in
other than stainless steel, (3) two- to four-stage centri-
fuging is needed to achieve solids-liquid separation, (4)
centrifuges are generally made in only 1 or 2 sizes in each
style so that multiple units must be used to achieve re-
quired throughputs, and (5) the high speeds of centrifuges
with such abrasive and corrosive material as brine sludge
could lead to high maintenance. For these reasons, rotary
vacuum filtration was selected instead.
The high solids loading indicated that pressure
filtration was not feasible. Vacuum filtration was ef-
fective on a bench scale. Two standard rotary vacuum fil-
ters were tried: a Komline-Sanderson 0.9 m (3 ft) diameter
x 1.5 m (5 ft) unit, and an Eimco 0.9 m (3 ft) x 0.3 m
(1 ft) unit. As shown in Table 10 and Figure 7, drum rate
and solids feed concentration are critical for filter ca-
pacity and cake dischargeability. Since filters are avail-
able in standard sizes, a filter with an area greater than
22
-------
Table 9. TEST RESULTS FROM BARRETT CENTRIFUGE ON
DEWATERING OF BRINE SLUDGE BY CENTRIFUGING
Sample
No.
Test Solids Insoluble Flow rate thru unit,
description by volume, % solids, %
As received
40
25.6
MODEL 912 CENTRIFUGE
2
3
1 pass
sludge
supernatant
21
71.8
6.0
MODEL 125 CENTRIFUGE (912 Supernatant)
4
5
6
7
1 pass 0. 7
supernatant
3 pass 0. 24
supernatant
5 pass 0. 11
supernatant
sludge
1.06
0.4
0.2
21.4
1
1
1
TEST METHOD
sludge
feed fc
'
'
Model
912
1
supernatant
t
V
Model
125
1 sludge
supernatant
23
-------
Table 10. DATA FROM THE EIMCO PILOT ROTARY VACUUM FILTER
to
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
Drum rate,
rpm
0.22
0.32
0.41
0.41
0.22
0.47
0.30
0,22
0.41
0.30
0.22
0.22
0.22
Feed Cake dry
suspended solids, solids rate
weight % kg/hr Ibs/hr
18.7
17.6
18.3
17.0
15.5
15.2
15.7
6.3
5.7
4.9
21.0
15.5
18.4
34.2
41.1
44. 1
32.3
16.2
21.5
19.1
8.3
24.3
16.6
56.2
38.6
48.5
75.4
90.5
97.3
71.3
35.8
47.3
42.0
18.4
53.6
32.1
124.0
85.0
107.0
Filter size
for 3.5 ton/day
m sq. ft.
3.4
2. 9
2.6
3.6
7.2
5.4
6.1
13.9
4.7
8.0
2.0
3.0
2.4
37
31
28
39
77
58
66
150
51
86
22
32
26
Cake thickness,
mm in.
5.
3.
3.
3.
2.4
1.6
2.4
1.6
1.6
1.6
8.
6.
4.
0.19
0.13
0.13
0.13
0. 09
0.06
0.09
0.06
0.06
0.06
0.06
0.25
0.16
-------
Figure 7. Eimco rotary vacuum pilot filtration of brine sludge
150
-------
7 m2 (75 ft2 ) was selected. The filters 0.9 m (3 ft) x
2.4 m (8 ft) and 1.8 m (6 ft) diameter x 1.2 m (4 ft) are
approximately 7 m2 (75 ft2 ) in area; the larger diameter
is preferred since there is greater control of drying time
and cake formation time. A 1.8 m (6 ft) x 1.8 m (6 ft)
filter would be a good investment since the small addi-
tional capital cost would provide 50$ more filter capacity
in nearly the same space. The filter area would be approx-
imately 10.4 m2 (112 ft2 ). At a filtration rate of 10 -
18 dry kg/m2/hr (2.3 - 4 dry lb/ft2 /hr) with a 33% sub-
mergence, the 1.8m (6 ft) x 1.8m (6 ft) rotary vacuum
filter will discharge 2.8 - 4.9 m tons (6,200 - 10,800 Ib)
of solids per 24-hour day. The extra capacity permits the
filter rpm to be slowed to build a thicker, drier cake if
cake discharge becomes a problem.
Equipment Sizing
The system was sized on the basis of 3-2 m tons
(3.5 s tons) per day to handle the expected solids from
the chlor-alkali plant at a chlorine production of l8l m
tons/day (200 s tons/day). This corresponds to a
sludge production of 1.37 m tons (1.5 s tons) per day from
the plant plus 1.83 m tons (2.0 s tons) per day from stock-
piled sludge and other Hg-containing solids.
The major pieces of equipment include a 3-7 m
(12 ft) diameter x 1.8 m (6 ft) high thickener, a rotary
vacuum filter and a 1.37 m (4.5 ft) l.d. 6-hearth multiple
hearth furnace (Figure 8). All decanted and filtrate brine
is recycled to the settling pond so that the small amounts
of solids remaining in these streams will not load up the
water handling system. In addition, if shower water is
needed to clean the filter cloth or sluice out sludge build-
ups around the filter, brine will be used and returned to
the sludge pit. No fresh water will be used for wash-down
to maintain the water balance.
WATER TREATMENT FOR Hg REMOVAL
Laboratory Methods
The test methods used to find the optimum water
treatment process involved many standard laboratory pro-
cedures. Solid particles such as the ion exchange resins,
activated carbons, and metal particles were packed in a
glass column of 1.9 cm (3/4 in) i.d. with a packing depth of
approximately 30 em (12 in). A constant liquid flow was
maintained by a head of liquid 7.6 cm - 15 cm (3 - 6 in)
above the top of the packing. Chemical tests to reduce Hg
26
-------
Figure 8. Proposed brine sludge handling system
NJ
FLUE CAS TO
REFRIGERATION
SYSTEM
HRINE FILTRATE TO
CURiriEX
I ^ sonns
X"-7*"-^v
<&&£s
UKINEMTED SOLIDS
TO IANDMLL
taZMRAlED
SOLIDS BIN
-------
ions with sodium borohydride, to precipitate Hg sulfide with
sodium sulfide, or to reduce Hg ions with powdered metals
such as zinc or aluminum were performed in 500 ml beakers
with magnetic mixer agitation.
The untreated Hg-contaminated water was first
added to the beaker. The pH was adjusted with hydrochloric
acid or sodium hydroxide to the desired point. The ap-
propriate chemicals were added and mixed for varying periods.
The liquid was then filtered through a 10 cm (4 in) or 15
cm (6 in) Buchner funnel precoated with about 6 mm (0.24
in) of diatomaceous earth. The untreated and treated so-
lutions were then analyzed for Hg by flameless AA.
To determine the solubility of mercuric sulfide
in solutions of varying pH and excess sulfide, solutions of
mercuric chloride and sodium sulfide were combined. The
resulting precipitate was collected, washed and weighed
into equal amounts. These samples of HgS were placed in
sealed containers of water at various pH and excess sulfide,
agitated and allowed to come to equilibrium. The super-
natant was then analyzed for Hg.
Alternative Methods Investigated
During the laboratory phase of the project, sev-
eral of the methods proposed for Hg removal from water were
tried including:
1. Ajinomoto ion exchange resin
2. Billingsfors - Langed ion exchange resin
3. Nuchar 722 activated carbon
4. Pittsburgh HGR activated carbon
5. Calgon Filtersorb 400 activated carbon
6. Zinc particles
7. Sodium borohydride
8. Stannous chloride
In general, these methods were not able to achieve
effluent Hg levels below 0.10 ppm for starting solutions of
2-20 ppm, or their capacity was limited so that their
effective life was greatly shortened by concentrated Hg
feeds. The ion exchange resins and activated carbons ap-
28
-------
pear to be most effective as polishing steps after the first
stage of treatment has removed the bulk of the Hg. They
are able to treat solutions in the range of 40 - 100 ppb
down to 1 - 5 ppb consistently. Appendix C details the
results of these tests .
Using another metal to reduce and adsorb Hg ions
while dissolving the second metal tends to trade one ef-
fluent problem for another, e.g. the zinc reduction method.
Of the methods tested, sodium borohydride,
appears to be the best alternative to the sulfide precipi-
tation for a primary Hg removal process. The equipment
necessary is very similar to the sulfide addition process
and with careful engineering a system could be built to
use NaBHn - sulfide interchangeably with only minor modifi-
cations. The only drawbacks found for the NaBHjj method for
this application appeared to be the slightly lower effic-
iencies found in the lab tests and the higher cost of NaBHjj.
Ventron Corporation holds a patent on the use of
NaBHjj for heavy metal removal (2).
In the last decade, the literature includes sev-
eral less common ways to. remove Hg from water or brine.
Among the methods proposed are: (1) solvent extraction with
high molecular-weight amines (3, 4); (2) electrolytic means
by passing the solution through a type of diaphragm cell
(5, 6); (3) hydroxide flocculation and filtration (7); and
(4) adsorption of Hg compounds by CaCp slag and flocculation
with PeSOh (8). Of these methods, the latter two seem to
have some promise although we did not investigate these
techniques.
Sulfide Precipitation
A number of publications have described the
use of sulfide ions for precipitation of Hg from water so-
lutions (9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, & 20).
It is generally agreed that a removal rat.e of 99-9$ can be
achieved with sulfides and this has been confirmed in our
test work on laboratory and pilot scales (Table 11).
The drawbacks to this method include: (1) the
formation of soluble sulfide complexes at high levels of
excess sulfide, (2) the difficulty of monitoring excess sul-
fide levels, and (3) the problem of sulfide residue in the
waste water discharge.
29
-------
Table 11. LAB TEST DATA ON SULFIDE PRECIPITATION FOR Hg REMOVAL FROM WATER
CO
o
Sample number
pH Range 7-14
1
2
pH Range 5-7
4
5
6
7
pH Range 1-5
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
PH
12.0
10.0
5.5
5.5
5.5
4.5
4.5
4.5
3.5
3.5
1.5
4.0
4.0
4.0
4.0
1.0
3.0
4.0
4.0
4.5
4.5
Reaction Filtration
time, Filter
min pad Precoat
60
60
1/2
10
60
60
60
60
60
60
120
30
60
60
50
60
60
60
20
30
30
Paper
Paper
Paper
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Glass-Fiber
Paper
Paper
Paper
D.E.
D.E.
D.E.
No
No
No
No
No
No
No
No
No
No
No
No
No
D.E.
D.E.
D.E.
Body
feed
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
D.E.
D.E.
D.E.
Acid
for pH
adjust.
None
None
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
HC1
H2S04
HC1
Initial
Hg content,
ppm
4.2
10.0
7.0
16.0
16.0
16.0
52.0
52.0
60.0
60.0
60.0
14.0
36.0
36.0
36.0
36.0
36.0
10.0
10.7
27.0
17.5
17.5
Final Hg ,
ppm
1.8
.36
0.11
0.63
.04
.03
5,6
.44
.03
.02
.01
.18
.002
.015
.006
.002
.012
.008
.12
.31
.15
.1
Hg removal
57.0
96.4
98.4
99.6
99.9
99.8
89.2
99.2
99.95
99.97
99.98
98,7
99.994
99.96
99.99
99.994
99.96
99.92
98.9
98.8
99.1
99.4
-------
From our work we have found that sulfide excess
is less critical so long as the HgS precipitate is filtered
out of the solution as soon as it is formed, as shown in
Table 12.
Table 12.
HG LEVELS IN WATER AFTER EXPOSURE TO
SULFIDE ION FOR 30 SECONDS TO 10 MINUTES
Mercury content Exposure time to Hg removal,
of solution, ppfa sulfide ion, min %
Starting
solution
51,500
After precipi-
tation & 9.7
filtration 15.0
15.0
14.0
0
0.5
2.0
5.0
10.0
99-98
99.97
99.97
9-9.97
Probably the sulfide comlex is formed more slowly than
the mercuric sulfide particle.
Hg
Hg
+2
2S
~2
> 2HgS + 2Hg
TT 0 _-2 , . slower TT -2
HgS + S (excess) > HgS0
•^ ti
In laboratory sulfide precipitation tests, 99-98$
of the Hg was reacted after only 30 seconds of contact.
For all practicle purposes, the reaction can be said to
go to completion in 30 - 60 seconds. As Figure 9 indi-
cates, even large excesses of sulfides did not reduce the
recovery of Hg significantly. Therefore, the accurate
monitoring of the exces's sulfide level is not necessary.
The third problem stated above, that of sulfides
discharged with the treated water, is easily solved in
a chlorine plant. There is normally a small amount of
residual chlorine in the main cooling water stream which
effectively oxidizes the sulfide ion. If residual
chlorine is not available in the treated discharge, then
31
-------
a small amount of sodium or calcium hypochlorite can.
be added to the filtered waste water to eliminate the re-
maining sulfide ion.
The most critical parameter in controlling the sul-
fide precipitation of Hg was found to be pH. Lab tests were
performed using mercuric sulfide which was precipitated from
mercuric chloride and sodium sulfide solutions (Figure 9).
A standard solution of sodium sulfide was prepared and
placed in full bottles of water containing various concen-
trations of sodium sulfide. The mixtures were agitated sev-
eral times and then allowed to settle 24 - 48 hours, enough
time for the Hg in the solid phase to approach equilib-
rium with the Hg in solution. The results appear in Figure
9 as initial excess sulfide level versus final dissolved Hg
level as a function of starting pH.
Since this experiment allowed a lengthy contact
time, it is likely, that the soluble mercuric sulfide ions,
HgS| , are formed at the higher initial sulfide levels so
that high dissolved Hg levels are due to both pH and excess
sulfide.
From the data presented, it appears that within
the pH range of 3 - 8 the level of excess sulfides does not
affect the amount of Hg sulfide which is redissolved. Less
than 0.1$ of the mercuric sulfide added to the sulfide con-
taining water redissolved at these pH conditions in 48 hours.
However, as high as 75% of the mercuric sulfide redissolved
above a pH of 10.
Possible methods of adding a controlled amount of
sulfide to the waste water stream are: (1) meter the desired
concentration from a concentrated NaHS solution, (2) add
the solid NaHS or Na2$ particles from a dry feeder, or (3)
pass the waste water through a bed of less toxic metal sul-
fide which has the proper solubility to release sulfides at
the 1-5 ppm level.
The first method has been used exclusively in our
test work because it is the easiest in batch tests in the
lab or pilot plant. However, on a continuous plant.scale
the second method would eliminate the need for an operator
to mix an exact solution concentration. Instead, he
could simply add a bag or two of the solid sulfide to the
dry feeder hopper when necessary. A possible problem with
this method is the rate of dissolution of the sulfide par-
ticles in the cold waste water. If sufficient agitation
and residence time were not provided, the sulfide would not
dissolve completely.
32
-------
Figure 9. Solubility of HgS in excess S
1000
CO
en
B.
(X
3
100
<
pH 6
pH 6 pH 7 pF8
pH ft pH 7
pH 8 pH 4,5 pH 5 pH pH
3 6
A
pH 6
Q-—
pK 13
0.1
12
KEY
0 G.P. data - R.P.
£D Chlorine Institute
data
A G.P. data - D.R.
At 0 ppm S
.001
.01
.1 1.0 10
" IN SOLUTION AT EQUILIBRIUM, ppm Hg
100
1000
-------
The third method proposed has been discussed but
not tested. In theory, since a certain excess sulfide must
be maintained, it should be possible to control this excess
chemically. By finding a sulfide compound of the correct
solubility and forming a packed bed of this material,, the
waste water passing through the bed would pick up the metal
and sulfide ions to saturation. As the sulfide ions are
consumed in the HgS precipitate, more sulfide would dis-
solve according to the solubility product of the specific
metal sulfide used. Possible problems with this method are:
(1) toxicity of the metal ion selected, (2) coating or de-
activation of the ted by contaminants in the waste water,
or (3) suppression of the solubility of the metal sulfide
if the metal ion concentration is high in the waste water.
The Selected Water Treatment Process
Of the alternatives considered, it appeared that the
sulfide precipitation method was the best choice for sev-
eral reasons. First, only five processing steps are needed
to achieve 99% Hg removal. These steps are: pH adjustment
to 5 - 7, addition of the sulfide ion, addition of the fil-
ter aid, filtration, and solids feed to the furnace. Most
other methods require 1-3 more steps (figures 10 - 13).
Secondly, the pH adjustment which is required of
the waste water changes the pH slightly to the acidic side
which will help to neutralize the basic cooling water ef-
fluent. Thirdly, the precipitated Hg is in a very concen-
trated form in the filter cake, 15 - 30$ Hg, and can be fed
into the furnace without adding significantly to the solids
load. Finally, the sodium sulfide is inexpensive and no
additional contaminating ions will be in the effluent after
the excess sulfides are eliminated by available chlorine in
the remaining effluent.
The process proposed is shown in figures 13 & 14.
The process begins with the collection of all the Hg con-
taminated waste water in a 170,000 1 (45,000 gal) agitated
tank. The pH is adjusted in this tank automatically and
continuously by the addition of spent sulfuric acid from
pH 11 down to pH 5 - 6. The waste water flows from this
vessel into a container in which the concentrated sodium
sulfide solution (or sodium hydrosulfide solution) is
added by a metering pump. The treated waste water is then
pumped through a system which adds a measured amount of
filter aid, diatomaceous earth, to the stream. It then
flows through a pressure filter which separates the Hg-
bearing solids from the water. The solids are discharged
in a slurry and pumped into the sludge dewatering filter.
34
-------
Figure JO. Schematic of the reduction method of Hg removal from water
REDUCING AGENT
LIQUID OR POWDER
w
en
WATER
'Hg VAPOR TO COND.
r*i
BED OF
I1EDUC-
ING PAR-j ^FILTER
TICLES
•»< NACE
ION EX. I
OR
ACT. C.
EFFLUENT
SOLIDS TO LANDFILL
-------
Figure 11, Schematic of the ion exchange process for Hg removal from water
CO
RECYCLE TO BRINE
PRECIPITATION CHEM.
SOLIDS TO
LANDFILL
-------
Figure 12. Schematic of activated carbon Hg removal from waste water
REPLACEMENT CARBON
GO
SECONDARY
CARBON BED
SOLIDS
TO
LANDFILL
-------
Figure 13. Schematic of the sulfide precipitation process for waste water
co
oo
FILTER
ION EXCHANGE
OR ACTIVATED
CARBON
EFFLUENT
Hg VAPOR TO
CONDENSER
SOLIDS TO
LANDFILL
-------
Figure 14. Proposed sulfide precipitation system
lit COKTAM- ^^mm^
INATED -X"
WATER 1
DM
I
K.jl
MISTING 40,000 |>1
Ftl TANK
Hx
TILTI*
1ACXUASR
RECEIVER
KF.S1S'
.\.~T1VA'
TED
IMC
BED
-------
The filtered water might then either go to the outfall or
pass through an optional activated carbon or resin bed
for a final polishing step before discharge.
Liquid/Solid Separation Techniques
This liquid-solids separation consists of removing
22 kg (60 Ib) per day of fine particles from a 150,000 1
(40,000 gal) water stream. Thus, the solids make up
0.02J6 by weight of the liquid stream or<0.01jS by volume.
A pressure filter is normally used, since centrifuges handle
high solids loading, rotary vacuum filters handle medium
solids loading, and pressure filters normally handle the
solids loading below 1035 solids by volume.
A number of alternatives were evaluated:
1. Dry or slurry solids discharge?
2. If dry discharge, vibrating or centrifugal
action cake removal?
3. Should the filtering elements be vertical
cylinders, vertical plates or horizontal
plates?
4. Should a precoat be added and what type?
5. Is continuous filter aid addition necessary?
6. Should the entire chamber open or one large
bottom outlet be provided?
In selecting the proper filter, the method of
cake removal is the most important feature to be examined.
Almost any filter will build up a suitable cake; the prob-
lem is frequently to remove the cake completely and restore
the original clear filtering area (21).
It is desirable for further handling of the Hg-
containing cake that the driest cake possible be discharged
from the filter. However, because of the small solids load-
ing, 22 kg (60 Ib) per day, the difficulty of cleaning the
filter properly without washing, and the extra cost in-
volved for the more complex equipment, $5,000 - $10,000 ad-
ditional for a 9 m2 (100 ft2 ) filter, we believe the
sluicing and settling technique is best for this applica-
tion.
40
-------
It is more difficult to sluice clean a horizontal
filter. Also, filtration occurs only on one side of the
horizontal leaf whereas both sides of a vertical are avail-
able for filtration, producing a smaller filter for the
same total area. For these reasons a vertical cylinder
filter was selected.
For the efficient filtration of fine Hg sulfide
particles, a precoat is advisable. Our pilot runs on a
batch pressure filter with horizontal leaves showed that
diatomaceous earth, cellulose fibers and activated carbon
are all effective at removing the Hg sulfide, although the
diatomaceous earth was slightly more effective (Figure 15).
Since the Hg sulfide particles are so fine, they tend to
form an impervious layer on the filtration media rather
rapidly. However, the addition of small amounts of dia-
tomaceous earth as a filter aid continuously to filter feed
water extended the cycle time by a factor of 10 - 20. The
diatomaceous earth was added at the rate of 0.7 gpl
(Figure 16).
The diatomaceous earth filter aid and Hg sulfide
will continue to build up on the filter elements until the
cake space between elements is full. It is preferable to
backwash the filter before the cake volume is completely
full for a more complete cleaning action.
-------
Figure )5. Data from pilot tests with Jl sq. ft. Niagara filter
CUJ
ffi
X)
a.
a.
H
W
O
u
ffi
H
D
J
EK
fc
W
10,000
J.OOO
JOO
10
KEY
Diatomite pll 5-7 0
Diatomite pH 1-3 D
Diatomite pH n-13 X
Pre-co-floc O
Nuchar 722 A
8
X
» 0
99%
O
O
9
a
a
t
99.9%
]0
12
14
J6
18
Hg LEVEL IN UNTREATED WASTE WATER, ppm Hg
42
-------
Figure 16, Pilot filter testa to determine cycle length with and without precoast and body feed
GO
No body feed or flow control
- coarse paper, no precoat
- fine paper, precoat of D.E«
- fine paper', precoat of
Nuchar
Body feed and flow control
- fine paper, precoat of D.E.
2000
4000
TOTAL FLOW THROUGH FILTER, gal
6000
-------
SECTION V
CONSTRUCTION
PROCEDURE
Following the selection of sulfide precipitation
for water treatment and sludge roasting for solids
treatment, the construction phase of the project began in
October, 1972. Construction consisted of several
phases:
1. Selection and ordering of long delivery
time items.
2. Design and construction of the supporting
structure.
3. Installation of major pieces of equipment.
4. Process piping.
5. Supplying of necessary utilities.
6. System start-up and modifications.
The equipment ordering began in October, 1972;
construction began in March, 1973. Start-up of the water
treatment system began November, 1973> and start-up of the
sludge system began in January, 197^.
Selection of Major Equipment
The longest delivery time item appeared to be the
multiple hearth furnace with a 32-week delivery. Three
suppliers were investigated: MSI Industries, Envirotech,
and Nichols Engineering. Envirotech was selected
on the basis of equipment quality and personnel experience
in Hg ore roasting.
The rotary vacuum filter was the other major piece
of equipment in the sludge system and the alternative manu-
facturers of plastic or rubber-lined rotary vacuum filters
for corrosive liquids were Ametek and Eimco. The Eimco unit
was selected because a one-month old unit of all plastic
construction was available at reduced cost with immediate
delivery.
44
-------
The only long delivery time item for the water
treatment system was the pressure filter. There are a
large number of domestic and foreign manufacturers of pres-
sure filters in various designs. The manufacturers investi-
gated were:
1. Buffalo Filters
2. De Laval
3. Durco - Enzinger
4. Niagara Filters
5. R. P. Adams
6. U. S. Filters
7. Votator - Schenck
The R. P. Adams filter was selected based on:
(1) operator and maintenance experience with this model,
(2) standardization of parts with existing filters, (3) ease
of cake removal with few moving parts, and (U) price per
square meter of filter area.
Const'ru'c'tion Mode 1
In order to assist in arranging the equipment for
this project, a student engineer was assigned the task of
constructing a model of the entire system (Figure 17). Due
to the solids handling problems, the elevations and equip-
ment and piping layouts were critical.
The model was also useful during the operator
training phase of the project. Before the construction was
complete the operators could see the location of the pipes
and valves during the training classes.
Modifications During Construction and Start-up
Originally, the design called for an anode crusher
and conveyor to process the spent cell anodes. However,
during the project, the decision was made to convert over
to metal anodes so the anode crusher was eliminated.
45
-------
Figure 17. The model of the Hg Recovery System
46
-------
Between the plant sludge collection system and the
rotary vacuum filter it was planned to have, hatch settlers.
However., since the other equipment was continuous, the
batch settlers were replaced with a continuous thickener.
In practice the thickener is not needed since the sludge is
fed in batches from it anyway.
47
-------
SECTXON VI
OPERATION AND EVALUATION
START-UP OP WATER SYSTEM
Objective
Following the construction stage, the start-up
crew began checking out the individual parts of the water
treatment process and training the operators. This phase
began in mid December, 1973 for the water treatment part of
the project. The goal was to reduce the chlorine plant
effluent to less than 45 g (0.1 Ib) Hg per day by January
1, 1974* in compliance with discharge permit T-3456, and
have the bugs worked out of the system.
Start-up
The start-up phase of the water treatment system
lasted approximately one month, from mid December, 1973 to
mid January, 1974 before control was turned over to the op-
erators. The major changes to the original design were:
(1) rerouting of the Hg-containing water through the pond,
(2) changes in the acid addition system, and (3) a change
in the sulfide storage and addition mechanisms.
Untreated water flow from the cell room was in-
termittent, from-a collection sump. This made pH adjustment
difficult. Therefore, the existing pond was used as a surge
tank to give a constant flow to the pH adjustment system.
The acid addition system was also modified to provide two-
stage dilution for pH adjustment of the waste water with
concentrated spent sulfuric acid. At present, the system
controls in the range of 6 - 8 pH.
Originally, the sodium sulfide was to be mixed
in a storage tank and then transferred to a small metering
tank to be metered into the process through a pump. This
was altered so that the sulfide flows directly from the
large storage tank to the mix tank through a rotameter.
Evaluation
At present, the water treatment system is oper-
ating as expected and requires approximately 30 - 45 min-
utes of operator time each shift. Following the initial
shakedown period, the operators seem pleased with the
48
-------
system. The only operator attention needed normally is to
add diatomaceous earth once per shift and backwash once
every 1-2 days.
Prom a design viewpoint, the system is oper-
ating as desired. The 50 ppb average effluent from the
system at an average flow; of 150 1/min (.40 .gpm) accounts
for only 20% of the maximum allowable 45 g (0.1 Ib) per day
Hg discharged. The Hg in the filter cake is fed back to
the sludge system as a slurry after each backwash for dis-
posal by incineration. As shown in Table 13, the average
Hg removal has been 97% the first 3 months of operation
and the effluent Hg content has averaged 49 ppb.
START-UP OP SLUDGE SYSTEM
Objective
As the construction neared completion, the start-
up crew began checking out each piece of equipment by moving
the sludge through each stage of the process. As expected,
the main problems encountered in this start-up were solids
handling. The start-up began in early March, 1974 and
the operators took over two months later.
Problems Encountered In Start-up
The majority of changes in the sludge system
during start-up were in the conveying system between the
rotary vacuum filter and the furnace. Originally a Moyno
pump was installed, but when the cake was dry enough to
discharge well from the filter, 65$ total solids, the cake
was too dry for the Moyno pump to handle. Next, a screw
conveyor was tried but it was still necessary to add a
little water to the filter cake to keep it from sticking
to the screw. Finally, a small belt conveyor was installed
which appears to be working well. At present, a 5-vane star
valve is being tested as an air seal where the sludge enters
the top of the furnace. The sludge did not plug this star
valve in the one-hour tests run to date.
Other changes have been made during the operation
to improve the system (Figure 18). It was found that a
large number of sticks were being pumped out of the clari-
fier and plugging the line to the thickener. Therefore,
the strainer was moved from just before the thickener to
just after the clarifier. A flow indicator was also in-
stalled in the line to the thickener so that the operators
could readily tell when this flow stops for any reason.
49
-------
Table 13. PARTIAL LIST OF DATA FROM START-UP OF FULL
SCALE SULFIDE PRECIPITATION SYSTEM
Date
12-9-73
12-9-73
1-2-74
1-8-74
1-9-74
1-15-74
1-23-74
3-12-74
3-15-74
3-20-74
3-26-74
Average
Hg content, ppb
Feed Filtrate (
820
740
2000
1400
1400
140
800
5000
1300
6000
5800
*
Minimum: 300
Max.
6000
16
40
48
125
50
18
42
68
96
50
51
49 ppb
10
125
Excess Flow rate
% removal pH Na2S, ppm 1/min. gpm
98.
94.
97.
91.
96.
87.
94.
98.
92.
99.
99.
96.
87
99.
0
5
6
0
4
1
7
6
6
2
1
8%
2
6.5
6.2
5.8
8.0
7.6
5.9
6.0
6.9
6.8
7.2
6.0
5.1
8.2
2
2
2
0
1
-
2
2
-
-
0
3
330
310
330
310
310
388
290
310
310
310
330
290
388
85
80
85
80
80
100
75
80
80
80
85
75
100
* The average values for 30 sets of data
50
-------
Figure l€. Installed brine sludge handling system
BRINE
FILTER
BACKWASH
MISC.
SLUDGE
PUMP
SLUDGE THICKENER
BELT CONVEYOR
DECANTED
BRINE
FILTER
u
1/1 ;
, -j
110"
RECVLE
WATER
RINE SHOWER
PUMP
MULTIPLE
HEARTH
FURNACE
1NE CIARIFIER
Hg
CONDENSERS
FLUE
Hg CAS TO
REFRIG-
ERATION
SYSTEM'
BRINE FILTRATE TO
CLARIFIER
DUST
INCINERATED SOLIDS
TO LANDFILL
INCINERATED
SOLIDS BIN
-------
Due to the small size and plugging potential of
the control valve between the thickener and the rotary
vacuum filter, the small continuous valve was replaced with
a larger intermittent valve. Rather than holding the filter
at a constant level, it is allowed to cycle over a 7.6 -
10 cm (3 - 4 in.) range.
The rotary vacuum filter was able to pick up a
thicker cake 6 - 10 mm (1/4 - 3/8 in.) than was found in our
pilot tests. As a result, the drive sprocket on the rotary
vacuum filter was reduced in size so as not to overload the
furnace. The Eimco rotary vacuum filter minimum rotation
speed was reduced from one revolution in 14 minutes to one
revolution in 26 minutes. The filter apparently has ample
capacitv for future needs.
Evaluation
At present, from a design standpoint, the system
is working better than expected. It runs consistently at
a feed rate of 6.4 m tons (7 s tons) of sludge-per 24 hours
and will remove 99. 8£ of the Hg from the sludge at 730°C -
760°C (1350°F - 1400°P). The system operates with a feed
content of 345 ppm Hg and a discharge (clinker) Hg content
of 0.5 - 0.8 ppm (Table 14). This feed rate is twice the
design rate. As a result of the high throughput, lower
sludge output from plant than expected, and no return of
sludge from the pond to date, it is only necessary to
operate the furnace 1 or 2 shifts every 3 days. At other
times, the temperature is lowered to 370°C - 480°C (700°P -
900°F) to reduce refractory stress. As lower quality salt
is processed, producing more sludge, and as sludge inventory
is reprocessed, the filter and furnace will operate for
longer periods.
Prom furnace tests, it has been found that
dredged material from bark sludge beds can only be fed at
about one half the rate of brine sludge through the furnace
because of its different handling characteristics. There-
fore, the furnace capacity drops to 3.5 m tons (3.8 s tons)
per day for dredged cellulosic material.
Prom an operator's standpoint, the process is
working well at this stage. The biggest problems seem to
be plugging of the strainer between the clarifier and
thickener, and conveying problems between filter and fur-
nace. The wood problem in the strainer may be eased by
52
-------
Table 14. DATA FROM START-UP OF THE FULL SCALE
BRINE SLUDGE TREATMENT SYSTEM
Sludge
Source
Brine Sludge
it it
it it
it ii
ii it
Bay dredging
Addition
rate
kg/hr Ib/hr
226
255
264
205
309
137
540
560
580
450
680
300
Temperature
•F °C
1400
1250
1350
1350
1386
1350
760
677
732
732
752
732
Feed
345
255
290
438
370
128
Hg
ppm.
Clinker
0.5-0.8
1.6-3.1
1.7-2.6
2-7.2
1.6
0.95-1.7
% removal
Range Avg.
99.8
98.7-99.2
99.1-99.4
98.3-99.5
99.6
86.7-92.1
99.8
98.9
53
-------
more careful wood removal by the screens in the brine flow
ahead of the clarifier. The sludge conveying problem into
the furnace is being solved by the start-up and construc-
tion crews as problems appear.
The Hg recovery from the air leaving the furnace
still requires modification to solve the dust plugging
problem in the condensers. At present the condensers must
be cleaned every 2-7 days.
The air leaving the condensers has been analyzed
to contain approximately 0.5 lb Hg per day. The stream is
routed to an existing chilled water heat exchanger and a
Brink demister for Hg recovery. This recovery system re-
covers over 90/5 of the mercury.
54
-------
SECTION VII
DISCUSSION
DESCRIPTION AND ANALYSIS OF WORK PERFORMED
Literature Search and Company Contacts
At the beginning of the project, an extensive
literature survey was conducted utilizing the Chemical Ab-
stracts, the Dow Chemical Company Keyword Index onEnvi-
ronmental Aspects of Mercury Usage, and others (22, 23, 24,
25).Pertinent articles from various periodicals and pa-
tents were collected for study. Also, 22 organizations in
or associated with the chlor-alkali industry, were contacted
by phone, mail, or in person to gather information on the
methods used or contemplated for Hg removal from solids and
liquids. These contacts are listed in Table 15.
Laboratory Tests, Process Design & Equipment Selection
As discussed in Section IV, a series of labora-
tory and pilot tests were conducted on brine sludge to
learn which method was. the most effective and practicle to
scale up to a full size plant. Of the methods tried, sludge
roasting resulted in the lowest Hg in the clinker by 2 - 3
orders of magnitude.
Concurrently with the sludge trials, tests were
performed in the laboratory on Hg contaminated waste water
to select a process capable of removing Hg to meet the Jan-
uary 1, 1971* limit of 45 g (0.1 lb) per day maximum Hg
discharged in the water effluent. The literature survey
revealed a large number of alternate water treatment methods
tried, proposed, or potentially effective. Several samples
of ion exchange resins and activated carbons claimed to re-
move Hg from water were purchased.
As the data show (tables 11, 16, 17, 18, 19 and
Figure 19) a large number of tests were performed varying
the parameters of concentration, reaction or residence time,
?H and filtration, methods. In addition to resins and car-
sons, reducing agents (both metallic and chemical), sulfide
Ion precipitation, and flocculating agents were tried. The
nost consistently effective and practical method from these
sxperiments was a combination pH adjustment and sulfide
precipitation followed immediately by filtration on a
Jrecoated filter.
55
-------
Table 15. NAMES AND ADDRESSES OF THE COMPANIES
CONTACTED FOR INFORMATION BY DIRECT
COMMUNICATION DURING THE PROJECT
Company name
a) Aktiebolaget Billingsfors-Langed
b) FMC Corporation
c) Weyerhaeuser Company
d) Stauffer Chemical
el Chemapec, Inc.
f) Crawford & Russell, Inc.
g) Rohm and Haas Company
h) Sobin Chlor-Alkali, Inc.
i) Ventron Corporation
j) Hoechst-Uhde Corporation
k) Diamond Shamrock
Address
S-660 11 Billingsfors
Sweden
Squamish, B.C.
Canada
Chlorine Plant
P. O. Box 188
Longview, Wash. 98632
Axis. Alabama 36505
1 Newark Street
Hoboken, N. J. 07030
Stamford. Conn. 06904
Ion Exchange Dept.
Independence Mall West
Philadelphia, Penn. 19105
P. O. Box 149
Orrington, Maine 04474
Congress Street
Beverly. Mass. 01915
550 Sylvan Avenue
Englewood Cliffs
New Jersey 07632
Deer Partk, Texas 77536
56
-------
Table 15. CONT.
Company name
1) Wyandotte Chemical Corp.
m) B. F. Goodrich Chemical Co.
n) Monsanto Company
o) British Petroleum (BP) Chemicals
p) Mo Och Domsjo
q) Finnish Chemicals Oy
Address
P. O. Box 161
Port Edwards, Wise. 54469
P. O. Box 527
Calvart City. Kentucky 42029
Sanget. Illinois 62201
Murgatroyd's Works
Sandbach, Cheshire
Husum, Sweden and
Ornskoldsvik, Sweden
Aetsa, Finland
57
-------
Table 16. DATA FROM LITERATURE ON ION EXCHANGE RESINS
Literature source
Hg level after this series of tests.
Initial Hg ppb
level. Pre- Polishing
ppb filtration I. E, resin resin
(5) Osaka Soda process «w 20, 000 /u 5. 000
(18) Dow Chemical patent
#3.083,079
(28) A.B. BiUingsfors -
Langed
(41) Terraneers process
(ion exchange or ad-
sorbant material not
specified)
(62) Ajinomoto Co. of N. Y. 1.000- Yes
15, 000 Not measured
15. 000
2. 000-
5,000
29. 000-
70, 000
None
None
None
150
2-5
300
100-200 10-20
110-1500
1-10
58
-------
Table 17. LAB TEST DATA ON ION EXCHANGE RESIN
FOR Hg REMOVAL FROM WATER
Resin Company
Ajinomoto
Ajinomoto
Ajinomoto
Ajinomoto
Ajinomoto
Ajinomoto
Billingsfors-Langed
Billingsfors-Langed
Billing sf or s - Lang ed
Billingsfors-Langed
Billingsfors-Langed
PH
11
11
1.5
6.0
1.5
6.0
11
11
1.5
6.0
6.5
Initial Hg.
ppm
13.5
1.8
.06
.087
189.00
205.00
13.5
1.8
189.00
205. 00
0.035
Final Hg,
ppm
0.38
.99
.005
.003
1.9
0.4
1.8
2.0
51.5
15
.001
Hg removal,
%
97.2
45.0
92.0
96.5
99.0
99.8
86.7
0
63
92.7
97.0
NOTE: All tests were performed in a glass column 3/4" i. d. with a
packing depth of 12 inches. Flow was controlled by main-
taining a head of liquid 3-6" above top of packing.
59
-------
Table 18.. LAB TEST DATA ON ACTIVATED CARBON FOR Hg REMOVAL FROM WATER
Activated Carbon
Nuchar 722
Nuchar 722
Nuchar 722
Huchar 722
Nuchar 722
Nuchar 722
Pittsburgh HGR
Pittsburgh HGR
Pittsburgh HGR
Pittsburgh HGR
Pittsburgh HGR
Calgon Filtersorb 400
Calgon Filtersorb 400
Initial
pH
11.5
11.5
6.0
1.5
6.0
1.5
11.5
11.5
6.0
6.0
1.5
11.5
11.5
Initial
Hg,
ppm
13.5
1.8
.087
.060
205
189
13.5
1.8
.087
205
189
13.5
1.8
Final
Hg,
ppm
0.23
.02
.006
.0045
.37
.1
0.43
.47
.020
3.1
16
0.73
.03
Hg
removal ,
%
98.3
98.9
93.0
92.5
99.8
99-. 95
96.8
73.9
77.0
98.5
92.0
94.6
98.3
NOTE: All tests were performed in a glass column 3/4" l.D. with a packing
depth of 12 inches. Flow was controlled by maintaining a head of
liquid 3-6" above top of packing.
60
-------
Table 19. LAB TEST DATA ON Hg REMOVAL FROM
WATER BY REDUCTION
Reduction agent
Zinc particles
Sodium borohydride
SnClo
Material
form
10 mesh
Liquid
Solution
PH
11.5
10.0
6.0
2.5
6.2
12.2
____
Initial Hg,
ppm
1.8
12.5
12.5
12.5
52.0
10.7
4.0
26.0
2.8
Final Hg,
ppm
0. 14
.83
.75
.47
0.09
0.22
.42
.82
.5
Hg removed,
%
92.2
93.4
94.0
96.2
99.83
98.0
89.5
96.85
82.0
61
-------
Figure 19. Lab tests on mercury removal from water using sulfide ppt.,
activated carbon, ion exchange resins and reduction chemicals
•a
a.
•
2
^H
w
H
**
05 g
K3 ^
10 a
F-
2
H
D
J
fa
fa
W
X X
-, '' y
• NallS pH 7-14 T X y
0 NaflS pll 5-7 Sulfide X .' A
- O NallS pll 0-5 Jj ^X / /
• Nuchar 722 ~~1 V X X
B Pittsburgh HGR — Carbon ^X QX ^X
D Filtersorb — J X X X
A Ajinomoto ~1_ Resin / / */
A Billingsfors-Langed_J x yr >
+ Zinc ~"| /f • AX .*
X SnCl2 1— Reduction xx xX /
- - NaBH4J / A x' / /'
X/ r* <* / '' ''
' / .1 x/° ^ /
v X / _— X X
x^ / x °x x
^ ./* °/°° +/'
&* / s S 9 / x
x x / x x
/ „ / °/ ";/ • /
/ B rA<>X • X 0 X
XX ^X /X X ° X
XX« Af/ / ° o XX
xxxj V x /
X A XX QAV Q^°X
X / CX' X CX'X
^X x ^ X ^>^ 00
/ A l/X IX Ix'' 1
STARTING Hg CONTENT OF THE SOLUTION, ppm
-------
Following the selection of the water and sludge
treatment processes, the preliminary process design was
drawn up and cost estimates made (figures 8 and 14) . During
the interval between the preliminary process design and full
scale plant start-up, there were several equipment and pro-
cess changes. The system in operation is essentially that
shown in figures 18 and 20.
The major pieces of equipment consisted of a
filter and furnace in the sludge system and a filter in the
water system. Due to the need for nearly complete solids
removal from the water phase and the high solids, content,
vacuum filtration was selected (Figure 21).
A number of different furnace designs were ^
sidered but this was narrowed to two basic designs for pilot
tests due to temperature limits and solids handling problems.
Tests were conducted on a rotary calciner, indirect fired
and on a multiple hearth furnace (MHF), direct fired. The
multiple hearth furnace was selected because it produced
lower Hg levels in the clinker (Figure 22).
The filter in the sulfide precipitation system
called for a type that would remove a small amount of fine
solids from, a water stream with minimum losses. This
narrowed the filter selection to a pressure filter and prob-
ably a precoated pressure filter due to the fine particle
size (Figure 23).
Const ruct ion
As equipment selection was made the first stages
of construction took place. Engineering drawings and a
plant model (Figure 17) were made. The support structure
and foundation were designed and construction was started.
As the major pieces of equipment arrived they were installed.
The multiple hearth furnace is. shown in Figure 22, the rotary
vacuum filter in Figure 21, 'the R. P. Adams filter in Figure
23, and the sludge thickener in Figure 24.
Once the major equipment, pumps and various tanks
were in place, the piping was laid by the contractor, and the .
electrical contractor was called in to wire the process. A
separate contractor was hired to design and install the in-
strumentation for the multiple hearth furnace. One major
piece, of construction involved piping the natural gas from
the nearest location to the furnace, a distance of 350 m
(1150 ft).
63
-------
Figure 20. Installed sulfide precipitation system for water treatment
OJ
PRESSURE
FILTER
TO SLUDGE
TREATMENT
FILTER
TO SEWER
DIATOMACEOUS
EARTH
DRY
FEEDER
Na2S
STOW
TANK
FILTER FEED PUMP
FILTER
AID AND
PRECOAT
MIX TANK
GE
9
MERCURY
WASTE
WATER
STORAGE
pH ADJUST
MIX TANK
CELL
ROOM
SUMP
-------
en
Figure 21. Installed 6' x 6' Eimco rotary vacuum filter for brine sludge dewatering
-------
CJ>
05
Figure 22. Installed 54" i. d. BSP multiple hearth furnace
-------
Figure 23. Installed R. P. Adams filter for HgS removal
67
-------
Figure 24. The 12' x 6' sludge thickener prior to the rotary vacuum
filter in the sludge treatment system
68
-------
Our plant instrument department designed and in-
stalled the remainder of the instrumentation for the sludge
and sulfide system. The nearly completed plant is shown in
Figure 25.
'Start-up-
The start-up team was selected during the con-
struction phase and the start-up leader began, operating and
checking out the equipment as soon as each piece was com-
pleted. An operating manual was written for use in operator
training (Appendix D).
As the start-up date approached, the start-up
leader held several oner-hour training classes with each op-
erating shift, going over the process, the model, the de-
sired operating procedures and the installation. Feedback
from the operators was valuable in correcting minor problems
apparent before start-up began.
Numerous problems were encountered during the
start-up and were corrected, as discussed earlier. The
start-up was divided, into two parts since the water and
sludge systems are nearly independent. The water treatment
system was started up about two months prior to the sludge
system.
INNOVATIONS AND NEW TECHNIQUES
Waste Water Treatment Innovations
In our laboratory work, and confirmed on a plant
scale, the critical operations in the sulfide removal
system are the control of pH and rapid filtration following
the sulfide addition. The sulfide and filter aid addition
equipment is shown in Figure 26.
Sludge Treatment Innovations
For our sludge, the furnace roasting process
achieved Hg levels, in the clinker 2 to 3 orders of magnitude
lower than the lowest Hg levels possible after chemical
treatment (Table 2 and Appendix A). Still further reduc-
tions could be achieved by acid treating the sludge before
roasting, as shown in Table 7, although, this method had the
disadvantage of generating foam and lowered the fusion
temperature of the clinker.
69
-------
Figure 25. The full scale Hg Recovery System as installed at
the Bellingham Chlor-Alkali Plant
70
-------
Figure 26. The mix tank in the sulfide precipitation system
where the D. E. and sulfide are added "
71
-------
Prior to this work, it was reported that the
brine sludge was too sticky to dewater with a rotary vacuum
filter alone; a precoat filter would have been required,
causing greater operating expense and requiring more opera-
tor attention. The pilot and full scale plants have
demonstrated that our sludge dewaters easily on a rotary
vacuum belt filter.
One further discovery was that the Hg present in
the sludge was concentrated in the graphite particles
present; the smaller particles had a much higher Hg content,
990 ppm, than the larger particles, 100 ppm,(Figure 27).
ECONOMIC ANALYSIS
Water Treatment System
The economics of the water treatment system are
shown in Figure 20 and Table 20. The system as installed
cost $1^3,900 to construct and will handle up to 570,000 1
(150,000 gal) per day. The operating costs include
chemicals, electricity, operator time and maintenance cost.
These costs total $510 per week or 13
-------
Figure 27.. Particle distribution in brine sludge
which was washed and screened to remove particles
0,007" diameter (Experiment 34)
80
§
H
•U
S5
W
1 40
Q
g
H
w 20
o
3
0
CO
0
\
\
\
\
\
\
\
\
\
\
\
i N
^ ^
X.
\ ^-v^
\ *"*•.
\ -—^
\. ^--_, -
"
•
-
- ^—
-
^
•
,— r~
''
I .
i\t\j\/
60
se
o.
ta
N
M
en
a
o
h-l
H
500 <
Ou
W
H
H
g
U
bO
0
i r
.05
i i r i i i
.10 .15
PARTICLE DIAMETER, inches
73
-------
Table 20. COST ESTIMATE- WATER TREATMENT SYSTEM
Item Description
Labor
Materials
Total
Filter and installation
Pumps
$ 3,000
4.700
Instrumentation and controls 9, 000
Tanks and vessels
Piping and valves
Electrical
Painting
Structure, ladders and
3, 100
41,000
7,000
1, 100
platforms 8, 700
Engineering at MH @
$16,000
4.000
6,700
5.600
14, 600
4.000
460
4,400
Subtotal
$10.00 MH
Total investment required
$19,000
8,200
15, 700
8,700
55,600
11, 000
1,560
13, 100
132,860
11,000
$143,900
74
-------
Table 21. COST ESTIMATE - SLUDGE SYSTEM
Item Description
Multiple Hearth Furnace
Rotary vacuum filter
Incinerated solids screw feeder
Sludge feed
Instrumentation and controls (System)
Furnace instrumentation and controls
(Union Heating)
Sludge piping and thickener
Heat exchangers and associated
off gas piping
Structure, ladders and platforms
Natural gas and water service
Site preparation and foundation
Pumps
Painting
Electrical
Labor
$ 9.000
2,700
1,200
4,000
6,200
28.000
6,300
31,000
14, 000
8.000
3.000
8,000
12,000
Engineering at MH
Materials
$58, 000
26, 000
4.800
7.500
2. 600
19.000
22,000
14. 000
11, 000
4,000
8,000
2.200
6,000
Subtotal
@ $10.00 MH
Total investment required
Total
$ 67,000
28,700
6, 000
11,500
8,800
18, 000
53, 000
28, 300
45, 000
25, 000
12, 000
11,000
10, 200
18,000
342,500
22,000
$364, 500
75
-------
SECTION VIII
REFERENCES
1. Botwick, E. J. and D. B. Smith. Mercury Recovery.
U. S. Patent 3,600,285. 1971.
2. Ventron Mercury Removal Process. Koertrol, subdi-
vision of Ameteck, cata. 7R. 1974. p. 4lB.
3. Caban, R. and T. W. Chapman. The Extraction of HgCl
From Acid Chloride Solutions With Trioctylamine. A.
I. Ch. E. Journal. 18(5):904, 1972.
4. Moore, F. L. Solvent Extraction of Hg from Brine So-
lutions with High Molecular-Weight Amines. Environ-
mental Science and Technology. 6:525-9, June, 1972.
5. Edwards, G. E. and N. J. LePage. Treatment of Brine
Solutions Containing Both Free and Available Chlorine
Plus Mercuric Chloride. U. S. Patent 3,102,035- 1963.
6. Edwards. G. E. and N. J. LePage. Treatment of Brine
Solutions. Great Britain Patent 885,818. 1961.
7. Hg Removal by Hydroxide Floe and Filtration. Hayward
Filter Co. Santa Ana, California. Lit. No. M-10,
No. 127, 44-A, 119. 1971.
8. Suhara, I., et al. Removal of Mercury from Waste
Liquid Containing Mercury. Japan Patent 3405. 1964.
9. Bergeron, G. L. and C. K. Bon. Mercury Recovery from
Electrolytic-Cell Brine Effluents. U. S. Patent
2,860,952. 1958.
10. Bouveng, H. 0. and P. Ullman. Sweden Tightens Up On
Mercury Wastes. Industrial Water Engr. pp. 24-6,
June, 1969.
11. Chlorine Institute, May publication on mercury state
of the art. Published by the Chlorine Institute. 1971.
12. Chlorine Institute. Minutes of the ad hoc mercury
committee. Published by the Chlorine Institute. 1970,
1971.
76
-------
13. Deriaz, M. G. Recovery of Mercury from Waste Brine
by Sulfide Precipitation. U. S. Patent 3,213,00.6.
1965.
14. Fulcher, R. A. Sulfide Ion Electrode. Personal Corn-
muni cat ion. 1971.
15. Gardiner, W. C. Mercury Problems in the Chlor-Alkali
Industry. Presented to Niagara Palls section of
Electro-Chemical Society. Published by the Electro-
Chemical Society. November 9, 1970.
16. Hirs, G. Private Communication to W. H. Hunt. July
28, 1970.
17. Hazards of Hydrogen Sulfide. Safety News Letter Pulp
and Paper Section. National Safety Council. November,
1971.
18. Wilkes, A. Private Communication. 1971.
19. Knepper, W. and S. Austin. Process for Recovering
Mercury from Waste Waters of Industrial Process. U.
S. Patent 3,695,838. 1972.
20. Dean, W. E. and C. M. Dorsett. Mercury Removal.
U. S. Patent 3,67^,428. 1972.
21. Maloney, G. F. Selecting and Using Pressure Leaf
Filters. Chemical Engineering. 79(11):88-94, 1972.
22. Bouveng, H. 0. Control of Mercury in Effluents from
Chlorine Plants. Presented at International Congress
on Industrial Waste Water (Stockholm), November 2-6,
1970. Butterworths Pub. (London). 1972.
23. Bouveng, H. 0. and P. Ullman. Reduction of Mercury in
Waste Waters from Chlorine Plants. Swedish Air and
Water Pollution Research Laboratory (Stockholm).
April, 1969.
24. Smith, S. B., et.al. Mercury Pollution Control by
Activated Carbon: A Review of Field Experience. West-
vaco Corp. No. M1002.01. 1971.
25. Yokota, Y. Recovery of Mercury. Soda and Chlorine.
19(3):87-95, 1968.
26. Tokawa, D. T. Treatment of Mercury Cell Waste. B. S.
Thesis. University of British Columbia. 1971.
77
-------
27- Glaeser, W. Method of Producing Mercury. U. S.
Patent 1,637,481. 1924.
28. Parks, G. A. and R. E. Baker. Mercury Process. U. S.
Patent 3,476,552. 1969.
29. Parks, G. A. and N. A. Fittinghoff. Mercury Extrac-
tion Now Possible Via Hypochlorite Leaching. Engin-
eering and Mining Journal, pp. 107-109, June, 1970.
30. Town, J. W. and ¥. A. Sttckney. Cost Estimates and
Optimum Conditions For Continuous-Circuit Leaching.
U. S. Dept. of the Interior Bureau of Mines report of
investigations. 6459. 1964.
31. Osaka Soda Mercury Recovery Process. Crawford &
Russell, Inc. Private Publication. November 5, 1970.
32. Gardiner, W. C. and P. Munoz. Mercury Removed from
Waste Effluent via Ion Exchange. Chemical Engineering.
79(19):57-59, August 23, 1971.
33. Scheiner, B. J., R. E. Lindstrom, D. E. Shanks, and
T. A. Henrie. Electrolytic Oxidation of Cinnabar
Ores for Mercury Recovery. U. S. Dept. of the Interior,
Bureau of Mines Metallurgy Research program. Technical
Progress Report - 26. June 1970. 11 pp.
34. Anon. Process Removes Hg in Plant Wastes. Chemical
& Engineering News. 48:48, December 14, 1970.
35- Allenbach, C. R. Reduction by Gallium, Aluminum,
and Mercury in Aqueous Solution. University Microfilms.
Ann Arbor, Michigan. No. 3927. 1952.
36. Gilbert, J. F. and C. N. Rallis. Recovery of Mercury
from Brines. U. S. Patent 3,039,865- 1959.
37- Karpink. R. S. and J. J. Hoekstra. Recovery of Hg
from Brine. U. S. Patent 3,029,143. 1962.
38. Karpluk, R. L. and J. J. Hoekstra. Mercury Recovery
Using Liquid Alakll Metal. U. S. Patent 3,029,144.
1962.
39. Neipert, M. P. and C. K. Bon. Reduction of Mercury
Ion to Metallic with Aldehyde. U. S. Patent 2,885,282.
1959.
78
-------
40. Rickard, M. D. and G. Brookman. Metal Reduction Aid
for Mercury. Water and Wastes Engineering, p. D-2,
July 1971.
41. Rhodes, D. W. and M. W. Wilding. Reduction of Hg in
Solution. TJ. S. Patent 3,^63,635. 1969-
42. Smith. W. W. Reducing Hg"*"1" and Hg+ to Hg with
and NaP. Chemical Engineering (London). Vol. 216, 1968.
43. Anon. Chelating Resin Separates Ions. Chemical and
Engineering News. 37:^9, January 26, 1959.
44. Anon. Paring Mercury Pollution. Chemical Engineering.
78(5):70-71, 1971.
45. Anon. Application of Resinous Mercury Adsorbent (Che-
lating Resin for Heavy Metal) to Sewage Disposal.
Ajinomoto Co., Inc. Tokyo, Japan. 1971.
46. Anon. Resinous Mercury Adsorbent. Ajinomoto Co., Inc.
Tokyo, Japan. Technical Data Bulletin No. 71927.
1970.
47. Calkins, R. C., et al. Removal of Mercuric Ions Prom
Electrolytic Solutions. U. S. Patent 3,083,079. 1963.
48. Grain, G. E. and R. H. Judice. Electrolytic Process
for the Recovery of Mercury. U. S. Patent 3,213,006.
1965.
49. Walton, H. P. and J. M. Martinez. Reactions of
Mercury (II) with a Cation-Exchange Resin. J. Phys.
Chem. 63:1318-19- 1959-
50. Puxelius, L. Ion-Exchange Resin for Removal of Heavy
Metal Ions in Waste Water. Presented at International
Congress on Industrial Waste Water (Stockholm), November 2-6,
1970. Butterworths Publ. (London). 1972.
51. Hokuetsu Tanso Kogyo, K. K. An Introduction to Mer-
cury Adsorbing Resins. Hikari Kogyo K. K., 11 Takara-
cho 2-chome chuo-ku, Tokyo, Japan.
52. Kraiker, H. Micro-Ionic Systems for Mercury Pollution.
L. A. Water Conditioning Bulletin No. 1100. 1970.
53. MacMillian. A. L. Private Communication. 1971.
79
-------
54. Morissette, B. G. Recovery of Hg from Brine. Cana-
dian Patent 595,813. I960.
55. Percival, R. W. Private Communication. 1970.
56. Rohm and Haas Co. Recovery of Mercury by Ion Exchange.
Private Communication from C. T. Dickert. 1970.
57. Rosenzweig, -M. D. Paring Mercury Pollution. Chemical
Engineering. 78(5):70-71, 1971.
58. Scholten, H. G. and G. E. Prielipp. Hg Removal by Ion
Exchange Resins. U. S. Patent 3,085,859. I960.
59. Selezrieva, N. A., et al. Separation of Selenium and
Mercury on Anion Exchangers. Inst. Yad. Fiz., Alma-
Ata (USSR). 19(!*):76-77S 1969.
60. Tsujiya, T. Private Communication. 1971.
61. Law, S. L. Methyl Mercury and Inorganic Mercury Col-
lection by a Selective Chelating Resin. Science.
17^:285-286, 1971.
80
-------
SECTION IX
PATENTS AND PURIFICATIONS
1. Patent Application - "Removal of Mercury from Mercury
Cathode Sludge", Donald A, Kaciior
and Richard A. Perry, Patent Ap-
plication Serial No. 354,983',
filed April 27, 1973.
2. Publication - Perry, R. A^, Mercury Recovery from
Process Sludges, Chemical Engineering
Progress, 70'(3') :73-86, 1974.
81
-------
SECTION X
GLOSSARY
1. Body feed - A filter aid added continuously to the sus-
pension to he filtered to keep the filter from
plugging.
2. Brine sludge - Sludge resulting from chemical addition
to sodium chloride brine to precipitate calcium
and magnesium compounds and other impurities.
3. Cell anode - One of the electrodes in a Hg cell, made
of graphite or metal.
4. Chlor-alkali plant - A plant producing chlorine and a
metal hydroxide.
5. Diatomaceous earth - A meterial used to precoat filters
and as a filter aid.
6. Effluent - The waste liquid discharged from a process.
7. Electrolytic oxidation - The generation of chlorine in
a brine with electricity to cause oxidation of a
desired material.
8. Hg - The chemical symbol for the element mercury.
9. Hg contaminated waste water - The waste water which
comes in contact with Hg or Hg-containing material
in a chlor-alkali process.
10.' Mercury cell - The unit producing chlorine and a metal
hydroxide from electricity and brine cathode.
11. Multiple hearth furnace - A direct fired furnace with
trays on a vertical shaft.
12. Precoat - A filter aid added to coat the filter ele-
ment before filtration begins.
13. Pressure filter - A filter which uses pressures
greater than atmospheric pressure on the unfiltered
side.
14. Polishing filter - A final filter which removes the last
traces, following a preliminary filter.
82
-------
15. Rotary calciner - An Inclined cylinder heated and ro-
tated; material is passed through the cylinder.
16. Rotary vacuum filter - A filter utilizing vacuum inside
a eye Under to pick up and de water a cake on the
outside of a drum.
17. Star valve - A rotating paddle wheel that allows solids
to pass through but seals the opening against air
leakage.
18. Thickener - A large continuously fed tank which concen-
trates or thickens a sludge to a higher total
solids.
19. Untreated sludge - Chlor-alkali plant sludge as it comes
from the process.
83
-------
SECTION XI
APPENDICES
Page
A. Hypochlorite, Chlorine and Electrolytic 35
Oxidation
B. Determination of Hg by Flameless AA 95
C. Experimental Data for Alternate Hg Re- 102
covery Methods from Water
D. Operating Manual 108
84
-------
APPENDIX A
HYPOCHLORITE, CHLORINE, AN0 ELECTROLYTIC OXIDATION
The object of the chemical oxidation trials was
to convert Hg to the soluble mercuric ion in the presence
of chloride ions to form the soluble mercuric tetrachloro
complex. The overall reactions involved are as follows:
Hg + CIO" + 3Cl" + H2O ^ HgCll + 2OH~
2Hg+ + CIO" + 7Cl" + H20 v 2HgCl* + 2OH~
Hg++ + 4C1" v HgCl^
Procedure
The trials were conducted using 250 ml to one 1
of brine sludge for each test. Trial conditions and results
are shown in Table 22.
The hypochlorite oxidation trials involved mix-
ing liquid sodium hypochlorite with a brine sludge in an
agitated beaker for the time period specified. The treated
sludge was filtered in a Buchner funnel and washed with
200 - 400 ml of distilled water before analyzing the washed
sludge for residual Hg (Figure 28).
The chlorine oxidation trials were similar ex-
cept that gaseous chlorine was sparged into the sludge from
a cylinder of liquid chlorine.
Results
From the results of bench scale trials, the hypo-
chlorite appeared to dissolve nearly as much sludge solids
(60 - 80$) as Hg (60 - Q6%). Thus, up to 90% of the Hg
was removed, but the solid residue contains 100 - 300 ppm
Hg. In the work of Tokawa (26), a one-stage hypochlorite
extraction removed 40 - 1Q% Hg and 4-8 stages were needed
to removed over 95% of the Hg (Table 23 and Figure 29).
Other investigators have suggested that the use
of chlorine gas as the oxidizing agent would reduce re-
sidual Hg in the treated sludge to 1 ppm. In the 9 experi-
85
-------
Table 22. OXIDATION OF BRINE SLUDGE USING SODIUM HYPOCHLORITE
CO
Exp.
No.
1A
IB
1C
ID
2A
213
3
4A
4B
5
6A
6B
6C
7
8
12
15
17
29
Hypo
concentration,
gpl C12
30
30
30
3D
30
30
30
90
90
90
90
90
90
87
89
190
150
150
100
PH
6.2
6. 2
8.5
8.5
5.7-10
5.2-10
7.3
11.2
11.2
6.8-10. 8
10.0
9.0
8.0
8.0
8.0
8.0
9.0
8.0
8.0
Reaction
time.
hr.
4
4
4
4
3
3
4
3
3
4
1
1
1
23
1
1
1
1
3
Temp..
•c
60
22
60
22
60
60
60
60
25
60
60
60
60
60
60
60
60
60
8-11
Initial
Hg content,
ppm
270
270
270
270
270
270
270
270
270
270
270
270
270
270
270
150
150
4200
260
Solids Remaining
dissolved. Hg removal. Hg in solids,
% % ppm, dry
304
197
169
175
110
A; 100
A^lOO
93
62%
64%
73.5%
168
«vl258
76% 77% 140
78% 86% 282
47
70% 76% 130
Comments
2 stage
4 stage
2 stage
Pressure
reaction
cell graphite
-------
Figure 28. Bench test set-ups for chemical oxidation of sludge
HYPO HCl
6
MAG MIXER
HYPO METHOD
Cl,
6
MAG MIXER
CHLORINE GAS METHOD
GRAPHITE PIATES
6
MAG MIXER
ELECTROLYTIC OXIDATION
METHOD
POWER SUPPLY
87
-------
Table 23. OXIDATION OF BRINE SLUDGE USING SODIUM HYPOCHLORITE
WORK PERFORMED AT UNIVERSITY OF BRITISH COLUMBIA
oo
00
Run No.
Stage No.
START
1
2
3
4
5
6
7
8
F
Hg „
recovered,
%
--
40%
67
81
87
92
96
98
99
Hg
remaining in
sludges, ppm
3280
1970
1080
620
430
260
130
.66
33
G
Hg
recovered,
%
-_
70
86
91
94
95
Hg
remaining in
sludges, ppm
931
280
130
84
56
47
I
Hg
recovered,
%
_ _
70
88
93
96
97.5
Hg
remaining in
sludges, ppm
146
44
18
10
6
4
-------
Figure 29. Effect of staging on mercury recovery by Tokawa
100
50
89
-------
ments performed in these trials, half of the residues
contained more Hg than the starting material (Table 2*1.)...
The chlorination of the sludge dissolved 70 - 9Q% of the
solids and approximately the same quantity of Hg so there
was no net reduction of Hg concentration in the residual
solids.
A. sample of brine sludge from a chlor-alkali plant
utilizing metal anodes was oxidized with hypochlorite and
chlorine gas. Over $8% of the Hg was removed, leaving a
solids residue containing 9.2 ppm Hg (Table 25). This
suggests that metal anode sludges are susceptible to
chemical conversion of Hg to the tetrachloro complex.
ALTERNATIVE Hg RECOVERY METHODS
A. Hypochlorite and Chlorine Oxidation
The initial work on extracting Hg from brine
sludge was performed using sodium hypochlorite. This
method was known to remove Hg as early as 1924 from
Glaeser's work (27). Extensive work has been done on
sodium hyopchlorite leaching of Hg from low grade ores
by Parks (28, 29), Town (30) and others. Although Parks
achieved a 96-4$ recovery, there still as 10 ppm Hg left
in the residue. Town was able to achieve a 99.8$ removal
by leaching of very concentrated ores (79#) but
there remained 24,000 ppm Hg in the residues.
In Japan, this process has been used by 4 chlor-
alkali plants for up to 5 years to remove Hg from brine
sludge. This process has been marketed in the United
States since 1970 by Crawford & Russell, who claim the
process will remove 95% or more of the Hg from brine sludge
(31). In a more recent publication, Crawford & Russell
claimed a reduction of Hg in the dry sludge from 50 - 4000
ppm to 0.1. ppm via the hypochlorite leaching process with
pH adjustment (32). Prom our work, the 95$ Hg removal
stated in their sales literature is more realistic than the
0.1 ppm Hg residual claimed.
The basis for this process is the conversion of
elemental Hg and insoluble Hg compounds to water soluble
mercuric Ions with the OC1. The soluble stable complex
HgCl]j is formed.
Tokawa found that multistaging the extraction
process could increase the Hg recovery to 99% with 8
stages (26). However, the maximum achieved in our labora-
tory in one stage was 86% recovery with a minimum final
90
-------
Table 24. OXIDATION OF BRINE SLUDGE USING CHLORINE GAS
Exp.
No.
14A
14B
23
24
25
26
27
28
30
Chemical
Chlorine/ Flotation
Chlorine / Flotation
Chlorine
Chlorine
Chlorine
Chlorine
Chlorine
Chlorine
Chlorine
Time, hr
...
2.5
5.0
5.0
1.75
7.0
7.0
7.0
Start
10.0
10.0
10.0
5.0
10.4
10.3
10.3
1.7
10.0
End
5.0
5.0
4.7
4. 1
4.5
4.0
3.6
1.5
4.5
Initial Hg
content, % solids
ppm reduction
150
150
150
150
150
150
150
150
150
-._-
70.0
86.5
86. 0
--..
88.0
% removal
...
...
74
'83
78
45
...
Remaining
Hg in solids.
ppm, dry Comments
114 Foam Hg content
102 Solids
200
163 Reducted pH to 5
,,o before Clo
.1 jO
314
A, 400
A/400 Acid addition to
117 PH 7
-------
CO
to
Table 25. OXIDATION OF BRINE SLUDGE USING COMBINATIONS OF HYPO,
CHLORINE, ELECTROLYTIC ACID TREATMENT AND ROASTING
Exp.
No.
9
10
11
18
29
20
21
22
36
54
Conditions of test
H2SO4 + Hypo
Electrolytic oxidation
Hypo + electioxidation
Hypo + chlorine gas
HC1 only
Hypo + chlorine gas
Hypo after acid treatment
Hypo of roasted
Leaching of roasted
Hypo + chlorine
Initial
Hg content,
ppm
137
150
150
150
150
4200
81
0.26
1.7
158
% solids
reduction
...
31
53
53
59
50
72.4
% recovery
88
—
—
56
72
85
54
17
--
98.4
Final
Hg in solids,
ppm. dry
65
111
120
43
81
150
100
0.2
3.0
9.2
Comments
Cell graphite
Acid treated
Roasted
Roasted
Metallic anode sludge
(Wyandotte)
-------
Hg content of 47 ppm. Due to the great difficulty and
expense required to separate the liquid from the fine solids
at each stage In a multistage extraction, a practical
system would be restricted to a one- or two-stage operation.
It became evident that the chemical equilibrium was not
favorable to remove the residual 5' - 1Q£ of the Hg in the
sludge in one stage. The use of chlorine injections did not
improve Hg conversion.
In the laboratory, graphite in the sludge was
concentrated as the sludge was digested. In addition,
the Hg concentration in the graphite particles was
found to be 6-8 times higher than the Hg level in
the remainder of the sludge (Table 26). Prom these
and other data, we hypothesize that the graphite from
the anodes and decomposer packing is a major contributor
to the residual. Hg after sludge digestion. This is further
supported by the results from a sample of metal anode
sludge (little graphite present) treated with hypo
and chlorine (experiment 54 and Table 25). The final
Hg level-was-' 9.2 ppm compared to a sample of our sludge
treated exactly the same (experiment 18, Table 25)
which contained 43 ppm Hg after treatment.
The distribution of Hg content in our sludge
of various graphite particle size groups is shown in
Figure 27. Over 80 weight percent of particles are 0.76 mm -
1.8 mm (0.03 - 0.07 in.) in diameter. Moreover, the Hg
content increases as the particle size decreases. This is
consistent with an adsorption mechanism of Hg on the
graphite, since the surface area increases with decreasing
particle size.
Attempts were made to remove graphite from sludge,
but abandoned for 3 reasons: (1) the sludge would still
have to be treated for Hg removal; (2) removal of the
very fine graphite particles with the highest Hg content
was difficult and costly; and (3) to remove the graphite
quantitatively from the sludge was not practical.
93
-------
CO
Table 26. Hg ANALYSIS OF BRINE SLUDGE SIZE FACTIONS
Sample Description
Brine sludge before separation
Large graphite particles in sludge
Remaining fines in sludge
Hg content
wet basis,
ppm
45
646
26
Total
solids,
%
35
40
10-15
Hg content
dry basis,
ppm
270
1600
175-250
-------
B. Electrolytic Oxidation
A modification of the hypochlorite and chlorine
oxidation methods for treating brine sludge is generation
of hypochlorite ions during sludge treatment. This
method has been used successfully by Scheiner (33) of the
U. S. Bureau of Mines to extract 90 - 95$ of the Hg from
ores containing 300 - 10,000 ppm Hg. The electrolytic
method simply uses dissolved sodium chloride in the sludge
mixture to liberate chlorine gas when a DC voltage is
applied across 2 graphite electrodes. Scheiner believes
the tiny chloride gas bubbles formed and the reaction
between Hg and chlorine at the surface of the electrodes
provides a more effective oxidation than simple injecting
chlorine gas or hypo into the slurry.
Our tests were performed in an apparatus consist-
ing of 2 graphite plates with an area of 13 cm2 (2 in2) in
a 1000 ml beaker spaced 2.5 cm (1 in) apart (Figure 20). A
DC power source supplied current to the brine-sludge solution
to generate chlorine gas from the sodium chloride present".
The tests were not significantly more successful than the
chemical oxidation methods; residual Hg levels in the re-
maining solids averaged 100 ppm.
95
-------
APPENDIX B
DETERMINATION OP Hg
BY FLAMELESS AA
The procedure used to analyze samples for Hg
during the course of this project was modified from an EPA
method published in 1970. The complete procedure is de-
scribed below.
I. Sampling
1. Rinse all glassware and polyethylene containers
with dilute nitric acid and then with distilled
water prior to use.
2. Acidify samples of water and effluent if they
will stand more than one day prior to determin-
ation of Hg. Add 10 ml concentrated HNOo per
1000 ml sample.
II. Equipment
1. Rinse all glassware with dilute HNOo and then
with distilled water prior to use.
2. Store glass beads in a small amount of concen-
trated HNO^ and rinse with distilled water prior
to use.
3. Modified A.O.A.C. digestion apparatus. Substi-
tute single-neck 250 ml or 300 ml boiling flasks
for the 3-neck digestion flask.
4. Perkin-Elmer Atomic Absorption Spectrophotometer,
Model 303. AA settings: wave length, 25^.5;
range, UV; slit width, 3; source current, 10 ma;
meter response, 1; scale, 1. Perkin-Elmer Re-
corder Readout: noise suppression, 2; scale ex-
pansion, x3. Airflow meter, set at *JO. Align
the gas absorption cell to allow maximum light
to pass through. Allow equipment to warm up at
least 20 minutes before using.
5. Aeration apparatus:
a. Use Anhydrone (magnesium perchlorate) as
the drying agent; change weekly or more
often as needed.
96
-------
CAUTION: If Anhydrone comes in contact with
skin or clothing, wash area immediately with
water. Magnesium perchlorate may cause
severe burns to skin or may cause fire when
in contact with clothing or combustible
material.
b. Clean the gas washing bottle biweekly with
a small amount of HF acid, rinse with water,
and clean again with dichromate cleaning
solution. Clean the sparger biweekly with
boiling dilute HC1. Rinse apparatus thor-
oughly with distilled water prior to use.
CAUTION: Hydrofluoric acid liquid and vapor
may cause severe burns which may not be im-
mediately painful or visible. Do not leave
glassware in contact with HF longer than is
absolutely necessary.
c. Do not allow moisture to collect in the 17
cm gas absorption cell. If moisture does
collect, dry cell thoroughly in a 105°C
oven and change the drying agent in the dry-
ing tube.
III. Procedure
A. Preliminary treatment of sample: Use modified A.O.A.C.
method for organic and solid samples and effluent: use
modified F.W.Q.A. method for inorganic aqueous samples,
caustic, and sulfuric acid.
1. Modified A.O.A.C. method:
Take suitable amounts of sample (not more than
100 ml or 5 g dry) to provide 0.1 - 1.5 ug Hg,
place in a single-neck flask and treat each ac-
cording to type of sample.
a. Samples
i. Mud, sludge, etc.
Add 10 ml distilled water to sample and
then add 10 ml concentrated HNOo per g
dry sample. Proceed with digestion as
below (IIIAlb).
ii. Effluent
Proceed with digestion as below (IIIAlb)
97
-------
b. Digestion procedure
To the single-neck flask containing the
sample, add 20 - 25 ml 1:1 HNCU -H2SOjj and
3-4 glass beads. Attach flask to modi-
fied A.O.A.C. digestion apparatus. Care-
fully heat sample until it refluxes steadily;
avoid losing gaseous NO too rapidly. Col-
lect condensate in extraction unit until
digest reaches incipient boiling or goes to
acid fumes.
If sample darkens or turns black, cool,
and add more concentrated HNOo.
Allow digest to cool; drain collected
liquids back into flask, and reflux for 10
- 15 minutes to rid apparatus and sample of
gaseous N02- (Add 25 ml distilled water to
sample through condenser if N02 is difficult
to remove. Reflux again for 10 - 15 minutes.)
Cool sample and rinse condenser with two 10
ml portions of water.
It may be necessary here to dilute the
sample to volume and take an aliquot of
sample before proceeding.
Proceed with F.W.Q.A. sample treatment
2. Modified P.W.Q.A. method: Take suitable amounts
of sample (not more than 100 ml) to provide 0.1
- 1.5 yg Hg, place in a 150 ml beaker containing
7 ml 1:2 HNOg -HpSOjj plus distilled water to make
a final volume of 100 ml. Treat each according
to type of sample.
a. H20, C12 plant effluent, NaOCl, samples
Proceed with modified P.W.Q.A. treatment as
below (IIIA2b).
b. Modified F.W.Q.A. sample treatment
Dilute sample aliquot to 100 ml with dis-
tilled water. Add 1 ml 5% KMnO^ and let
sample. stand for at least 15 minutes. Add
"2 ml 5% K2S2C>8, allow sample to stand at
98
-------
least 30 minutes and proceed with aeration
step as below.
B. Aeration Procedure
Connect aeration apparatus to spectrophotometers ;
adjust spectrophotometer, flow meter, etc., as in
After allowing apparatus to warm up, adjust baseline
and 100% absorption line with stopcock in bypass po-
sition. Proceed with aeration of sample, treating each
sample individually as below. Carry out each step with
as little delay as possible between steps:
1. Destroy excess permanganate with 2 ml 10%
HC1, and immediately wash the clear sample into
gas washing bottle.
2. Add 5 ml 10% SnCl2 to gas washing bottle. Im-
mediately replace gas washing bottle in the aer-
ation apparatus and turn stopcock to aeration
position.
3. After pen has returned to within 2% absorption,
turn stopcock to bypass; rinse gas washing bottle
and proceed with next sample.
C. Calculations
A series of 6 standards ranging from 0.10 - 1.5 ug Hg
is treated as for H20 and C12 plant sewer samples and
is run each time the spectrophotometer is operated.
Plot a calibration graph on semi-log paper with yg Hg
on the linear scale and percent absorption on the log
scale. Convert percent absorption of the sample to
pg Hg and determine Hg content as follows :
(yg Hg from graph) (dilution in ml)
Hg, ppm = _
(g sample, note 1)
IV. Notes
1. Assume specific gravity for volumetric samples
to be 1.0 for dilute liquids, 1.5 for 50% caustic
and 1.84 for sulfuric acid.
V. Reagents
1. Nitric acid-sulfuric acid, 1:1 mixture. Slowly
99
-------
add 250 ml concentrated H2SOij to 250 ml HN03
with constant stirring. Allow to cool before
using; store in glass container. Caution: Wear
safety glasses and gloves at all times during
preparation of acid solution.
2. Nitric aeid-sulfuric acid, 1:2 mixture. Follow
procedure above using 150 ml concentrated HNO^
and 300 ml concentrated H2SC>4.
3. Potassium permanganate, 50 gpl. Weigh 50 g re-
agent grade KMnOjj into a 150 ml tall-form beaker.
Add approximately 70 ml distilled water and stir
for about 20 seconds. Allow the KMnOjj crystals
to settle, and decant the supernatant liquid into
a one liter volumetric flask. Repeat the oper-
ations of dissolving and decanting until all the
KMnOjj has dissolved. Dilute to volume, mix, and
store in a brown bottle in a dark place.
4. Potassium persulfate, 50 gpl. Dissolve 20 g
in 400 ml distilled water.
5. Hydroxylamine hydrochloride , 100 gpl. Dissolve
40 g NH2OH-HC1 in 400 ml distilled water.
6. Stannous chloride, 100 gpl. Dissolve 20 g SnCl2-
2H20 in 20 ml concentrated HC1 on the hot plate.
Cool and add 180 ml distilled water. Prepare
weekly or more often as needed. If solution
becomes discolored, cloudy, or turns the sample
solution cloudy upon addition (prior to aeration),
discard and prepare a fresh solution.
7. Stock Hg solution, 1000 ppm. Dissolve 0.6768 g
mercuric chloride (HgCl2) in a 500 ml volumetric
flask. Add 5 ml concentrated HNOo and dilute to
mark with distilled water.
8. Working Hg standard, 10 ppm. Dilute 5.0 ml 1000
ppm Hg to 500 ml with distilled water plus 5 ml
HNOo. Prepare bimonthly.
9. Working Hg standard, 0.5 ppm. Dilute 25.0 ml
10 ppm Hg to 500 ml with distilled water plus
10 ml HNOo. Prepare monthly.
VI. References
1. William Horwitz, Ed. "Official Methods of Anal-
100
-------
ysis of the Association of Official Agriculture
Chemists," 9th edition, Association of Official
Agriculture Chemists, Washington, D.C., I960,
pp. 327-330.
2. Federal Water Quality Administration, Provisional
P.W.Q.A. Method for Hg Determination by Flameless
AA, 1970.
3. Dow Chemical Company, Determination of Mercury
by Atomic Absorption Spectrophotometric Method,
1970.
101
-------
APPENDIX C
ALTERNATE Hg RECOVERY METHODS
Reduction Methods
A method much discussed in the "literature, and in
commercial operation, is the reduction of the mercuric ion
to the metallic state followed by physical removal of the
Hg particle by filtration (2, 25, 3^, 42). Diverse
materials may be used to perform this reduction but all
rely using a suitable reducing agent. Some of the chemicals
proposed or used are:
1. Hydrazine hydrate
2. Aldehydes
3. Sodium borohydride
4. Sodium amalgam
5. Metals: zinc, iron, bismuth, tin, nickel,
magnesium, manganese, copper,
aluminum, tin chloride.
The Ventron process utilizes sodium borohydride
as the reducing agent. This process was installed at the
Sobin Chlor-Alkali Plant in Orrington., Maine, and at the
Ventron Plant in Wood^-Ridge, New Jersey. A. 99-5/6 Hg removal
efficienty was reported. In lab tests, we were not able to
achieve Hg removals as great (Table 19). The differences
may be explained by varying conditions between our tests
and Ventronfs or by the difference between our waste water
and the Sobin waste water.
In any case, the equipment required is similar
to that needed for sulfide precipitation: a pH adjustment
system, a reducing chemical addition, and a filtration
step with or without filter aid (Figure 10).
The main advantage of the Ventron process is
that Hg can be recovered in the metallic state and reused
without further processing. However, to achieve the
99>5% recovery as claimed, the reduction step must be
followed by a carbon bed and a resin bed for polishing.
In our laboratory tests, the reduction step using sodium
borohydride alone produced recoveries in the 95 - 98?
range (Figure 19). With these efficiencies, the reduction
102
-------
and filtration process could be used alone in plants
producing 100 - 200 tons per day of chlorine but large
plants would have to add the polishing step (Table 27) .
The cost of the sodium borohydride is about
$l6.50/kg ($7. 50/lb). Excess addition of chemicals or con-
centrations of other ions which consume NaBH^ could create
high operating costs. Theoretically, one kg of NaBHjj could
reduce up to 21 kg of Hg if no intefering substances are
present. However, any oxidizing chemicals such as avail-
able chlorine or metal ions capable of being reduced would
consume
Other reduction methods tried successfully have
involved a number of chemicals. One of the most common is
zinc. In work performed at Merck, Sharp and Dohme by
Rickard and Brookman (^0), a 99% Hg removal was reported
using a dosage level of 3.8 kg zinc per kg of Hg. In
our laboratory work, we have achieved recoveries of 95 -
99.8/E using zinc particles in a column, followed by filtra^
tion. To separate the Hg from the zinc, a distillation step
is required, in common with most other methods of Hg pre-
cipitation or adsorption. An additional problem associated
with this method is residual dissolved zinc in the effluent,
ranging from a few to a few hundred ppm zinc depending on
the pH of the effluent. The background level of zinc in
seawater is 0.01 ppm, and as with other heavy metals, bio-
logical concentration has been reported up to 1500 ppm.
Therefore, if the zinc process is to be used, some method
of zinc ion removal would be required. Such a process
would add to the cost and complexity of the system;
therefore, no further studies are contemplated on zinc
treatment systems .
Many other metals have been tried with results
similar to those reported for zinc, but the toxicity prob-
lem of dissolved metal ions is present to varying extents
for each alternative. Of the least toxic metals tried,
such as magnesium and iron, the cost of the metal is high
or its effectiveness low (Table 28).
Laboratory studies are reported with other
reducing chemicals such as hydrazine hydrate, aldehydes
and others. Although we have not studed these, we believe
the same problems and advantages hold as for sodium
borohydride.
103
-------
Table 27. Hg REMOVAL RATES NECESSARY FOR VARIOUS SIZE
CHLORINE PLANTS TO ACHIEVE 45 gm (0. I Ib) PER
DAY MERCURY IN THE EFFLUENT
Plant size
Cl2/day
m ton
90
180
360
740
1450
s ton
100
200
400
800
1600
Estimated1 Hg
contaminated
water volume.
I/day gpd
75,000 20.000
150. 000 40. 000
300. 000 80. 000
600.000 160.000
1.200.000 320.000
Calculated2
final Hg
level.
ppb
600
300
150
75
38
Reduction^
through
treatment.
%
94.0
97.0
98.5
99.3
99.6
Volume estimated on the basis of 75. 000 I/day (20, 000 gpd^ of
contaminated waste water per 90 m ton (100 s ton) chlorine production
per day.
2
Maximum effluent concentration to achieve level of 45 gm/day (0. 1 Ib/day)
Hg in effluent.
3
Assuming the starting Hg level in waste water was 10 ppm Hg.
104
-------
Table 28. COMPARISON OF SUBSTANCES USED OR CONSIDERED
FOR REDUCING MERCURY ION IN SOLUTION
Compound
or metal
Sodium borohydride
Bismuth
Tin
Nickel
Hydrazine hydrate, 85%
Magnesium
Copper
Maganese
Aluminum
Zinc
Iron
4
Sodium sulfide
Effectiveness
High
?
Low
Low
High
High
Low
High
High
High
Low
High
Cost2
$/kg
16.50
19.80
9.35
3.56
1.50
.84
1.50
.84
.68
.77
.22
.15
$/lb.
7.50
9.00
4.25
1.62
.68
.38
.68
.38
.31
.35
.10
.07
3
Toxicity
potential
--
--
Medium
High
Low
High
Low
Medium
High
Low
High
Based on Standard Oxidation Reduction Potentials.
2 From"Chemical Marketing Reporter", April 8, 1974;
"Metals Week", April 29, 1974.
3 Subjective information from Water Quality Criteria. F. W. P. C. A.,
U. S. Department of the Interior, April 1968.
Not a reducing agent.
105
-------
The use of sodium amalgam to reduce Hg in brine
or waste water has been tried by Karpink with limited
success (35, 36). The 78% reduction reported is too low.
for this system.
Ion Exchange & Che'lating "Resins
Another method for removing Hg from waste water
that appears frequently in the literature and has been
used in several plants in Japan is the use of ion exchange
or chelating resins (43 - 6l).
The literature states that starting with Hg levels
in the 2 ^- 30 ppm range, after, one stage of resin treatment,
the effluent contains 0.1 - CL.5 ppm Hg. With the addition
of a polishing resin step, the effluent can reach 0.001 to
0.020 ppm (Table 16.). Similar results have been achieved
in our laboratory tests. With a starting solution of
10 ppm, the effluents range in concentration from 0.3 -
1.8 ppm. But when the starting solution is low in Hg,
less than 0.1 ppm, which simulates a polishing step, the
final Hg levels are 1 - U ppb (Figure 19 and Table 17).
Resins tested in our laboratory work were from the
Billingsfors-Langed and AJinomoto companies and were
specifically designed for Hg removal. Of the two resins
tested, the Ajinomoto resin gave more consistent results.
Activated Carbon
Another means of removing Hg from waste water
streams is to pass the water through a bed of activated
carbon to adsorb the Hg onto the carbon particles. This
principle has been used extensively for. the removal of Hg
from caustic soda using a finely divided carbon,
such as Nuchar KD Special, as a precoat on a pressure fil-
ter. In this application, the Hg concentration is lowered
from 2000 ppb down to 100 ppb.
Although the literature contains fewer references
to work with activated carbon than resin, the experience in
our laboratory indicates that activated carbon achieves
nearly the same Hg removal rates as ion exchange resins
(Figure 19). Effluent levels of 100 - 300 ppb Hg were
achieved with starting Hg levels >10 ppm. However, with
starting Hg levels below 0.1 ppm, the effluent contained
5-7 ppb. Of the 3 carbons tried, the Westvaco Nuchar 722
gave the lowest Hg levels in the effluent (Table 18). The
bed capacities of the carbons were not determined.
106
-------
As with, the ion exchange resin, there are several
problems which must be considered with such a system. They
include: CD periodic regeneration or replacement of carbon,
(2) Hg recovery process from regenerant or spent carbon,
(3) prefiltration of the treated stream to minimize bed
plugging, and (4) determination of bed capacity and Hg
leakage point.
Thus, the use of activated carbon for Hg re-
moval seems more appropriate as a secondary polishing
step rather than a primary process.
107
-------
APPENDIX D
START-UP MANUAL
SLUDGE TREATMENT SYSTEM
Start-up
Note: Start-up requires a controlled sequence
to have each piece of equipment ready
when needed. It :takes 8 hours after the
starting the sludge pump to the thickener
before the rotary vacuum filter and.furnace
will have to handle product. The furnace
takes 48 hours to preheat to operating
temperatures, so plan your time accord-
ingly.
I. Sludge Dewatering (Assume brine clarlfier is in opera-
tion. )
A. Brine Sludge Thickener
1. Close drain valve on thickener.
2. After determining there are no potential ob-
structions, start the rake on the thickener.
3. Open manual valves before and after sludge
pumps.
4. Start pump by adjusting air valve on pump.
Adjust valve on oil reservoir.
B. Gas Cooler Condensers
1. Open all manual valves so the gas can
pass through bodies 1, 2, and 3 from top to
bottom.
2. Start cooling waiter to each body.
3. Start induced draft fan.
C. BSP Envirotech Furnace
Upstairs Control Room
108
-------
1. Turn on master control switch, at the remote
station.
2. Turn all burner control swi'tchs to "on"
positions at the remote station.
3. Manually adjust controller valves to "0"
supply.
Downstairs at Furnace
4. Start shaft rotation.
5. Start shaft cooling fan.
6. Start combustion air fan.
7. Push reset switch, (indicating shaft rotation
has been reset).
8. Turn master gas control switch to "automatic"
Purge timer light should come on; timer is
set for 5 minutes.
9. Reset low and high pressure gas meters.
When purge complete light comes on, proceed
to next step.
10. Open manual gas valve.
11. Start the burner on low fire on No. 6
hearth.
12. Adjust temperature controller to 400 - 500°F.
When stabilized, adjust controller upwards
slowly (about 100F° per hour) until No. 6
hearth has a temperature of 10QO°F.
13. Start burners on No. 5 hearth.
14. Adjust temperature controller to 400 - 50Q°F.
When stabilized, adjust controller upwards
slowly (about .10OF0 per hour) until No. 5
hearth has a temperature of 1QOO°F.
15. Start main burners on No. 4 hearth.
109
-------
16. Adjust .temperature controller to 400 - 500°F.
When stabilized., adjust controller upwards
slowly (about 100F° per hour) until No. 4
hearth has a temperature of 1000°P.
17. Start burners on No... 3 hearth.
18. Adjust temperature controller to 400 - 5000F,
When stabilized, adjust icontroller upward
slowly (about lOOF0 per hour) until No. 3
hearth has a temperature of 1000°F.
19. Start increasing temperature on all four
hearths at the rate of 5QF° per hour.
Operating temperature is between 1400 -
1500°F.
Upon reaching operating temperature of 1400 -
1500°P, prepare rotary vacuum filter for
operation.
D. Eimco Rotary Vacuum Filter
1. Close the filter vat drain valve.
2. Open the wash water line to the cloth and
rolls.
3. Begin taking up the slack in the filter
belt, being sure to adjust the ends of the
takeup rolls equally.
4. Start the filter drum drive and completely
soak the cloth, while retensioning the
belt. When the takeup roll is at normal
operating position, stop the filter drive.
5. Start the vat agitator. This should be op-
erated at all times when the sludge is in
the vat.
6. Turn on the seal water to. the vacuum pump
and filtrate pump.
7. Start discharge conveyor from furnace.
8. Start furnace feed conveyor.
9. Turn on feed to filter.
110
-------
10. When sludge level in the vat reaches 30 #
full, start the vacuum pump, filtrate pump,
and filter drive.
11. Adjust the filter drum speed and rate of
feed as may be necessary for cake thick-
ness, cake dryness, and removal from cloth.
It may take *IO - 60 minutes before sludge begins
coming out the discharge conveyor. Check clinker
for dryne&s and plugging of furnace.
SHUTDOWN
I. Sludge Dewatering
A. Brine Sludge Thickener
1. Shut manual valve from bottom of clarifier
before pump. Flush fresh water through
pump inlet and outlet to thickener.
2. After clear water appears at thickener.
shut off sludge pump and drain water from
line.
3. Continue to dewater sludge until consistency
is too low for good filter performance. Di-
vert the rest of the sludge to pond by shut-
ting off filter feed and opening line to pond,
B. Eimco Rotary Vacuum Filter
1. Fully open all wash water to the filter belt.
2. Open vat drain valve. Flow will divert to
pond.
3. Stop filtrate pump and vacuum pump, and shut
off seal water.
4. Rotate drum drive at least 5 revolutions
until filter cloth is clean. Then stop
filter drive and shut, off wash water.
Note: Never leave a cloth to dry unless it
is washed thoroughly.
5. Release the tension on the filter cloth by
111
-------
turning the takeup roll cranks equally.
Note the numfrer of turns of the adjusting'
cranks so that the takeup roll can be re-
turned to its. original position at start-
up.
6. Plush the wash trough with a small amount of
fresh water. If shutdown will be longer
than 8 hours, flush the filtrate tank and
wash down the filter.
Note: It will take 2 hours after the belt
conveyor has delivered the last little
bit of solids into the furnace, before
the last clinker is discharged by the
clinker conveyor. When the clinker
conveyor is empty, begin shutting down
the furnace.
BSP Envirotech Furnace
1. Start reducing the temperatures on all four
hearths at the rate of 50°P per hour. When
temperatures have stabilized at 900 - 1000°F
continue to next step.
2. Adjust the temperature controller on No. 3
hearth so the temperature drops at the
rate of 100°P per hour. When temperature
stabilizes at 400 - 500°F turn off burner
on No. 3 hearth.
3. Adjust the temperature controller on No. 4
hearth so the temperature drops at
the rate of 100°P per hour. When tempera-
ture stabilizes at ^00 - 500°F turn off
burner on No. 4 hearth.
4. Adjust the temperature controller on No. 5
hearth so the temperature drops at
the rate of 100°F per hour. When tempera-
ture stabilizes at 400 - 50Q°F turn off
burner on No.5 hearth.
5. Adjust the temperature controller on No. 6
hearth so the temperature drops at
the rate of 1QO°F per hour. When tempera-
ture stabilizes at ^00. - 500°F turn off
burner on No. 6 hearth.
112
-------
Furnace temperature should be between 300 -
liOO0?. A decision at that time will be made
whether to shut off the pilot or not. It
is advantageous to keep the furnace at this
temperature if .possible.
D. Gas Cooler Condenser
As long as furnace is running on pilots or burn-
ers, the cooler condensers will remain in oper-
ation.
WATER TREATMENT
Start-up
I. Water Treatment
A. Acid Mix Tank
1. Open valve on the inlet to Hg waste water
storage tank.
2. Open recycle valves to pH mix tank.
3. Open valve from spent acid stream.
4. Adjust level controller.
5. Open valves on inlet and outlet of the pond
pump to mix tank.
6. Start flow from pond.
7. Adjust pH controller and start acid flow.
Liquid in tank will continue to recycle
until appropriate level is reached.
8. Start the agitator in the Hg waste water
storage tank when the level is above the
agitator.
B. Na2S Storage Tank
1. Open valves on the outlet of the
storage tank.
2. Adjust rotometer to required gph.
113
-------
C. R. P. Adams Pressure Filter
1. Open accept valve to sewer.
2. Open valve for filter feed.
3. Make sure bottom valve to backwash, tank is
closed.
D. Filter Aid and Precoat Mix Tank
1. Open valves on outlet and inlet of feed
pump to filter.
2. Adjust level controller in mix tank.
3. Adjust BIF feeder to the rate of 1 oz/100 gal.
4. Adjust flow indicator from Hg waste water
storage to mix tank. Remember, the flow
should agree with the setting on the Na2S
addition system.
5. Start flow from Hg waste water storage tank.
6. Adjust Na2$ flow on rotometer.
7. Start BIF feeder. initial start-up requires
50 Ib. of diatomaceous earth.
8. Start agitator.
Note: When level has reached the controller set
point, it will begin feeding filter at the rate
you set at the flow indicator from the Hg waste
water storage tank.
WATER TREATMENT
Shut down
I. Water Treatment
A. Acid Mix Tank
1. Shut off pond pump to mix tank.
2. Shut off other flows to mix tank.
114
-------
3. Shut off acid flow.
Note: If shutdown is only temporary, you can
pump for a short time to: the waste water storage
tank. Be sure to treat with acid first.
B. Waste ¥ater Storage Tank
1. When level is- bellow agitator, shut off
agitator.
2. Close valve out. of tank.
Note.: If waste water is to be stored in tank,
leave on agitator.
C. Filter Aid and Freeoat Mix Tank
1. Turn off BIP feeder.
2. Turn off rotometer from Na2S storage tank.
3. Adjust level controller such that you can
pump the remainder of tank through the fil-
ter. Then turn off pump.
4. Add fresh water and flush lines and pumps.
D. R. P. Adams Pressure Filter
1. Main objective now is to backwash filter:
a. Close the feed valve from precoat mix
tank.
b. Open dump valve to backwash tank. Fil-
ter will drain to decant tank and ma-
jority of precoat should fall off by
reverse flow.
c. Refill tank 3/4 full with fresh water.
Turn off water flow.
2. Pressurize filter with about 50 psi, then
turn off air.
3. Open backwash valve quickly. The compressed
air head should push the liquid in the
reverse direction, thoroughly purging the
filter tubes of any remaining cake.
115
-------
Note: The receiver dump will be full of water
and filter aid. Decant off water and open dump
receiver to filter. Time required will be de-
termined after initial sfrart^up.
116
-------
OPERATING NOTES
I. Sludge Dewatering
A. Thickener
1. A low level alarm on the thickener may in-
dicate an Insufficient pumping rate. If
problem cannot be quickly- corrected, notify
the Tour Foreman. Note time period of
trouble in log book.
2. There is a screen on top of the thickener
to prevent larger size particles from enter-
ing the thickener. This should be cleared
once a shift.
B. Rotary Vacuum Filter
1. In general, vacuum will be kept at a maximum
and not varied in order to achieve maximum
dryness.
2. Vat level and drum speed determine cake thick-
ness and production rate. Level will gener-
ally- be kept cons-tant and speed varied.
High speeds will tend to decrease dryness.
3. If filter will not pick up cake:
a. vat consistency may be too high
b. vat level may- be too low
c. vacuum may be too low
d. the cloth may not be getting cleaned
properly
4. In order to obtain dryer cake:
a. slow down the filter
b. decrease cake thickness (lower at vat
level)
c. increase vacuum
117
-------
5. The vacuum pump can be severely damaged if
it is operated" without .seal water. Thus,
the seal water should be adjusted or checked
every 4-6 hours.
6. If the filtrate is not removed from the re-
ceiver, it will carry over to the vacuum
pump. Check to. see that the valve on the
filtrate pump is wide open at all times.
7- Check cloth appearance frequently. Improper
slack in cloth or misalignment can cause lack
of vacuum or tearing of cloth. It is very
important to correct these problems immedi-
ately.
C. BSP Envirotech Furnace
1. Monitor temperatures on all four hearths.
It is important that we maintain an oper-
ating range of 1*100 - 1500°F. Problems of
insufficient temperature could be:
a. combustion air fan
b. improper gas to air ratio
2. Check shaft cooling fan regularly. This
is vital for good operation of furnace.
3. It is important that we maintain an even
flow to and from the furnace. If feed rate
is too fast, plugging of upper hearth can
and will be a problem. If this does hap-
pen, discontinue feed to furnace until in-
cinerated solids conveyor is empty - pos-
sibly 4 hours - then continue operation.
D. Cooler Condensers
1. It is important that we receive the maximum
amount of cooling from each condenser. If
scale begins building up within the condens-
ers, poor heat transfer will result in higher
air temperature at the induced draft fan.
These temperatures will be monitored each
shift until a temperature range for opera-
tion is established.
118
-------
2. Check Induced draft fan frequently. Note
excessive vibrations, or other problems which.
might develop.
II. Liquid Treatment
A. pH Adjustment
1. If pH becomes a problem, check the following:
a. acid feed pump
b. acid fitters
c. automatic control valve
d. make sure agitator is operating
Note: pH should be maintained between 5-8.
As you approach the higher pH's, Hg becomes more
soluble and tends to pass through the filter
media more easily. Also, lack of pH control tends
to disrupt sewer pH.
B. Sulfide Precipitation
1. Excess sulfide is needed to precipitate Hg
in the mix tank. Therefore, It is important
that we maintain a proper flow and the right
concentration of Na?S to the mix tank. Check
these often.
C. R. P. Adams Filter
1. It is important that the tubes within the
filter are properly precoated. Improper
precoating can cause tubes to plug and
eventually break when backwashing.
119
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
1. Report No.
2.
w
4, Title
MERCURY RECOVERY FROM CONTAMINATED WASTE
WATER AND SLUDGES
7. Author(s)
Richard Perry
Georgia-Pacific Corporation, Bellingham Division
5. Report Bat*,
6.
8 Performing Organization
Ri-"cr! Ma.
12040 HDU
12,
v p? of Report and
;',:dod Covered
organization EPA, Water Quality Office
_.___._______:--^—. was Designed, built and operated to remove Hg from
waste water and sludge produced by a mercury cell chlor-alkali plant. Mercury content
of the waste water ranged from 300 - 18,000 ppb mercury while mercury content of the brii
fr-nm 150 to 15QQ ppm Hg. Other sludges processed include sludges from
near our plant outfall with a Hg content of 10 - 25 ppm Hg.
From a variety of removal techniques tried in the lab, the methods selected were
sulfide precipitation for the water treatment and high temperature roasting for the sludj
treatment. The sulfide precipitation consists of collecting the various water streams,
adjusting the pH from 5 - 8 with spent sulfuric acid, settling the large solid particles
in a surge tank, adding sodium sulfide to a 1-3 ppm excess, adding diatomaceous earth at
the rate of 0.07 gpl in an R. P. Adams pressure filter. The effluent Hg levels range
from 10-125 ppb with an avesage of 50 ppb Hg for an 87-99% removal, averaging 96.8%.
The 4.8m3 filter handles 280-280 liters/min adequately with an approximate 48 hour cycle
time between backwashings. Capital costs were $143,900 and operating costs were 50C/378.
The sludge system contains a collection system, 3.7 m diameter thickener, 1.8 m
diameter rotary vacuum filter, 1.37 m i.d. multiple hearth furnace, and 3 stainless stee
condensers 21 m2 each. Processing rate for the sludge is 140-320 kg/hr, dry basis.
At present we are processing approximately 18 m tons of sludge per month for our Chlor-
Alkali Plant. Operating temperatures ranged from 540 C - 760 C, feed Hg content ranged
from 290 to 440 ppm Hg (dry basis), and clinker Hg content after treatment contained
0.5 - 7.2 ppm Hg, for a removal rate of 98.3 to 99.8%. Waterway sediments containing
12.8 ppm were roasted at 730 C and clinker contained 0.95 - 1.7 ppm Hg for an 87-92%
removal. Capital costs were $364,500 and operating costs were $32 per m ton of dry
ol 11/I c
Ja. Descriptors Wflter Pollutionj Metals, Sludge Disposal, Waste Water Treatment,
Sulfides, Chemical Precipitation, Hydrogen Sulfide, Filtration, Electrolysis,
Activated Carbon, Ion Exchange, Reduction (chemical) Particle Size, Oxidation,
Ghlorination
i7b. Identilfer*
Mercury, Chlor-alkali Cells, Separation, Recovering, Thickening, Clarificati
Roasting, Sodium Hypochlorite, Methyl Mercury, Sodium Borohydride, Vacuum Filtration
'$, S?*UTJVy Class,
. Security Class
(Page)
21. No', of
Pages
22. Price
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CCNTKR
U.S. DEPARTMENT Of THE INTERIOR
WASHINGTON. OjC. 20*40
U. S. GOVERNMENT PRINTING OFFICE: 1974-697-649 /6I REGION 10
------- |