EPA R2-73-200
MAY 1973 Environmental Protection Technology Series
Recondition and Reuse of
Organically Contaminated
Waste Sodium Chloride Brines
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
U. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-R2-73-200
May 1973
RECONDITION AND REUSE OF
ORGANICALLY CONTAMINATED WASTE SODIUM
CHLORIDE BRINES
R. D. FOX
R. T. KELLER
C. J. PINAMONT
PROJECT 12020-EAS
PROJECT OFFICER
CLIFFORD RISLEY
OFFICE OF RESEARCH AND MONITORING
REGION V, U. S. ENVIRONMENTAL PROTECTION AGENCY
CHICAGO, ILLINOIS 60606
PREPARED FOR
OFFICE OF RESEARCH AND MONITORING
U. S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D. C.
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402
Price $2.10 domestic postpaid or $1.75 OFO Bookstore
-------
EPA Review Notice
This report has been reviewed by the U. S. Envi-
ronmental Protection Agency and approved for pub-
lication. Approval does not signify that the con-
tents necessarily reflect the views and policies
of the Environmental Protection Agency, nor does
mention of trade names or commercial products
constitute endorsement or recommendation for use.
11
-------
ABSTRACT
A plant of 100 gal/min capacity was constructed and operated
for one year to demonstrate the feasibility to remove and
recover phenol and acetic acid from an 18% sodium chloride
brine by adsorption on fixed beds of activated carbon. The
purified brine was used for production of chlorine and caus-
tic soda. Separate electrolytical test-cell evaluation of
the purified brine showed it to be equivalent to pure brine.
Regeneration of the carbon was accomplished by desorption
with dilute sodium hydroxide. The phenol desorbed was re-
cycled to the phenol manufacturing plant while the acetate
regenerant was processed to underground disposal wells. More
than 23 million gallons of brine were purified. Fourteen
cycles of phenol adsorption and regeneration and 105 cycles
of acetic acid adsorption and regeneration were completed
with no significant deterioration of carbon performance.
Phenol removal to <1 ppm was accomplished at 5O-140 gal/min
and 15-70°C with an effective carbon capacity of 0.167 Ib/lb.
Optimum regeneration was with 4% NaOH at 55-70°C. Removal
of 90% of the acetic acid from brine requires <80 gal/min
flow rate and <40°C temperature, the resultant loading is
0.04 - 0.06 Ib/lb of carbon. The projected net cost of puri-
fying this waste brine for reuse was $1.32 per 1000 gallons.
This report was submitted in fulfillment of Grant No. 12O2O-
EAS between the Office of Research and Monitoring of the U. S.
Environmental Protection Agency and The Dow Chemical Company.
This project was proposed, designed, and performed by the
Environmental Research Laboratory, Dow Chemical U.S.A. at its
plant site in Midland, Michigan.
111
-------
CONTENTS
SECTION PAGE
I CONCLUSIONS 1
II RECOMMENDATIONS 3
III INTRODUCTION 5
IV DESIGN OF THE DEMONSTRATION PLANT 9
V DEMONSTRATION PLANT OPERATION 15
Carbon Selection 21
Startup 21
VI DEMONSTRATION PLANT PERFORMANCE 23
Phenol Adsorption and Regeneration 23
Acetic Acid Adsorption and Regeneration ... 32
Materials of Construction Evaluation .... 53
Costs 55
VII CHLORINE TEST CELLS 59
VIII LABORATORY STUDIES 65
Phenol Adsorption and Regeneration ...... 65
Acetic Acid Adsorption and Regeneration ... 74
Physical Properties of Carbons 80
Acetic Acid Recovery 88
IX MATHEMATICAL SIMULATION OF ACTIVATED CARBON
ADSORPTION SYSTEMS 91
X ACKNOWLEDGEMENTS 101
XI REFERENCES 103
XII PUBLICATIONS 105
XIII GLOSSARY 1O7
XIV APPENDIX 109
-------
FIGURES
PAGE
1 PHOTOGRAPH OF DEMONSTRATION PLANT ADSORBERS
AND CONTROL BUILDING 13
2 PHOTOGRAPH OF ENTIRE DEMONSTRATION PLANT 14
3 SCHEMATIC FLOWSHEET OF ADSORPTION PROCESS .... 17
4 SCHEMATIC FLOWSHEET OF PHENOL COLUMN REGENERATION
PROCESS 18
5 SCHEMATIC FLOWSHEET OF ACETIC ACID COLUMN
REGENERATION PROCESS 20
6 BREAKTHROUGH CURVE FOR NUCHAR WV-G PHENOL
ADSORBER, CYCLE NO. 8 . 24
7 BREAKTHROUGH CURVE FOR WITCO 718 PHENOL ADSORBER,
CYCLE NO. 10 25
8 PHENOL DESORPTION CURVES SHOWING EFFECT OF TEM-
PERATURE AND % CAUSTIC 30
9 ACTUAL DEMONSTRATION PLANT LOADING OF ACETIC ACID
ON WITCO 718 CARBON AND NUCHAR WV-G CARBON AS A
FUNCTION OF THE NUMBER OF CYCLES 38
10 FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO 718
ACETIC ACID ADSORBER, CYCLE NO. 11 40
11 FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR
WV-G ACETIC ACID ADSORBER, CYCLE NO. 11 41
12 FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO
718 ACETIC ACID ADSORBER, CYCLE NO. 25 42
13 FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR
WV-G ACETIC ACID ADSORBER, CYCLE NO. 22 43
14 FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO
718 ACETIC ACID ADSORBER, CYCLE NO. 30 44
15 FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR
WV-G ACETIC ACID ADSORBER, CYCLE NO. 30 45
16 FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO
718 ACETIC ACID ADSORBER, CYCLE NO. 53 46
vi
-------
PAGE
17 FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR
WV-G ACETIC ACID ADSORBER, CYCLE NO. 52 47
18 FEED AND EFFLUENT CONCENTRATION CURVES FOR WITCO
718 ACETIC ACID ADSORBER, CYCLE NO. 52, AND NUCHAR
WV-G ACETIC ACID ADSORBER, CYCLE NO. 51 48
19 REGENERATION CURVE FOR WITCO 718 ACETIC ACID
ADSORBER, CYCLE NO. 22, USING FRESH CAUSTIC ... 51
20 REGENERATION CURVE FOR NUCHAR WV-G ACETIC ACID
ADSORBER, CYCLE NO. 21, USING FRESH CAUSTIC ... 52
21 REGENERATION CURVES FOR ACETIC ACID ADSORBERS USING
RECYCLE OF FORTIFIED WEAK REGENERANT 54
22 EQUILIBRIUM PHENOL LOADING FOR VARIOUS VIRGIN
ACTIVATED CARBONS 66
23 EQUILIBRIUM LOADING OF PHENOL ON CARBON VS pH . . 67
24 EQUILIBRIUM LOADING OF PHENOL ON CARBON VS
PERCENT CAUSTIC 68
25 EQUILIBRIUM LOADING OF PHENOL ON CARBON VS
PERCENT CAUSTIC AT TWO PHENATE CONCENTRATIONS . . 69
26 EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED
WITCO 718 CARBON IN NEUTRAL 2O% SODIUM CHLORIDE
SOLUTION 70
27 EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED
NUCHAR WV-G CARBON IN NEUTRAL 2O% SODIUM
CHLORIDE SOLUTION . 71
28 EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED
WITCO 718 CARBON IN 2-4% CAUSTIC 72
29 EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED
NUCHAR WV-G CARBON IN 2-4% CAUSTIC 73
30 EQUILIBRIUM ACETIC ACID LOADING FOR VARIOUS
VIRGIN ACTIVATED CARBONS 75
31 EQUILIBRIUM LOADING OF ACETIC ACID ON WITCO 517
CARBON VS pH 76
32 EQUILIBRIUM LOADING OF ACETIC ACID ON WITCO 718
AND NUCHAR WV-G CARBONS 78
VI
-------
PAGE
33 PORE VOLUME DISTRIBUTION CURVE, NUCHAR WV-G,
ACID-WASHED . 84
34 PORE VOLUME DISTRIBUTION CURVE, NUCHAR WV-G ... 85
35 PORE VOLUME DISTRIBUTION CURVE, WITCO 718,
ACID-WASHED 86
36 PORE VOLUME DISTRIBUTION CURVE, WITCO 718 .... 87
37 BATCH ADSORPTION OF PHENOL ON WITCO 718 CARBON
PREDICTED BY MATHEMATICAL MODEL COMPARED TO
EXPERIMENTAL DATA 96
38 PHENOL DESORPTION FROM WITCO 718 CARBON PRE-
DICTED BY MATHEMATICAL MODEL COMPARED TO ACTUAL
DATA FROM PHENOL ADSORBER REGENERATION NO. 4 . . 97
39 BREAKTHROUGH CURVE PREDICTED BY MATHEMATICAL
MODEL FOR DEMONSTRATION PLANT PHENOL ADSORBER . . 99
V111
-------
TABLES
KO. PAGE
1 SUMMARY OF PHENOL REGENERATION DATA ON WITCO 718
CARBON 27
2 SUMMARY OF PHENOL REGENERATION DATA ON NUCHAR
WV-G CARBON 28
3 ACETIC ACID ADSORPTION AND DESORPTION, SUMMARY
OF CONDITIONS AND PRODUCT QUALITY FOR WITCO 718
CYCLES 33
4 ACETIC ACID ADSORPTION AND DESORPTION, SUMMARY
OF CONDITIONS AND PRODUCT QUALITY FOR NUCHAR WV-G
CYCLES 35
5 SUMMARY OF OPERATING RESULTS FROM ACETIC ACID
ADSORPTION ON WITCO 718 SHOWING EFFECT OF TEM-
PERATURE AND FLOW RATE 5O
6 PROJECTED COSTS FOR PURIFICATION OF 100 GAL/MIN
OF PHENOL PLANT WASTE BRINE 57
7 AVERAGE ELECTROLYTIC TEST CELL OPERATING DATA . . 61
8 LABORATORY COLUMN EVALUATIONS OF DEMONSTRATION
PLANT CARBON COMPOSITES 79
-------
SECTION I
CONCLUSIONS
1. Waste sodium chloride brine from a phenol manufacturing
plant was purified by activated carbon adsorption and
recycled to a diaphragm-type electrolytic cell to produce
chlorine and caustic soda. The projected net cost of
treatment was $1.32/M gal.
2. Purified brine from the activated carbon adsorption proc-
ess showed no significant difference from pure salt brine
when tested in a 2660 amp diaphragm-type electrolytic cell.
3. Separation of organic byproducts from waste brine by ad-
sorption on activated carbon and their recovery by chemi-
cal regeneration of the carbon has been demonstrated on
an engineering scale in a 100 gal/min Demonstration Plant.
4. Fixed bed activated carbon adsorption columns removed and
recovered the phenol from the waste brine down to <1.0 ppm
through 14 cycles of adsorption and caustic regeneration.
5. Flow rates of 50-140 gal/min (1.0 to 2.8 gal/min/sq/ft)
and temperatures of 15-70°C had no significant effect on
the performance of the phenol adsorption system.
6. The effective capacity of the activated carbons tested
for phenol adsorption averaged 0.167 Ib/lb and no sig-
nificant deterioration in performance was detected after
14 cycles.
7. In regeneration of the phenol adsorbers, 4% caustic soda
solution at temperatures of 50-65°C at a flow rate of 30
gal/min (O.6 gal/min/sq/ft) were better than higher caustic
strength and lower temperatures. The peak phenol con-
centration in the regenerant averaged 6% and the volume
of regenerant was 6% of the feed brine treated. The NaOH
required for regeneration was 20 Ib/M gal brine purified.
8. No difference between Witco 718 and Nuchar WV-G granular
activated carbons was observed in phenol adsorption or
caustic regeneration.
9. Fixed bed activated carbon columns removed >90% of the
acetic acid from the waste brine at temperatures below
4O-45°C and flow rates below 80 gal/min (1.6 gal/min/sq/f t),
The effective carbon capacity ranged from O.O4 to O.O6
Ib/lb.
-------
10. Operation of the acetic acid adsorbers at temperatures
greater than 40-45°C resulted in an effective carbon
capacity too low to be practical. Poor effluent qual-
ity also resulted at these temperatures.
11. Each acetic acid adsorber operated through 105-cycles of
adsorption and regeneration. No abnormal decrease in
carbon capacity was noted.
12. Witco 718 produced better effluent quality and loading
under comparable conditions than did Nuchar WV-G in
acetic acid adsorption.
13. Regeneration of the acetic acid adsorbers required 3600-
4200 Ib of NaOH as 10% NaOH. The peak acetic acid con-
centration in the regenerant was 55-65 g/1, and the volume
of regenerant was 6000 gal, or ~6% of the feed brine
treated. The caustic usage was 36 Ib NaOH/M gal.
14. The tail portion of the regenerant, being weak in acetic
acid, can be saved, fortified with 3600 Ib of NaOH and
reused as regenerant. The volume of regenerant is thereby
reduced to 3500-4000 gal.
15. A method of recovering acetic acid from the regenerant,
involving solvent extraction and vacuum distillation, was
developed in the laboratory, but the amount and value of
acetic acid available for recovery was too low to justify
capital expenditure.
16. No correlation of the measured surface properties of the
two carbons tested could be made with their performance
in the Demonstration Plant.
-------
SECTION II
RECOMMENDATIONS
The activated carbon byproduct recovery and reuse approach to
the reduction or elimination of wastes has been demonstrated
on an engineering scale and can be recommended for application
to the reuse of waste sodium chloride brines to manufacture
chlorine and caustic soda. Application to phenol or phenol-
type compounds has also been demonstrated using activated
carbon adsorption and caustic regeneration. Further studies
of phenol adsorption on an engineering scale are recommended
to determine the ultimate service life of activated carbon
by extending the number of cycles run, to further optimize
the regeneration process by studying countercurrent regen-
eration, and to determine carbon performance at higher flow
rates and higher temperatures.
The same activated carbon technique of byproduct recovery is
a satisfactory method for dealing with waste brines contami-
nated with acetic acid. The process requires low temperatures
and flow rates. Better methods of treating byproducts con-
taining acetic acid and other similar small organic molecules
are needed that are operable over wider temperature ranges.
-------
SECTION III
INTRODUCTION
Waste brines contaminated with organic materials are one of
the major disposal problems in the chemical industry. Brines
are a particular concern because the standard biological treat-
ment plants are usually upset by the high salt level which
hinders biological oxidation. The brine also passes through
the treatment plant and becomes a dissolved solids addition
to the receiving waters, thereby, affecting the uses to which
the water can be put. At present, brines are disposed of in
three ways. They are discharged into seawater, if convenient,
or diluted and discharged into fresh water streams, or injected
into disposal wells. Discharge into fresh or seawater is
coming into increasing scrutiny by pollution control organi-
zations, resulting in more stringent requirements for the
brine discharge. Disposal wells are a third way to handle
briney wastes, especially for inland plants, but they require
suitable geology, and a properly installed, operated, and
monitored disposal well is expensive. In addition, subsurface
disposal of organically contaminated brines should not be
considered a satisfactory long-term solution to the waste
disposal problems of the chemical industry.
Chlorine and caustic soda plants all over the United States
use salt brine as a raw material source. Thus, it seems log-
ical and feasible to use waste brines as raw material sources
after suitable removal of contaminants.
As the beginning of a program on contaminated brine reuse,
the waste brine from a phenol manufacturing facility was
selected for demonstration studies. The waste brine from
this plant is about the same strength as the raw production
brine used in the chlorine-caustic soda cells. It is thus
an obvious candidate for recycle to chlorine-caustic soda
production. The nominal composition of this waste brine is:
Sodium chloride, % 18
Phenol, mg/1 150-750
Sodium acetate, mg/1 1400-2000
pH 8
Temperature, °C 108
Insolubles, Fe(OH)3, ppm <1 to 10
It also contains benzene and acetone, which have been reduced
to low levels by steam stripping.
-------
Range finding experiments to test the effects of phenol and
sodium acetate on a 6 ampere laboratory scale diaphragm-type
chlorine cell had shown that 200 mg/1 of phenol is unacceptable,
but sodium acetate may be acceptable when judged by their ef-
fect on the operation of the diaphragm cell. Phenol is de-
stroyed in the electrolytic cell and therefore affects current
efficiency, whereas, acetate passes through the chlorine cell
intact. Therefore, it was concluded that substantial recycle
of such a contaminated brine requires organic removal to pre-
clude problems in chlorine production and caustic soda finish-
ing.
In considering processes to remove phenol and sodium acetate,
each organic component was examined separately. Research showed
that phenol can be removed from brines by stripping, solvent
extraction, and adsorption. Steam stripping requires a good
distillation column and a high heat load. Solvent extraction
requires an efficient contactor and solvent recovery. Ad--
sorption can be on activated carbon or other adsorbents, but
activated carbon had a much greater capacity for phenol than
any other adsorbent.
In order to remove sodium acetate from an aqueous system, it
is necessary to convert it to acetic acid by acidifying to
at least pH 3. The following processes were examined without
success: solvent extraction, azeotropic distillation, and
extractive distillation. Acetic acid was adsorbable on acti-
vated carbon at relatively low but still practical loadings.
In view of the failures and problems of other unit operations
considered, adsorption on activated carbon became the only
practical approach for purifying this byproduct brine. How-
ever, the key to the successful use of the unique adsorptive
properties of activated carbon in industrial waste treatment
is the regeneration process.
Industrial wastes, having organic concentrations orders of
magnitude greater than municipal wastes, saturate carbon much
more quickly. Cycle times are short, carbon requirements are
high. Thermal regeneration, with its 5 to 15% carbon loss
per regeneration, means the carbon is replaced every 7 to 20
cycles. For industrial wastes, this cost becomes prohibitive.
What is needed are regeneration processes giving a much longer,
useful carbon life.
Further research defined the activated carbon process as
follows: phenol is removed from the brine at pH 7; the brine
is then acidified with HC1 to pH 3 and passed through another
carbon bed to remove acetic acid. As each carbon bed becomes
saturated, it is regenerated by chemical desorption of the
organics with caustic soda solution. This research and dev-
elopment grant for the demonstration of the recondition and
reuse of this organically contaminated brine had the following
project objectives:
-------
a) Develop and demonstrate a chemical-adsorption process
for the treatment of process wastewater from a phenol
manufacturing plant.
b) Demonstrate the requirements for byproduct recovery
from the subject wastewater, which will include the
net recovery of both organic and inorganic pollutants.
c) Demonstrate wastewater reuse on an engineering scale.
d) Research and develop methods for: 1) the regeneration
of adsorption materials to be used in the proposed
process, and 2) establishing requirements and limi-
tations on the use of renovated brine wastewater for
chlorine-caustic production.
The project was to be completed within a period of two years
(beginning June 30, 1969) in three major phases: Phase I -
Design and Engineering (6 months), Phase II - Construction
(6 months), and Phase III - Operation of the Demonstration
Plant (12 months).
The purpose of the Phase III operational portion was to dem-
onstrate plant operability on an engineering scale and to
define and optimize cost-sensitive operating parameters for
the activated carbon adsorption process for removing phenol
and acetic acid from a sodium chloride brine. Removal of
these organic contaminants must be sufficient to allow re-
cycle of the brine to chlorine-caustic soda production. The
level of contaminants in the adsorption plant effluent depends
on flow rate, contact time, temperature, pH, and the conditions
of the carbon after regeneration. Costs are also affected
by the above as well as by the carbon bed life.
Operating conditions which most affect the adsorption and
regeneration of each bed were defined first in a supporting
research program. The Demonstration Plant then operated at
these conditions to provide data on an engineering scale for
determination of performance and costs. The supporting re-
search studies also studied methods of recovering acetic acid
from the regenerant and investigated the physical properties
of the activated carbons for correlation with plant performance.
In order to determine precisely, on a semi-commercial scale,
the effects of the reconditioned brine on the manufacture of
chlorine and caustic soda in diaphragm-type electrolytic cells,
two specially instrumented 2660-amp test cells were operated
in the chlorine production area during Phase III. These test
cells were smaller than production cells but large enough to
supply reliable pilot plant data. One cell was fed seven tenths
of a liter per minute of treated brine from the Demonstration
Plant. The second test cell operated on pure brine to serve
as a control for comparison.
-------
SECTION IV
DESIGN OF DEMONSTRATION PLANT
The waste brine purification facility was designed to remove
the organic components from stripped Phenol Plant brine in
order to recycle the 18% NaCl brine to Chlorine-Caustic Pro-
duction. This Demonstration Plant was designed to process
100 gal/min of stripped feed brine of the following nominal
composition:
NaCl - 18%
Phenol - 200 mg/1
Benzene - 10 ppm
Acetone - 10 ppm
Sodium Acetate - 14OO mg/1
It was also designed to handle traces of finely divided in-
organic sludge, typically iron hydroxide. The design called
for a product brine containing <5 mg/1 phenol and approximately
140 mg/1 sodium acetate, when operated at temperature of 30°C.
The NaCl concentration would not be changed in this facility
and would be the same as the feed brine (nominally 18%).
The plant was located outdoors in an area across the street
from the Phenol Production Plant. A control house was built
adjacent to the plant for instrumentation and plant utilities.
Research pool operators were used to man the Demonstration
Plant 24 hr per day, 7 days per week.
Unit ratios for design were
Peak Demand Rate Avg. Use
Brine 200 gal/min 100 gal/min
Lake water 100 gal/min 0.216 gal/gal
Caustic as 100% NaOH 45 Ib/min 0.162 Ib/gal
HC1 as 100% 30 Ib/min 0.0018 Ib/gal
Spent Phenate Regenerant 100 gal/min 0.10 gal/gal
Utilities
Peak Demand Rate Avg. Use
Electrical 125 KW 0.016 KWH/gal
Steam (150 psig) 7,000 Ib/hr 0.19 Ib/gal
The plant should purify 100 gal/min of waste phenate brine
for use in the chlorine cells. This 100 gal/min of brine
would not go to subsurface disposal. Ten gal/min of spent
acetate regenerant are to be sent to the disposal wells giving
a net reduction of 90 gal/min to subsurface disposal.
-------
Laboratory data as well as the literature indicated that 2
gal/min/sq ft superficial velocity was a reasonable condition
for design of carbon adsorbers. These rates give pressure
drops of about 2" water per foot of carbon on a clean bed.
The question of height of bed was decided on the basis of
loading of the least well adsorbed impurity. We could expect
a loading of 0.05 Ib acetic acid/lb carbon at our operating
conditions and this would result in saturation of an 18 ft
bed in about 24 hours. The design specified two carbon columns
for phenol adsorption and two columns for acetic acid adsorp-
tion. They were operated downflow in series except when one
was off-line for regeneration.
There is a definite cost advantage to buying multiple equip-
ment of the same design so the phenol adsorption towers were
designed to be the same size as the acetic acid towers. This
results in a much longer adsorption cycle for phenol. At a
loading of 0.15 Ib phenol/lb carbon and a feed concentration
of about 300 mg/1 phenol, six-day cycles could be expected
with an 8' bed height. It was decided to buy vessels that
had a potential carbon height of 20 ft, giving the flexibility
to load any quantity of carbon in order to attain a reasonable
number of cycles during the grant operating period.
Carbon vendors advised as to the galvanic type corrosion of
steel in contact with carbon beds. This, coupled with our
corrosive solutions, led us to the choice of rubber-lined
steel as material of construction.
Regeneration was also designed to be carried out downflow, or
cocurrent to adsorption. This simplified the piping and carbon
support designs. The bed supports were simple in design.
Structural support was furnished by channels welded to the
vessel and then rubber coated. Perforated steel plates coated
with vinyl plastisol were laid on the channels. Then a poly-
propylene filter cloth that could retain 20-40 mesh carbon
was laid over the perforated plates. Sealing to the vessel
sides was with a silicone RTV sealant.
It was recognized from lab data that careful pH control was
essential to adsorption of acetic acid. Therefore, it was
decided to use a stirred tank with one hour residence time
(6000 gal) and continuous recirculation of a sample through
an inline pH sensing device. The pH sensor was to operate a
control valve that regulates the HC1 addition rate. Operating
data have proven that this is an excellent system controlling
at pH 3.0 ± 0.1 pH unit.
The treated brine neutralizing tank was sized at 1800 gallons
because the pH control is not so critical. Since the pipeline
to the Chlorine Plant was a steel pipeline, the purified brine
effluent from the acetic acid adsorbers at pH 4 had to be
neutralized to the pH range of 7.5-8.5.
10
-------
The clarifier was designed on the basis of settling rates at-
tained on Phenol Plant brine with the addition of Purifloc®
A-23 flocculant at 0.2 ppm. This resulted in the vessel di-
mensions of 30 ft dia by 10 ft high with a standard overflow
weir.
The tank volumes were selected on the basis of projected use
rates in the case of reagents and accumulation rates for proc-
ess vessels. Caustic and HC1 storage were sized to be refilled
once a shift. Lake water was contained in a surge tank of
1000 gal capacity to stabilize a repressurized system. Phenate
regenerant was to be continuously pumped at 50 gpm so 750
gallon was an adequate size to give a 15 minute surge capacity.
Acetate regenerant was to be recycled and reused, and based
on the regenerant volume projected, a 12,000 gallon tank was
selected.
The materials of construction used in the Demonstration Plant
were in the following categories:
Service
Material
18% NaCl Brine above 80°C, pH 7-9
pipe
pumps
valves
heat exchanger
18% NaCl Brine below 80°C, pH 6-9
pipe
pumps
valves
18% NaCl Brine, pH 2-3
tanks
agitators
pipe
pumps
Sodium Hydroxide Sol'ns, 4%-3O%
pipe
pumps
agitators
Hydrochloric Acid Sol'ns 17-36%
pipe
pumps
Sched 80 steel
Ductile Cast Iron
Forged Steel
Forged Steel
Sched 40 steel
Ductile Cast Iron
Forged Steel
Butyl Rubber-Lined Steel
Butyl Rubber-Lined Steel
Propylene-Lined Steel
Polypropylene-Lined Steel
Sched 40 Steel
Ductile Cast Iron
Forged Steel
Polypropylene-Lined Steel
Polypropylene-Lined Steel
Standard design practices were used for foundations, vessel
supports, pipe supports, etc. These standards have been
developed by long term experience in construction of chemical
processing plants.
11
-------
Complete engineering details and drawings of the Demonstration
Plant have been recorded in the Design and Engineering Report
submitted to the Environmental Protection Agency. The design
and construction of the facility was done from June 30, 1969,
to August 27, 1970. Construction was done from February 1970
through August 27 by outside contractors with client inspection
as required. Photographs of the completed facility are shown
in Figures 1 and 2. Figure 1 shows the activated carbon ad-
sorption columns and the control house. Figure 2 is an overall
photograph of the Demonstration Plant. The concrete clarifier
is seen on the left, raw materials tankage and the air cooler
in the middle foreground, the Dempster stations and control
house on the right. In the far background can be seen portions
of the Phenol Plant where the brine originates.
12
-------
U>
Figure 1
PHOTOGRAPH OF DEMONSTRATION PLANT ADSORBERS AND CONTROL BUILDING
-------
T
1 «,
Figure 2
PHOTOGRAPH OF ENTIRE DEMONSTRATION PLANT
-------
SECTION V
DEMONSTRATION PLANT OPERATION
After startup in late August 1970, the Demonstration Plant
was operated for thirteen months, during which time over 23
million gallons of the actual byproduct brine from the Phenol
Manufacturing Plant was purified. During this period of time,
the conditions under which the plant was run were varied to
determine the effects on certain variables on both the ad-
sorption and the regeneration processes. In addition, the
plant operation also experienced both the variations in con-
ditions caused by a complete annual cycle of Central Michigan
weather and also the variations that are caused by the oper-
ation and misoperation of the phenol process equipment.
The quality of the feed brine received from the Phenol Plant
was highly variable. From the beginning, the specification
on feed brine analysis used as the design basis for the plant
was too restrictive. Phenol, acetone and benzene concentration
were frequently higher than the specifications. Acetic acid
concentration did not vary as greatly or rapidly as the con-
centration of these other orga'nics. The feed brine specifi-
cations were eventually adjusted to accommodate the real capa-
bilities of the Phenol Plant equipment, especially the brine
stripper. The acceptable maximum concentrations were raised
to 1500 mg/1 for phenol, 50 microliters/1 for acetone, and 15
microliters/1 for benzene. The incoming feed brine was sampled
every two hours to monitor the concentration of these organics.
Failure to consistently meet feed brine specifications was
caused by two major problems.
The first problem was the inability of the Phenol Plant to meet
the specified organic contaminant level in the stripped brine.
It was necessary to improve the steam stripping operation to
hold volatile organics (acetone and benzene) at the required
levels. Another problem was upsets in the extractor operation
which resulted in intolerable phenol levels above 2000 ppm.
The correction to this problem was more precise pH control and
increased solvent to brine ratios in the extractors in the
Phenol Plant. For effective phenol extraction and adsorption,
the pH must be less than 9. Sufficient contact between the
phenol bearing brine and the benzene extractant must be main-
tained for good extraction. Toward the end of the operating
period both problems had been corrected, and throughputs of
130-140 gpm of quality brine were being attained. These prob-
lems serve to illustrate the variability of feed quality with
which an industrial waste treatment process must be designed
to cope.
15
-------
The second operating problem was the inability of the Chlor-
Alkali Plant to accept the purified 18% NaCl brine when it
was short of solid NaCl to saturate the brine. When the
Chlor-Alkali Plant could not saturate purified brine, the
Demonstration Plant was operated to collect data, and the
product brine was pumped underground for disposal.
The adsorption process shown schematically in Figure 3 is
operated as follows:
The feed brine to the unit is received from the stripping
still at the Phenol Plant at 108°C. Purifloc A-23 floc-
culant is added to the hot brine, and the insoluble floes
are allowed to settle in the clarifier. The clarifier
overflow is cooled in an air cooled exchanger to the
temperature selected for study in the range of 30 to 70°C.
Brine is pumped at a controlled flow rate down through
two activated carbon beds in series for removal of phenol.
The brine, after passing through phenol adsorption, is
acidified to a pH of 3 by adding HC1 on pH control. The
acidified brine is then pumped at a controlled flow rate
down through two activated carbon beds in series for
removal of acetic acid. The treated brine is collected
in a surge tank where it is neutralized by addition of
NaOH on pH control. The brine is pumped to the brine
treating plant in the Chlor-Alkali complex. A continuous
sample flow of the treated brine stream is collected in
a Dempster tank for transport to the chlorine test cells.
About half-way through- the operational phase, it was con-
cluded that more brine could be processed through the
phenol adsorbers, but that the capacity of the acetate
adsorbers would be exceeded, especially as operating tem-
peratures rose in the summer. In order to process more
brine through the phenol adsorbers but have only a portion
of the dephenolated brine go to the acetic acid adsorbers,
a by-pass and control system (shown in Figure 3) was in-
stalled to divert to the treated brine tank any brine not
pumped to the acetic acid adsorbers. This allowed opera-
tion at rates of to 140 gal/min, which was the limit of
the Phenol Plant stripper, through the phenol adsorbers.
The flow to the acetic acid adsorbers was restricted to
60-80 gal/min. This change occurred at Cycle No. 70 on
the acetic acid absorbers.
Regeneration of the phenol beds, shown schematically in Figure 4,
is done as follows:
When the first phenol bed becomes saturated as monitored
by outlet phenol concentration approaching the inlet con-
centration, it is taken off line for regeneration; the
second bed continues to process feed brine. Twenty-seven
percent plant caustic is diluted with filtered lake water
to 4% by use of a flow ratio controller to serve as re-
16
-------
RAW BRINE
FROM
PHENOL
PLANT
108°C
pH 7
PURIFLOC
k.
INJECTION
^
(
r
s^
F
~1
\
r
0s) PHENOL
J ADSORBER
1
1
BRINE
COOLER
L
r
.j L.
CLARIFY
SLUDGE
<*—
r
• v
1
r
'•i
^
r
PHENOL
ADSORBER
HCI
N
)
r^
p
^
"
^
r
ACETIC
ACID
ADSORBER
^
r
ACETIC
ACID
ADSORBER PURE BR|NE
"1 FC J
V V x-x
BYPASS "-
\*
41
l-i
LJ
1
i
ACIDIFY
V.
i
)
I
-•^•C.
i
_'
1
1
r*
<-{->
N^
PK.
~t
Na
L)H
r*
j
^••H
^
n
rSTp)
^ J \ 1
L
y'
|
V
PRODUCT TO
pRnnnrrinM
CELLS
w
_. 1
i
|
NEUTRA-
LIZE
i
1
N.-^-"-...'
SAMPLE
FLOW TO
TEST CI2
CELLS
r*n
r rN
V 7
TO LANDFILL
Figure 3
SCHEMATIC FLOWSHEET OF ADSORPTION PROCESS
-------
FEED BRINE
CO
SPENT
PHENOL
ADSORBER
FRESH
PHENOL
ADSORBER
TO ACETATE
REMOVAL
SECTION
I
SODIUM
PHENATE
TANK
TO PHENOL
PLANT
Figure 4
SCHEMATIC FLOWSHEET OF PHENOL COLUMN REGENERATION PROCESS
-------
generant solution. The regenerant is pumped on flow
control down through the bed at 1/3 to 1/2 the adsorption
flow rate. The displaced feed brine is pumped to the
disposal well line to prevent adding salt to the phenate
regenerant. All regenerant solution but the initial dis-
placed brine is collected in the phenate regenerant tank
and returned to the Phenol Plant as dilute caustic for
their use. In order to achieve good effluent quality in
the subsequent adsorption cycle, regeneration must be
continued until the level of phenol in the column ef-
fluent reaches 500 mg/1. The regenerated bed is put on
line down stream from the partially spent bed, and the
cycle is repeated as the beds saturate alternately.
Regeneration of acetate beds, shown schematically in Figure 5,
is done as- follows:
When one acetate bed becomes saturated as monitored by
outlet acetic acid concentration approaching the inlet
concentration, this bed is taken off line for regeneration;
the second bed continues to process feed brine. A se-
quence of regeneration steps using caustic are then per-
formed to optimize the concentration of sodium acetate
in the regenerant and minimize the volume of regenerant.
Two methods were used differing only in the caustic re-
generant make-up. The first method was to make up 10%
NaOH solution by mixing the 27% caustic supply with water
in a flow system. The second method utilized the weak
regenerant at the end of the cycle as a dilution source,
and 27% caustic was added to the regenerant tank to give
a 10% caustic solution. After the caustic was pumped
downflow into the bed, the carbon bed was sequentially
washed with water and acidified with HC1 to prepare it
for adsorption service. The regenerant solution that
has a high acetate concentration was pumped to the dis-
posal well system. In the first method, the regenerated
bed was placed in adsorption service when the pH of the
effluent dropped below 3, and all regenerant was pumped
underground. In the second method, regenerant was pumped
out until the sodium acetate concentration in the re-
generant sample was less than 2 g/1 and .then the weak
regenerant was saved as makeup solution for the next
cycle. The regenerated bed was put on line downstream
from the partially spent bed, and the cycle was repeated
as the beds saturated alternately.
Simple operating control analyses were done by Research Operators
operating the adsorption facility. Complex analyses were per-
formed by the Analytical Services Laboratories.
19
-------
to
O
SPENT
ACETATE
ADSORBER
DEPHENOLATED
BRINE
FRESH
ACETATE
ADSORBER
MA RECYCLE
REGENERANT
I
ACETATE
REGENERANT
n
TREATED
BRINE
10%
NaOH
STRONG
REGENERANT
TO WELL
DISPOSAL
NaOH
J^ TREATED
/-Sj BRINE TO
I f J CELLS
y—x
Figure 5
SCHEMATIC FLOWSHEET OF ACETIC ACID COLUMN REGENERATION PROCESS
-------
Carbon Selection
As determined in the laboratory studies, commercially
available carbons do not show much difference towards
phenol selectivity. The principal basis for selection
of eligible suppliers was the initial capacity or loading
for acetic acid. Bids were solicited on three grades of
carbon: Witco 718, 12 x 30 mesh; Westvaco Nuchar WV-G,
12 x 40 mesh; and Pittsburgh CAL, 12 x 40 mesh. Witco
petroleum coke base carbon, and Nuchar bituminous coal
base carbon, were selected as suppliers of 1200 cu ft
of carbon each. The original plan was to load 800 cu ft
of each in the acetic acid adsorbers, and only I/2-fill
the phenol adsorbers with 400 cu ft of each. This arrange-
ment would provide a long term use comparison of two dif-
ferent types of carbon.
Startup
After check-out of the mechanical aspects of the system
using water, carbon loading began in late August. A jet
eductor was used to suck dry carbon from a hopper into a
100 gpm recirculating water stream and lift it to the
top of the water-filled adsorption column. The carbon
was allowed to fill the column by free-fall through the
water. Water flow was also downflow through the support
screen, so no backwashing of fines from the carbon was
accomplished. The phenol adsorbers were filled to 8'
height and the acetic acid adsorbers were filled with the
balance of the initial carbon shipments. They had ~18'
of carbon at the beginning. Brine treatment commenced in
late August, and shortly thereafter, the appearance of
insoluble hydroxides in the treated brine made it necessary
to stop operation after a few hours and acid-wash the
carbon. This was done by circulating acid brine through
the columns, and after neutralization, back to the clar-
ifier for solids removal.
The first 2 cycles on the Witco 718 phenol adsorption
column and the first cycle on the Nuchar WV-G phenol
column were run with the 8' of carbon bed height. The
operational problems of getting plug-flow of regenerant,
with a column only half-full of carbon, were such that
it was decided to fill the columns completely. This was
accomplished in early November. Similar problems in
achieving plug-flow of regenerant were also experienced
with the nearly-full acetic acid adsorbers, so they were
also filled up with carbon. This brought the amount of
carbon in each adsorber to the following weights:
21
-------
Witco 718 phenol adsorber : 27,500 Ib
Nuchar WV-G phenol adsorber : 26,750 Ib
Witco 718 acetic acid adsorber : 29,500 Ib
Nuchar WV-G acetic acid adsorber: 26,600 Ib
All four columns then contained 19' of carbon. This in-
creased quantity of carbon in each phenol adsorption
column has had the effect of more than doubling the
length of the cycle time of each column. It was on the
basis of this carbon column configuration that the per-
formance of the Demonstration Plant was evaluated.
22
-------
SECTION VI
DEMONSTRATION PLANT PERFORMANCE
As a result of processing 23 million gallons of brine during
the one year operating period, 14 cycles for phenol adsorption
and regeneration and 105 cycles for acetate adsorption and
regeneration for each type of carbon were completed,,
The degree of freedom in choosing temperature variables for
adsorption were regulated by the air-cooled heat exchanger..
For a given ambient air temperature, the adsorption temperature
of the carbon columns was a function of flow rate» The sys<=
tern was actually operated at flow rates ranging from 50=140
gal/rain and at temperatures ranging from 15° to 70°C. With
an empty bed volume of 7100 gal and a void volume of 75%, the
range of residence time for the above flow rates was 100 min
to 40 min.
Phenol Adsorption and Regeneration
The quality of the effluent from the phenol adsorbers
under stable operating conditions after several cycles
of cocurrent regeneration ranged in phenol concentration
from 0,5 to 400 mg/l0 The higher values are believed to
have resulted from incomplete regeneration of the column/
The phenol adsorption system, after 13 cycles of adsorp-
tion and regeneration, produced effluent brine containing
0.7 mg phenol/1 when treating 140 gal/min of brine con-
taining 300 mg phenol/1 at 70°C. It can be concluded
that a properly regenerated carbon system can produce an
effluent containing <100 mg phenol/1 with good carbon
capacity at flow rates up to 208 gal/min/sq ft and tern-
peratures up to 70°C.
Two typical phenol adsorption breakthrough curves at 105=
110 gal/min and 60°C for each type of carbon are presented
in Figures 6 and 70 These data were collected when the
flow rate through the adsorbers remained fairly constant
throughout the cycle. The phenol concentration in the
feed was not uniform; these curves illustrate the feed
concentration fluctuations which were encountered. In
both of these cycles, the effluent from the columns was
less than 1 mg/1 until breakthrough. Figure 7 also serves
to illustrate other pertinent points. The effluent from
the carbon column after being placed in service following
regeneration is not low. This can occur if the regener-
ation is not complete. The shape of the breakthrough
curve is also noteworthy. Breakthrough occurs when the
column is ~40% loaded and the column slowly begins to
pass ~20 mg/1 phenol as the column is treating strong
feed (>800 ppm phenol). However, once the feed compo-
23
-------
1400
CYCLE NO. 8
FLOW: 105 gpm
TEMP: 60°C
1200
1000
z~
g
< 80°
£C
LJJ
o
8 600
LJJ
I
0.
400
200
100
800 900
200 300 400 500 600 700
THOUSAND GALLONS OF EFFLUENT
Figure 6
BREAKTHROUGH CURVE FOR NUCHAR WV-G PHENOL ADSORBER,
CYCLE NO. 8
24
-------
1400
1200
CYCLE NO. 10
FLOW: 110gpm
TEMP: 60 - 65°C
1000
FEED
800
600
E
O
cc
z
w
u
z
O
O
O
Z 400
LU
I
a.
200
100 200 300 400 500 600 700 800 900
THOUSAND GALLONS OF EFFLUENT
Figure 7
BREAKTHROUGH CURVE FOR WITCO 718 PHENOL ADSORBER,
CYCLE NO. 10
25
-------
sition dropped down, the column was able to hold its own
until it saturated. The loading was calculated from these
graphs for comparison with equilibrium data and regener-
ation data.
Because of the long phenol cycle time periods, during
which time the conditions of flow rate, temperature, and
feed concentration varied considerably, the calculation
of loadings from the adsorption curves was not an accurate
approach. In order to track changes in the adsorptive
capacity of the phenol adsorbers with repeated regener-
ation cycles, the loading for each cycle was measured by
integrating the sodium phenate desorption curve.
/
The regeneration data for all cycles are summarized in
Tables 1 and 2 for Witco 718 and Nuchar WV-G, respectively.
These tables list the regeneration conditions employed,
the amount of phenol removed, and the effects observed
on the total regenerant effluent. From the amount of
phenol removed, an effective loading for each cycle was
calculated.
Most of the regenerations were carried out with ~30 gal/min.
(0.6 gal/min/sq ft) of 4% caustic, because equilibrium data
had suggested the optimum to be around 4% and slow flow
rates do provide conditions for maximum phenol desorption
with minimum regenerant volume. The region below 4% NaOH
was not tested because of concern that sufficient alkalinity
would not exist for raising the pH and reacting with phenol.
The rate at which the Phenol Plant could accept the regen-
erant recycle in their process also restricted the flow
rate options. As seen in the Tables, there were a few
occasions when the caustic soda strength and flow rate
did vary. The chief regeneration variable examined was
temperature, which was varied from 15° to 65°C. Data
from 3 regenerations were misplaced or judged to be in
obvious error.
One of the first effects to be noted in the Tables is the
effect of bed height on the regenerant effluent and ef-
fective loading. The first cycle on each carbon column
contained only 8' of carbon. A comparison of 4% NaOH vs
14-19% NaOH as regenerant was made on the first cycle.
The effluent volume and the loading was the same, but the
shape of the elution curves was much different as evidenced
by the much lower peak phenol concentration with 14-19%
NaOH. The second cycle on the Witco 718 column also had
only 8' of carbon. All three indicators - regenerant
volume, peak phenol concentration, and phenol removed were
worse as compared to the first cycle.
It was at this point that both carbon columns were filled
with carbon to 19' height to facilitate plug-flow during
26
-------
TABLE 1
SUMMARY OF PHENOL REGENERATION DATA ON WITCO 718 CARBON
Cycle
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Bed
Height,
ft.
8
8
19
19
19
19
19
19
19
19
19
19
19
19
Re gene rant
Feed
% NaOH
4
4
4/15
4
4
4
4
4
4
4
4
4
4
4/12
Flow, gpm
36
25
30/5
33
30
30
30
30
30/12
30
30
30
30
30/17
Temp., °C
25
15
15/32
30
50
55
60/50
55
60
55
50-60
50-60
Effluent
M
gal .
29
56
no
108
peak phenol
cone. , mg/1
50 ,000
20,000
61 ,000
61 ,000
Ib
phenol
1660
740
6550
5820
bad data
60
48
61
54
68
54
45
62
98,000
62,000
62,000
37,000
46 ,000
58,000
57,000
42,500
6160
4600
6000
5900
3865
3990
3995
6200
4640
Carbon
Loading
Ib phenol/lb
0.138
0.062
0.250
0.220
0.233
0.174
0.210
0.207
0.142
0.145
0.145
0.236
0.175 ,
27
-------
TABLE 2
SUMMARY OF PHENOL REGENERATION DATA ON NUCHAR HV-G COLUMN
Cycle
No.
1
2
3
4
5
6
7
8
9
10
n
12
13
14
Bed
Height,
ft.
8
19
19
19
19
19
19
19
19
19
19
19
19
19
Regenerant
Feed
% NaOH
14-19
4
4
4
4
4
4
8/4
4
4
4
3-11
7
18/4
Flow, gpm
15
30
33
30
30
25
30
30
30
30
30
6/30
30
27/32
*
Regenerated before saturated
Temp. , °C
25
10
10
30
55
60
55
55
25/60
65
60
60
60
Effluent
M
gal .
29
140
140
135
66
72
72
52
54
63
65
80
48
peak phenol
cone. , mg/1
10,500
59,000
42,000
49,000
64,500
88,000
34,000
69,000
61,500
48,000
64,000
21 ,000*
58,000
Ib
phenol
1300
5700
4420
4180
3900
4450
4500
3100
4100
3350
3090
4930
2510*
3360
Carbon
Loading
Ib phenol/lb
0.145
0.226
0.176
0.165
0.154
0.175
0.177
0.116
0.153
0.125
0.116
0.184
0.10*
0.125
28
-------
regeneration. From the data collected with this column
configuration, three desorption curves have been plotted
in Figure 8 to show the effect of temperature. The main
temperature effect is on the volume of regenerant, with
the higher temperature clearly desorbing phenol down to
500 ppm faster. Thus, high regeneration temperatures
become most important to efficient phenol desorption.
The observed peak'phenol concentration of 98,000 mg/1
obtained at 55° was the highest level attained, but no
cause could be ascertained. More typical values were
around 60,000 mg/1. Another effect on regeneration, that
of caustic soda concentration has also been depicted in
Figure 8.
The discontinuities in the 15°C curve are due to changes
in conditions during the regeneration. They occurred
when the dilution water for making 4% caustic stopped
and the caustic strength pumped into the column rose to
14-16%. Even though there was a corresponding decrease
in flow rate with the attendant increased contact time,
the effect of strong caustic on phenol elution is apparent.
The higher strength caustic markedly slowed down the rate
of desorption. This effect was noted whenever higher
caustic strengths were used, as can be seen by comparing
the peak phenol concentration and regenerant volume for
Nuchar WV-G cycles 8, 11, and 13. On these cycles, the
phenol removed was nearly the same, but the cycles using
7-8% NaOH were slower in desorbing. For Cycle No. 14 on
Nuchar WV-G, 18% NaOH was used at the beginning and the
peak phenol concentration was high, but after 5 hr of
regeneration, the caustic strength was lowered to 4%. •
The result was a definite second phenol peak on the de-
sorption curve of 33,500 ppm. This again showed the
advantage of 4% NaOH as regenerant. The volume of re-
generant produced with 30 gal/min of 4% NaOH at 55°-65°C
ranged from 45,OOO gallons to 72,OOO gallons, with an
average being ~60,000 gallons. This represents 8.5 bed
volumes of regenerant.
From the data collected on these 28 cycles of phenol re-
generation, the optimum regeneration conditions can be
specified. At 30 gpm of 4% NaOH in cocurrent flow and
temperatures of 55-65°C, phenol is desorbed from a 19'
high carbon column with 60,000 gallons of regenerant
being produced. A peak phenol concentration of ~60,000
mg/1 can be expected, and the average phenol concentration
will be ~10,000 mg/1.
Examination of the loading values calculated by integration
of the desorption curves showed no significant deteriora-
tion in the effective working capacity of either carbon
during the processing of over 23 million gallons of brine.
29
-------
100,000
60,000
REGENERATE: 30 gpm OF 4% NaOH
EXCEPT WHERE NOTED
PHENOL REMOVED:
• 6550 Ib
-- v -- 6160 Ib
--- • --- 5820 Ib
12% NaOH
@ 5 gpm
300
15 30 45 60 75
THOUSAND GALLONS OF REGENERANT
Figure 8
PHENOL DESORPTION CURVES SHOWING EFFECT OF TEMPERATURE
AND % CAUSTIC
30
-------
The variations observed in Tables 1 and 2 are attributed
to variable feed compositions and starting regeneration
before true saturation of the carbon. The loadings cal-
culated for the two breakthrough curves in Figures 6 and 7
compare favorably with those obtained from the regener-
ation curves for those cycles. For Cycle No. 8 on Nuchar
WV-G, the breakthrough curve loading was 0.144 Ib/lb while
the phenol removed was 0.116 Ib/lb. For Cycle No. 10 on
Witco 718, these values were 0.132 Ib/lb and 0.140 Ib/lb.
The loading for these cycles can also be compared with a
calculated equilibrium loading obtained from the isotherms
in Figures 26, 27, 28, and 29 under Laboratory Studies.
The average feed concentration for Cycle No. 8 on Nuchar
WV-G was ~800 mg/1; the effective loading from equilibrium
data on virgin carbon at this concentration is 0.28 Ib/lb.
Similarly, Cycle No. 10 on Witco 718 had a feed concentra-
tion of ~600 ppm; the equilibrium value is 0.20 Ib/lb.
Although this comparison shows that Witco 718 achieved a
loading closer to equilibrium than Nuchar WV-G, the wide
fluctuations in phenol concentration in the feed, evident
in Figures 6 and 7, make such a conclusion ill-advised.
From examination of the data in Tables 1 and 2 and from
observations during operation of the Demonstration Plant,
there does not appear to be any significant difference
in the performance of the two carbons tested in the phenol
recovery program. The service loading, effluent quality,
and regeneration characteristics of the two were essentially
equivalent.
The total amount of phenol recovered and recycled from 23
million gallons of brine ill 28 cycles of adsorption and
regeneration was 116,100 Ib. Of this, 112,700 Ib was re-
covered in the 19' high carbon columns in 25 cycles (13
on Nuchar and 12 on Witco). These figures represent an
overall average loading in the 19' high columns of 0.167
Ib phenol/lb carbon. The overall average feed concentra-
tion for the entire demonstration period was 535 ppm.
Using the average loading and feed concentration, the
average number of gallons processed per cycle in the 19'
high columns can be calculated as follows:
0.167 Ib phenol 27,125 Ib 1, OOP, OOP Ib brine 1 gal brine _
Ib carbon column 535 Ib phenol 9.15 Ib brine
928,000 gal/column cycle
At 60,000 gal of 4% NaOH regenerant, the caustic require-
ments are 20 lb/1000 gal. This average performance data
will be used in determining the cost of treatment in a
subsequent section.
31
-------
Acetic Acid Adsorption and Regeneration
The acetic acid adsorption section of the Demonstration
Plant was operated at flow rates of 50-105 gpm (1.0-2.1
gal/min/sq ft) and temperatures of 15° to 70°C. A total
of 105 cycles of adsorption and caustic regeneration were
carried out on each carbon column. The data on feed brine
composition, operating conditions, and product brine
quality for cycles 16-100 are summarized in Tables 3 and
4 for Witco 718 and Nuchar WV-G, respectively. Also
listed are loadings calculated from the loading curves
for some of the cycles.
In the tabulation of feed brine composition, it should be
noted that the average phenol concentration in the feed
brine began to vary more widely as the operational phase
progressed. Similarly, there was a definite decrease in
acetic acid concentration in the feed starting around
cycle 70. Whereas, 1500-2000 mg/1 acetic acid had been
the concentration range, it declined to 900-1200 mg/1.
The product brine quality was determined by analyzing 2-
hr composite samples of both the effluent from the second
or backup acetic acid adsorber when both columns are on
line in series flow and the effluent from the one column
on line while the other is being regenerated. Therefore,
the summary for Witco 718 in Table 3 is the quality of
the brine effluent from the Demonstration Plant when that
carbon column was backup for the Nuchar WV-G column, or
onstream alone. Similarly, the Nuchar WV-G data in
Table 4 represent the quality of brine after passing
first through the Witco column and then through the
Nuchar column or through the Nuchar column alone. These
tables are indicative of the product quality of the
treated brine produced by the method of operation de-
scribed earlier.
As reported previously, the phenol content of the ef-
fluent brine leaving the phenol adsorbers was in the
range 0.5 to 4 mg/1. This phenol is subsequently ad-
sorbed in the acetic acid adsorption columns resulting
in a product containing less than 0.1 mg/1 phenol.
Benzene is strongly adsorbed by carbon and by butyl
rubber, so that between the rubber liners on equipment
and carbon in the adsorption system, the product contains
less than 0.1 mg/1 benzene. Acetone is adsorbed at its
equilibrium conditions on the carbon in all four adsorbers.
The tables reflect an equilibrium amount of acetone pass-
ing through the process. There is some disappearance of
acetone. We postulate that during regeneration with
caustic, some condensation products of acetone (diacetone
alcohol and highers) are formed and are' permanently held
on the carbon and not displaced by the caustic regener-
ation procedure.
32
-------
TABLE 3
ACETIC ACID ADSORPTION AND DESORPTION
SUMMARY OF CONDITIONS AND PRODUCT QUALITY
FOR WITCO 718 CYCLES
FEED ANALYSIS
Cycle
No.
17
Flow i 00H : HOAc
Rate i Ads Ads
gpm ! T °C i T °C
90 ! 45
18 ' 80 50
19 ] 80
20 65
24 '. 50
25 ; 60
26 ' 90
27
28
29
30
31
32
33
34
35
36
37
38
39
40
42
43
44
45
46
47
48
49
50
51
52
53
54
55
J56
57
58
iJL
80
85
75
90
90
88
78
80
85
72
80
78
62
70
70
55
66
100
98
95
65
60
95
95
90
97
99
55
105
80
85
65
80
60
38
45
30
40
52
50
52
45
50
40
43
42
35
35
28
34
35
33
40
36
44
36
67
67
66
32
25
50
50
44
• 50
77 .
58
53
40
45
30
38
30
80 45
42
46
34
40
24
33
49
47
45
43
471
37
40
38
30
30
24
30
31
29
36
30
30
31
60
62
55
27
20
46
46
38
45
68
42
52
26
35
25
25
20
35
DMK Bz
(.1/1 i i.l/1
13 1 10
40
15
24
36
25
20
15
60
50
65
62
75
50
35
5
3
24
3
12
2
0
0
9
5
5
18
23
0
46
0
0
14
5
5
0
25
42
55
00H ' HOAc
mg/1 mg/1
500 1500
5 1500
3 , : 1600
4 i 165 . 1800
1 i 175 ' 1700
10
4
10
15
15
15
20
20
10
5
2
1
41
1
1
4
0
0
5
1
1
1
1
0
2
0
0
8
1
1
0
0
0
0
140' 1700
125 1800
250 1700
200
180
125
150
-
-
275
240
155
230
190
198
180
225
200
210
165
270
305
210
190
-
265
650
455
440
340
270
290
275
160
1800
1700
1400
1750
1500
1140
1400
1350
1550
1440
1500
1410
1370
1650
1770
1960
1620
1770
1600
1510
1300
1220
1560
1690
1820
2300
1800
PRODUCT ANALYSIS
DMK Bz
,,1/1 pl/1
23 <1
18 <1
23 <1
19 <1
14 <1
16 <1
19
<1
12 <1
10
9
8
8
7
7
4
8
7
13
7
20
17
7
14
8
9
10
8
8
8
8
8
5
3
7
11
1950 5
2060 i 5
2000 10
2050 10
<1
<1
<1
<1
<1
'1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
00H
mg/1
<1
<1
<1
<1
<1
<1
<1
<]
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
,<1
<1
<1
<1
<1
<1
<1
j HOAc
mg/1
240
60
58
23
60
90
150
68
220
135
190
250
277
150
139
219
284
66
82
145
130
225
-100
-200
-600
853
383
211
127
258
292
90
273
505
95
245
59
145
72
176
Total] Unacc
Carbon1 Carbon
mg/1 mg/1
166
% Re-
moval
HOAc
84.0
90 j 96.0
107 . j 96.4
71 ; 98.7
108
58
63
81
88
132
448
198
132
91
153
183
280
167
263
76
36
45
40
23
34
433 48
25
27
13
27
28
102
39
32
35
45
62
53
58
36
136
76
101
55
111
227
96.5
94.7
90.6
96.0
87.8
92.0
86.5
85.7
81.5
86.9
90.0
83.8
81.7
95.4
94.5
89.7
91.5
86.4
56.5
76.4
88.1
92.1
82=19
77.5
82.5
70.1
94.8
89.3
96.7
92.6
96.5
91.2
78.9
.
HOAc
loading
Ib/lb
0.056
0.056
0.061
0.045
0.040
0.046
0.054
0.022
0.020
0.043
0.043
0.046
0.026
0.042
0.055
0.055
0.042
33
-------
TABLE 3 (Cont'd.)
FEED ANALYSIS ', PRODUCT ANALYSIS !
Cycle
No.
60
61
62
63
64
65
66
67
68
69
*70
71
7.
73
**74
75
76
77
78
79
80
81
82
83
84
85
86
87
89
90
91
92
93
94
95
96
Flow
Rate
gpm
90
90
75-90
90
90
95
85
95
98
92
92
92
50
60
50
23
60
50-60
50-60
60
60
60
60
60
60
60
60
60
60
60
80
80
80
80
00H
Ads
T °C
55
55
42-48
50
48
56
50
55
40
50
55
55
60
65
60
40
50
55
62
55
65
66
67
62
65"
60
64
66
72
68
68
HOAc
Ads
T °C
50
50
30-35
43
40
50
45
48
35
45
50
50
53
55
20
20
50
35
40
45
55
45
55
55
55
58
62
56
60
64
68
64
64
60 64 \ 60
DMK
1.1/1
60
66
120
65
102
70
70
70
60
60
7
32
60
46
42
10
8
30
22
45
60
75
70
80
50
5
2
20
10
20
50
10
80 72 57 : 25
Bz
nl/1
0
0
0
2
4
1
2
10
12
10
0
0
0
9
3
0
0
5
13
6
25
2
10
10
4
1
0
2
1
5
10
00H
mg/1
270
270
230
190
280
300
285
350
240
1500
275
400
NW
530
400
490
660
620
960
880
850
730
1200
1000
800
500
200
300
300
400
350
HOAc
mg/1
1520'
1810
1550
DMK
Ml/1
11
15
12
1500, 14
1680; 8
1510 8
1480
1520
1500
1400
1100
1400
1000
1380
1120
1460
1300
1050
1100
1100
1150
950
1160
1020
1200
1200
1400
1200
1000
1100
1000
850
1000
1 200 1 1100
2 j 700 '• 1000
80 66 ; 64 3 i 0 800 \ 900
97 80 • 66 64 40 i 10 400 ' 950
98 80 i 56 54 j 30 ' 3 500 ; 900
99 80 ' 60 58 40 ' 5 500 _ 900
100 80
45 . J3 60 5 . 400 i 900
9
7
1
5
3
3
1
14
12
23
1
5
1
20
10
12
12
Bz
i)l/l
„,
<1
--1
<1
<1
<1
<1
<1
<]
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
00H
mg/1
«,
<]
<1
<1
<1
<1
<1
<1
<1
<1
-.}
'1
<1
<1
.
-------
TABLE 4
ACETIC ACID ADSORPTION AND DESORPTION
SUMMARY OF CONDITIONS AND PRODUCT QUALITY
FOR NUCIIAR WV-G CYCLES
FEED ANALYSIS
Cycle
No.
16
17
18
19
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
Flow
Rate
gpm
80
75
85
80
85
70
70
88
80
85
75
90
80
80
80
82
88
68
80
80
95
65
70
70/55
100
95
95
60
95
95
95
90
105
00H
Ads
T °C
37
38
39
42
45
40
45
50
50
52
45
50
41
40
25
35
36
30
35
38
48
38
36
45
70
66
66
30
48
50
45
44
77
55 j 55
55 ' 58
85 ', 40
90 1 45
65 30
60 ; 30
HOAc
Ads
T °C
35
35
36
38
40
36
41
47
47
46
42
47
37
36
19
31
31
25
31
35
45
35
30
20
62
55
55
26
43
46
DMK
i.l/l
5
15
40
5
55
43
10
30
15
60
50
65
55
50
35
7
5
4
6
1
2
13
0
0
0
5
5
0
20
23
3« 71
38 i 50
68 ' 0
42 0
42 | 14
32 5
40 I 5
25 0
20
60-85 j 40 [ 25
58
Sz
i.l/l
1
2
10
i
8
10
6
8
10
15
15
15
10
10
5
3
2
2
1
1
1
2
0
0
0
0
1
0
2
1
16
11
00H
mg/1
400
340
-
-
320
125
140
150
250
200
180
125
150
-
275
225
215
180
195
190
200
260
225
150
400
180
270
375
350
210
315
-
0 650
0
8
345
'155
3 470
1 ' 340
0 , 290
t
0
40 : 0
420
255
HOAc
mg/1
PRODUCT ANALYSIS
DMK
i.l/l
1400
1500
1600
1600
1700
1600
1900
1600
1700
1800
1700
1400
1560
1340
1400
1500
1500
1640
1440
1500
1320
1300
1650
1820
1860
1640
1680
1560
1520
14
11
13
11
8
10
10
9
16
14
12
11
15
8
8
8
8
14
18
22
27
31
4
4
7
13
7
2
4
1510 ' 8
1330 il 18
1340 '• 4
1690 ' 5
1780
7
1820 • -
Bz
i.l/l
,!
<1
<1
<1
-1
-1
<1
<]
--1
<1
<1
,1
-1
•-1
<1
<1
<1
<1
-.-1
-1
<1
-!
•0
<1
<1
-1
<1
.1
<1
<1
-1
00H
mg/l
•0
vl
<1
•0
<1
<1
•1
<1
<1
•0
<1
<1
'1
•0
<]
<1
<1
<1
<1
'1
-1
<1
•0
<1
'1
<1
--1
-1
<1
-1
<1
<1 •'!
<1 -0
<1
<1
-1
HOAc
mg/1
90
60
175
95
106
115
23
150
46
270
300
340
221
356
277
388
320
285
242
235
382
127
126
443
258
1009
448
89
134
397
493
238
180
248
•'I
1980 ' 9 ;
-------
TABLE 4 (Cont'd.)
FEED ANALYSIS !
Flow 00H HOAc >
Cycle Rate Ads Ads DMK Bz 00H HOAci
No. gpm T °C T °C pl/1 ul/1 mg/1 mg/1 1
59 80 48 40 58 0 170 2060;
60 90 55 50 66 0 270 1810(
61 90 55 50 135 0 270 17501
62 75-90 42-48 30-35 120 0 230 1550!
DHK
yl/1
16
16"
9
9
63 90 50 40 45 .1 260 1410J 1Z
64 90 53 46 45 1 410 1500JJ 8
65 90 55 48 70 1 300 1510
66 90 50 45 70 10 350 1520
67 95 55 48 33 0 - 1480
68 100 40 35 85 20 500 1370
69 92 50 45 55 20 250 1380
70 92 55 50 7 0 325 1200
*71 94-50 55 45 44 0 400 1300
72 60 60 50 100 7 635 1250
**73 50 20 500
74 60 55 50 100 50 500 1000
75 60-50 25 20 0 0 1000
76 60 50 40 5 0 500 1140
77 60 50 40 25 10 650 1100
78 60 60 50 5 0 750 1200
79 60 60 52 35 15 1200 900
80 60 55 45 32 5 975 1160
81 60 65 55 7 0 750 1200
82 60 67 55 100 5 - 1250
83 60 62 58 80 10 1000 1200
84 60 65 62 50 4 800 1400
85 60 60 56 5 1 500 1200
86 60 60 56 20 2 400 1400
88 80 66 64 20 2 300 1100
89 80 72 68 10 1 300 1000
90 80 68 64 20 5 400 850
91 80 68 64 50 10 350 1000
92 60 64 60 10 1 200 1100
93 80 72 68 25 2 700 1000
94 80 66 64 3 0 800 900
95 80 66 64 40 10 400 950
96 80 56 54 30 3 500 900
97 80 60 58 40 5 500 900
98 80 45 43 60 5 400 900
99 80 62 58 80 5 500 1000
100 80 60 56 60 2 600 1100
8
12
7
3
1
1
1
9
1
7
3
14
14
23
12
14
1
3
7
5
12
5
2
0
10
12
5
5
PRODUCT ANALYSIS
Bz
ul /I
0.019
0.025
0.021
0.022
0.022
0.016
0.011
0.01
0.01
0.021
0.011
36
-------
A calculation of the organic carbon accounted for in the
acetone and acetic acid in the product is less than the
total organic carbon determined by TOC analysis. This
unaccounted for carbon is probably trace amounts of or-
ganic or inorganic contaminants other than phenol, benzene,
acetone, and acetic acid. The unaccounted for total or-
ganic carbon in the product was in the range of 30 to
100 mg/1.
The data in Tables 3 and 4 have been combined with the
plotting of breakthrough curves to define the operating
and performance limits of the application of activated
carbon adsorption to acetic acid and to compare the per-
formance of Witco 718 and Nuchar WV-G in this application.
Since a distinguishable difference in the two carbons
was obtained, any conclusions as to the operating and
performance limits of the system must be prefaced by the
comparison of the carbons.
Charts comparing the loading of Witco 718 and Nuchar WV-G
as a function of the number of cycles are presented in
Figure 9. These graphs do not distinguish any variables
in the operation and cover all conditions experienced in
the plant. They are an indication of the service life
of each carbon. The general trend of the points in both
plots is downward, and this reflects both the results of
operation at higher temperatures and lower feed concen-
trations in the later cycles and some not unexpected
deterioration in carbon performance. The low loadings
seen in these figures are clearly due to temperatures
>50°C, since the higher loadings, particularly in the
earlier cycles, were obtained at moderate to cool tem-
peratures. In order to get an indication of the capac-
ity of both carbons at lower temperatures after 105
cycles, samples of the carbon in each acetic acid ad-
sorber were withdrawn from a point 4' above the support
plate. Batch equilibrium loadings at 25°C and two ace-
tic acid concentrations were measured. The equilibrium
loadings after 105 cycles for each carbon were as follows:
Equilibrium cone. Loading
Carbon mg acetic acid/1 Ib acetic acid/lb carbon
Witco 718 1270 0.029
1800 0.040
Nuchar WV-G 1075 0.032
1700 0.043
These results, when compared with actual loadings obtained
in the Demonstration Plant under similar conditions, con-
form that both carbons, after 105 cycles, have nearly the
same capacity that was obtained earlier at moderate to
37
-------
00
.07 ,.
.06
.05
.04
I -03
.01
o
I-
UJ
CJ
.n
O
O
O
.07
.06
.05
.04
.03
.02
.01
10
WITCO 718
I I I I I I I I I I I I I
20 30 40 50
60 70
80 90
NUCHAR WV-G
1 1
1 1 1 1
1 1 1 1 1 1
10 20 30 40 50 60
NO. OF CYCLES
70
80
90
100
100
Figure 9
ACTUAL DEMONSTRATION PLANT LOADING OF ACETIC ACID ON WITCO 718
CARBON AND NUCHAR WV-G CARBON AS A FUNCTION OF THE NUMBER OF CYCLES
-------
cool temperatures. There has been very little continuing
deterioration; the effective loading has stabilized.
Early in the operational period, during the first 10-15
cycles, the performance of the two carbons was similar,
generally exhibiting high loadings and good effluent
quality (<100 ppm acetic acid). This is shown in Figures
10 and 11, plots of the breakthrough curves of Cycle No.
11 for both carbons at the same conditions. In these
and subsequent figures, the alignment of the acetic acid
adsorbers is indicated by the notations at the bottom.
The Witco 718 column is 501 and the Nuchar WV-G column is
502. In Figure 10, the Witco column 501 is at first on-
stream behind the Nuchar column 502, then it is online
alone as 502 is regenerated, and then 501 is the lead
column when the regenerated 502 is put back on line.
Shortly after Cycle 11, the quality of the effluent from
the Nuchar WV-G column began to worsen, as evidenced by
levels of acetic acid in the effluent of 200-500 ppm.
Plots of breakthrough curves for the two columns at 70
gal/min and 35°C are presented in Figures 12 and 13; the
former is Cycle No. 25 for Witco 718; Figure 13 is Cycle
No. 22 for Nuchar WV-G. The Witco 718 column continued
to produce high quality effluent with good loading, while
the Nuchar WV-G removed only 70% of the acetic acid.
This trend was observed to continue, and further examples
of the differences in the performance of the two carbons
at equivalent conditions are presented in Figures 14 and
15, breakthrough curves on Cycle No. 30 for both carbons.
It should also be noted that at these more severe con-
ditions, 86 gpm and 45°C, the Witco 718 column is not as
effective as previously, but still better than the Nuchar
WV-G column.
At 55-57 gal/min and 42°C, the Nuchar WV-G column broke
through more quickly and had a much lower loading than
the Witco 718. Curves for these conditions are shown in
Figures 16 and 17, Cycle No. 53 for Witco 718 and Cycle
No. 52 for Nuchar WV-G respectively. A final comparison,
obtained when the plant operated at 95-99 gal/min and
70°C, is presented in Figure 18 and again shows Witco
718 to be better. These nine curves (Figures 1O-18) were
used to illustrate the fact that the Witco 718 column
was able to adsorb acetic acid more completely and in
higher loading than the Nuchar WV-G column over the range
of operating conditions tested. Referring back to Figure
9, the higher loadings achieved with Witco 718 can also
be seen. There was no indication from laboratory equilib-
rium tests that Nuchar WV-G would not perform equal to
or better than Witco 718. Laboratory column studies did
indicate that Witco 718 gave higher loadings at all tem-
peratures.
39
-------
o
z
o
o
g
o
<
o
111
o
Time, hr 0
AVERAGE CONDITIONS
FLOW: 80 gpm
TEMP: 35°C
ACETIC ACID REMOVED = 1340 Ib
LOADING =0.045 Ib/lb
EFFLUENT
501 ALONE
502-»501
Figure 10
FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO 718
ACETIC ACID ADSORBER, CYCLE NO. 11
40
-------
3.0
2.5
2.0
u
o
o
§«
o
LU
O
1.0
0.5
Time, hr 0
AVERAGE CONDITIONS
FLOW: 80 gpm
TEMP: 35°C
ACETIC ACID REMOVED = 980 Ib
LOADING = .042 Ib/lb
EFFLUENT
. ....
12
18
24
30
501-* 502
OFF
502 ALONE
Figure 11
FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR WV-G
ACETIC ACID ADSORBER, CYCLE NO. 11
41
-------
o>
o
2.0
1.5
FEED
AVERAGE CONDITIONS
FLOW: 70 gpm
TEMP: 35° C
ACETIC ACID ADSORBED = 1800 Ib
LOADING = 0.061 Ib/lb
EFFLUENT
.1-1.1.1.1
Time,
Flow,
hr 0
gpm 56
502-*5C
4 8 12 16 20
44 65 67 70 72
1 T-501 ALONE
• /*>
24 28
78 72
501-^502
|
32 31
68 6(
FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO 718
ACETIC ACID ADSORBER, CYCLE NO. 25
42
-------
2.5
2.0
1.5
o
Cycle No. 22
Average Conditions
Flow: 70 gpm
Temp: 35°
Acetic Acid Adsorbed = 1430 Ib
Loading = 0.048 Ib/lb
FEED
Flow, gpm 98
4
80
8
78
T-501-
12
71
- 502
16
74
20 24 28
50 71 72
T-502 ALONE
32
72
36
62
501-
40
96
-502
44
109
Figure 13
FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR WV-G
ACETIC ACID ADSORBER, CYCLE NO. 22
43
-------
2.5
2.0
AVERAGE CONDITIONS
.FLOW: 86gpm
, J TEMP: 45°C
AC'ETIC ACID ADSORBED = 1168 Ib
LOADING = 0.04lb/lb
FEED
Time, hr 0
Flow, gpm 84
502
30
501 ALONE
501-^502
STOP,
REGENERATE
Figure 14
AND; EFFLUENT CONCENTRATION CURVE FOR WITCO 718
\ACETIC ACID ADSORBER, CYCLE NO. 30
44
-------
2.5
2.0
AVERAGE CONDITIONS
FLOW: 90 gpm
TEMP: 45°C
ACETIC ACID ADSORBED = 800 Ib
LOADING = 0.03 Ib/lb
r 1.5
o
z
o
o
9
o
o
I-
LLJ
Time, hr 0
Flow, gpm 85
1.0
501-*>502
502 ALONE
502-^501
Figure 15
FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR WV-G
ACETIC ACID ADSORBER, CYCLE NO. 30
45
-------
2.0
1.5
u
Ill
AVERAGE CONDITIONS
FLOW: 57 gpm
TEMP: 42°C
ACETIC ACID ADSORBED = 1250 Ib
LOADING = 0.042 Ib/lb
FEED
T-501 EFFLUENT
T-502 EFFLUENT
Time, hr 0
Flow, gpm 57
8
54
16
58
24
57
32
502-^501
501 .ALONE
Figure 16
FEED AND EFFLUENT CONCENTRATION CURVE FOR WITCO 718
ACETIC ACID ADSORBER, CYCLE NO. 53
46
-------
3.0
2.5
2.0
01
h
6
O
o
1.5
LU
u
1.0
AVERAGE CONDITIONS
FLOW: 55 gpm
TEMP: 42°C
ACETIC ACID ADSORBED = 710 Ib
LOADING =0.026 Ib/lb
Time, hr 0
Figure 17
FEED AND EFFLUENT CONCENTRATION CURVE FOR NUCHAR WV-G
ACETIC ACID ADSORBER, CYCLE NO. 52
47
-------
3.0
2.5
2.0
0)
O
O
§1.5
O
l-
LU
O
1.0
0.5
CYCLE NO.
FLOW, gpm
TEMP., °C
ACETIC ACID
ADSORBED, Ib
LOADING, Ib/lb
WITCO 718
(T-501)
52
99
70
750
.026
NUCHAR WV-G
(T-502)
51
95
70
480
.018
501-*502
502 ALONE
501 ALONE
Figure 18
FEED AND EFFLUENT CONCENTRATION CURVES FOR WITCO 718
ACETIC ACID ADSORBER, CYCLE NO. 52, AND NUCHAR WV-G
ACETIC ACID ADSORBER, CYCLE NO. 51
48
-------
Having established the better performance of Witco 718
carbon in acetic acid adsorption, the operating and per-
formance limits for it can now be defined. The capability
of the acetic acid adsorption system is best represented
by the data collected in the first 6 mos. of the grant
during colder weather when cooling of the brine was the
maximum. As expected from the laboratory data, the most
important variables affecting adsorption of acetic acid
were temperature and flow rate. A summary of the oper-
ating results on the Witco 718 column is presented in
Table 5, showing the effect of temperature and flow rate
on acetic acid adsorption. At temperatures greater than
50°C, the loading is too low (0.02-0.03 lb/lb) to be
practical. Even at low flow rates of 50-60 gal/min, the
low loading produces rapid breakthrough and poor effluent
quality. At temperatures of 4O-45°C, the loading is high
enough and is practical (0.04 Ib/lb) and acetic acid
levels<100 mg/1 can be achieved at flow rates between 50
and 80 gal/min. At temperatures below 40°C, good loading
and effluent quality are enhanced. The only significant
regime not tested in this grant period was temperatures
below 40°C, and flow rates greater than 9O gal/min. It
can be concluded that the acetic acid adsorption system
can be operated on brine containing 1500-180Q ppm acetic
acid at flow rates up to 80 gal/min and temperatures up
to 40-45°C; loadings will be ~0.04-0.06 Ib! acetic acid/lb
of carbon, and the effluent will contain less than 200
mg/1 of acetic acid, and quite likely less than 100 mg/1,
at the less severe conditions. This performance data can
be used to calculate some plant capabilities. For a feed
brine containing 1500 mg/1 acetic acid adsorbed on 8' x
19' carbon column at ^80 gpm and £40°C, a loading of 0.04
Ib/lb can be attained and 100,000 gal of brine will be
purified to lower than 100 mg/1 acetic acid concentrations.
Regeneration of the saturated acetic acid adsprption columns
was by desorption with 10% caustic at flow spates of 30 to
60 gal/min. Studies were directed toward minimizing the
amount of caustic used and the volume of regenerant. On
those cycles when good loadings of acetic acid (0.04-0.06
Ib/lb) were achieved, it was necessary to use at least
360O to 4200 Ib of NaOH. When lower loadings (0.02-0.03
Ib/lb) were experienced, 2400 Ib of NaOH was sufficient to
regenerate the column. Curves showing the desorption of
acetic acid as sodium acetate and the concentration profile
of the caustic in the regenerant are presented in Figures
19 and 20 for the Witco 718 and Nuchar WV-G adsorbers.
These are typical of the results obtained when using fresh
caustic as the regenerant. The peak acetic acid cpncen-
tration in the regenerant, expressed in g/l'> was 55 to
65. This is the normal range for columns having a high
49
-------
TABLE 5
SUMMARY OF OPERATING RESULTS FROM ACETIC ACID ADSORPTION
ON WITCO
Temp, °C
>50°
40-45°
<40°
718 SHOWING
Flow Rate
gal/min
60
80
100 0.
60
80
80-90
90-100
<80 0.
80
EFFECT
Loading
Ib/lb
0.02
0.02
02-0.026
0.042
0.04
0.04
04-0.06
0.045
OF TEMPERATURE AND
Acetic Acid
Concentration
in Effluent mg/1
300-6OO
30O-600
~5OO
25-100
6O-100
220-280
ISO- 330
60-225
<1OO
FLOW RATE
Reference
Cycle No.
in Table 3
90-100
78-85
43,52,65
20,24,53
18,27
28,32
26,30,31,48, 51,
63,64
24,25,33,38,40,
42,46,47
11
Note: 10O gal/min is equivalent to 2 gal/min/sq ft and a
detention time of 60 minutes.
50
-------
70
60
D)
50
<
DC
Z
ill
O
z
o
o
p
o 30
o
40 -
u
<
20
10
REGENERANT: 4200 GAL 10% NaOH
FLOW RATE: 35 gpm
TEMP: ~ 40°C
ACETIC ACID REMOVED = 1740 Ib
ACETIC ACID
SODIUM HYDROXIDE
— 12
10 z
g
H-
-------
14
REGEIMERANT: 4200 GAL 10% NaOH
FLOW RATE: 50 gpm
TEMP.: ~ 40°C
ACETIC ACID REMOVED = 860 Ib
ACETIC ACID
SODIUM HYDROXIDE
1234567
THOUSAND GALLONS OF REGENERANT
Figure 20
REGENERATION CURVE FOR NUCHAR WV-G ACETIC ACiD ADSORBER,
CYCLE NO. 21, USING FRESH CAUSTIC
52
-------
acetic acid loading, and it was unaffected by flow rate.
When the loading is less, the peak acetic acid concen-
tration is lower. The volume of regenerant was ~6,000
gal. This is a good average number, and it is not changed
appreciably when 2400 Ib NaOH are used on low loading
cycles. The sodium hydroxide content in the regenerant,
as anticipated, was essentially a mirror image of the
acetic acid concentration.
The presence of the unused caustic in the tail of the re-
generant led to the studies on saving the last few thou-
sand gallons of regenerant, fortifying it with strong
caustic to 10%, and reusing it as regenerant on the sub-
sequent cycle. Typical regeneration curves resulting
from this method of operation are shown in Figure 21.
The curves are somewhat flatter than those in Figures 19
and 20 even though there was not much difference in the
amount of acetic acid desorbed. Several combinations of
regenerant saved and the amount of strong caustic used
to fortify it were tested. The total amount of caustic
regenerant fed to the column was varied from 2300 gal to
4500 with no appreciable effect on the net volume of re-
generant for disposal. It was typically 35OO-4000 gallons.
Thus, it was possible to reduce the amount of regenerant
from each cycle to 0.5-0.6 bed volumes or 4-4.5% of the
feed volume.
Recycle of weak regenerant has one drawback - the level
of acetic acid in the weak regenerant can begin to build
up and make the regeneration less efficient. This is best
handled by periodically using fresh regenerant for a few
cycles.
The amount of caustic soda used per cycle is not affected
by recycle of regenerant. The optimum for a column having
a high loading of acetic acid was ~3600 Ib NaOH. This
represents a caustic requirement for regeneration of 36
Ib NaOH/1000 gal. The amount of HC1 used per cycle was
8 lb/1000 gal for pH adjustment and 6 Ib HC1/1000 gal for
regeneration.
Evaluation of Materials of Construction
In one year of operation, the following failures in equip-
ment were experienced:
1) Ductile cast iron pump in 100°C, 18% NaCl, pH
7-9 service failed in three months by erosive
and corrosive attack. It was replaced with a
cast 316 stainless steel pump and very low ero-
sive or corrosive wear were experienced in the
next nine months.
53
-------
70
60
50
en
J 40
O
u
o 30
H
UJ
a
20
10
REGENERANT: 2400 GAL OF WEAK REGENERANT
FORTIFIED WITH 1200 GAL OF 30% NaOH
ACETIC ACID
CYCLE NO. CARBON FLOW TEMP REMOVED, Ib
• 29 WITCO 718 52 40°C 1650
25
NUCHAR WV-G 53
40°C
1050
0 1234567
THOUSAND GALLONS OF REGENERANT
Figure 21
REGENERATION CURVES FOR ACETIC ACID ADSORBERS USING
RECYCLE OF FORTIFIED WEAK REGENERANT
8
54
-------
2) Ductile cast iron pump in 60°C, 18% NaCl, pH 6-9
service failed in 11 months by erosive and corro-
sive attack. It was replaced with a cast 316
stainless steel pump.
3) The rubber-lined agitator in acid brine service
(pH 2-3) failed because a pinhole developed,
caused by flexing of the agitator shaft. The
steel was then attacked and one blade fell off
in 10 months service. A replacement rubber-lined
agitator failed by the same mechanism in 6 weeks.
Rubber lining is unsuitable for agitator blade
service although it is satisfactory for tank lin-
ings. A replacement, all-titanium agitator was
ordered and installed. In the interim, the tank
was agitated by recirculating its contents using
a polypropylene-lined centrifugal pump,
4) Chlorinated organics periodically present in the
crude HC1 caused failure in polypropylene-lined
pipe. This pipe was replaced with FEP polymer
lining.
5) Although the air cooled heat exchanger did not
suffer corrosive failure, finely divided in-
organic sludge that did not settle in the clar-
ifier plugged the passages and increased pressure
drop and decreased heat transfer efficiency. This
problem was corrected by a standard acidizing
treatment.
Corrosion test coupons installed in a 4" pipeline con-
taining 100°C salt brine at pH 7-9 gave a corrosion rate
of 0.007 inches/year. Therefore steel is a suitable
material for pipelines with moderate velocites. However,
centrifugal pumps which operate at high velocity are
subject to a combined erosive and corrosive attack and
fail in the same service. At the corrosion rates ex-
perienced, we would project a service life in excess of
10 years for the major equipment installed in the plant.
Costs
Capital
The total installed capital cost for this 100 gal/min
Demonstration Plant was $592,000. This was broken down
as follows:
55
-------
Major equipment purchases
Installation of equipment
Instrumentation
Engineering
Carbon, initial charge
$138,100
273,700
72,000
75,OOO
33,200
$592,000
Approximately one-half of this cost can be allocated to
the phenol adsorption system(and one-half to the acetic
acid adsorption system. An additional $44,200 was spent
in constructing and installation of the chlorine test
cells.
Operating
The significant items affecting the operating costs for
this adsorption system are carbon life (or carbon makeup
requirements), chemicals for pH changes and regeneration,
labor, maintenance, utilities, and depreciation. Carbon
life has a linear effect on carbon makeup costs and is
important to the economics of wastewater purification by
activated carbon with caustic regeneration.
Phenol Plant Brine Purification
The Demonstration Plant does not have adequate cooling to
operate at 1OO gal/min capacity. A feed brine cooler ex-
changing heat between the cool product brine and the hot
feed brine would be a required additional capital expense
to assist the air cooler particularly in the summer. Also,
it would be desirable to add instrumentation to automate
the system to eliminate the need for full-time operators.
We project that labor needs could be met with incremental
labor. The additional capital required is estimated to
be $100,000, bringing the total to $692,000.
With these modifications, the cost of purifying 1OO gal/min
of Phenol Plant brine by a two-stage carbon adsorption
prpcess can be projected.
Plant capacity =1(g.gal x 1440min x 36 5 days x O.9O operating
day
= 47 , 500,000 gal/yr
factor
From previously derived figures for the volume of brine
processed per cycle (928,000 gal for phenol and 100,000
gal for acetic acid) the annual number of cycles required
for each of the phenol jind each of the acetic acid ad-
sorbers processing 47.5M gal of brine is 25 and 237.
56
-------
The Demonstration Plant operated through 14 cycles on
phenol and 105 cycles on acetic acid and was still ef-
fective. It is reasonable to assume that the effective
carbon life for the optimized process would be one year.
For the phenol columns, the carbon life can quite likely
be even longer. Since each carbon column contains $8300
worth of carbon, the carbon makeup cost if replaced after
one year would be $0.35/M gal for phenol and $0.35/M gal
for acetic acid adsorption.
The projected chemical costs for the optimized two-stage
adsorption system are:
NaOH for phenol regeneration: 20 Ib/M gal® $70/ton=$0.70/M gal
for acetic acid regener-
ation
HC1 for process
36 lb/Mgal@ $70/ton=$1.26/M gal
14 Ib/M gal® $30/ton=$0. 21/M gal
$2.17/Mgal
By recycle to the Phenol Plant, the sodium phenate and
the caustic used for phenol regeneration are recovered.
The recovery value of the phenol is $0.35/M gal and
caustic is $0.70/M gal; therefore, the recycle of the
phenol regenerant reduces net raw material costs by
$1.05M/gal.
Actual maintenance and utilities costs were measured over
six months of routine operation, and the annual maintenance
costs (materials and labor) were projected to be $25,000.
The annual utilities costs were projected to be $8600.
The operating costs are tabulated in Table 6 per 1000 gal
of brine purified.
TABLE 6
PROJECTED COSTS FOR PURIFICATION OF 100 GAL/MIN OF
PHENOL PLANT WASTE BRINE
Depreciation @ $69,200/yr
(10 yr straight line)
Maintenance @ $25,000/yr
Utilities @ $8600/yr
Chemicals
Labor(incremental with automation)
Carbon Makeup
Total Projected Costs
Credits
Recovery of phenate regenerant
Salt recycle
Reduction in brine disposal cost
Total Credits
Net Operating Cost
$71000 gal
1.45
0. 52
0.18
2.17
0
O. 7O
$5.02/M gal
1.05/M gal
1.00/M gal
1.65/M gal
$3.70/M gal
$1.32/M gal
57
-------
The major cost items are seen to be depreciation, the
caustic soda required for acetic acid regeneration, and
carbon.
The above cost data can be manipulated to develop treat-
ment costs for separate removal of phenol and acetic acid.
By assigning one-half of the capital to the phenol ad-
sorption system and one-half to the acetic acid adsorption
system, the operating cost per Ib of phenol recovered
develops as follows:
Basis: Recovery of 212,000 Ib phenol/yr from 100 gal/min
waste containing 535 ppm
$/lb phenol
Depreciation @ 10 yr straight line
$692,000 T 2 x l/10=$34,600/yr $0.163
Maintenance, $12,000/yr 0.056
Utilities, $4,300/yr 0.020
Chemicals, no cost if recovered 0
Labor 0
Carbon Makeup @ $0.35/M gal 0.078
Recovery cost/lb phenol $0.317
Similarly, the unit cost for removing acetic acid from a
wastewater by carbon adsorption can be calculated:
Basis: Removal of 9O% of the acatic acid in 10O gal/min
waste containing 1500 mg/1, or 535,000 Ib/yr.
$/lb acetic acid
Depreciation, $34,600/yr $0.065
Maintenance, $12,000/yr 0.022
Utilities, $4,300/yr O.008
Chemicals, $1.47/M gal 0.125
Labor 0
Carbon Makeup® $O.35/M gal O.O31
Treatment cost/lb acetic acid $0.251
58
-------
SECTION VII
CHLORINE TEST CELLS
Construction
The two 2660 ampere chlorine cells for testing phenate
brine were built using standard cathode pockets and
graphite anodes. Scaling down of the cells to test-size
was accomplished by reducing the number of pockets and
anodes. The cells were set up at the Chlorine Production
Plant. D.C. power was supplied by a Udylite model 4ADW-
5M012-101 constant current rectifier. Each cell had a
saturator in the feed brine system.
Feed brine for Cell 1 operating on pure brine was made
by dissolving salt in steam condensate and required no
treating. Cell 2 which operated on purified phenate
brine had a 200 cu ft batch settling tank for the removal
of calcium, magnesium and iron impurities, and a storage
tank for finished brine.
Additional equipment associated with the cells included
automatic head controls with a brine pump and a heater
for each cell and a Beckman 315A infrared continuous C02
analyzer for the chlorine.
Operation
The two cells were started up September 1, 1970, using
regular chlorine plant feed brine. After a three-week
checkout period to make sure the cells were operating
normally, feed streams to the cells were switched to
the test brines.
Cell 1 - Pure brine made from salt and condensate to
serve as a reference.
Cell 2 - Phenate brine purified in the Demonstration
Plant.
The cells were switched to regular chlorine plant feed
brine for three weeks in January and again in June to
check for permanent changes caused by the test brine.
Results
Based on six ampere laboratory cell test results as well
as plant experience attempting to run unpurified phenate
brine through production cells, the most likely problem
areas were defined as follows:
59
-------
1. Brine treating
2. Diaphragm plugging
3. Loss of current efficiency
4. Voltage
5. Graphite wear
6. Product quality
Analyses of the test brines, chlorine and caustic as well
as other pertinent cell variables averaged over the entire
test period are given in Table 7. These data are based
on feeding undiluted phenate brine, so that any dilution
of the phenate brine with brine from other sources would
produce a corresponding reduction in the effects on cell
operation.
The most likely problems listed above are discussed in
the following paragraphs. These were the only areas
where the purified phenol brine was observed to have
any effect on cell operation.
1. Brine treating
The feed brine analyses in Table 7 indicate that
there was some difficulty in treating the iron
out of the phenate brine. The high iron resulted
from the tendency of organics to complex iron and
reduce settling rates. The problem is less with
purified phenate brine than with raw phenate
brine, and the brine plant has the capability to
produce brine with less than 1 ppm iron when puri-
fied phenate brine is recycled. No difficulty
was experienced in removing either calcium or
magnesium.
2. Diaphragm Plugging
The slightly higher head and caustic strength of
Cell 2 indicate some diaphragm plugging. This
amount of plugging is normal for brine which con-
tains high iron. There is no evidence that the
organics in the phenate brine directly plugged
the diaphragm. When the cells were switched from
the test brines to regular plant feed brine there
were no sudden changes in caustic strength or
voltage.
3. Loss of Current Efficiency
The current efficiency of the two cells was iden-
tical, and the current efficiency loss associated
with raw phenate brine in earlier tests was com-
60
-------
TABLE 7
AVERAGE ELECTROLYTIC TEST CELL OPERATING DATA
Type feed brine
Brine Analyses
NaCl g/1
SO,, g/1
Ca^ ppm
Mg++ ppm
Fe ppm
Organic Carbon ppm
Inorganic Carbon ppm
Sodium Acetate ppm
Phenol ppm
Acetone ppm
Days Operated
Volts
Current, amperes
Temp. °C
Head, inches
Chlorine Current Efficiency
C12 Analyses (air free)
%C12
%C02
%02
%H2
%COC12
Caustic Analyses
NaOH g/1
NaCl g/1
NaC103 g/1
Organic Carbon ppm
Sodium Acetate ppm
Phenol ppm
Acetone
Cell 1
Pure brine
311
0.83
1.7
0.2
0.3
16
14
<5
0
<2
206
3.01
2660
69
6.5
97.90
97.87
0.49
0.60
1.04
nil
97
197
0.084
29
18
1.4
<2
Cell 2
Purified phenate brine
310
1.6
2.5
0.2
1.4
280
47
553
4
8
210
3.11
2660
70
8
97.81
98.49
0.85
0.32
0.38
0.001
1O2
191
0.107
223
48O
1.9
<2
61
-------
pletely eliminated. Experience with raw phenate
brine recycle to chlorine cells showed that oxi-
dation of phenol and acetone .in the feed brine
produce most of the efficiency loss. These com-
pounds - particularly phenol - are very effec-
tively removed in the Demonstration Plant.
4. Voltage
The higher voltage of Cell 2 was associated with
difficulty in removing the iron in brine treat-
ing and subsequent plugging of the diaphragm.
Dilution of the brine in plant operation improves
the brine treating and reduces the tendency for
the voltage to increase.
5. Graphite Wear
There was no significant difference in graphite
wear between the cells. The slightly higher C02
content of the chlorine with purified phenate
brine (which normally is formed from the graphite
and indicates graphite wear) resulted from the
oxidation of organics in the feed brine. The
extent to which this takes place is measured by
the difference in organic content of the feed
brine and the cell effluent.
6. Product Quality
Infrared analyses of the chlorine for organics
showed similar spectrum for the cells operating
on pure brine and purified phenate brine. Chlo-
rinated aromatics, which would cause a great
deal of trouble in the chlorine compression equip-
ment were not present in the chlorine. Although
a small quantity of phosgene is produced with
purified phenate brine, the amount will not be
a problem.
The organic contamination of the caustic is largely a
problem of meeting the specifications of the product
for sale. Samples of the cell effluent produced no
unusual foaming problems when evaporated to 50%. The
caustic from Cell 1 which operated on pure brine con-
tained 29 ppm organic carbon, up slightly from the 16
ppm in the feed brine. Oxidation of the graphite anode
is the source of the additional organic carbon. The
organic carbon in the caustic from Cell 2 comes pri-
marily from the feed brine. Much of this organic car-
bon is in the form of sodium acetate. About 90% of the
sodium acetate in the feed brine enters into no reaction
in the cell and ends up as a contaminant in the caustic.
62
-------
The sodium acetate content of the purified phenate brine
fed to Cell 2 varied from 77 to 1000 ppm during the run,
depending on the flow rates and conditions of the towers
at the Demonstration Plant. The high acetate level seemed
to have no harmful effect on cell operation other than
contamination of the caustic soda effluent.
Conclusions
Waste phenate brine can be used as chlorine cell feed
without severe operating penalties if it is purified
equivalent to that treated in the Demonstration Plant.
Problems that were experienced when raw phenate brine
was used, but are eliminated with the purified brine are:
1. Current efficiency loss.
2. Organic contaminants in the chlorine causing
plugging of compression equipment.
Problems that are reduced but not completely eliminated
by purification in the Demonstration Plant are:
1. Interference of organics with brine treating
(particularly iron).
2. Organic contamination of caustic.
63
-------
SECTION VIII
LABORATORY STUDIES
Phenol Adsorption and Regeneration
Batch equilibrium studies were conducted to provide guid-
ance for the selection of the type of carbon and for the
operational phase of the Demonstration Plant as to the
effects of certain critical variables. Carbon from 8
manufacturers, and a total of 12 types were initially
screened for phenol capacity with the results shown in
Figure 22. Most of the carbons were equivalent in these
virgin carbon tests. Because of its superior performance
for acetic acid adsorption, Witco 517 was selected for
further studies. Phenol adsorption and desorption ki-
netics were measured, and 90% equilibrium was reached in
1 hour regardless of particle size. Rates at 60°C were
equivalent to rates at 25°C. The effect of pH on phenol
loading was examined over the range of 0.5 to 12.0. The
results in Figure 23 shows a maximum loading in the pH
range 6.0 to 8.5 at 25°C. Next, the effect of caustic
strength on phenol loading was tested on virgin Witco 517.
The optimum caustic soda strength at 25° and 60°C for
desorption was determined to be around 4% as seen in
Figure 24. This batch equilibrium test was conducted
by shaking a fixed initial concentration with the same
amount of carbon in various caustic soda concentrations.
Thus, the equilibrium concentration of phenol varied.
When this effect was recently checked again, keeping the
equilibrium phenol concentration constant, the optimum
caustic strength was less pronounced as shown in Figure 25.
The optimum range has broadened to at least 2 to 8% NaOH
at phenol concentrations of 400 and 1600 mg/1.
With the selection of Witco 718 and Nuchar WV-G activated
carbons as those to be loaded into the Demonstration
Plant columns, and the subsequent discovery of the need
to acid-wash the acid soluble compounds from the carbon,
phenol adsorption and desorption isotherms at various
temperatures were determined. The isotherms for phenol
adsorption on acid-washed Witco 718 in 20% brine at pH 7
at 25°, 40°, and 60°C are presented in Figure 26. The
40° and 60°C lines coincide for all practical purposes.
Similar isotherms at 25° and 60°C for acid-washed Nuchar
WV-G are shown in Figure 27. Noteworthy is the observed
crossing of the two lines.
The isotherms for sodium phenate adsorption (phenol de-
sorption) on acid-washed Witco 718 and Nuchar WV-G are
presented in Figures 28 and 29 respectively. Temperature
65
-------
CARBON TYPE
NUCHAR WV-G
12 x 40 MESH
NUCHAR WV-L
8x30 MESH
NUCHAR WV-W
20 x 50 MESH
FISHER-COCONUT
50-200 MESH
PITTSBURGH CAL
PITTSBURGH SGL
COLUMBIA
TS-570
PITTSBURGH CPG
14 x 40 MESH
DARCO
20 x 40 MESH
WITCO 517
20 x 40 MESH
BARNABEY-
CHANEY
SK-8623
IONAC
P-50
Y////////////////////////A
V7/77/7/////////////////A
/////////////////////////A
Y////////////////////////A
Y////7///////////////////A
''//////////////////A
Y///////////77/7/////////A
V/////////////////A
0 .10 .11 .12 .13 .14 .15
EQUILIBRIUM LOADING AT 15-50 mg/l , WT. PHENOL/WT. CARBON
Figure 22
EQUILIBRIUM PHENOL LOADING FOR VARIOUS
VIRGIN ACTIVATED CARBONS
66
-------
.16
.14
.12
cc
o .10
J£ .08
Q.
Z .06
Q
.04
.02
EQUILBRIUM
CONC. = 40 mg/l
TEMP. 25°C
WITCO 517, 20 x 40 MESH
I I
2 4 6 8
pH AT EQUILBRIUM
Figure 23
EQUILIBRIUM LOADING OF PHENOL ON CARBON vs pH
10 12
67
-------
WITCO 517 CARBON, 20 x 40 MESH
INITIAL PHENOL CONCENTRATION: 1450 mg/l
EQUILIBRIUM
CONG. = 750 mg/l
EQUILIBRIUM
CONC. = 1100 mg/l
12
% CAUSTIC
Figure 24
EQUILIBRIUM LOADING OF PHENOL ON CARBON vs % CAUSTIC
68
-------
.08
o
CO
cc
o
(-'
LU
I
Q.
(D
Z
O
<
o
cc
00
a
LU
WITCO 718 25°C
.07
1600 mg/l PHENATE
.06
.05
.04
400 mg/l PHENATE
.03
.01
0
1
8
2 34567
WT. % MaOH
Figure 25
EQUILIBRIUM LOADING OF PHENOL ON CARBON vs % CAUSTIC
AT TWO PHENATE CONCENTRATIONS
10
69
-------
1.00
.80
.60
z
o
CO
cc
.40
.20
LU
I
Q.
§.10
CD
Z .08
Q
<
O .06
£.04
O
LU
.02
.01
0 1 10 100 1000 10,000
mg PHENOL/I IN SOLUTION AT EQUILIBRIUM
Figure 26
EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED WITCO 718 CARBON
IN NEUTRAL 20% SODIUM CHLORIDE SOLUTION
70
-------
z
o
OQ
DC
O
1.00
.80
.60
.40
.20
LU
X
Q.
Z .08
O .06
2
D
CC nyi
O
111
.02
• 25° C
• 60° C
I
I
1
10,000
10 100 tOOO
mg PHENOL/I IN SOLUTION AT EQUILIBRIUM
Figure 27
EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED NUCHAR WV-G
CARBON IN NEUTRAL 20% SODIUM CHLORIDE SOLUTION
71
-------
1.00
.80
.60
.40
z
o
CD
CC
<
O
O
Z
LU
X
a.
.20
- .10
g
o .08
.06
2 .04
cc
CO
O
ui
.02
.01
• 25°C
o 40° C
• 60° C
I I
I I
I I
10 100 1000 10,000 100,000
mg/l PHENOL IN SOLUTION AT EQUILIBRIUM
Figure 28
EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED WITCO 718 CARBON
IN 2-4% CAUSTIC
72
-------
1.00
.80
.60
.40
O
CO
cc
<
o
LU
I
Q.
z
o
D
O
HI
.20
.10
.08
.06
2 .04
cc
QQ
25°C
60°C
.02
.01
100 1000 10,000
mg/l PHENOL IN SOLUTION AT EQUILIBRIUM
Figure 29
EQUILIBRIUM LOADING OF PHENOL ON ACID-WASHED
NUCHAR WV-G CARBON IN 2-4% CAUSTIC
100,000
73
-------
has no effect for Witco 718. From these four curves,
the loading that can be expected in column operations
can be calculated if nearly equilibrium conditions are
attained. The effective loading for each carbon is the
difference between the adsorption and desorption loadings
at the appropriate conditions. For example, if the phe-
nol columns are regenerated at 60°C until the phenol con-
centration in the effluent is 500 mg/1, then the effective
loadings for each carbon in near equilibrium with 200
mg/1 feed at 60°C is calculated as 0.16 Ib/lb for Witco
718 and 0.15 Ib/lb for Nuchar WV-G. At 600 mg/1 feed at
60°, the values are 0.2 for Witco 718 and 0.24 for Nuchar
WV-G.
These batch equilibrium data on phenol adsorption and re-
generation, were available for use in comparison with the
actual performance of the Demonstration Plant.
Acetic Acid Adsorption and Regeneration
Similar to the program followed in laboratory studies of
phenol adsorption, batch equilibrium studies on acetic
acid adsorption were conducted to determine the best
types of carbon to use in the Demonstration Plant and
to examine further the acetic acid adsorption information
which had been developed by earlier research and led to
the selection of a two-stage carbon adsorption and caus-
tic regeneration process for purification of this waste
brine. The same sources and types of virgin carbon were
initially screened for acetic acid capacity at 25°C with
the results shown in Figure 30. Contrary to the results
with phenol, significant differences to adsorb acetic
acid were observed, with Witco 517, Nuchar WV-G, and
Pittsburgh CAL showing the best results. Witco 517 was
selected for further tests to check the effect of sev-
eral variables. The kinetics of acetic acid adsorption
and desorption were measured, and as expected with a
small molecule such as acetic acid, they were rapid.
Adsorption was essentially complete in 15 minutes for
two different mesh sizes. Desorption was a little slower,
requiring 30 minutes to reach equilibrium. The effect
of pH on acetic acid loading was examined over the pH
range 0 to 12, with the results shown in Figure 31. The
optimum pH range is between 3 and 4. This curve also
shows that acetic acid (as sodium acetate) loading on
carbon at high pH values is very low,
-------
CARBON TYPE
FISHER-COCONUT
50-200 MESH
NUCHAR WV-G
12 x 40 MESH
NUCHAR WV-L
8 x 30 MESH
NUCHAR WV-W
20 x 50 MESH
PITTSBURGH SQL
PITTSBURGH CAL
DARCO
20 x 40 MESH
COLUMBIA
TS-570
WITCO 517
20 x 40 MESH
WITCO 718
12 x 30 MESH
IONAC
P-50
PITTSBURGH CPG
14 x 40 MESH
BARNABEY-
CHANEY
SK-8623
/////////A
0 .01 .02 .03 .04 .05 .06
EQUILIBRIUM LOADING AT 2000-2300 mg/l
WT. ACETIC ACID/WT. CARBON
Figure 30
EQUILIBRIUM ACETIC ACID LOADING FOR VARIOUS VIRGIN
ACTIVATED CARBONS
.07
75
-------
.08
.07
.06
z
o
CO
cc .05
<
(J
D
0-04
<
o
LJJ
O
.03
.02
.01
• EQUILIBRIUM CONCENTRATION =
1400 mg ACETIC ACID/I
WITCO 517 CARBON, 20 x 40 MESH
25°C.; HCI OR NaOH ADDED
0 24 6 8 10 12
pH
Figure 31
EQUILIBRIUM LOADING OF ACETIC ACID ON WITCO 517 CARBON vs pH
76
-------
With the selection of Witco 718 and Nuchar WV-G carbons
for the Demonstration Plant, adsorption isotherms were
measured. At first, the carbons were run at 25°C and
as received. The resultant isotherms are the top two
curves in Figure 32. These data along with laboratory
column evaluations formed the basis for comparison of
the actual performance of these carbons in the Demon-
stration Plant. Recently, isotherms at 25°C and 60°C
were measured on Witco 718 and Nuchar WV-G that had been
acid-washed; they are also plotted on Figure 32. Sur-
prisingly, these results indicate a marked decrease in
the capacity of both carbons for acetic acid when they
are acid-washed, and in all cases, Nuchar WV-G showed up
better than Witco 718. These equilibrium data on the
acid-washed carbons are not in agreement with the results
obtained in the Demonstration Plant at lower temperatures.
The loadings obtained at 25°-40°C were 0.04 to 0.06, much
higher than the isotherm would predict.
Early in the grant period, acetic acid adsorption on
Witco 517 and desorption with caustic was studied in a
1" I.D. x 8' high fixed bed. When treating a brine con-
taining a high quantity of acetone (~350 ppm) the per-
formance of the carbon with repeated adsorption-regen-
eration cycles slowly deteriorated. Loading decreased
and the concentration of acetic acid in the effluent in-
creased. This was postulated to be the result of the
adsorption of acetone on the carbon and its subsequent
conversion by a condensation reaction under caustic re-
generation conditions to diacetone alcohol and larger
molecules, resulting in gradual filling of the carbon
pores with molecules not desorbable with caustic. This
problem with acetone in the feed led to the setting of
tight feed brine specifications on acetone content.
A composite sample was made of each of the truckload
shipments of the actual carbons loaded into the Demon-
stration Plant columns. Part of this composite was
evaluated in 1" I.D. x 8T high columns in the laboratory.
The conditions used in this evaluation were a brine feed
containing 1500 mg/1 acetic acid at a flow rate of 2.0
gal/min/sq ft at pH 2 to 4. The results are presented
in Table 8. In both evaluations, the carbon was not
acid-washed prior to use. With Witco 718, this presented
no problems. When this step was omitted for the Nuchar
WV-G, the effluent brine from the first 3 cycles was of
poor quality due to high pH. By the 4th cycle, perform-
ance was as expected. The data in Table 8 clearly show
that the effective loading at a given temperature soon
reaches a fairly constant value. Witco 718 exhibits a
higher capacity than Nuchar WV-G. The effect of temper-
77
-------
z
o
OQ
CC
o
01
o
z
o
<
O
CC
m
a
ai
.10
.09
.08
.07
.06
.05
.04
.03
25°
.02
WITCO 718
• WITCO 718 ACID
WASHED
• IMUCHAR WV-G
ACID WASHED
A NUCHAR WV-G
.01
II
100 200 400 1000 2000 4000
mg ACETIC ACID/I IN SOLUTION AT EQUILIBRIUM
Figure 32
EQUILIBRIUM LOADING OF ACETIC ACID ON WITCO 718
AND NUCHAR WV-G CARBONS
10,000
78
-------
TABLE 8
LABORATORY COLUMN EVALUATIONS OF
DEMONSTRATION PLANT CARBON COMPOSITES
Column size: 1" x 8f
Flow : 2 gal/rain/sq ft
WITCO 718, 12 x 30 Mesh, Composite
Acetic Acid Effluent Quality
Cycle No. Temp. Loading, Ib/lb ppm NaOAc ppm TOG
1 25 0.072 51 28
2 25 0.063 65 41
3 25 0.058 67 41
4 25 0.055 150 64
5 25 0.046 321 120
6 25 0.054 108 55
7 70 0.036 281 100
8 70 0.033 125 55
NUCHAR WV-G. 12 x 40 Mesh. Composite
Acetic Acid Effluent Quality
Cycle No. Temp. Loading, Ib/lb ppm NaOAc ppm TOC
1 25 0.025 327 120
2 25 0.034 282 100
3 25 0.032 483 180
4 25 0.044 30 25
5 25 0.042 42 30
6 25 0.043 12 21
7 70 0.021 45 33
8 70 0.017 92 54
9 25 0.034 142 68
10 50 0.029 74 45
79
-------
ature on loading is also apparent. However, temperature
has no real significant effect on the quality of the ef-
fluent, the results at 50° and 70°C being nearly equiv-
alent to those at 25°C. This observation over a relatively
few laboratory cycles does not correlate with the results
obtained in the Demonstration Plant at high temperature
and a great many cycles.
In all of the cycles run in this laboratory column eval-
uation of composites, the regeneration procedure was the
same: 250-300 ml of 10% caustic, 100 ml of water, and
45 ml of 5N HC1 were added sequentially. This 0.44 liters
of total regeneration recipe represented one-third of an
empty bed volume. The quantity and strength of these
various reagents had been varied in several earlier 8'
column runs to optimize the concentration of sodium ace-
tate in the regenerant. In general, it was not affected.
Peak sodium acetate concentrations of ~7% were achieved.
The long, slow desorption experienced with phenol does
not occur with acetic acid regeneration. After the peak
concentration is reached, the acetic acid desorption
curve has a good downward slope to low levels. In the
laboratory columns, the volume of regenerant was 1.00 to
1.40 liters, or 7 to 10% of the volume of brine treated.
Physical Properties of Carbons
Mechanical Properties
The mechanical hardness or abrasion resistance of carbon
is not as important in our fixed bed in-place regenera-
tion application as it is in a process involving thermal
regeneration. There the saturated carbon must be mechan-
ically removed from the bed, transported to the furnace,
passed through the furnace, and returned to a packed bed.
In the course of obtaining isotherm data, a qualitative
measure of abrasion resistance can be made when the gran-
ular carbon is tested as received. Since equilibrium is
attained in a few hours with acetic acid and overnight
with phenol, we did not feel that it was necessary to
pulverize the granular carbon to obtain isotherm data.
In the course of obtaining isotherm data with granular
carbon, we noted that some carbons gave clear supernatant
liquid as soon as removed from the shaking apparatus,
while other contained suspended carbon particles after
standing for an hour. These suspended particles caused
a general background absorbance when analyses were done
spectrophotometrically, which changed little with wave
length. Presumably a test could be developed by stan-
dardizing the variables such as time and vigor of shaking,
80
-------
time of settling, path length and wave length. In gen-
eral, lignite base carbon gives the most fines, petroleum
coke base carbons the least, with bituminous base car-
bons intermediate. The manufacturers' representatives
are quick to indicate that fines produced from their car-
bons are a small fraction of the total carbon and merely
represent the rounding of sharp projections on some gran-
ules. However, Witco carbon, for example, has nearly
spherical granules as shipped and very few fines are gen-
erated by mechanical action in water suspension.
Composite samples of the Witco 718 and Nuchar WV-G used
in the Demonstration Plant were tested for particle size
distribution and comparative strength. The latter tests
were conducted in the following manner. The carbon (200 g)
was placed on a US No. 35 sieve. A US No. 40 sieve and
pan were placed below the No. 35 sieve. A rotap shaker,
built by W. S. Tyler Company, was used to impart mechan-
ical energy to the tester. After 5 minutes of agitation,
the tester was stopped and the amount of material on the
No. 40 sieve and on the pan was determined. Agitation
was again applied for an additional 55 minutes. After
weighing the sieve and pan, 408 g of stones were added
to promote breakage. The test was continued another 15
minutes. The results shown below indicate that there
are differences between products with the Witco 718 com-
posite being the better product both in particle size
distribution and strength.
Sieve Analyses, Composite Samples
US Sieve
Number
Weight Percent
on Sieve
Nuchar Witco
8 nil
10 0.03
12 0.69
14 9.98
16 23.84
18 27.39
20 19.14
30 16.05
50 2.63
100 0.03
•100 0.22
Breakage Tests
nil
nil
1.81
22.15
34.59
25.15
10.72
4.82
0.56
0.06
0.14
Product
Test Time
5 min
60 min
add 406 gm stones
15
Nuchar
-35 + 40
2.17
2.27
5.35
81
Weight
-40
0.56
1.22
8.97
% On
Witco
-35 + 40
0.34
0.35
1.22
-40
0.24
0.31
4.60
-------
Surface Properties
Composite samples were also tested for physical properties
normally measured on solid catalysts. The results are given
below:
Surface Area True Density Total Porosity
mVgm g/cc %
Witco 718 1176 2.42 66.9
Nuchar WV-G 1226 2.16 63.3
It is interesting that the carbon with slightly higher
surface area has less total porosity. The apparent den-
sity, as measured with mercury at 1.8 psia, was the same,
0.89 g/cc, for both carbons. Thus, the Witco particle
probably has denser walls.
Activated carbons are usually characterized by reference
to their capacity to adsorb a particular molecule under
specified conditions. Some reference molecules include
phenol, carbon tetrachloride, chloropicrin, methylene
blue and iodine. There is sentiment among workers in
the field of advanced waste treatment that iodine number
may correlate with desirable carbon utility in their
area. However, A.S.T.M. is still working on standardizing
the conditions for this one reference molecule.
Thus, it is obvious that it would be highly desirable if
some physical property could be measured for all commer-
cially available carbons which would enable one to at
least select a small group of carbons for evaluation each
time adsorption is indicated for treatment of a new in-
dustrial byproduct stream. Such a physical property
might be pore size distribution. One could scale data
on available carbons and select those possessing a large
percentage of their pore volume in the size range ap-
propriate to the molecule of interest.
Mercury Porosimetry is a standard way to obtain infor-
mation on pore size distribution. The 5OOO psi pene-
82
-------
trometer is believed to provide data to 350 A pore
diameter1. In the case of small molecules like acetic
acid and even phenol, very little adsorption would occur
in pores this large. These are probably the passageways
to smaller pores. Specifically, only 22% of the pore
volume of Witco 718 and 34% of the volume of Nuchar WV-G
is available to mercury at 5000 psi.
A mercury porosimeter usable to 60,000 psi was also avail-
able. This^is believed to correlate with pore diameters
down to 30 A1, which is still probably large for acetic
acid and phenol. The data showed a reasonably smooth
increase in penetration with pressure for both carbons,
with Nuchar having more penetration than Witco in the
range of 500 to 60,000 psi. Witco was beginning to gain
on Nuchar above 30,000 psi. Acid-washing the carbon in-
creased the penetration at all pressures in the above
range, with the effect being more pronounced for Nuchar.
Since Nuchar has more inorganic ash, this result with
acid-washing is plausible.
By determining the "complete" nitrogen desorption iso-
therm, information on pore diameters down to about 15
Angstroms can be obtained2. This is a much slower pro-
cedure than the standard B.E.T. surface area measurement,
requiring about a week per sample, providing no experi-
mental mishaps such as leakage occur during the long ex-
periment. From the desorption curve, a pore volume
distribution curve can be derived. This information
can also be presented in the form of a cumulative pore
volume curve.
Four samples were submitted; Witco 718 and Nuchar WV-G,
both as received and acid-washed. The pore volume dis-
tribution curves are given in Figures 33 through 36,
showing the increase in pore volume with increasing pore
diameter. Acid-washing eliminates some pores in the 50
to 100 Angstrom radius range with both carbons. This
may be responsible for the drop in capacity for acetic
acid or phenol observed when both carbons are acid-washed.
With both acid-washed carbons, the bulk of the pore
volume is at very small radius. It is not understood
what causes the negative pore volume indicated with
Witco carbon at about 10 A radius. In conclusion, no
83
-------
00
o
I— ooT
o:
o
CVJ
00
o
a-
CM —
CM
to
CO .
f\J
•4
#
o -I
FIGURE 33
PORE VOLUME DISTRIBUTION CURVE
NUCHAR WV-G, ACID WASHED
*7.SO~ ll7.SO ^7.50 ^27.50 ^67.SO 207* SO ^7.50 ^87-50
Pore radius, r (Angstroms)
-------
o
o.
(O —
o
o
o
o
FIGURE 34
PORE VOLUME DISTRIBUTION CURVE
NUCHAR WV-G
0=C
cn
oo
O
o
(M _
O
O
O
o
7.50 Ii7.50 87.50 127.50 167.50 207.50 247.50 287.50
Pore Radius, r (Angstroms)
-------
to
o»
FIGURE 35
PORE VOLUME DISTRIBUTION CURVE
WITCO 718, ACID WASHED
00
in
»—<
O
a;
o
Q.
0=C
cn
u
u
CD
0>
CO
O
(O -
CO
«n
•
sr -
CD
CJ
CJ _
CO
o
O -
7.50 1*7.50 87.50 127.50 167. SO
PORE RADIUS, r (Angstroms)
207.50 247.50 287.50
-------
=!•
o> —
00
CO
or
I—
t/i
3
ce
o
Q.
u
u
L.
-------
satisfactory correlation between physical properties
and performance was obtained. The "complete'1 nitrogen
isotherm seems the most likely source of useful infor-
mation at this time. Similar data on a variety of car-
bons should be obtained and compared with data obtained
on reference compounds.
Acetic Acid Recovery
To complete the picture on byproduct recovery from the
waste brine, laboratory studies were conducted on the
recovery of a useful byproduct from the sodium acetate
regenerant. There is very little internal plant con-
sumption of sodium acetate or acetic acid esters that
might be synthesized from the regenerant. If an anhy-
drous acetic acid could be derived from the regenerant,
there would be an internal use.
For any recovery process, it would be advantageous to
build up the acetate concentration as high as possible
in the regenerant. Modifications were made in the
Demonstration Plant to save the last half of regenerant,
add fresh caustic, and reuse this as the first regen-
erant on the next column. This was accomplished in a
single tank where two would have been desirable. How-
ever, the volume of regenerant was reduced 3O4O%. The
acetate concentration in the strong regenerant was in-
creased to 40-60 g/1 as acetic acid by these changes.
The first approach considered to producing acetic acid
was to boil off as much of the water as possible, then
add as strong HC1 as possible, preferably anhydrous HC1.
The mixture of water and acid could be distilled and
dried by a well known azeotropic system3. Since large
amounts of NaCl are formed when the regenerant is acidi-
fied, and the salt precipitates out during the boil down
of the acid brine, special equipment would be required
to handle the salt separation. Since the azeotropic
drying system necessarily involves some kind of organic
solvent to act as a water entrainer, solvent extraction
is a reasonable alternate and was the second approach
considered.
One of the standard solvents used in azeotropic drying
of acetic acid is ethylene dichloride. It is not a good
extraction solvent for acetic acid from water. Butyl
acetate is a better extractant and a good water entrainer,
but since it can be present in only limited amounts in
the drying system, it can not be used in the quantity
required to do a good solvent extraction. The two best
extractants found in our earlier work were t-butyl alcohol
and tributyl phosphate. The only azeotropic data avail-
88
-------
able on a butanol-acetic acid system are for the normal
isomer which boils too close to acetic acid. Since pro-
panol is reported as non-azeotropic with acetic acid,
it is possible that the lower boiling butyl alcohols are
also non-azeotropic with acetic acid. However, any sol-
vent which boils lower than acetic acid has the disadvan-
tage that all of the solvent would have to be vaporized
leaving acetic acid as a bottom product.
A high boiling solvent, such as tributyl phosphate had
the advantage that acetic acid can be stripped as an
overhead product without expenditure of energy for vapor-
ization of the solvent. A simplified block diagram of
such a process is given below.
Str
Ace
Rege
ong
tate
nerai
HCX
1
it
\c ]
h
Vacuum „
Still ^
\
t
HCl
1
^ nH
Adjust
X.
7»
S20 + HOAc
Atm.
Still
* 1
Solvent
Extractor
•
s
~^
>To pH adjustment tank
in
Demonstration Plant
Solvent
Storage
}
V
Strong acetate regenerant is acidified to pH 3 with HCl,
then contacted with the solvent to remove ~90% of the
acetic acid. The raffinate phase, a strong brine con-
taining acetic acid, is pumped to the pH adjustment tank
where it is combined with the regular brine flow and put
through the acetic acid adsorber. The extract phase is
distilled to remove water, then acetic acid. The water
overhead, rich in acetic acid, is recycled to the strong
acetate tank. Tributyl phosphate (TBP) the solvent of
choice, is not stable in the presence of acetic acid
above about 200°C. Thus, a vacuum still would be required
to strip acetic acid from the solvent.
Since regeneration is a batch operation, the recovery
system could also be a batch process rather than con-
tinuous. In this case, one tank could serve as the
strong acetate wash storage tank and the pH adjustment
tank and batch extractor. Also, a single still could
serve both distillation functions indicated on the block
diagram. This scheme would reduce the capital require-
ments at the expense of operating costs.
89
-------
The brine effluent from the solvent extractor will be
saturated with TBP; in brine its solubility is only ~50
ppm, whereas, in water, it is ~6000 ppm. At 50 ppm,
1.3 Ib/day of solvent is lost. This is too little to
justify a stripper to remove the TBP. If fed back into
the Demonstration Plant at the pH adjustment tank, TBP
would adsorb on the top of the acetate beds. Here it
would be hydrolyzed at the extremes of pH it would ex-
perience. This proposed acetic acid recovery process
has two unique features compared to the practice in wood
distillation or pulp mill black liquor recovery. A sol-
vent is used which is insoluble enough in brine to elim-
inate the necessity for stripping the raffinate. The
other feature is that the solvent is high boiling so that
acetic acid can be distilled from the solvent rather than
the usual practice of distilling the solvent, always the
larger quantity, away from the residual acetic acid. The
reason for a volatile solvent in the two industries men-
tioned is doubtless the necessity to avoid solvent loss
in the tars they must contend with. Our regenerant con-
tains no tar.
Assuming optimum regeneration efficiency, the scale of
this acetic acid recovery system, at 100 gpm in the Demon-
stration Plant, could be as low as 4 gpm in the brine
circuit and 2 gpm or less in the solvent cycle.
At an acetic acid concentration in the brine of 1500 mg/1
and 6^/lb credit for acetic acid, there is $100 worth of
acetic acid per day to be recovered. After paying oper-
ating costs, there would not be enough value to support
the capital cost of a recovery unit. With the acetic
acid level in the incoming brine having recently decreased
to ~1OOO ppm, acetic acid recovery is even less favorable.
90
-------
SECTION IX
MATHEMATICAL SIMULATION OF ACTIVATED
CARBON ADSORPTION SYSTEMS
The design of adsorption equipment and the prediction of its
performance for a range of operating conditions have been
based on prior experience and large amounts of experimental
data. The unsteady-state and nonlinear nature of adsorption
systems made the evaluation of scale-up factors and perform-
ance prediction difficult. In this section, a mathematical
model is described that allows reliable prediction of the
performance of large scale systems on the basis of data from
relatively simple and inexpensive laboratory experiments and
from related experiments and correlations reported in the
literature. The ability to predict performance is a major
step in the progression of the design of adsorption equipment
and the specification of operating conditions from an art to
a science.
The mathematical model is merely a mathematical description
of the various inter-relationships that exist among the proc-
ess variables. In the development of the model approximations
are made to simplify the model and to speed up its numerical
solution. The approximations are judiciously selected such
that simplicity and computing speed are not attained at the
expense of model accuracy. The first approximation is the
familiar lumped parameter approximation4. That is, the ad-
sorption column is divided into axial sections, and the prop-
erties (temperature, concentration, etc.) of each volume
section are lumped into average values for the section. In
essence, the adsorption column is considered to be a series
of ideal stirred tank reactors. The lumped parameter approx-
imation transforms the mathematical model from a system of
partial differential equations to a system of ordinary dif-
ferential equations. An additional benefit of the lumped
parameter approximation is that the same model can be used
to predict the performance of both batch adsorbers and column
adsorbers. A batch adsorber is simply a column with one axial
section.
The mass balance for the jth axial section reduces to:
f!i + !jb d_ w - /_• (c - c ) (i)
dt e dt J vje J x J
j = 1,..., NAS
91
-------
where j = axial section number, number consecutively in
direction of flow
C. = concentration of adsorbate in solution in jth
^ axial section (mass/volume)
t = t ime
p, = bulk density of adsorbent (mass/volume)
e = bed void fraction (volume/volume)
, w. = weight of adsorbate per weight of adsorbent in
J jth axial section (mass/mass)
F = flow rate (volume/time)
V. = volume of jth axial section (volume)
J
NAS = number of axial sections
The next step is to specify how the loading (weight of ad-
sorbate per weight of adsorbent) changes with time, dWj/dt.
This is accomplished by specifying the controlling mechanism
and then expressing that mechanism in terms of the process
variables. Three resistances can be identified for the ad-
sorption of adsorbate onto an adsorbent - transport across
an external "stagnant" film, diffusion within the adsorbent,
and surface reactions5.
It is assumed that the surface reactions are relatively rapid,
and that the rate of adsorption is controlled by resistance
to mass transfer.
Transport of adsorbate from the bulk liquid across the "stag-
nant" film to the external surface of the adsorbent is related
to process variables by means of a mass transfer coefficient6.
(C. - C.*) (2)
. .
dt p^ 3 3
where k = liquid phase mass transfer coefficient
(moles/area/time/mole fract.)
a = interfacial area per unit volume (area/ volume)
p.. = solution molal density (moles/ volume)
C.* = concentration of adsorbate in solution adjacent
^ to external surface of adsorbent (equilibrium is
assumed at the solution-adsorbent interface)
(mass/ volume)
92
-------
Transport of adsorbate within the adsorbent occurs by both
diffusion in the fluid occupying the intra-particle voids and
diffusion along the walls of the internal voids. A homoge-
neous solid diffusion mechanism is assumed to account for
the transport of adsorbate within the adsorbent7.
In this model, the solid is taken to be homogeneous with no
distinction made between the solid itself and the intra-
particle voids. Application of the mass balance to the homo-
geneous solid diffusion model results in the familiar second-
order parabolic partial differential equation for diffusion
in homogeneous medium8.
Again, the lumped parameter approximation is used to reduce
the partial differential equation to a system of ordinary
differential equations. The adsorbent particles are assumed
to be spherical and of uniform size. The lumped parameter
approximation d.ivides the adsorbent particles into equal
volume spherical shells and assumes uniform conditions within
each shell. The mass balances for adsorbate on the adsorbent
reduce to
(3)
J
r
1,
1,
where r
NAS
NRS
spherical shell number, number consecutively
from center of particle out
weight of adsorbate per weight of adsorbent
in rth radial shell in jth axial section
(mass/mass)
= o
q.,NRS+l
J
a
D
adsorbate loading in equilibrium with
solution at solution-adsorbent interface
D A
s r
DR DV
r r
Ds Ar-l
DRr-l DVr
homogeneous diffusivity (area/time)
area of outer surface of rth spherical
shell (area)
93
-------
DR
DV
NRS
= average of the thickness of the rth and
r+lth spherical shells (length)
= volume of rth spherical shell (volume)
= number of spherical shells
The total adsorbent loading is related to the spherical shell
loadings as follows
NRS
NRS
r=l
(4)
j = 1 , . . . , NAS
An additional relationship is required to complete the spec-
ification of the rate of change of loading and that is the
equilibrium relationship that exists at the solution- adsorbent
interface
j = 1,..., NAS
where E is the adsorption equilibrium relationship
To complete the model, it is necessary to specify initial
conditions and boundary conditions. The appropriate initial
conditions are
C^t-OJ-CJ
C.*(t-0)-C»
J J
w_.(t=0)=w..0
j = 1 , . . . , NAS
r = 1 , . . . , NRS
(6)
and the boundary conditions are
Co(t) - Cf(t)
where Cf is the concentration of adsorbate in the feed.
(7)
The mathematical model thus consists of equations 1 to 7. The
numerical solution of this model is accomplished by starting
94
-------
with the initial conditions and predicting conditions within
the column and in the effluent by marching forward in time.
By linearizing the equilibrium relationship an extremely ef-
ficient computing algorithm can be developed.
The following quantities are required as input data for the
model:
1) dimensions of carbon bed
2) physical properties of carbon - particle size
bulk density
bed void fraction
3) physical properties of solution - density
4) equilibrium between solute in solution and on the
carbon
5) transport parameters - mass transfer coefficient
solute diffusivity in carbon
The mass transfer coefficient comes from a literature
correlation while the diffusivity is determined from
batch kinetic experiments
6) initial conditions in carbon bed
7) feed conditions as a function of time
Three applications of the model to the phenol/Witco 718 system
are illustrated in Figures 37, 38, and 39 respectively: cor-
relation of batch kinetic data, prediction of a regeneration
curve, and prediction of a breakthrough curve.
Batch kinetic experiments (both adsorption and desorption)
were run to determine parameter values for diffusion within
the carbon particles. The data from one kinetic experiment
are illustrated in Figure 37 and compared with the predictions
of the model. In this experiment, 0.5 g of phenol-free Witco
718 carbon was initially shaken with 10 ml of water to drive
air out of the internal voids. Then, 90 ml of 1000 mg phenol/1
solution was added. The points on Figure 37 are the experi-
mentally measured phenol levels at 25°C as a function of the
time elapsed after the addition of the phenol solution to the
carbon. The curve on Figure 37 is the change in phenol level
that is predicted by the model. The correlation is excellent,
indicating that the program is a good model of the actual
physical adsorption process.
On Figure 38, the measured phenol levels in the Witco 718
phenol adsorber effluent during the fourth regeneration are
indicated by the points. The curve on Figure 38 is the ef-
fluent phenol level that is predicted by the model. Again,
the agreement between the actual data and the model is good.
The peak phenol concentration is predicted and the shape of
the curve is close to actual.
95
-------
in
O-,
Carbon:
Feed ;
Temp. :
D
0.5 gm Witco 718
100 ml - 900 mg phenol/1
25°C
measured
predicted
u 5 6
TIME CHOURSD
FIGURE 37
BATCH ADSORPTION OF PHENOL ON WITCO 718 CARBON
PREDICTED BY MATHEMATICAL MODEL COMPARED
TO EXPERIMENTAL DATA
96
-------
10 -
Carbon: Witco 718
Flow : 30 gpm of
Temp. : 30°C
A measured
predicted
Caustic
12 18 2U
TIME CHOURSJ
FIGURE 38
PHENOL DESORPTION FROM WITCO 718 CARBON
PREDICTED BY MATHEMATICAL MODEL COMPARED TO
ACTUAL DATA FROM PHENOL ADSORBER REGENERATION NO. 4
36
97
-------
On Figure 39, a predicted breakthrough curve for the Witco
718 phenol adsorber is illustrated. For this curve, the
feed conditions are 100 gpm, 250 ppm phenol, and 25°C, No
actual breakthrough curve under these conditions is avail-
able for comparison.
On the basis of data from only small laboratory experiments
and literature correlations, the model does a very good job
of predicting the performance of large columns. The model
is a valuable tool for use in the design of adsorption sys-
tems and in the evaluation of operating conditions.
98
-------
o
!/>_
C\J
V>
rvj_
(M
o
o.
(M
Carbon:
Flow :
Witco 718
100 gal/min
250 ppm phenol
Temp. : 25 °C
o
in-
o
o.
Q_
o
o.
tn
r-
in
(VI
16 20 30 «0 50
TIME CHOUR5) X10"1
70
80
FIGURE 39
BREAKTHROUGH CURVE PREDICTED BY MATHEMATICAL
MODEL FOR DEMONSTRATION PLANT PHENOL ADSORBER
99
-------
SECTION X
ACKNOWLEDGEMENTS
This project was performed by the Environmental Research Lab-
oratory at the Midland, Michigan Division of the Dow Chemical,
U.S.A. Dow personnel participating in the project were Dr.
R. A. Gaska, Project Manager, Mr. R. D. Fox, Project Super-
visor, Mr. C. J. Pinamont, Superintendent of the Demonstration
Plant, Dr. R. T. Keller, Senior Research Chemist in charge of
laboratory studies, Mr. C. K. Bon, in charge of the chlorine
test cell evaluation, and Dr. G. G. Hoyer, who developed the
mathematical model.
Design engineering for the Demonstration Plant was performed
by Dow Process and Plant Engineering, with Mr. W. L. Bennett
serving as Project Engineer, and Mr. Lyle Martz as Process
Engineer. The site and foundation contractor was Collinson
Construction Corp of Midland, Michigan. The general and
mechanical contractor was Kaighin, Hughes, and Paulin of
Toledo, Ohio.
We wish to acknowledge the helpful cooperation, assistance,
and guidance provided by the Project Officer, Mr. Clifford
Risley, Jr., of the Office of Research and Monitoring, Region
V, U.S. Environmental Protection Agency, Chicago, Illinois.
101
-------
SECTION XI
REFERENCES
1. Frevel, L. K., and Kressley, L. J. , Anal. Chem., 35,
1492 (1963).
2. Barret, E. P., Joyner, L. G., and Halenda, P. P.,
Journal of the American Chemical Society, 73, 373-380.
3. Kirk-Othmer, Encyclopedia of Chemical Technology, 2nd
Edition, Interscience Publisher, New York, 1968, Vol. 8,
391.
4. Smith, C. L. , Pike, R. W. , and Murrill, P. W. , Formulation.
and Optimization of Mathematical Models, International
Textbook, 1970.
5. Vermeulen, T., Advances in Chemical Engineering, Vole II,
Separation by Adsorption Methods, Academic Press, 1958.
6. McCune, L. K., and Wilhelm, R. H., Ind. Eng. Chem., 41,
1124 (1949).
7. Magtoto, E. R., PhD thesis, Univ. of Md., 1966, and Miller,
C. 0. and Clump, C. W. A.I.ChE Journal, 16, 1970.
8. Bird, R. B., Stewart, W. E., and Lightfoot, E. N.,
Transport Phenomena, John Wiley & Sons, Inc., New York,
1960.
103
-------
SECTION XII
PUBLICATIONS
1. "Purification of a Waste Brine by Carbon Adsorption with
Emphasis on Wastewater Reuse", R. D. Fox, R. T. Keller,
C. J. Pinamont, and J. L. Severson, presented at the 25th
Purdue Industrial Waste Conference, May 6, 1970.
2. "Brine Purification and Byproduct Recovery by Activated
Carbon Adsorption", R. D. Fox, R. T. Keller, C. J. Pinamont,
presented at the 64th Annual Meeting, American In-
stitute of Chemical Engineers, November 30, 1971.
105
-------
SECTION XIII
GLOSSARY OF ABBREVIATIONS
M - one thousand
HOAc - acetic acid
NaOAc - sodium acetate
DMK - acetone
00H - phenol
Bz - benzene
£ - less than or equal to
mu - millimicrons
|al - microliter
TOC - Total Organic Carbon
1O7
-------
SECTION XIV
APPENDIX
Analytical Procedures
Analysis for Acetic Acid
At concentrations greater than about 5OO ppm, titration
with a base was satisfactory since no significant amount
of other weak acids were present in our sample. The
Demonstration Plant operators performed the titrations.
Below about 500 ppm, the samples were sent to the Ana-
lytical Laboratory for nuclear magnetic resonance analy-
sis using a computer average transient for greater sen-
sitivity and tetrolic acid as an internal standard.
A Sargent- Welch recording titrator, Model DG was used in
the pH mode at 1O pH units span and automatic rate. A
10 ml burret was used which gives 5 inches of chart per
ml. Sodium hydroxide N/10 was the titrant. The pH was
adjusted to about 2 with hydrochloric acid solution in
a dropping bottle, IN for neutral or slightly acid sam-
ples and 5N for regenerant samples. The sample size
varied from 20 ml to 0. 5 ml depending on the amount of
acetic acid present with the object of having 1 to 3
feet of chart paper between inflections. The titrator
was adjusted to shut itself off at pH 10 so that no
operator attention was required after the titration was
started. The inflections between strong and weak acid
and between weak acid and strong base were determined
visually, and the inches of chart paper between inflections
was counted on the chart paper. The concentration of
acetic acid was then calculated as follows:
TT/-.A /-, inches between inflections 6
g HOAc/1 = - ml of sample - x 5
Analysis for Phenol
These analyses were done by the operators using a Beckman
DBGT spectrophotometer. The samples were made strongly
basic and the absorbance was measured at 307, 287, and
267 mu in a 1 mm cell. The net absorbance was calculated
as follows:
Anet = A287 ~ A307 ~ 1/2 (A267 ~ A3O7)
The mg phenol/1 was then read from a plot of concentration
vs net absorbance covering a range of zero to 300 mg phe-
109
-------
nol per liter. If the reading was off-scale, dilution
was made to bring the net absorbance below 0.6. This
method was unreliable below abo\it 5 mg phenol/1. Such
samples were sent to another laboratory where they were
extracted with chloroform and the extract was then anal-
yzed by ultraviolet spectrometry in up to 100 mm cells.
Analysis for Acetone and Benzene
These analyses were done by gas chromatography on 6' x
1/8" of Porapak® Q column support at 175° and 30 psi of
helium carrier gas pressure. A three nl sample was in-
jected and peak heights were compared with a standard
of 10 ul/1 benzene and 50 |dl/l acetone in water. Under
these conditions, the acetone peak appeared in 3 and
the benzene in 6 minutes. Phenol also gave a peak in
about 30 minutes but the recorder was usually on standby
at this time, and the information was not used. The
operators ran this analysis using an Aerograph GOOD
c hroma tograph.
110 »U.S. GOVERNMENT PRINTING OFFICE: 1973 514-156/343 l-»
-------
• 1 Accession dumber
w
r
Organization
2
•Environmental
Subject Field & Croup
050, 08A,
Research 1
08C
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
.aboratory
Dow Chemical U.S.A.
An Operating Unit of the Dow Chemical Company. Midland, Michigan 48640
Title
RECONDITION AND REUSE OF ORGANICALLY CONTAMINATED WASTE SODIUM CHLORIDE BRINES
To
Authors)
Fox, Robert D.
Keller, Richard T.
Pinamont, Carl J.
16
Project Designation
EPA Grant No. 12020-EAS
Note
22
Citation
Environmental Protection Agency report
number, EPA-R2-73-200, May 1973.
oo I Descriptors (Started Fitst)
J *Brine Disposal,
*Adsorption Activated Carbon,
*Industrial Wastes, *phenols, operating costs,
*Recovery, Reuse of waste brines
*Chlorine production
*Mathematical model
Identifiers (Starred Firs')
*Acetic acid
*Phenols
*0rqanics
Abstract
A plant of 100 gal/min capacity was constructed and operated for one year to dem-
onstrate the feasibility to remove and recover phenol and acetic acid from an 18%
sodium chloride brine by adsorption on fixed beds of activated carbon. The puri-
fied brine was used for production of chlorine and caustic soda. Separate electro-
lytical test-cell evaluation of the purified brine showed it to be equivalent to
pure brine. Regeneration of the carbon was accomplished by desorption with dilute
sodium hydroxide. The phenol desorbed was recycled.to the phenol manufacturing
plant while the acetate regenerant was processed to underground disposal wells.
More than 23 million gallons of brine were purified. Fourteen cycles of phenol
adsorption and regeneration and 105 cycles of acetic acid adsorption and regenera-
tion were completed with no significant deterioration of carbon performance. Phenol
removal to <1 ppn was accomplished at 50-140 gel/min and 15-70°C with an effective
carbon capacity of 0.167 lb/lb. Optimum regeneration was with 4% NaOH at 55-70'C.
Removal of 90% of the acetic acid from brine requires <80 gal/min flow rate and
,<40'C temperature, the resultant loading is 0.04 - 0.06 lb/lb of carbon. The pro-
jected net cost of purifying this waste brine for reuse was $1.32 per 1000 gallons.
Icfor
Carl 0. Pinamont
Institution
The Dow Chemical Company - Midland, Michigan
(RllOJ (REV. JULY !««»>
SEND, WITH COPY OP DOCUMENT. T Ol WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C. 20240
* OPOI 1870-369-930
-------
|