WATER POLLUTION CONTROL RESEARCH SERIES • 12010 DUL 02/71
Limestone Treatment of Rinse Waters
from
Hydrochloric Acid Pickling of Steel
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollu-
tion of our Nation's waters. They provide a central source
of information on the research, development, and demon-
stration activities of the Water Quality Office, Environ-
mental Protection Agency, through inhouse research and grants
and contracts with Federal, State, and local agencies, re-
search institutions, and industrial organizations.
Inquiries pertaining to the Water Pollution Control Research
Reports should be directed to the Head, Project Reports
System, Office of Research and Development, Water Quality
Office, Environmental Protection Agency, Washington, B.C. 20242.
-------
LIMESTONE TREATMENT OF RINSE WATERS
FROM HYDROCHLORIC ACID PICKLING OF STEEL
Arraco Steel Corporation
703 Curtis Street
Middletown, Ohio 45042
for the
Water Quality Office
Environmental Protection Agency
PROJECT 12010 DUL
February, 1971
-------
EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily 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.
For sale by the Superintendent ol Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price $1.60
-------
ABSTRACT
Two hydrochloric acid picklers for cleaning steel strip
at Armco Steel Corporations Middletown, Ohio, Works
produce up to 1,500 gpm of acid rinse waters which con-
tain up to 0.5 g/1 free hydrochloric acid and up to
O.g? g/1 ferrous chloride. A facility for disposal of
these rinse waters was designed, based on a process
developed at bench scale by Armco research scientists.
This process utilizes limestone for neutralization plus
aeration and sludge recirculation to oxidize ferrous
iron and form soluble calcium chloride. This report
describes the investigation of process variables at
pilot scale and the optimization and demonstration of
the process at full scale.
Pilot studies indicated that a 25 percent excess of
limestone was normally required, sludge recirculation
was very beneficial, and temperatures above 120°F led
to higher sludge filtration rates and system capacities.
Operation at lower temperatures required larger lime-
stone excesses, higher aeration rates, higher sludge
recirculation rates, and limestone of smaller particle
sizes.
The^full scale facility provided 100 percent neutrali-
zation of free acid and over 99 percent removal of iron
using a 50 percent excess of limestone. A very dense,
easily filtered sludge was produced. Although influent
temperatures as low as 59°F were encountered, game fish
populations were maintained in the treated water.
Capital costs for a facility to treat 1,500 gpm acid
rinse water were $1,360,000. Operating costs were
24.00/1,000 gal. or 4.38#/ton of steel pickled.
This report was submitted in fulfillment of Project
12010 DUL under the partial sponsorship of the
Environmental Protection Agency.
Key Words: Acid rinse water, aeration, clarification,
ferric hydroxide, ferrous iron, flocculation,
limestone, neutralization, pickle rinse
water, sludge recirculation.
-------
CONTENTS
Section
I
II
III
IV
V
VI
VII
VIII
IX
X
Conclusions
Recommendations
Introduction
Pilot Studies
Field Studies
Equipment Design Evaluation.
Acknowledgments
References
Glossary
Appendixes
Paee
v
vii
1
2
10
21
27
23
29
31
ii
-------
FIGURES
No. Page
Schematic Flow Diagram of Pilot
Plant
II Effect of pH on Ferrous Iron
Concentration in Flocculation
Tank 13
III Percent of Period Ferrous Iron
Concentration in Flocculation
Tank 3 mg/1 or Less vs. pH 14
IV Total Iron in Effluent and Flow... 22
iii
-------
TABLES
Page
I Results of Pilot Scale Studies ..... 6
II Capacity of Plant Under Various
Conditions ........................ 11
III Typical Treatment Plant Operating
Results (Daily Average Data) ...... 16
IV Typical Treatment Plant Operating
Data (April 1970) ................. 19
V Typical Treatment Plant Operating
Costs ........... . ................. 20
VI Treatment Plant Capital Costs ...... 20
iv
-------
SECTION I
CONCLUSIONS
This process clearly demonstrated an ability to
effectively, efficiently, and economically treat
large volumes of hydrochloric acid rinse waters.
The conclusions stated in this section have been
derived from pilot scale and full scale studies.
1. The treatment facility described in this report
provided 100 percent neutralization of free acid
and over 99 percent removal of iron from the
acid rinse waters. Temperatures were normally
above 95 °F (35°C). Temperatures as low as 59°F
were encountered briefly; while the process was
still effective, the reaction rate was appreci-
ably reduced. Typical data are found in Tables
III and IV (p. 13, 19).
2. The process produced a very dense, easily filtered
sludge. The sludge was typically filtered at a
rate of SO Ib wet cake/sq ft/hr with a moisture
content of 30 percent and a cake thickness of
about 3/4 in. (p. 24).
3. With 90,000 gal. surge capacity ahead of the
aeration-mix tank, the treatment facility was
operated satisfactorily by one man using manual
control; however, automatic control would
probably improve efficiency substantially by
precisely regulating limestone dosage and might
reduce the surge capacity requirements (p. 21).
4. Typical treatment costs were 24.0^/1,000 gal. or
4*3s£/ton of steel pickled (p. 19]
5. The facility was designed to handle 6,900 Ib
iron/day. The normal load, however, was only
about 2,500 Ib/day. The plant capacity was
determined experimentally to be in the range
of 7,000 to 9,000 Ib/day, depending on temper-
ature, when using one 250 SCFM blower. Using
two blowers, the capacity was increased about
25 to 50 percent (p. 11).
6. Game fish populations were maintained in the
treated effluent water (p. 25).
-------
7. A minimum amount of recirculated sludge was needed
to improve floe quality. Exceeding this minimum
amount of sludge did not affect floe quality
(p. 16).
8. Practically all the iron in the acid rinse water
was present in the ferrous form (p. 9). This led
to higher filtration rates than anticipated, and
the filter capacity provided was more than required
(p. 24).
9. A minimum pH of 6.6 in the effluent from the
aeration-mix tank was required for complete
oxidation of the iron; this is a prime indicator
of proper plant operation (p. 12).
10. The sludge was too heavy to be handled with
conventional A-frame flocculators. This problem
was solved by substitution of turbine-type
mixers (p. 23).
11. Higher filtration rates were achieved at higher
temperatures (p. 8).
12. The amount of air required for oxidation decreased
noticeably with increasing temperature (p. 7).
13. The aeration-mix tank was adequately sized. Both
the temperature and the flow rate were less than
anticipated. The 45 to 50 min retention time
available was sufficient at the temperatures
experienced; pilot studies indicated 30 min to be
sufficient only above 120°F (p. 15).
14. The flocculation tank was designed for 30 min
retention time at 1,500 gpm. This appeared to be
excessive, since the floe formed quickly and was
very dense (p. 23)•
15. The clarifier was designed with a rise rate of
0.3 gpm/sq ft at 1,500 gpm. This proved to be of
sufficient size (p. 23).
VI
-------
SECTION II
RECOMMENDATIONS
1. The aeration-mix tank should be sized with the
operating temperature in mind. Additional
retention time is required at lower temperatures.
2. A flocculator - clarifier should be evaluated as
a substitute for separate flocculation and
clarification.
3. If a separate flocculation tank is used, variable-
speed turbine-type flocculators are recommended.
4. Automatic monitoring of the plant influent for
total acid is recommended. Automatic control of
limestone addition should be considered. Manual
monitoring and control require sufficient surge
capacity ahead of the aeration-mix tank to pre-
vent upset of the plant between tests. Automatic
control of flow rate might also be considered.
5. Vacuum filter capacity should be based on the
ferric/ferrous ratio of the iron in the acid
rinse water. The presence of ferric iron in the
waste substantially reduces the filtration rate
of the sludge.
6. The clarifier effluent normally had a pH of about
6.9 and contained less than 5 rag/1 total iron.
However, a small lagoon is recommended as a back-
up for the facility.
vii
-------
SECTION III
INTRODUCTION
Ferrous metals develop an oxide coating in the course
of manufacture. This coating is customarily removed
from low carbon steels by pickling in acid and rinsing
with water. Two different waste streams result:
spent acid from the pickling baths, and once-through
rinse water. Two new hydrochloric acid pickling lines
were constructed at Armco Steel Corporation's Middle-
town, Ohio, Works. Disposal of the spent pickling
acid was provided, but a facility to dispose of the
acid rinse water was needed.
Armco research scientists had developed at bench scale
a hydrochloric acid pickling rinse water treatment
process that was believed to be economical and con-
trollable. This process is unique in that it utilizes
limestone neutralization along with aeration and
sludge recirculation to oxidize ferrous iron and
produce a readily filterable precipitate. The
reactions for this process are:
2HC1 + CaCO} -» CaCl2 + H20 + C02
4FeCl2 + 4 CaC03 + 6H20 + 02 -* 4CaCl2 + 4Fe(OH)3 + 4C02
A facility for this type process was needed immediately,
so plant design was based only on bench level informa-
tion. Several of the design criteria were based on
experience gained in the design of previous similar
structures. Pilot scale tests were carried out
concurrently with construction so that, upon completion
of the facility, the pilot level information would be
available for use at the treatment facility.
On May 16, 196&, Armco Steel Corporation received an
Environmental Protection Agency grant to: develop at
pilot scale the process to convert large volumes of
hydrochloric acid rinse waters into good quality water
and an easily handled sludge; design and construct a
full scale installation to demonstrate the process
developed in the laboratory; develop and determine a
simple, economical, efficient, and controllable method
of operation of the process; and, demonstrate the
efficiency, economy, and controllability of the
process.
-------
SECTION IV
PILOT STUDIES
EXPERIMENTAL
The pilot plant was designed for a maximum flow rate of
0.5 gpm. A schematic diagram of the plant is shown in
Fig. I. A brief description of equipment and operating
practice follows:
Acid Rinse Water
The acid rinse water (ARW) to be treated was programmed
to contain about 1.0 g/1 of total hydrochloric acid (0.5
g/1 of hydrochloric acid plus O.S? g/1 of ferrous
chloride-1-). The decision was made to prepare a more
concentrated solution of ARW and meter it into the pilot
facility. With this approach, reaction temperatures
could be controlled more easily by dilution of the ARW
stock solution with heated tap water.
A stock solution of ARW was synthesized by dissolving a
specified weight of powdered iron in a known amount of
hydrochloric acid. The solution was then diluted to 55
gal. with water, and samples were taken to be analyzed
for total acid and ferrous iron.2
The ARW stock solution Was metered into the system by
means of a variable-flow tubing pump which could be
precisely controlled to deliver 0 to 940 ral/min.
Limestone Slurry
Prior to a run, 20 gal. of limestone slurry were pre-
pared. Limestone concentration in the slurry was
determined by oven-drying a sample overnight. The value
1 0.3? g/1 of ferrous chloride is equivalent to 0.5 g/1
of hydrochloric acid.
2 Samples of this solution were analyzed for both ferrous
and ferric iron during the first experimental runs.
Despite exposure to the atmosphere over a period of
several weeks at room temperature, ferrous chloride
remained unoxidized.
-------
AIR
LIMESTONE SLURRY
ACID RINSE WATER
HOT TAP WATER
SLUHRY
MIX TANK
t
AERATION
TANK
SLUDGE RETURN
POLYMER
0 D
f
f -*
T
\— -~\
1 -r
t- 1
\ — J-> — -J V •*— •• 1
L
\
FLOCCULATION
TANK
EFFLUENT
CLAEIFIER
SLUDGE
HOLDING
TANK
FIGURE I.
SCHEMATIC OF PILOT PLANT
-------
obtained determined the rate at which the slurry was
pumped. The slurry was metered into the process by
means of a variable-flow tubing pump, also having 0 to
940 ml/min capability.
Air
Air was supplied to the process by means of a 3 SCFM
blower. Air flow rate was indicated by one of two
rotameters having ranges of 0 to 3 and 0 to 13 1/min.
Equipment
The pilot plant consisted of five tanks. The first tank,
the slurry mix tank, was used to blend the ARW stock
solution and heated tap water with the limestone slurry.
Retention time in this tank was normally 30 sec, but a
90-sec retention time was also possible.
The ARW-limestone-water mixture flowed from the slurry
mix tank into the aeration-mix tank. This tank was 22
in. in diameter and was baffled with four steel plates
(each 12 in. high by 2 in. wide) set against the out-
side wall of the tank 90 deg apart. The reaction
mixture was agitated by a 4-in. diameter flat-blade
turbine set 2 in. above the center of the bottom of the
tank. The turbine had six blades and was powered by a
1/4 hp variable-speed (0 to 350 rpm) motor. Normal
operating speed was 350 rpm. Air was introduced into
the tank at a point directly below the turbine.
Retention time of the tank was 30 rain.
From the aeration-mix tank, the treated ARW flowed by
gravity into the flocculation tank. An organic floccu-
lating aid was added at this point. The tank itself
was divided into two chambers, each having a 15-min
retention time. The contents of each chamber were
mixed with a 4 by li in. paddle. Tip speeds were 6
ft/sec in the first chamber and 2 ft/sec in the second
chamber.
The reaction products passed from the flocculation tank
to the center well of the clarifier by means of an over-
flow weir and a siphon which was secured 3 in. below
the liquid level in the second chamber of the floccu-
lation tank.
The clarifier was a cylindrical polyraethylmethacrylate
tank with an effective height of 4 ft and an inside
-------
diameter of 9§ in. The center well was a polymethylmetha-
crylate tube 2 ft long and 4 in. I.D. At a flow rate of
0.5 gpm, the clarifier had a rise rate of 0.155 ft/min
(1.15 gpm/sq ft). The clarified effluent overflowed a
peripheral weir and was directed to a drain. Settled
sludge was removed from the bottom of the clarifier daily.
Part of this sludge was stored in a holding tank from
which it was recirculated back into the aeration-mix tank
with a variable-flow tubing pump.
Operation
At the full scale facility, the variables which are con-
trollable are type and dosage of limestone and the sludge
recirculation rate. Variables such as flow rate,
temperature, and acid concentration depend upon pickler
operation.
At the time the pilot scale studies were commenced, ARW
temperatures in the range 150° to 1?0°F were expected.
No bench scale work had been carried out at temperatures
above 120°F, and temperature effects had been explored
only briefly. Therefore, investigation of temperature
effects was incorporated into the pilot scale studies.
Significant differences in filtration rates were expected
when varying the temperature, limestone type, limestone
excess, and sludge recirculation rate. The effect of
these variables must be known to be able to establish
optimum treatment plant operating procedures. A designed
experiment was set up to evaluate these four parameters.
The schedule called for 24 runs to include the following
.conditions:
A. Temperature: 120°, 135°, and 150°F (49°, 57°, 66°C)
B. Limestone:
1. YA Grade^ - 25 and 100 percent excess of
theoretical.
2. #2 LS Grade^ - 25 and 100 percent excess of
theoretical
C. Sludge Recirculation Rate: 5 and 15 g solids/min
Table I summarizes the results of these runs.
£4 to S7% through 200 mesh; 65 to 70% through 300 mesh
99.6$ through 200 mesh; 99.0$ through 300 mesh
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
DISCUSSION
The principal functions of this waste treatment process
are to remove iron and neutralize the acid originating
in pickling rinse waters. A proposed criterion for
effluent quality is that the treated rinse waters shall
not contain more than 5 mg/1 of ferric hydroxide or
approximately 3 mg/1 of iron. Further, turbidity other
than that originating from the ferric hydroxide should
be negligible.
Limestone Dosage
The filtration tests and the completeness of the treat-
ment (as evidenced by the amount of iron in the
effluent), indicate there is little to be gained by
using a 100 percent excess of limestone instead of a
25 percent excess at the temperatures investigated.
If the concentration of total acid in the waste to be
treated varies only slightly, a 25 percent excess of
limestone should be a sufficient quantity. Highly
variable concentrations of total acid should be treated
by a somewhat larger excess to accommodate sudden acid
peaks.
Aeration Rate
Generally, as the temperature of the ARW was increased,
smaller quantities of air were required to completely
oxidize the ferrous iron.
The values given for the aeration rate actually repre-
sent about a 50 percent excess over the minimum amount
required as determined by trial and error. In no way
are they meant to show the amount of air required for
oxidation. The full scale plant should have an adequate
air supply under the same conditions.
Sludge Recirculation Rate
The return of sludge to the aeration tank is known to
decrease sludge volumes (3>6,7). The process in this
case consists of the addition of precipitation nuclei
in the form of ferric hydroxide (and excess limestone)
from the clarifier underflow to the aeration-mix tank.
-------
The ratio of precipitation sites to the amount of
precipitating material might affect particle growth.
Variation of recirculation rate is the only means of
controlling this variable at pilot plant and full
scale.
The necessity of sludge recirculation in this process
has been demonstrated on bench scale (Appendix A). The
quantity of recirculation could not be determined from
a batch operation. Therefore, one of the functions of
the pilot plant was to ascertain the effects of varying
recirculation rates. The rates 5 and 15 g sludge/rain
were selected for the designed experiments in antici-
pation that significantly different filtration test
values would result.
AS can be seen from the data in Table I, a rate of 5 g
sludge/min was a satisfactory recirculation rate in
itself, and increasing the rate to 15 g/min did not
measurably affect the filtration rate.
The filtration rate was influenced markedly by the
absence of sludge recirculation. Two one-week pilot
runs were carried out; these runs were identical, ex-
cept that during the first week no sludge was
recirculated. Conditions were: Temperature - 150°F;
limestone - 175 percent of theoretical, YA type; air -
6 I/rain; recirculation rates - 0 and 3 g/min. «.s is
shown in Table I, sludge recirculation improved the
wet filtration rate by more than 1? percent and the
dry rate by over 30 percent while lowering the cake
moisture almost 12 percent. Material balance calcu-
lations indicate that a sludge return rate of 3 g/min
would increase the solids concentration in the
aeration tank 90 percent (from 1.77 g/1 to 3.36 g/1).
Optimum sludge return rates were not determined during
the course of the pilot plant experiments. However,
observation of many pilot plant experiments suggested
the optimum recirculation rate to be less than 3 g/min.
Temperature
The comparison of filtration tests (Table I) obtained
from runs at 120°, 135°, and 150°F indicated that
higher temperatures produced higher filtration rates.
Another apparent advantage is that a smaller excess of
air is required.
-------
At temperatures above 120°F (49°C), the pilot process
satisfactorily removed iron and produced a filterable
sludge using YA grade limestone. At 120°F, the #2LS
grade limestone (finer, therefore more surface area)
was required, together with a higher limestone dosage
and/or a higher sludge recirculation rate.
Ferric/Ferrous Ratio
The full scale facility was started up several weeks
before the studies described above were completed. As
was the case with the synthetic ARW used for pilot in-
vestigations, the ferric iron concentration in the
treatment plant influent was found to be negligible.
The ferric/ferrous ratio could, then, no longer be
considered a variable in this plant and was not
investigated at pilot scale.
9
-------
SECTION V
FIELD STUDIES
A facility was designed and constructed (Appendix B)
to apply the limestone-aeration process to the rinse
waters from the hydrochloric acid pickle lines at the
Middletown, Ohio, Works. The optimization and
demonstration of the facility are described below.
Capacity of Plant
The treatment plant was programmed to handle a 1,500
gpm flow of acid rinse water (ARW) containing 1.0 g/1
total hydrochloric acid at temperatures in excess of
150°F (66°C). About half of the total acid was ex-
pected to be in the form of ferrous chloride (0.8?
g/1 ferrous chloride or 0.3# g/1 iron, or 0.5 g/1
hydrochloric acid). This gives a plant design load
of 6,900 Ib Fe/day.
Since start-up in May 1969> flow rates have seldom
exceeded 1,100 gpm. The temperature of the ARW has
never reached that which was expected; generally,
temperatures are in the range 104 to 113 °F (45 to
50°C). The quantity of iron the plant could handle
under these conditions was determined in order to
ascertain whether design modifications were necessary.
By injecting spent pickling acid into the ARW, the
concentration of ferrous iron in the ARW could be
controlled at high levels for the purpose of deter-
mining plant capacity. The temperature of the ARW
was not controllable but remained constant during
the course of any given test. Flow rates were
adjustable, but higher rates were available for a
limited time only.
The capacity of the plant was determined for various
flows and temperatures by metering spent pickling
acid into the ARW until ferrous iron was detected in
the effluent from the aeration-mix tank. The presence
of 3 mg/1 ferrous iron at this point was taken as
evidence that the capacity of the system for oxidizing
iron had been reached. A limestone dosage of 60
Ib/min was used for all tests; at no time was this
less than a 50 percent excess.
Table II presents the empirically determined plant
capacities for various flow rates, temperatures, and
aeration rates.
10
-------
Table II
CAPACITY OF PLANT UNDER VARIOUS CONDITIONS*
Oxygen Plant
Test Flow Temp Influent Air Transfer Capacity
No. (ppm) (°F) Fe,(mg/l) (SCFM) SffV (%} (Ib Fe/day)
1
2
3
4
5
6
7
;igr
470
goo
1,200
gOO
1,200
g50
350
i 1,500
113
113
113
113
113
102
102
150-170
1,700
950
goo
1,400
1,000
700
900
330
230
230
230
470
470
230
470
230
27.6
27. g
30.2
lg.9
20.2
20.6
12. g
20.5
9,600
9,100
11,500
13,400
14,400
7,100
9,200
6,900
^Limestone dosage - 60 Ib/min
As these experiments indicate, the capacity of the plant
to oxidize ferrous iron increased 30 to 45 percent with
an 11-deg increase in temperature. Doubling the
aeration rate resulted in only a 25 to 50 percent
increase in plant capacity.
Limestone Requirements
The limestone used for this process is produced by
Armcofs Piqua Quarries in Piqua, Ohio, under the
designation "YA Stone Dust". This limestone contains
g4 to g? percent calcium carbonate and 11 to 14 per-
cent magnesium carbonate. The available alkalinity
averages 55.5 percent as calcium oxide.
Screening specifications are:
99 to 99.5$ through a 100 mesh screen
g4 to $7% through 200 mesh
65 to 70% through 300 mesh
11
-------
Bench scale studies (Appendix A) had previously shown
that an excess of limestone is necessary to raise the
pH in the aeration tank sufficiently high to bring
about oxidation of ferrous iron.
Full scale studies were carried out between September
and December 1969 to determine the minimum pH at which
the process should be maintained. This was done by
monitoring the pH and ferrous iron concentration in
the flocculation tank. Ferrous iron concentrations
above 3 mg/1 at this point resulted in both a turbid
clarifier and more than 5 mg/1 total iron in the
clarifier effluent.
Figure II shows the median and maximum ferrous iron
values in the flocculation tank at various pH values
during the period. Figure III shows the percentage
of the period that the ferrous iron concentration in
the flocculation tank was 3 mg/1 or less. To deter-
mine the minimum pH to be maintained, these two
figures should be used together. At a pH value of
6.8, the ferrous iron concentration in the flocculation
tank is maintained at 3 mg/1 or less for 97 percent
of the time; if the pH were to drop to 6.6, the iron
concentration would remain satisfactory #0 percent of
the time*
Due to the variable nature of the ARW a limestone
excess of 50 percent was chosen. Using this excess,
minor increases in the acid concentration could be
easily handled. If the acid concentration suddenly
decreased without detection, the quantity of lime-
stone added above the 50 percent excess would not
adversely affect the process; the filter cake would
contain less moisture and iron.
Air Requirements
The treatment plant is equipped with two identical
blowers capable of delivering 470 SCFM of air when
combined in parallel. This quantity of air is
theoretically able to oxidize 52 lb Fe/min. The
plant design calls for the capability of oxidizing
4.7 lb Fe/min (1,500 gpm at 3&> mg/1 iron).
The effect of temperature on the air requirements for
the plant is threefold. Higher temperatures decrease
the solubility of oxygen in water, thus decreasing the
driving force in the system (the difference between
12
-------
7.5
Median Values
— — — — Maximum Values
7.0
Period 9/26 to 12/9/69
pH 6.5
6.0
5.5
0
10
20 30 40 50 60 ?0 30 90
Ferrous Iron Concentration in Flocculation Tank, mg/1
FIGURE II
EFFECT OF pH ON FERROUS IRON CONCENTRATION IN FLOCCULATION TANK
100
-------
100
90
go
70
"8 60
•H
0)
50
c 40
0)
O
£ 30
20
10
0
Period 9/26 to 12/9/69
6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0
FIGURE III
PERCENT OF PERIOD FERROUS IRON CONCENTRATION IN FLQCCULATION TANK 3 mg/1 OR LESS VS. pH
-------
the saturation value of oxygen in the waste and the
actual dissolved oxygen). However, higher temperatures
increase the ability of oxygen to pass through the
liquid film from the atmosphere. The net result of
these two effects is a general increase in oxygen
transfer with an increase in temperature. Finally,
higher temperatures increase the reaction rate, thereby
decreasing the dissolved oxygen concentration and in-
creasing the driving force.
At higher temperatures, e.g. 150°F, the oxidation of
ferrous iron would probably progress toward completion
in less than the design aeration time of 30 min. Since
start-up, the treatment plant has never encountered a
waste temperature higher than 122°F (50°C). The normal
range is 104° to 113°F. Pilot scale experience with a
very similar process indicates a definite "trade-off"
between temperature and aeration time. Apparently, the
full scale facility successfully handled wastes at
lower temperatures because higher aeration tank
retention times were experienced.
During periods of sub-freezing weather, ARW temperatures
as low as 59°F (15°C) were recorded. Fortunately, these
low temperatures lasted only 6 to 12 hours at a time and
were accompanied by low flow rates and iron concentra-
tions; there was no apparent effect on the process.
Effect of Sludge Recirculation
The advantages of returning sludge to the aeration-mix
tank have been discussed (Section IV and Appendix A).
Others (3>6,7) have commented on the effectiveness of
recirculation on improving sludge settleability and
filterability.
An attempt was made to quantify the amount of sludge
recirculated by running at a selected rate for one week
followed by one week during which no sludge was
recirculated. Variations in the solids concentration
in the clarifier underflow, which can be directed
either to the aeration-mix tank or to the filters, made
it difficult to operate at a specific rate; these con-
centrations have been observed to vary from 120 - 800
g/1 over a three-day period. The solids concentration
in the aeration-mix tank did not exceed 2 g/1 without
recirculation, and was never less than 3 g/1 (average
7 g/1) with recirculation.
15
-------
Observations made over the sixteen month period indicate
three important benefits from recirculating sludge:
a. There was a noticeable red color, due to finely
divided ferric hydroxide, in the clarifier
effluent each time recirculation was halted.
On restoration of sludge recirculation, the red
color disappeared.
b. Sludge recirculation improved sludge settle-
ability. When recirculation was stopped the
flocculator tip speed required to keep the
sludge suspended was reduced by about 25
percent.
c. During the period when the total acid concen-
tration in the ARW far exceeded the capacity
of the facility, the alkalinity values in the
recirculated sludge (from the excess limestone)
contributed towards neutralizing the acid.
Sludge recirculation has no effect on the final pH when
the facility is operating within the design limits.
Flocculator Operation & Polymer Practice
The original A-frame flocculators did not provide
sufficient agitation to keep the floe particles in
suspension (p. 23). Turbine-type flocculators were
substituted for the A-frames; when these mixers were
operated at their maximum tip speed of £.1 ft/sec
(24 rpm), excellent agitation was obtained. As
little as a 4 rpm decrease resulted in the formation
of a sludge layer several feet below the surface of
the tank. While this sludge layer was still being
kept in motion by the mixers, the flow from the bottom
discharge pipe was impaired by the increase in the
solids concentration. Restoring the maximum mixer
setting lifted the sludge layer.
To aid in flocculation of the precipitated iron and
excess limestone, facilities were incorporated for
the preparation and dispensing of a water-soluble organic
polymer to the flocculation tank. The polymer, or
flocculating aid, is prepared in one of two 2,400 gal.
tanks at a concentration not exceeding 0.1 percent.
Positive displacement pumps transfer the solution to any
of three distribution points: across the top of either
flocculation chamber or into the discharge from the
aeration tank. Normal practice during a period of low
16
-------
flow rate, (less than 700 gpm) is to add the polymer
to the last chamber of the tank. This minimizes the
disintegration of the floe particles that results at
this low flow rate from extended contact with the
turbine mixers. At higher flow rates, addition of
the polymer into the aeration-mix tank discharge has
been satisfactory.
During the bench-scale investigation, a nonionic
polyacrylamide was found to be the most efficient
flocculating aid. At pilot scale the nature of the
precipitate was apparently different and an anionic
polyacrylamide was superior to the one previously
used. At the conclusion of the pilot plant investi-
gation, after approximately fifty polymers had been
evaluated, "Polymer A", a highly anionic polyacrylamide
derivative, was selected for full-scale trial.
Despite poor agitation in the flocculating tank,
performance was good; a 1.5 to 2.0 mg/1 dosage was
required for good flocculation.
After the breakdown of the flocculators and the
subsequent replacement of same with turbine mixers,
a higher dosage, (# to 10 mg/1) of polymer was re-
quired for flocculation. Jar testing at the site
suggested a change in polymer be made. "Polymer Bn,
also an anionic polyacrylamide derivative but having
a lower charge density than "Polymer A", was chosen.
The optimum dosage rate for "Polymer B" was found to
be about 3.0 to 3.5 mg/1.
Demonstration
This process clearly demonstrated an ability to
effectively, efficiently, and economically treat
large volumes of hydrochloric acid rinse waters.
The sludge produced was easily de-watered and readily
handled. Operating results for one typical month
during the demonstration period are presented in
Tables III and IV. Except during periods of plant
upset, the water produced by this process consistently
had a total iron concentration less than 5 mg/1, a
turbidity less than 15 J.T.U. and a pH of 6.8 to 7.1.
The average amount of chlorides and hardness added to
the waste stream by the picklers and treatment plant
can be calculated from the data in Tables III and IV.
These calculated values indicate that the picklers
add 475 mg/1 chlorides to the process water which
contains 55 mg/1 chlorides. The treatment plant adds
670 mg/1 hardness as CaC03 to the 200 mg/1 hardness
of the influent river water used at the picklers. The
game fish, such as large-mouth bass, channel catfish,
and bluegill, that have been thriving in the waste
treatment lagoon are further evidence of the quality
of the effluent.
17
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
Table IV
TYPICAL TREATMENT PLANT OPERATING DATA (APRIL 1970J
Influent
Avg. Flow - 730 gpm
Avg. pH - 2.5
Avg. Temperature - 102° F
Avg. Total Acid - 490 mg/1
Process Requirements
Limestone @ 50$
Excess - 4.5 tons/day
Air - 250 SCFM
Flocculating Aid
©3.5 mg/1 - 30.7 Ib/day
Sludge Recircu-
lation2 - 60 gpm
Effluent
Avg. Total Iron - 4.0 mg/1
Avg. pH - 6.9
Avg. Turbidity - 10 JTU
Sludge Characteristics
Quantity - 4.6 tons/day (wet)
Moisture - 31$
Iron - 27fo (dry)
The treatment facility was easily operated by one man.
The primary functions of the operator included:
periodic tests of influent and effluent concentrations,
limestone feeder adjustment, flocculating aid make-up,
and sludge filtration. A regular inspection of the
mechanical equipment, general clean-up, and preventive
maintenance items were also included. Maintenance
personnel were required only when equipment trouble
developed.
The cost data presented in the following paragraphs
are based on the facility in Middletown. Costs will
vary for other installations depending on several
factors, such as geographical location and the volumes
and characteristics of the rinse waters.
Using the data given in Table IV, operating costs for
the demonstration period, excluding supervision and
sludge disposal are calculated to be $7,530/mo (Table
V). This is equivalent to 4.3^0/ton of steel pickled
or 24.00/1,000 gal. ARW.
1 Pickling Rate 240 tons/hour
2 Solids concentration varied between 10 and 75$
19
-------
Table V
TYPICAL TREATMENT PLANT OPERATING COSTS
Limestone, 135 tons @ $5.52/ton (delivered) $ 750
Flocculating aid, 920 Ib @ $1.38/lb 1,270
Operating personnel, one man @ 4.50/hr 3,200
(21 turns/wk)
Maintenance personnel, 185 hr/rao @ $6.94/hr 1,280
Repair material 400
Supplies 350
Utilities 280
Total operating costs/mo $ 7,530
Table VI presents the capital costs required to con-
struct this treatment plant at the Middletown Works
during the period of November 1967 through May 1969.
Table VI
TREATMENT PLANT CAPITAL COSTS
Yard Work $ 15,000
Structures 124,000
Equipment (Installed):
Collection Tanks & Pumps 30,400
Surge Tanks 70,400
Aeration-Mix Tank & Mixer 30,700
Flocculation Tank & Mixers 26,700
Clarifier 89,400
Limestone Feed Equipment 33,300
Vacuum Filters 41,300
Misc. Equipment 74,800
Instrumentation 105,000
Final Lagoon 142,000
Piping 190,000
Electrical & Lighting 144,000
Indirect Costs 243,000
Total Capital Costs 1,360,000
Changes, such as reducing the size of the back-up
lagoon, could reduce capital costs. The relative
location of the treatment facility to the waste
source could also affect piping costs significantly.
Variations in construction labor rates may also
affect capital costs.
20
-------
SECTION VI
EQUIPMENT DESIGN EVALUATION
Since construction of the treatment plant was begun
directly from bench level information, many of the design
criteria were made empirically. During the field studies
and demonstration period, some design errors became
apparent. The basic units of the existing facility have
been evaluated in the following sections.
Surpe Capacity
The ARW is transferred to the treatment plant with one or
both of two 750-gpm pumps. Therefore, the flow rate to
the plant is either zero, 750, or 1,500 gpra. To give some
equalization of flow, and acid concentration as well,
90,000 gal. of surge capacity were provided to receive the
ARW prior to treatment. This consists of two 45,000-gal.
reinforced fiberglass tanks arranged in parallel.
In practice, the ARW level in the surge tanks was main-
tained in the intermediate range by means of a throttling
valva downstream from the tanks. Adjustment of this valve
was rarely necessary. Although hour-to-hour variations in
flow through the treatment plant did occur (Fig. 4), these
changes were relatively gradual. In this plant, influent
conditions were checked periodically and chemical additions
manually adjusted. The effluent iron concentrations
remained satisfactory (also shown in Fig. 4), so additional
surge capacity is apparently unnecessary.
Aeration-Mix Tank
The aeration-mix tank is the reaction zone for the process.
Limestone is added here to neutralize free acid and raise
the pH to the appropriate level; air is mixed in to
oxidize ferrous to ferric iron. Research data from bench
level studies (Appendix A) indicated a 30-min retention
time was necessary for this tank. A square 45,000-gal.
aeration-mix tank was installed. Following equipment
manufacturer's recommendations, the aeration-mix tank was
equipped with a 30-hp submerged turbine type aerator and a
blower rated at 250 SCFM.
The tank was apparently sized correctly for 1,500 gpm at
150°F. The temperature of the ARW actually received was
21
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
considerably less than 150°E; fortunately, the flow was
considerably less than 1,500 gpm, and the 45,000-gal.
tank proved sufficient. (See Section V.)
Flocculation Tank
This tank is designed to promote flocculation of the
ferric hydroxide particles with the addition of an organic
polyelectrolyte. The tank is rectangular with a 45,000-
gal. capacity and 30-ruin retention time. A baffle was
mounted mid-way across the width of the tank to prevent
short-circuiting or stratification. It originally had
two A-frame oscillating flocculators, one in each chamber,
with a maximum tip speed of 1.6 ft/sec. Although bench
scale investigations (Appendix A) utilized a flocculation
time of 15 min, previous experience in water clarification
indicated a period of 30 min would be more favorable.
Upon start-up of the treatment facility, immediate
problems were encountered in the flocculation tank.
Sludge settled in the bottom of the tank, impeding and
eventually stopping the movement of the flocculators. Air
lances were used to break up the sludge formations; how-
ever, this damaged the floe particles, resulting in a
turbid effluent due to high iron carryover. Attempts were
made to alleviate the problem by using a finer grade lime-
stone and reinforcing the A-frame, but excessive settling
continued. Eventually one of the A-frames broke at an
axle weld.
The A-frame flocculators were replaced with turbine-type
flocculators. The turbine blades were 7$ in. in diameter
with a 45 deg pitch, and could be driven from 6 to 24 rpm
by means of a 5 hp variable-speed motor.
A flocculation time of 30 min was much greater than
actually necessary. The floe has proven to form quickly
and to be very dense. A flocculator-clarifier would seem
to be applicable here.
Clarifier
The 50-ft diameter clarifier was designed with a rise rate
of 0.8 gpm/sq ft at 1,500 gpm. This rise rate was chosen
by previous experience in water clarification.
If the rise rate were too high, iron floe particles would
be carried over in the clarifier effluent. Over a 10-day
period, the flow and total iron concentrations in the
23
-------
effluent were observed. The results are shown in Fig.
IV. During this period the flow varied between 400 and
1,200 gpm, and the total iron in the effluent was less
than 5 mg/1 for 95$ of the time; it exceeded 7 mg/1
only once, and then only for one hour. This indicates
that the clarifier is large enough for the maximum flow
achieved during the period. Due to the difficulty in
obtaining large flows for extended periods, the maximum
flow that this clarifier will handle is unresolved.
Sludge Filters
The filters were sized from bench scale data; assuming
an ARW temperature of 120°F and a maximum ferric/ferrous
ratio of 1/3, a filtration rate of IB Ib dry cake/sq ft/
hr was anticipated (Table A-II, experiment no. 7).
Material balance calculations indicate at design flow
and acid concentration,, and assuming the use of a 50
percent excess of limestone, a sludge accumulation rate
of 12.8 tons dry cake/day, or 1,060 Ib/hr. Therefore, a
minimum of 60 sq ft of filter area would be needed if
the filters were run continuously; a minimum of ISO sq
ft would be needed if the filters were run only one turn
daily. Two belt-type rotary vacuum filters were in-
stalled. Each filter has a 6-ft diameter and a 6-ft
wide face with approximately 113 sq ft of filtration
area. A vacuum of 5 SCFM/sq ft at 20 in. Hg is required.
Initially, sludge buildup on the belt due to inadequate
washing resulted in a tracking problem. This was
eliminated by the installation of spray headers to
provide a thorough wash on both edges of the belt.
The sludge accumulation rate experienced at the full-scale
facility was about 4 tons dry cake/day. The actual
filtration rate achieved was about SO Ib wet cake/sq ft/hr
with an average of 30 percent moisture; the difference
between the 56 Ib dry cake/sq ft/hr experienced and the
18 Ib value expected is explained by the practical
absence of ferric iron in the ARW. The operation of one
filter a few hours at a time several times per week was
sufficient to handle the sludge from this facility. The
filter cake density was approximately 100 Ib/cu ft so
removal of the cake from the plant site required about
one gondola car (1775 cu ft) per week.
By varying filter belt speed, cake thicknesses of 1/2 in.
to 2 in. were produced. If a filter cake thickness of
2 in. were maintained, however, undue stresses were added
to the belt and tracking problems resulted. A thickness
-------
of 3/4 in. appeared to be the optimum based on moisture
content and ease of operation.
Polypropylene belts were normally used on these filters;
nylon also performed satisfactorily, but is less
resistant to low pH values. Failure of the belt
generally occurred at the seam, probably due to the
abrasiveness of the limestone in the sludge.
Final Lagoon
The 16-acre 30 million gal. final lagoon was originally
proposed as a back-up for the treatment facility in
anticipation that some water not suitable for direct
discharge to a public waterway might be produced during
the testing program.
During the demonstration period it became apparent that
the final lagoon was unnecessary for the process. Armco
research scientists determined that stocking the final
lagoon with game fish would be feasible. Largemouth
basses, bluegills, and channel catfishes were obtained
from the Federal hatchery at Hebron, Ohio, or purchased
privately by Armco. Stocking of the final lagoon,
which contains only treated acid rinse water, has proven
to be a highlight of the demonstration phase of this
project.
Automatic Acidity Analyzer
To insure efficient operation of the process, the lime-
stone feed rate must be proportional to both the ARW
flow rate and to the total acidity. An automatic
acidity analyzer was originally provided for control
of the limestone feed rate. The analyzer receives a
constant sample flow of ARW and automatically and
continuously titrates this sample to a selected pH
value with a caustic (sodium hydroxide) solution. A
ratio setter is provided to compensate for variations
in caustic concentration. The amount of caustic flow
is measured and electronically converted to a signal
which is sent to a multiplier. A signal from a magnetic
flow meter measuring ARW flow is also sent to the
multiplier. The multiplier combines these two signals
and in turn sends a signal to the limestone feeders
which is proportional to limestone demand.
Since a continuous acidity analyzer for this particular
application was not commercially available, a custom-
designed instrument was built. It proved to be a highly
25
-------
sophisticated instrument, too delicate for plant
operations. Frequent adjustment and cleaning were
needed. Instead, the operator determined ARW conditions
and manually adjusted the feeders.
An acidity analyzer could be worthwhile if it were built
simply enough to allow inexperienced personnel to handle
the instrument.
26
-------
SECTION VII
ACKNOWLEDGMENTS
The research studies and field activities reported
herein were carried out by the following employees of
the Armco Steel Corporation: Mr. J. E. Barker, Project
Director; Mr. G. A. Pettit, Technical Consultant;
Messrs. R. J. Bendure, M. Dannis, and V. Y-7. Foltz,
Project Services; Mr. S. F. Melzer, Project Chemist;
Mr. T. L. Taubken, Project Engineer; and Mr. J. L.
Bauer, Jr., Assistant Project Engineer.
The employees of the Middletown, Ohio, Works of Armco
Steel Corporation are recognized for their cooperation
throughout the course of this study.
The partial support of this study by the Federal Water
Quality Administration, Project 12010 DUL, and the
advice and assistance of Mr. Robert L. Feder, Project
Officer is hereby acknowledged.
27
-------
SECTION VIII
REFERENCES
1. Anon., "Acid Mine Water Plus Limestone Equals Clean
Stream," Coal Age, 24 1 No. 2, 112 (1969).
2. Deul, M. , "Limestone in Mine Drainage Treatment,"
Mining Gongr. Journal, 5J3., No. 11, 83 (1969).
3. Faust, S.D., Orford, H.S., and Parsons, W.A.,
"Control of Sludge Volumes Following Lime
Neutralization of Acid Waste," Sewage Ind. Wastes,
28, No. 7, 872 (July, 1956).
4. Hill, D.W., "Neutralization of Acid Mine Drainage,"
Jour. Water Poll. Control Fed.. 4.1, 1702 (1969).
5. Hoak, R.D., "Disposal of Spent Sulphuric Pickling
Solutions," Ohio River Valley Sanitation Commission,
Oct., 1952, p. 25.
6. Judkins, J.F., Jr., "Crystal Seeding for the
Control of Sludge Properties," Thesis, Master of
Science, Virginia Polytechnic Institute, Blacksburg,
Va. ( 1964) »
7. Judkins, J.F., Jr., and Parsons, W.A., T! Optimization
of Acid Waste Sludge Characteristics/' Jour. Water
Poll. Control Fed.. £1, No. 9, 1625, (Sept., £969).
8. Mihok, E.A., Deul, M. , Chamberlain, C.E., and
Selmeczi, J.G., "Mine Water Research, The Limestone
Neutralization Process," Report of Investigations
No. 7191, Bureau of Mines, U.S. Dept. of the
Interior, Sept., 1968.
9. Rathmell, R.K., U.S. Patent 3,261,665 (1966).
10. Unifloc LTD and Samuel, J.D., British Pat. 1,046.04.1
(1966); Chem. Abs.. 66, 40609 (1967).
28
-------
SECTION IX
GLOSSARY
LIST OF TERMS
Acid Rinse Waters - Low acid and iron mixtures resulting
from rinsing of steel strip after hydrochloric acid
pickling; and fume scrubber water.
Oxygen Transfer Efficiency - The ratio expressed as a
percentage, of oxygen consumed to oxygen supplied.
SJLudge - Precipitated ferric hydroxide and excess lime-
stone.
29
-------
LIST OF ABBREVIATIONS
ARW - acid rinse waters
cps - centipoises
gpm - gallons per minute
hp - horsepower
in. Hg - inches of mercury
SCFM - standard cubic feet per minute
TDH - total dynamic head
30
-------
SECTION X
APPENDIXES
A,
B,
Bench Scale Research & Development
Work
Page
32
Table A-I:
Neutralization of Acid
Rinse Water with
Doloraitic Lime - Bench
Scale ,
Table A-II: Neutralization of Acid
Rinse Water with Lime-
stone - Bench Scale ..,
Detailed Engineering Report & Drawings
Figure B-I: Schematic Flow Sheet ,
Figure B-II: Plot Plan ,
Figure B-III: General Arrangement ,
39
40
41
46
47
31
-------
APPENDIX A - BENCH SCALE RESEARCH & DEVELOPMENT WORK
Int _r_p du c t io n
Because hydrochloric acid pickling is a relatively new
process, methods for a practical chemical treatment of
the wastes were not known. A considerable number of
treatment processes have been proposed for the disposal
of spent sulfuric acid. Though most of these do not
appear to be economically feasible, the initial approach
employed for our investigations into the disposal of
iron chloride wastes was a modification of previously
proposed methods. These methods treat acid waste with
lime. Free acid is neutralized, and dissolved metal
salts are precipitated. The resulting sludge may be
separated by filtration or by settling, and the
effluent discharged into a public waterway.
A plant has been in operation at Armco's V/orks at Butler,
Pa., for seven years for treatment of nitric acid-
hydrofluoric acid spent pickle liquor resulting from
pickling of stainless steel. During development of this
treatment process, various alkaline reagents had been
tested. Some of these, such as sodium hydroxide,
ammonium hydroxide, ammonium carbonate, high-calcium
lime, and magnesium oxide, precipitated the metals but
did not produce a filterable sludge. Others, such as
dolomitic lime and limestone, produced a filterable
sludge. Limestone was discarded because complete
precipitation of metals was impossible. Dolomitic lime
is now used in this process.
For treatment of sulfuric acid spent pickle liquors,
two methods (5,9) use lime to neutralize the free acid
and precipitate the iron with controlled aeration of
the mixture at elevated temperatures. This converts the
ferrous iron to precipitated ferroso-feric oxide,
FeO-FeoO-3. Optimum efficiency is claimed when a
ferric/ferrous ratio between 2 and 5 is maintained.
Description of Problem
Rinse waters from full-scale pickling operations were
expected to contain 1.0? grams of hydrochloric acid
equivalent per liter, half as free acid and half as
iron chlorides. Further, the wastes were expected to
enter the treatment plant at about 100°F. The
temperature range of 75°F to 120°F was therefore used
in this investigation.
32
-------
Inventj.^atipn^ of Treatment with Dolomitic Lime
To determine the ferric/ferrous ratio at which ferroso-
ferric oxide is formed from hydrochloric acid solutions,
waste waters containing ferric/ferrous ratios of 1/0,
3/1, 2/1, and 1/1 were treated with 120 percent of the
theoretical amount of dolomitic lime (alkalinity 92.5
percent as CaO). Two solutions, one ferric and the
other ferrous chloride, were prepared such that any
combination of the two, totaling 50 ml and diluted to
one liter, would be equivalent to the free acid and
iron chloride expected in the waste water.
This investigation was carried out on a relatively
small scale, in beakers, by batchwise addition of
neutralizing chemical to the simulated waste. The
procedure was as follows:
Fifteen hundred milliliters of tap water at the
desired temperature were added to each of four
beakers. A 500 ml buret was positioned above
each beaker; each was filled with 100 ml of con-
centrated acid-iron chloride solution and 400 ml
of tap water at slightly above the desired
temperature. As the burets were emptied into the
beakers, a slurry of neutralizing chemical was
added in greater than stoichiometric amounts.
The contents of the beakers were stirred at 120
rprn (tip speed 2 ft/sec) during the additions.
After 15 min reaction time, 1.5 mg/1 flocculating
aid was added. This was followed by mixing at 120
rpm for one min and at 60 rpm for 15 min. The
mixer was then shut off and the sludges permitted
to settle.
After siphoning off the supernatant solutions,
the sludges were combined and accumulated by
repeating the above procedures.
Variation in this procedure consisted of running
each neutralization in the presence of "seed"
sludge from previous batches.
A test leaf filter having an effective area of 0.1 sq
ft was used with a nylon filter cloth. Filtration
tests were made when sufficient sludge was collected.
Data obtained were: total solids in the accumulated
sludge before filtration, weight of wet and dry filter
33
-------
cake, water in the wet cake, iron in the dry cake,
thickness of cake and volume of filtrate.
Sludge produced by lime treatment of Butler waste pickle
liquor was used as a control. The results obtained for
both the hydrochloric acid and control solutions are
given in Table A-I. The data indicate that, at best,
only one-third as much dry filter cake is produced with
chloride wastes as with the control sample. Apparently,
weak hydrochloric acid waste did not react in the same
manner and another approach was necessary.
In no case was a ferroso-ferric oxide observed.
The size and density of the precipitated metal compounds
affect both settling rates and filtering rates. The
high activity of lime apparently results in the formation
of many small particles of ferrous hydroxide when
chloride wastes are treated. This suggests that the best
filtering characteristics might be produced when the
hydroxyl ion is introduced slowly, resulting in slower
precipitation and agglomeration of the metal compounds.
Limestone, because of its slower reaction rate, might
serve this purpose. Subsequent tests were carried out
using limestone.
Investigation of Treatment with Limestone
Limestone alone does not convert ferrous iron into
insoluble Fe(QH)2» However, by passing air through the
solution while carrying out the treatment with limestone,
iron can be oxidized and quantitatively precipitated
from the solution as ferric hydroxide. Consequently,
the test procedure described previously was modified as
follows:
1. Air was bubbled into the solution through
gas dispersion tubes during a 30 min
neutralizing and precipitating period.
When the iron was initially present in the
ferric form, no air was used and a 15 min
reaction time allowed. Then aeration was
discontinued and flocculating aid was
added.
2. Initial and final pH of the solution were
measured.
3. A test leaf filter having an effective area
of 0.0256 sq ft was substituted for the 0 1
sq ft filter.
-------
The alkalinity of the limestone was 55.5 percent (as
CaO) ; the chemical analysis was 6?.5 percent CaCO-j,
10.2 percent MgCCh, 1.3 percent Si02, and 0.45 percent
A1203.
Ideally, the limestone requirement would be that required
to just neutralize an acid solution. In practice, the
physical combination of magnesium carbonate in the lime-
stone matrix affects its reactivity, and greater than
stoichiometric quantities are required. A limestone
excess of 50 percent was necessary for solutions
containing ferrous iron, whereas a 20 percent excess
appeared suitable for wastes with all of the iron in
the ferric form.
All iron in the acid rinse water should initially be in
the ferrous form. However, through exposure to
atmosphere a portion of the ferrous iron might be
converted to ferric iron. The ratio of ferric to
ferrous iron at the time of treatment was not known
until full-scale pickling operations were begun.
Consequently, this variable was included in the
laboratory studies.
Values obtained were for waste waters having ferric/
ferrous ratios of 0/1, 1/3, and 1/0. Accumulating
sufficient amounts of sludge for filtration tests was
tedious because the sludge produced with limestone was
more dense and occupied less volume than that produced
with lime. This problem was ameliorated by substituting
a 0.0256 sq ft test leaf filter for the 0.1 sq ft
filter.
Various chemicals to aid in the flocculation of the
precipitated iron were evaluated. Nonionic polymers
were found to be satisfactory when used at a concen-
tration of 1.5 mg/1.
The best sludge was produced from the solution contain-
ing no ferric iron. The specific gravity of this
sludge after standing 16 hr was 1.5 g/ml at 21°C. The
viscosity at 21°C, using a Brookfield Model RVF-100
viscometer with the no. 4 spindle, was 2,200 cps at 20
rpm and 1,700 cps at 50 rpm. The settling rate of
this sludge was too high to measure. This sludge
yielded filtration rates as high as 114 Iba dry cake/sq
ft/hr and 30 Ibs iron/sq ft/hr (Table A-II).
35
-------
When a solution having a ferric/ferrous ratio of 1/3
was treated, the filtration rate values dropped
dramatically. The waste having no ferrous iron gave
the lowest values; this sludge also exhibited the
poorest settling rate, about 3 in/min.
To oxidize ferrous iron to ferric with molecular oxygen
and precipitate the iron as ferric hydroxide, a pH in
excess of 5.7 is critical. The first few batches of
all experiments involving ferrous iron failed to achieve
this critical pH. When enough excess limestone had been
accumulated from preceding batches to push the pH above
5.7, the ferrous iron was oxidized and precipitated.
The total iron in the supernatant solution then dropped
to less than 1.0 mg/1.
Studies of temperature effects were carried out on
waste waters having ferric/ferrous ratios of 0/1 and
1/3 at 100° and 120°F. *t a ferric/ferrous ratio of
1/3, a 25 percent reduction in the filtration rate was
experienced when the temperature was increased from
100° to 120°F. The temperature effect was quite
different at a ferric/ferrous ratio of 0/1, however.
The sludge prepared at 120°F was exhausted after only
30 sec filtration, one-half of the usual filtration
time. Had sufficient sludge been available for a one
rain filtration, the yield probably would have been
higher than that produced when the same solution was
treated at 100°F.
Normally, the runs were made with limestone containing
particles of which 92 percent were less than 115 mesh
and 74 percent were less than 200 mesh. A coarser
grade of limestone, one with particles less than 100
mesh and larger than 200 mesh, was used to neutralize
solutions having ferric/ferrous ratios of 1/3 and 1/0.
At a ferric/ferrous ratio of 1/3, it was difficult to
get the pH up to 5.7 in order to oxidize the ferrous
iron. A waste water solution having all ferric iron
was successfully treated. Switching from the finer to
the coarser limestone produced a 60 percent increase
in filtration rate.
Summary
Laboratory experiments have been made to determine the
parameters necessary to precipitate iron from a dilut-3
hydrochloric acid pickling waste so as to obtain both"a
filterable sludge and water of acceptable quality.
36
-------
In terms of f ilterability, the sludge produced by lime-
stone treatment of dilute hydrochloric acid wastes is
obviously superior to that produced by lime. The best
sludge results from treatment of wastes containing no
ferric iron; the difference is more dramatic in the case
of limestone than with lime.
Conclusions
1. Limestone, in amounts up to 150 percent of the
stoichiometric amount required, has been success-
fully used to produce a filterable sludge. Lime
completely precipitates the iron but yields a
sludge generally unsuitable for filtering.
2. Aeration and a minimum pH of 5.7 is necessary to
oxidize ferrous iron and precipitate it as ferric
hydroxide.
3. Optimum filtering characteristics are obtained if
all of the iron is initially present as ferrous
iron. This sludge has a specific gravity of 1.5
g/ml (21°C) and a viscosity of 2,200 cps at 20 rpm
and 1,700 cps at 50 rpm. The settling rate is too
fast to measure. Filtering rates of 114 Ibs dry
cake/sq ft/hr or 30 Ibs iron/sq ft/hr have been
achieved on a laboratory scale.
4. Minimum filtering rates are obtained if all of the
iron is initially present as ferric iron.
5. The recirculation of sludge is necessary to produce
an acceptable filtration rate.
6. A nonionic organic polymer at a concentration of
about 1.5 mg/1 is required as a flocculating aid.
7. A better sludge was produced at 100 °F than at 120°F
only when the iron was present in the ferric form.
8. A larger limestone particle size increases the
rate of filtration only if all of the iron is
initially present as ferric iron.
9. Total iron in the treated waste was less than 1.0
1/min.
37
-------
Basis forJPlant Design
Based on this investigation the following operations
\vill be necessary for the treatment plant:
1. Aeration-Mix Tank - Dilute waste acid will enter
the plant at this point, where it is reacted with
a limestone slurry and previously precipitated
ferric hydroxide in the form of settled sludge.
Molecular oxygen is continuously introduced into
the solution in the form of air. Retention time
in this tank should be not less than 30 min.
2. Flocculating Tank - The mixture will flow to a
second tank where flocculation takes place.
Flocculating aid will be added at this point.
Retention time should be at least 15 min, and
slow mixing will be necessary.
3. Clarifier - The precipitate from the flocculating
process will be permitted to settle in a clarifier.
Means for returning a portion of the settled
sludge to the aeration-mix tank will be provided.
4. Vacuum Filter - After settling, the sludge will
be de-watered by filtration.
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
APPENDIX B - DETAILED ENGINEERING REPORT & DRAWINGS
Introduction
The terminal treatment plant is designed to prevent
stream pollution due to contaminated rinse waters from
two new hydrochloric acid picklers.
The facility is designed to neutralize free acid, oxidize
ferrous iron, then flocculate the resulting ferric
hydroxide and separate it from the water in a clarifier.
The process neutralizes 100 percent of the free acid
and oxidizes essentially 100 percent of the ferrous iron,
which is collected as ferric hydroxide sludge. This
effectively solves the problem normally associated with
hydrochloric acid rinse water (ARW).
General Description
ARW, including fume scrubber water, received at a
maximum of 1,500 gpm (750 gpm per pickle line) is
collected in one of two 12,000-gal. tanks, one adjacent
to each pickle line. The tanks are interconnected to
pumps which transfer the rinse waters through an over-
head pipeline, approximately 1,300 ft long, to two
45,000-gal. surge tanks. These surge tanks are elevated
to provide gravity flow through the treatment facilities.
The ARW then flows to an aeration-mix tank where lime-
stone is added and air is blown through to effect
oxidation. This neutralizes the free acid and reacts
with the ferrous chloride which forms insoluble ferric
hydroxide, when oxidized, and soluble calcium chloride.
The ferric hydroxide and the insoluble limestone solids
are then flocculated with turbine type flocculators
and, with the addition of a flocculating aid, are
agglomerated into a readily settleable mass which is
subsequently separated in a clarifier.
The clarifier effluent flows to a 16-acre, 30-million
gal. final lagoon which discharges to a public waterway.
The ferric hydroxide sludge collected in the clarifier
is pumped to rotary vacuum filters for de-watering.
The filter cake is collected in gondola cars and is
hauled away for storage. Clarifier sludge is also
recirculated to the aeration-mix tank.
Two 230-SCFM air blowers, two cylindrical limestone
storage bins with 60 deg conical bottoms, and two
41
-------
flocculating aid dilution tanks with 2,400 gal. capacity
each are required to support the above operation. Each
flocculating aid dilution tank has suitable mixing
equipment for dissolving the chemicals in water.
A storage area for a 30-day supply of flocculating aid
is also provided.
Control
The influent line to the aeration-mix tank has a magnetic
flow meter that measures the flow which is recorded in
the control room. Influent pickle rinse water acid and
iron tests are normally conducted each hour, and the
chemical feeders are adjusted manually by the operator.
Clarifier sludge rakes have a torque meter which indi-
cates amount of sludge in the clarifier. ARW surge
tanks, flocculating aid dilution tanks and limestone
storage bins have level gauges.
Jill pumps are controlled manually. This includes the
clarifier sludge recirculation pump and clarifier sludge
pump to the vacuum filters.
Enclosure
The enclosure for the terminal treatment plant is BO ft
long and 70 ft wide. The enclosure contains the lime-
stone storage and feed bins, flocculating aid dilution
tanks, dry storage area, air blowers, and rotary vacuum
filters.
MAJOR EQUIPMENT
Pickler Storage Tanks
One 12,000 gal. reinforced fiberglass storage tank is
located at each of the two HC1 picklers. Two 750-gpm
pumps transfer the ARW from these storage tanks 1,300
ft in an overhead line to the surge tanks located at
the terminal treatment plant.
Surge Tanks
Two 45,000-gal. reinforced fiberglass tanks provide
equalization and surge capacity for the treatment plant.
These tanks are elevated to provide gravity flow for
the remainder of the system. These tanks are inter-
connected and have a manually operated valve so that
flov; through the plant may be regulated. They provide
42
-------
a one-hour minimum retention time. Effluent from the
surge tanks flows by gravity to the aeration-mix tank.
Aerat ion-Hi_x_ Tank
The aeration-mix tank is a 23 ft square by 15 ft high
concrete walled tank with a 30-min minimum retention
time. The aeration-mix tank has a 30-hp turbine-type
aeration mixer designed to mix limestone and recirculated
sludge into the waste water and maintain the suspension.
Air is blown to the aeration mixer by a 25-hp blower
capable of 230 SCFM to oxidize the iron from ferrous to
ferric. Sludge is recirculated to the aeration-mix tank
from the clarifier by an air-operated diaphragm-type
pump capable of pumping 60 gpm. Effluent from the
aeration-mix tank flows by gravity to the flocculation
tank.
Limestone Storage Bins
The limestone is stored in one of two 75-ton capacity
bins equipped with screw operated dry feed mechanisms.
Each bin is 13 ft in diameter with 9-ft straight sides
and a 60 deg conical bottom. Bins are filled by
pneumatic piping from an outdoor truck loading station.
The bins are equipped with a dust collector. The
limestone is dry fed into a trough in which it is
slurried with water from recirculated clarifier water.
The slurry is then fed to the aeration-mix tank.
Flocculation Tank
The flocculation tank is 30 ft long, 16 ft wide, and
14 ft deep. It is concrete walled and has a design
minimum retention time of 30 min. The tank is divided
into two sections by a baffle in the middle of the
tank. This extends from the surface down to 7 ft from
the bottom of the tank and is designed to prevent
stratification and short-circuiting in the flocculation
tank. Two 5-hp, variable-speed turbine-type
flocculators keep the floe particles in suspension.
Flocculating aid is added in the flocculation tank at
the surface of the first half or the second half, or
both halves of the tank. The flocculating aid is
mixed in one of two 2,400-gal. steel solution tanks
equipped with dispersion funnels and mixers. The
flocculating aid is then fed to the flocculation tank
by a 1-hp proportioning pump capable of 225 gal./nr*
43
-------
jClarifier
The clarifier is a steel shell with sludge rake assembly,
It has a diameter of 50 ft and a side water depth of 15
ft. At 1,500 gpm the rise rate is 0.3 gpm/sq ft. The
clarified water flows to a final lagoon. The sludge is
periodically filtered. Clarifier effluent water is
provided to the terminal treatment plant as process
water.
Vacuum Filters
Some of the sludge which settles in the clarifier is
constantly recirculated to the aeration-mix tank.
When necessary to filter sludge, it is pumped to one of
two rotary vacuum belt-type filters by means of an air-
operated diaphragm sludge pump capable of about 60 gpm.
The filters are 6 ft in diameter and 6 ft long with
filtrate receivers, filtrate pumps, and vacuum pumps
with 30-hp motors. The filter cake drops into gondola
cars and is hauled away for storage. The gondola cars
can be moved limited distances within the immediate
terminal treatment plant area by a car puller driven
by a 10-hp motor.
Major Equipment
Three (3) transfer pumps for pumping ARW from the
pickle lines to surge tanks, 750 gpm @ 100 feet TDK
60-hp motor
Three (3) sump pumps, 75 gpm @ 40 feet TDH, 5-hp motor,
at pickle line collection areas
Two (2) sump pumps, 75 gpm @ 70 feet TDH, 15-hp motors
at terminal treatment plant
Two (2) ARW collection tanks, 12,000 gal. each, rubber-
lined steel, at pickle line
Two (2) influent surge tanks, 45,000 gal. each,
reinforced fiberglass, at terminal treatment plant
Two (2) limestone storage bins, 3-day supply (75 tons)
each, 13 ft diameter with 9 ft straight sides and 60
deg cone bottom, with pneumatic piping from outdoor
truck loading station
Two (2) limestone feeders, dry volumetric, 60 cu ft/hr
44
-------
One (1) aeration mixer with gear reducer and 25-hp motor
Two (2) flocculators, turbine type, 5-hp motor
One (1) clarifier, 50 ft diameter by 15 ft side water
depth, steel shell with sludge rake assembly, rated @
0.8 gpm/sq ft for 1,500 gpm
Two (2) sludge pumps, air-operated diaphragm type, 60 gpm
(2) sludge filters, rotary vacuum belt-type, 6 ft
diameter by o ft long drums, with filtrate receivers,
filtrate pumps, and vacuum pumps with 30-hp motors
One (1) flocculating aid proportioning pump, 225 gal./hr,
with 1-hp motor
Two (2) flocculating aid solution tanks, 2,400 gal.
steel with agitators
Two (2) clarified water pumps, horizontal centrifugal,
200 gpm @ 200 feet TDH
One (1) car puller, with rope guides and return sheaves,
with 10-hp motor
45
« V. S. GOVERNMENT PRINTING OFFICE : 1971 O - 4Z2- 345
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
|