Environmental Protection Technology Series
ACTIVATED CARBON TREATMENT
OF KRAFT BLEACHING EFFLUENTS
industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-119
June 1977
ACTIVATED CARBON TREATMENT OF
KRAFT BLEACHING EFFLUENTS
by
E. W. Lang
J. W. Stephens
R. L. Miller
St. Regis Paper Company
Cantonment, Florida 32533
Grant No. R-803270
Project Officer
John S. Ruppersberger
Food and Wood Products Branch
Industrial Environmental Research Laboratory-Cincinnati
Corvallis, Oregon 97330
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory - Cincinnati, U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory-
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
This report presents the findings of a pilot scale research project on
the removal of color and organic contaminants from a kraft pulp bleaching
plant effluent. The results will interest both industry and regulatory agen-
cies in considering alternatives for effluent polishing and color removal.
Cost estimates are presented. For further information contact the Food and
Wood Products Branch, Industrial Environmental Research Laboratory-Cincinnati.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
ill
-------
ABSTRACT
The removal of color and organic contaminants by adsorption on activat-
ed carbon from the effluent of a kraft pulp bleaching plant was investigated
in a pilot plant. The caustic bleach effluent, which contains 80% of the
color from pulp bleaching, was decolorized successfully when it was adjusted
to pH 2.5. The.spent carbon was regenerated with caustic solution for an
average of 11 adsorption-regeneration cycles before thermal regeneration was
required. Variables studied included pH of feed, feed rate, effluent from
bleaching of hardwood and softwood, caustic requirements for regenerating
the carbon, and concentration of color in feed. Capital and operating cost
estimates for a full-scale plant are presented. The cost effects of varia-
tions in design and operating conditions are also discussed.
Conclusions are that the process is technically sound, that it will re-
move 94% of the color and 84% of the total organic carbon from caustic
bleach effluent from the bleaching of softwood, but that it has slightly
higher capital and operating costs than alternative methods for reducing
color in bleach effluents (resin adsorption, ultrafiltration, or bleach
sequence modifications, for example).
This report was submitted in fulfillment of Grant R-803270 by St. Regis
Paper Company under the partial sponsorship of the Environmental Protection
Agency. This report covers the period September, 1974 to October, 1975 and
work was completed as of November 30, 1975.
-------
CONTENTS
Foreword
Abstract iv
Figures vli
Tables viii
Acknowledgments ix
I Introduction 1
II Conclusions 4
III Recommendations 6
IV Pilot Plant Facilities and Procedures 7
Pilot plant 7
Normal Operating Procedures 10
Adsorption Loading Runs 11
Regeneration Sequence 12
V Results 13
General 13
Pilot Plant Adsorption Runs 16
Pilot Plant Regeneration Runs 22
Equilibrium Isotherms 22
VI Discussion 26
Adsorption (Loading) Runs 26
Regeneration Runs 35
St. Regis Activated Carbon for Treating
Caustic Bleach Effluent. 38
Reversion of Color of Treated Caustic
Bleach Effluent 40
Reversion of Color of Acid Bleach Effluent 40
Use of Strong Caustic Bleach Effluent 41
Effect of Increased Temperature for Caustic
Regeneration 42
-------
CONTENTS (Con't.)
Treatment of Caustic Bleach Effluent from
Bleaching of Hardwood Pulp 43
Preliminary Design for Full-Scale Installation... 44
Preliminary Estimates of Costs for Full-Scale
Installation 46
References 52
Appendix 53
V1
-------
FIGURES
Number Page
1 Flow diagram of carbon adsorption pilot plant 8
2 Pilot plant adsorption column and regeneration tanks 9
3 Results from Series A loading runs 17
4 Results from Series B loading runs 19
5 Results from Series C loading runs 21
6 Concentrations in recycled strong eluate during Series B.. 23
7 Average color isotherm using ICI Powdered Darco S-51 in
Pensacola mill caustic bleach effluent at pH 2.5 24
8 Average TOC isotherm using ICI Powdered Darco S-51 in
Pensacola mill caustic bleach effluent at pH 2.5 25
9 Percentage removal of color for four runs of Series B 29
10 Accumulated percent removal of color during a prolonged
run.
30
11 Color concentration within column vs elapsed length of
loading run 32
12 Color concentration in column vs bed depth after various
lengths of operating time 33
-------
TABLES
Number Page
1 Summary of Operating Conditions and Results of Loading
Runs 14
2 Summary of Operating Conditions and Results of Regenera-
tion Runs 15
3 Conditions and Results from Series B Loading Runs 20
3
4 Capital Costs for Plant Treating 4730 m /day (1.25 mgd) of
Caustic Bleach Effluent by Carbon Adsorption 47
3
5 Operating Costs for Plant Treating 4730 m /day (1.25 mgd) of
Caustic Bleach Effluent by Carbon Adsorption 48
6 Effect of Changes in Operating Conditions and Equipment on
Estimated Capitol and Operating Costs for Plant Treating
Caustic Bleach Effluent from Bleaching 227 Metric Pulp Tons
per Day 49
viii
-------
ACKNOWLEDGMENTS
Most of the operation of the pilot plant and associated laboratory work
was performed by Mr. D. W. Barnes. ICI United States, Inc., provided
estimates of costs for carbon adsorption and thermal regeneration. The
Project Officer, Mr. John Ruppersberger, of the U.S. Environmental Protection
Agency, provided valuable suggestions to the program and to this report.
ix
-------
SECTION I
INTRODUCTION
A process for removing color from caustic extraction stage bleach
effluent by adsorption with activated carbon was investigated in a 25-gpm
pilot plant at the St. Regis pulp and paper mill at Pensacola, Florida. The
objectives were to determine the technical feasibility of the process for
treating the caustic bleach effluent (CBE), to obtain design data from a
reliable scale of operation over a prolonged operating period, and to deter-
mine the economic feasibility of the process on the basis of a preliminary
design for a full-scale installation.
The effluent from the bleach plant of a kraft mill that bleaches all of
the pulp produced contains 70% to 85% of the total color loading in the mill's
effluent. Typically, about 25% of the volume of bleach plant effluent is
from the caustic extraction stage but contains about 80% of the color load-
ing from the bleach plant and about 60% of the color from a totally bleached
kraft pulp mill. The removal of color from this relatively small stream
from the caustic extraction stage could therefore effect a substantial re-
duction of color in the total mill effluent.
Many investigations of methods for the removal of color from the CBE and
the total bleach effluent have been carried out in recent years. The treat-
ment processes that do not include basic changes in pulping or bleaching in-
clude coagulation using lime or alum; membrane separation processes, such as
ultrafiltration and reverse osmosis; adsorption processes, such as activated
carbon and polymeric resins that may or may not have ion-exchange properties.
The adsorption processes have shown considerable promise because they permit
recovery of most of the organic materials for disposal in the kraft mill
black liquor chemical recovery system. These processes all use caustic for
regenerating the adsorption resin or carbon and acidification of the feed
water or adsorbing medium. Both the caustic and acid (sulfuric acid or acid
bleach effluent) are available within a pulp mill.
Sanks (1) investigated the use of activated carbon and ion-exchange
resins for treating bleach effluents and found that carbon was not very
effective for removal of color from CBE at its natural pH (pH 10) but that
1. Sanks, R. K., "Ion-Exchange Color and Mineral Removal from Kraft Bleach
Wastes". EPA Report R2-73-255 (May 1973).
-------
some commercially available resins showed promise. Fuchs (2) and
McGlasson (3) showed that activated carbon was effective in removing color
from CBE at pH values of 2 to 4 but that the required large dosage of carbon
would make the process very costly.
The Rohm and Haas Company has evaluated the removal of color from bleach
effluents with their XAD-8 granular polymeric adsorption resin that has no
ion-exchange properties (4). In this process, CBE is combined with all or
part of the chlorination stage effluent to provide a feed having a pH of
about 2.5. Using a feed of total bleach effluent at a color of 2785 CU, the
process removed 84% of the color to give a product water volume of 23 bed
volumes having a color of 437 CU. The resin is regenerated with pulp mill
white liquor (about 10% NaOH) or weak wash (about 1% NaOH) and these solutions
are returned to the mill's chemical recovery system. A similar process is
the Uddeholm-Kamyr process (5), which uses a granular ion-exchange resin to
adsorb the color from CBE at its normal pH of about 10 but with the resin bed
previously rinsed with sulfuric acid to put the resin in its acid form.
About 90% of the color of the CBE was removed from a feed containing 14,000
CU. The product water (15 bed volumes) had a color of about 1400 CU and a
pH of about 6. The spent resin is regenerated with about 8% NaOH and then
rinsed with 3.4% H?SO, to restore the ion-exchange capacity of the resin.
A third resin adsorption process is under development by the Dow Chemical
Company (6) at pulp mills. A mixture of CBE and chlorination stage effluent
at pH 5.7 and a color of 10,420 CU was fed at a rate of 7.8 bed volumes per
hour to give 10 bed volumes of product water at 90% removal of color (1063
CU) . The spent resin is regenerated with 1 bed volume of weak wash (1% NaOH)
followed by about 10 to 15 bed volumes of chlorination stage effluent to re-
acidify the resin bed.
St. Regis Paper Company, with the partial support of EPA, evaluated
several treatment sequences involving carbon adsorption for treating un-
Fuchs, R. K., "Decolorization of Pulp Mill Bleaching Effluents Using
Activated Carbon", National Council for Stream Improvement Technical
Bulletin No. 181 (May 1965).
McGlasson, W. G., Thibodeaux. L. J., and Berger, H. F., "Potential Uses
of Activated Carbon for Wastewater Renovation" Tappi 49 (12) 521 (1966).
Rock, S. L., Bruner, A., Kennedy, D. C., "Decolorization of Kraft Mill
Effluents with Polymeric Adsorbents" Tappi 57 (9) 87 (1974).
Anderson, L. B., et al "A New Color Removal Process: A Field Report"
Tappi 57 (4) 102 (1974).
Chamberlin, T. A., ^t al "Color Removal from Bleached Kraft Effluents"
In: Proceedings 1975 TAPPI Environmental Conference, TAPPI, Atlanta,
1975. pp 35-45.
-------
bleached kraft pulp mill effluent to make it suitable for reuse in the mill
(7,8). Largely because of the experience from that pilot plant program and
the availability of the pilot plant carbon adsorption columns, a laboratory
study was carried out by St. Regis to determine whether it would be feasible
to treat bleach plant effluents with activated carbon. The unpublished re-
sults of this investigation showed that the color and a large portion of the
TOG of total bleach effluent could be removed when the effluent was passed
at high flow rates through a column filled with granular Darco 20x40 carbon.
It was shown that CBE alone could be decolorized equally well if it was
first adjusted to a pH of about 2.5. The laboratory study also showed that
the carbon could be regenerated by treatment with dilute caustic. However,
the carbon lost a large part of its capacity for removal of color after 10
to 20 cycles of adsorption and caustic regeneration. Therefore, these
studies indicated that thermal regeneration would be required after the 10
to 20 cycles using caustic regeneration. The laboratory studies showed that
CBE from the bleaching of hardwood could be treated even more effectively
than that from bleaching of pinewood.
Preliminary cost estimates for carbon adsorption indicated that it would
be more economical to treat acidified CBE because of its low volume than to
treat the total bleach effluent. The cost estimates indicated that the
process possibly would be lower in cost than other processes suggested for
this application, such as ultrafiltration, lime treatment, and resin adsorp-
tion processes.
With these encouraging results from the laboratory study, plans were
made for the pilot plant study reported here to be carried out in the plant
previously used for treatment of unbleached kraft effluent. A pipeline was
run from the bleach plant to the effluent treatment pilot plant, a distance
of 1220 m (4000 ft), to provide CBE for this program. It was decided that
only CBE would be treated, that the pH would be adjusted to the range of
2 to 4 by the addition of sulfuric acid, and that regeneration would be by
use of 1 to 2 bed volumes of 1% to 4% NaOH followed by a rinse and acidifica-
tion of the carbon bed with 1 to 2 bed volumes of 1% to 2% H2SO,. The major
variables to be investigated included: the feed rate of CBE to the carbon
column, the volume of water that could be treated before the product
cumulative color reached about 200 CU, the number of cycles of adsorption-
regeneration that could be obtained before the volume decreased to an un-
acceptably low level, the effect of minor changes in the regeneration sequence,
the effectof using CBE from hardwood bleaching, and the effect of treating a
CBE having a higher color concentration than that form the Pensacola bleach
plant.
7. Lang, E. W., Timpe, W. G. and Miller, R. L. "Activated Carbon Treatment
of Unbleached Kraft Effluent for Reuse" EPA-660/2-75-004, April 1975.
8. Timpe, W. G. and Lang, E. W., "Activated Carbon Treatment of Kraft Mill
Effluent for Reuse" Water-1973. AIChE Symposium Series 70 (136)579(1974).
-------
SECTION II
CONCLUSIONS
This study has demonstrated that activated carbon can be used to remove
94% of the color from CBE to a color of 200 CU. The process also removed a
large portion (84%) of the TOG to a concentration of 34 mg/1. When the
treated CBE is mixed with the acid bleach effluent (400 CU) at the Pensacola
mill, the total bleach effluent color is reduced by 72% to 350 CU.
To obtain high rates and percentage removals of color, the CBE must be
adjusted to pH 2.5. Such adjustment can be done by adding sulfuric acid or
acid bleach effluent at about the same overall operating cost.
In full-scale operation, the feed rate to an adsorption column should be
0.081-0.163 m3/m2 min (2 to 4 gpm/ft2) which results in a volumetric flow of
1 to 4 bed volumes per hour when bed heights of 3 to 6m (10-20 ft) are used.
The temperature of the pilot plant feed averaged only 23°C, due to cool-
ing in the storage basin. The normal temperature of CBE (about 55°C) can be
expected to increase the rate and degree of removal of color by 30 to 100%
on the basis of equilibrium adsorption tests in this study and dynamic adsorp-
tion tests reported in the literature.
The inclusion of hardwood CBE reduces the feed color and the operating
costs. The use of a low-volume high-color CBE did not significantly improve
the adsorption performance and would not result in reduction of operating
costs.
Regeneration of the spent or loaded carbon with recycled 2% caustic
solution followed by acidification of the carbon bed with 2% sulfuric acid
gave satisfactory adsorption performance for about 11 cycles. Thermal re-
generation of the carbon is then needed to restore its adsorption capacity.
The caustic regeneration solution can be reused (with additions of
caustic) to increase the concentrations of sodium and organics. The re-
cycled caustic solution is suitable for addition to the pulp mill weak black
liquor system for partial (30% to 60%) recovery of the sodium used in re-
generation. The amount of chloride in the caustic solution to be sent to
black liquor is low and would not cause an appreciable increase in the
chloride content of the mill's chemical recovery system.
3
A plant for treating 4730 m /day (1.25 mgd)of CBE from the Pensacola
mill would cost approximately $2,380,000 and the operating cost would be
about $8.16/metric pulp ton ($7.41/short ton). The operating costs for the
-------
conservatively designed plant contain all costs, including amortization.
With cost reductions from the inclusion of 34% hardwood CBE, recovery of
31% of the sodium in the regenerations, elimination of prefiltration through
the use of up-flow columns, use of acid bleach effluent for acidification of
the column after regeneration, and operating at 60°C, the capital cost was
estimated to be $1,590,000 and the total operating costs were $5.31/metric
pulp ton ($4.83/short ton).
The indicated costs are generally greater than published costs for resin
adsorption processes, ultrafiltration, and modifications of the bleach
sequence that achieve about the same reductions of color from bleach eff-
luents .
-------
SECTION III
RECOMMENDATIONS
The carbon adsorption process should be considered carefully for treat-
ing CBE if is is not technically or economically feasible to achieve suffi-
cient color reductions through changes in the bleaching sequence.
The process should not be considered for treating the total bleach
effluent because of the high costs.
If other color reduction alternatives prove to be unattractive, the
carbon process should be developed further to reduce the major cost items.
In particular, further investigation is needed on the use of high temperature
(60°C) adsorption, high temperature (90°C) caustic regeneration, up-flow
mode of adsorption, thermal regeneration and reuse of the carbon, and re-
duction of losses of sodium.
-------
SECTION IV
PILOT PLANT FACILITIES AND OPERATING PROCEDURES
PILOT PLANT
The pilot plant used for this study was that used in a prior EPA-St.
Regis project on the treatment of effluent from an unbleached kraft mill by
carbon adsorption to make the water suitable for reuse in the mill (7,8).
A flow diagram of the pilot plant used for this study is given in
Figure 1, and a picture of the adsorption column and regeneration solution
tanks is shown in Figure 2.
The equipment used in the prior EPA-St. Regis project was modified to
permit operation at the conditions selected from laboratory studies and from
preliminary plant designs and cost estimates. The diameter of the adsorption
bed, 0.915 m (3 ft), was fixed by that of the existing adsorption columns.
The feed rate was selected to give a fairly high lineal velocity of 0.082
to 0.144 m-Vm^min (2.0 to 3.5 gpm/ft ) which would be used in plant columns.
A bed height of 3.05 m (10 ft) was selected for all runs because it provided
a satisfactory height:diameter ratio and reasonably high values of BV/hr.
The volume of the carbon bed was 2.01 m3 (530 gal) and the dry carbon bed
density was 0.37 kg/1 (23 Ib/ft3).
Caustic regeneration was used throughout this study because a previous
study by St. Regis had indicated that it would provide much lower operating
costs than thermal regeneration. The volumes of the caustic regeneration
solution (strong eluate), slop cuts (weak eluate), and acid rinse were
selected at 1 BV on the basis of the prior laboratory work. Three of the
four adsorption columns in the pilot plant were used for these regeneration
solution tanks. Their capacities were 3.030 m3 (800 gal) each. An additional
tank of 2.270 m3 (600 gal) was installed for storage of strong eluate during
the regeneration sequence.
A system was installed for continuous adjustment of the pH of the CBE
feed to the adsorption column. A cost estimate for a full scale plant in-
dicated that the overall costs would be lower if purchased sulfuric acid were
used for pH adjustment rather than the required large volume of acid bleach
effluent, which has a pH of about 1.8. If acid bleach effluent were used for
7. Lang, E. W., Timpe, W. G. and Miller, R. L., "Activated Carbon Treatment
of Unbleached Kraft Effluent for Reuse" EPA-660/2-75-004, April 1975.
8. Timpe, W. G. and Lang, E. W., "Acitvated Carbon Treatment of Kraft Mill
Effluent for Reuse" Water-1973. AIChE Symposium Series 70 (136) 579 (1974) .
-------
00
EFFLUENT FROM
BLEACH PLANT
STORAGE BASIN
FOR FEED
PH
ADJUST
TANK
ACID
TANK
V Vr
PRODUCT
WATER TO
SEWER
REGENERATION SOLUTION
TANKS
ADSORPTION
COLUMN
Figure 1. Flow diagram of carbon adsorption pilot plant.
-------
Figure 2. Pilot plant adsorption column and regeneration tanks.
9
-------
adjusting the pH of the feed, the volume of the feed would have to be
tripled, which would increase substantially the size of the adsorbers and
the capital costs of a full scale plant. Therefore, a sulfuric acid feed
tank holding 0.378 m3 (100 gal) was installed, and air pressure was used to
transfer the acid through a pH control valve to the 0.890 m3 (235 gal) pH
adjust tank.
3
An existing 708 m (187,000 gal) stprage basin was used to hold CBE feed
for the pilot plant. This basin was needed so that the CBE could be solely
from the bleaching of softwood and to ensure a non-fluctuating quality of
CBE.
The basin was refilled with softwood CBE at intervals of 1 to 2 weeks.
Even though the change of BOD during storage was not determined, the changes
in conductivity, TOC, pH, and color indicated that the only significant change
in the quality of the CBE in storage was a dilution by rainwater. The BOD
of CBE as it was fed from the adsorption column averaged only 31 mg/1 during
four runs of Series B.
A 1220 m (4000 ft) run of 6.4 cm (2 1/2 in.) PVC pipe was installed for
pumping the CBE from the bleach plant cuastic extraction seal pit to the
pilot plant basin.
The CBE from the basin that was used as feed to the pilot adsorption
column was not filtered and there was no problem with pressure build-up in
the column.
Most of the suspended solids of the CBE were removed by sedimentation in
the feed basin. In a commercial unit, a multimedia filter probably would be
needed to remove suspended solids ahead of the adsorption column.
The pH, color, conductivity and temperature of the feed and product water
were monitored and recorded continuously, and composite samples were collected
automatically.
The adsorption column was provided with five sample ports located 0.6 m
(2 ft) apart. The column and the solution tanks were provided with sight
glasses extending the length of the tanks to indicate liquid levels during
regeneration. The carbon was supported by a graded sand and gravel bed
0.6 m (2 ft) deep. The column and the associated tanks used in regeneration
have quick-disconnect hoses for inlet and outlet flows, which greatly
faciliated the switching of flows among the tanks during the regeneration
sequence.
NORMAL OPERATING PROCEDURES
The feed basin was filled with CBE while the bleach plant was bleaching
softwood (pine) pulp, the basin was refilled with softwood CBE whenever the
level fell below about 60% full, or at intervals of 1 to 2 weeks.
10
-------
The pilot plant was operated solely with softwood CBE because: (1)
hardwood CBE has a much lower color concentration and is more readily de-
colorized than softwood CBE (the elimination of this variable made it easier
to evaluate the results); (2) hardwood constitutes only about 34% of the
bleached pulp production at the Pensacola mill. The relative effects of
treating hardwood and softwood CBE were determined in a small glassware
adsorption column, as discussed later in this report.
The St. Regis Pensacola mill bleaches about 250 tons of pulp per day
using the CEHDEH bleaching sequence. The CBE is made up of the water from
the two E (extraction) stages and the D (chlorine dioxide) stage and amounts
to about 4,730 m3/day (1.25 mgd), or 20.8 nrYmetric ton (5000 gal/short ton)
of bleached pulp. The chlorination or acid bleach effluent is from the C
(chlorine) and H (hypochlorite) stages and amounts to about 14,190 m /day
(3.75 mgd), or 62.4 m3/metric ton (15,000 gal/short ton).
Adsorption or Loading Runs
The pilot plant normally was operated only during the day shift, 5 days
per week. In an adsorption run CBE was pumped from the feed basin to the agi-
tated pH adjustment tank where 93% sulfuric acid was added at a rate controlled
by a pH controller to reduce the pH from about 10 to the pH selected for the
adsorption run (pH 2.3 to 2.9). The pH adjusted CBE was then pumped through
an orifice-type flow indicator to the top of the adsorption column. The flow
rate during Series A was 1.6 BV/hr, or 53.7 1/min (14.2 gpm), and the flow
rate during Series B was 2.83 BV/hr, or 94.6 1/min (25 gpm). The water from
the bottom of the column was run to the sewer after a sample stream was fed
to a composite sampler and the monitoring sensors, where pH, conductivity,
and color were measured and recorded continuously. The pH, color, and temper-
ature of the feed were also recorded continuously. Analyses were made by
the procedures given in the Appendix.
In all runs, the adsorption column was charged with 727 kg (1600 Ib) of
ICI Granular Darco 20x40 mesh activated carbon, which has a particle size
range of 0.84 to 0.42 mm. This quantity of carbon gave a bed depth of 3.05 m
(10 ft) in the 4.6 m (15 ft) column, and the volume of the carbon bed was
2.01 m3 (530 gal). The carbon bed had a height-diameter ratio of 3.3, which
is large enough to ensure plug-flow of feed.
ICI Granular Darco carbon was used in these runs because prior studies
had indicated that it was superior to other commercial carbons for removal of
the relatively large color molecules that exist in kraft pulping effluents.
The particle size range of 0.84 x 0.42 mm (20 x 40 mesh) was used because it
gave more external area and higher adsorption rates than the often used
larger sizes of 1.68 x 0.42 mm (12 x 40 mesh) or 2.38 x 0.59 mm (8 x 30 mesh).
As will be noted later, the greater pressure drop of the 20 x 40 mesh carbon
was not large enough to be a concern.
11
-------
The loading run was continued until the color of the product water
reached a cumulative color of 200 CU which gave a color removal of over 94%.
Some runs were continued until much higher product colors were reached to
provide additional information.
Regeneration Sequence
When the selected cumulative color had been reached, the loading run was
stopped and the loaded (spent) carbon was caustic regenerated.
In the regeneration sequence, the water remaining in the column was
lowered to the top level of the carbon and maintained at this level during
the regeneration sequence to minimize the amount of intermixing of regenera-
tion solutions. One bed-volume (BV) or 2.01 m (530 gal) of strong eluate
containing 2% NaOH was passed downward through the carbon at a rate of 1
BV/hr. The strong eluate was transferred from its tank to the column by air
pressure applied to the strong eluate tank and the flow was manually con-
trolled to give 1 BV/hr on the inlet flowmeter.
When the water from the column reached a conductivity of 5000 micromhos
and a pH of about 4, as indicated by the values that were being continuously
monitored and recorded, the water from the column was diverted from the pro-
duct water stream to the empty weak eluate tank. After 1 BV of strong eluate
was fed to the column, 1 BV of weak eluate from the previous regeneration was
passed through the column at 1 BV/hr. When the conductivity of the outlet
water from the column reached 10,000 to 15,000 micromhos and the pH reached
10, the outlet flow was diverted to the strong eluate tank until 1 BV had
been collected. The outlet flow was then diverted back to the weak eluate
tank until the two slop cuts amounted to 1 BV. One BV of rinse water was
then fed to the column at 1 BV/hr following the weak eluate. In later runs,
after the rinse water had been added, the column was agitated with air ad-
mitted to the bottom of the bed for a few minutes and then backwashed for
30 min with 3 BV of previously treated water. The rinse and backwash waters
in a large installation would be returned to the feed surge tanks for re-
processing.
An acid rinse, consisting normally of 1 BV of 2% H2SO,, was then passed
through the bed followed by 0.5 BV of fresh water. The next loading run was
then started by feeding pH-adjusted CBE to the column. The acid rinse water
coming from the column was sufficiently low in color that it could be in-
cluded in the product water.
The recovered 1 BV of weak eluate and 1 BV of strong eluate were used in
the subsequent regeneration to recover their chemical values. The recovered
strong eluate was normally quite low in free NaOH (about 0.45%) but quite
high in Na content. The free NaOH concentration was again increased to 2%
by the addition of 50% caustic. In a commercial plant the strong eluate
would be reused until its total solids content had been increased to about
10% and then some of it would be sent to the pulp mill's weak black liquor
for recovery of the organic (fuel) and sodium values. Weak eluate would be
added to the strong eluate tank to bring it back to 1 BV and the free NaOH
content would be increased to 2%.
12
-------
SECTION V
RESULTS
GENERAL
Three series of loading-regeneration cycles (runs) were made in the
pilot plant to investigate the effects of feed rate, effects of minor changes
in regeneration procedure, and primarily to determine how much water could be
treated per run and how many cycles could be made in a series before the
carbon would need to be regenerated thermally. Fresh carbon was charged into
the adsorption column at the start of each series. Conditions that were
common to all of the loading adsorption runs included: a bed depth of 3.05 m
(10 ft), 2.01 m (530 gal) of carbon (1 BV), 727 kg (1600 Ib) of ICI Granular
Darco 20 x 40 mesh activated carbon, and CBE from bleaching of pine pulp.
The results of the loading (or adsorption) runs are given in Table I
as averages for each series. The average results do not include aborted
Run 4, Series A and Run 3, Series C for the reasons discussed below under
Pilot Plant Adsorption Runs. The average results from the caustic regenera-
tions for the three series of runs are given in Table 2. Results and evalu-
ation of each series of loading and regeneration runs are given below. A
general discussion of the results of pilot plant studies, supporting labora-
tory studies, and plant cost estimates are given in SECTION VI.
Series A was made at a feed rate of 1.61 BV/hr or 53.7 1/min (14.2 gpm)
which was selected to give a fairly high areal flow rate of 0.082 m /m min
(2.0 gpm/ft ) such as would be used in full-scale units. The conditions used
in regeneration were not changed during this series. Series A was discon-
tinued when the volume of water treated per cycle decreased below 20 BV.
Series B was made at the same conditions as Series A except that a
higher feed rate of 2.83 BV/hr, or 5680 1/min (25 gpm), was used to show the
effect of feed rate on the number of runs before the quantity of water
treated again decreased to 20 BV/run. The results of Series B indicated that
the higher flow rate reduced the number of BV/run which might give higher
operating costs than the use of 1.61 BV/hr. Therefore, Series C was made to
recheck the results of Series A at 1.61 BV hr under improved conditions of
caustic regeneration discussed below. Series C was terminated after four
loading runs because it had provided sufficient information and because it
was necessary to apply the remaining available effort to small-scale studies
of other variables.
13
-------
Table 1.
SUMMARY OF OPERATING CONDITIONS AND RESULTS OF LOADING RUNS
Item
ftin
(gpm/ft2)
(kgal)
Conditions:
Jf'eed rate, 1
Feed rate m:
Length of run, total,hr
Feed Composition:
Color, CU
TOC, mg/1
Temperature, °C
Results: ,
Total product vol., vf'
per series
per run
BV/run, total
BV/run, to 200 cum. CU
Product water composition
color, at end of run, CU
color, avg. for run, CU
TOC, avg. for run, mg/1
pH, avg.
Na, avg.
total Cl, avg.
Loadings: .
color, total run, CU/g
color, to 200 cum. CU, CU/g
TOC, total run, mg/g
TOG, to 200 cum. CU, mg/g
Removals:
color, to 200 cum. CU, %
color, total run, %
TOC, total run, %
100# H^O^ used, g/1
Sl BV = 2.006 m5 (530 gal.)
CU/g or~rr,g/g x 0.36 = kg/nr
Series A
12/6/74-2/21/75
1-3, 5-8 = 7
1.61
0.082(2.0)
32.2
3^90
270
2.9
18
723(191)
103(27.3)
52
42
—
314
45
4.o
—
-
443
381
32
26
94
91
83
0.62
Series B
2/25-5A3
1-10 = 10
2.83
0.144(3-5)
19.0
2856
210
2.4
22
1080(285)
108(28.5)
54
38
810
318
33
2.6
355
362
360
280
26
18
93
89
84
0.55
Series C
6AO-7/30
1,2,4 = 3
1.61
0.082(2.0)
79
3133
220
2.3
28
764(202)
254(67.3)
127
107
944
214
23
2.5
364
258
1087
866
70
58
94
93
89
0.68
Series C
Extrapolated
1-8 * 8
1.61
0.082(2.0)
39
3133
220
2.3
28
1000(264)
125(33)
62
62
w
200
23
2.4
-
-
496
496
34
34
94
94
89
0.64
-------
Table 2.
SUMMARY OF CONDITIONS AND RESULTS OF RB3ENSRATION RUNS
Item
Caustic make-up per regeneration:
100& NaOH, kg/ra^db/ft3)
100# NaOH, gA of CBS treated b
Strong eluate:
NaOH concentration to column, %
NaOH concentration to column, %
No. of regenerations
Final strong eluate
TBS, mg/1
TOG, mgA
Color, CU
conductivityi cicrozhos/cm
Cl, ionized, mgA
Cl, total, mgA
Weak aluate:
final weak eluate,
TDS, mgA
TOG, mgA
Cl , total
H SO, (100#) for rinse:
concentration, %
added per regen., kg/m carbon(lb/ft )
added, gA of CBE treated b
Scries A
19.3(1.20)
0.375
2.3
0.41
5
41,200
8,600
92,000
32,000
-
-
23,400
4,700
-
1.53
15.4(0.96)
0.32
Scries B
15.1(0.94)
0.362
2.0
0.48
10
59,400
18,100
222,000
30,000
3,056
4,465
24,000
7,500
4,400
2.0
20.0(1.25)
0.48
Series C
29.5(1-84)
0.206
4.2
1.8
3
61,800
9,100
57,000
89,000
-
1900
29,000
5,800
1360
2.7
27(1.7)
0.19
Volume of strong eluate, weak eluate, and acid rinse was 1 BV (2.01 nr) in all regenerations.
CBE treated to 200 cum. CU in one more run than number of regenerations.
-------
PILOT PLANT ADSORPTION RUNS
In a typical adsorption run, the color of the product water remained at
less than 100 CU for 75 to 95% of the run length and then increased rapidly
as the color compounds broke through. Since a product color of 200 CU was
considered acceptable quality for reuse or discharge, the performance of the
adsorption runs was compared on the basis of bed volumes of water treated
when the cumulative color reached a level of 200 CU. This cumulative color
is the color of all product water if it were accumulated to that run length.
As seen in Table 1 the average run of Series A was continued until the cumu-
lative, or average product color, reached 314 CU, Series B was continued to
a cumulative color of 318 CU, and Series C to a cumulative color of 214 CU.
Therefore the percentage removal for a total run was lower than the percentage
removal if the run had been stopped at 200 cumulative CU.
The results from Series A (Figure 3) showed that the volume of CBE that
can be treated per run before breakthrough of color varied considerably be-
tween runs. Run 4 gave poor removals, presumably because of inadequate con-
trol of feed pH, and was aborted after 8.9 BV, or 5.6 hr of operation. The
volume of CBE treated to a cumulative color of 200 CU decreased from 78 BV
for the first run to about 20 BV after 8 runs and averaged 42 BV/run, ex-
cluding Run 4.
The percentage removal of color for the total run lengths during
Series A remained in the range of 81 to 97% and averaged 91%. When the cumu-
lative product color reached 200 CU, the color removal was 94%. The removal
of TOG was unexpectedly high - ranging from 70 to 92% and averaging 83%. The
average concentration of TOC in the product water was 45 mg/1, which is about
half the TOC concentrations obtained by resin adsorption processes when treat-
ing pulp bleaching effluents (6 and unpublished reports) . The value of
45 mg/1 of TOC is also about half that of secondary effluents from pulp
mills.
The accumulated amount of color or loading of color on the carbon is cal-
culated by:
f&ed conc'n — av product conc'n
T j-j ~ _ _.. -.!...._ ™~/~ — av &e concn — av p
loading on carbon, mg/g - rt%'£T,V dosage, log A
It is assumed that a color unit is equivalent to 1 mg/1 which permits loading
of color to be expressed as CU/g of carbon. It is desirable to obtain as
high color loading as practical to minimize the frequency of caustic re-
generation. The loading of color on the carbon for the total run was 600
CU/g on Run 2, Series A, decreased to 240 CU/g by Run 8, and averaged 381
CU/g during Series A. The exhaustion, or dosage, rate of carbon is found
from the loading on the carbon and the change in concentration by the above
expression. In series B the average exhaustion or dosage rate for removal
of color to 200 cumulative CU per cycle was (2856-200) /360 = 7.4 g/1.
6. Chamberlin, T, A. , et al "Color Removal from Bleached Kraft Effluents"
In: Proceedings 1975 TAPPI Environmental Conference, TAPPI, Atlanta,
1975. PP 35-45.
16
-------
§
g
03
94
80
60
40
20
0
80
60
40
20
0
600
400
200
0
4000
§ 3000
00
n
u
u
z
PS
3
o
CJ
oi
o
2000
1000
EXTRAPOLATED
.TOTAL RUN
TO 200 CUM. CU
COLOR
TOC
PRODUCT
4 5
LOADING RUN NO.
Figure 3. Results from Series A loading runs.
All plots are for total run except in
top graph as indicated.
17
-------
The pH of the feed during the loading runs of Series A was maintained
between 2.6 and 3.6 and averaged 2.9. Passage of the water through the bed
caused the pH to increase to an average of 4.0, possibly due to ion-exchange
on the carbon as discussed in the following section. The pH of the feed was
decreased as the series progressed to improve, the adsorption performance. The
data of Figure 3 indicate that the low pH value on Run 5 was largely re-
sponsible for the improved adsorption of this run. The higher pH of Run 6
evidently caused the indicated increased color of product water and lower
loading of color and lower value of BV/run.
The adsorption performance of Series B runs (see Figure 4), made at a
higher feed rate of 2.83 BV/hr or 0.144 m3/m2 min (3.5 gpm/ft2), was similar
to that of Series A. The regeneration procedure was altered slightly as
Series B progressed, but apparently these changes did not have a pronounced
influence on the loading runs. These changes in regeneration procedure are
discussed below. Detailed data from Series B are presented in Table 3.
Loading Run 4 gave poor results for undetermined reasons. The length of
Run 6 (60 BV) was longer than that of previous runs, but this was due mainly
to the lower color of the feed water caused by a heavy rainfall on the storage
basin. After Run 8, the bed volumes of water treated to 200 cumulative CU
decreased steadily to 20 BV/run for Run 11 and to 6 BV/run for Run 13. Run
11 was continued until the product color reached 1650 CU for a total of 115 BV
and a cumulative color of 780 CU. The overall (total) removals were 72%
of the color and 65% of the TOC, and the color loading on the carbon was 634
CU/g. This extra large loading on the carbon did not affect adversely the
color removal of the following run any more than a normal loading on the
carbon (Figure 4).
As shown in Figure 4, the pH of the feed was decreased during Series B
from 2.7 to about 2.3 in an effort to obtain improved performance. The pH
of the product water was greater than that of the feed for the first seven
runs. The effect was noted also in Series A runs and is discussed later.
Series C was started with a fresh charge of carbon after the volume of
water treated per run in Series B had declined to low values (less than 10
BV/run). A feed rate of 1.6 BV/hr was used to recheck the results of
Series A when using the modified regeneration procedure of added holding
time of strong eluate and the use of backwashing of the bed. Results from
Series C runs are shown graphically in Figure 5. Loading Run 1 of Series C
did exceptionally well. A total of 235 BV was fed and the cumulative product
color was only 65 CU and TOC was only 14 mg/1.
In the next three loading runs, the BV/run to 200 cumulative CU de-
creased to 62 in Run 2, to 15 in Run 3, and increased to 45 in Run 4. The
only apparent reason for the long length of Run 1 was that the pH of the
carbon bed was reduced to a lower than usual level with the acid rinse
prior to starting the run, as evidenced by the fact that the product water
pH was 1.7 during the first 2 hr of operation. The pH increased slowly to
2.7 and did not exceed 3.2 for the remainder of the run. The number of bed
18
-------
I
s
en
120
100
80
60
" 40
20
0
80
I 60
* 40
fe
20
0
oc 600
i 400
200
3000
a
o
g 2000
1000
0
COLOR
TOC'
PRODUCT
PRODUCT
56789
LOADING RUN NO.
10 11 12
13
Figure 4. Results from Series B loading runs.
All plots for Total Run except as indicated.
19
-------
NJ
O
fable 3. CONDITIONS ASD RESOLTS FKOM SmTKS B LOADING RUNS
Teed rate=2.83 BV/hr, 0.144 m3/m2 mln (3.54 gps/ft2)
Run number
Item
Length of ran, total hr
To 200 CUB. CU, hr
?ecd composition :
Color, CU
TCC, rj/l
T-H to coluan
Ttspsriture, C
Sesiilts :
Length of rua, hs
Total product, n
Total produ:t, ksal«
SV/rua, total
BV/run, to 2DQ cua. CO
Product water cocpositioc:
Color at end of run, CU
Color, cue., CO
TOC, CUD., B^l
pH, cux.
Ka, c-i=., Bg/1
Total Ci, cum., sgA
Color, total run, CO/g
Color, to 200 cua. CO, CO/g
TOC, total run, Bg/g
ICC, to 200 cua. color, mg/g
Hiccvals, total:
Color, %
TOO, *
1
19.5
19.4
3737
275
2.7
18
19.5
110.9
29.3
55-3
55
1075
160
15
2.7
4oo
—
546
537
40
39
96
95
2
15.0
9.9
3664
266
2.5
14
15.0
85.2
22.5
42.5
28
1350
460
34
3.1
412
— -
376
268
27
18
87
87
3
16.3
12.4
3775
233
2.5
18
16.3
92.7
24.5
46.2
35
1230
400
50
3.4
..
—
430
345
24
20
89
79
4
17.7
1.4
2745
181
2.4
17
17.7
100.7
26.6
50.1
4
1020
496
36
2.7
350
246
311
28
20
2
82
80
5
19.6
13.4
2646
158
2.5
20
19.6
111.3
29.4
55-4
38
555
281
16
2.7
334
2?0
362
25?
22
14
89
90
6
21.2
23.7
1848
146
2.5
20
21.2
120
31.7
59-9.
67°
98
118
24
2.6
279
385
286
305
20
23
94
84
7'
22.7
14.8
2532
212
2.3
26
22.7
128.7
34.0
64.2
42
705
292
41
2.4
344
297
270
30
21
88
81
8
23.1
15-5
2409
194
2.3
25
23.1
131
34.6
65.3
4t
750
323
44
2.2
346
376
267
27
20
87
77
9
20.8
12.5
2350
170
2.4
24
20.8
118.1
31.2
58.8
35
653
309
31
2.4
316
275
331
210
23
14
87
82
10
14.3
10.0
2850
258
2.3
30
14.3
81
21.4
40.3
2o
653
339
42
2.3
3S7
334
279
207
24
19
88
84
11
40.8
7.0
2770
246
2.2
25
40.8
231.3
61.1
115.3
19.8
1650
778
86
2.2
382
—
634
140
50.9
10.7
72
65
12
14.3
3.2
3193
2C9
2.2
29
14.3
81.4
21.5
4C.6
9.1
1020
615
59
2.3
350
283
75
16.3
3.8
81
72
13
8
2.1
3COO
216
2.4
30
8
45.4
12.0
22.6
5-9
664
234
30
2.3
~
173
46
11.8
3.0
93
86
100* HgSO^ used, g/1 0.66 0.52 0.53 0.56
mextca?olaced, Run 6 was stopped before 200 CO was obtained.
0.4?
0.46
0.73
0.52
1.00
0.73
0.59
-------
a
H
200
150
100
50
0
80
§ 60
o
I 40
*t
20
0
BO 1600
8
o 1200
3 800
g
400
3000 -
g 2000
8
1000 -
3
-a-
PRODUCT
TOTAL RUN
TO 200 CUM. CU
COLOR
FEED
•a
PRODUCT
FEED
345
LOADING RUN NO.
Figure 5.
Results from Series C loading runs. All
plots are for total run except as indicated.
21
-------
volumes per run for Run 3 was low, apparently because the bed was not acidi-
fied sufficiently after regeneration (as evidenced by the fact that the pH
of the product was approximately 6). Since Run 3 length was abnormally low
in comparison with Runs 1,2, and 4, the data of Run 3 were not included in
the average results of Series C.
As mentioned previously, Series C was terminated after four loading runs
because it had provided sufficient information and because it was necessary
to apply the remaining available effort to small-scale studies of other
variables. To permit the results from Series C to be compared to those of
the other two series, a smoothed curve was drawn through the plot of BV/run
vs number of runs and extrapolated parallel to the corresponding plots from
Series A and B. This extrapolated curve predicted that the number of BV/run
would reach 20 after 8 runs. The data in Table 1 for "Series C extrapolated"
are based on 8 runs and are believed to be conservative estimates of what
Series C would have given if continued for 8 runs.
PILOT PLANT REGENERATION RUNS
The regeneration procedure was altered slightly throughout the pilot
plant operation in an attempt to determine the best procedure to provide
maximum lengths of loading runs. The general procedure was described in the
section on "Pilot Plant Facilities and Operating Procedures". During Series
B, the concentration of the acid rinse was increased from 1% to 2% H«SO,
(Run 5) to ensure a lower pH of the bed before starting the loading run.
Starting with Run 2 of Series B, the strong eluate and the acid rinse were
held in the column for an additional 0.5 to 1 hr to provide more time for
caustic extraction of the color and for acidification of the carbon.
Starting with Run 4 of Series B, the carbon bed was backwashed for 30 min
with tap water after the weak eluate had been eluted from the column with
tap water* The results of the regenerations are summarized for each series
of cycles in Table 2 and discussed in the following section. The build up
of concentrations of various components in the strong eluate as it was re-
cycled during Series B is given in Figure 6.
EQUILIBRIUM ISOTHERMS
Equilibrium adsorption isotherms were prepared by the procedure given
in the Appendix using powdered Darco S-51 carbon and eight samples of CBE
obtained during the pilot plant program. Darco S--51 is stated by the manu-
facturer to have the same adsorption properties as the Granular Darco used
in the pilot plant runs. The average results from five isotherms for CBE
at pH 2.5 are represented by the curves given in Figure 7 for color ad-
sorption and in Figure 8 for TOC adsorption. Individual values of adsorp-
tion loadings varied as much as + 30% from the average values shown by
these isotherms. The color isotherm indicates that the Darco carbon at
equilibrium with CBE at 4000 CU at pH 2.5 will have an average color loading
of 1200 CU/g.
22
-------
oo
W
Q
PS
O
ae
o
ff
*>
4000
2000
20,000
10.00C .
TOTAL
SOLUBLE
TOC
Na
2468
NO. OF CAUSTIC REGENERATIONS
10
Figure 6. Concentrations in recycled strong eluate during Series B,
23
-------
5000
,M 2000
»
§
§
o
z
1000
500
200
100
50
100 200 500 1000 2000
COLOR CONCENTRATION IN REMAINING SOLUTION, CU
4000
Figure 7. Average color isotherm using ICI Powdered Darco S-51
in Pensacola mill caustic bleach effluent at pH 2.5.
24
-------
500
M
e
•>
o
CO
200
§
O
z
M
1
100
50
20
10 ' 20 50 100 200
TOC CONCENTRATION IN REMAINING SOLUTION, mg/1
500
Figure 8. Average TOC isotherm using ICI Powdered Dareo S-51
in Pensacola mill caustic bleach effluent at pH 2.5.
25
-------
SECTION VI
DISCUSSION
Preliminary cost estimates made early in this study indicated that
capital costs are primarily affected by the number and size of the adsorbers
needed for given daily throughput and by the frequencey of thermal regenera-
tion of the carbon. The major factors in the direct operating costs are
labor associated costs, make-up of carbon lost during thermal regeneration,
net caustic consumed in caustic regeneration, and sulfuric acid used in
operating at the low pH. The primary objective during the pilot plant runs
was to determine how to minimize these costs and to provide data needed in
designing a full-scale plant. To achieve these goals, it was necessary to
determine the economic optimum pH to use in adsorption. A lower pH would
give longer runs and reduce many of the costs, but acid needed for pH adjust-
ment below pH 2.5 would become very costly. It was desirable to run as long
as possible between caustic regenerations to minimize the time lost in re-
generation and to minimize the cost of caustic and acid used in regeneration.
It was particularly important to extend the number of loading-regeneration
cycles per series before thermal regeneration would be required with its
high attendant costs.
ADSORPTION (LOADING) RUNS
The amount of water treated per run (before caustic regeneration was
needed) was quite high when compared to results from three resin adsorption
processes for treating caustic bleach effluents (4,5,6). As shown in Table 1,
the number of BV/run to 200 cumulative CU ranged from 38 for Series B to 62
for Series C (extrapolated). Data reported for the three commercially
developed resin processes range from 14 to 40 BV of caustic bleach effluent
treated per cycle at flow rates of 4 to 8 BV/hr to give 90% removal of color.
The number of bed volumes treated per run in the carbon decolorization pro-
cess could be increased by decreased flow rate (longer retention time), by
decreased feed concentration, by reduced pH of the feed, by the conditions
of caustic regeneration and probably by increased temperature of the feed.
Rock, S. L., Brunner, A., Kennedy, D. C., "Decolorization of Kraft Mill
Effluents with Polymeric Adsorbents" Tappi 57 (9) 87 (1974).
Anderson, L.G., e± al_ "A New Color Removal Process: A Field Report"
Tappi 57_ (4) 102 (1974).
Chamberlin, T. A., et al "Color Removal from Bleached Kraft Effluents"
In: Proceedings 1975 TAPPI Environmental Conference, TAPPI, Atlanta,
1975. pp 35-45.
26
-------
The higher flow rate of Series B (2.83 BV/hr) reduced the contact time of
the CBE in the carbon bed to 21 min from the contact time of 38 min for
Series A and C. When the lengths of the runs are adjusted to the feed color
of Series A (3490 CU), the BV/run to 200 cumulative CU would be 42 for Series
A, 31 for Series B, and 55 for Series C (extrapolated). These results in-
dicate that the length of a run is indirectly proportional to the flow rate.
The optimum flow rate would be determined by the relative cost of providing
more adsorption volume vs the cost of the regenerations.
The number of BV/run for Series A probably was adversely affected by the
use of a slightly greater pH of the feed (2.9 vs 2.4 and 2.3 for Series B
and C). Series A also was affected adversely by less optimum conditions of
caustic regeneration. The first run of the C Series was unusually long—235
BV to 200 cumulative color—and this run certainly biased the adsorption per-
formance of Series C on the favorable side.
Overnight shutdowns (16 hr) and weekend shutdowns (64 hr) during loading
runs probably influenced the adsorption performance. In a typical loading
run, the color of product water from the column after an overnight shutdown
was lower than the day before when the feed was stopped. However, after a
few bed volumes, the product color was as great or greater than the final
color at the previous shutdown. A study of the data indicated that the length
of a loading run was perhaps 5-20% grater than it would have been with no
shutdowns.
The temperature during adsorption was expected to be important to the
rate and extent of color removal. It was not practical to alter the tempera-
ture of the adsorption runs in the pilot plant. CBE as it leaves the bleach
plant has a temperature of about 55°C. However, the CBE temperature had
decreased to almost ambient temperature when fed to the carbon column because
of the long transfer piping to the pilot plant and the prolonged time the
CBE was held in the storage basin. As indicated in Table 1, the average
feed temperature was 18°C for Series A (winter time), 22°C for Series B, and
28°C for Series C. The possible benefit of the higher temperatures of the
later runs was masked by other changes made in these runs. Therefore, no
conclusion on the effect of temperature can be drawn from the pilot plant data.
The effect of temperature on equilibrium loadings of color and TOG was
determined in isotherm tests made with total bleach effluent at pH 2.9 using
Darco S-51 powdered carbon. The loading of color increased 20% and the load-
ing of TOG increased 30% when the temperature was increased from 25°C to 60°C.
Chamberlain (6) found that the volume of bleach effluent treated with a Dow
adsorption resin to a given cumulative color increased as much as sevenfold
when the temperature was increased from 22°C to 60°C. Fuchs (2) found that
the equilibrium loading of color on carbon increased 40% when the temperature
of CBE was increased from 38°C to 93°C. It appears evident that a full-scale
2. Fuchs, R. E., "Decolorization of Pulp Mill Bleaching Effluents Using
Activated Carbon", National Council for Stream Improvement Technical
Bulletin No. 181 (May 1965).
6. Chamberlin, T. A. ejt al "Color Removal from Bleached Kraft Effluents"
In: Proceedings 1975 TAPPI Environmental Conference, TAPPI, Atlanta,
1975, pp 35-45.
27
-------
unit using carbon adsorption would obtain much better performance, when
treating CBE at 55°C, than was obtained in the pilot plant runs.
The removals of color and TOC were very good. The average removal of
color was 94% to a product color of 200 cumulative CU and 91% for the total
run to a cumulative color of 282 CU. The average removal of TOC was 85% for
total runs. On the basis of BOD analyses of cumulative feed and product
samples from four loading runs of Series B, carbon adsorption reduced the BOD
concentration from 31 mg/1 to 18 mg/1, for a reduction of 42%. The variation
of accumulated percentage removal of color with length of run (BV) is illus-
trated in Figure 9 which presents data from Runs 1,3,7 and 10 of Series B.
The percentage removal vs length of run for Run 11, Series B is shown in
Figure 10. In this run the accumulated percent removal was 85% at 55 BV
treated and 70% at 110 BV. As Series B progressed, the percentage removals
decreased somewhat (see Figure 4) . The percentage removals for Run 4 were
low mainly because the feed concentration was about 24% below that of the
first three runs.
The average feed for the B Series contained 284 mg/1 soluble chloride and
78 mg/1 of organic chloride, and the product water contained 263 mg/1 of
soluble chloride and 63 mg/1 organic chloride. The removals by the carbon
were 7% of the soluble and 10% of the organic chlorides in the feed water.
Of the average sodium content of 360 mg/1 in the feed water for Series B
and C, only 8% was removed by the carbon. On the basis of composite samples
of the feed and product from a single run (Run 4, Series C), there was
generally little change in the concentrations of ten metal ions normally pre-
sent in bleach effluents:
Concentration, mg/1
Na Ca Fe _K Si Al Mg Cu Mn Cr
Feed 338 11.0 1.4 1.1 1.2 0.5 0.6 0.8 0.1 <0.05
Product 343 10.7 1.3 1.4 0.9 2.8 0.7 0.2 0.2 0.24
The loading of color per weight of carbon is another measure of perfor-
mance. A high loading means less frequent regeneration. As see in .Table 1
and Figures 3,4 and 5, the average total loadings per run were 433 CU/g for
Series A, 360 CU/g for Series B, which was low because of lower feed con-
centration, and 469 for Series C (extrapolated). The average for the three
series was 433 CU/g for the total run lengths and 386 CU/g if all runs had
been stopped at 200 cumulative CU. At the end of Run 10, Series B, when
color had broken through the column, the top 10% of the column had a loading
of about 800 CU/g, which is 73% of that expected at equilibrium (Figure 7) .
The bottom 20% of the column had a loading of only 40 CU/g which was only 6%
of that expected at equilibrium. Run 1, Series C, which was exceptionally
long, had a final loading of 1950 CU/g, which is greater than that pre-
dicted by the equilibrium isotherm. In general, the loadings of color at the
end of the runs were about half of the equilibrium isotherm loading at the
mean concentration in the column.
The loadings of TOC for the total lengths of runs ranged from 26 mg/g
for Series B to 34 mg/g for Series C (extrapolated) . The loading expected at
28
-------
100
90
80
70
H &J
gS 60
o 5
0 50
!^
i§
40
30
20
10
RUN 3
FEED COLOR:
RUN 1 3737 CU
RUN 3 3775 CU
RUN 10 2350 CU
RUN 7 2532 CU
I
10
20 30 40
BED VOLUMES TREATED
50
60
RUN 7
70
Figure 9. Percentage removal of color for four runs of Series B.
29
-------
100
80
al
3
o
CJ
ft.
O
I 60
id
g 40
Q
20
FEED COLOR:
2770 CU
20
40 60 80
BED VOLUMES TREATED
100
120
Figure 10. Accumulated percent removal of color during a
prolonged run (Run 11, Series B).
30
-------
equilibrium from the average isotherm (Figure 8) would be 65 mg/g at the
mean concentration of 80 mg/1 for Series B. Therefore, the TOC loadings
in the pilot plant runs were also about half of those expected at equili-
brium.
The above results on loadings show that the top (feed) end of the column
approached or exceeded the loadings obtained in equilibrium isotherm tests.
However the discharge end of the column had very low loadings, partly be-
cause of the low concentrations of color but primarily because of the low
rates of adsorption on the granular carbon at the low concentrations.
Greater average loadings of a column could be obtained by using two or more
columns in series and regenerating the feed column when the color breaks
through the last column. Such a system gains greater loadings and lower
regeneration costs but at the expense of greater capital costs for the
additional columns and increased complexity of piping.
The variation of color concentration with the length of the carbon bed
was investigated in several of the runs in Series B and C. At various
elapsed times of operation, samples of the water being treated were ob-
tained through sample ports at 0.6 m (2 ft) intervals down the column.
The results from Run 10, Series B are given in Figure 11 as smoothed data
plots of color vs elapsed length of loading run for the five sections of
the carbon bed. A cross plot of the above data is given in Figure 12 as
concentration vs bed depth at various volumes of water treated. A plot
on semi-log graph paper normally gives a straight line that permits ex-
trapolation of the data to greater lengths of column to predict the con-
centrations in a much deeper bed.
Concentration profile data within the column during Run 1, Series C
showed that the color of the water in the top 40% of the column actually
increased, for unknown reasons, above that of the feed after 110 BV had
been treated (concurrent changes in pH are discussed below) . After 212 BV,
the color of the water 0.6 m (2 ft) from the top had increased from a feed
color of 3000 CU to a color of 4400 CU. In spite of these increases of
color within the column, the color was readily adsorbed in the lower part
of the column in this unusually long run.
The concentration profiles in the column were used to prepare plots of
service time to breakthrough of 200 cumulative color vs bed depth. Such
plots are useful in predicting the service time of full-scale adsorption
beds (9). A service time vs bed depth plot for Run 10, Series B, indicated
that the wave front of 200 CU moved through the first 0.6 m of the bed
immediately, through the next 1.2 m in 2 hr, and through the last 1.2 m
in 10 hr. In Run 1, Series C, there was a similar slowing of rate of
movement of the wave front as it progressed through the column. The wave
front broke through 0.6 m of bed immediately through the next 1.2 m in
30 hr, and through the last 1.2 m in about 50 hr. This slowing of the wave
front indicated that longer beds (or two columns operated in series) will
permit the carbon to be loaded to greater values of CU/g than was possible
9. Hutchins, R. A., "Design of Activated Carbon Systems" Chemical Engineer-
ing 80 (19) 133 (August 20, 1973).
31
-------
3000
2500 -
2000 -
3
o
1500 -
1000 -
500
6 8
ELAPSED TIME, HR
10
]2
Figure U. Color concentration within column vs elapsed
length of loading run (Run 10, Series B).
32
-------
3000
2000
1000
500
200
FEED COLOR
13 hr
10
12
BED DEPTH, FT.
Figure 12. Color concentration in column vs bed depth
after various lengths of operating time (Run 11,
Series B, total length of run =14.3 hr).
33
-------
in the single 3 m (10 ft) bed used in the pilot plant.
*y
Pressure drop through the pilot plant column was only 0.14 to 0.28 kg/m
(2 to 4 lb/in2) during Series A and C at a flow rate of 0.082 m3/m2min
(2.0 gpm/ft2). At the higher flow rate of 0.144 m3/m2min (3.5 gpm/ft2) dur-
ing Series B, the pressure drop was about 0.32 kg/m2 (4.5 lb/in2). Even
with much longer beds in a full-scale plant, the pressure drop with this
carbon will be a minor consideration if the water is well clarified or
sand filtered. The feed water to the pilot plant column was not filtered,
but the prolonged holding of the feed CBE in the storage basin probably
permitted most of the suspended solids to be removed. In a plant installa-
tion, the pressure drop could be expected to be greater because thermal
regeneration of the carbon normally reduces the average particle size.
Pressure drop would not be a concern if up-flow beds are used. Up-flow beds
were chosen for the full-scale design discussed later in this report.
The ion-exchange properties of the carbon after repeated caustic re-
generations were investigated since an ion-exchange mechanism possibly
contributes to the removal of color bodies from CBE. The results from four
tests indicated that the carbon after three regenerations had an ion-
exchange capacity for NaOH of 0.3 to 0.5 meq/g and for H2S04 of 0.3 meq/g.
Ion-exchange capacities reported in the literature for various activated
carbons ranged from 0.14 to 0.61 meq/g using NaOH and 0.23 to 0.4 meq/g
using acid. In the pilot plant regenerations, the amount of acid used to
acidify the bed to a pH of about 1.5 after it had been backwashed with tap
water was 1.0 and 1.2 meq/g for Series A and B, respectively. The amount
of NaOH used in regeneration (total added less that found in the strong
and weak eluates) was 0.80 and 0.87 meq/g for Series B and C, respectively.
It was concluded that the carbon has an ion-exchange capacity of approxi-
mately 0.3 meq/g.
The mechanisms responsible for color removal under the conditions used
in the pilot plant are not clear, but removal evidently is due to a com-
bination of physical adsorption, chemi-sorption on ion-exchange sites,
precipitation and coagulation due to release of protons from ion-exchange
sites, and concentration-precipitation of color bodies at the surface of
the carbon.
The previously discussed effects of pH on removal clearly show that at
low pH values the color bodies become less soluble, presumably because the
hydrogen form is less ionized than the sodium form which makes it more
insoluble. The pilot plant runs showed that the carbon bed must be at a
pH of less than 2.5 for efficient removal of color even though the feed
CBE is already reduced to pH 2.5. The decreased solubility at pH 2.5 does
not cause the color bodies to be coagulated to the extent that they can be
removed on an 0.8 micron Millipore filter that is used in the standard
color measurement, procedure (see Appendix) . Fuchs (2) found that CBE
2. Fuchs, R. E., "Decolorization of Pulp Mill Bleaching Effluents Using
Activated Carbon", National Council for Stream Improvement Technical
Bulletin No. 181 (May 1965).
34
-------
having a high concentration of color of 8000 CU would adsorb on carbon
much more readily at pH 3 than at pH 4 and that some of the color bodies
would precipitate from solution at pH 2.
It may be significant that pH varied considerably through the bed as the
loading runs progressed. In Run 1, Series C, after feeding 21 BV of CBE
at pH 2.3, the pH of the bed at 0.6 m from the top (feed end) increased to
4.8 and then decreased to 2.5 at 1.5 m and at the outlet of the column. As
the run progressed, this increase of pH remained in the top part of the
column but the value of pH gradually decreased to near that of the product
water. Similar pH increases within the column were noted during Runs 2
and 4, Series C. In these runs, a peak pH of 5.0 occurred at 0.6 m from
the feed end and another occurrred at 2 m from the feed end. These changes
of pH perhaps indicate that chemical or chemi-sorption reactions were taking
place in the bed. There was no measurable increase of temperature from top
to bottom of the bed which might be expected if chemical reactions were
taking place.
REGENERATION RUNS
The effect of the concentration and amount of caustic in the regeneration
sequence was investigated over a limited range. Normally, the spent carbon
from a loading run was regenerated by passing 1 BV of strong eluate contain-
ing 2% free NaOH through the bed over a period of 1 hr. In Regeneration 6,
Series A, the bed was regenerated with 2 BV of fresh strong eluate contain-
ing 1% NaOH to see if the doubled volume and contact time with 1% NaOH
would regenerate the bed more completely. There was no apparent effect
on the subsequent loading run. In Regeneration 11, Series B, the procedure
was changed from recycled strong eluate to 2% NaOH in tap water. As seen
in Figure 4, there was no apparent change in performance from the trend line
of the preceding runs.
The effect of using 1 BV of 4% NaOH was investigated in the regeneration
preceding Run 13, Series B. The results from loading Run 13 indicatd that
doubling the concentration of the strong eluate had no effect on the sub-
sequent adsorption run. The use of 4% free NaOH in the recycled strong
eluate was continued in Series C, but the possible influence of the higher
strength strong eluate was camouflaged by several other changes that
apparently had a greater effect on adsorption performance.
When 1 BV of 2% NaOH in the strong eluate is used, the addition of
caustic amounts to 0.055 g NaOH/g carbon. As discussed later, some of the
added sodium is recovered in the strong eluate when it is added to the
pulping black liquor in a full-scale installation.
The degree of acidification of the carbon following caustic regeneration
and rinsing is important to the adsorption performance, of the regenerated
carbon. Whenever less than 1 BV of 2% H2S04 was used, the pH of the
product water from the adsorption run increased to about 3.5 and the ad-
sorption performance decreased. In these cases an additional dose of half
the normal amount of rinse acid through the column generally would increase
the performance to a satisfactory level.
35
-------
The use of a fresh water rinse and backwash of the bed after the weak
eluate had come through caused a lower pressure drop through the bed in the
subsequent loading run, removed caustic from the carbon, and reduced the
amount of acid for acidification. In a full-scale unit, the backwash water
would be returned to the feed surge tank for re treatment. Backwashing
consisted of running 0.114 m3/min (30 gpm) of tap water for 30 min. In
later runs, 0.5 BV of tap water was passed through the bed between the acid
rinse and the start of CBE for the next loading run. The results indicate
no particular benefit from this rinse removal of excess acid remaining in
the carbon bed. The acid water that is displaced from the column at the
start of the next loading run is low in color and is of product water
quality. In a full scale plant, the collection of product water would start
immediately after the backwash when the acid rinse is started to the column.
On the basis of these pilot regenerations, the best procedure to use is
the following:
Bed Elapsed
volumes time, hr Source Disposition
1 1 strong eluate strong eluate tank
1 1 weak eluate weak eluate tank
1 0.5 backwash water backwash tank
2 0.25 backwash water backwash tank
1 12% H SO product water
In this procedure, after the weak eluate has been added, one bed volume
of water from the backwash storage tank is used to displace the weak eluate
over a 0,5 hr period and then the bed is backwashed at a high rate of about
0.3 m-Vrn min (7 gpm/ft^) for 15 min. The backwash water is returned to a
second backwash tank for settling and reused in backwashing the multiple
adsorption columns as well as the sand filters that will probably be re-
quired for the CBE before it is fed to the columns. This regeneration
procedure would require 3.75 hr.
Chlorination stage effluent could be used to acidify the bed instead of
using 2% H2S04. During some small-column tests described later, the acid
bleach effluent at pH 1.7 was used to acidify the bed. A total of 29 BV
was required to reduce the pH of the water coming from the bottom of the
bed to 1.9. The results from the subsequent adsorption run were poorer
than usual which indicated that the bed was not sufficiently acidified by
the acid bleach effluent. Owing to poor results and large amount needed,
it was concluded that the use of acid bleach effluent in place of 2% H2S04
should not be investigated further.
The buildup of Na, Cl, and organics in the recycled strong eluate was
of prime interest in this project. Since the strong eluate in a full-scale
plant would be sent to weak black liquor for disposal, it is necessary for
the strong eluate to be as concentrated as possible to minimize an addition-
al evaporation load. The amount of chloride in the strong eluate should be
low to avoid increased corrosion and problems in the recovery furnace. As
shown in Figure 6, the concentration of total solids in the strong eluate
during the regenerations of Series B increased to about 60,000 mg/1, or 6%,
36
-------
after about five runs and then leveled out. The total chloride content in-
creased to 4400 mg/1 of which 3000 mg/1 was soluble chloride and 1400 mg/1
was organic chloride (32% of the total). The concentrations of these
materials in the weak eluate were about 1/3 of those in the strong eluate.
The results of Series B regenerations indicate that, in a full-scale plant,
about 0.5 BV of strong eluate should be drawn off and sent to weak black
liquor after five regenerations and replaced with 0.5 BV of weak eluate.
The volume of CBE treated in five runs at 50 BV/run would be 250 BV. There-
fore the volume of strong eluate sent to black liquor is (0.5/250) (100) =
0.2% of the volume of CBE treated. The concentrations of materials in the
strong eluate would be about the same as those found in the strong eluate
in Series B after 10 runs.
The amounts of water, organics, and chloride that would be introduced
into black liquor in a full scale plant can then be estimated. At the
Pensacola St. Regis bleach plant, CBE amounts to 20.8m3/metric ton of
bleached pulp (5000 gal/short ton). The volume of strong eluate sent to
black liquor would be 0.002 x 20.8 or 0.042 m3/metric ton (10 gal./short
ton). The weight of contained materials per ton, on the basis of Series B
results, would be:
added to black liquor from
concentration, strong eluate,kg/metrie
material mg/1 ton(lb/short ton) of pulp
total solids 60,000 2.50 (5.0)
Na 15,000 0.62 (1.25)
Cl, total 4,400 0.18 (0.37)
TOC 18,000 0.75 (1.50)
water 94% 39 (78)
Typically, the amount of black liquor solids from kraft pulping is 1500
kg/metric ton (3000 Ib/short ton). Therefore, the total chloride in-
troduced in the pulping chemicals cycle by the strong eluate would be (0.18/
1500)x(100) or 0.012% of the black liquor solids if all of the pulp is
bleached.
In the case of the Pensacola mill, which bleaches 28% of the pulp
produced, the chloride introduced to the pulping chemicals cycle would be
0.0034% of the black liquor solids from the total mill. The chloride level
in the black liquor solids would build up as more chloride is added, and
some would be purged along with the losses of solids from the pulping
chemicals cycles. In a southern kraft mill, the white liquor (regenerated
pulping liquor) contains 3 to 4 g/1 of NaCl. If all of the pulp is bleached
and the strong eluate from the carbon decolorization is added to the black
liquor, the equilibrium NaCl concentration in the white liquor would in-
crease by about 1 g/1. This approximately 25% increase of NaCl concentra-
tion in white liquor is an acceptable increase. Since weight balances in-
dicated that 84% of the chloride removed from the water was unaccounted for
(not recovered in the strong and weak eluates), the chloride content of
strong eluate in a full-scale plant might run greater than that found in the
pilot plant.
37
-------
The water added by the strong eluate to the weak black liquor would be
39 kg/metric ton (78 Ib/short ton) of which 37.5 kg (75 Ib must be removed
in the evaporators to increase the solids concentration to 63%, which is
the concentration of the black liquor burned in the recovery boiler. The
amount of steam required to evaporate this water in a large recovery furnace
is 8.3 kg/metric ton (16.7 Ib/short pulp ton) and the steam recovered from
burning the organic matter of the strong eluate would be 7.5 kg/metric ton
(15 Ib/short ton). There the steam used in evaporation is essentially
balanced by the steam generated by burning the organic matter.
Weight balances for each series were prepared to determine where the
chloride, sodium, color, and TOG went since very little was recovered in
the strong eluate. On the basis of average results from the three series
of runs, only 11% of the total chloride in the feed CBE (343 mg/1) was
adsorbed on the carbon. Of the amount adsorbed by the carbon, only 16%
was recovered in the strong and weak eluates and 84% was unaccounted for.
Of the total sodium in the CBE feed (360 mg/1) , only 10% was adsorbed on the
carbon. Of the sodium removed by the carbon and added to the strong eluate,
31% was recovered in the strong and weak eluates, 26% was recovered in the
acid rinse, and 43% was unaccounted for.
Of the color removed by the carbon (91% of that of the feed), 14%
appeared in the strong and weak eluates, and 86% was unaccounted for. Of
the TOG removed from the feed by the carbon (86%), 29% appeared in the strong
and weak eluates, and 71% was unaccounted for.
Of the unaccounted-for materials that were apparently retained by the
carbon, some might have been removed from the column during the backwash
(the backwash water was not analyzed) . However, it appears that most of
the unaccounted for materials remained on the carbon. If so, the average
loading of TOG at the end of a series would be 192 mg/g of carbon, or 19%
by weight. Such a loading does not appear impossible but is almost twice
the equilibrium loading (115 mg/g) predicted by the average isotherm
(Figure 8) at a concentration of 210 mg/1 of TOG in the feed to the B Series
of runs.
ST REGIS ACTIVATED CARBON FOR TREATING CAUSTIC BLEACH EFFLUENT
The suitability of activated carbon made by St. Regis from black liquor
char was evaluated in several small-scale experiments. The St. Regis carbon
was prepared in August, 1975 (Run 2 Condition 3) in another EPA-St. Regis
project (Grant No. 12040 EJU) to demonstrate the St. Regis hydropyrolysis
process for making char and activated carbon from pulping black liquor.
Isotherm Tests
Isotherm equilibrium adsorption tests were made with the St. Regis carbon
pulverized to -0.044 mm (-325 mesh) and with ICI powdered Darco S-51 carbon
using 2 to 4 g of each carbon in five one-liter quantities of CBE maintained
at pH 2.5 and 40°C for 15 min. The results of these isotherms showed that
the St. Regis carbon loaded to 800 CU/g at 3000 CU and the Darco carbon
38
-------
loaded to 1200 CU/g at this concentration. In a previous isotherm test
under these conditions, the Darco carbon loaded to 950 CU/g at 3000 CU.
These isotherm tests indicated that the St. Regis carbon had an equilibrium
adsorption capacity slightly below that of the Darco carbon for removing
color from CBE at pH 2.5. For TOC removal, the St. Regis carbon gave
higher loadings than the Darco carbon at less than 60 mg/1 and lower load-
ings at higher concentrations.
Dynamic Column Tests
Dynamic column adsorption runs were made in side-by-side small
(1.3 cm i.d.) columns, using CBE adjusted to pH 2.5 as the feed at a rate
of 5 BV/hr. The two columns were charged to a depth of about 30 cm with
18.9 g (dried weight) of each carbon screened to 0.84 x 0.42 mm (20 x 40
mesh). When the columns became loaded, they were regenerated with 1 BV of
4% NaOH and rinsed with 2 BV of 2% H2S04- Three loading runs were made with
each carbon.
The results of these column tests showed that the St. Regis carbon, at
the time of breakthrough of color, had removed only 15 to 20% as much color
and TOC as did the Darco carbon. These comparative isotherm and dynamic
column tests indicated that the St. Regis carbon had almost as great ad-
sorptive capacity as the Darco carbon but that the rate of penetration of
the color bodies into the pores of the carbon was much lower in the case of
the St. Regis carbon.
Dynamic Stirred Adsorption Tests
A third comparison was made between these carbons. Five grams of each
carbon screened to 0.84 x 0.42 mm (20 x 40 mesh) were added to one-liter
quantities of CBE maintained at pH 2.5 and stirred for 21 hr. Samples were
analyzed for color and TOC after various lengths of stirring time.
The results showed that the Darco carbon reduced the concentrations of
color and TOC at substantially greater rates during the first few hours and
then the rate of removal was greater for the St. Regis carbon. After 21 hr,
both carbons had reduced the color by 99.9% and the TOC by 91%. The results
from all of these comparison tests indicate that the St. Regis carbon had a
finer mean pore size that inhibited the movement and adsorption of the
organic compounds present in CBE at pH 2.5 but the finer pore sizes did not
prevent the St. Regis carbon from attaining about the same equilibrium load-
ing as the Darco carbon.
It has been demonstrated in other work that the mean pore size and pore
size distribution in the St. Regis activated carbon can be altered through
the selection of activating conditions. New batches of carbon were produced
to provide a larger mean pore size but these carbons were not available in
time to include the results in this report.
39
-------
REVERSION OF COLOR OF TREATED CAUSTIC BLEACH EFFLUENT
In a full-scale application of carbon adsorption treatment for CBE, the
product water would be sent to the sewer to be blended with acid bleach
effluent and then end up in the final treatment of the total mill effluent.
The color of pulping effluents from other types of decolorization pro-
cesses, such as lime treatment, have been found to increase with time, es-
pecially during the 5 to 10 days of biological oxidation in aerated lagoons.
The stability of color of the decolorized CBE from the pilot plant was
determined by holding two 30-liter samples of product water from loading
Run 1, Series C after 200 BV had been treated. Both samples were stored
in covered polyethylene containers in sunlight for 21 days and were sampled
at 10 intervals of time. One sample was not aerated but the dissolved
oxygen remained in the range of 4 to 6 mg/1. The other sample was bio-
logically oxidized by adding the normal levels of nitrogen and phosphorous
nutrients, biological seed from the mill secondary treatment system, and
aerated to give a dissolved oxygen concentration of 6.4 to 8.0 mg/1.
The color of the bio-oxidized sample increased from an initial value of
83 CU to 207 CU after 7 days and then decreased to 120 CU after 21 days.
The color of the other sample also increased, but to a maximum of 120 CU
after 1 day and then decreased to a constant value of about 50 CU from 5
days to 21 days.
This color stability experiment indicated that the carbon-treated CBE
can increase by as much as 124 CU during bio-oxidation but that the net
increase after the usual retention time in secondary treatment will pro-
bably be an acceptable amount. With a lesser amount of aeration and without
the addition of nutrients, the carbon treated CBE could be expected to de-
crease in color by 50% after a few days of retention. The magnitude of
these indicated increases of color during bio-oxidation or holding are pro-
bably not great enough to be a major concern in applying carbon adsorption
to the treatment of CBE. However, it does indicate that color increases
might be a problem if the retention is less than about 5 days and should be
investigated further.
REVERSION OF COLOR OF ACID BLEACH EFFLUENT
It is expected that the carbon-treated CBE from a full-scale installa-
tion would be treated to an average color of 200 to 300 CU and blended with
acid bleach effluent to form the total effluent from the bleach plant.
Alternatively, the treated CBE would be reused as dilution or wash water in
the chlorine or hypochlorite stages of bleaching and then exit from the plant
in the acid bleach effluent. Numerous measurements of the color of the acid
bleach effluent from the Pensacola mill showed that it normally had a color
of about 100 CU and therefore would contribute little to the color loading
of the bleach plant effluent. Because of this low color of the acid bleach
effluent, it was ass'umed at the start of this project that the total bleach
plant effluent color would be under 200 CU if the CBE were treated to 200 CU.
Since the volume of the acid bleach effluent is about three times that of the
CBE, the exclusion of the acid bleach effluent from treatment was important
40
-------
to minimizing the costs of reducing the color of the bleach plant. Samples
of acid bleach effluent obtained during the latter half of this project in-
dicated the color was not always below about 100 CU but was as great as
300 to 500 CU on occasions. Therefore a more extensive program of sampling
and investigation was started to determine the color variation and stability
of the acid bleach effluent.
It was found that the acid bleach effluent as it leaves the chlorination
stage is normally low in color but that it is quite unstable and increases
rapidly in color within 1 to 3 hr. During one study, samples of the acid
bleach effluent were obtained for each of six successive days and held in
glass bottles. The color was measured immediately and then hourly for 7 hr
and at the end of 24 hr. Five of the six samples increased from an average
color of 57 CU to 350 CU in 7 hr, which is an increase of six fold. Almost
all of this color increase occurred during the first 1 to 3 hr. The sixth
sample had an initial color of 503 CU and a color of 345 CU after 7 hr.
A second experiment with samples obtained during a 2-week period showed
the same degree of color increase. An earlier sampling study, in which
twelve samples of acid bleach effluent from bleaching of both softwood and
hardwood pulp were analyzed, showed that within several hours the acid
bleach samples from softwood bleaching had increased to an average of 863
CU and those from hardwood had increased to 283 CU.
It appears evident that acid bleach effluent from the Pensacola bleach
plant is quite unstable in color, that the color of fresh acid bleach effluent
is less than 100 CU except when this effluent has a noticeable lack of
chlorine odor, and that the color will increase within a few hours to about
350 CU. This finding of color reversion of acid bleach effluent indicates
that plans for applying any process for decolorizing bleach effluents must
consider the possible reversion of the acid bleach effluent color if it is
not to be treated. When acid bleach effluent at 350 CU is blended 3:1 with
CBE that has been treated to 200 CU, the resulting color of the total bleach
effluent will be about 312 CU, which leaves a fairly large amount of color in
the mill's effluent.
USE OF STRONG CAUSTIC BLEACH EFFLUENT
At the Pensacola St. Regis bleach plant, the caustic extraction stage
effluent is rather dilute because the water from the subsequent chlorine di-
oxide and second extraction stages is cascaded through the first extraction
stage. Typically, the color of the CBE is 3800 CU and the volume is 21 mj/
metric ton (5000 gal/short ton). In some bleach plants, the extraction
liquor is not diluted by water from the subsequent stages, which results in
a color concentration of about 15,000 CU and the volume per ton of pulp is
about 25% as great. The treatment of this strong caustic bleach effluent by
carbon adsorption would be expected to be much lower in cost because of the
small volume and because the greater concentration should result in much
greater loadings of color per gram of carbon. Therefore, isotherms and
column adsorption tests were carried out to determine the loadings of color
and TOC on the carbon when treating strong CBE.
41
-------
Samples of the strong CBE were obtained from the Pensacola bleach plant
as the pulp and liquor came from the first caustic extraction tower. The
strong CBE had a color of 18,000 CU and TOG concentration of 1630 mg/1.
Isotherm tests were made with powdered Darco S-51 carbon and with the strong
CBE held at pH 2.5. The isotherms were found to be extensions of those ob-
tained with more dilute CBE used previously (Figures 7 and 8). The equili-
brium loading was 1100 CU/g at 3500 CU and 1400 CU/g at 18,000 CU. The
equilibrium loading of TOC was 250 mg/g at 1630 mg/1 as compared to 120 mg/g
at the normal concentration of CBE. These data indicate that the maximum
loadings using strong CBE rather than normal CBE would be 27% greater for
color and 200% greater for TOC.
The removals of color and TOC were then checked under dynamic conditions
in a 1.3 cm dia. glass column filled with 21 g of Darco 20 x 40 carbon to a
depth of 40 cm. The strong CBE used for feed contained 20,600 CU and 1427
mg/1 TOC and was adjusted to a pH of 2.5. In the first loading run at 3
BV/hr, the removal of color was 97% at 9 BV and the loading was only 448
CU/g. The carbon in the column was regenerated with 1 BV of 4% caustic and
acidified with 2 BV of 2% I^SO^ In the second run at a lower feed rate
of 1.2 BV/hr, the BV of water treated and loading of color were about 12%
less than those for the first run. After three additional runs at 1.7
BV/hr, the BV/run had decreased to 6.6 at 92% removal of color, the loading
of color was 314 CU/g, and the loading of TOC was 19 mg/1. Two additional
runs were made after the column was regenerated with hot caustic, as dis-
cussed later. The use of hot caustic did not materially improve the adsorp-
tion performance of these two runs.
For the five-cycle series, the average volume of water treated per cycle
was 7.5 BV. The color removal was 95% to a color of 1080 CU and the TOC
removal was 92% to 109 mg/1. The average loadings on the carbon were 370
CU/g and 25 mg/g of TOC. These results indicate that the use of strong CBE
would reduce the volume of carbon in the adsorption columns for a full-scale
unit by a factor of 3 to 4 but the costs for thermal regeneration would be
greater and the color in the total bleach effluent would be greater than
when treating the normal low-concentration CBE.
EFFECT OF INCREASED TEMPERATURE FOR CAUSTIC REGENERATION
The effect of using a hot caustic solution in the regeneration sequence
was evaluated in two small-scale experiments. All pilot plant regenerations
were made within a fairly narrow temperature range of 20 to 26°C. It was
expected that regeneration with caustic at a higher temperature would in-
crease the rate of regeneration and increase the adsorption capacity in the
following loading run. If effective, the use of heated caustic for re-
generation could be used in a full-scale unit at very little extra cost.
In one experiment, 20 g quantities of spent (loaded) carbon from pilot
plant Run 9, Series B, were regenerated under three conditions in stirred
beakers:
42
-------
A. 100 ml of 2% NaOH at 25°C for 1 hr
B. 200 ml of 2% NaOH at 25°C for 1 hr
C. 100 ml of 2% NaOH at 80°C for 1 hr
Each quantity of caustic-treated carbon was washed with tap water, with
2% H2SO^, and again with tap water. A 10 g (dry weight) portion of each re-
generated carbon was then added to 1 liter of CBE that had been adjusted to
pH 2.5 and stirred. The color of the water was measured and recorded until
there was no further decrease in color (165 min maximum). The rate and ex-
tent of removal of color from the CBE was then used as a measure of the
effectiveness of the regeneration condition that had been used for each
sample. The adsorption test showed that the use of Condition C (100 ml of
NaOH and 80°C) increased the rate of removal of the first 30% of the color
by a factor of 4 over that of Condition A and increased the loading of color
on the carbon by a factor of 2. However, the use of Condition B (200 ml
NaOH and 25°C) gave a 5 fold increase in rate and a 2.6 fold increase of
color loading. Therefore, this experiment indicated that an increase of
both temperature and quantity of the caustic regeneration solution would
greatly increase the color adsorption properties of the regenerated carbon.
As discussed earlier, the results from tests in the pilot plant indicated
that there was no benefit from increasing the volume or concentration of the
caustic regeneration solution.
Another experiment on the effect of hot caustic regeneration was made in
conjunction with a small-column study of treating strong CBE. In that study,
the carbon bed was regenerated three times by use of 1 BV of 4% NaOH at
23°C followed by rinse water and 2 BV of 2% H2SO^ for acidification. The
fourth regeneration was the same except the caustic solution and the column
were held at 93°C. The length of the next run (Run 5) to a common cumulative
color removal of 94% was about 10 to 20% greater than expected on the basis
of the loading runs using caustic regeneration at room temperature. The
carbon was then regenerated by removing it from the column and stirring it
with 2 BV of 4% NaOH at 93°C and rinsing and acidifying as before. The
length of the subsequent loading run (Run 6) to 94% cumulative color removal
was only half that of Run 5.
This experiment on the effect of hot caustic regeneration indicated that
hot caustic regeneration was slightly beneficial. There appears to be suf-
ficient gain from hot caustic regeneration to warrant its use in a full-
scale installation.
TREATMENT OF CAUSTIC BLEACH EFFLUENT FROM BLEACHING OF HAREWOOD PULP
Since about 34% of the pulp bleached in the Pensacola mill is from hard-
wood pulping, a series of adsorption-regeneration cycles was carried out to
determine adsorption performance with hardwood CBE as feed. A small glass-
ware column 1.3 cm i.d. was used that contained 18.9 g of fresh Darco 20x40
carbon, which gave a bed height of 32.5 cm. The feed was adjusted to a pH
of 2.4. The flow rate was 11.3 BV/hr for the first two loading runs and 4.7
BV/hr for the next three runs. The regeneration procedure was the same as
for prior tests in this column using softwood CBE (1 BV of 4% NaOH and 2 BV
43
-------
of 2% H2SO,) . The hardwood CBE had a color of 790 CU and TOC concentration
of 128 mg/I.
The average product color from five cycles was 231 CU (71% removal), the
average TOC of the product was 42 mg/1 (67% removal), and the average length
of the loading runs was 158 BV. The length of successive runs decreased in
about the same manner as the runs with softwood CBE. The average loading
of color on the carbon per run was 168 CU/g, which was only 26% of the
loading obtained in equilibrium isotherm tests at the feed concentration.
The runs using softwood CBE (2962 CU) averaged only 103 BV/run as com-
pared to 158 BV/run with hardwood CBE, but the average loading of the soft-
wood CBE runs was 568 CU/g, as compared to 168 CU/g for the hardwood CBE
runs.
This comparison of operation with hardwood and softwood CBE indicates
that hardwood CBE also can be decolorized readily under the same operating
conditions and that caustic regeneration is equally effective. The only
effect of operating with periodic feeds of hardwood CBE will be to increase
the length of the loading cycle because of the lower feed color and to de-
crease the frequency of caustic and thermal regeneration of the carbon.
PRELIMINARY DESIGN FOR FULL-SCALE INSTALLATION
A preliminary design was prepared for carbon adsorption treatment of CBE
from the Pensacola bleach plant to determine the size and complexity of the
required installation and to determine the capital and operating costs. To
be conservative, this treatment plant was designed primarily on the basis
of average operating results from the pilot plant runs (Table 1)- The
design conditions, the properties of the CBE before and after treatment,
and the properties of the resulting total bleach effluent are as follows:
Bleach plant production - 227 metric tons/day (250 short tons/day)
CBE from softwood bleaching
flow - 4731 m3/day (1.25 mgd)
color - 3160 CU
TOC -233 mg/1
pH - adjusted from.10 to 2.3 with IL^SO^
temp. - 23°C
Product water
color-200 CU (94% removal)
TOC - 34 mg/1 (85% removal)
PH - 2.6
Flow velocity - 0.163 nrYnrmin(4 gpm/ftz) or 3 BV/hr
Cycles per thermal regeneration - 11
Loss of carbon per thermal regeneration - 7%
Bed volumes per cycle - 48
Loading time - 16 hr, caustic regeneration - 4 hr, off-time (excess
capacity) - 4 hr
Price of carbon - $0.95/kg ($0.43/lb)
Bulk density of carbon - 368 kg/m3 (23 lb/ftj)
44
-------
Caustic regeneration
1 BV 2% NaOH + 2 BV recycled rinses + 1 BV 2% H2S04 over 4 hr
NaOH added per cycle - 17.1 kg/m3 (1.07 lb/ft3) based on adsorber
bed volume
Final bleach plant effluent
treated CBE blended 1:3 with acid bleach (at 400 CU) to give total
bleach effluent of 350 CU, 143 mg/i XOC, pH 2.1
Overall % reductions for total bleach effluent
72% of color, 26% of TOG
The major independent design criteria for this plant are (1) the color of
the feed and product water, (2) the loading of color on the carbon, (3) the
flow velocity, and (4) the number of cycles that can be made before the
carbon must be thermally regenerated. The conditions for caustic regenera-
tion were those found best from the pilot plant runs. The plant equipment
was sized to handle 20% greater flow and amount of color than the average
design values given above.
The features of the plant are summarized below:
Adsorption columns
. 5 columns, 3 active, 2 being thermally regenerated, each 3.66 m>
(12 ft) in diameter
. bed height - 3.14 m (10.3 ft)
. bed volume - 32.9 m3 (1161 ft3)
. parallel feed (single stage) to the 3 active columns
up-flow feed, glass-reinforced polyester columns
provided with associated piping, cycle controls, etc.
. carbon in system - 61,000 kg (134,000 Ib)
Caustic regeneration
. two 37.9 m (10,000 gal.) tanks for strong (caustic) eluate
and weak eluate
rinse - 2% acid from in-line mixer and pump
backwash pump - 4.3m /min (1130 gpm) using water from backwash
storage tank
Thermal regeneration furnace
. operating 3 shifts at 3307 kg/day (7275 Ib/day) with 20% excess
capacity
. 7 hearths, total area 16.3 m2 (175 ft2)
associated controls, dewatering equipment, storage tanks
Surge feed tank for 12 hr of feed,165 m3 (625,000 gal.)
Multimedia filters for feed water (3), each 3.2 m (10.5 ft) diameter
Backwash tanks (2), each 132 nr (35,000 gal.) and pump for 7 m3/min
(1740 gpm) 9
Building, 520 m (5600 ft )
Single-stage adsorption columns made of glass reinforced polyester with
up-flow of feed were chosen for minimum capital cost. All equipment in
contact with low-pH solutions is designed for that service. The feed is
assumed to be caustic bleach effluent from the bleaching of softwood be-
cause of the predominance of softwood bleaching in kraft mills. It is
assumed in this design that the treated CBE at 200 CU will be mixed 1:3 with
the acid bleach effluent at 400 CU which would give a color of 350 CU in the
45
-------
total bleach effluent. In some bleach plants the treated CBE could probably
be used in place of fresh water or paper machine Whitewater now used for
shower water on washers or for regulation of stock consistency at other
stages of bleaching. If the treated CBE is sent to the sewer and the total
mill effluent cannot tolerate the low pH of the water, lime would be added
to neutralize this stream. The amount of lime needed would be 1.2 metric
tons per day to adjust the pH to 7.
In the caustic regeneration of the carbon, it was assumed that the caustic
solution is reused (as strong eluate) after the free NaOH content is in-
creased to 2% by addition of 50% NaOH. Since about half of the sodium value
of the added caustic is recovered, the net cost of the caustic is a minor
item in the total operating costs. The causticity of the strong eluate
could also be maintained at 2% by addition of white liquor (containing 10%
NaOH) from the mill's chemical recovery system. If weak wash (about 1%
NaOH) were used for regeneration, the eluate would not be reused in the
regeneration and would be too dilute to be disposed of in the black liquor.
For some mills the use of weak wash might be advantageous over the use of
caustic-augmented strong eluate.
It was assumed for this design that the strong eluate is sent to the
mill's weak black liquor when the total solids content increases to 6-10%
and that it contains 31% of the sodium added as NaOH or white liquor to the
strong eluate for regeneration. The average recovery of sodium in the pilot
plant was 31%, but it is expected to be greater in a full-scale plant.
COST ESTIMATES FOR TREATING 4730 m3/DAY (1.25 mgd) OF CAUSTIC BLEACH
EFFLUENT BY CARBON ADSORPTION
Preliminary, or order-of-magnitude, cost estimates were prepared for the
plant described in the previous subsection. The costs of the carbon ad-
sorption and thermal regeneration sections of the plant were provided by
ICI United States Inc. for the conditions of this design. The costs of
multimedia filters and all tanks were based on estimates from a major tank
fabricator. Costs for other equipment, for engineering, installation, con-
tractor's profit, and 12% contingency were from St. Regis experience. The
costs are based on October, 1975 costs (Engineering News Record Construction
Cost Index = 2294).
The estimated capital costs for the plant based on the basic conditions
listed previously are give in Table 4. The total estimated cost is
$2,380,000 which is equivalent to $10,470 per daily metric ton of bleached
pulp production ($9,520/short ton). The major costs are for adsorption
columns, thermal regeneration, and feedwater filtration and storage.
Operating costs for the basic design are given in Table 5. The major
cost items are amortization (22% of total), caustic, acid and carbon (38%),
and labor (9%). The'total cost including amortization is $8.16/metric ton
of bleached pulp ($7.41/short ton} This cost is rather high compared to
those published by vendors of resin adsorption processes (4,5,6) and from
estimates for ultrafiltration (10,11) for similar removals of color from
bleaching effluents.
46
-------
Table 4, CAPITAL COSTS3 OF PLANT TREATING 4730 m3/DAY (1.25 mgd)
OF CAUSTIC BLEACH EFFLUENT BY CARBON ADSORPTIONb
Item Total cost % of total
1. Adsorption columns
2. Thermal regeneration
3. Caustic regeneration
4. Feedwater filters including
backwash facilities
5. Surge feed tank
6. Building and electrical costs
7. Carbon in system
Total cost of investment
$934,000
500,000
82,000
240,000
354,000
212,000
58,000
2,380,000
40
21
3
10
15
9
2
100
Includes installation, engineering, aid contingencies.
Bleached pulp production = 227 metric tons (250 short tons) per day.
The capital costs would be reduced substantially if the carbon were dis-
carded instead of being regenerated thermally. However, the resulting cost
of make-up carbon would make this mode of operation unacceptably costly.
4. Rock, S. L., Bruner, A., Kennedy, D. C., "Decolorization of Kraft Mill
Effluents with Polymeric Adsorbents" Tappi 5_7_ (9) 87 (1974) .
5. Anderson, L. G., et al^ "A New Color Removal Process: A Field Report"
Tappi 57. (4) 102 (1974) .
6. Chamberlin, T. A., et_ aJL "Color Removal from Bleached Kraft Effluents"
In: Proceedings 1975 TAPPI Environmental Conference, TAPPI Atlanta,
1975. pp 35-45.
10. Fremont, H. A., Tate, D. C., Goldsmith, R. L., "Color Removal from Kraft
Mill Effluents by Ultrafiltration" EPA-660/2-73-019, December 1973.
11. Anon. "Ultrafiltration Looks Feasible for Kraft Effluent Color Removal"
Paper Trade J. 158 (17) 26 (April 29, 1974).
47
-------
Table 5. OPERATING COSTS FOR PLANT TREATING 4730 m3/DAY
(1.25 mgd) OF CAUSTIC BLEACH EFFLEUNT BY CARBON
ADSORPTION
Item
Caustic, 1538 kg/day NaOH @ $0.154/kg
(3388 Ib @ $0.07/lb
Acid, 5200 kg/ day 93% H SO @ $0.046/kg
(11,415 Ib @ $0.021/lb)
Carbon makeup, 231 kg/day @ $0.95
(509 Ib @ $0.43/lb)
Utilities
Maintenance and materials @ 3% of TCI/yrb
Labor, 24 hr/day @ $7.00/hr
Plant overhead, @ 75% of labor
Ins. and taxes @ 2% TCI/yr
Total less amortization
Amortization @ 6.25% of TCI/yr
Total operating cost per day
Total per metric ton of pulp
$/day
237
240
219
119
196
168
126
130
1445
408
1853
8.16 (7
% of total
13
13
12
7
11
9
7
7
78
22
100
. 41/ short ton]
3
Total per m of caustic bleach 5.62 (1.48/1000 gal.)
Bleached pulp production = 227 metric tons (250 short tons) per day.
Total cost of investment (TCI) = $2,380,000
Other conditions of operation and possible process improvements can
materially affect the capital and operating costs for this basic design.
Estimated costs for the plant using 13 modifications were prepared by adjust-
ments of the costs for the basic design. The capital and total operating
costs for each of these modifications are given in Table 6 along with the
change in costs compared to the basic design conditions.
If all of the bleach plant effluent were treated, the volume would be
increased four-fold, no acid would be required for pH adjustment, and most
operating and amortization costs would be increased about two-fold (Condi-
tion 2, Table 6). The color of the total bleach effluent would be improved
(200 CU vs 350 CU for the basic conditions). However, the very high costs
would make this mode of operation very unattractive.
48
-------
TABLE 6. EFFECT OF CHANGES IN OPERATING CONDITIONS AND EQUIPMENT ON ESTIMATED CAPITAL AND
OPERATING COSTS FOR PLANT TREATING CAUSTIC BLEACH EFFLUENT FROM BLEACHING OF 227
METRIC PULP TONS PER DAY
Capital cost, $106
Condition
No.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Change over basic design
Basic design, no change
Treat total bleach effluent
Treat concentrated CBE
Credit for 31* of Na
recovered
Credit for 60Z of Na
recovered
CBE from bleaching 34Z
hardwood
Don't need preflltration
Use acid bleach effluent
for rinse
CBE treated at 60°C
Operate with half as much
Total
2.38
4.94
2.56
2.38
2.38
1.93
2.14
2.51
2.15
2.38
Change from
basic design
0
+2.56
+0.18
0
0
-0.45
-0.24
+0.13
-0.23
0
Total
cost,
Total
8.15
14.52
9.75
7.82
7.52
6.93
7.51
7.82
7.83
7.50
operating
$/metric ton
Change
0
+6.37
+1.60
-0.33
-0.63
-1.22
-0.64
-0.33
-0.32
-0.65
labor
11. Obtain doubled no. of 2.13 -0.25 6.84 -1.31
cycles per series (per
thermal regen)
12. Loss of carbon per regen- 2.38 0 7.88 -0.27
eratlon is 5Z ys 7%
13. Feed surge tank is not 2.03 -0.35 7.76 -0.39
included
14. Neutralize acidic treated 2.43 +0.05 8.44 +0.29
CBE
The costs for treatment of CBE at a high color concentration (20,000 CU)
and low volume (Condition 3) were based on the results from the laboratory
column tests. The capital costs for the adsorption and regeneration sections
of the plant were greater than for the basic design because the amount of
color to be removed and the loading of color on the carbon were essentially
unchanged and a larger portion of the total carbon in the system was in a
second, inactive column. Therefore, there seems to be no cost advantage of
treating a more concentrated CBE.
49
-------
The treated CBE could probably be used as dilution or shower water in
other stages of the bleaching process. If so, credits might be given the
decolorization process for savings in fresh water, for savings in energy,
and for reducing the contaminant loading to the mill secondary treatment
system. Credits for these would be difficult to estimate and were not in-
cluded in these cost estimates.
The treated water will reduce the pH of the bleach plant effluent. If
the mill effluent is adversely affected by this added load of low pH water,
lime addition could be used to neutralize the acid in the water. The cost
for lime addition was estimated to add $0.29/pulp ton to the operating costs
(Condition 14).
The costs for the basic design did not include a credit for the sodium re-
covered in the strong eluate. At 31% recovery and credit for sodium (Con-
dition 4), the operating cost would be reduced by $0.33/ton. At 60% recov-
ery (which should be possible), the operating cost would be reduced by $0.63/
ton.
The inclusion of CBE from bleaching of hardwood in the feed to the carbon
adsorption plant reduces the color concentration of the feed and materially
reduces the costs of operation, based on the same throughput of 4731 m /day
(1.25 mgd) from bleaching of 227 metric tons/day (250 tons/day) of which
34% is hardwood. As shown by Condition 6 of Table 6, the inclusion of hard-
wood CBE reduced the capital cost to $1,930,000 and operating costs to $6.93/
pulp ton.
Prefiltration of the feed CBE will probably not be needed for the up-
flow columns. If so, the capital cost is reduced by $240,000 (Condition 8).
Acid bleach effluent can be used to acidify the carbon bed after caustic
regeneration which eliminates the cost of acid for this purpose but requires
a large volume of acid bleach effluent and larger equipment. The use of acid
bleach for rinse provided a slight reduction of capital and operating costs.
CBE comes from the bleach plant at about 60°C as contrasted to 23°C used
in the pilot plant and for the basic design conditions. The experimental
program indicated that the use of the higher temperature might increase ad-
sorption rates as much as 25%. If so, the capital costs would be reduced by
10% and the operating costs by 47% (Condition 9).
It appears plausible that a plant could be operated under the combined
improvements of Conditions 4,6,7,8 and 9. If so, the capital costs would
be reduced to $1,590,000 and the operating costs to $5.31/metric pulp ton.
More optimistically, it might be safe to assume that 60% of the sodium is
recovered (Condition 5), that the amount of labor needed is half as great
(Condition 10), that the number of cycles can be doubled before thermal
regeneration is needed (Condition 11), that the loss of carbon will be 5%
rather than 7% per thermal regeneration (Condition 12), and that a feed surge
tank is not needed (Condition 13).
50
-------
If all of these reductions in cost were possible, which is an optimistic
assumption, the combined effects of Conditions 5 through 13 would be to re-
duce the capital cost to $990,000 and operating costs to $2.39/metric pulp
ton ($2.17/short ton). In this minimum cost estimate, the capital cost is
60% and the total operating cost is 70% below those for the basic design.
With these optimistic assumptions, the costs for decolorizing the
caustic bleach effluent by carbon adsorption are generally comparable to
those reported for resin adsorption processes (4,5,6) and ultrafiltration
(11,12). Therefore carbon adsorption is one of the options to be considered
for treating bleach effluents from pulp mills.
51
-------
REFERENCES
1. Sanks, R. L., "Ion-Exchange Color and Mineral Removal from Kraft Bleach
Wastes". EPA Report R2-73-255 (May 1973).
2. Fuchs, R. E., "Decolorization of Pulp Mill Bleaching Effluents Using
Activated Carbon", National Council for Stream Improvement Technical
Bulletin No. 181 (May 1965).
3. McGlasson, W. G., Thibodeaux, L. J., and Berger, H. F., "Potential Uses
of Activated Carbon for Wastewater Renovation" Tappi 49_ (12) 521 (1966) .
4. Rock, S. L., Bruner, A., Kennedy, D. C., "Decolorization of Kraft Mill
Effluents with Polymeric Adsorbents" Tappi ,57^ (9) 87 (1974) .
5. Anderson, L. G., et_ al^ "A New Color Removal Process: A Field Report"
Tappi ,5_7 (4) 102 (1974).
6. Chamberlin, T. A., ^t al^ "Color Removal from Bleached Kraft Effluents"
In: Proceedings 1975 TAPPI Environmental Conference, TAPPI, Atlanta,
1975. pp 35-45.
7. Lang, E. W., Timpe, W. G. and Miller, R. L., "Activated Carbon Treatment
of Unbleached Kraft Effluent for Reuse" EPA-660/2-75-004, April 1975.
8. Timpe, W. G. and Lang, E. W., "Activated Carbon Treatment of Kraft Mill
Effluent for Reuse" Water-1973. AIChE Symposium Series 70 (136) 579
(1974).
9. Hutchins, R. A., "Design of Activated Carbon Systems" Chemical Engineer-
ing 80 (19) 133 (August 20, 1973).
10. Fremont, H. A., Tate, D. C., Goldsmith, R. L., "Color Removal from
Kraft Mill Effluents by Ultrafiltration" EPA-660/2-73-019, December 1973,
11. Anon. "Ultrafiltration Looks Feasible for Kraft Effluent Color Removal"
Paper Trade J. 158 (17) 26 (April 29, 1974).
12. "Standard Methods for Examination of Water and Wastewater", 13th Edition
APHA, New York (1971).
52
-------
APPENDIX
PROCEDURES FOR ISOTHERMS AND ANALYSES
PROCEDURE FOR MAKING CARBON ADSORPTION ISOTHERMS
1. Pulverize a representative sample of the carbon (10-20g) and
screen it through a 325-mesh sieve. Oven-dry the carbon for
3 hr at 150°C.
2. Weigh out 5 amounts of carbon in disposable weighing dishes.
For pulp mill effluents, use 0.3, 0.5, 1.0, 2.5, and 5.0 g per
500 ml of sample (0.6 to 10.0 g/1).
3. Add 500 ml of the water to be tested to a 1-liter beaker, heat
to 40°C, add one of the weighed dosages of carbon, stir vigorously
(100 rpm on a Phipps and Bird gang-stirrer) for 15 minutes while
maintaining the water at 40°C. Filter about 100 ml of the water
through Whatman No. 2 filter paper using a presure-type lab filter
(to prevent loss of volatile organics).
4. Repeat for the other dosages of carbon with separate 500-ml
quantities of water to be tested.
5. Analyze the filtrate from each dosage of carbon for the impurity
of interest and express in mg/1 or APHA color units. Normally,
isotherms were prepared for both TOG and color.
6. For each carbon dosage, subtract the final from the initial im-
purity concentration, and divide this difference by the dosage
of carbon in g/1. This number is the loading of impurity in mg/g
of carbon used. On 3x3 cycle log-log graph paper, plot loading
on the y-axis versus remaining concentration on the x-axis. The
isotherm is completed by drawing a straight line through the
plotted values. (Sometimes the line has two slopes and has a dog-
leg shape.) Extrapolate the line to the initial concentration.
The loading at this intercept is the ultimate capacity of the
carbon for that effluent.
ANALYTICAL METHODS
The methods used for analyses of samples in the work covered by this re-
port are listed below. The standard method number given refers to the method
in the 13th edition of APHA Standard Methods (12). Methods modified by
St. Regis Paper Company are available on request.
Color - Standard Method 206A modified by NCASI and adopted by the pulp and
paper industry. The pH of the sample is adjusted to 7.6 and the sample
filtered through a 0.8 micron Millipore filter and light transmittance is
measured at 465 nm on a Spectronic 70 spectrophotometer. Color is determined
from a calibration curve using a cobalt chloroplatinate standard, where
equivalent mg/1 of cobalt chloroplatinate is termed color units or "CU".
53
-------
TOG and TIC - Standard Method 138 using a Beckman 915 total carbon analyzer;
sample filtered through Whatman No. 2 paper filter.
jjH - Standard Method 144A
Turbidity - Standard Method 163A using a Hach Model 2100 turbidimeter.
Total Suspended Solids (TSS) - Standard Method 224C using Whatman GF/C glass
fiber filter discs.
Conductivity - Standard Method 154 using Yellow Springs Instrument Co. Model
31 conductivity bridge.
BOD-5 - Standard Method 219.
Dissolved Oxygen (DO) - Standard Method 218F using a Weston and Stack
dissolved oxygen analyzer Model 300.
Chloride - Standard Method 203C.
Total chlorine - by X-ray diffraction.
Organic chloride - by difference of total chlorine less chloride.
Metal ions - Standard Method 129A using atomic absorption with the Perkin-
Elmer Model 403 spectrophotometer.
54
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-77-1iq
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Activated Carbon Treatment of Kraft
Bleaching Effluents
5. REPORT DATE
June 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
E. W. Lang, J. W. Stephens, R. L. Miller
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
St. Regis Paper Company
Cantonment, Florida 32533
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
R-803270
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab-Cin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The removal of color and organic contaminants by adsorption on activated carbon
from the effluent of a kraft pulp bleaching plant was investigated in a pilot plant.
The caustic bleach effluent, which contains 80% of the color from pulp bleaching,
was decolorized successfully when it was adjusted to pH 2.5. The spent carbon
was regenerated with caustic solution for an average of 11 adsorption-regeneration
Cycles before thermal regeneration was required. Variables studied included pH
of feed, feed rate, effluent from bleaching of hardwood and softwood, caustic
requirements for regenerating the carbon, and concentration of color in feed.
Capital and operating cost estimates for a full-scale plant are presented. The
cost effects of variations in design and operating conditions are also discussed.
Conclusions are that the process is technically sound, that it will remove 94%
of the color and 84% of the total organic carbon from caustic bleach effluent from
the bleaching of softwood, but that it has slightly higher capital and operating
costs than alternative methods for reducing color in bleach effluents (resin
.adsorption, ultrafiltration, or bleach sequence modifications, for example).
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Industrial Waste Treatment*, Activated
Carbon Treatment*, Chemical Removal,
Biochemical Oxygen Demand, Industrial
Wastes, Pulp Mills, Cost Estimates,
Pilot Plants
Pulp Bleach Mill Effluent
Color Removal, Tertiary
Treatment, Bleached Kraft
Wastewater Treatment,
Caustic Bleach Effluent,
Activated Carbon Regenera
Wastewater Treatment, Tot&l
13/B
tion,
Organic Carbon
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
65
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
55 &U.S. GOVERNMENT PRINTING OFFICE: 1977-757-056/6i.5
------- |