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

-------