&EPA
United States
Environmental Protection
Agency
Industrial Environmental Research EPA-600/2-80-058
Laboratory March 1980
Cincinnati OH 45268
Research and Development
Removal of
Fluorides from
Industrial
Wastewaters Using
Activated Alumina
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
qories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4, Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental 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 document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
1
EPA-600/2-80-058
March 1980
REMOVAL OF FLUORIDES FROM INDUSTRIAL WASTEWATERS
USING ACTIVATED ALUMINA
by
Irwin Frankel
and
Eric Juergens
Versar, Inc.
Springfield, Virginia 22151
for
The Feldspar Corporation
Spruce Pine, North Carolina 28777
Grant No. R-804377
Project Officer
Mary K. Stinson
Industrial Pollution Control Division
Industrial Environmental Research Laboratory
Edison, New Jersey 08817
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
Ohis report has been reviewed by the Industrial Environmental Research
Laboratory -Cincinnati, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
nent or recommendation for use.
11
-------
FOREWORD
When energy and material resources are extracted, processed, converted
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly more efficient pollution
C?ncii T £-„?%"?• -The Industrial Environmental Research Laboratory -
^h Tf "• i -Cl).^S1StS ln develoPin§ «nd demonstrating new and improved
methodologies that will meet these needs both efficiently and economically
This report is a product of the above efforts. These studies were
undertaken to perform a laboratory-scale investigation to remove excess
concentrations of fluoride ion from wastewaters emanating from plants that
separate _ feldspar from ore by wet-process methods. The fluoride removal
from the feldspar-sand
,vf*UCh inflation will be of value both to EPA's regulatory program
(Effluent Guidelines Division) and to the industry itself in arriving at
meaningful and achievable discharge levels. Within EPA's R&D program, the
information will be used as part of the continuing program to develop and
^1^ C°Stly technol°Sy to minimize industrial waste
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
111
-------
ABSTRACT
A four-step, bench-scale process has been developed that is capable of
removing at least 92 percent of the soluble fluoride from feldspar flotation
SScess wastewaters afa projected cost of $1.03/Kkg ($0.93/ton) _of feldspar.
For a 163,000 kkg/year (180,000 ton/year) plant, the initial capital expendi-
ture would be about $200,000. The wastewater is pretreated with sodium
hydroxide, line, and calcium chloride, which removes an initial 50 to 60 per-
cent of fluoride. The clarified water is then passed through a bed_of _
activated alumina for further fluoride removal. The activated alumina is
regenerated with a 1-percent sodium hydroxide solution, and fluoride in
the caustic liquor is effectively precipitated with calcium chloride. The
fluoride can be recovered in concentrated form as insoluble calcium fluoride
filter cake.
This report was submitted in fulfillment of Grant No. R-804377 by the
Feldspar Corporation under the sponsorship of the U.S. Environmental Pro-
tection Agency. This report covers the period May 10, 1976, to January 31,
1977, and work was completed as of March 22, 1977.
IV
-------
CONTENTS
Foreword . ...
.... 111
Abstract .
• iv
Figures ........
^ vi
Tables : :
• ,. .... . , . vi
Acknowledgment •••-.....
-.-•»-. . vxi
1. Introduction -,
2. Conclusions 1 .... 4
3. Recommendations 5
4. Experimental ^ 6
5. Results and Discussion 8
6. Summary of Process for Removing Fluoride from Feldspar
process wastewater ..... 16
7. Economics -,q
Bibliography 22
Appendices
A. Summary of Experimental Tests on Alcoa Activated Alumina ... 23
B. Analytical Methods 25
C. Trip Reports 28
D. Cost Estimate Work Sheets 41
v
-------
FIGURES
Number
1
2
3
Bench scale system for removal of fluoride from feldspar
process wastewater «
Partial renoval of fluoride from feldspar process waste-
water by calcium hydroxide
Feldspar wastewater defluoridation process flow diagram .
14
18
TABLES
Number
1
2
3
Removal of Fluoride from Alumina Regenerant Liquor
Economics Summary
Estimated"Costs for Major Equipment Items
12
20
21
VI
-------
AOKNOWLEDGMENT
The authors gratefully acknowledge the help and cooperation of Mr. Larry
Sparks, Research Director, Mr. Charles Wiseman, Chemist; and Mr. Richard L
Barber, Vice President of Production; Feldspar Corporation personnel in
Spruce Pine, North Carolina; and also Mr. Ben Robinson, Manager of the
Middletown, Connecticut, plant of the Feldspar Corporation. Frequent contact
was made with Mr. Sparks and Mr. Wiseman during the course of this study to
obtain samples of wastewaters and of flotation agents, or to obtain needed
information on plant operations, and their assistance wasi always forthcoming.
Mr. Robert E. Anderson, Senior Public Vtorks Operator of the alumina
treatment plant in Desert Center, California, was of considerable assistance
in his detailed directions pertinent to the regeneration and reactivation of
our activated alumina columns.
Keen interest in this project was displayed by the EPA Project Officer,
Mrs. Mary K. Stinson, of Edison, New Jersey. Mrs. Stinson gave us detailed
instructions for the preparation of this report and helped greatly in keepinq
the project active to its ultimate conclusion.
vii
-------
-------
SECTION'!
INTRODQCTION
BACKGROUND
Five major and several minor feldspar ore processing plants in the
United States use multi-stage flotation methods to separate feldspar from
other ore constituents. Hydrofluoric acid is a necessary, key reagent in the
feldspar-sand flotation step of the process. Other processes or more environ-
mentally acceptable reagents have proven relatively unsuccessful for a
variety of reasons. Some of the plants, including the three operated by the
Feldspar Corporation, have made some in-process changes for isolating and
reusing flotation water containing hydrofluoric acid. However, a major part
of the fluoride introduced during feldspar flotation is eventually diluted
and discharged with other plant wastewaters.
The Feldspar Corporation plant at Spruce Pine, North Carolina, as
presently operated, produces a combined effluent from the feldspar flotation
of 450 to 1,140 liters/min (120 to 300 gpm), with a fluoride content ranging
from 60 to 135 ppm. This effluent is combined with other process effluent
streams with a flow rate of about 6,430 liters/min (1,700 gpm), containing
about 0.25 ppm of fluoride. The combined stream is treated with lime to a
pH of 9.5 to 10.0 to flocculate the suspended solids. The lime flocculant
does not precipitate a significant amount of fluoride, and the discharge from
the plant may contain as much as 23 ppm dissolved fluoride;.
Discharge of plant effluents containing large concentxations of fluoride
ion into waterways creates a problem because these streams! may serve as
sources of water for downstream corrnunities and for wildlife. Unless dilution
by the water is sufficient (to about 1 ppm or less), the residual fluoride
concentration is considered detrimental. Too much fluoride intake by animal
life, including man, can result in permanent bone and teeth deformities in
the young. At a minimum, fluorosis (mottling of the teeth) can occur and
lead to early loss of thl teeth. On the other hand, some discharge of
fluoride into waterways can be permitted if the background, level of fluoride
in the water is quite low, and if dilution by the waterway is great, nil-such
areas as Spruce Pine, North Carolina, where there are three wet-process
feldspar plants operating, the downstream fluoride levels are closely
monitored.
Numerous attempts to remove fluoride from water have generally met with
poor to mediocre success. These attempts were based on the use of alum,
phosphates, magnesium chemicals, bone char adsorption, and other materials.
However, one process presently in use at three known locations in the
-------
United States is successfully and economically removing low concentrations of
fluoride from otherwise potable water supplies. This method is based on the
adsorption of fluoride on activated alumina.
In the activated alumina process for drinking water, the fluoride-
containing water is passed through one or more fixed beds of activated
alumina granules, and fluoride adsorbs on the alumina surface. The treated
wastewater is then ready for use. Regeneration of fluoride-saturated beds
is accomplished by passing dilute sodium hydroxide solution through the beds,
followed by washing with dilute acid. Usually two or more fixed beds are
used so that the beds can be alternated between adsorption and regeneration
cycles.
This process has been used successfully for years in reducing fluorides
to the level of 1 mg/liter in potable water supplies. This proven fluoride
removal process should be applicable to the feldspar industry problem and
also to fluoride-bearing wastewaters of other industries.
Activated alumina is prepared by controlled thermal treatment of
granules of hydrated alumina. It is primarily a porous aluminum oxide with
a large surface area (a typical grade of activated alumina used for water
treatment purposes has a specific surface area of 210 m2/g). The large
surface area and the relative polarity of activated alumina make it suitable
for adsorptLve removal of fluoride from aqueous solution.
The above process for removing fluoride from otherwise potable water is
both technically successful and economically feasible. But transferring the
application of this simple technology to an equally successful industrial
process could require considerable research effort for the following reasons:
• Industrial process wastewaters could contain components that
would interfere with fluoride adsorption or alumina regeneration,
or otherwise make the process inefficient or uneconomical.
• Because the process wastewater contains large concentrations of
fluoride, the activated alumina becomes saturated and requires
regeneration very quickly. Since regeneration is the costliest
step in the process, frequent regenerations escalate the process
cost.
• A litre pretreatment of wastewater will significantly reduce fluoride
concentrations and thereby reduce the load on^the activated alumina.
This step is less expensive than the activated alumina step, but
it will require study.
• In the three known plants that produce potable water by activated
alumina removal of fluoride, the dilute caustic soda regenerant
liquor is simply returned to the environment--i.e., it is dumped
on the ground or into river beds that are dry most of the time.
This approach cannot be used by industry. High concentrations of
fluoride must be recovered and suitably disposed of without hazard
to the environment. This need may add a costly step in terms of
-------
additional equipment, chemicals, and manpower.
On April 14, 1976, the Feldspar Corporation in Spruce Pine
Carolina, was awarded a grant fromlhe SvlrcnnenSl aScSS'a
ss
.
tion deczded not to pursue Phase II and notified EPA of
dis.
. ^ Sarly date that activated alumina could in fact
w«TO 2chn^cal ^.economic optimization of two fluoride-reaving technioues
were therefore indicated and were the primary objectives of ou "^
-------
r
SECTION 2
CONCLUSIONS
A multi-stage process has been successfully developed on a laboratory
scale to reduce the fluoride ion content of feldspar process wastewater to
5 ppm or less.
The initial fluoride removal step, which consists of treating^the
feldspar flotation process wastewater with caustic, lame, and calcium
chloriS and then removing the insoluble calcium fluoride in a clarifier,
results in 50 to 60 percent fluoride removal.
Cn a full plant-scale basis, the process will cost on the order of
$1.03/kkg ($0.93/ton) of feldspar product, with an omtial capital expenditure
on the order of $200,000.
The process is not adversely affected by the presence of most flotation
agents, but the activated alumina adsorbent cannot be regenerated if fuel
oil is present in. the process wastewater.
The undesirable fluoride ion in the process wastewater is recovered as
solid, insoluble calcium fluoride. When recovered as a filter cake, calcium
fluoride can be sold or stored in regulated landfills.
-------
SECTION 3
RBCCM/1ENDATIONS
The process research in this phase has demonstrated the applicability
and suitability of removing fluoride from feldspar process! wastewater by
means of activated alumina. But it has left unanswered the questions per-
taining to optimization. Specifically, the areas of concern are:
1.
2.
Pretreatment of wastewater with calcium chloride and spent
regenerant liquor.
Precipitation of fluoride from regenerant liquor with
calcium chloride.
3. Recycling of treated process wastewater.
4. Recycling of regenerant.
All four items above are best dealt with on a pilot-plant scale, as
they can either be defined much more satisfactorily on a srale larger than
the laboratory Citems 1 and 2), or they require long-term operations (items 3
and 4} before any effects of recycling streams become apparent.
Consequently, the extension of this project to a pilot-plant scale is
reconmended, since the proposal is probably applicable to the wastewaters
from a number of other industries. Furthermore, the economics will probably
not be as onerous to industries that manufacture products with higher unit
values than feldspar.
-------
SECTION 4
EXPERIMENTAL
A bench-scale test unit was constructed based on two 6-on (2.5-in)
diameter, 76-on (30-in) high transparent columns containing a bed of acti-
vated alumina about 25 cm (10 in) high. In addition, other laboratory equip-
ment such as pumps, valves, tubing, tanks, etc., was purchased as needed.
A schematic diagram of the assembled equipment is shown in Figure 1.
For fluoride and pH analyses, an Orion 407A specific ion meter with
appropriate electrodes was chosen and purchased. Versar-owned equipment
was utilized as required, especially such analytical equipment as a
Beckman Acculab IR spectrophotometer and a Heath 704 spectrophotometer.
Process wastewater was shipped from the feldspar plants in Spruce Pine,
North Carolina, and Middletown, Connecticut, in plastic-lined, 208-liter
(55-gal) drums. Samples of activated alumina were obtained from Kaiser
Chemical Corporation, Filtrol Corporation, and Alcoa. Samples of the various
reagents utilized in the feldspar flotation process were furnished by
Feldspar Corporation. These included frother, oleylamine acetate collector
(General Mills "Alamac IIC"), and Williams Company slaked lime.
To evaluate the performance of the activated alumina column, an average
(weighted) effluent concentration of 5 ppm of fluoride was selected by
Versar and Feldspar Corporation personnel as a practical end point. When
the present flow of treated process wastewater (1,140 liters/min or 300 gpm)
at a fluoride concentration of 5 ppm is mixed with approximately 6,430 liters/
min (1,700 gpm) of other plant effluent (which has a background fluoride
level of 0.25 ppm), the total plant effluent should be less than 1 ppm.
Versar personnel made trips to the water treatment plants at Desert
Center, California, and Bartlett, Texas, and to the Spruce Pine, North
Carolina, plant of Feldspar Corporation. Some detailed data regarding the
design and operation of activated alumina treatment plants for potable
water supplies were obtained. In addition, better knowledge of a feldspar
wet-processing plant resulted for the Versar personnel, and dependable lines
of communication were established between Feldspar and Versar. The detailed
trip reports appear in Appendix C.
-------
LIME-TREATED
WASTEWATER
RESERVOIR
[~1M|XING MOTOR
U& IMPELLER
DIAPHRAGM
PUMP
A COLUMN IN OPERATING MODE
B 'COLUMN IN REGENERATION (BACKWASH) MODE
• VALVE, CLOSED
O VALVE, OPEN
222 ACTIVATED ALUMINA
ESJ GRAVEL
Figure 1. Bench scale system for removal of fluoride
from feldspar process wastewater.
-------
SECTION 5
RESULTS AND DISCUSSION
INITIAL TESTS
In the initial test, Spruce Pine process wastewater, at a pH of about
2.5, was neutralized with hydrated lime. This step was necessary because a
low pH prevents adsorption of the fluoride by the alumina. After the waste-
water was neutralized with lime, a heavy yellow precipitate was formed, and
the fluoride content of the wastewater was decreased from 110 to 45 ppm.
When, the precipitate was isolated by filtration and submitted for analysis,
it proved to be roughtly 80 percent calcium fluoride (CaF2) and 20 percent
oleylamine acetate (OAA)*, the feldspar flotation reagent.
After removal of the precipitate, the supernatant was pumped to the top
of the column of activated alumina and allowed to flow down through the
alumina (Alcoa F-l, 28-38 mesh). The run was made with a space velocity of
0.0017 sec"1 (residence time of 10 min). It was continued past the initial
fluoride breakthrough (81.6 volumes of wastewater per volume of activated
alumina) and then past the point at which the average fluoride concentration
was 5 ppm (about 120 volumes). The run was arbitrarily halted at 136 volumes,
although the alumina was still removing some fluoride.
Regeneration of the alumina was initiated using a 1-percent sodium
hydroxide (NaOH) solution as the regenerant. Instrumental analysis of the
effluent caustic solution indicated that regeneration was not being accom-
plished, although regeneration conditions were the same as those established
at Desert Center (3 volumes of the above regenerant per volume of media over
a period of 20 min).
The above test was repeated, however, without proceeding to saturation
of the alumina, and regeneration was again attempted without success.
Two difficulties were recognized as a result of these tests: First, the
Orion 407A specific ion meter and fluoride electrode would not detect
fluoride ion accurately in the high-pH environment of a 1-percent caustic
solution; and second, the activated alumina had co-adsorbed an unknown
organic material that could not be removed during caustic regeneration and
that prevented regeneration of the alumina. The latter problem proved to be
* The formula of OAA is:
CH3 (CH2) 7CH = CH(CH2)
CH3COO
MW = 327.6
-------
quite^difficult, and considerable project time was spent: on its eventual
solution. The following section discusses these efforts: along with a number
of other ^ aspects, of the process that required study before pilot plant design
was_feasible. These aspects included removal of fluoride from regenerant,
minimum use of caustic in regeneration, process wastewater pretreatment, and
acid wash of alumina.
RESEARCH CDNDUCTED
Detection of Fluoride Ion in a High-pH Environment
t The high-pH environment of a 1-percent caustic solution prevented the
Orion_407A specific ion meter and fluoride electrode from accurately detecting
fluoride ion. This problem was solved by neutralizing such solutions before
analysis with a mineral acid that corrected the resulting fluoride analysis
by the amount of dilution introduced by neutralization.
Identification of Organic Co-Adsorbent
Oleylamine acetate (QAA), the feldspar flotation agent, was initially
suspected as the adsorbed organic compound, and therefore a study of the
extent of QAA adsorption by alumina was initiated. A general analytical
method for amines was obtained from the Ashland Oil Co. and was modified
slightly to be specific for OAA. This method was used to measure the con-
centration of QAA in the neutralized and filtered wastewater fed to the
alumina column, in wastewater discharged from the column, and in process
wastewater received from the feldspar plant.
After considerable effort, it was concluded that OAA was not the alumina
co-adsorbent since:
1. Alumina that removed fluoride from synthetic solutions based on
fluoride and OAA only was regenerable, and
2. Alumina that removed fluoride from feldspar process wastewater that
was several weeks old was not regenerable. This wastewater contained
no OAA, since by that time, the compound had completely biodegraded.
A search for other interfering adsorbates was undertaken. In one
approach, process wastewater (neutralized and filtered) was contacted with
activated charcoal. The charcoal was then washed with freon 113, and this
extract was analyzed by infrared spectrophotometry. The most significant
peak in this extract was characteristic of a large number of methylene
groups. The only compound with this characteristic known to be added to the
process at any point was fuel oil. ' Although the fuel oil was added to the
mica flotation step as a flotation agent, this step was followed by two
dewatering steps before the feldspar flotation. It therefore appeared
unlikely that the fuel oil was passed on in significant quantities.
A visit was made to the Spruce Pine plant where onsite experiments were
conducted to identify this contaminant positively and to locate its source.
The principal tool used was a Beckman Acculab infrared spsctrophotometer.
-------
As a result of this effort, the interfering material was positively identi-
fied as fuel oil, and very shortly thereafter, Spruce Pine personnel con-
firmed that significant quantities of No. 2 fuel oil were added to spar
float slurry as a pumping lubricant. Part of the resulting process waste-
water stream is recycled to the spar-sand flotation step. This stream con-
tains fuel oil concentrations of 100 ppm or more, depending on the partition
of the oil into sand, feldspar, and water from the spar float and from sand
overflow.
Fluoride Rsmoval by activated Alumina in the' Absence of Fuel Oil
A decision on how best to continue the process research work was
necessary. Should a process be sought to remove fuel oil, or should the
effort concentrate on an oil-free wastewater from another source?
Laboratory experiments showed that activated carbon removed the fuel
oil very effectively, but costs for this route were prohibitively high.
However, fuel oil addition to process wastewater is not necessarily repre-
sentative of the entire feldspar wet-processing industry. For example,_
Feldspar Corporation's plant at Middletown, Connecticut, no longer requires
a slurry pump. Their spar float flows by gravity to a vacuum^pan filter
where a very good spar^water separation is achieved. The decision was
therefore made to continue process research on oil-free wastewater. To do
this, synthetic solutions and process wastewaters from the Feldspar Corpora-
tion's Connecticut plant and from the small Plant #2 at Spruce Pine, North
Carolina, were used. These two locations do not use fuel oil as a pumping
lubricant, so the use of these plants as a source^of process wastewater
eliminated problems relating to alumina regeneration.
Connecticut plant wastewater contained about 125 ppm of fluoride and no
detectable fuel oil. When it was limed with 0.4 g/liter (0.0033 Ib/gal) of
calcium hydroxide and allowed to stand for 24 hr, the soluble fluoride con-
tent was analyzed at 25.6 ppm. After the calcium fluoride was filtered off,
the wastewater was passed through a column of activated alumina. After
saturation, the column was regenerated, and two more complete cycles were
run with reproducible results. The alumina capacities attained were 5.00,'
4.85 and 5.00 kg F-/m3 (0.312, 0.303 and 0.312 Ib F~/ft3) of alumina at a
weighted average effluent concentration of 5 ppm of fluoride. Similar
reprcducibility was obtained using synthetic fluoride solutions. These
results established the feasibility of using activated alumina to remove
fluoride from process wastewater that contains no fuel oil.
Rsmoval of Fluoride from Spent Regerierant
The 1-percent sodium hydroxide solution used for regeneration was
essentially unchanged after contacting the alumina, except for a high fluo-
ride content of 500 to 600 ppm or more. Clearly, this solution could not
be discharged as waste without reduction of the fluoride level. This step
can be accomplished by addition of calcium hydroxide, which precipitates
calcium fluoride, as in the pretreatment of the feldspar process wastewater.
It was hoped, however, that calcium chloride could be used to accomplish the
same purpose at a lower cost.
10
-------
Accordingly, a nurtber of experiments were carried out to test the
efficacy of calcium chloride and to provide some design data. The results
are shown in Table 1, where calcium chloride is compared 'with lime.
Various amounts of calcium chloride and lime were added to spent
regenerant liquor. The amount of reagent used was calculated in terms of
the ratio of the stoichicmetric amount required for formation of calcium
fluoride. Spent liquor and reagent were mixed initially and allowed to
settle, as was previously done with the liming of process wastewater.
The superiority of calcium chloride over lime is evident and can be
attributed to the solubility of the chloride. The low fluoride contents
achieved with 10 times the stoichiometric lime requirement probably result
from adsorption of fluoride by the insoluble lime, but sucft quantities are
not economically feasible.
Optimal Regeneration of Alumina
During the early phases of this program, process parameters were modeled
after those in use at the two drinking water installations visited, and the
activated alumina was regenerated with three volumes of 1-percent sodium
hydroxide solution per volume of alumina. Attempts were made to decrease
caustic soda requirements, because this chemical is the largest single
item of cost in the process. Most attempts resulted in incomplete regenera-
tion of the alumina, but a scheme was conceived that accomplishes successful
regeneration with one-sixth less caustic. The details of this process step
are given in Section 6 of this report. Briefly, the new procedure consists
of using all caustic solutions twice. Chronologically, 5,0 volumes of once-
used, 1-percent caustic solution are recycled for 30 min through the spent
alumina. The stream is then drained off the alumina, and half of it is
treated to precipitate fluoride as calcium fluoride. The remaining caustic
soda solution is made up to 5.0 volumes by addition of 2.5 volumes of fresh
1-percent solution, which is also recycled continuously through the partially
regenerated alumina for 30 min. In the next regeneration cycle, the latter
5.0 volumes of caustic are recycled-through the spent alumina.
Further optimization is believed possible using similar techniques. For
instance, only 2 volumes of partially spent solution would, be discarded from
a prior 4-, 5-, or 6-volume caustic batch. By so doing, still another
fractional saving of caustic might be possible.
A series of three runs was made using soda ash (i.e., sodium carbonate,
Na2CO3) as a regenerant in lieu of caustic soda. During the third cycle,
55+ volumes of 62 ppm F- wastewater were passed through the alumina media
before the cumulative effluent concentration was 4.8 ppm F~. This corres-
ponds to an alumina capacity of ~3 kg F~/m3 (-0.2 Ib F~/ft3) alumina. The
use of the less expensive soda ash instead of caustic soda is promising, but
since the requirement for caustic soda has already been reduced once and
further reduction is probable, it would have been fruitless to pursue a new
process direction at that time. Moreover, the presence of carbonate might
preclude the use of regenerant in neutralizing the process wastewater because
of foaming.
11
-------
OJ
a
s
1
o
1
•H
§
"
1
£
1 — •
1
•r.
4.
(1
$
O
O
•
O ^O 00
rH m "
o
o
• vo **o
oo M) m
o
o
• <3< O
o
"I co ro
rH <*D VO
0
O
. -5T 00
rH 00 W5
O
0
o ro oo
H 00 CN
o
o o m
. 00 CN]
00 rH rH
O
O O O
• ^ oo
OJ rH rH
o o o
• in "*
rH H rH
fj £_|
o i S i S
rH (I)
8 S
W 0)
O SH •
(d oo
S ^ rH
%
O ^ |
W W
ft ft °^
_d cd
r< ei O
* +
12
-------
Process Wastewater Pretreatmant
Since calcium chloride precipitates fluoride as calcium fluoride in
spent regenerant liquor, it should perform equally well in removing fluoride
from the process wastewater. The wastewaters are quite acid, however (pH
2.0 to 3.0), and must be neutralized before calcium fluoride would be
expected to precipitate rapidly.
ISfeutralization may be accomplished using the defluoridated, spent regen-
erant liquor, which is about pH 12 after defluoridation. Mien process waste-
water from the Feldspar Corporation's Connecticut plant (125 ppm F-) was
neutralized with spent regenerant liquor (68 ppm F-), the volumetric ratio of
wastewater to regenerant was 137:1. The initial pH of the wastewater was 3.8
and that of the regenerant liquor was 11.8. This neutralized wastewater was
treated further by the addition of powdered calcium chloride for a calcium/
fluoride equivalence of 4/1. The resulting pH of the solution was 6.7, and
there was little or no precipitation of calcium fluoride at that point.
However, when the pH was adjusted to 7.3 using less than 0.1 slaked lime
[Ca(OH)2] per liter of solution, the fluoride was precipitated as calcium
fluoride.
Additional experiments with calcium chloride and lime shoved similar
results. Fluoride does not precipitate appreciably until the solution is
made alkaline, but one can conclude that calcium chloride is more effective
than lime as the precipitant because of its greater solubility. Also, the
nature (composition as related to source, age, etc.) of the process waste-
water appears to have some effect on the efficacy of fluoride precipitation.
To obtain design data for pretreatment, a "worst case" was studied in
which lime alone was used to remove the fluoride and neutralize the waste-
water^ (the limited solubility of the lime would give the lowest rate of
fluoride precipitation and therefore the largest size for a clarifier or
settling tank). Connecticut plant wastewater was limed in the laboratory at
room temperature using 0.4 g/liter (0.0033 Ib/gal) calcium hydroxide
[Ca(OH)2]. Two cases were studied: one with agitation and the other without
agitation except for initial mixing. Fluoride ion was monitored with the
Orion 407A apparatus.
As shown in Figure 2, reduction of fluoride was rapid initially, but was
not complete for 6 to 8 hr. About 50 percent, or slightly more, of the
fluoride was precipitated during the first 30 min. Surprisingly, the un-
agitated case appears superior to the continuously agitated case.
Optimization of the use of spent regenerant in combination with calcium
chloride would be better studied in a pilot plant environrrent where rate
studies can be made as a function of residence time and where age of the
process wastewater is more nearly representative of actual conditions.
Acid Washing of Activated Alumina
A number of acid washing techniques were used for neutralizing the
activated alumina after caustic regeneration. These included the use of
13
-------
I
•a
CC UJ
= 9
f— Q
< <
IX 1M
Ul =
Q. I
s 2
UJ :s
i- 0
•3
•a
O
1
O
14
-------
hydrochloric acid and sulfuric acid solutions up to 2 percent. Neither was
technically superior to the other, so dilute sulfuric acid was selected
because:
— It is cheaper,
— In concentrated form, it can be handled in carbon steel equipment,
and
— Feldspar plants generally use sulfuric acid in o'bher steps
and already have acid storage facilities.
Washing time and wash water velocity appear to be more important than
acid strength in neutralizing the beds. Accordingly, the following proce-
dure is recommended: backwash regenerated activated alumina for no less than
45 min using wash water maintained at pH 6.8 (+0; -0.4), at a space velocity
of 0.00417 sec"1 (4 min residence time). Washing cycles in excess of 45 min
can do no harm. Wash water can be made up by diluting 5 volumes of neutral
water with 1 volume of 1-percent sulfuric acid.
Activated Alumina Testing
Initial experimental work using activated alumina was based on the
Alcoa F-l product (28-38 mesh), since this was the alumina used at the
Desert Center, California, and Bartlett, Texas, installations. Two other
coranercially available activated alumina products were oljtained from the
Filtrol Corporation, Jackson, Mississippi and from the Kaiser Aluminum and
Chemical Corporation, Baton Rouge, Louisiana. Both products were screened
for use in this process but proved inferior to the Alcoa product. No
further effort was expended in this direction.
15
-------
SECTION 6
SCMYI&RY OF PROCESS FOR REMOVING FLUORIDE FROM FELDSPAR PROCESS WASTEWATER
1. Process wastewater at pH 2.0 to 3.0 is treated with three chemicals—
line, calcium chloride and a once-used alumina regenerant (1-percent sodium
hydroxide solution in an in-line mixer). The result is an increase in pH to
about 7.3 and formation of calcium fluoride precipitate.
2. Calcium fluoride (CaF2) settles out as a slurry at the bottom of a
settling tank. Residence time in the tank is a minimum of 30 to 35 min and
preferably 1 hr. Periodically calcium fluoride slurry is pumped from the
bottom of the tank to a filter.
3. Settling tank overflow passes through a column of activated alumina
and then leaves the plant, proceeding either to the river or to mixing with
other effluents. This water can also be recycled in the feldspar processing
plant, as it is of good quality.
4. While one of the two columns of activated alumina is processing
wastewater, the other is being regenerated. About 48 volumes of pretreated
wastewater at 60 ppm fluoride will saturate 1 volume of activated alumina.
5. When being regenerated, a column of activated alumina is first back-
washed with 5 volumes of previously used 1-percent sodium hydroxide (NaOH)
solution per volume of activated alumina for 30 min with a space velocity of
0.00417 seer1 (residence tune of 4.0 min).
6. Then 2.5 volumes of the used sodium hydroxide solution are pumped to
a settling tank where they are mixed with 1.5 moles of calcium chloride (in
aqueous solution) per mole of fluoride (F~). After the calcium fluoride has
settled out, the clear supernatant sodium hydroxide solution is used either
to neutralize incoming fluoride-containing process wastewater (see step 1)
or to neutralize other non-fluoride plant effluents before discharge. Calcium
fluoride slurries from the bottom of the settling tank or clarifier are
pumped to the same filter as in step 2 for concentration of the calcium
fluoride precipitate. The recovered calcium fluoride is either sold and
shipped, or stored in a secured landfill.
7. The column of activated alumina is then further backwashed with 5
volumes of 1-percent sodium hydroxide solution for 30 min with a space
velocity of 0.00417 sec-1 (4.0 min residence time). Half of this solution
was left over from the prior regeneration cycle, and half is fresh make-up
solution. After this backwashing, the entire 5 volumes of partially spent
hydroxide solution are drained off the alumina and retained; 2.5 volumes of it
16
-------
are used in the initial regeneration step of the next cycle (step 5) .
?-- f-
in
feldspar
defluoridatim process is given
17
-------
RIVER WATER
PROCESS
WaSTEW
300 torn
1N-UNE MIXER
1
1
1
1
1
FOR&BpH VWTER
t
3> WASTEWATER
ni... p.(
SET
TUNG
TANK
-^^
ISLURRY
r
!
5
r
FOR IV.NoOH
r__J
I
£1
• 1
H
|l
1
'
— —
(ACTIVATED
«-4 ALUMINA ^»
( COLUMNS
t
T
— 1
1 t
t 4
kn^J %m~ml
-»• PURGE
4
Ogom _ ,__
40ppmF i
pHILB j
-7l
FIL-
Cc
TO DK
^~~ •
SLURRY
!
;
i
.
ft
POSAL
• 1,
1
1
1
TREATED WATTFWATPR
SOOgpm SppmF pA7Z
DEFLUORID&TED
IN-LINE MIXER
REGENER
pH
— -. SETTLING
TANK
ANT
>ppmF~
IIB
1
THICKENED CoF? SLURRY 1
TO RECYCLE
OR RIVER
TO ACID STREAMS
FROM FELDSPAR
PLANT
• CONTINUOUS OPERATION
• INTERMITTENT OPERATION DURING REGENERATION
Figure 3. Feldspar wastewater defluoridation process flow diagram.
18
-------
SECTION 7
ECONOMICS
An economic evaluation has been prepared for the feldspar wastewater
defluoridation process. Capital and operating costs for the proposed pro-
cess are summarized in Table 2, and equipmsnt costs are shown, in Table 3.
For a 163,000 kkg/year (180,000 ton/year) plant, the initial capital
expenditure was estimated to be about $200,000. The projected unit cost -for
removing at least 92 percent of the soluble fluoride from feldspar flotation
process wastewaters using the four-step process outlined here is $1.03/kkg
($0.93/ton). :
Some reductions in operating costs appear possible wrbh process optimi-
zation, particularly those associated with the caustic registration of
activated alumina. Other possible economies include:
—Substitution of calcium chloride for the more expensive lime in the
initial treatment of process wastewater;
I
—Partial recycling of spent regenerant liquor during alumina
regeneration to conserve sodium hydroxide;
—Recycling of once-used regenerant liquor to neutralize incoming low-
pH feldspar process wastewater; and
—Use of excess treated alumina regenerant liquor (sodium hydroxide
solution) as partial or whole replacement of presenliy used lime
for pH adjustment of the total plant effluent.
Cost reductions that could result from these steps have not been calculated,
nor have those that might accrue from the sale of the calcium fluoride cake.
19
-------
TABLE 2. ECONOMICS SUMMARY
Item
Cost
Cost/kkg feldspar
(Cost/ton feldspar)
Estimated initial capital expenditure
for fluoride removal from process
wastewater $200,000
Estimated annual operating costs for
fluoride removal:
[Capital recovery], 10% interest rate,
10-year depreciation 33,400
Maintenance, 6% of capital/year 12,000
Taxes and insurance, 2% of capital 4,000
Power, 46 hp @ 2.2£/kwh 5,700
Labor, 1 operator/shift,
$300/week each including
supervision, overhead, etc. 47,000
Chemicals:
Activated alumina \
if*1"8,.. * ( $0.4010/kkg
Caustic soda > ($0.3637/ton) 65/0oo
Sulf uric acid I
Calcium chloride /
$0.2046 (0.1856)
0.0735 (0.0667)
0.0245 (0.0222)
0.0346 (0.0314)
0.2867 (0.2600)
0.0252
0.0198
0.2862
0.0119
0.0578
(0.0229)
(0.0180)
(0.2596)
(0.0108)
(0.0524)
~$1.03 p$0.93)
Total
$168,000
Based on a production rate of 163,000 kkg/year (180,000 ton/year)
of feldspar.
20
-------
TABLE 3. ESTIMATED COSTS FOR MAJOR EQUIPMENT ITEMS
Number
required
1
1
2
1
1
1
1
1
1
5
1
2
2
Description
In-line mixer or mixing
1,000-gal
tank (agitated) ,
P.W.W. settling tank, slanted bottom
w/bottom slurry discharge, 10,500-gal
Columns, activated alumina 125 ft3 media
Filter, vacuum, disc
Tank, storage, 50% NaOH,
Tank, recirculation, 1%
Tank, recirculation, 6.8
Tank, storage, 40% CaCl2
Tank, settling, 1% NaOH,
w/bottom outlet
Pumps, 300-gpm
Feeder, dry lime
Pumps, 10-gpm
Pumps, 5-gpm
4,000-gal
NaOH, 5,000-gal
% pH, 5,000-gal
, 11,000-gal
6,000-gal
Estimated cost
installed
$ 9,000
16,500
48,400
14,600
8,200
10,000
10,000
17,000
11,200
6,600
3,000
2,300
2,300
Total
$159,000
21
-------
BIBLIOGRAPHY
Aluminum Company of America. Alcoa Activated Aluminas. Pittsburgh, Penn-
sylvania.
Anon. New Fluoride Removal Msthod Cuts Costs. Engineering News Record,
June 12, 1952.
Colorado School of Mines Research Institute. Capital and Operating Cost
of a Suggested Process for the Removal of Fluoride Ion from Tailings
Water. Golden, Colorado. 1973.
Lee, J.A. Cheap Way to Remove Fluorides from Water. Chemical Engineering,
July 1952.
linke, William, ed. Solubilities of Inorganic and Metalorganic Compounds.
4th ed. American Chemical Society, Washington, D.C., 1958.
Maier, F.J. Partial Defluoridation of Water. Public Vforks Magazine,
November 1960.
Parker, C. L. and Fong, C. C. Fluoride Removal: Technology and Cost Esti-
mates. Industrial Wastes, Nbv/Dec 1975.
Savinelli, E.A. and Black, A.P. Defluoridation of Water with Activated
Alumina. Journal A.W.W.A., January 1958.
Zabban, W. and Helwick, R. 30th Annual Purdue Industrial Waste Conference.
West Lafayette Indiana. May 1975.
22
-------
•<* m m oj t*- m
r^oQooor**o!ocoo ooo
r-t ,(N ^J* rH r—(
cncncncncncotocncncn tn en co
i 04 CU P4 C G C C C C C g .5 do*
ico coco wwtnww cowcn <*> OP dP aP
! Qji-lr-tiHiHtHkO^OUSkOVOiOOOOOO
m m ro m vo *
O
as1
-------
APPENDIX B. ANALYTICAL METHODS
Determination of Fluoride Con<3entration in Aqueous Solutions
Application:
Principle:
Apparatus:
Reagents:
Procedure:
Note:
Ihis direct readout method for fluoride determination was
followed so that activated-alumina-treated waters could be
analyzed for fluoride concentration rapidly.
Direct readout of fluoride as free species is possible using
the logarithmic scale of an Orion 407A meter after the scale
has been calibrated with known standards.
Orion 407A meter
Fluoride-specific electrode
total ionic strength adjustment buffer (TISAB)
Make fluoride standards of 1, 10, and 100 ppra and dilute 1:1
with TISAB. Place electrode in 10-ppm standard and turn
calibration control to get a center-scale reading.
Rinse electrodes, blot dry and place in the 100-ppm standard.
Turn the temperature compensator until needle points to right-
hand full scale line.
Rinse electrodes, blot dry, place in unknown solution that has
been diluted 1:1 with TISAB. Multiply the reading on the
logarithmic scale by the center-scale value to determine
the unknown concentration.
With the various standards, it is possible to set center scale
at 1, 10, or 100 ppm, depending on the fluoride concentration
range of unknowns.
Determination of Fluoride Concentration of Highly Alkaline Streams
Containing muminum 'long ' (Al-f-H-)
Application:
Principle:
Apparatus:
Reagents:
Regenerant streams, typically 1% NaOH, were found to contain
~360 ppm aluminum. A1+++ interferes with accurate fluoride
readings obtained using fluoride-specific electrode.
TISAB contains 4 g/liter 1,2-cyclo-dinitrilo tetra-acetic acid
(CDTA), which preferentially chelates small amounts of A1+++.
By increasing CDTA to ~45 g/liter in TISAB, the Al-m- inter-
ference limit is increased to 38 ppm.
Orion 407 meter
Fluoride-specific electrode
Modified TISAB (modified in-house)
24
-------
Procedure:
Dilute samples 1:10 with distilled water, and again 1:1 with
modified TISAB. Compare millivolt reading from expanded
millivolt scale to standard curve obtained by the same method.
Infrared Spectroscopic Analysis of Freon Extracts of Process Wastewater
Application:
Principle:
Apparatus:
Reagents:
Procedure:
Run No.
1
2
3
4
5
6
7
9
10
11
One or more of the organic flotation reagents routinely used
in the Spruce Pine facility, or an unsuspected organic
material, was blinding the activated alumina and preventing
alumina regeneration.
A great many organic compounds have characteristic infrared
spectroscopic profiles, and such profiles can be used to
identify organics.
Separatory funnels, 125-ml
Beckman Acculab IR
Two matched adsorption cells, 1-an, quartz
1- x 30-an glass columns (alumina)
1- x 30-cxn glass columns (activated carbon)
Freon 113
To a 125-ml separatory funnel, add 10 ml of sample and 10 ml
of Freon 113. Shake well and let stand for 10 min. (Ratio of
10 ml sample/10 ml Freon may be insufficient if organic concen-
tration is too minute). Drain off Freon layer into one of the
matched cells. Fill the other (reference) cell with Freon.
Place reference cell and sample cell in respective positions
in the Acculab. Scan from 2.5 microns (4,000 cm.-1), to 5
microns (2,000 cm,-1).
Gompounds or Materials Tested
Polypropylene glycol
Sulfonated white oil
Alcohol type frother
No. 4 fuel oil
Tap water
Freon extract of process wastewater
Freon extract of process wastewater after lime treatment (pH
adjustment)
(a) Freon extract of unused activated carbon (Filtrasorb 400)
(b) Freon passed through activated carbon (Filtrasorb 400)
after process wastewater
Freon passed through activated carbon after process wastewater
(a) Process wastewater extracted with Freon after passage
through activated alumina
(b) Freon passed through activated alumina eifter process
wastewater
Freon passed through alumina after process wastewater
25
-------
Determination of Trace amounts of amines in Aqueous Solutions
Application:
Outline:
This irethod can be used to determine trace amounts of amine
in aqueous solutions to a level as low as 0.5 ppm.
The amine complexes with methyl orange to form a yellow solu-
tion which is extracted into chloroform and measured in the
visible range on a spectrophotometer at 420 mu. The amount of
amine in the sample is determined by comparing it with stand,
ards run by the same method.
Specifications A.
and
Directions:
B.
Apparatus
1. Separatory funnels, 250-ml
2. Beckman DK-2 spectrophotometer or equivalent
3. Ttoo matched absorption cells, 1-cm, quartz
Reagents
1. Methyl orange - acid buffer reagent. Dissolve 0.1 g
of methyl orange in distilled water. Dissolve 29.6 g
of sodium acetate trihydrate and 50 g of potassium
chloride in water. Combine the two solutions, add
100 ml of glacial acetic acid, and dilute to 500 ml
with water
2. Chloroform AR
3. Isopropyl alcohol AR
4. 1 percent NaOH solution.
C. Procedure
Titrate test solution with 1 percent NaOH to a pH of 7.0.
Note amount of NaOH required for computation of dilution
factor. (This step will negate the possible effect of the
HP acidity of the solution on the indicator.)
In a 250-ml separatory funnel, add 500 ml of the sample to
be tested. Add 5 ml of the methyl orange reagent, shake,
and set aside for 10 mm. Add 20 ml of chloroform and
shake for 5 min.
Allow to separate for 3 min. Drain off the chloroform
layer, add 0.5 ml of isopropanol, and measure absorbance
(A) immediately on a spectrophotometer at 420 mu. It is
important to be consistent in the tinning because the color
is very unstable.
NOTE: If necessary, dilute extracted layer with chloroform before reading.
Make standard solutions of the amine to be measured (1.0,
3.0, 5.0 and 7.0 ppm of amine). Run these by the above
procedure and plot a standard curve of the absorbances (A).
The ppm of amine in the unknown can be read from the curve
and multiplied by the dilution factor.
26
-------
APPENDIX C. TRIP REPORTS
TO: R. Shaver, R. Smith, E. Juergens
File, Feldspar Corp. (2 copies)
FROM: I. Frankel
Date: 8 June 1976
Subject: Trip Report, Visit to Alumina Treatment Plant, Riverside County
Service Area 51, Desert Center, California 92239
5-27-76
Contact: Mr. Robert E. Anderson
Sr. Public ftforks Operator
Phone: (714) 227-3203
The writer was with Mr. Anderson during a 4- to 5-hr period during
which the operation and equipment of the alumina treatment plant was discussed.
The plant itself was inspected, as was the sewage treatment system. Mr.
Anderson had prepared a two-page process description in adveance of my visit,
and this provided an excellent basis for our further discussion. This
material is reproduced in Figure C-l.
Desert Center is a small southern California cormunity on Interstate 10,
about 89 km (55 mi) west of Arizona. The area is desert and sparsely popu-
lated. The community is divided into two parts: A business district with
stores, restaurants, and service stations at the exit from 1-10, and a
pleasant residential area about 2.4 km (1.5 mi) north of the highway. The
latter is irrigated and therefore is grassy with some trees., A few farms
nearby grow dates and citrus fruit by means of irrigation.
The residential area was planned and financed by the Kaiiser Steel Co. to
acconmodate mine supervisory and management personnel and their families.
The nearby Iron Mountain mine is a large source of iron ore from Kaiser's
western steel operations. Water for the community is obtained from deep
wells and is of potable quality except for fluoride content.
The incoming well water has a fluoride content of 7.5 ppm, which is re-
duced to 0.8-1.2 ppm in the subject plant. The lower fluori.de value is
obtained in the summer, and the higher value in winter. Thtis the human in-
take is adjusted to an average year-round value of 1.0 ppm.
Fluoride removal is accomplished by passing the well water through a bed
of activated alumina (media). During the period before the need for media
regeneration, the F~ concentration in output water is as low as 0.2 to 0.3
ppm. Bbwever, as indicated above, the average value of distributed water is
maintained at 0.8 to 1.2 ppm. In 7 years of operation there has been no
requirement for replacement of media. Media addition to the. extent of 15 to
20 percent has taken place twice during that period.
27
-------
Operation of the defluoridation system is sirtple and consists of four
cycles, as follows:
(a) Gravity flow of pH adjusted water (5.0-6.0) through one or both
parallel madia tanks. Flow can be continuous, or part time, depending on the
water level in the reservoir.
(b) Mien the media in tank begin to indicate breakthrough (by an
increasing level of F- in the product water), the water flow is reversed and
the madia bed is expanded (to approximately 200 percent of its settled
volume). After washing out silt and iron oxide, the bed is aerated for a bit
less than 5 minutes. The aeration tends to enhance or help sustain media
activity toward F~ ion removal.
(c) Regeneration is accomplished by downflow of caustic solution during
a period of 30 minutes.
(d) She bed is then neutralized by gravity feeding of low pH (2.5-3.0)
water. Excess exposure to acid waters, or excess acidity should be avoided,
else the madia itself will become acidic and will not remove fluoride without
a reverse neutralization. Specific operating and performance data for the
above steps are given on sheets attached to this report.
At Desert Center, the itedia tanks are constructed of steel, and the acid
storage tank is made of fiberglass-lined steel tanks. Underground piping is
concrete-lined steel. None of these materials would be recommended for these
applications in a new system. The media tanks should be fiberglass-lined
steel, the acid tanks should be carbon steel, and no concrete-lined pipes
should be used.
Fluoride content of treated and untreated water is determined with a
Hach DR/2 spectrophotometer. A water sample is mixed with a reagent, placed
in a glass cell, inserted into the unit, and the F--ion concentration^is read
directly. Any given analysis takes only a minute or less. When originally
constructed, this plant was highly instrumented and automated, and a Beckman
Model 900 fluoride ion analyzer was purchased for process control. Ihis
instrument was probably one-of-a-kind, and it was not accurate. It was _never
used except during the start-up tests. Other initial installed automation
devices were also discarded, modified, or brought into use over a period of
years, as opposed to the initial design mode.
Other water quality analyses are performed by Mr. Anderson or his
assistant, as needed. A copy of their operating log sheet is attached.
Furthermore, county health officials check water quality once per month and
arrive unannounced to sample onsite.
Another F~-ion removal plant was built in 1973 in a suburban Tucson,
Arizona, development. Its capacity is about 25 percent of the Desert Center
unit, but its design was based on and benefited by the experiences of the
Desert Center operation, particularly with respect to materials of construc-
tion. It is alleged to be operating satisfactorily, and it also uses Alcoa
F-l, 28 to 48 mesh media.
28
-------
Present costs for fluoride removal at Desert Center are 5£/l,000 liters
(19<:/1,000 gal). Ihe costs include labor, chemicals, and power, but no
depreciation. A breakdown of this cost will be provided by Mr. Anderson in
the near future. Labor application is approximately equivalent to one-third
of a man-day. Oily two men are used for operation and maintenance of the
deep wells, the fluoride removal plant, the sewage plant, and the water dis»
tribution and sewage collection systems. Heavy maintenance is performed by
outside contractors.
Figures C-2, C-3 and C-4 show, respectively, a schematic flowsheet of
the Desert Center defluoridation plant, a crude sketch of i±e construction of
a media tank, and a sketch of the plastic spargers mounted on the steel plate
on the bottom of the media bed. These spargers were purchcised from the Eimco
Corporation, Salt Lake City, and they are quite effective in allowing de-
fluoridated water to be collected without loss of media and in permitting
good backwashing and aeration during the initial stages of media cleanup and
regeneration.
Desert Center, California, Operating arid Design Parameters
1. Defluoridation
a. Two tanks in parallel flow, used sequentially or at the same time.
b. 17 m3 (600 ft3) media (activated alumina) in each tank.
Co Each 0.028 m3(ft3) of media will treat about 16,000 liters (4,200 gal)
of water at a rate of -134 liters/m3 rain (~1 gal/ft3-min). Removal of
F- is 3,999 g/m3 (1,700 g/ft3) (maximum of 130 g (2,000 gr), which is
equivalent to 3.9 kg F-/m.3 (0.243 Ib F-/ft3).
d. Media is 24 to 48 mesh Alcoa F-l activated alumina.
e. Incoming water is pretreated to 5.0 to 6.0 pH with controlled input
of sulfuric acid.
f. Defluoridation is not continuous; the system is operated in line with
water demand and reservoir level.
g. Other than pH control, incoming well water is not pretreated. After
defluoridation, the only post-treatment of the water is the use of
NaOH to bring the pH to ~7.0.
2. Backwashing and Aeration
a. Start backwashing at 1,900 liters per min (500 gpm). Increase to
3,800 liters per min (1,000 gpm) for several minutes. Decrease the
rate to 1,900 liters per min (500 gpm) and aerate for ~5 min.
b. Total backwashing water is 76 to 114 m3 (20,000 to 30,000 gal).
3. Regeneration
a. Regenerating fluid is a 1-percent solution of caustic soda (NaOH).
b. Use 3 bed volumes ( 5 m3 or 1,800 ft3) of regenerant over a period of
30 min (S.V. = 0.00167 sec-1, or residence time of 10 min).
29
-------
LAKE TAMARISK DEFLOURIDATION PLANT
RIVERSIDE COUNTY SERVICE AREA #51
DESERT CENTER, CALIF.
TANK #_
RUN I
OPERATING LOG
DATE
PRE-HEGENERATE RINSE TO LAND
Meter-end =_
Meter-start , =~
Total Gals. FTO=
REGENERATE SOLUTION
Mater-end
Meter-start
Voluma Caustic
(50% NaOH)
NEUTRALIZATION RINSE TO LAND
Tims-end -
Tine-start =
Total Min. x gpm = GAL.
Time-end =
Time-start?1
Total Min.
Fluoride Av.
x gpn = GAL.
TOTAL ViASTEWATER SUMMARY
Total Land
Total Ponds
Total Waste"
Gals.
'Gals.
' Gals.
SERVICE TO RESEVOIR
Meter-end
Mater-start
Quantity Treated_
Gals.
Total Water used
Gals.
Percent Waste
TREATED WATER PROGRESSIVE LOG
rai mi (0 (Dl
DATE
meter
reading
1000
Gals.
treated
cumulate
column
(A)
fluoride
reading
average
fluoride
in
column
(A)X(D)
W\ (G)
column
OB)
aver.
fluoride
(m(B)
oper.
Figure C-l. Process description.
30
-------
1
CYCLONE
,, .,. currnv
DEEP WALL
\*
l
SETTLING
TANK
SANO&
JS *^
H2S04
GRAVEL
H20 @ 5.0 - 6.0 pH
LAND
'FILL
NaOH
H20 @ 7.0 pH
TANK
RESERVOIR
(BELOW GRADE)
Figure C-2. Schematic flow plan for defluoridation of
Desert Center California, Potable water.
31
-------
' V • SHAPED " LAUNDER
.- BED EXPANSION
.., SPACE
8" PIPE
Figure C-3. Sketch of media tank design.
32
-------
(0
I
1" PIPE THREAD
i" I.D., THIN-WALL
PLASTIC TUBING
Figure C-4. ELMCO plastic sparger (probably nylon).
33
-------
LAKE TAMARISK DEFLUORIDATION PLANT
DESERT CENTER, CALIF.
For many years much research and experimentation has been done to find means
of removing excessive fluoride from water from domestic use.
Many types of processes for the removal of fluoride have been established;
however, most of them have not proven practial.
After the installation of the Lake Tamarisk fluoride removal plant at Desert
Center, California, many months of trials and errors of experimentation
resulted in an established process that ended the search for a simple and
practical method of fluoride removal from domestic water.
The Lake Tamarisk pland has now been in operation for over seven years and is
operating far above original expectations.
During a normal, average process run, each cubic foot of media will treat
4200 gallons of raw water.
The actual grains of fluoride remover per cubic foot of media will average
1700 grains with a high in excess of 2,000 grains per cubic foot of media.
The labor and chemical costs (including pumping costs) for producing water at
the Lake Tamarisk Defluoridation Plant is nineteen cents per 1000 gallons.
The sole purpose of the Lake Tamarisk Defluoridation Plant is to change the
fluoride content of well water with a fluoride level of 7.5,ppm to water with
an acceptable fluoride level of 0.8 ppm or less.
The actual chemistry of removal of fluoride in this process has not been com-
pletely defined. However, it is believed to be a combination of ion exchange
and adsorption.
The fluoride removal by this process is accomplished by passing raw well
water through a media bed of activated alumina, (supplied by Alcoa)
As the water passes through the activated alumina bed, the fluoride ion is
exchanged and trapped in the media bed. The resulting effluent is water
with an acceptable fluoride level of 0.8 ppm or less.
Figure C-5.
Description of process at Desert Center, California,
provided by Robert E. Anderson.
34
-------
The flow through the bed can continue until such time as this alumina media
is saturated to a point where there is not sufficient ion exchange and/or
adsorption and the resulting effluent is above acceptable levels. At this
point, the loaded media bed necessitates regeneration.
The regeneration process is preceded by first cleansing the media bed of
foreign matter such as fine silts and iron oxides that accumulate on top
of the bed during the process run. This is done by an up-flow rinse of raw
water through the bed. No addition of chemicals is needed at this point.
Regeneration consists of passing a solution of Sodium Hydroxide (NaOH)
through the media bed for a required length of time. During this regenerate
period the release of fluoride by the activated alumina is .immediate. However,
to remove the released fluoride from the bed requires a bed neutralization
rinse.
The neutralization rinse continues until such time as the effluent fluoride
level has dropped to a required range. The effluent flow can then be turned
into the potable water supply, (reserviors, etc.)
After the regeneration process and during the process run, PH control of the
raw water is very important. This is accomplished by installing a chemical
injection point in the raw water influent line prior to entering the media.
The pH control of the treated water effluent may or may not be a factor.
In any event, this also is easily controlled by injection of the proper
chemical in the treated water effluent prior to entering storage reserviors,
etc.
Raw water pH control is continued through out the process run -until the acti-
vated alumina bed is again saturated with fluoride and the fluoride removal
is insufficient. At this point the production cycle is ended and another
begins.
i
These cycles consist of four parts, namely: Pre-regenerate rinse, passing
of the regenerate fluid through the media bed, bed neutralization rinse and
the process run.
Submitted by:
Robert E. Anderson
P.O. Box 495
Desert Center, California
Figure C-5. (continued)
35
-------
4. Msdia Neutralization
a. Media bed is washed with water at 2.5 to 3.0 pH at 1,100 to 1,300
liters per min (300 to 400 gpm) for 1.5 hr (S.V. = 0.0011 to 0.0014
sec"1, or residence time of 15 to 11.2 min).
b. Acid vised to lower pH is sulfuric
c. Secondary washing is with water at ~4.0 pH, to extent of ~380 m3
(-100,000 gal).
TO: R. Shaver, R. Smith, E. Juergens
FROM: I. Frankel
Date: 14 June 1976
Subject: Trip Report, Visit to Alumina Treatment Plant, Bartlett, Texas
5/25/76
Contact: Mr. Billy White
Utilities Superintendent
(817) 527-3557
Bartlett, Texas, is a small community, approximately 72 km (45 mi) south
of Waco. The community's water supply is obtained from deep wells and is
potable as pumped except for fluoride content. The well water presently
obtained has a fluoride content of only 3.0 ppm of F~. However, when the
alumina treating plant was first constructed (1949) , the well in use at that
time produced water having an 8 ppm. F- content.
The treatment is quite simple. Untreated well water passes by gravity
through a bed of media (Alcoa F-l activated alumina, 28 to 48 mesh) . The
fluoride content is reduced to an average of ~1 ppm of F~ and is pumped to
storage without additional treatment.
The treatment tank is constructed of steel but is coated on the inside
with a durable organic coating ("Vinoline" from 1NEMEC Co., North Kansas
City, Missouri). The coating is now 4 years old and is generally holding up
well; it is alleged to have an 8-year life. The treatment tank is vertical
(2.9 m or 9 '6" high) and cylindrical (3.35 m or 11' in diameter) , and thus
has a total volume of 26 nr (6,870 gal) . Tank drainage or backflow is handled
by an X pipe grid on the bottom. This is covered by 30 cm (12 in) of graded
gravel, then by an alleged 1.52- to 1.60-m or 5- to 5%-ft bed of media. This
depth is equivalent to about 1.9 m3 (500 ft3) of media, but I do not believe
that the bed was more than ~0.9 m (~3 ft) deep on the day of my visit. There
is about 0.9 m (3 ft) of free space above the media and an X-shaped launder
that discharges backwash water. The tank has no cover, although it and its
accessory equipment are in a fully enclosed building.
The plant handles about 680 m3 (180,000 gal) of water per day; the
media is regenerated every 378.5 m3 (10 6 gal) , and regeneration occurs every
4 days in summer and about once a week in winter. About 227 kg (500 Ib) or
36
-------
~2.3 m (~9»1 ft3) of media is added per year as make-up for miscellaneous
losses.
Regeneration is accomplished by dissolving 227 liters (60 gal) of 50-
percent caustic soda to 19 m3 (5,000 gal) of solution (0.9-percent) and flow-
ing this selection by gravity through the media. The bed is .then backwashed
using 76 m (20,000 gal) of well water. Washing is continued using 151 m3
(40,000 gal) on a gravity flow basis. Acid washing by gravity flow removes
final traces of free caustic [295 kg or 650 Ib of 93-percent sulfuric acid is
added to sufficient water to make 1,830 kg or 4,030 Ib (-.1,665 liters or -440
gal) of 15-percent acid, which in turn is diluted to 19 m3 or 5,000 gal of
~1.5-percent acid]. The acid is washed out with 151 m3 (40,000 gal) of well
water by gravity flow at a rate of -1.14 m3 per min (-300 gpm). All waste
streams are sewered, and the sewage plant overflows to a normally dry river
bed.
Analyses of water samples were allegedly done with a Klett-Summerson
spectre-photometer. However, this instrument was so dusty and so obviously
unusable that the writer doubts that routine analyses are performed. Infre-
quent state water analyses are probably the basis for water quality judgments.
Labor application on the plant is stated to be about 0.5 man-days per
regeneration. Costs of operating the plant were 2<=/l,000 liters (8£/l,000
C7dI3 Y*G a^**T/™\ r\1 1^~ 4—T"\irNT T ^"fciVN •t^x'x-l— •wij—*.* T*"M«
-------
Feldspar Corporation currently employs an Orion specific ion meter
(Model Kb. 407A) in conjunction with a fluoride confoination electrode (Orion
Model No. 96-09) for fluoride determinations. Double concentration total
ionic strength adjustment buffer (TTSAB) is utilized to negate possible
interference caused by, the presence of aluminum (A1+3) and/or ferric (Fe+3)
ions.
The influent to feldspar flotation conditioning is approximately 1,135
liters per min C300 gpm). The pH at this point is 2.5 to 3.0, and condition-
ing time is about 2 min. Conditioning reagents added at this point are HF,
oleylamine acetate (Alamac IIC-General Mills), polypropylene glycol, and a
trace amount of sulfuric acid. Upon occasion, fuel oil is added at pump
boxes to reduce clogging throughout the system. Sulfonated white oil may be
present as a result of carryover from the previous flotation process.
Fluoride concentration is ~100 to 150 ppm. • Feldspar Corporation indicated
that further analysis and characterization such as suspended solids, total
solids, etc. would be performed and results would be forwarded to Versar.
The effluent from feldspar flotation is separated into two portions,
each with a flow of 570 liters per min (150 gpm)—sand underflow and feldspar
float. After respective sand and feldspar removal, the effluents are recom-
bined. The recombined stream is expected to contain some of the above
mentioned conditioning reagents and a significant amount of floating/suspended
solids—roost of which are 80 to 200 mesh feldspar.
Total plant effluents are currently combined and treated with lime
(hydrated lime, Williams lime Co.) until a pH of 9.5 to 10 is reached. The
purpose of lime treatment of total plant effluent is to elevate the pH to
the desired discharge level and to act as an aid to flocculation. At the
point of discharge, the flow is ~7,600 liters per min (~2,000 gpn). _At
present, line treatment does not promote calcium fluoride precipitation.
If Versar's research and development effort on effluents from the
feldspar-sand flotation (separation) step is successful, process wastewaters
from each stream (sand underflow water and feldspar float water) would be
combined. The combined streams from each of two adjacent plants would also
be combined, resulting in a total flow of 2,270 liters per min (600 gpm) of
fluoride-containing water. An expected 50 percent of this water can be
recycled directly to the flotation units with no treatment. Unfortunately,
recycling results in an eventual buildup of flotation reagents to unacceptable
levels. It is therefore necessary to purge a fraction of the recycling
water, in this case up to a maximum 1,135 liters per min (300 gpm) for both
plants. This is the stream that would require defluoridation.
Since fluoride-containing wastewaters would be at a pH of 2.5 to 3.0
(below the range needed for effective removal of fluoride by activated
alumina), liming would be required at this point. If calcium fluoride preci-
pitation is desirable, it would require sufficient liming to raise the pH
above 9. The necessity for calcium fluoride precipitation has yet to be
determined. There is a recognized necessity to pretreat activated alumina,
destined wastewater to remove or significantly reduce the amount of floating/
suspended feldspar for the elimination of clogging of the activated alumina.
38
-------
Mr. Sparks' research gave evidence that fluoride ions are adsorbed on the
feldspar (floating/suspended) to further complicate the effective renoval
of fluoride. It was concluded that removal of this feldspar from the
effluent is both functionally necessary and economically desirable.
Feldspar Corporation's mode of operation is alleged to be similar
enough to other existing feldspar operations that a system: designed for
Feldspar Corporation could be readily adopted by the rest of the industry.
39
-------
APPENDIX D. COST ESTIMATE WORK SHEETS
Cost of Chemicals, 1977 Prices
1. Activated Alumina (media)
Assume: 163,000 kkg/year (180,000 tons/year) feldspar production
1,135 liters per min (300 gpm) flow - process wastewater (PWW)
4-min residence time on media
2 columns of media - in parallel
833 to 881 kg/m3 (52 to 55 lb/ft3) - bulk density of media
66
-------
Calcium chloride is also used to remove F~ from regenerant caustic.
125
ft3
2.5 volumes
volume
62.4
ft3
Ib
600
X
io-6
5.3 kkg
removed
(11.7
eve:cy
lb)F- to
2.5 hr
be
11.7
111.1
19 x 2
24
2.4
312
$30
2,000
19 = atomic weight of F~
111.1 = molecular weight of CaCl2
If we use 1.2 times stoichiometric values of CaCl2 for F-, and if F~ concen-
tration averages about 600 ppm, then
1.2 _ $0.0114/kkg ($0.0103/ton)
180,000 ~
Ihe sum of PWW pretreatment and fluoride removal from regenerant is
$0.0103
+0.0421
$0.0578/kkg or $0.0524/ton = CaCl2 costs
4. Caustic Soda
Assume that soda is purchased as a 50-percent solution,, 76 percent Na20
basis. Recent quote is $154 to $192/kkg ($140 to $175/ton), fob works,
freight equalized (C.M.R.).
Assume that cost is $176/kkg ($160/ton) delivered.
Lab data required 2.5 volumes of 1-percent NaOH/volume of media every 2.5 hr
125 ft3
195
2.b
2
2.5 volumes
4
voli.
312
ime
$160
62.4 Ib
2,000
ftj
.01 88.5 kkq (195 2b) NaOH/2.5 hr
$0.2862/kkq ($0.2596/ton)
180,000
5. Sulfuric Acid
Use 5 volumes/volume of acidulated water per regeneration, discharging
50 percent of it after each cycle. Prepare acidulated water by adding
5 volumes natural water to 1 volume 1 percent H2SOi>.
125 ft3
2.5 volumes
volume
7.48 gal
ft3
8.34 lb| 24
gal [ 2.5
312
1
b
.01
L $4°J
2,000
180,000
$0.0119/kkg ($0.0108/ton) acid cost
Labor and Supervision
The rate of pay at Spruce Pine is ~$265/week. Ihis rate includes benefits,
fringes, overhead, and average overtime. Equivalent supervisory costs are
$30/week. ^
41
-------
Assune that one operator per shift is required (may be somewhat less than
this with proper process instrumentation).
«% Use three operators per day with supervisor for a total of approximately
$300/week per operator
$300
52 $0.29/kkg ($0.26/ton) = labor and supervisory cost
180,000
Equipment Calculations
1. Activated Alumina Column
3.54 m3 (125 ft3) madia
J- (D2) (H) = 125
let D3= H
D = 125
= 159+
D - -1.64 m (-5.4 ft) = 165 cm (65 in)
Use 80 percent free space above nedia for backwashing. Aid 0.3 m (1 ft)
below media support for liquid collection.
.•.Media columns are 1.64 m (5.4 ft) I.D. and 3.26 m (10.7 ft) high.
;
2. 50-percent Caustic Storage Tank
30 = 21,800 kg (48,100 lb) NaOH/
month
1.53 - specific gravity of 50 percent caustic
(X gal) (1.53) (8.33) = 21,800 kkg (48,100 lb)/month
X = 14,300 liters (3,780 gal) 50-percent caustic regenerated per month
.•.Plan on 15,000-liter (4,000-gal) tank.
3. 1-percent Caustic Storage Tank
125
2.5 1 7.48
1
8.34
.01
24
2.5
6
7
30
125 ft3
5.0 volumes
7.48 .= 13,700 liters/cycle (4,675 gal/cycle)
volume
/.Plan on 19,000-liter (5,000-gal) storage tank.
4. Acidulated Water Storage Tank
Use same size tank as for 1-percent caustic: 19,000 liters (5,000 gal).
42
-------
5. Regeneration liquor treatment tank
125
2.5
| 7.48
I
6 hr
2.5
6 hr = 21,300 liters (5,610 gal)
Use 22,700-liter (6,000-gal) tank with bottom outlet for calcium fluoride
slurry, carbon steel, coated.
6. PWW In-line Mixing Tank, Agitated
1,135 liters per rain (300 gpm) PWW flow
t
Use 3,785-liter (1,000-gal) tank for 3+ min residence, wilth agitation.
7. Calcium Chloride Storage Tank
Assume 40-percent concentration of purchased calcium chloride
Specific gravity = 1.47 for 40-percent solution
0.00375 Ib CaCl2
gal PWW
300 gpm
6.4
1,440 min
day
26 day
month
= 32,500 liters
8.33 x 1.4T per month
11.7 Ib F-
cycle
111.1
19 x 2
24
2.5
26
.40
8.33
1.47
(8,600 gal/month)
of 40-percent
CaCla regene-
rated
__^= 6,600 liters (1,743
1.47 gal) of 40-percent
CaCla/month for
regenerant liquor
8,600
1,740
Total - 10,340 gal/month or 39,140 liters per month
Purchase 42-m3 (11,000-gal) tank for monthly filling of 40-percent CaCl2
tank. Carbon steel OK; could use Haveg, special concrete,, etc.
8. Treated PWW Settling Tank
Use 30 to 35 min settling time. Bottom slightly conical, with bottom
slurry discharge.
1,135 liters per min (300 gpm) (35) = 40-m3 (10,500-gal) tank.
Vacuum Disc Filter Cost
(10-6) (160-5) (300 gpm) (8-33) (1,440) (||) = 340 kg/day (750 :ib/day) of CaF2
As 50 percent cake, would be ~454 kg/day (~1,000 Ib/day)
w/e = (2) (63) = 2,020 kg/m3 (126 lb/ft3)
43
-------
1,000 = 0.23
"~~
(8
Assume 0.09 ma (1 ft2) discs, with 5-cm (2-in) cakes on both sides of a disc.
thick
(8 ft3) =
3
sides discs cycles/day
(2) (5) X
index
adj.
(1.65) = $14,600
installing
tanks
3 cycles/day = X
($550/ft) (10 lb/ft2)
per ft2 in 1970
Calcium Fluoride Credit
(10-6) (160-5) (300) (8.33) (1,440) (312) (Z|j
$0.0519/kkg ($0.0471/ton) of feldspar
Credit not taken at this point because of uncertainty of market.
44
-------
1. REPORT NO.
EPA-600/2-80-058
TECHNICAL REPORT DATA
(t'lease read Instructions on the reverse before completing)
2.
I. TITLE AND SUBTITLE
Removal of Fluorides from Industrial Wastewaters
Using Activated Alumina
^UTHOR(S)
Irwin Frankel and Eric Juergens
Versar, Inc., Springfield, Virginia 22151
ORG'XNIZATION NAME AND ADDRESS
The Feldspar Corporation
Spruce Pine, North Carolina 28777
12. SPONSORING AGENCY NAME AND ADDRESS
industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
15. SUPPLEMENTARY NOTES
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
March 1980 issuing date
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1BE610
11. CONTRACT/GRANT NO.
R-804377
13. TYPE: OF REPORT AND PERIOD COVERED
Final 5/10/76-1/31/77
14. SPONSORING AGENCY CODE
EPA/600/12
A four-step,.bench-scale process has been developed that is capable of removing
at least 92 percent of the soluble fluoride from feldspar flotation process waste-
waters at a projected cost of $1.03Akg C$0.93/ton) of feldspar. For a 163,000 kkg/
year (180,000 ton/year) plant, the initial capital expenditure would be about $200,000.
The wastewater is pretreated with sodium hydroxide, lims, and calcium chloride, which
removes an initial 50 to 60 percent of fluoride. Ihe clarified water is then passed
through a bed of activated alumina for further fluoride removal,. The activated alumina
is regenerated with a 1-percent sodium hydroxide solution, and fluoride in the caustic
liquor is effectively precipitated with calcium chloride. The fluoride can be
recovered in concentrated form as insoluble calcium fluoride filter cake.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Adsorption
Fluorides
Regeneration
Wastewater
3. DISTRIBUTION STATEMENT
Release to public
.• Form 2220-1 (9-73)
b.lDENTIFIERS/OPEN ENDED TERMS
Activated alumina
Adsorption column
Backwash
Settling
Precipitation
19. SECURITY CLASS (This Report)
Unclassified
!O. SECURITY CLASS (Thispage)
Unclassified
c. COS AT I Field/Group
13B
21. NO. OF PAGES
53
22. PRICE
45
* U.S. GOVERNMENT PRINTING OFFICE: 1980-657-146/5635
-------
-------
-------
-------
-------
03 O
• DO
-
«
031
J-i *< ""** 15
W
CO
Ul
oo
rr o o
Hi
ffl O
r1 <
0.5"
rn > T3 m
TJ2 S S.
ra c
------- |